Combining ICT with Eco-driving Concepts to Improve Corporate Car

Transcription

Combining ICT with Eco-driving Concepts to
Improve Corporate Car Drivers’ Fuel Efficiency DISSERTATION
of the University of St. Gallen,
School of Management,
Economics, Law, Social Sciences
and International Affairs
to obtain the title of
Doctor of Philosophy in Management
submitted by
Johannes Tulusan
from
Germany
Approved on the application of
Prof. Dr. Elgar Fleisch
and
Prof. Dr. Rolf Wüstenhagen
Dissertation no. 4246
Difo-Druck GmbH, Bamberg, 2013
The University of St. Gallen, School of Management, Economics, Law, Social Sciences and International Affairs hereby consents to the printing of the present dissertation, without hereby expressing any opinion on the views herein expressed.
St. Gallen, October 21, 2013
The President:
Prof. Dr. Thomas Bieger
Acknowledgements
I
Acknowledgements
The research for this dissertation was jointly conducted with the Bits to Energy Lab
(University of St. Gallen and ETH Zurich) and SAP Research from January 2010 to
January 2013. The cross-institutional approach helped to enhance my research from
both a theoretical and practical perspective through the application of theoretical
concepts to a true global problem. Colleagues from both institutions provided valuable feedback and guidance, which enabled me to complete my work.
I would like to thank my supervisor, Prof. Dr. Elgar Fleisch, for his ongoing support. He helped create a rewarding work environment that nurtured this form of research. My sincere gratitude to Prof. Dr. Thorsten Staake, the Bits to Energy lab
director, who enabled me to narrow down my research area, thoroughly discussed
the experiments, and who provided regular feedback in a resourceful manner. And
to Prof. Dr. Rolf Wüstenhagen, who already offered guidance in the first year of my
doctoral studies.
From the industry I would like to thank my manager, Dr. Uli Eisert, the CFO of
SAP Switzerland AG, Thomas Scheer, and the sustainability manager, Jonas
Dennler; without their support and facilitation, research with the case study company, including access to the 340 corporate car drivers’ data sets would not have been
possible.
I had fruitful discussions with, and received immense support from my colleagues at
the Bits to Energy Lab and SAP Research. My sincere thanks to: Dr. Claire
Michelle-Look, Dr. Tobias Graml, Michael Baeriswyl (from the Bits to Energy
Lab), and Dr. Ali Dada, Dr. Marc Brogle, and Dr. Felix von Reischach (SAP Research). I hope these friendships will remain beyond my days as a doctoral student,
and that we continue to inspire one another.
Finally, to my parents and two siblings, which have been invaluable during this
challenging and timely journey. And to my fiancée, Barbara Kuzman, thank you for
your unconditional encouragement, understanding, and love.
Munich, May 2013
Johannes Tulusan
II
Abstract
Abstract
The steady increase in energy consumption, largely due to the road transport sector,
necessitates immediate action to curtail CO2 emissions. Regulations implemented by
the European Commission dictate that the automotive industry should manufacture
vehicles with reduced emissions output. In addition to these governmental regulations, individuals can address this problem by adopting a more eco-friendly driving
style, as up to 30% of the fuel consumption rate is determined by their driving practices.
This thesis aims to outline a way to influence corporate car drivers to improve their
fuel efficiency and reduce CO2 emissions by using eco-feedback technologies. This
sample is unique because unlike private car drivers who respond well to financial
incentives for adopting sustainable driving behavior, corporate drivers are not financially motivated, as their fuel costs are reimbursed by their employer. Two feedback technologies that provided details about the fuel consumption of each driver,
an eco-driving smartphone application and a mobility information system, were applied in an experimental setting. These technologies were tested in two field tests
over a period of three months, involving a total of 340 corporate car drivers.
Analysis of the datasets, which included more than 20,000 fuel efficiency data
points, showed an average improvement in fuel efficiency of 3.23% in the experiment with the eco-driving smartphone application and 1.24% respectively 1.69% in
the second experiment with a mobility information system. Since no further managerial guidelines or financial incentives were given, these findings indicate that ecofeedback technologies can play an important role in improving the overall fuel efficiency of a large corporate fleet. In total, the case study company could save up to
1.4 million Euros on fuel costs and reduce emissions by up to four tons of CO2 per
year.
In 2013, the initiative will be rolled out to 11,000 corporate car drivers based in
Germany. Furthermore, the technology will be prototyped, implemented as a service
offered in fleet management software, and commercialized. This solution will be
attractive to other corporations with large or expanding car fleets that are eager to
reduce energy consumption and fuel costs by modifying the driving behavior of
their employees without making large financial investments.
Zusammenfassung
III
Zusammenfassung
Der anhaltende Anstieg des Energieverbrauchs durch den Verkehrssektor verlangt
nach sofortigen Maßnahmen, um den steigenden Ausstoß von CO2-Emissionen zu
senken. Regulierungen seitens der Europäischen Kommission verlangen von der
Automobilindustrie eine Produktion von Kraftfahrzeugen mit einem geringeren
CO2-Ausstoß. Neben diesen Regulierungen können Autofahrer dem Problem entgegenwirken, indem sie einen nachhaltigeren Fahrstil erlernen. 30% des Benzinverbrauchs kann vom Fahrverhalten jedes Einzelnen abhängen.
Das Forschungsvorhaben dieser Doktorarbeit ist es zu evaluieren, ob es möglich ist,
Firmenwagenfahrer durch Feedback Technologien zu beeinflussen, den Benzin- und
CO2-Verbrauch zu senken. Firmenwagenfahrer stellen eine besondere Gruppe dar,
da die Benzinkosten von der Firma bezahlt werden und daher kein tatsächlicher Anreiz für ein nachhaltigeres Fahren besteht. Bei Privatwagenfahrern ist die finanzielle
Einsparung der wichtigste Grund für eine nachhaltigere Fahrweise. In den Experimenten wurden zwei Feedback Technologien, die individuelle Angaben zum Benzinverbrauch für jeden einzelnen Fahrer anzeigten, verwendet. Diese Technologien
wurden in zwei Feldtests über eine Zeitspanne von drei Monaten und insgesamt 340
Firmenwagenfahrern getestet.
Die Analyse der Datensätze beinhaltet mehr als 20.000 Tankvorgänge. Eine Verbesserung des durchschnittlichen Benzinverbrauchs um 3,23% wurde in dem ersten
Experiment mit einer eco-driving Smartphone Applikation und 1,24% bzw. 1,69%
in dem zweiten Experiment durch ein Mobilitätsinformationssystem erzielt. Da es
keine Richtlinien seitens des Managements und finanzielle Anreize für die Teilnehmer gab, zeigen diese Ergebnisse, dass eco-feedback Technologien eine wichtige Rolle bei der Verbesserung des durchschnittlichen Benzinverbrauchs großer
Firmenwagenflotten spielen können. Insgesamt könnten bei der Fallstudie bis zu 1,4
Millionen Euro an Benzinkosten und bis zu vier Tonnen an CO2-Emissionen pro
Jahr eingespart werden.
In diesem Jahr (2013) wird die Initiative bei 11.000 Firmenwagenfahrern in
Deutschland erweitert. Zusätzlich wird ein Prototyp dieser Technologie entwickelt
und als ein weiteres Serviceangebot in der bestehenden Flottenmanagement Softwarelösung angeboten. Dieses Angebot wird vor allem für Firmen mit einem immer
größer werdenden Fuhrpark relevant sein. Insbesondere kann - ohne eine hohe finanzielle Investition zu tätigen - die Senkung des Energieverbrauchs und der Benzinkosten durch die Veränderung des Fahrverhaltens der Mitarbeiter erzielt werden.
IV
Table of Contents
Table of Contents
I
Introduction ...................................................................................................... 1
I.1
I.2
I.3
I.4
I.5
Motivation ................................................................................................. 1
Objectives and Research Questions ............................................................ 7
Research Methodology in relation to Research Questions .......................... 8
Scope of the Thesis ...................................................................................16
Structure of the Thesis...............................................................................22
II Eco-driving and Eco-Feedback Technologies Related Work ............................25
II.1
II.2
II.3
II.4
II.5
II.6
II.7
III
III.1
III.2
III.3
III.4
III.5
III.6
IV
IV.1
IV.2
IV.3
IV.4
IV.5
IV.6
IV.7
IV.8
V
V.1
V.2
V.3
V.4
V.5
Overview of Eco-driving ...........................................................................25
Differences between Private and Corporate Car Drivers ............................28
Eco-Feedback Technologies ......................................................................30
Smartphones Market Share........................................................................36
Impact of Eco-Feedback Technologies on the Driving Behavior ...............37
Feedback Intervention Theory ...................................................................39
Research Gap, ...........................................................................................41
Private and Corporate Car Drivers’ Preferences for Eco-feedback Types .....43
Overview and Research Question ..............................................................43
Research Design and Research Methodology ............................................43
Data Evaluation and Findings ....................................................................44
Discussion .................................................................................................56
Conclusion ................................................................................................58
Limitations and Future Research ...............................................................58
Direct Feedback by an Eco-driving Smartphone Application, .......................61
Overview ..................................................................................................61
State of the Art and Related Work .............................................................62
Research Questions and Hypothesis ..........................................................68
Research Design and Research Methodology ............................................70
Data Evaluation and Findings ....................................................................73
Discussion .................................................................................................84
Conclusion ................................................................................................86
Limitations and Future Research ...............................................................86
Indirect Feedback by a Mobility Information System ....................................89
Overview ..................................................................................................89
State of the Art and Related Work .............................................................90
Research Questions and Hypothesis ..........................................................98
Research Design and Research Methodology ..........................................101
Data Evaluation and Findings ..................................................................104
Table of Contents
V.6
V.7
V.8
VI
V
Discussion ............................................................................................... 106
Conclusion .............................................................................................. 107
Limitations and Future Research ............................................................. 108
Recommendations to Increase Corp. Car Drivers’ Intrinsic Motivations ....109
VI.1
VI.2
VI.3
VI.4
VI.5
VI.6
Overview and Research Questions ..........................................................109
Extrinsic and Intrinsic Motivation ........................................................... 109
Research Design and Research Methodology ..........................................110
Data Evaluation and Findings ..................................................................112
Conclusion .............................................................................................. 119
Limitations and Future Research ............................................................. 119
VII Development of an Eco-driving Dashboard System ....................................121
VII.1
VII.2
VII.3
VII.4
System Architecture, Data Structure, and Data Processing ......................121
Mock-ups ................................................................................................ 123
Mock-up incorporated in SAP´s Portal ....................................................125
Drawbacks of the current eco-driving System .........................................126
VIII
Conclusion, Recommendations, Outlook and Future Research ................129
VIII.1
VIII.2
VIII.3
VIII.4
VIII.5
VIII.6
Key Findings ........................................................................................... 129
Theoretical Contributions ........................................................................133
Practical Implications ..............................................................................135
Recommendations and High Level Concept ............................................137
Limitations .............................................................................................. 140
Future Research ......................................................................................141
Appendix .............................................................................................................143
References ...........................................................................................................171
Curriculum Vitae .................................................................................................181
VI
List of Figures
List of Figures
Figure 1: EU CO2 Emissions per sector (left) and transport mode (right) ................. 1
Figure 2: Research Process and Relation of Research Questions to Methods ......... 13
Figure 3: 2011 SAP AG’s global distribution of CO2 Emissions ............................ 17
Figure 4: Fuel Efficiency Calculation..................................................................... 19
Figure 5: Research Perspectives ............................................................................. 20
Figure 6: Structure of the Thesis ............................................................................ 22
Figure 7: Feedback Technologies and Response of Driver ..................................... 31
Figure 8: Ford’s SmartGauge in-car Interface14 ...................................................... 31
Figure 9: Eco-driving Applications Market Share .................................................. 32
Figure 10: Bliss Trek Application Interface ........................................................... 33
Figure 11: greenMeter Application Interface .......................................................... 34
Figure 12: Color changing Feedback ...................................................................... 34
Figure 13: Green Driving App ............................................................................... 35
Figure 14: Smartphone Market Share Outlook ....................................................... 37
Figure 15: Research Design ................................................................................... 44
Figure 16: Categorization Process of Open-ended Questions ................................. 45
Figure 17: Awareness of Feedback Devices ........................................................... 46
Figure 18: Changes while using a Feedback Device ............................................... 47
Figure 19: Preferences for Feedback Types ............................................................ 48
Figure 20: Preferences for Feedback Technologies ................................................ 50
Figure 21: Reasons for Favoring the Color-changing Feedback System ................. 51
Figure 22: Reasons for Not Favoring the PC Application Feedback System .......... 52
Figure 23: Suggestions for Additional Eco-efficient Feedback ............................... 53
Figure 24: Average Kilometers driven per Month .................................................. 54
Figure 25: Preference of Feedback Types to improve Fuel Efficiency .................... 55
Figure 26: Understanding of the Meaning of 100 grams of CO2 per km ................. 56
Figure 27: Preferences for Feedback Types, Private vs. Corporate Car Drivers ...... 57
Figure 28: Overview of DriveGain Application Interface ....................................... 65
Figure 29: Settings Screen ..................................................................................... 66
Figure 30: Select Vehicle Screen ........................................................................... 66
Figure 31: Advanced Savings Feedback ................................................................. 66
Figure 32: Fuel Savings Feedback ......................................................................... 66
Figure 33: Detailed Journey View shown in Online Driving Portal ........................ 68
Figure 34: Between-subjects Experimental Design ................................................ 71
List of Figures
VII
Figure 35: Influence on Driving Behavior .............................................................. 76
Figure 36: Direct – vs. Indirect Feedback............................................................... 80
Figure 37: Frequency of Journey Score reviewed in the ODP ................................ 81
Figure 38: Preference of when Feedback should be received ................................. 82
Figure 39: Preference of how Feedback should be provided................................... 83
Figure 40: Preference of Social Comparisons ......................................................... 84
Figure 41: Verbal Descriptor Scale (VDS) ............................................................. 94
Figure 42: Visual Analogue Scale (VAS) ............................................................... 95
Figure 43: Verbal Graphic Rating Scale (GRS) ...................................................... 95
Figure 44: Verbal and Numerical Graphic Rating Scale (GRS) .............................. 95
Figure 45: Normative Feedback to improve Energy Consumption ......................... 96
Figure 46: Energy Efficiency Rating Scale (EU directive 2002/91/EC) .................. 96
Figure 47: Categorical and Continual Feedback Scale Formats .............................. 97
Figure 48: Experimental Design ........................................................................... 101
Figure 49: Monthly Feedback Email with Discrete or Continual Feedback .......... 102
Figure 50: Interview Data Analysis ...................................................................... 111
Figure 51: Eco-driving Dashboard Architecture ................................................... 122
Figure 52: Driver´s Master Data Types ................................................................ 122
Figure 53: Eco-driving Dashboard Architecture ................................................... 123
Figure 54: Corporate Car Driver´s Eco-driving Dashboard .................................. 124
Figure 55: Management´s Corporate Fleet Dashboard.......................................... 125
Figure 56: Management´s Corporate Fleet Dashboard.......................................... 126
Figure 57: Audi navigation system (under development) ..................................... 144
Figure 58: Ford Gauge ......................................................................................... 145
Figure 59: greenMeter iPhone application ............................................................ 145
Figure 60: Gauge in cars dashboard ..................................................................... 147
Figure 61: Screen on top of the car dashboard ...................................................... 148
Figure 62: iPhone Application ............................................................................. 148
Figure 63: PC Application ................................................................................... 149
Figure 64: iPhone Application ............................................................................. 149
Figure 65: Audi navigation system (under development) ..................................... 150
Figure 66: Description in iTunes Store ................................................................. 154
Figure 67: Main Screen ........................................................................................ 154
Figure 68: Setting Screen ..................................................................................... 154
Figure 69: Add Vehicle ........................................................................................ 154
Figure 70: Enable Upload of Data ........................................................................ 155
VIII
List of Figures
Figure 71: Shift to N Gear.................................................................................... 155
Figure 72: Shift to 1st Gear ................................................................................... 155
Figure 73: Shift to 2nd Gear .................................................................................. 155
Figure 74: Shift to 3rd Gear .................................................................................. 156
Figure 75: Fuel Savings Meter ............................................................................. 162
Figure 76: Advanced Savings Meter .................................................................... 162
List of Tables
IX
List of Tables
Table 1: Mixed-method Research Methodologies .................................................... 9
Table 2: SAP Corporate Car Drivers’ Data Set Example ........................................ 19
Table 3: Advantages and Disadvantages of Eco-Driving ........................................ 26
Table 4: Differences between Private and Corporate Car Drivers ........................... 29
Table 5: Overview of Eco-driving Applications ..................................................... 35
Table 6: Five Arguments Feedback Intervention Theory ........................................ 39
Table 7: Private Car Driver’s Awareness of CO2 Emissions ................................... 49
Table 8: Sample Demographics ............................................................................. 74
Table 9: Group Statistics ........................................................................................ 78
Table 10: Independent Samples Test (t-test for Equality of Means)........................ 78
Table 11: Independent Samples Test (Levene’s Test for Equality of Variances) .... 78
Table 12: Correlation of Direct or Indirect Feedback vs. Duration ......................... 79
Table 13: Illustration of Hypothesis 1 .................................................................... 99
Table 14: Illustration of Hypothesis 2 .................................................................... 99
Table 15: Illustration of Hypothesis 3 .................................................................. 100
Table 16: Fuel Efficiency Rating and Intervals .................................................... 103
Table 17: Paired Samples T-Test TG1 ................................................................. 104
Table 20: Categories of Structured Analysis ........................................................ 112
Table 21: Summary of Feedback Intervention Theory Arguments ....................... 115
Table 22: Recommendations and High-Level Concept ......................................... 138
X
Abbreviations
Abbreviations

ANOVA
Cronbach's Alpha
AT
Avg_fe_y
Attitude
Average fuel efficiency total year
Avg_fe_trtmt_t
Average fuel efficiency treatment time
Baseline
Baseline
BI
BIC
B2E
Behavioral Intention
Behavioral Intention to Change
Bits to Energy Lab
CAN-bus
Controller Area Network Bus
CO2
CG
CPM
date_fil
Df
Carbon Dioxide
Control Group
Computer Performance Monitoring
Date of the filling
Degree of freedoms in a t-test
diff_fe_percentage
Percentile difference between fe_filling and fe_oem
DS
EMEA
ETH
Driving Style
Europe, the Middle East and Africa
Swiss Federal Institute of Technology
EU
EUR
FE
fem1
fem2
European Union
Euro
Fuel Efficiency
First Feedback EMail
Second Feedback EMail
Analysis of Variance
fem3
fe_filling
Third Feedback EMail
Fuel Efficiency in Liter per 100km (I_fil / km_drivn) x
100=fe
fe_filling_oem_difference Difference between fe_filling and fe_oem
fe_oem
Average Fuel Efficiency defined by car manufacturer
fe_type
FI
FIT
Fuel Type of the Filling
Feedback Intervention
Feedback Intervention Theory
FT
Feedback Technologies
F&S
Frost and Sullivan
Abbreviations
XI
Gart
gCO2/km
Gartner
Gram CO2 per kilometer
GPS
Global Positioning System
GWh
Gigawatt hours
HCI
Human Computer Interaction
HSG
ICT
University of St. Gallen
Information Communication Technology
Id
IS
IT
ID number of a driver
Information System
Information Technology
km_fil
kTons
Total amount of km at the time of the filling
Kilo Tons
l/100km
l_fil
Mil
Mtoe
Liters of fuel needed for 100km
Amount of liters filled up (in liters)
Million
Million Tons of Oil Equivalent
N
Total amount
n (%)
n_fillings
pers_max
pers_satis
Total amount in percent
Number of fillings per driver
Maximizer Personality
Satisficer Personality
P
PBC
PRQ
Q
Significance level
Perceived Behavioral Control
Principal Research Question
Quarter
RPM
RQ
Sig.
SQ
Std. Dev.
Revolution per Minute
Research Question
Significance
Sub Question
Standard Deviation
survey_answd
T
TG1
TG2
Survey answered
Time
Treatment Group 1
Treatment Group 2
TPB
Theory of Planned Behavior
TRA
Theory of Reasoned Action
XII
Abbreviations
Trtmt
Treatment
USD
US-Dollar
Introduction
1
I Introduction
I.1
Motivation
I.1.1 Practical Relevance
In 2010, energy consumption increased by 5.5% to 12.9 Mtoe (Million Tons of Oil
Equivalent) and reached the highest consumption rate to date (Enerdata, 2011). One
approach to address this alarming trend would be to reduce the consumption of fossil fuels in the transport sector (European Commission, 2010). In Europe, for example, 23% of carbon dioxide (CO2) emissions were attributable to the transport sector
(see Figure 1, left pie chart), and 71% of these resulted from road transport (see
Figure 1, right pie chart). In comparison, marine and aviation transport only contribute 15% and 12%, respectively.
Figure 1: EU CO2 Emissions per sector (left) and transport mode (right)1
Within road transport, much of the contribution comes from drivers who use their
car for business-related travel; this group is known as corporate car drivers. As
highlighted in the latest report by Dataforce (2011) (“Fuhrparkmanagement in deutschen Unternehmen 2011”), one quarter of the three million newly registered
1
European Commission. (2010), “Road transport: Reducing CO2 emissions from light-duty vehicles.” Retrieved from http://ec.europa.eu/clima/policies/transport/vehicles/index_en.htm
2
Introduction
cars in Germany are corporate cars, representing a rise from 700,000 in 2011 to
750,000 in 2012. In Switzerland the ratio is even higher: in 2011 one third of the
319,000 newly registered cars were corporate cars, a rise of 8.4% compared to 2010
(Bahnmüller, 2012; Ullmenstein, 2012). The number of corporate cars is predicted
to rise further in Europe in the coming years (Dataforce, 2011). Corporate car drivers in Europe drive 35,000 kilometers per year on average, which is 21,500 kilometers per year more than private car drivers. Their fuel costs are reimbursed by their
company (DTLR, 2001), so they lack the financial motivation that influences private car users to improve their fuel efficiency.
Government legislation is one efficient means of reducing CO2 emissions in the
personal transportation sector. In Europe, Environmental Zones help reduce traffic
congestion in city centers. Germany has forty-two Environmental Zones spread
across city centers; an obligatory environmental badge specific to the vehicle type,
known as a green badge, is required to enter these designated ‘green’ zones (Climate Company, 2011). In 2007, a new European regulation set limits on the average
CO2 emissions permitted from the yearly total sum of new cars produced by each
car manufacturer. By 2012, each car manufacturer should not exceed more than 120
grams per km (gCO2/ km) for their yearly total sum of produced cars. This is a reduction of 25% when compared to emission rates in 2006 (European Commission,
2010). A further reduction target for manufacturer’s average fleet emissions has been set at 70 gCO2/ km by 2025.
A report from the National Research Council (2010) appraised technologies that
could improve vehicles’ fuel economy in order to meet the European Commission’s CO2 fleet regulations. Factors affecting fuel consumption include vehicle weight,
engine efficiency, transmission type, automatic start / stop function, brake regeneration, tire type, air conditioning, and cruise control (National Research Council,
2010). Accommodating all of these features would require purchasing a new vehicle, a step many private car drivers may be unwilling or unable to undertake solely
for the benefit of reducing overall CO2 emissions (Gardner and Stern, 2008). In corporations such as SAP, eligible corporate-car-driving employees are allowed to purchase a new car every four years and drivers must purchase a car falling under a
maximum CO2 emissions threshold of 190 gCO2/ km. Drivers willing to buy a car
Introduction
3
with emissions below 150 gCO2/ km receive a financial incentive in the form of
reduced price for the car (e.g., ‘EcoBonus’ offers a 5% reduction in list price).
In addition to government and corporate legislations and improvements in vehicle
technology, drivers can influence their CO2 emissions rate without purchasing a new
car by adopting a more eco-friendly driving style. Driving style can influence fuel
consumption either positively or negatively; aggressive driving habits, for example,
can lead to a 30% increase in fuel consumption (Romm and Frank, 2006). One example of an intervention aimed at improving individuals’ driving styles is an ecodriving training program in which drivers learn techniques to minimize fuel consumption (Boriboonsomsin et al., 2010). Eco-driving is an economical approach to
reducing fuel consumption that can be utilized regardless of the vehicle type and is
typically taught through classroom instruction and a practical training session over
the course of a single day. The underlying ethos of eco-driving is to introduce positive attributes into an individual’s daily driving habits, thus encouraging an average reduction in fuel consumption of up to 15% (Onoda, 2009). Additional benefits
could include reductions in greenhouse gas emissions, petrol expenses, wear and
tear of the vehicle, and improved road safety resulting in fewer vehicle collisions
(GreenRoad, 2008; Onoda, 2009). However, some studies have uncovered a 5 10% deterioration in fuel efficiency one month after participating in eco-driving
training, as drivers resorted back to their old driving habits (Wahlberg, 2007). Consequently, considerable post-training measures are needed after eco-driving courses
to ensure that over the long term drivers can change actions that have become automatic and are performed without conscious thought.
In addition to classroom and onsite training, eco-driving advice can be provided
through brochures, driving simulators, websites, and in-vehicle eco-driving systems.
The results of a field test conducted with twenty drivers who received instantaneous
feedback from an in-vehicle eco-driving system indicated an improvement in fuel
efficiency of 1% on highways and 6% on city streets (Boriboonsomsin et al., 2010).
With the increasing popularity of smartphone technologies (as demonstrated, for
example, by an estimated increase in mobile-phone market share from 11% in 2008
to 40% in 2014 (Gartner, 2010)), one cost-effective intervention may be eco-driving
smartphone applications. These applications inform drivers about their driving style
either following eco-driving training or in the absence of such training, and those
4
Introduction
currently on the market cost less than five Euros per installation. The latest
smartphone processors and sensing technologies, such as GPS and accelerometer
sensors, are able to provide feedback according to the driving context.
Another way to provide feedback is by increasing drivers’ awareness of fuel consumption via email, to which every employee has access. This could be beneficial in
a corporate environment in which drivers are normally not informed about their average fuel consumption, fuel costs, or CO2 emissions unless they independently note
such details. Some corporations also have fleet management systems in place to
track their fleet’s contractual agreements (e.g., fuel costs, car maintenance cycles, fuel tank data, and accident reports) but do not share this information with their employees.
Evidence suggests that a reduction in CO2 emissions brings both ecologic and economic advantages. The German Federal Environmental Agency (2009) estimated
that an improvement in fuel efficiency of one liter per 100 km leads to total financial savings of 2,500 to 4,200 Euros (for yearly travel of 15,000 km or 25,000 km,
respectively) over the twelve-year lifespan of a car. This calculation assumed an
estimated petrol price of 1.40 Euro per liter; if fuel prices per liter increase in the
upcoming years, the savings would also increase. These savings are particularly
relevant for companies with large car fleets.
In summary, to reduce the CO2 emissions by lowering fossil fuel consumption in the
road transport sector, corporate car drivers should be targeted because their mileage
exceeds that of other groups. Even if vehicles with the latest fuel efficiency technologies are purchased, a major fuel consumption factor that remains unaddressed is
drivers’ driving style (Romm and Frank, 2006). To attain a better understanding of
how to motivate corporate car drivers to drive more sustainably with the support of
an eco-driving feedback information system (IS), it is important to consider the increasing size of this driver group in Europe.
Introduction
5
I.1.2 Theoretical Relevance2
From a theoretical perspective, studies exist that describe social, psychological, and
structural factors influencing changes in transportation modes (see Bamberg et al.,
2003; Forward, 2004; Graham et al., 2011), but difficulties persist in making a clear
link between a given feedback and a behavioral modification. Kluger and DeNisi
(1996) used the Feedback Intervention Theory (FIT) to evaluate how feedback interventions (FI) can be applied to influence employees’ task performance and initiate changes in their behavior within an organizational setting. They defined the term
‘feedback-standard gap,’ which denotes the comparison between the given feedback and the existing standard, such as a common practice in an organization. If the given feedback does not align with the given standard, employees tend to reject the FI
(Kluger and DeNisi, 1996). Kluger and DeNisi (1996, p.278) highlighted that further research needs to be conducted “to establish the circumstances under which positive FI effects on performance are also lasting and efficient.”
Alder (2007) extended the FIT by using computer-mediated feedback through a
computer performance monitoring information system (CPM IS) to evaluate changes in employee performance in the workplace. He found that changes were only effective if the CPM FI was not perceived as a control mechanism but rather as a support tool (Alder, 2007). A CPM IS has key benefits for employers and employees: it
is objective and improves performance evaluation (Angel, 1998; Worsnop, 1993).
On the downside, a CPM IS may be seen as an invasion of employees’ privacy, contribute to decreased job satisfaction, and increase stress (Greengard, 1996). In a
study conducted by Van Mierlo et al. (2004), subjects demonstrated the ability to
reduce their fuel consumption by up to 6% with the support of a prototype feedback
device that provided clear and concise advice on a screen without creating more
work for the driver. Although this example is limited to a simulated test environment, it offers important evidence of the potential of feedback technologies for
promoting fuel efficiency. These technologies could enable drivers to gain a better
understanding of eco-driving practices without attending a daylong training session.
2
Published paper: Tulusan, J., Steggers, H., Staake, T., Fleisch, E., Supporting eco-driving with ecofeedback technologies: Recommendations targeted at improving corporate car drivers’ intrinsic motivation to drive more sustainable, Energy Informatics 2012 (EI 2012), Atlanta, Georgia, United States, October 2012.
6
Introduction
In contrast, Lee et al. (2010) found a deterioration in fuel efficiency as drivers’ task loads were increased as a consequence of using an eco-driving support system.
In addition to feedback technologies, the appropriate time at which feedback should
be given to corporate car drivers needs to be further explored. The literature on energy savings in households shows that direct or real-time feedback has a positive
impact on residents, as they get an immediate response to their energy consumption
and are able to act accordingly (Graml et al., 2010; Winkler and Winett, 1982). In
relation to driving, direct feedback can improve fuel efficiency but also distract the
driver, increase the risk level, and even lead to increased fuel consumption
(Donmez, 2007; Meschtscherjakov et al., 2009). Indirect feedback, on the other
hand, refers to decontextualized types of feedback such as a detailed breakdown of
a driver’s fuel consumption, CO2 emissions, or acceleration/braking patterns. This
type of feedback is offered after driving behavior has been monitored and analyzed.
A study of college students by Graham et al. (2011) targeted a reduction in their
daily car usage, and the results revealed that providing them with an online IS
showing their financial and CO2 emissions savings facilitated an overall reduction in
their car usage.
Finally, a driver’s motivation has a large impact on reducing their overall fuel consumption. Among the studies researching the effectiveness of eco-driving interventions, Johansson et al. (2003) found that the pre-existing motivation of individuals
to reduce their fuel efficiency was essential. Since improvements in fuel efficiency
do not directly result in financial benefits for corporate car drivers, different motivations need to be evaluated in order to understand which factors encourage these
drivers to adapt their driving style.
To summarize, this research builds on the extant literature by evaluating the feedback provided by feedback technologies. Feedback through information communication technology (ICT) is recognized as part of the concept of a CPM IS, as evaluated by Nussbaum and duRivage (1986) and Alder (2007). A CPM IS enables organizations to provide their employees with information about their personal performance by collecting relevant data when they are working. This improves their
awareness of areas for development and enables them to modify their behavior
when necessary (Grant & Higgins, 1989). The FIT from Kluger and DeNisi (1996)
Introduction
7
can be used to evaluate changes in task performance (i.e. changes in driving behavior) as they relate to feedback technologies. Yet there remains a gap in this literature, as it does not yet consider how feedback technologies such as smartphone applications or a CPM IS can promote eco-driving amongst corporate car drivers by
informing them about their individual fuel consumption through direct or indirect
feedback. In this way, corporate car drivers’ intrinsic motivation must be targeted in order to encourage long-term sustainable driving behavior.
I.2
Objectives and Research Questions
These theoretical and practical motivations suggest a discrepancy surrounding the
potential for feedback technologies to instill eco-friendly driving habits in their users. While the approval of many in-vehicle feedback systems largely reflects a positive general outlook for the technology, only a small number of current studies have
assessed feedback delivered through external feedback technologies such as
smartphones or fleet management information systems. There are no studies that
investigate the use of these feedback technologies with corporate car drivers over a
longer period of time. Corporate car drivers maintain their unique status as research
subjects because their vehicle expenses, such as fuel and car maintenance costs, are
covered by their employing organization. Monetary incentives and faster car depreciation have little authority to encourage them to behave sustainably. As these influencing factors are insignificant within this sample population, the extent to which
feedback technologies stimulate eco-friendly driving can be directly evaluated.
Using eco-driving feedback technologies in the corporate context has great potential
to support the success of corporate eco-driving that can reduce a company’s overall CO2 emissions and fuel costs. This is especially important for companies with large
car fleets and is relevant at the continental European level, given the constant increase in the number of corporate cars (Dataforce, 2011). The growing acceptance
of smartphone technologies within organizations and universal employee access to
email present a cost-effective and time-efficient alternative to participating in ecodriving training. This research study has both theoretical and practical relevance and
draws on identified gaps in the body of research appraised to formulate the following principal research question:
8
Introduction
How can eco-feedback technologies influence the average fuel efficiency of corporate car drivers when monetary incentives are not given?
To assist in answering the principal research question, the following sub-questions
(SQ) must be addressed:
SQ1: Which eco-driving feedback technologies and feedback types are preferred
by private and corporate car drivers?
SQ2: How does an eco-driving smartphone application influence corporate car
drivers’ average fuel efficiency?
SQ3: Which feedback type, direct or indirect, do corporate car drivers prefer
when using an eco-driving smartphone application?
SQ4: How does social-normative feedback about a driver’s fuel consumption via email influence corporate car drivers’ average fuel efficiency?
SQ5: What recommendations can help an organization increase corporate car drivers’ intrinsic motivation to improve their fuel efficiency?
I.3
Research Methodology in relation to Research Questions
I.3.1 Mixed-Methods Research Methodology 3
This chapter offers a description of the epistemological viewpoint and research
methodologies that prove most appropriate for the research. Epistemology refers to
a system of assumptions about the nature of knowledge, its empirical material, and
how it can be acquired (Flick, 2006). Positivistic, anti-positivistic, interpretivistic,
and critical forms all represent different paradigms of research epistemologies
(Orlikowski and Baroudi, 1991). These different approaches not only influence the
way in which a researcher views reality but also have implications for the choice of
research aims and methods (King et al., 1994). The positivistic/quantitative school
of thought claims that all true knowledge is scientific and measurable (Black, 1999).
Research methodologies in this field observe, prove, and measure phenomena by
3
Course work: Tulusan, J., MP-02 Qualitative Research Methods Lecture, Judge Business School, Cambridge University, England, October 2008.
Introduction
9
using statistical or mathematical techniques and testing theories according to hypothetical-deductive models, laboratory experiments, or surveys.
On the contrary, Merton (1968) argues that not everything is scientifically proven or
possible to prove. Toulmin (1960, p.190) claims that a modern scientific researcher
should “not only concentrate on abstract and universal questions but to treat again specific, concrete problems….” Researchers should thus demand the explanations
behind the numbers and seek outcomes that are applicable to social reality (Lawson,
2008). The interpretivistic/qualitative school of thought is founded on the assumption that “systematic enquiry must occur in a natural setting rather than an artificially constrained one such as an experiment” (Marshall and Rossman, 1989, p. 1011). Interpretivistic approaches search for meaning based on the social phenomena
and context in which the research takes place. An interpretivist tries to explore phenomena such as human actions by making sense of the meanings symbolized by
those actions (Flick, 2006).
In this research, a mixed-methods research approach based on both positivistic/
quantitative and interpretivistic/ qualitative paradigms was chosen to: a) quantify
the empirical data from the experimental research and b) obtain a better understanding of eco-feedback technologies from the point of view of the driver and in the
context of broader social phenomena. Mixed-methods research approaches apply
different research methodologies within the same research design, and the chosen
methods should complement each other (Flick, 2006; Strauss and Corbin, 1998).
The research methodologies detailed in Table 1 were therefore applied.
Table 1: Mixed-method Research Methodologies
Qualitative Methodologies
Quantitative Methodologies
Desk research to obtain an overview of existing feedback technologies and relevant theoretical
insights.
Two experimental quantitative
field tests to measure the impact
of feedback technologies.
Case study to obtain a detailed
long-term understanding of a
company’s fleet management.
Pre- and post-experiment online
surveys to support the findings
with quantifiable results and
proven scales.
Semi-structured interviews to ac-
10
Introduction
quire a better understanding of
drivers’ opinions.
Pre- and post-experiment online
surveys to support the findings
with qualitative descriptive results.
Desk Research
Desk research enables the researcher to attain an initial understanding of the research area by reviewing available secondary literature, including practical and theoretical streams. The practical aspect focused on obtaining a better understanding of
different types of feedback technologies available on the market and the technical
capabilities required to support eco-driving concepts. The theoretical aspect drew on
various concepts, schools of thought, and studies that highlight the potential impact
of feedback technologies on promoting sustainable behavior. The keywords used to
search for relevant articles in research databases such as EBSCOhost, Mendeley,
and Google Scholar. These keyword searches identified articles in the fields of HCI,
energy reduction through feedback technologies and green IS, eco-driving concepts,
design of eco-feedback technologies, and sustainability-increasing behavioral
change.
Case Study4
Case studies are applied to describe an interpretivistic approach to social research
whereby a researcher can, over a period of time, observe, ask questions, and collect
the data required to probe issues that are central to the research (Eisenhardt, 1989).
Cornford and Smithson (1996) state that “the strength of the case study is in the richness of data that can be obtained” (Cornford and Smithson, 2006, p. 49).
King et al. (1994) distinguish between the numbers of variables measured and total
observations versus the number of case studies with which the researcher should be
involved. These aspects need to reflect the purpose of the study and “the amount of information a study brings to bear on a theoretical question” (King et al., 1994, p. 52). Throughout this research the researcher was able to work
4
Introduction
11
with a global company with a large corporate fleet serving a total of 19,100 corporate car drivers out of 60,000 employees. The company, which had the internal goal
of reducing its overall energy consumption, supported the researcher by granting
access to relevant data and providing management support. For this research, a
smaller sample size of 450 corporate car drivers from the Swiss region was chosen
for conducting the experiments. Several meetings with responsible employees, such
as sustainability officers, fleet managers, industry experts, and fleet management
software product owners took place. The researcher was also part of the company’s internal sustainability project team. Findings from the thesis were used to support
the company’s internal sustainability strategy and are currently implemented as part of a fleet management software solution.
Semi-Structured Interviews5
Semi-structured interviews assist in gaining first-hand information by analyzing the
rich responses of interviewees and enable the researcher to compare different viewpoints (Yin, 2008). It is important to prepare interview questions relevant to the
field and test them in advance to ensure that the questions used during the interview
yield credible responses (Burton-Jones and Gallivan, 2008). Semi-structured interviews help to maintain a balanced flow of conversation between questions and answers as opposed to imposing a rigid question-and-answer dialogue or being too
open (J. Hussey and R. Hussey, 1997). Follow-up semi-structured interviews were
conducted with corporate car drivers who participated in the field tests in order to
obtain a better firsthand understanding. Besides providing feedback, interviews
helped define and reconfirm recommendations and guidelines for instituting organizational eco-driving initiatives supported by feedback technologies. These were
then reconfirmed through workshops and interviews with relevant company stakeholders.
5
12
Introduction
Online Surveys
Surveys enable the collection of opinions using a standardized approach (such as an
online questionnaire using proven scales related to environmental behavior) and
provide answers that can be evaluated using descriptive or quantitative methods (J.
Hussey and R. Hussey, 1997). A pre-survey was conducted with both private and
corporate car drivers to get a better understanding of their preferences for feedback
technologies that support sustainable driving and the types of feedback offered by
these technologies. A post-experiment survey evaluated the opinions of corporate
car drivers who had participated in the experimental field tests, including whether
direct or indirect feedback was preferred for promotion of sustainable driving. Additionally, these surveys were used to control for drivers’ attitudes towards environmental behavior and their technological affinity in order to strengthen the generalizability of the findings. For measuring these behavioral attributes, the surveys
used constructs and scales that had been applied in the prior literature on energy
reduction and behavioral economics.
Experiments
The core research methodology centered on two experiments conducted on corporate car drivers working within the case study company. As the researcher is working for the company, it was possible to run these experiments while guaranteeing
the confidentiality of any sensitive data collected during them. The company management made the decision to pilot these experiments within a smaller group of 450
corporate car drivers to establish whether the technological approaches to ecofeedback had an impact on their employees’ overall fuel efficiency. Two experiments were conducted between the third quarter of 2011 and the first quarter of
2012. The first opt-in experiment included fifty corporate car drivers, of which
twenty-five used a smartphone eco-driving application for eight weeks. In the second opt-out experiment (n = 240), corporate car drivers received a monthly email
that detailed each driver’s individual fuel efficiency on a discrete or continuous feedback scale, illustrating how ‘good’ or ‘bad’ their fuel efficiency was in comparison to their colleagues (social comparison). The second experiment ran for three
months and enhanced the researcher’s understanding of the extent to which email feedback on fuel consumption influenced drivers’ fuel efficiency.
Introduction
13
I.3.2 Relationship to Research Questions
A summary of the overall research process is illustrated in Figure 2, helping to reflect the relationships between the research questions, chosen methodologies, and
findings. Findings obtained through this mixed-methods research approach made it
possible to address more than one sub-research question using the same data set;
they have been highlighted in the relevant chapters.
Figure 2: Research Process and Relation of Research Questions to Methods
14
Introduction
Initial research focused on secondary data collection/ desk research to acquire a
thorough understanding of feedback technologies currently on the market, as well as
obtaining a theoretical grounding in this topic area. This information was used to
formulate two online surveys to investigate what drivers expected from feedback
devices and which feedback types they preferred or found most relevant to them.
The first survey was answered by private car drivers (n = 139), the second by corporate car drivers (n = 131). This allowed for comparing preferences across the different user groups.
The preliminary research into SQ1 assisted in designing the first empirical study.
This experiment probed how an eco-driving smartphone application influences the
fuel efficiency of corporate car drivers. Fifty drivers participated in the opt-in experiment, of which twenty-five used the smartphone application for eight weeks, completed an online post-experiment survey, and participated in semi-structured interviews. The quantitative evaluation of driver’s fuel efficiency together with the qualitative approach of the survey and interviews enabled exploring how the use of an
eco-driving smartphone application changes over time. These findings support the
argument that an eco-driving smartphone application can have a positive impact on
the fuel efficiency of corporate car drivers who have not taken part in eco-driving
training.
SQ3: Which feedback type, direct or indirect, do corporate car drivers prefer when
using an eco-driving smartphone application?
The Feedback Intervention Theory of Kluger and DeNisi (1996) proposed that varying the features of the feedback delivered could manipulate the extent to which individuals would change their behavior. As highlighted by Kluger and DeNisi (1996)
“a device is required that gives the driver immediate and accurate fuel consumption
Introduction
15
information, yet is not a distraction from safe driving.” Meschtscherjakov et al.
(2009) highlighted that the distraction of such a technology can induce inflated risk
levels during driving. Immediate and context-related feedback is integral to promoting drivers’ fuel efficiency as they drive. Additionally, the ability to control the frequency of the feedback received has been proven to motivate users to change their
behavior. This is known as “feedback control” and allows individuals to control the amount and timing of feedback with which they are provided (Meschtscherjakov et
al., 2009). The eco-driving smartphone application allowed corporate car drivers to
choose between direct or indirect feedback types and recorded how long each feedback type was used during each journey. These data were used to statistically evaluate which feedback type was most popular; this was validated in the semi-structured
interviews.
SQ4: How does social-normative feedback about driver’s fuel consumption via email influence corporate car drivers’ average fuel efficiency?
Receiving feedback from a smartphone application depends on certain prerequisites,
such as owning a smartphone and activating the application during driving. These
hurdles can reduce the acceptance of such feedback technologies. An alternative
solution that is available to every employee in a corporation is providing feedback
via email, known as indirect/ offline feedback. This type of feedback is easily accessed by a large number of company drivers, as most of them have access to an
email account. Corporate car drivers have limited knowledge of their monthly fuel
consumption unless noting it autonomously, but monthly emails make this information readily available to them and assist with improving their awareness of fuel
consumption and how to improve their overall fuel efficiency. Hence, the second
experiment provided greater insight into how drivers’ fuel efficiency changes as a result of receiving an email detailing their fuel consumption and comparing this to
that of their colleagues (social comparison). The data needed to calculate the fuel
efficiency is automatically collected from the company’s petrol credit cards during the process of paying for each tank filling.
16
Introduction
Findings from the previous research questions must be considered within the correct
context to guarantee the successful application of eco-feedback technologies in an
organizational setting. Management must understand which factors, intrinsic or extrinsic, motivate their employees(Beswick, 2007). Extrinsic motivation refers to the
use of external rewards (e.g., financial incentives or extra days off) and/ or recognition (e.g., participation in important meetings) to stimulate change (Zahorsky,
2010). Intrinsic motivation stems from within the person; satisfaction thus derives
from completing the task itself, thereby improving the driver’s self-esteem
(Zahorsky, 2010). Because external motivations such as monetary incentives to improve fuel efficiency are irrelevant for this sample group, greater understanding is
needed to establish how to motivate corporate car drivers. Their intrinsic motivation
must be drawn upon so that they feel that sustainable driving is partly their responsibility. How this can be achieved will be explored and defined with clear recommendations in this chapter; this is done using information collected via the semistructured interviews, alignment meetings, and the online survey.
I.4
Scope of the Thesis
The goal of this doctoral thesis is to attain a better understanding of how ecofeedback technologies can be used to educate corporate car drivers about their fuel
efficiency. Data on tank refills and the number of kilometers driven monthly were
evaluated to determine the impact of the eco-friendly smartphone application or
monthly email on drivers’ fuel efficiency. These findings and the resulting recommendations should help corporations ascertain whether or not emissions reduction
concepts need to be supported by these feedback technologies to maximize their
impact. If so, corporations will be able to reduce their overall CO2 emissions and
fuel costs with minimal investment.
Introduction
17
I.4.1 Project Context6
These research activities were part of the Bits to Energy Lab (B2E), a collaboration
between the University of St. Gallen (HSG), Swiss Federal Institute of Technology
(ETH Zurich), and SAP Research. The B2E Lab took the lead on the academic research activities, receiving some management support from the case study partner,
and the researcher had access to the primary data generated during the field tests.
This was possible because the researcher is working in cooperation with both organizations, the B2E Lab and SAP Research.
The case-study company has 60,000 employees and a total of 19,100 corporate car
drivers (Q4, 2011). The number increased by 10% in 2011 and is predicted to rise
again in 2012. The total CO2 emissions in 2011 were 490 kTons, of which 23%
were attributed to corporate car drivers and 38% to business flights (see Figure 3).
Electricity accounted for 10%, and commuting and data centers from the case-study
company were each responsible for 8%.
490.0
kTons
Figure 3: 2011 SAP AG’s global distribution of CO2 Emissions7
The high CO2 emissions and energy consumption of the 19,100 corporate cars indicate an urgent need to determine effective and sustainable mobility solutions. The
first approach was to reduce the guideline CO2 emissions cap for new corporate cars
6
7
www.sustainabilityreport.com
18
Introduction
from 220 gCO2/ km down to 190 gCO2/ km. However, this guideline is only applicable to newly purchased cars; as corporate cars are driven for up to five years or
150,000 kilometers, implementation is gradual over time. A more immediate and
cost-efficient approach is thus required to modify driving behavior. Past initiatives
have focused on classroom training strategies such as eco-driving training, which
require drivers to invest time and the company to invest money.
I.4.2 Financial Saving Potential
Based on the potential fuel saving figures for corporate car drivers from the German
Federal Environmental Agency (2009), the case study company could save approximately four million Euros per year if their total fuel efficiency improved by one
liter per 100 km. This would represent a savings of nearly 10% of the total 2011
petrol expenditure of 42.5 million Euros. With a moderate average travel distance of
15,000 kilometers per year, the savings would be 1,250 Euros per corporate car8.
Since corporate car drivers drive an average of 35,000 kilometers per year, the savings could be much higher; nevertheless, a conservative estimate of potential fuel
savings was chosen.
I.4.3 Example of Data Set
The dataset used for calculating the fuel efficiency of each corporate car driver, an
example of which is shown in Table 2, includes each corporate car’s ID, the driver’s employee ID, monthly total kilometers driven, monthly total liters filled, total number of kilometers driven, and how long the driver has owned the vehicle (in
months).
8
Calculation is based on 19,100 corporate cars with a lifespan of six years and a constant petrol price of 1.40
Euro per liter.
Introduction
19
Table 2: SAP Corporate Car Drivers’ Data Set Example
Car Employee
ID
ID
Monthly
total km
Monthly
total liters
filled
Monthly
fuel efficiency in
l/ 100km
Total
Total ownership
km in months
1
1
3398
260
7,7
26290
11
2
2
2132
195
9,1
19545
7
3
3
1433
121
8,4
24245
10
4
5
2368
146
6,2
33390
13
These data were automatically collected from the drivers’ petrol credit cards during the process of paying for tank refilling and made available to the researcher. It was
then possible to verify monthly changes in the overall fuel efficiency of each driver
participating in the field tests. The fuel efficiency is calculated from these data by
dividing the total amount of liters filled per month by the total distance driven per
month and multiplying by 100 (see Figure 4).
  liters 
  100
FE  
  kilometers 


Figure 4: Fuel Efficiency Calculation
Further data were collected via the online pre- and post-experiment surveys using
open-ended questions and likert scales ranging from most to least preferable. To
extend the understanding gleaned from these purely numerical values, semistructured interviews were conducted to gain deeper insights into why users preferred certain feedback technologies over others. This enabled the researcher to gain
a better understanding of why eco-feedback technologies influence the driving behavior of corporate car drivers.
I.4.4 Research Context
Focusing on the technical rationale of this topic would provide a limited understanding of why drivers adapted their driving after receiving feedback from ecofeedback technologies. Difficulties would have arisen when applying these findings
to a practical business setting. As such, this research builds on two research streams
20
Introduction
to analyze changes in fuel efficiency from both social/ behavioral and technological
perspectives (see Figure 5). This is defined in the literature as the socio-technical
perspective (Bostrom and Heinen, 1977); emphasis is placed on evaluating the
technical and social artifact (the corporate car driver) to gather a better understanding of why new technologies have the potential to change organizational processes
(Keen, 1981).
Figure 5: Research Perspectives
The technological perspective builds on the literature in the field of feedback information systems, specifically persuasive mobile eco-feedback technologies. Fogg
(2002) classified persuasive technologies as “any interactive computing systems designed to change people’s attitudes or behaviors.” These technologies can bridge
the gap between one’s lack of environmental awareness and everyday behavior;; this has been termed the ‘environmental literacy gap’ by Froehlich et al. (2010). Published research has claimed statistically significant reductions in household energy
use (Geller et al., 1982; Graml et al., 2011; Loock, Graml, et al., 2011). Interaction
and usage behavior have also been explored within the Human Computer Interaction (HCI) research domain. Authors such as Froehlich et al. (2010) and Lee et al.
(2010) are particularly interested in how humans interact with these devices (e.g.,
how corporate car drivers use the eco-feedback technologies over time) and how the
technology in these devices should be developed, designed, and applied.
The behavioral perspective, with its origin in social-psychological concepts
(Cialdini et al., 1991), assesses the effectiveness of these devices in stimulating behavioral changes, such as changing fuel efficiency by evoking more ecological driving behaviors. Past research has evaluated why people speed and how to encourage
seat belt use to improve safety (Geller et al., 1982). Determining how (whether via
numbers, symbols, or scales) and when (direct/ real-time vs. indirect/ after driving)
eco-feedback should be presented is important for reducing a driver’s cognitive Introduction
21
workload and level of distraction when using a feedback device (Donmez, 2007).
The motivations for employees within an organization to adapt their behavior on
work-related tasks has been appraised using the concepts of extrinsic and intrinsic
motivational factors (De Young, 1985; Zahorsky, 2010).
The business value perspective builds on the social/ behavioral technological findings to suggest recommendations for industry stakeholders and present the first
practical design for supporting eco-driving via feedback technologies. These recommendations and concepts can be used by organizations to promote energy reductions or to define new CO2 emissions reduction policies within an organization. As
an example that directly reflects this research, relevant findings are currently being
implemented as part of the overall CO2 reduction strategy of the case study company. Specifically, the company is offering the eco-driving smartphone application to
all interested drivers and is providing information to all drivers via a monthly feedback email. The latter compares the details of an individual driver’s fuel efficiency and CO2 emissions to that of their colleagues and offers eco-driving tips. These services will also be added to a fleet management software solution and could be offered to other corporations. Further details are outlined in Chapter VI.
I.4.5 Target Audience
The thesis is targeted at both practitioners and members of the academic community
who want to gain a better understanding of how feedback technologies enable corporate car drivers to improve their overall fuel efficiency. For practitioners (i.e. sustainability managers in organizations, corporate fleet managers, or consultants) these findings could extend and support a sustainability strategy aimed at reducing
their company’s overall energy consumption. Those who develop products such as smartphone applications or fleet management software solutions also have a vested
interest in how feedback technologies should be enhanced to improve services to
their clients. Findings are particularly relevant for corporations with a large corporate car fleet.
Academics in the fields of green IS, HCI, and behavioral economics will find that
the outcomes of this research contribute to the technical research domain and enhance knowledge of eco-feedback technology usage and user interaction. The be-
22
Introduction
havioral economics research focus aims to understand why corporate car drivers’ behavior changes even if monetary incentives are not relevant. Given the present
importance of energy reduction, this research will be of strong interest to the research community.
I.5
Structure of the Thesis
Figure 6: Structure of the Thesis
The last section of this chapter will briefly explain the content of each chapter (see
Figure 6).
The introduction outlines the practical and theoretical relevance of the thesis, after
demonstrating the overall significance of the topic. The research questions, subquestions, and mixed-methods research methodology are outlined, indicating the
direction of the ensuing thesis.
The second chapter reviews ‘state of the art’ concepts, the theoretical foundations of eco-driving and eco-feedback technologies and discuss studies that have indicated
how these technologies enable drivers to adopt an eco-friendly driving style.
Introduction
23
Chapter third, a preliminary study, evaluates drivers’ preferences for feedback technologies based on an explorative investigation, indicating which feedback types are
preferable for influencing the driving behavior of both private and corporate car
drivers.
Chapter fourth explains the eco-driving smartphone application experiment, in
which corporate car drivers used an eco-driving application for eight weeks. Statistical analysis is used to analyze usage patterns of the eco-driving application over
time and which of the feedback meters shown on the application were preferred.
Chapter fifth describes the second experiment, which evaluated the difference in
fuel efficiency between groups of drivers’ using two different feedback scales. This information was delivered to the participants via monthly emails that highlighted
their fuel consumption figures. This field test involved 240 corporate car drivers.
Chapter sixth outlines insights and recommendations for increasing intrinsic motivations of drivers to improve their fuel efficiency. The five basic arguments from
the FIT from Kluger and DeNisi (1996) have been applied for a structured analysis.
Guidelines were given on the findings of the post-survey and validated through
semi-structured interviews.
Chapter seven provides an outlook of a prototype fleet management solution, data
structure and the system´s architecture by illustrating current mock-ups of the solution. Drawbacks are also highlighted.
Finally, the last chapter highlights key findings and summarizes the theoretical,
practical, and technological implications. Furthermore, provides recommendations
and the first conceptual model of how to apply eco-driving programs in organizations with the support of eco-feedback technologies. Limitations and future research
are also considered and an outlook of a prototype fleet management solution illustrated.
Eco-Driving and Eco-Feedback Technologies Related Work
25
II Eco-driving and Eco-Feedback Technologies Related
Work
This chapter provides an overview of eco-driving and related techniques that can
improve drivers’ fuel efficiency. The potential fuel savings achievable through ecodriving are also illustrated with examples from previous research. The difference
between private drivers and the user group focused on in this research, corporate car
drivers, is also discussed.
II.1 Overview of Eco-driving9
Different driving styles result in fluctuations in fuel efficiency; for example, aggressive drivers consume 30% more fuel than conservative drivers (Romm and Frank,
2006). Aggressive drivers who express bodily or psychological aggression towards
drivers, passengers, or pedestrians also demonstrate risky behavior (Tasca, 2000).
Risky driving behavior is chiefly characterized by speeding, changing lanes frequently, and maneuvering without signaling (Dula and Geller, 2003). In contrast,
economic driving behavior, referred to as hypermiling, has the ultimate goal of consuming the least possible volume of fuel, even if this means compromising on safety to a degree (Chapnick, 2007). Another behavior is eco-driving; this does not
compromise on road safety and can lead to average reductions in fuel consumption
of up to 15%, lower greenhouse gas emissions, improved road safety, and reduced
accident rates (GreenRoad, 2008). Eco-driving depends on individuals’ behavior;; its principles thus aim to alter and discourage driving practices that do not conform to
eco-driving techniques.
EcoDrive (2012) defines several practices that can alter driving behavior:
 Anticipating traffic flow and looking ahead in order to be able to react quickly and avoid sudden starts or stops.
9
Published paper: Tulusan, J., Soi, L., Paefgen, J., Staake, T., Eco-efficient feedback technologies: Which
eco-feedback types prefer drivers most?. World of Wireless, Mobile and Multimedia Networks IEEE Conference, Lucca, Italy, June 2011.
26
 Shifting into a higher gear as soon as possible once the threshold of 2500 or
2000 revolutions per minute (rpm) for petrol and diesel cars, respectively, is
reached.
 Using the highest gear possible at a low RPM and maintaining a steady
speed.
 Smoothing deceleration by releasing the accelerator in time and leaving the
car in gear.
 Using fuel-saving devices onboard vehicles, such as cruise control or a navigation system’s ecoRoute setting.
 Considering switching off instruments that consume extra energy, such as
air-conditioning.
 Checking tire pressure at least once a month.
Advantages and disadvantages of such eco-driving are listed in Table 3.
Table 3: Advantages and Disadvantages of Eco-Driving
Advantages
Disadvantages
Vehicle type unimportant
Driver has to be willing to
change driving behavior
Easy to learn
Eco-driving training required
Raises awareness of ecological driving practices
Few training possibilities
available
Low effort (one day training)
Might not have a longterm effect without further
interventions
Low investment (from 100
Euros)
Loss of time due to longer journey time
Advantages of eco-driving are its independence from the vehicle type and the ease
of learning the driving techniques with low financial and time investment (a one day
training starts from 100 Euros). Among the studies investigating the efficiency of
eco-driving, however, one demonstrated that the majority of drivers who improved
their fuel economy were already motivated to do so before training; this motivation
appeared to have an substantial impact on reducing fuel consumption (Johansson et
27
al., 2003). During and shortly after training, drivers achieved maximum improvements in fuel efficiency of up to 15%; these gains declined in the long term if ongoing prompts were not given (EcoDrive, 2012; Onoda, 2009). This gradual deterioration in maintaining fuel efficiency indicates the necessity for substantial ongoing
measures to modify activities that are performed automatically. Potential interventions include ongoing feedback, regulatory actions, economic incentives, and feedback technologies that inform drivers after they have participated in the classroom
training.
II.1.1 Eco-driving Training Practices 10
Campaigns, such as the pan-European ECODRIVEN campaign that ran from 2006
to 2008, are one approach to introducing eco-driving. That campaign helped raise
awareness of eco-driving by offering lessons during regular one-day classroom
training courses that included theoretical and practical sessions (EcoDrive, 2012). In
the theoretical classroom session, eco-driving guidelines and recommendations
were taught. In the practical session, drivers were able to use a driving simulator
and/ or their own vehicle in a designated driving area to practice the eco-driving
techniques. Even though further classroom education is welcomed, it is insufficient
to effectively alter driving habits that have become entrenched in individuals
through years of practice, as drivers tend to return to previous driving habits soon
after training is complete. Therefore, classroom training should be supported by
internet-based training material and/ or reoccurring practical training sessions that
explain the eco-driving principles again or remind participants of the most relevant
and up-to-date driving practices. Ideally, changes in driving style that promote ecodriving behavior should be maintained on a long-term basis.
10
28
II.1.2 Eco-driving Fuel Saving Potentials 11
The transport sector report produced by the International Energy Agency (2009)
analyzed the short- and medium-term impact of eco-driving initiatives. Immediately
after eco-driving training, average fuel efficiency improved by up to 15 % (Onoda,
2009). In the long-term (i.e. more than one year later), an improvement of approximately 5% was maintained without further feedback and an improvement of 10%
was achieved with continuous post-training feedback (i.e. tips and driving style
analysis) (EcoDrive, 2012). A five-percent improvement in fuel efficiency can be
sustained by using in-car eco-driving equipment or an onboard computer that provides feedback on fuel consumption, correct gear changes, or cruise-control functions. Corporate studies conducted by EcoDrive (2012) with 350 service car drivers
from the Canon company showed an overall reduction in fuel consumption of 6.1%,
with 35% fewer car accidents. Corporate car drivers working for the German company Hamburger Wasserwerke achieved a 6% decline in fuel consumption and a
25% decrease in the number of vehicular collisions. The Austrian bus company
NIGGBus was able to improve its fuel efficiency by 5% in 2000 and 7% in 2001.
These figures show that eco-driving training can have a positive impact on the driving behavior of corporate car drivers. The following sub-chapter introduces the differences between two driver groups: private and corporate car drivers.
II.2 Differences between Private and Corporate Car Drivers12
Corporate car drivers are defined as those who drive a corporate car, largely for
work-related travel but also for private reasons. The nature of their occupations requires daily travel to customer meetings (i.e. to pitch sales or provide consulting
services) and is integral to the company’s generation of leads and income. 11
Published paper: Tulusan, J., Staake, T., Fleisch, E., Direct or indirect sensor enabled eco-driving feedback: Which preference do corporate car drivers have?, Internet of Things 2012 – Third International Conference on the Internet of Things (IoT 2012), Wuxi, P.R. China, October 2012.
12
29
Table 4: Differences between Private and Corporate Car Drivers
Travel Reasons
Average travel distance
(in Europe)
Payment of petrol costs
Payment of car maintenance
Incentives to drive
ecologically
Passenger car registrations in 2012 (Germany)
Passenger car registrations in 2011 (Switzerland)
Private Car Drivers
Private
Corporate Car Drivers
Work and/ or private
13,500 km/ year
35,000 km/ year
Self-payment
Paid by the company
Self-payment
Paid by the company
Save petrol costs
No incentives for the
driver
2,250,000
750,000
213,000
106,000
In Europe, corporate car drivers drive an average of 35,000 kilometers per year (see
Table 4). In comparison, private car drivers drive an average of less than half this
distance, 13,500 km per year (DTLR, 2012). Private car drivers must pay their own
petrol and car maintenance costs, whereas corporate car drivers pay a monthly fee
(a leasing fee or a fee for the company) for their car of between 300 and 600 Euros.
This fee covers petrol costs for business-related mileage and maintenance and service costs, depending on the car type and country. The car’s CO2 emissions have an
impact on the size of the monthly fee; the higher the emissions, the greater the fee.
In some countries, such as Germany, drivers have to pay an additional monthly tax
to the government of 1% of the total purchase price of the corporate car. Petrol is
paid through corporate credit cards, and the total number of kilometers driven must
be typed into the credit card terminal during the payment process at the petrol station. The petrol companies provide the company with the tank filling data each
month, but the corporate car drivers do not receive a monthly overview of these
costs or of the volume of petrol consumed. Hence, drivers are unaware of these
monthly costs unless they maintain a personal record.
For private car drivers, incentives to drive more ecologically include direct savings
in fuel costs, whereas corporate car drivers have their fuel and vehicle maintenance
costs paid by the company. As such, economic and eco-friendly driving concepts do
not directly lead to financial rewards for these individuals. An improvement in their
30
fuel efficiency would, however, result in financial savings for their employing company.
The total number of corporate cars in Germany indicates that a quarter (750,000 of
3,000,000) of cars newly registered in 2012 were corporate cars (Dataforce, 2011).
For Switzerland in 2011, the fraction was one third (106,000 of 319,000)
(Bahnmüller, 2012; Ullmenstein, 2012). Bearing this in mind, it is important to understand how corporate car drivers can be motivated to adopt more sustainable driving behavior.
II.3 Eco-Feedback Technologies
II.3.1 Eco-Feedback Technologies in Cars 13
GreenRoad (2008, p.12) suggested “a device is required that gives the driver immediate and accurate fuel consumption information, yet is not a distraction from
safe driving”. A limited number of in-car eco-feedback devices have been pioneered for vehicles that use petrol. In spite of ongoing advancements in technology,
a large percentage of feedback devices are utilized for hybrid or electrical vehicles,
as the most effective concept of ecological driving behavior extends across this
range of vehicle types (Romm and Frank, 2006).
Eco-feedback technologies in cars aim to generate the driver’s awareness of their driving style and consumption of petrol (Formosa, 2009). The driver needs to receive information visually through a feedback technology, whereas the technology
needs to know what the driver is doing. This interplay (see Figure 7) of the information received and the response of the driver could influence the driving style
(Formosa, 2009).
13
Published paper: Tulusan, J., Brogle, M., Staake, T., Fleisch, E., Becoming a Sustainable Driver: The
Impact of Mobile Feedback Devices. 4th ERCIM eMobility WWIC Conference, Lulea, Sweden, June 2010
31
Figure 7: Feedback Technologies and Response of Driver 14
Car manufacturers focus on enhancements of on-board communication instruments,
which support the driver during the journey. These functions focus more on
telematic solutions or car2car communications, such as calculating the correct distance to a car in front and automatically slowing down the car to keep a safe distance. Ford, with his SmartGauge concept, worked with the service design company
IDEO to define a concept for an on-board instrument to make the driver aware of
eco-efficient driving. The notion that design can influence behavior underlines the
importance of developing systems that are relevant to the driver (Formosa, 2010).
The SmartGauge concept connects the instrument with the controller area network
(CAN) bus of the car to store relevant data, such as average speed, acceleration,
braking, liters used, etc., from the car. From this information the system is able to
understand the driving behavior and displays real-time feedback by displaying
leaves of the fuel usage (see Figure 8 right side).
Figure 8: Ford’s SmartGauge in-car Interface14
14
Formosa, D. (2009), “Ford SmartGauge: Designing an Extra 9 MPG?,” Smart Design. Retrieved from
www.smartdesignworldwide.com/pdf/SmartDesign_SmartGauge.pdf
32
II.3.2 Mobile Eco-Feedback Technologies15
Through the advancements of smartphone technologies by companies, such as Apple, Google, HTC, and Samsung, the latest eco-feedback technologies are ecodriving applications that run on mobile devices. Eco-driving applications use the
GPS sensor and the accelerometer technologies of the smartphone and empower the
driver by analyzing their driving style to provide recommendations about how they
can drive in a more eco-friendly manner.
In 2009, only one application was available on the market (see Figure 9). In 2012,
29 offer the iOS and/ or Android operating systems. Since it was not possible to
compare the iOS ratings with the Android application and user ratings were only in
relation to the latest version of the application, it was not feasible to provide a valid
ranking for the least and most favored applications. The overall trend indicates that
more and more applications are available on the market, but the frequency use during driving and the impact that these applications have on driving behavior remain
unclear.
Figure 9: Eco-driving Applications Market Share
15
Published paper: Tulusan, J., Brogle, M., Staake, T., Fleisch, E., Becoming a Sustainable Driver: The
Impact of Mobile Feedback Devices. 4th ERCIM eMobility WWIC Conference, Lulea, Sweden, June 2010
33
One example of a current application is the Bliss Trek (see Figure 10) eco-driving
application, which tries to persuade the driver to modify their driving in a playful
way by using speed and acceleration data. Real-time feedback with respect to ecological driving behavior is illustrated using a green meter represented by leaves and
a driving score. The goal is to reach 1000 points to reach the next level. When the
driving score deteriorates clouds appear on the screen and the sky darkens
(Stoppani, 2009).
Figure 10: Bliss Trek Application Interface16
Another example is the greeenMeter application (see Figure 11), which is more
straightforward and displays the average speed in an eco-driving efficiency gauge.
Depending on the average speed and roads driven the software displays the area for
eco-efficient driving. Besides this gauge, further feedback meters that show the CO 2
emissions, fuel efficiency, and fuel costs are able to be selected.
16
Stoppani, S. (2009), “Bliss TrekTM,” Overview of the BlissTrek Application. Retrieved August 5, 2012,
from http://blisstrek.com/
34
Figure 11: greenMeter Application Interface17
This interface changes colors according to your driving behavior; it turns dark blue
when you consume more fuel and bright green when you drive more efficiently.
Figure 12: Color changing Feedback18
While you drive efficiently the silver ball remains into the green area. If you start
consuming more fuel or accelerate aggressively, the ball moves to the yellow and
17
18
Hunter, C. (2009), “greenMeter: iPhone/iPod Eco-driving App.” Retrieved September 5, 2012, from http://hunter.pairsite.com/greenmeter/
Source: The Game (Changer), motive, 2009
35
red areas. In addition to this you get a driving score, marking your driving behavior
at each trip.
Figure 13: Green Driving App19
A selection of further applications currently on the market that have similar ecodriving functions are shown in Table 5:
Table 5: Overview of Eco-driving Applications
Logo
19
App.
Name
Company
Release
Date
Target
Market
Oper.
System
goDrive
Green
ProLogi
c, Inc.
Jan. 12th,
2011
USA
iOS
EcoDriv
e
Applications
Co. Ltd.
May 19th,
2012
China,
Japan
iOS/
Android
Eco
Bccard
Ltd.
Sep. 4th,
2012
Korea
Android
DriveGa
in
DriveGa
in Ltd.
Jul. 20th,
2010
UK,
Netherlands
iOS
Tokyo
Smart
Driver
maysun
soft Co.
Ltd.
Oct. 27th,
2011
Japan
iOS
Source: GreenGasSaver, The Apple Seed Story, 2009
36
II.4 Smartphones Market Share
Looking at the smartphone market share penetration (see Figure 14), Frost & Sullivan (2009) estimated a market share of smartphones of the total mobile phones by
2014 of 37% and Gartner (2010) of 40%. In 2008, the smartphone market share was
only 12% (Frost & Sullivan, 2009) or 11% (Gartner, 2010). The total mobile
phones’ sales rate in 2014 is predicted to be 1,516 million (Frost & Sullivan, 2009) or 1,908 million (Gartner, 2010) units. These estimates support the importance ecodriving smartphones application could have in the upcoming years.
20
21
Total
mobile
phones
Smart
phones
(F&S)20
Market
Share
(F&S)
Total
mobile
phones
Smart
phones
(Gart.)
Market
Share
(Gart.)
2008
1178
139
12%
1222
139
11%
2009
1223
190
16%
1211
172
14%
2010
1267
251
20%
1400
260
19%
2011
1313
321
24%
1500
374
25%
2012
1375
398
29%
1600
484
30%
2013
1451
478
33%
1750
621
35%
2014
1516
555
37%
1908
770
40%
21
Frost & Sullivan. (2009), 2010 Outlook & Forecast: Mobile & Wireless Communications.
Gartner. (2010), Forecast: Mobile Devices, Worldwide 2003-2014, 2Q10 Update
37
40%
800
30%
600
25%
500
19%
400
300
200
40%
35%
700
14%
11%
12%
29%
45%
37%
33%
35%
30%
25%
24%
20%
20%
15%
16%
10%
100
5%
0
Smartphone market share
Total smartphones (millions)
900
0%
2008
2009
2010
2011
2012
2013
2014
Total smartphones (Frost&Sullivan)
Total smartphones (Gartner)
Smartphone market share (Frost&Sullivan)
Smartphone market share (Gartner)
Figure 14: Smartphone Market Share Outlook
II.5 Impact of Eco-Feedback Technologies on the Driving Behavior
The absence of a significant body of literature in the field of mobile eco-efficient
feedback technologies does not allow for one to confidently speculate the potential
success of similar systems. At the discretion of one’s knowledge, currently no research exists, besides studies in which an eco-driving navigation or on board system
have been used that directly measure the influence Green IS or smartphone technology have on driving behavior. However, feedback technologies “seems to be a promising approach to be researched in automotive environments, when it comes to
fostering a more fuel-efficient driving behavior” (Meschtscherjakov et al., 2009,
p.82). Froehlich et al. (2009) operated this type of mobile technology when recording the daily transportation behavior of participants, i.e. if the person used the bus or
bike, which presented a positive outcome for a small sample group (n=13) and field
test duration of only three weeks.
Monetary, environmental, and technological factors were found to be prominent in
inspiring drivers to modify their driving style, however, caution must be taken when
38
generalizing these finding as many of the studies analyzed how to influence drivers
to use other modes of transport or were limited to a laboratory setting. Foxx and
Hake (1977) acquired a 20% reduction in students’ mileage to campus by offering cash bonuses as a consequence of encouraging them to use their cars less and travel
on alternative modes of transport, i.e. carpooling, bike or bus, for an experimental
duration of one month. Deslauriers and Everett (1977) persuaded people to use the
bus by paying passengers $0.10 for each journey; however, this was only successful
for the period of time when reimbursement was given. Van der Voort et al. (2001)
conducted a laboratory study dedicated to improving drivers’ fuel efficiency. Results showed a reduction of 7% in fuel consumption when participants tested a fuelefficiency support tool in a driving simulator. The support tool provided drivers
with eco-driving information in response to their driving style and driving conditions. However, this device has yet to be trialed in the field.
Studies by Boriboonsomsin et al. (2010), Lee et al. (2010), Meschtscherjakov et al.
(2009), Pace et al. (2007) and Tester et al. (2000), among others, have evaluated
eco-driving on-board feedback systems. An improvement in fuel efficiency by 1%
on highways and 6% on city streets was shown in a field test with 20 drivers in
Boriboonsomsin et al. (2010) study. Participants participated in an eco-driving
training program and had access to a vehicle with an on-board feedback device,
which
provided
them
with
instantaneous
fuel
consumption
information
(Boriboonsomsin et al., 2010). Siero et al. (1989) applied feedback interventions to
urge postal-workers in the Dutch Postal Service to be more eco-friendly when driving by providing them with instant feedback via an installed tachometer in their delivery trucks and agreeing on a goal of reducing fuel consumption by 5%. Results
revealed a 7.3% reduction which was also maintained six months thereafter (Siero
et al., 1989). Situational and personality variables such as self-esteem, altruism, environmental believes, anxiety, attitude towards driving, and positive or negative
expectation of achievement are additional variables that could have had an impact
on behavioral change. For this reason, one cannot claim which feedback proposed
had the strongest impact on reducing fuel consumption. Siero et al. (1989, p. 417)
recognized this short-coming in their research and suggested that “…it would be desirable to split up such an experiment into a number of sub experiments to examine the effect of each individual element”.
39
II.6 Feedback Intervention Theory
Although there is evidence to suggest that it is possible to reduce energy consumption, “there does not appear to be substantial research on how specific behavioral changes lead to measurable reductions in GHG emissions” (Doppelt et al., 2009, p.
7). Many studies describe social psychological and structural factors when determining which transport mode to take, but struggle to make the link between a specific behavioral change and one’s attitude. Difficulty persists when attempts are made to relate behavioral modification with feedback interventions; quite often the
results have been contradictory and feedback has not necessarily improved performance (Balcazar et al., 1986). Further research is necessary to get a better understanding of the relationship between behavioral modification and feedback interventions.
Van Velsor et al. (1997, p.36) define feedback as “information about a person’s performance or behavior, or the impact of performance or behavior, that is intentionally delivered to that person in order to facilitate change or improvement”.
Kluger and DeNisi (1996) applied the Feedback Intervention Theory (FIT) to appraise how feedback interventions provoked driving motivation and behavior, determined by comparing feedback to goals, standards, or norms. They defined feedback interventions (FIs) as “actions taken by (an) external agent(s) to provide information regarding some aspect(s) of one’s task performance” (Kluger and
DeNisi, 1996, p.255). Henceforth, FIs are applied to influence individual’s task performance and motivate them to change their actual behavior; this is referred to as
‘knowledge of results'. Interventions may encompass a broad spectrum of tasks that
are elicited by intent from an external agent. Kluger and DeNisi (1996) proclaimed
five basic arguments (see Table 6) for the FIT:
Table 6: Five Arguments Feedback Intervention Theory22
Behavior is regulated by comparisons made from feedback obtained by
goals or standards.
Goals and standards are organized in a hierarchy.
22
Kluger, A.N. and DeNisi, A. (1996a), “The effects of feedback interventions on performance: A historical
review, a meta-analysis, and a preliminary feedback intervention theory.,” Psychological bulletin, Vol. 119
No. 2, p. 254.
40
Attention is limited therefore, only feedback-standard gaps that receive
attention actively participate in behavior regulation.
Attention is limited, and therefore only feedback-standard gaps that receive attention actively contribute to behavior regulation.
FIs change the focus of attention and consequently, have an impact on
behavior.
Individuals use feedback to verify their performance in relation to their goals or
standards (Froehlich et al., 2009). Goals and standards are organized in a hierarchy;
the difference between the feedback obtained by the FI and the goal or standard set
is defined as the ‘feedback-standard gap’. The feedback-standard gap allows individual’s to reflect on their actual (positive or negative) performance with regards to the specified goal or standard (Kluger and DeNisi, 1996). Individuals may respond
to the feedback-standard gap in one of four ways: behavioral change, rejecting the
feedback, amending the goal/ standard set, or rejecting the goal/ standard set altogether. Standards can include multiple facets, such as norms, prior or past performance levels, and performance of other groups. Hence, multiple standards can influence the performance of one’s response to the FIs in variable ways;; for this reason, it is not as straightforward as claiming that behavioral change is solely a result
of the goals or standards appointed.
II.6.1 FIT and Computer Aided Feedback
Alder (2007) applied the FIT by Kluger & DeNisi (1996) to appraise how feedback, provided through a computer performance monitoring (CPM) IS, influences
the performance of employees within the company. Alder (2007) argued that computer mediated feedback allows one to focus on the tasks and results in improved
performance. Constructive feedback given with regards to performance was only
effective in promoting behavioral modification when the CPM was not perceived as
a control device by employees. Feedback from the CPM device in conjunction with
face-to-face feedback delivered the most benefits (2007). The value of CPM to organizations is undeniable as it produces performance data at an efficient and frequent rate, regularly drawing using awareness to their personal performance figures
and enabling them to modify their behavior when necessary (Grant and Higgins,
1989). Electronic monitoring provides employers and employees following benefits:
41
it is objective (Worsnop, 1993), it may help employees to avoid stress associated
with evaluation uncertainty (Alder and Tompkins, 1997), and it improves performance feedback and evaluation (Angel, 1998; Henriques, 1996). On the downside,
CPM may be seen as an invasion of employees’ privacy, contribute to decrease job
satisfaction and increase stress (Greengard, 1996). CPM has to be seen as one component of a complete performance management system, which is influenced by other elements that are also part of the process (Alder, 2007). In this research the assumption is made that “FIT predicts that computer-mediated feedback will lead to a
focus on the task” (Alder, 2007, p.163) and could improve the driving behavior.
II.7 Research Gap23, 24
Greater insight is needed into how eco-feedback technologies, such as CPM and/ or
mobile feedback technologies (e.g., smartphones), can be used to influence drivers
to adopt a sustainable driving style without classroom training. Feedback technologies can inform the driver about eco-driving techniques and provide information
about their fuel consumption. Improving drivers’ understanding of their fuel consumption could stimulate their motivation to drive more sustainably. The feedback
intervention type must be displayed so that obstruction is minimal, and that users
are able to decipher relevant information easily using their peripheral vision without
much effort or increased cognitive workload (Meschtscherjakov et al., 2009). Motivating participants to improve their driving behavior will play a significant role in
promoting fuel efficiency. When deliberated within the correct context, ecofeedback must correlate with drivers’ personal beliefs and culture to influence them in modifying their behavior (Foxx and Schaeffer, 1981).
The FIT of Kluger and DeNisi (1996) will be applied as a theoretical foundation to
evaluate changes in task performance (i.e. change of driving behavior) in response
to eco-feedback technologies. Feedback through ICTs is recognized as part of the
concept of CPM evaluated by Alder (2007) and Nussbaum and duRivage (1986).
23
24
42
CPM enables organizations to provide their employees with information about their
personal performance by collecting relevant data when working. This improves
their awareness of areas for development and enables them to modify their behavior
(Grant and Higgins, 1989). How these concepts were applied to this research and
evaluated via two field tests with corporate car drivers is explained in the following
chapters.
Private and Corporate Car Drivers’ Preferences for Eco-feedback Types
43
III Private and Corporate Car Drivers’ Preferences for Eco-feedback Types
III.1 Overview and Research Question
Based on the eco-driving fundamentals and technologies introduced in Chapters 0,
this chapter evaluates the preferences of private and corporate car drivers for feedback types presented by eco-feedback technologies. Studies by Graham et al. (2011)
have indicated that feedback detailing financial savings (i.e. reduced fuel costs) and/
or CO2 savings influence the driving behavior of private car drivers. Feedback related to financial savings had a stronger influence on reducing car usage than that related to CO2 savings, and drivers achieved greater fuel reductions when both types
were given, as opposed to only one type (Graham et al., 2011). However, how feedback given by eco-feedback technologies influences corporate car drivers has not
yet been explored. The following research sub-question was therefore identified:
Findings from this study provide insights into private and corporate car drivers’ preferences concerning eco-feedback types. Outcomes also confirm Graham et al.’s (2011) findings that eco-feedback must be easily understood and that feedback
types illustrating financial savings have the strongest impact on private car drivers.
Corporate car drivers preferred receiving feedback that detailed their fuel efficiency
and included eco-driving tips, as this enhanced their understanding of how their
driving could impact fuel efficiency. As expected, financial savings were not relevant. Feedback on sustainable driving techniques was received well by both driver
groups but especially by drivers who had no understanding of eco-driving concepts
or techniques.
III.2 Research Design and Research Methodology
The research design centered on two online surveys: one for private car drivers and
one for corporate car drivers (see Figure 15). Findings were compared across these
two driver groups, and results were used as a foundation for designing the experiments conducted with corporate car drivers (see Chapter IV and Chapter V).
44
Figure 15: Research Design
The first online survey ran for four weeks in August 2010 and was completed by
private car drivers (n = 139). The goal was to obtain a better understanding of how
feedback technologies function and which eco-feedback types are best for stimulating eco-driving behavior. The second online survey (n = 131) ran for three weeks in
August 2011 and was completed by corporate car drivers prior to the field test. This
survey was used to acquire a better understanding of which eco-feedback types
could best encourage this sample population to adapt their driving habits to drive
more sustainably. Questions enquired about participants’ preferences for the feedback types they would find helpful to see on an eco-driving feedback device. The
complete surveys are shown in Appendixes A and B.
III.3 Data Evaluation and Findings
III.3.1 Data Evaluation
The survey included three different question types: five-point likert scale, multiplechoice, and open-ended. The five-point likert scale questions required participants
to rate each item’s favorability between one and five, with one being the least fa-
45
vorable and five being the most. A rating higher than three (the scale mean) was
interpreted as favoring the selected item. This enabled the researcher to make a
comparison between the specified items and ascertain the most popular or wellreceived item in each category. For each item, a dependent two-tailed t-test was
used to determine whether the deviation from the scale mean was statistically significant.
Multiple-choice questions were analyzed by calculating the overall percentage of
respondents choosing each item (i.e. by summing the number of times each item
was chosen and dividing this sum by the maximum number of times the item could
have been selected). Open-ended questions gave participants the opportunity to express themselves more uniquely than allowed via the predefined five-point likert
scales or multiple-choice answers. The unstructured answers received from the
open-ended questions were evaluated to identify similar opinions and ideas (i.e. factors), which were then grouped into categories (see Figure 16).
Open-ended Questions
Evaluation and Reduction
Factors
Factors
Factors
Categories
Factors
Categories
Ranked Factors and
Categories
Figure 16: Categorization Process of Open-ended Questions
Cronbach’s alpha was calculated for the five-point likert scale questions, revealing a
value  = 0.73 for the first survey and  = 0.79 for the second. This suggests decent
consistency between the overall answers (Bernstein and Nunnally, 1994): as a
46
measure of the meaningfulness of a set of data, Bernstein and Nunnally (1994) recommend a Cronbach’s alpha value above 0.7. The normality of the likert scale responses was tested using the Shapiro-Wilk normality test. The result was nonsignificant, confirming that it was valid to compare answers between the two driver
groups. The next two paragraphs describe the findings from the two online surveys;
differences between private and corporate car drivers are highlighted.
III.3.2 Findings from Survey Conducted with Private Car Drivers
Of the 139 participants who completed the full survey, 64% were male and 36%
female; 66% were between 25 and 34 years old. The number of years for which participants had been driving varied, with 43% of drivers having five to nine years experience followed by 28% having zero to four years experience. Fifty-eight percent
of participants drove on a daily basis, and sixty-eight percent declared that they
were aware of feedback devices that targeted improvements in driving behavior. As
shown in Figure 17, navigation systems were the most widely known devices, with
35% of respondents being familiar with them, followed by trip computers (25%),
on-board gauges (23%), and smartphone applications (only 13%). This demonstrates that only a minority of drivers is aware that there are smartphone applications available to help them improve their driving style. Other in-car technologies
known to the participants were flat/ short shift and cruise control.
100
90
80
70
60
50
40
30
20
10
0
89 (35%)
64 (25%)
57 (23%)
33 (13%)
9 (4%)
Navigation
System
Trip
Computer
On-board
Gauge
Smartphone Other (please
Application
state)
n (%); N=252
Figure 17: Awareness of Feedback Devices
47
Fifty-one percent of the drivers had used feedback devices while driving, and 68%
of these drivers reported that their driving style had changed as a result. Changes in
driving behavior were explored in detail via the open-ended questions. Analysis of
this data identified four categories (see Figure 18): a) more sustainable driving, b)
car handling during driving, c) perception of the interface, and d) improved driving
Question: Could you describe the changes in the
way you drove while using a feedback device?
experience.
Reduce fuel consumption as a motivation (24)
Environmentally friendly as a motivation (2)
Driving more smoothly (13)
Switching gears more often (5)
Distraction (5)
Gaming aspect (1)
More comfortable with routing decisions (4)
No difficulty finding destination (3)
Time saving (2)
Sustainable
driving
Car handling
Perception
of the
interface
Driving
experience
GPS usage
Figure 18: Changes while using a Feedback Device
A strong indication of more sustainable driving behavior was a motivation to reduce
fuel consumption or fuel costs, which aligns with the literature on why private car
drivers improve their driving style (Johansson et al., 2003). An additional contributing factor was car handling: drivers aimed to drive their car more smoothly and
improved gear changes to up-shift the gears more rapidly. This supports the goal of
possibly slowing down the car’s eventual deterioration (EcoDrive, 2012). How participants perceived the interface used for the feedback device was an influential factor; an interface that was too much of a distraction from driving was considered unfavorable from a safety point of view. The ease of interpreting the information provided by the feedback device was thus important; this was also found by
Meschtscherjakov et al. (2009). The final category defined was the impact of GPS
technology on driving experience. Factors relevant to this category were improved
driving experience and finding destinations easier because a GPS-enabled system
provided routing options.
48
Having introduced the topic area, the next survey section queried participants’ preferences regarding specific feedback types. These items were graded using a fivepoint likert scale, and there was an overall tendency to favor most feedback types
(see items with p < 0.05 in Figure 19).
1 Strongly Disagree
2
3
AVG fuel efficiency (monthly) 2%
1%11%
Personalized eco-driving tips 6%5%
11% 11%
17%
0%
Items
23%
18%
20%
5
5
5
5
3
3
3
3
41%
28%
38%
30%
35%
32%
26%
22%
29%
27%
22%
25%
25%
27%
40%
17%
24%
60%
80%
15%
100%
Deviation
Sig.
from Scale
(2-tailed)
Mean
4.47
0.86
1.47
.000***
4.02
1.12
1.02
.000***
3.86
1.16
0.86
.000***
3.79
1.19
0.79
.000***
3.46
1.14
0.46
.000***
3.39
1.24
0.39
.000***
3.16
1.24
0.16
.092
3.03
1.30
0.03
.775
Note: * p < 0.05; ** p < 0.01; *** p < 0.001
Mode Mean
AVG fuel efficiency (monthly)
AVG fuel savings (monthly)
Personalized eco-driving tips
Alternate. eco-efficient route
Goal setting
AVG CO2 emission (monthly)
Social comparison
Rewards
35%
21%
Goal setting 5% 15%
Rewards
65%
22%
Altern. eco-efficient route 6% 8%
Social comparison 10%
5 Strongly Agree
22%
AVG fuel savings (monthly) 6%3% 13%
4
Std.
Dev.
Figure 19: Preferences for Feedback Types
Statistically insignificant items were: a) fuel consumption feedback in comparison
to other users who drive the same distance (item - social comparison) and b) a reward when the driver has achieved an efficient driving behavior (item - rewards).
49
Since the results are not statistically significant, no clear preference for these two
feedback types can be concluded from the answers of this survey.
All other feedback types were relevant for private car drivers, with a statistically
significant deviation from the scale mean (p < 0.05). Participants most preferred to
see their own average fuel efficiency (4.47 out of 5), followed by the potential savings they could have achieved by driving sustainably (4.02 out of 5). Receiving
feedback consisting of personalized eco-driving tips, an alternative eco-efficient
route, goal setting for fuel efficiency improvements, or monthly average CO 2 emissions was not as important. In sum, private car drivers had the strongest preference
for feedback that detailed their fuel consumption (87% agreement), followed by that
which noted potential financial savings (76%).
When asked how the feedback should be provided, 52% of the respondents desired
visual feedback (i.e. shown by a symbol), 43% preferred both visual and auditory
(i.e. through a voice), and 6% auditory only. This shows that feedback should be
presented visually or visually and audibly. Multiple-choice questions were used to
probe participants’ awareness of the impact of CO2 emissions on the environment
and their views of the importance of reducing these emissions (see Table 7). When
asked if they understood what was meant by ‘100 grams CO2 emission per km,’ 54% were unable to translate this into a meaningful context. This clearly indicates
drivers’ lack of understanding of the impact of CO2 emissions on the environment.
Of the 54%, 71% felt that improving their understanding of CO2 emissions would
better motivate them to adapt their driving style to be more eco-friendly. Sixtyseven percent of participants agreed that knowing their CO2 emissions would be
necessary for them to improve their driving style.
Table 7: Private Car Driver’s Awareness of CO2 Emissions
Understand ‘100 grams CO2
emission per km’
Motivated to improve FE if understanding of CO2 improved
Relevance providing CO2 emissions feedback to improve FE
Yes
No
46%
54%
71%
29%
67%
23%
50
Section three of the questionnaire asked participants to order different types of
feedback technologies from most to least preferred (see Figure 20). The technologies to rank were three on-board devices, two smartphone applications, and one application running via the internet. The objective of this line of questioning was to
obtain a better understanding of which feedback technologies were preferred.
Least preferable
Most preferable
Color-changing screen 5%10% 14%
Leaves
Navigation system
Silver ball iPhone app.
CO2 emissions iPhone app.
13% 14%
16%
11%
16%
PC application
Color-changing
screen
Leaves
Navigation System
Silver ball iPhone app.
CO2 emissions
iPhone app.
PC application
24%
16%
15% 12%
23%
19%
21%
21%
12%
22%
19%
20%
40%
32%
18%
20%
37%
0%
Items
16%
16%
19%
18%
24%
18% 11%
16% 8%
16% 11% 9% 8%
60%
80%
100%
Mode
Mean
Std. Dev.
Deviation
from Scale
Mean
Sig.
(2-tailed)
6
4.40
1.52
0.90
.000***
5
3.74
1.67
0.24
.085
6
3.71
1.77
0.21
.136
2
3.41
1.56
-0.09
.500
3
3.24
1.54
-0.26
.040
1
2.60
1.64
-0.90
.000***
Note: * p < 0.05; ** p < 0.01; *** p < 0.001
Figure 20: Preferences for Feedback Technologies
The most favored feedback system was a color-changing dashboard (mean rating:
4.4 out of 6) that showed the fuel consumption according to the driving style; the
51
deviation from the scale mean showed statistical significance (p < 0.05), with 72%
of respondents preferring it (four or more scale points). The least important was
found to be the PC Application (mean rating: 2.6 out of 6), with 72% giving it lower
preference (three or fewer scale points). This feedback technology provided indirect
feedback about an individual’s CO2 emissions after the driver manually entered
their fuel consumption into their PC. Other feedback technologies did not show a
statistically significant outcome (p > 0.05), thus not eliciting a clear preference from
the participants.
Open-ended questions were utilized to explore participants’ reasoning behind their preferences for feedback systems. Once the coding process was completed and relevant influential factors were determined, it became clear that a color-changing
dashboard was the most favored because this enhanced drivers’ ease of interpreting the feedback they received (see Figure 21). Using three different colors to illustrate
their driving style was considered to be the most straightforward and non-distracting
Question: What were the reasons you liked the
Colour Changing Dashboard the most/least?
method for presenting the necessary information to drivers.
Easy to understand (34)
No distraction (11)
Easy to track (8)
Attractive design (6)
Most
Relation to driving behavior (4)
Real-time feedback (2)
Lack of information (3)
Too basic (3)
Least
Figure 21: Reasons for Favoring the Color-changing Feedback System
A PC application was considered to be least favored because of the timing of the
feedback (after the event); the application was also found to be unattractive and too
complicated in design (see Figure 22). Drivers preferred real-time feedback, as they
immediately received direct feedback regarding their driving behavior.
Question: What were the reasons you liked the PC
Application the most/least?
52
Easy to understand (4)
Attractive design (2)
Feedback type (2)
Most
Relation to CO2 savings (2)
User friendly (2)
No real-time feedback (18)
Unattractive design (17)
Complicated (12)
Distraction (4)
Least
No useful feedback (3)
Inappropriate concept (2)
Not user friendly (2)
Figure 22: Reasons for Not Favoring the PC Application Feedback System
The next open-ended question asked participants’ about their views on which feedback types would be beneficial to display on an eco-feedback device in order to assist them to drive more sustainably. Since this question was not mandatory, fewer
answers were given than to other questions. Unstructured answers were coded, defined into factors, and grouped into categories reflecting similar content or meaning.
Several themes were apparent (see Figure 23). Appearance was found to be important, as user-friendly feedback should not distract the driver during driving. Additional information that participants found to be useful was more detailed material,
such as long-term data on fuel efficiency. This information would help participants
acknowledge the financial benefits of sustainable driving in addition to its environmental impact.
Question: Do you have any further ideas to display
information about eco-efficient driving?
53
No distraction (9)
User friendly (3)
Appearance
More efficient information display (2)
More information (6)
Focus on fuel efficiency (4)
Long-term data acquisition (2)
Financial related (8)
Environmental related (7)
Gear shift recommendation (2)
Visual support (2)
Information
provisioning
Visualization
Support
Figure 23: Suggestions for Additional Eco-efficient Feedback
III.3.3 Findings from Survey Conducted with Corporate Car Drivers
The second survey was completed by 131 (29%) of the 450 corporate car drivers
working for the case study company. The survey ran in August 2011 for three
weeks, prior to the field tests. The demographic questions revealed that 77% of respondents used their car every day or at least three to four times per week. Seven
percent did not know their average monthly distance driven, 27% drove between
2,001 and 3,000 kilometers, and 16% above 3000 kilometers per month (see Figure
24), which is consistent with the average yearly total distance driven by corporate
car drivers in Europe (DTLR, 2001).
54
Figure 24: Average Kilometers driven per Month
As corporate car drivers have their fuel costs reimbursed by their company, 26% of
drivers did not know their monthly fuel costs, while 37% estimated this cost to be
between 201 and 400 CHF and 25% estimated it to be between 0 and 200 CHF.
Respondents rated the importance of receiving a certain type of feedback for improving fuel efficiency using a five-point likert scale, and results are shown in Figure 25.
1 Strongly Disagree
2
3
4
5 Strongly Agree
Personalized eco-driving tips 6%7% 22%
28%
14% 5%12%
27%
18% 13%
25%
AVG km driven (monthly)
23%
25%
0%
37%
42%
28%
14% 16%
20%
18%
40%
16%
31%
21%
60%
17%
18%
19%
80%
100%
Scenario
Personalized ecodriving tips
AVG fuel efficiency
(monthly)
AVG CO2 emission
(monthly)
AVG km driven
(monthly)
AVG fuel savings
(monthly)
55
Mode
Mean
Std. Dev.
Deviation
from Scale
Mean
5
3.83
1.18
0.83
.000***
5
3.77
1.41
0.77
.000***
4
3.11
1.32
0.11
.319
4
3.06
1.43
0.06
.603
1
2.89
1.45
-0.11
.364
Sig.
(2-tailed)
Note: * p < 0.05; ** p < 0.01; *** p < 0.001
Figure 25: Preference of Feedback Types to improve Fuel Efficiency
Drivers most preferred receiving tips to improve their driving habits, with a mean
value of 3.83 out of 5, 65% agreement (four or more scale points), and a statistically
significant deviation from the scale mean (p < 0.05). Receiving information about
their average fuel efficiency was a close second (3.77 out of 5), with 69% agreement. As fuel costs were reimbursed, drivers considered feedback regarding their
average monthly fuel costs to be least useful (2.89 out of 5), though this result is not
statistically significant.
When asked if they understood what was meant by ‘100 grams of CO2 per km,’ 74% of respondents answered ‘yes,’ but 37% were unable to translate the value into a feasible context (see Figure 26). Twenty-seven percent did not understand this
value at all, and 60% of these drivers said they would have a better understanding if
a vivid comparison was given (e.g., ‘100 grams of CO2 per km is similar to using a
computer for 4.5 hours.’) Finally, only 53% believed that knowing their CO 2 emissions figures was influential in improving their average fuel efficiency.
56
Figure 26: Understanding of the Meaning of 100 grams of CO2 per km
III.4 Discussion
Only 13% of the private car drivers participating in the study had knowledge of the
emerging eco-feedback smartphone applications available on the market. This
aligns with the fact that smartphone device penetration was only 19% in 2010,
though this is expected to rise to 40% by 2014 (Gartner, 2011). Better-known feedback technologies included navigation systems (35%), trip computers (25%), and
on-board gauges (23%).
A strong preference for undisruptive feedback technologies was found, which confirms the findings of Meschtscherjakov et al. (2009). The on-board unit that changes
color from blue to green according to the driving style was most popular, as it was
efficient and effective at indicating drivers driving style in a real-time setting. The
PC application was least favored due to the offline nature of its feedback and its
complicated design format, implying that drivers like receiving direct feedback
when driving.
As shown in Figure 27, private car drivers preferred receiving information about
their personal fuel consumption (4.47 out of 5) and average fuel savings (4.02 out of
5) to receiving personalized tips for a more ecological driving style (3.86 out of 5).
In comparison, corporate car drivers preferred to receive tips that helped them improve their driving style to become more eco-friendly (3.83 out of 5), closely followed by information about the average fuel efficiency (3.77 out of 5). The least
interesting information for this user group was, not surprisingly, the average month-
57
ly total cost of petrol (2.89 out of 5). This is in line with the existing literature: financial savings are important reasons for private car drivers to change their driving
behavior (Graham et al., 2011), whereas corporate car drivers are reimbursed by
their company and therefore consider financial savings less important.
5,00
4,47
4,02
3.77
4,00
3,86 3,83
2,89
3,00
2,00
1,00
0,00
Personal fuel
consumption
Financial savings
Private Car Drivers
Items
Private
Car
Drivers
Corp.
Car
Drivers
AVG fuel efficiency
(monthly)
AVG fuel savings
(monthly)
AVG fuel efficiency
(monthly)
AVG fuel savings
(monthly)
Corpor. Car Drivers
Mode
Mean
Std.
Dev.
Deviation
from Scale
Mean
Sig. (2tailed)
5
4.47
0.86
1.47
.000***
5
4.02
1.77
1.02
.000***
5
3.86
1.61
0.86
.000***
5
3.77
1.41
0.80
.000***
1
2.89
1.45
-0.11
.364
5
3.83
1.18
0.83
.000***
Figure 27: Preferences for Feedback Types, Private vs. Corporate Car Drivers
Both groups agreed that information about their CO2 emissions was less important;
private and corporate car drivers ranked feedback about CO2 emissions at 3.39 out
of 5 and 3.11 out of 5, respectively. Limited understanding of the meaning of CO2
58
emission figures may be one reason why participants did not value this information.
This is supported by the fact that only 46% of private car drivers understood the
meaning of the CO2 emissions figures given, compared to 74% of corporate car
drivers. Of the latter group, though, only 37% were able to relate it to their own fuel
efficiency.
III.5 Conclusion
This chapter’s findings provided a first insight into the potential impact feedback technologies could have on driving behavior. They also identified the ideal feedback types to be given using feedback technologies and to be tested during the field
tests. Both driver groups preferred to receive information about their actual fuel
consumption. Details related to financial savings were relevant for private car drivers, but CO2 emission figures were not as important due to drivers’ difficulties in relating this information to their driving styles. Consistent with findings from Graham et al. (2011), feedback about financial savings was more relevant for private
car drivers than was information about CO2 savings. Further education is required to
raise drivers’ understanding of CO2 emissions as they relate to fuel consumption
figures, enabling drivers to relate this meaningfully to energy efficiency. For example, CO2 emissions can be compared to those resulting from the electricity consumption of a laptop. Based on these findings, it can be argued that eco-feedback
technologies have the potential to change driving behavior, particularly when the
feedback is real-time, unobtrusive, and easy to understand. Offline feedback (i.e.
after driving) has the advantage of providing detailed data about driving behavior
but must be presented in an easy-to understand way.
III.6 Limitations and Future Research
The first survey with private car drivers was conducted in 2010, whereas the survey
with corporate car drivers was conducted in 2011. To make an ideal comparison of
the driver groups, it would have been better to conduct the surveys at the same time.
However, these preliminary findings were still valuable and could be incorporated
into the design of the field tests. Graham et al. (2011) found that showing CO2 saving figures to private car drivers had a positive impact on reducing their car usage.
The preferred feedback types identified in these surveys will be used in the field
59
tests to attempt to alter corporate car drivers’ driving behavior, as explained in the following chapters. Whether corporate car drivers improve their fuel efficiency
needs to be further evaluated; this is especially important as the use of smartphone
technologies is continuously rising (Gartner, 2010; Morgan Stanley, 2010).
The next chapter, Chapter IV, evaluates how direct/ real-time feedback, provided by
an eco-driving smartphone application, influences the fuel efficiency of corporate
car drivers. Chapter V then explains how indirect feedback, provided via a monthly
fuel consumption email, impacts corporate car drivers’ fuel consumption.
Direct Feedback by an Eco-driving Smartphone Application
61
IV Direct Feedback by an Eco-driving Smartphone Application25, 26
IV.1 Overview
The personal transport sector constitutes an important target of energy conservation
and emission reduction programs. In this context, eco-feedback technologies that
provide information on the driving behavior have shown to be an effective means to
stimulate changes in driving in favor of both, reduced costs and environmental impact. This field test extends the literature on eco-feedback technologies as it demonstrates that a smartphone application can improve fuel efficiency even under conditions where monetary incentives are not given, i.e. where the drivers do not pay for
fuel.
The field test, which took place with 50 corporate car drivers, demonstrates an improvement in the overall fuel efficiency by 3.23%. The theoretical contribution underlines the assumption that direct/ real-time feedback provided by an eco-driving
smartphone application can favorably influence behavior even without direct financial benefits for the agent. Receiving real-time feedback was also preferred compared to indirect feedback (accumulated over time or offline feedback). Given the
large share of corporate cars, findings are also of high practical importance and motivate future research on eco-driving feedback technologies.
25
Published paper: Tulusan, J., Staake, T., Fleisch, E., Providing eco-driving feedback to corporate car drivers: what impact does a smartphone application have on their fuel efficiency, 14th ACM International Conference on Ubiquitous Computing (UbiComp), Pittsburgh, Pennsylvania, United States, September 2012.
26
62
IV.2 State of the Art and Related Work
IV.2.1 Feedback Types27
As mentioned in Chapter 0, feedback geared towards promoting sustainable driving
can be presented using various mediums, such as: email, brochures, classroom training, driving simulator, videos, online portals, and mobile devices. It is important to
consider when and what type of feedback is required to ensure the success of an
eco-driving intervention. Direct feedback should not endanger the driver’s safety (Donmez, 2007). Distraction is a crucial factor which must be considered because
this can lead to increased risk levels and even fuel efficiency (Meschtscherjakov et
al., 2009). Indirect feedback, on the other hand, can be proclaimed to be ‘richer’. For instance, feedback presenting electricity consumption for households includes
data reflecting the actual consumption shown on tables, graphs, charts, or scales
(Graml et al., 2010).
In the following section, a list of feedback types relevant for this research will be
introduced and should instill a fundamental understanding about how and when ecodriving feedback can be utilized to motivate drivers to reduce energy consumption.
Context-sensitive feedback is defined by Dey et al. (2001, p. 108) as “any information that can be used to characterize the situation of entities (e.g. person, place,
or object) that are considered relevant to the interaction between a user and an application, including the user and the application themselves”. Arroyo et al. (2006)
used context-sensitive feedback with a sensor enabled logging device to remind individuals to appropriate driving techniques to reduce their stress and improve safety. Studies by Meschtscherjakov et al. (2009), Hutton et al. (2001), and Smiley et al.
(1989) have demonstrated an impact on drivers’ performance when providing feedback about their driving ability through a support system. Consequently, they found
that the driving performance can be enhanced by providing feedback in relation to
the specific context.
27
63
Direct or real-time feedback also known in the literature as ‘just-in-time’ feedback provides immediate feedback within a certain situation (Fogg, 2002), and has been
applied in various fields to motivate behavioral change, e.g. encourage seat belt use
(Geller et al., 1982) and to reduce electricity consumption (Graml et al., 2010; Winkler and Winett, 1982). With the emergence of sensor-enabled mobile devices and
their capabilities of collecting data through GPS and acceleration sensors, it is now
possible to provide context-sensitive just-in-time feedback to drivers (Froehlich et
al., 2010). It highlights various driving habits, e.g. aggressive or smooth acceleration and braking, which is immediately illustrated through screen color changes or
by an fuel/ CO2 emissions consumption scores (Froehlich et al., 2010). One must
consider the vital notion of ensuring that just-in-time feedback does not compromise
the safety of its user by increasing their cognitive workload and distracting their
attention from the act of driving (Donmez, 2007). As driving conditions have the
tendency to change rapidly, feedback can impede on the driver’s ability to respond to the task demands and impair their task performance (Arroyo et al., 2006). This
must be taken into consideration when providing eco-driving real-time feedback to
corporate car drivers.
Information providing indirect feedback or not in real-time is collected over a duration of time, which may range from several minutes to one or more driving cycles,
affording drivers an insight into their general driving style and how this may evolve
over time. It can be shown in different forms, e.g. in an accumulated form or offline
(Graml et al., 2010). Accumulated feedback can be imparted using either fuel consumption figures or using schematics, e.g. leaves growing (Formosa, 2009). The
idea of rewards is conceptualized by accumulated feedback, as the more efficient
the driver evolves the more advanced the figures or schemas become. Examples of
systems that can be attributed to this category are Ford’s SmartGauge (Formosa,
2009) and Honda’s EcoGuide (Froehlich et al., 2010). A prototype sensor enabled
mobile tool, evaluated by (Froehlich et al., 2009), illustrated that users were capable
of positively perceiving accumulated feedback regarding green transportation practices. Acknowledgment can be given to the rewarding nature of advanced technologies that inspire users to adapt their behavior and continue to reduce their energy
consumption (Froehlich et al., 2010). Offline feedback is a decontextualized type of
feedback;; e.g. a detailed historical breakdown of driver’s fuel consumption, CO 2
64
emissions, acceleration, braking, and gear shifting patterns is formulated once their
general driving behavior is monitored and analyzed. Evidently, feedback is not provided during the act of driving and so can be combined with social networks, where
challenges from within the online community could potentially improve behavior.
This form of feedback is known as social normative feedback, as defined by
Cialdini (2003), and applied to change environmental behavior contributing to littering, towel reuse (Goldstein et al., 2008), and energy consumption (Siero et al.,
1996). Fiat’s eco:Drive is an example of where this type of feedback is used. Drivers are able to compare their own fuel efficiency with their peers on Fiat’s eco:Drive online community.
Indirect feedback can also be provided prior to driving, as this category includes
systems aimed at informing the driver about the planned route. This information is
largely delivered in the form of navigation systems or internet websites through
which multiple routes are presented to drivers regarding distance, travel time, or
fuel optimization. This enables drivers to make an informed decision about the best
possible route to meets their situational demands.
IV.2.2 DriveGain Eco-Driving Smartphone Application
An eco-driving smartphone application called DriveGain, a product of DriveGain
Ltd. in UK (DriveGain Ltd., 2012), was used for the purpose of the first field test.
After extensive evaluation of the aforementioned products and the feedback received from the user study introduced in Chapter III, this application was chosen
because of the quality of eco-driving feedback and access to the data collected by
the application from the drivers participating in the field test. Besides a very comprehensive car model database of more than 25,000 vehicle types, the application
provides different feedback types related to the eco-driving concepts, such as correct gear change during acceleration and braking, and most efficient average speed
depending for each vehicle type. Similar feedback types were also tested by van der
Voort et al. (2001), they demonstrated that drivers were able to reduce their fuel
consumption by 16% when using the support tool. It is important to acknowledge
that this test was confined to a driving simulator. Feedback types identified to reduce fuel consumption were correct gear changing during acceleration and smooth
acceleration. These feedback types were also implemented within the DriveGain
65
application. The DriveGain application runs on the Apple operating system (iOS),
therefore, only iPhone models 3G, 3Gs, 4, 4s and 5 are compatible. These models
have a GPS and acceleration sensor that enable the DriveGain application to obtain
data from these components.
Journey Score
Recommended
Gear
Type of Vehicle
Feedback Meter Types
Application Settings
Figure 28: Overview of DriveGain Application Interface 28
The top third of the interface screen (see Figure 28) illustrates: optimal gear change
(recommended gear), journey score, and type of vehicle. Feedback meters are located in the central third of the screen, and below this, functions to activate music, reset the journey scores, upgrade with new feedback meters and settings of the application. Once the application has been started, the car type chosen (see Figure 29 and
Figure 30) and a GPS signal received, it will log and monitor the driving behavior
for this specific route. The application provides several feedback meter types which
provide individual feedback towards acceleration, braking, CO2 emissions, or costs
(see Appendix C). To focus on the eco-driving feedback concepts and to evaluate
the difference direct or indirect feedback can have on a driving behavior, only two
feedback meter types were activated for this experiment:
28
DriveGain Ltd. (2012), “DriveGain.” Retrieved September 19, 2012, from http://drivegain.com
66
Figure 29: Settings Screen
Figure 31: Advanced Savings Feedback
Meter (direct feedback)
29
Figure 30: Select Vehicle Screen
Figure 32: Fuel Savings Feedback
Meter (indirect feedback)29
67
The first known as ‘Advanced Savings’ (see Figure 31) provides direct/ real-time
feedback that combines acceleration, braking and the current vehicle speed. Aggressiveness or smooth acceleration together with braking is shown in an efficiency
score which is calculated in a percentage value (100% being best). Additionally, the
current speed is shown on a scale. The second meter, known as ‘Fuel Savings’ (see
Figure 32) provides indirect accumulated feedback in a three-minute interval for the
average figures for acceleration, braking, and vehicle speed. Acceleration, braking,
and vehicle speed is measured by the GPS receiver. All three feedback measurements are displayed on a scale categorized red to green (green being most ecological), as well as a numerical score from 0 (least) to 100 (representing most ecological).
The journey score visible on the top right corner of the screen, 0 (least) and 100
(most ecological), is calculated from data collected regarding acceleration, braking,
and speed values with respect to each car type. The recommended gear feature
prompts drivers with a manual gearbox when to change gear. Only two gears, ‘P’ for parking and ‘N’ for driving were represented for cars with automatic gearboxes.
Relevant data for this study was collected from this application: 1. Journey start and
end location (measured by the GPS receiver); 2. When and duration application was
used in seconds; 3. Distance travelled in meters per journey; 4. Duration each type
of feedback meter was used in seconds; 5. Values for acceleration, braking and vehicle speed.
Besides the eco-driving information displayed on the application, additional data
sets were recorded by the application and automatically transmitted to the company's server after each journey. This data were accessible for each driver on an online
driving portal (ODP). This information was ‘richer’ in detail, as the raw data were evaluated using an information system and made visible to each individual user on
their ODP account. An overview cockpit summarized recent journeys driven, fuel
usage, fuel efficiency, high speeds, and CO2 emissions on the first page (see Figure
33). Further details were broken down on a dedicated sub-page for each section.
This allowed appraisal of offline feedback during the experiment, as every driver
had to register as a user to the ODP where they were then able to review their information after any driving episode.
68
Figure 33: Detailed Journey View shown in Online Driving Portal 30
IV.3 Research Questions and Hypothesis
The literature indicates a discrepancy about the impact an eco-feedback technology
has on driving behavior. Generally speaking, the approval of numerous in-vehicle
on-board feedback systems is largely reflected on in a positive manner (Boriboonsomsin et al., 2010; Meschtscherjakov et al., 2009; H. Lee et al., 2010; S. Siero et
al., 1989). Meschtscherjakov et al. (2009) highlighted that inflated risk levels can be
induced by distraction through inappropriate feedback. This was further illustrated
by Lee et al. (2010) who found that drivers, as a result of amplified cognitive workload and stress levels, ignored feedback provided by a non-user-friendly design or a
poorly positioned system on their dashboard during a field test.
30
69
To understand if an eco-driving smartphone application has an impact on the fuel
efficiency, the following hypothesis was defined to answer sub-research question 2.
Hypothesis 1: Corporate car drivers who use an eco-driving smartphone application
improve their average overall fuel efficiency, with µ1 representing the liters of petrol per 100 kilometer without and µ2 with eco-driving feedback.
(H1) H0 = µ1 ≤ µ2;; H1 = µ1 > µ2.
Real-time and context related feedback is integral when promoting fuel efficiency
of individuals when driving. The ability to control the frequency of the feedback
received has been proven to motivate users to change their behavior. This is known
as “feedback control,” e.g. individuals are able to control the amount and timing of the feedback they receive (Corbett and Anderson, 2001). The DriveGain application
allows corporate car drivers to choose between different feedback types, such as
‘Advanced Savings’ with direct feedback that provides real-time feedback about
one’s actual driving style and ‘Fuel Savings’ with a three-minute interval feedback
that provides the feedback in an accumulated format. The second hypothesis was
defined to answer sub-research question 3.
SQ3: Which feedback type, direct or indirect, do corporate car drivers prefer when
using an eco-driving smartphone application?
use real-time feedback for longer periods of time when compared to accumulated
feedback, with µ1 representing average time in seconds regarding real-time feedback and µ2 representing average time in seconds of which accumulated feedback
was used.
(H2) H0 = µ1 ≤ µ2;; H1 = µ1 > µ2.
Drivers were also able to review their eco-driving data offline, known as indirect
feedback, and summarized in the ODP. Literature in the field of energy savings in
70
households reflect that feedback related to energy consumption should also be provided immediately using a non-obstructive form (Van Houwelingen and Van Raaij,
1989; Mountain, 2008). This enables individuals to relate feedback directly to a
specific behavior and prompts an immediate response (Darby, 2006; Graml et al.,
2010). The importance of indirect feedback in relation to eco-driving is not clear
and so hypothesis three was defined to get a better understanding to appraise SQ3.
prefer to receive feedback during driving in comparison to offline feedback provided via a ODP, with µ1 representing the preference to receive feedback during driving through the smartphone application and µ2 representing the preference of receiving the feedback via a ODP.
(H3) H0 = µ1 ≤ µ2;; H1 = µ1 > µ2.
IV.4 Research Design and Research Methodology
IV.4.1 Research Design
It was deemed appropriate to use surveys and a field experiment for this experiment
in order to review drivers’ interaction with technology. This research design was founded on a sequential multi-method data collection approach that incorporated an
assortment of approaches within a sequential order, permitting data from one level
to be utilized in the next (Mingers, 2001). Before and after the field test, drivers
were requested to complete a survey that allows monitoring and control of factors
that relate to individual differences, such as: driving habits, environmental believes,
and preference of feedback types. The opt-in experiment was announced in August
2011 by email by one board member of the case study company to 450 corporate
car drivers based in Switzerland (see Figure 34).
71
Figure 34: Between-subjects Experimental Design
Sixty-two employees responded to the announcement email. The criterion to participate was ownership of an iPhone model 3G, 3Gs, 4 or 4s. Seven drivers with other
phones had to be removed from the sample. Tank-refill data of the remaining 50
drivers were available prior to and during the introduction of the field test for each
driver. This made it possible to calculate the drivers’ average baseline fuel efficiency from January to October 2011 (prior to implementing the experimental condition). Once the average fuel efficiency (= baseline) was calculated, a random selection of 25 drivers for the Control Group (CG) and 25 for the Treatment Group (TG)
72
took place. The normality of the distribution was met, as shown by the not significant outcome of the Shapiro-Wilk test for normality (with P > .05, skeweness = .579 and kurtosis = 1.739). Participants in the TG used the eco-driving application
for a duration of eight weeks from 24th of October to 16th of December 2011. As it
would be in a non-study environment, it was left to the drivers in TG to decide how
often and when the application would be used; however, they were informed that
the application should be used on a regular weekly basis within the designated eight
weeks.
The drivers could choose between the two different feedback types, ‘Advanced Savings’ with direct/ real-time feedback and ‘Fuel Savings’ with indirect/ three-minute
interval feedback. At the end of the experiment the recorded data from the application were evaluated (see Appendix D). Data were transmitted through the GSM
network directly from the smartphone to the DriveGain Ltd. server after each recorded route and provided to the researcher after the experiment (see Appendix E).
Participant’s agreement to automatically transmit the collected data for each route
was mandatory to participate in this field test. This it made it possible to verify,
without self-reporting errors, the frequency-use of the application together with the
duration each type of feedback meter was used by each driver.
IV.4.2 Research Methodology
The pre-survey (n = 131) conducted in August 2011, gathered details about the general understanding of drivers’ fuel efficiency, CO2 emissions, and transportation
modes (see Appendix B). The post-survey (n = 24) conducted in March 2012, inquired about the environmental attitude of the driver, eco-driving feedback usefulness of the eco-driving application, feedback types, usage patterns, and recommendations for improvement (see Appendix F). To get a richer understanding, semistructured interviews with 15 drivers were conducted from March to April 2012
(see Appendix G), which investigated details about changes in usage patterns, i.e.
how often and how long the DriveGain application was utilized by each driver
throughout the eight weeks. Details about the relevance of feedback types, e.g. realtime vs. accumulated vs. offline feedback and the combination of a smartphone application with an ODP were also explored in-depth. Each interview lasted 20
minutes.
73
Besides the qualitative approach, quantitative findings were based on the data collected from the DriveGain application and fuel efficiency data collected from the
petrol cards of each corporate car driver. Drivers’ monthly tank-refill details were
collected from petrol credit cards provided by the company. For each tank-refill, the
total number of kilometers driven and volume of petrol were recorded during the
payment process as a normal prerequisite to get permission to refuel on the company account (a measure to prevent drivers from refueling other cars at the company’s expense). Tank-refill data were sent to the company through a mobility operator.
This allowed the researcher to calculate and compare the fuel efficiency per tankrefill more objectively, as drivers only had to record the total number of kilometers
shown on their in-vehicle system at the time of payment. The total number of liters
required to calculate the fuel efficiency were automatically reported by the system
from the petrol station, thus, eliminating the need for further self-reported input
from drivers. At the end of the field test more than 7000 fuel fillings from the Control (n = 25) and Treatment Group (n = 25) were recorded and more than 800 journey data sets collected from the TG’s application usage analyzed.
The between subjects statistical evaluation for comparing the means of fuel efficiency between the CG (without use of smartphone application) and TG (with use
of smartphone application) was an independent samples t-test. A bivariate Pearson
correlation coefficient was computed to identify the relationship between the total
duration of application usage and duration of direct or indirect feedback type (’Advanced Savings’ or ‘Fuel Savings’).
IV.5 Data Evaluation and Findings
IV.5.1 Descriptive Evaluation
The descriptive evaluation of the survey (see Table 8) shows that 75% of the participants were male and 25% female. Twenty nine percent were aged between 45 and
49 years, 25% between 40 and 44, and 17% between 30 and 34, which shows a bias
towards people above 40 years. Seventy five percent of participants used their car
every day, 21% five to six times per week, and 92% drove on average more than
74
24,000 kilometers per year, which conforms with the corporate car drivers criteria
(Department for Transport, 2011).
All drivers used the eco-driving application on a weekly basis; 37% used it seven to
eight times, 25% three to four times, and 17% five to six times. When asking which
routes and travel purpose the application was used for, of the 119 answers (multiple
selections were possible) 95% used the application on motorways and business related travel and 92% for distances longer than 40 kilometers. Fifty percent used the
application for private travel and 46% for short distances.
Table 8: Sample Demographics
Variable
Gender
Age
Years of
Experience
Car Usage
Category
Frequency
In %
Female
6
25
Male
18
75
Under 30
1
4
30-34
4
17
40-44
6
25
45-49
7
29
50-54
3
13
55-59
3
13
0-9
0
0
10-14
5
21
15-19
1
4
20-24
6
25
25-30
8
33
More than 30 years
4
17
Less than once a month
0
0
1-2x per month
0
0
Every week
0
0
1-2x per week
0
0
3-4x per week
1
4
5-6x per week
5
21
Every day
18
75
Mode
Male
45-49
25-30
Every day
Variable
Total kilometers
per year
Usage of
application
per week
Routes/
Travel
Purposes
Category
75
Frequency
In %
Less than 6000 km
0
0
Between 6000 – 12.000 km
0
0
Between 12.000 – 18.000 km
0
0
Between 18.000 – 24.000 km
2
8
Between 24.000 – 30.000 km
4
17
Between 30.000 – 36.000 km
10
42
More than 36.000 km per year
8
33
Never
0
0
1-2x
2
8
3-4x
6
25
5-6x
4
17
7-8x
9
38
9-10x
2
8
More than 10x
1
4
Business related travel
23
96%
Motorway
23
96%
For long distances (more than
40 km)
22
92%
Country / rural road
18
75%
Private related travel
12
50%
For short distances (less than 40
km)
11
46%
Inner cities
10
42%
Mode
Between
30.000 –
36.000 km
7-8x
Business
related travel; Motorway
The next question asked participants about the influential aspects of the eco-driving
smartphone application regarding their driving behavior on a 7-point likert scale,
one representing least and seven most favorable (see Figure 35).
76
Question: The feedback given by the application influenced me in…
1 Strongly Disagree
…shifting gears earlier.
2
3
10%
4
30%
17%
…smoother braking. 4% 8% 8% 4%
27%
...reducing average speed
on motorways.
7 Strongly Agree
25%
18%
18%
0,2
36%
13% 4%
0,4
0,6
21%
0,8
4%
1
7
7
6
5.45
5.13
4.83
1.75
1.90
1.52
Deviation
from Scale
Mean
1.45
1.13
0.83
6
4.36
1.69
0.36
.492
2
2.83
1.55
-1.17
.001**
Mode Mean Std. Dev.
…shifting gears earlier.
…a smoother acceleration.
…smoother braking.
…driving in the highest gear.
...reducing average speed on
motorways.
29%
46%
42%
0
40%
29%
17%
Items
6
10% 10%
…a smoother 4% 13% 4% 8%
acceleration.
…driving in the highest gear.
5
Sig.
(2-tailed)
.020*
.008**
.013*
Note: * p < 0.05; ** p < 0.01; *** p < 0.001
Figure 35: Influence on Driving Behavior
Sixty percent of drivers with a manual gearshift felt that the feedback given by the
application influenced them to shift the gear earlier (5.45 out of 7). Seventy one
percent of all drivers accelerated smoother (5.13) and 75% decelerated smoother
(4.83). All items were also statistically significant (p < 0.05). The least influential
factor was to reduce the average speed on motorways (2.83) with 72% not in favor
(three or less scale points), which could relate to the absence of speeds limit on
German motorways.
77
IV.5.2 Statistical Evaluation
The initial findings of the demographical evaluation provided an overview of the
sample size and an understanding about how the eco-driving application promoted
eco-driving techniques. The following chapter answers the sub-questions SQ2 and
SQ3 with a statistical evaluation using the software packages STATA and SPSS.
The first step included the identification of outliers. Five drivers were identified as
outliers using the Grubb’s test due to very low or excessively high baseline consumption and so were removed from the sample. As each participant’s data reflecting each tank filling from the beginning of the year were available, and to control
for seasonal effects, the fuel efficiency of each driver was calculated for each tank
filling from the start of January 2011 until the treatment started, and during the
treatment phase. Once the average fuel efficiency was calculated for these two
timeframes and changes in percentage calculated, a paired samples t-test comparison between participants’ (CG vs. TG) changes in fuel efficiency was made. (H1) Before the t-test was conducted three outliers in the CG and two outliers in the
TG in terms of atypical changes of consumption due to false entries of kilometers
driven during the experiment phase were identified using Grubb’s outlier procedure. Three additional data sets were removed in the TG to control for periods where the
application was not used. A total of CG = 22 and TG = 20 were used for the t-test.
The outcome showed that all participants’ fuel efficiency changed, be it positively
or negatively. The maximum improvement in the CG was -4.94%, and -12.15% in
the TG. The minimum change in the CG was 8.40% and 1.92% in the TG. On average the corporate car drivers in the TG had improved their overall fuel efficiency
more so by the end of the treatment phase (M = -2.2, SE = 0.87) compared with the
CG (M = 1.03, SE = 0.70). This difference between the means of CG and TG1 result in a 3.23% fuel efficiency improvement (see Table 9) and was statistically significant p < 0.01 with t(40) = 3.23, p = 0.006 (see Table 10). As the Levene’s test was not significant (p = 0.459) (see Table 11), the equality of variances can be assumed and the H0 rejected in favor of the H1.
78
Table 9: Group Statistics
N
Mean
Std. Deviation
Std. Error
Mean
0
22
1.034
3.291
.702
1
20
-2.202
3.875
.867
Id_trtmt_grp
diff_trtmt_avg_fe_y
Table 10: Independent Samples Test (t-test for Equality of Means)
t-test for Equality of Means
t
df
Sig.
(2-tailed)
Mean
Difference
Std. Error Difference
Diff_trtmt_avg_fe_y
Equal variances assumed
2.926
40
.006
3.236
1.106
Equal variances not assumed
2.902
37.495
.006
3.236
1.115
Table 11: Independent Samples Test (Levene’s Test for Equality of Variances)
Levene’s Test for Equality of Variance
F
Diff_trtmt_avg_fe_y Equal variances assumed
Sig.
.558
.459
(H2) The analysis of driver data sets, inclusive of 800 recorded journeys, were used
to compare the variations in duration-use of direct versus indirect feedback. It was
evident that the application was used mainly during the week and at daytime. Furthermore, an average distance per route of 33 kilometers was travelled with an average usage of the application at 28 minutes per route. The total average journey score
rating of 75 indicates a good ecological driving behavior.
The total time the application was used by all participants during the field test was
416 hours. Data collected revealed that the real-time feedback meter was used
85.11% (354 hours) of the time compared to the accumulated feedback with only
2.65% (11 hours). For the remaining 51 hours (12.24%) the application was sus-
79
pended due to no GPS signal, e.g. driving through a tunnel, which is quite common
in Switzerland.
A bivariate Pearson correlation coefficient was computed to identify the relationship between the total duration of application usage and duration of direct or indirect feedback type used. Findings showed a strong positive correlation with statistical significance between the total duration of application usage and usage of direct/
real-time feedback (Adv._Sav._Meter), r = .878, n = 759, p < .000. When comparing the total duration of application usage and usage of indirect/ accumulated feedback (Fuel_Sav._Meter), a weak correlation with no statistical significance (p >
0.05) was shown with r = .190, n = 95, p = .065 (see Table 12).
Table 12: Correlation of Direct or Indirect Feedback vs. Duration
Direct Feedback
(Adv._Sav._Meter)
Indirect Feedback
(Fuel_Sav._Meter)
score
distance
total_duration
Pearson
-.178
.800
.878
Correlation
Sig. (2-tailed)
.000
.000
.000
N
759
759
759
Pearson
.205
.205
.190
Correlation
Sig. (2-tailed)
.046
.046
.065
95
95
95
N
Qualitative interpretation of data collected via the online post-survey (24 answered
from the total of 25 drivers) were evaluated in percentage (1% least preference to
100% highest preference) or on a scale from one to seven. An additional 15 interviews were analyzed following an axial coding process to be able to compare the
rich answers.
The post-survey supports the findings from the bivariate Pearson correlation, as
88% of the participants preferred to use real-time (“Advanced savings”) feedback
compared to only 19% who opted to use accumulated (“Fuel savings”, every three minutes) feedback (see Figure 36).
80
Question: Which type of feedback did you choose to receive?
1 Strongly Disagree
“Advanced savings” (real-time feedback)
2
3
12%
Real-time feedback
6% 11%
was useful.
“Fuel Savings” (every three minutes)
6
7 Strongly Agree
35%
17%
53%
11%
33%
22%
44%
44%
22%
Three minutes
interval…
“Advanced savings” (real-time feedback)
Real-time feedback
was useful.
“Fuel Savings” (every three minutes)
Used both feedback
meters evenly.
Three minutes interval feedback was useful.
5
38%
Used both feedback
meters evenly.
Items
4
13% 6%
11%
63%
13%
0,4
0,6
17%
13%
0,8
6%
13%
0
0,2
1
Mode
Mean
Std. Dev.
Deviation
from Scale
Mean
Sig. (2-tailed)
7
6.06
1.60
2.06
.000***
6
5.06
1.86
1.06
.028*
2
2.41
1.94
-1.59
.004*
1
2.22
1.48
-1.78
.000***
1
2.00
1.73
-2.00
.000***
Note: * p < 0.05; ** p < 0.01; *** p < 0.001
Figure 36: Direct – vs. Indirect Feedback
When considering the usefulness of the application in stimulating eco-friendly driving behaviors, real-time feedback ranked 5.06 and accumulated feedback (three
minutes interval feedback) ranked only 2.0. All values have a statistical significance
of p < 0.05.
The post-experimental interviews were valuable in attaining an insight into why
participants favored real-time feedback; the immediate nature of real-time feedback
81
enabled participants to directly link eco-driving principles to their driving practices.
Accumulated feedback lacked any significance with regards to usefulness when
driving on motorways, as very similar feedback was given by the application to
drivers with a constant average speed and when there was not much traffic.
With respect to these findings one is able to reject the H0 in favor of the H1; to conclude, real-time feedback was more eminent and favorable compared to accumulated feedback.
(H3) H2 was appraised using results from both the post-survey and interviews. The
post-survey indicates that 71% of participants reviewed their journey scores on the
ODP once or twice per week and only 12.5% three to four times (see Figure 37).
18
16
14
12
10
8
6
4
2
0
17 (71%)
3 (13%)
1-2x
3-4x
2 (8%)
Never
1 (4%)
1 (4%)
5-6x
7-8x
0 (0%)
0 (0%)
9-10x
More than
10x
N = 24; n (% of total N)
Figure 37: Frequency of Journey Score reviewed in the ODP
In fact, they preferred to receive real-time feedback during a driving episode (88%)
or in a weekly summary (63%) (see Figure 38). Multiple selections were possible
for this question. The multiple selections were further evaluated using a 7-point
likert scale question format, which asked drivers how thy preferred the feedback to
be provided (see Figure 39).
82
25
21 (88%)
20
15 (63%)
15
12 (50%) 11 (46%)
10
5
3 (13%)
3 (13%)
2 (8%)
Once
every 3
months
Once a
year
Once a
day
0
Real-time Once a
during
week
After
driving
Once a
month
n (%); N = 67
Figure 38: Preference of when Feedback should be received
Visualization (6.00) was most favorable with 88% agreement followed by weekly
emails (4.88) with 61% agreement. Both items showed a statistical significant value
of p < 0.05. Feedback through an online portal had the lowest mean with 3.21 and
63% disagreement, but, the statistical significance was above 0.05.
Question: I would like to receive feedback from the application about how I drive…
1 Strongly Disagree
2
3
…by visualisation 4% 8% 4%
…by a weekly email 25%
…by a computerized voice
29%
0
Items
22%
6
7 Strongly Agree
0,2
Mode
38%
13%
25%
33%
…by an online portal
5
46%
9% 4%4%
…by a computerized voice and…
4
8%
21%
0,4
Mean
39%
4%
13%
17%
13%
13%
0,6
Std. Dev.
9%
8% 8%
8%
8% 4%
21%
17%
0,8
Deviation
from Scale
Mean
13%
4%
8%
1
Sig.
(2-tailed)
…by visualization
…by a weekly email
…by a computerized voice and visualization
1.22
1.75
83
6
6
6.00
4.88
2.00
0.88
.000***
.023*
2
1
1
3.33
2.14
-0.67
.141
3.33
2.12
-0.67
.137
3.21
2.15
-0.79
.084
Note: * p < 0.05; ** p < 0.01; *** p < 0.001
Figure 39: Preference of how Feedback should be provided
The interviews clearly identified that participants felt the ODP was beneficial in
getting a summary of the journeys driven, supporting an improved understanding of
driving behavior/ patterns, but did not play an influential role in modifying their
actual driving style. As the data collected during the post-survey can only be descriptive statistically validated and the interviews interpreted, this hypothesis cannot
be accepted or rejected. However, a trend towards receiving direct feedback through
visualization compared to the offline feedback through an ODP is reflected.
Besides these findings related to hypotheses (1) to (3), another question queried
about the preference to which feedback items the driving style should be compared
with (see Figure 40).
Question: My driving style should be compared with…
1 Strongly Disagree
2
3
4
5
…with the same car 8% 13%
model.
29%
…with same driving style.
13%
…from my organization. 0
13%
8% 4% 8% 13%
21%
38%
0,2
29%
63%
29%
17%
7 Strongly Agree
50%
...who drive a
4%4%4% 13%
similar amount of km.
...with the same
job role in organization.
6
0,4
38%
8% 8% 13% 8%
13% 8%
0,6
21%
0,8
4%
1
84
Items
Mode
Mean
Std. Dev.
…with the same car model.
...who drive a similar
amount of km.
...with the same job role in
organization.
…with same driving style.
…from my organization. 6
6.00
0.88
Deviation
from Scale
Mean
2.00
6
5.63
1.17
1.63
.000***
6
3.83
2.24
-0.17
.719
2
2
3.42
1.91
-0.58
.148
2.92
1.56
-1.08
.002*
Note: * p < 0.05; ** p < 0.01; *** p < 0.001
Sig.
(2-tailed)
.000***
Figure 40: Preference of Social Comparisons
A preference from drivers towards a comparison with the same car model (6.00)
with 92% agreement and similar number of kilometers (5.63) with 89% agreement
was identified. Both means were statistical significance (p < 0.05). Other feedback
items for a comparison with other drivers, such as the same job role in the organization or the same driving style, were not as important and did not show a statistical
significance (p > 0.05). The least important comparison item was drivers from the
same organization with 68% disagreement and a mean of 2.92 (p < 0.05).
IV.6 Discussion
Comparing the improvement of corporate car drivers’ fuel efficiency of 3.23% with
existing studies referred to in this thesis strengthens the underlying assumption that
smartphone technologies can have a positive impact in promoting the eco-driving
behavior of corporate car drivers. Studies with greater improvements in fuel efficiency where either conducted in a driving simulator (Van der Voort et al., 2001) or
with a company where the management set a fuel reduction target of 5% (Siero et
al, 1989). Existing studies did not evaluate improvements in corporate car drivers’ fuel efficiency where monetary incentives are not relevant. The data recorded by
this application provides important insights into driving behavior and, in line with
the body of knowledge, showed that a strong acceleration and shifting gears too late
contributes strongly to fuel consumption, which supports the findings of the experiment conducted in the driving simulator by Van der Voort et al. (2001).
A strong positive correlation between total application and direct/ real-time feedback usage underlines the assumption that real-time feedback was used more than
85
indirect/ accumulated feedback. Findings are in accordance with literature in the
field of energy savings, which also identified direct/ real-time feedback as being
favorable to participants (Darby, 2006; Graml et al., 2010). Explorative findings
from the interviews provides additional support, as influencing driving styles to incorporate more eco-friendly behaviors required immediate prompts when directly
involved in the act of driving. This was especially important for the more experienced drivers, e.g. corporate car drivers, as their driving habits are that much more
internalized (Stenner, 2009).
Existing findings, from the energy savings in household literature (Graml et al.,
2010), revealed that providing energy consumption feedback through an online portal made it possible to convince stakeholders to change their electricity consumption
habits, particularly when accompanied by social psychological concepts (Mountain,
2008). When considering sustainable driving, findings from the post-survey and
interviews indicate that the ODP has a role in promoting participants awareness and
understanding of eco-friendly driving concepts (e.g. details about acceleration/
braking per journey, total kilometers driven, or average speed), but do not strongly
influence behavioral modification to encompass eco-friendly driving styles.
It is apparent that the feedback source together with the frequency of feedback underpins the relevance sensor enabled services can have in shaping corporate car
drivers’ behavioral change towards eco-driving practices. When considering participants’ control over the level of feedback they received, e.g. frequently through realtime feedback or intermittently via accumulated/ offline feedback, existing findings
in organizations recommend that by bestowing this ownership on participants enhance their desire to modify their behavior at work (Corbett and Anderson, 2001).
Furthermore, if feedback from mobile feedback technology sources were given in
conjunction with face-to-face feedback, e.g. by management, it can be predicted that
a greater improvement in fuel efficiency may be found. This assumption is corroborated by Siero et al. (1989), who provided feedback to their postal lorry drivers via a
mobile device as well as face-to-face and achieved a long-term fuel efficiency improvement of 7.3%.
86
IV.7 Conclusion
The findings highlight how mobile feedback technologies, which provide context
related eco-driving feedback during driving, improve corporate car drivers’ fuel efficiency. An existing smartphone application from the company DriveGain Ltd.
(DriveGain Ltd., 2012) with eco-driving functionalities was selected and tested in
an experimental setting with one company’s pool of drivers. The improvement in
fuel efficiency of 3.23% with t(40) = 3.23, p = 0.006 demonstrates that sensor enabled mobile feedback technologies can play an important role in reducing a company’s overall CO2 emissions and petrol costs. Findings also strengthen existing literature, eco-feedback technologies support the notion that feedback provided through
sensor enabled mobile devices can empower their users to drive sustainably
(Boriboonsomsin et al., 2010; Siero et al., 1989). Our findings extend the current
research area, as participants in this study were not motivated by financial rewards.
This study appraised which type of feedback, direct or indirect feedback, was preferred by the users. The application provided direct/ real-time or indirect/ threeminute accumulated interval eco-driving feedback to participants. Additionally, participants were able to review further driving related data in greater detail using an
ODP. It is justified to conclude that direct/ real-time feedback provided during a
driving episode was largely preferred over indirect feedback delivered via accumulated or offline feedback. The importance of an ODP within this eco-driving context
appears to be limited, as participants indicated a greater preference to attaining direct feedback when driving; simply reviewing a summary of driving related figures
on the ODP was not enough to warrant a change in their driving style. The main
function of the ODP appeared to be in its summary of the driving routes available.
IV.8 Limitations and Future Research
Generalization of these findings may be limited due to the sample size of n = 50 and
short field study duration. However, other studies in this research domain have had
even smaller sample sizes (n < 20) (Froehlich et al., 2009) or the treatment period
was shorter (H. Lee et al., 2010) or based on a driving simulator (H. Lee et al.,
2010; van der Voort et al., 2001). Findings which entail a lager sample size were
mainly conducted with truck drivers and did not evaluate an eco-driving smartphone
87
application with different feedback meters in combination with ODP, rather a fixed
installed on-board system that constantly monitored the driving behavior. Therefore,
this field study provides a preliminary insight into how and when feedback should
be provided using an eco-driving application with sensor enabled feedback technology, within a corporate context to promote improvement of fuel efficiency. Future
studies with a larger sample size and longer duration should evaluate long-term effects how both feedback types combined, direct and indirect have an impact on
driving behavior compared to each individual approach.
As findings from qualitative research methods are more difficult to generalize, findings from the interviews, where n = 15, should be met with caution. Hence, the validity of H2 requires further investigation by means of an additional experiment,
which specifically evaluates the impact of indirect feedback provided via an online
portal vs. direct feedback provided during driving.
Participants for this study were selected from one company’s pool of drivers. As it was an opt-in field test, participants who chose to take part in the experiment may
have already had a pro-environmental attitude or were technological affine, exposing a degree of bias in the sample selection. It was possible to control this aspect by
analyzing their environmental beliefs and technological affinity using validated
scales from energy saving literature in the post-survey. Findings indicated that drivers only had moderate (= partly agree) pro-environmental attitude (5.02 out of 7.0)
as well technological affinity (5.1 out of 7.0); this reflects that results of this study
can be related to other corporate car drivers with a similar environmental attitude
and technological affinity.
Indirect Feedback by a Mobility Information System
89
V Indirect Feedback by a Mobility Information System31
V.1 Overview
The Green information systems (IS) discipline has the potential to support environmental sustainability in corporations by drawing employees’ attention to their energy consumption. A sample of 240 corporate car drivers from one company were
selected for the second field test; over a duration of three months their fuel efficiency was calculated using a mobility information system from which additional qualitative data was used to provide a detailed monthly email reflecting customized
feedback for each participant. The emails summarized driving related figures such
as total kilometers driven per month, CO2 emissions, and tips to incorporate ecofriendly driving practices. A comparison of their average monthly fuel efficiency
rating against that of their colleagues was made and illustrated using either a categorical or continual scale. This provided them with social normative feedback with
the desire to explore if this had a greater impact on reducing participants’ fuel consumption.
Results indicate that descriptive social normative feedback had a significant impact
in motivating participants to improve their fuel efficiency. However, a difference in
the overall fuel consumption savings between the groups who received the feedback
by a categorical - or continual feedback scale was not apparent. These findings are
vital in endorsing the importance of maximizing the potential use of feedback information systems in combination with social normative feedback to promote ecofriendly driving within the corporate driving world. Ultimately, the second experiment supports previous findings and the capacity for corporations to reduce their
CO2 footprint and fuel costs by improving their employees’ (i.e. car drivers) awareness of fuel consumption using a mobility information system to collect relevant
data, which can then be appraised against ecological concepts, and customized
feedback sent to each driver via a monthly email. Hence, this optimizes the attributes of Green IS technologies whilst keeping the need for corporate investment to a
minimal.
31
Published paper: Tulusan, J., Staake, T., Fleisch, E., Providing eco-driving feedback to corporate car drivers: what impact does a smartphone application have on their fuel efficiency, 14th ACM International Conference on Ubiquitous Computing (UbiComp), Pittsburgh, Pennsylvania, United States, September 2012.
90
V.2 State of the Art and Related Work
V.2.1 Green IS
Green IS proposes a medium through which environmental sustainability can be
advocated, whether this is addressed by improving individuals’ awareness of ecological issues or modifying current practices and patterns of human behavior. A recent swift away from Green IT towards Green IS has been made; this movement
involves reinforcing the business process that promotes ecological concepts (i.e.
supporting a sustainable supply chain), and providing an evidence-based decision
support system which facilitates organizational strategies to meet their CO2 reduction goals (Thambusamy and Salam, 2010; Watson et al., 2010). Many studies evaluate the conceptual and organization levels, but few appraise the application of
Green IS in enhancing employees’ motivation to be progressively energy efficient (Jenkin et al., 2011; Loock, T Staake et al., 2011). Studies by Froehlich et al.
(2009), Loock et al. (2011), Holmes (2007), and Mankoff et al. (2007), amongst
others, generated promising results that suggest Green IS can have an influential
role in modifying individuals’ behavior to reduce energy consumption. Fogg (2002) defined the term ‘Captology’, which refers to information communication technologies (ICT), such as websites and information systems combined with
behavioral changing concepts. He classified persuasive technologies as “any interactive computing system designed to change people’s attitudes or behaviors”
(Fogg, 2002, p. 1). Froehlich et al. (2010, p.1999) expanded on this concept and
characterized eco-feedback technologies as “technology that provides feedback on individual or group behaviors with a goal of reducing environmental impact”. Ecofeedback technologies have the potential to bridge the gap between individuals’ lack of environmental awareness and how their everyday behavior, such as driving to
work, impacts the environment;; this is termed ‘environmental literacy gap’ (Froehlich et al., 2010). When considering energy saving in households, Geller et al.
(1982) has shown that eco-feedback technologies have been effective strategies in
reducing energy consumption.
Despite the substantial body of literature in areas, such as energy savings in households using smart meters, and recently, online IS portals to motivate users to reduce
energy consumption (Graml et al., 2011; Loock, T Staake, et al., 2011), limited research proclaims the impact of IS in improving driving habits. Hence, one can only
91
speculate the potential success IS can have on modifying human behavior within the
energy saving context. Graham et al. (2011) adopted an approach to research in the
field by applying concepts from social psychology with Green IS. This provided
them with a strong theoretical platform from which to assess the effectiveness of an
on-line intervention in encouraging 128 college students to reduce their daily car
usage. Using a 2x2 experimental design they discovered that a combination of both
feedback types, representing drivers’ monetary savings and reduction in CO2 emissions, had the greatest bearing in reducing daily car usage, followed by feedback
regarding monetary savings alone. Environmental feedback in isolation had the least
impact. Participants were required to report their car usage details and savings in
miles, when using alternative modes of transport, on an online portal. This data enabled the researcher to provide personalized ecological feedback to the participants
using Green IS to stimulate behavioral modification.
V.2.2 Cognitive Intentions and Behavioral Change in Transportation
On reviewing the existing literature concerning behavioral change in transportation,
attempts have been made to apply theories from social psychology research to influence driving behavior, but with variable success. Social psychology theories,
such as the Rational Choice Theory (Simon, 1955), the Theory of Reasoned Action
(Ajzen and Fishbein, 1977), and the Theory of Planned Behavior (Ajzen, 1991),
were applied to evaluate travel mode choices of individuals (Bamberg et al., 2003;
Gardner and Abraham, 2010) or in analyzing interventions to reduce speeding
(Goldenbeld et al., 2008; Stead et al., 2005). One critique of the Rational Choice
Theory is that it assumes that individuals are able to comprehend their preferences
and therefore, able to make an informed decision having considered all their options. This assumption has been scorned by several authors, such as Baron (2008),
Tversky and Kahneman (1986), Schwartz et al. (2002). For instance, Schwartz et al.
(2002) argued “the assumption of complete information that characterizes rational
choice theory is implausible” (Schwartz et al., 2002, p.1178).
The Theory of Planned Behavior (TPB) on the other hand, has been applied in several transportation research papers and is claimed to be “the most widely applied model of modifiable cognitive antecedents of travel mode choice” (Gardner and
Abraham, 2010, p. 832). TPB claims that behavior is triggered by intent and per-
92
ceived behavioral control (PBC). Intention is defined by attitudes (i.e. perceived
outcomes of the behavior in question) and subjective norms (i.e. perceived social
approval from others of the behavioral act performed) (Ajzen, 1991). Intention
therefore motivates people to behave in a specific manner and is equivalent to a favored behavioral act (Fujii and Gärling, 2003). Perceived behavioral control is defined as the professed ease with which a new behavior can be implemented (Ajzen,
1991), i.e. how easy it could be to drive sustainably. Newnam et al. (2004),
Goldenbeld et al. (2008), and Stead et al. (2005) used the TPB theory to direct their
interventions at reducing speeding, whereas Bamberg et al. (2003) and Forward
(2004) utilized it to encourage drivers to travel by bus through means of providing
them with prepaid bus tickets.
Additionally, one’s environmental beliefs reinforce attitude and intention. Environmental beliefs elicit moral obligations and consequently influence intentions, which
“have been shown to contribute variance in travel-mode choice over and above
TPB cognitions” (Gardner and Abraham, 2010, p. 833). Transportation research
appraised the importance of reducing journey times and financial savings; however,
with energy reduction being a priority, recent studies have acknowledged environmental beliefs too. For further information on these types of studies refer to Bamberg and Schmidt (2003), Steg and Sievers (2000) and Polk (2003). It is evident that
additional research investigating the relationship between sustainable driving and
one’s environmental beliefs, personal moral norms, and attitudes is required. Besides cognitive intentions, motivational factors of employees, i.e. why employees
are motivated to act more sustainable, needs to be further understood.
V.2.3 Descriptive Social Normative Feedback and Boomerang Effect32
Social norms, defined as conventional rules or behaviors imposed by society
(Cialdini et al., 1991), have elicited a strong impression on both attitudes and behavior, specifically in the context of social responsibility. This form of feedback
was successfully embraced when evaluating energy conservation in households
(Loock et al., 2011; Schultz et al., 2007), fuel consumption (Siero et al., 1989), re-
32
93
ducing alcohol consumption among college students (Agostinelli et al., 1995), or
towel reuse in hotels (Goldstein et al., 2008). All studies recognized that social
normative feedback had a strong influence on participants’ behavior, and hence, could also play an influential role in motivating corporate car drivers.
Cialdini et al. (1991) and Reno et al. (1993) extended the concept of social norms in
their focused theory of normative conduct by differentiating between descriptive
norms (individual’s perception of what is commonly executed in a certain situation) and injunctive norms (individual’s perception of what is commonly approved or
disapproved by society), and identifying which circumstances these norms affect
behavior. The norm must be a focal point for the individual, i.e. command their attention when in a certain situation at that present time to influence their behavior;
this is defined as normative focus (Cialdini et al., 1991).
Feedback provided by social norms does not always have a positive impact. Schultz
et al. (2007) defined a crucial example of this called the ‘boomerang effect’. In their
eight-week study of 290 households in California, household energy conservation
was addressed by providing descriptive normative feedback. Each household’s average energy consumption was presented together with social normative feedback
indicating the average energy consumption of the neighborhood. While households
with above-average figures decreased their consumption, some individuals with below-average energy consumption increased their consumption towards the average.
This reflects that individuals with energy consumption below the average for their
neighborhood adjust their consumption accordingly, resulting in a negative change.
The boomerang effect was offset by adding injunctive feedback, which showed a
smiling face to support the below-average group as consuming less energy than the
average households shown in the descriptive normative feedback (Schultz et al.,
2007). After adding this injunctive message to the descriptive normative feedback,
the energy consumption of the below-average households stayed at the desired lower level. Evidently, social-normative feedback without injunctive feedback can produce mixed outcomes; this creates a need for further investigation into the potential
impact of descriptive social normative feedback on corporate car drivers’ fuel efficiency.
94
V.2.4 Layout of Rating Scales
Recent studies in the field of energy savings analyzed the impact of social normative feedback on household energy consumption by, for example, providing descriptive normative feedback or injunctive and descriptive normative feedback (Loock et
al., 2011; Schultz et al., 2007). Little attention (at least not in the scientific literature) has been paid to the design and layout of feedback scales through which social
normative feedback can be given. Some relevant insights can be found in earlier
research in psychology and psychophysics from Freyd (1923) and Froberg and
Kane (1989), who assessed the design and layout of rating scales. Their research
focused on the statistical validity and reliability of rating scales in surveys
(Svensson, 2000). The definitions and layouts of these scales provide an overview
how feedback scales, in combination with social normative feedback, could be designed and tested.
Verbal Descriptor Scales (VDS) display items in “ordered lists of verbally described and/ or numerically labeled response categories” (Belz and Kow, 2011, p. 231). The
number of descriptive items correspond to the values of the ordered list and can
vary from 2 to 11 items (Svensson, 2000). Ordered lists include distinct rating categories in either verbal (e.g., Poor, Fair, Good, etc.) or numerical values. An example
of a VDS scale with a verbal ordered list with five items is shown in Figure 41:





Poor
Fair
Good
Very Good
Excellent
Figure 41: Verbal Descriptor Scale (VDS)
Visual Analogue Scales (VAS) show the responses on a continuous line, typically
on a horizontal line but in some cases on a vertical one (Scott and Huskisson, 1977).
The endpoints on the line represent the extreme values of the variables; for example, ‘poor’ and ‘excellent’ (Belz and Kow, 2011). These formats are used to provide
feedback via a scale ranging from 0 to 100 in a continuous form without distinct
rating categories; other ranges (e.g., from 0 to 200) have also been applied (Aitken,
1969). Figure 42shows an example of a VAS scale:
95
Figure 42: Visual Analogue Scale (VAS)
The combination of VDS and VAS scales is called a Graphic Rating Scale (GRS)
and combines the verbal or numerical items with a continuous line. The GRS displays the categorical values, either verbally (e.g., from poor to excellent) or in units
ranging from 0 to 100 on a continuous line, with pre-defined categories marked, as
shown in Figure 43 and Figure 44.
Figure 43: Verbal Graphic Rating Scale (GRS)
Figure 44: Verbal and Numerical Graphic Rating Scale (GRS)
Comparative studies between VDS and continuous VAS or GRS scales have reported conflicting outcomes, with no firm evidence of which type of scale provides a
better outcome in terms of validity and reliability (Svensson, 2000). Svensson
(2000) found that the VDS and GRS have a stronger impact than the VAS format,
as the VAS format shows a continuous line and does not provide any categorical or
pre-defined line items. However, Aitken (1969) describes the VAS scale with more
items (e.g., any value between 0 and 100) as more sensitive than discrete scales.
This has been disproved by Svensson (2000), who showed that an increased number
of possible responses does not lead to a more valid and reliable outcome.
V.2.5 Layout of Social Normative Feedback Scales
The discussion in the previous chapter illustrates the lack of clarity in the design of
scales, adding to the gap in research on how feedback scales should be designed to
provide social normative feedback. Considering the types of scale used in the energy-savings research domain, social normative feedback could be presented in various formats. Common formats include scales in which the average value for a population is displayed in comparison to an individual’s own consumption (social normative feedback). For instance, in the field of energy savings, low, average, and
96
high values of electricity consumption in a neighborhood are used as benchmarks to
inform households in the same area with similar demographic structures about their
electricity consumption in comparison to the benchmark values (Baeriswyl et al.,
2011). This type of scale is similar to a VAS format, and Figure 45 illustrates these
factors via a linear scale with categorical benchmark items for low, average, and
high energy consumption and showing the consumption of the individual
household.
Figure 45: Normative Feedback to improve Energy Consumption
(Baeriswyl et al., 2011)
Another example of a standardized rating scale, which includes elements of a VDS,
is defined in the EU directive 2002/91/EC, as shown in Figure 46:
Figure 46: Energy Efficiency Rating Scale (EU directive 2002/91/EC)
The CO2 emissions energy efficiency rating scale is used to provide an index of
how much CO2 a household is currently emitting and compare this value via an ordered categorical list, shown as a vertically oriented GRS (from category ‘A’, very low CO2 emissions, to category ‘G’, very high CO2 emissions). Each category is
further defined in numerical ranges (i.e. 100-92, 91-81, etc.). The CO2 emissions
efficiency is also color coded, with the lowest CO2 emission category dark green
and the highest dark red. Finally, the feedback scale provides an index for potential
97
reductions from the current CO2 emissions. A scale with the same layout but different color coding, from light blue to dark blue, was defined by the EU Commission
for rating electricity consumption of households.
These examples demonstrate that rating scales in energy-reduction research are similar to the rating scales from psychology and psychophysics that were evaluated by
Freyd (1923) and Froberg and Kane (1989). No study in this research domain evaluates VAS and GRS scales in relation to social normative feedback and energy reduction. Research from Cialdini et al. (1991), Loock et al. (2011), and Schultz et al.
(2007) supports the importance of providing social normative feedback in combination with injunctive feedback for improving energy reduction. The latter helped to
prevent the energy consumption of below-average households from rising above the
average once the average consumption of similar households was known (Schultz et
al., 2007). In terms of the design of scales, VAS have many items on a continuous
linear scale, whereas VDS and GRS have pre-defined categorical/ discrete items.
Whether a certain type of scale has a stronger influence on environmental behavior
needs to be further evaluated.
For the purpose of this study, two different feedback scales that follow the GRS and
VAS design principles and provide descriptive normative feedback were applied.
The feedback displays the driver’s own fuel consumption in comparison to the average fuel consumption of other drivers from the same company. The GRS (Figure
47, left) provides feedback in discrete/ categorical areas (defined in three fuel consumption areas: low, middle, and high), whereas the VAS (on the right of the figure) shows the feedback in a continuous format, with the lowest fuel consumption
on the left endpoint and the highest consumption on the right endpoint. Color coding emphasizes the fuel consumption on both types of scales, with dark green denoting low fuel consumption and dark red denoting high fuel consumption.
Figure 47: Categorical and Continual Feedback Scale Formats
98
V.3 Research Questions and Hypothesis
Studies have identified an improvement in participants’ fuel efficiency when monetary rewards were awarded, i.e. financial savings as a measure of driving infrequently or reimbursement when another transportation mode was used (Deslauriers and
Everett, 1977; Graham et al., 2011; Thogersen and Moller, 2008). Further clarity is
required to ascertain how feedback interventions that are not driven by financial
rewards can motivate corporate car drivers to adopt sustainable driving habits, and
hence, improve their company’s carbon footprint and reduces fuel costs. Reflecting
on the literature currently available, it is apparent that there is limited research, besides truck or van drivers, that investigates this client group. Furthermore, the role
Green IS reinforced by social normative feedback in promoting eco-friendly driving
habits is a relevant gap in research that will be investigated.
This study focuses on feedback delivered by a mobility IS (MIS) to improve the
fuel efficiency of corporate car drivers. As employees working for this corporation
had regular access to their emails and a corporate mobile phone, it was feasible to
extend the parameters of the investigation from the first field test to maximize the
capacity of the existing MIS, which stored relevant data from each driver, by sending customized feedback via email. Sending feedback by email enabled the researcher to access a larger sample of the research corporations driving fleet, more
than 50% of their drivers. Concepts of Green IS were employed to focus on the
technical aspect of delivering feedback to participants, and findings from environmental social psychology used to appraise how the intervention and which type of
feedback scale best influenced drivers. The fourth sub-question from the thesis is
stated as follows:
SQ4: How does social-normative feedback about a driver’s fuel consumption via email influence corporate car drivers’ average fuel efficiency?
To assist in answering this sub-question, three hypotheses were defined:
The first hypothesis evaluates the question, how does providing eco-driving techniques/ tips by email, without any eco-driving training or use of an eco-driving
smartphone application, effect driving behavior (see Table 13).
99
Table 13: Illustration of Hypothesis 1
Corporate Car
Drivers’ FE
Eco-driving techniques
feedback by email
↑ ↓ =
The first hypothesis is defined using µ 1 to represent fuel efficiency without any
feedback and µ 2 to denote receiving eco-driving tips by email:
H1: Providing details about ecological driving techniques by email improves corporate car drivers’ fuel efficiency.
(H1) H0 = µ 1 ≤ µ2; H1 = µ 1 > µ 2.
Having acquired a better understanding of how eco-driving tips by email impacts
behavior, it was also important to understand to what extent individual feedback
about driver’s fuel consumption and social normative feedback influenced behavioral change. Findings from Graham et al. (2011) have shown that increasing the
environmental awareness of individuals by providing feedback about their reduction
in CO2 emissions reduced their car use, and therefore their overall fuel consumption. When considering the social psychological theory perspective, it was evident
that the Theory of Planned Behavior by Ajzen (1991) had been extensively applied
to review travel mode choice, i.e. taking the bus instead of the car. For this study,
the application of the Feedback Intervention Theory (FIT) by Kluger & DeNisi
(1996) introduced in Chapter II provided a foundation from which to reflect on
how feedback interventions influence individuals’ task performance (i.e. driving style). The feedback intervention applied descriptive social normative feedback by
presenting drivers’ fuel efficiency rating with a comparison against the fuel efficiency rating of colleagues from the same company (see Table 14).
Corporate Car
Drivers’ FE
Social Normative
Feedback by email
↑ ↓ =
100
No current findings exist that reflect on how social normative feedback format influence corporate car drivers’ fuel efficiency, which raises the second hypothesis by
µ 1 representing the fuel efficiency without and µ 2 with receiving social normative
feedback by email.
H2: Providing social normative feedback about driver’s fuel consumption by email improves corporate car drivers’ fuel efficiency even when monetary incentives are
not relevant.
(H2) H0 = µ 1 ≤ µ2; H1 = µ 1 > µ 2.
It is evident that the setting and design of social-normative feedback encompasses
several attributes and can generate mixed outcomes. One example, known as the
boomerang effect, was tested by Schultz et al. (2007) and Loock et al. (2011). Freyd
(1923) and Froberg and Kane (1989) evaluated the reliability and validity of categorical versus continuous scale formats. Svensson (2000) identified that discrete/
categorical scales provide a better outcome compared to continuous scales. To this
end, the third hypothesis aimed to appraise what difference the type of feedback
scale (i.e. discrete/ categorical or continuous) has on participants’ response to the social normative feedback they received (see Table 15).
Scale Format
Discrete
Continual
FE rating
Corporate Car
Drivers’ FE
Low
↑ ↓ =
Medium
↑ ↓ =
High
↑ ↓ =
1
↑ ↓ =
to
↑ ↓ =
12
↑ ↓ =
The third hypothesis is defined with µ 1 representing the fuel efficiency shown on a
discrete and µ 2 representing a continual social normative feedback scale.
101
H3: Descriptive social normative feedback shown on a discrete/ categorical scale
has a stronger impact on corporate car drivers’ fuel efficiency versus that shown on a continual scale.
(H3) H0 = µ 1 ≥ µ 2; H1 = µ 1 < µ 2.
V.4 Research Design and Research Methodology
The opt-out experiment (see Figure 48) was announced in August 2011 via email by
management to 450 corporate car and car allowance drivers. Two hundred and forty
participants from this sample were then randomly selected to take part in the experiment and assigned to three groups: Treatment Group 1 (TG1), Treatment Group 2
(TG2), and Treatment Group 3 (TG3).
Figure 48: Experimental Design
Evaluation of participants’ fuel efficiency was based on true fuel consumption figures from sales data provided by the petrol credit cards of each corporate car driver.
These data were transferred to the company’s MIS. Data included: total number of kilometers driven, liters filled, total cost for each tank filling, petrol type, and location of petrol station. The raw data were then used by the MIS to calculate the fuel
102
efficiencies of each driver after each tank filling by dividing the total number of
liters of petrol filled by total number of kilometers driven since the last tank filling.
To establish a baseline for each group, the average fuel efficiency prior to the intervention (01/01/2011 to 21/09/2011) was calculated, this allowed for a comparison
to be made within subjects from each group (TG1, TG2, and TG3). The TG1 received a monthly email with only tips how to drive more ecological. Participants in
the TG2 and TG3 received a monthly email (09/22/2011, 24/10/2011, 24/11/2011)
detailing their: individual fuel efficiency data, difference from the previous month,
average kilometers driven during the month, CO2 emissions, and eco-driving tips
regarding how to improve their fuel efficiency (see Figure 49).
Figure 49: Monthly Feedback Email with Discrete or Continual Feedback
Two scales, discrete and continual, were used for to reflect how ‘good’ or ‘bad’ participant’s fuel efficiency was in comparison to their peers (descriptive social norma-
103
tive feedback). The discrete scale presented fuel efficiency in three categories: low,
middle, and high (see Figure 49 left scale). The continual scale (see Figure 49 right
scale) illustrated this information on a low to high continuum, which was divided
into 12 sections and indicated the driver’s fuel efficiency with an arrow. The average fuel efficiency percentage of the total sample (n = 240) prior to the initiation of
the experiment was 13.57% above the stated fuel efficiency from the car manufacturers, a value higher then claimed by the car manufacturer confirming their fuel
efficiency figures were somewhat optimistic. This average fuel efficiency percentage value from the total sample size was used to define the mean value as the baseline for each fuel efficiency rating scale (see Table 16).
Table 16: Fuel Efficiency Rating and Intervals
Type of
Scale
Discrete
Continual
FE Rating
Fuel Efficiency Intervals
in %
Low
< 7.09%
Middle
Between 7.09% and 18.53%
High
> 18.53%
1
< -7.94%
2
-7.94% to -3.00%
3
-3.01% to 3.52%
4
3.53% to 7.08%
5
7.09% to 10.68%
6
10.69% to 13.91%
7
13.92% to 15.83%
8
15.84% to 18.52%
9
18.53% to 20.68%
10
20.69% to 23.54%
11
23.55% to 28.47%
12
> 28.47%
The individual fuel efficiency rating on the scales were based on the difference between the actual fuel efficiency of the driver for the month and fuel efficiency of the
car type as stated by the manufacturer. For instance, if a driver had an average
monthly fuel efficiency of 6.7 liters per 100 kilometers (l/100 km) and the fuel effi-
104
ciency stated by the car manufacturer was 5.6 l/100 km, a difference of 1.1 l/100 km
or 19.72% is evident. This percentage value of 19.72% was compared with the average fuel efficiency value of 13.57% and the driver’s position on the feedback scale assigned. Furthermore, comparison between participants’ changes in monthly
fuel efficiency and different car types with varying engine sizes, and against the rest
of the total sample was possible.
V.5
Data Evaluation and Findings
V.5.1 Findings for Hypotheses H1, H2, and H3
Of the 240 participants selected for the experiment, 20 exercised their right to optout leaving in the TG1 = 72, TG2 = 75, and TG3 = 73. An additional three participants were removed from the sample, one from the TG1 and two from TG3, as they
were identified as outliers using the Grubb’s test due to very low or excessively high changes in their overall fuel efficiency. The paired t-test statistical method was
used for each group to validate the comparison before and after the treatment phase.
(H1) The results of the paired t-test for TG1 (N = 71) indicated that the average
fuel efficiency (M = 7.852, SE = 0.160) marginally increased during the treatment
phase (M = 7.861, SE = 0.160). A difference of +0.009 l/100 km or +0.12% was
calculated and found to be statistically insignificant (p > 0.05 with t (0.204) = 0.009,
p = .839 (see Table 17).
Table 17: Paired Samples T-Test TG1
Paired Differences
avg_fe_y –
avg_fe_trtmt_t
avg_fe_y avg_fe_trtmt_t
Mean
Std. Deviation
Std. Error
Mean
-.009
.374
0.444
t
df
Sig. (2-tailed)
-.204
70
.839
95% Confidence Interval of Difference
Lower
Upper
-.098
.079
105
As there was no statistically significant difference between the baseline and the
treatment, we can conclude that providing feedback about ecological driving techniques by email does not change corporate car drivers’ fuel efficiency, and the Ho cannot be rejected.
(H2) In order to address the second hypothesis a paired t-test was applied to TG2
and TG3. Statistical significance (p < 0.05) was confirmed for both groups; the average fuel efficiency for TG2 (N = 75) improved by -0.093 l/100 km or -1.24% with
the mean value of M = 7.501, SE = .174 before the treatment and M = 7.408, SE =
.186 during the treatment. Results found to be statistically significant (p < 0.05 with
t (2.068) = -0.093, p = .042 (see Table 18).
Paired Differences
avg_fe_y –
avg_fe_trtmt_t
Mean
Std. Deviation
Std. Error
Mean
.093
.388
.045
t
df
Sig. (2-tailed)
2.068
74
.042
Lower
Upper
.004
.182
The average fuel efficiency for TG3 (N = 70) improved by -0.134 l/100 km or 1.69% with the mean value of M = 7.907, SE = .166 before the treatment and M =
7.773, SE = .176 during the treatment. Results found to be statistically significant (p
< 0.05 with t (2.214) = -0.134, p = .030 (see Table 19).
106
Paired Differences
avg_fe_y –
avg_fe_trtmt_t
Mean
Std. Deviation
Std. Error
Mean
.134
.505
.060
t
df
Sig. (2-tailed)
2.214
69
.030
Lower
Upper
.013
.254
These outcomes indicate that provision of social normative feedback related to the
participant’s fuel consumption by email improves corporate car drivers’ fuel efficiency even when monetary incentives are not relevant. Therefore, the Ho can be
rejected in favor of the H1.
(H3) An independent t-test was used to compare the difference between the discrete
scale (TG2) and continual scale (TG3). The result showed that, on average, the continual feedback scale had a greater impact on participants’ fuel efficiency with a
difference of 0.73%; however, this was statistically insignificant (p > 0.05 with t
(144) = 0.73, p = .579). Therefore, it cannot be concluded that descriptive social
normative feedback shown on a discrete scale has a stronger impact than a continual
scale on corporate car drivers’ fuel efficiency, and the H0 cannot be rejected.
V.6
Discussion
The application of social normative feedback produced promising and statistically
significant results when a within group comparison was made in groups TG2 and
TG3. In both treatment groups an improvement in fuel efficiency was apparent; in
TG2 -1.24% and in TG3 -1.69%. However, simply providing feedback about techniques endorsing sustainable driving habits by email does not motivate a change in
participants’ task performance, i.e. driving style, to improve fuel efficiency. This
assumption was also stated by Kluger & DeNisi (1996) and extended by Alder
107
(2007) in the application of the FIT, which identified that feedback interventions
should be postulated in the correct context to trigger changes in task performance.
When comparing the discrete and continual feedback scales, no statistical significant difference between TG2 and TG3 was discovered. Therefore, we cannot conclude that descriptive social normative feedback illustrated on a discrete scale generates a stronger or weaker impact in improving fuel efficiency when compared to a
continual scale. This echoes current literature findings that explain the design, reliability or validity of feedback scales, but do not disclose what impact the different
scales have on an environmental behavior. Nevertheless, we can assume that indirect feedback by email and social normative feedback provided in the correct context can have an influence on driving behavior, which further confirms findings by
Siero et al. (1989). In their study one short-coming was the application of direct and
indirect feedback types; hence, they were not able to clearly conclude which of the
feedback types had a stronger influence on the overall improvement in fuel efficiency (Siero et al., 1989).
V.7
Conclusion
To summarize, an opt-out field experiment was conducted over three months using
drivers from one corporation. Access to participants’ MIS enabled rich data relevant to fuel consumption to be collected. Research within the fields of feedback technologies and social psychology appraised how the potential use of Green IS can be
maximized by using emails with social normative feedback. Feedback IS coupled
with social psychological concepts have a valuable contribution to make in research
approaches to modifying drivers’ behavior to stimulate improved fuel efficiency (Froehlich et al. 2010; Jenkin et al. 2011b; Meschtscherjakov et al. 2009).
Findings of the study imply that extended use of Green IS by sending emails that
provide fuel consumption figures and social normative feedback had an influence in
improving corporate car drivers’ fuel efficiency without any financial rewards. This
affords support to recent literature that claims Green IS can stimulate behavioral
modification to reduce energy consumption (Graml et al., 2010; Loock et al., 2011)
and so bridging the ‘environmental literacy gap’ by enhancing individuals’ environmental awareness (Fogg, 2002; Froehlich et al., 2010). It is important to
108
acknowledge that providing social normative feedback on a discrete/ categorical or
continual scale format did not show a statistically significant difference and that
further research is required. In practice, organizations with large car fleets looking
for cost effective strategies to reduce their CO 2 emissions and petrol costs, Green IS
supported by social normative feedback proposes an attractive solution.
V.8
Limitations and Future Research
As all groups received some level of feedback, with TG1 receiving the minimal of
driving tips to promote sustainable driving, limited control over seasonal effects
was possible. As weather conditions vary during the course of a year, it is honest to
recognize that this can have an impact on fuel efficiency. A solution to address this
would have been to assign a control group that did not receive any feedback, however, the average fuel efficiency of TG1 prior to the field test was calculated from
data collected over a duration of time (January to September) and only marginal
fluctuations in fuel efficiency were detected in comparison to the treatment time
(+0.12% per l/100 km). This suggests that seasonal effects had the slightest impact
on fuel efficiency in 2011. During the period of the field test no extreme weather
changes were apparent and so it is fair to claim that weather conditions had a minimal impact on fuel efficiency data collected for this study.
There were opportunities for further data analysis. Changes in fuel efficiency could
have been evaluated on a monthly basis instead of accumulatively at the end of the
field test. This would have provided information about how participants’ responded to feedback over time; i.e. was there a greater improvement in their fuel efficiency
after the first email compared to the third email.
Recommendations to Increase Corporate Car Drivers’ Intrinsic Motivations
109
VI Recommendations to Increase Corp. Car Drivers’ Intrinsic Motivations33
VI.1 Overview and Research Questions
The eco-driving smartphone application provided direct and indirect feedback to the
driver when driving and resulted in an average improvement in fuel efficiency of
3.23% (see Chapter IV). The MIS provided indirect feedback through a monthly
fuel consumption email, which led to a fuel efficiency improvement of up to 1.69%
(see Chapter V). In addition to analyzing whether feedback technologies have a positive impact on corporate car drivers’ fuel efficiency, behavioral and motivational aspects that help corporate car drivers improve their fuel efficiency must be addressed. The case study company currently has 19,100 corporate cars globally and
is interested in the behavioral aspects of motivating employees without giving financial incentives as it seeks to encourage eco-driving via feedback technologies.
The main goal of this chapter is to derive recommendations by evaluating corporate
car drivers’ motivations for fuel savings and opinions about implementing a fuel savings program supported by feedback technologies within their organization.
Specifically, extrinsic and intrinsic motivations are evaluated to answer the following research question:
VI.2 Extrinsic and Intrinsic Motivation
Understanding what motivates an individual is a complex task. Management must
understand which factors, intrinsic or extrinsic, motivate their employees (Beswick,
2007). Extrinsic motivation involves external rewards (i.e. financial incentives or
extra days off) and/ or recognition (i.e. participating in important meetings)
(Zahorsky, 2010). The disadvantage of this form of motivation is that employees
focus on the reward, not on the action itself (Beswick, 2007). Intrinsic motivation
33
110
stems from within the person, and satisfaction derives from completing the task itself (Zahorsky, 2010), contributing to the driver’s self-esteem. When applied correctly, extrinsic motivation can enhance intrinsic motivation. De Young (1985) suggested that intrinsic motivation encourages environmentally responsible behavior;
individuals do not wait for rewards to be environmentally responsible but rather
“seem to derive personal satisfaction from the very activities that others so often try
to externally reinforce” (De Young, 1985, p. 289).
Studies on home energy savings discovered that people were more motivated to
save energy and be environmentally friendly when they received personalized information about their energy consumption (Shipworth, 2000). Users need to understand the information (Shipworth, 2000) and feel that they have freedom of choice
and control (Rotter, 1966). It is necessary to support these individuals in developing
their own intrinsic motivation (e.g., to drive sustainably), as recommended by De
Young (1985). Siero et al. (1989) concluded that although it is important for management to set feasible goals and provide information when making task assignments, they must also capture employees’ attention by reinforcing that energysaving driving behavior is an integral part of their job. However, caution must be
exercised to avoid external control (Amabile, 1993): the driver must feel a sense of
autonomy, and information must be clear enough that they can understand and utilize it effectively. Implementing an organizational eco-driving initiative supported
by eco-feedback technologies requires management to understand what motivates
their employees and to meet their needs.
VI.3 Research Design and Research Methodology
In total, fifteen semi-structured post-experiment interviews (see Appendix G), each
lasting 20 - 25 minutes, were conducted with field test participants. Before the interview, all interviewees filled out the post-experiment online survey (see Appendix
F). This helped to obtain a better understanding of the interviewees before the interviews took place. The semi-structured interviews were used to give interviewees a
forum for expressing their experiences and opinions in a structured manner. Questions for the semi-structured interviews were selected based on the interviewees’ answers in the post-experiment survey, completed directly after the field tests. This
enabled the researcher to highlight specific post-experiment survey answers and
111
open questions that were in need of clarification during the interview. Together with
obtaining feedback about the feedback technology, the post-experiment interviews’ purpose was to evaluate opinions and derive conceptual recommendations to be
considered when implementing a fuel efficiency improvement or CO2 reduction
initiative within an organization.
Structured analysis was used to evaluate the unstructured data collected in the follow-up interviews (see Figure 50). The process of structured analysis is defined as,
“microanalysis, which consists of analyzing data word-by word” while determining
“the meaning found in words or groups of words” (Strauss and Corbin, 1998, pp.
65–68). The interview data was transcribed and classified into topics to unveil
emergent factors; this is known as the coding process. The subsequent process, axial
coding, grouped similar factors into categories. This enabled a systematic analysis
and constant comparison of factors and categories. Finally, factors were ranked according to the frequency of answers and assigned to their category. This process
enabled a structured approach to interpreting the answers.
Figure 50: Interview Data Analysis
112
After evaluating and categorizing the interview data, the FIT (Kluger and DeNisi,
1996) was applied to obtain a better understanding how feedback interventions using eco-feedback technologies enable corporate car drivers’ to modify behavior. In detail, five basic arguments drawn from the FIT and the literature on extrinsic and
intrinsic motivation were compared with the categorized factors from the interviews
and answers from the post-experiment survey.
VI.4 Data Evaluation and Findings
VI.4.1 Opinions and Concerns of Drivers
The structured analysis process of the interviews revealed several categories which
are defined from the ranked factors. Categories with most ranked factors are summarized in Table 20 (all answers are shown in Appendix G) and further defined in
two levels: Driver and Management Level. The Driver Level focuses on the opinions of the drivers about eco-feedback technologies, whereas the Management Level on specific aspects a management should consider by rolling out an eco-driving
concept supported with feedback technologies.
Table 20: Categories of Structured Analysis
Questions
1. Driver Level
1. What do you think Awareness:
was the greatest bene- - Less about feedback,
fit of the application?
more about using the
application (15)
- Raise awareness (13)
- Sensitize (11)
2. Besides the feedFeedback:
back you have re- Real-time feedback (15)
ceived, which addi- Should not distract from
tional feedback criteri- driving (12)
on are relevant to you? - Transparent feedback (12)
2. Management Level
113
3. Do you think a
monthly e-mail with a
comparison of your FE
against another driver
would be useful?
If yes, what kind of
comparison would you
prefer?
- Yes (15)
4. Which other aspects
would be relevant for
you to use the application regularly in your
corporate car?
Time importance:
Provide information:
- Everyday on the road (15)
- Provide personalized
- Important to be home fast (15)
information (14)
- Many customer visits (10)
- Expose fuel costs (13)
Importance of valid comparison:
- Same car model (15)
- Same average total distance
driven annually (14)
- Transparent comparison (10)
Incentives:
- High influence of financial
incentives (12)
5. How should the
management roll out a
fuel improvement/ CO2
reduction initiative
within the organization?
i) What steps do you
feel are crucial for implementation?
ii) Where do you see
problems?
Rewards:
- Realistic/ fair rewards (12)
Punishment/ control vs.
awareness:
- No punishment (14)
- Unfair to punish drivers
who need their car for
daily business (13)
- No control from management (12)
Goal setting :
- Goal setting necessary (12)
- Realistic goals related to
corporate car drivers (11)
All participants agreed that using the eco-driving smartphone application raised
their awareness, encouraging them to think about actual driving habits in comparison to a more sustainable driving style (question one). The application’s abilities to raise awareness and sensitize the driver to their personal driving style were considered its biggest benefits (stated 13 and 11 of 15 times, respectively). Question two,
which asked about which feedback criteria were important to the drivers, revealed
that real-time and transparent feedback that did not distract from driving were favored.
The feedback provided by a monthly email (question three) was also considered
very important (stated 15 times). Respondents felt feedback should be given in the
form of a social comparison with other drivers; participants specified that the com-
114
parison had to be valid, explicit, and transparent (stated 10 times). Drivers identified
the most suitable social comparison at one between drivers who drove the same vehicle type model (stated 15 times) and approximately the same average distance per
year (stated 14 times).
Greater focus was placed on cost savings (petrol-cost reduction or reduced deterioration of the vehicle) and time savings than on environmental benefits (question
four). Employees working in sales or consulting are required to visit several customers per day to generate leads or offer consulting services. Drivers are thus unable to completely take control of their fuel efficiency, as this is partly dependent on
where their customers are located and how many customers they have to visit on a
given day. Depending on customers’ geographical locations, time pressure, and traffic conditions, employees adapted a different driving style regardless of the impact
this had on their fuel efficiency. Participants stated that this was the main reason
why management should restrain from penalizing drivers who were unable to meet
fuel efficiency goals. It is necessary to consider the sales pressure and sales targets.
Fifteen statements were made highlighting that when driving long distances daily, it
is more important for the employee to reach home earlier than to improve their fuel
efficiency.
In terms of the management’s role, they should provide each driver with personalized feedback that recognizes drivers’ progress and provides individual advice
about how they can continue to improve. Drivers also favored knowing their monthly total fuel costs and having realistic and fair rewards defined for improving their
fuel efficiency. Providing the total sum of monthly fuel costs can also act as an indirect incentive for employees, especially when the company reimburses petrol costs
for private journeys. This is the case for all the corporate car drivers employed by
the case study company in Germany. The interviews revealed that the participants
had a positive attitude towards implementing eco-driving but had concerns about
management’s response to drivers who did not improve their fuel efficiency (question five). It is important that the management not punish or control the drivers, especially if they need their car daily to perform their job. Aside from these concerns,
participants had a positive attitude towards goal setting by management. They stated
that setting realistic goals was necessary to enhance their motivation to drive more
115
sustainably (stated 12 times). In contrast, only three drivers felt that management
intervening by defining fuel improvement goals was not desirable.
These answers provide valid insights into corporate car drivers’ opinions and valuable information for management. In order to better verify how feedback could trigger employees’ motivations, the five basic arguments of the FIT (Kluger & DeNisi,
1996) were applied and compared with the respondents’ answers in the next section.
VI.4.2 Analysis of Findings and Discussion
The FIT of Kluger & DeNisi (1996), introduced in Chapter II, is composed of five
basic arguments that can be applied to explain behavioral modification using computer-mediated feedback interventions, e.g., through an eco-driving smartphone application or an MIS. The results of this research will be discussed in relation to these five basic arguments as well as employees’ extrinsic and intrinsic motivations, as defined in Chapter VI.2. These results are based on findings from the postexperiment survey and the semi-structured interviews that were conducted, coded,
and evaluated after the field tests. A high-level summary of key findings is given in
Table 21.
Table 21: Summary of Feedback Intervention Theory Arguments
Arguments
Argument 1
Argument 2
Argument 3
Argument 4
Argument 5
Findings
- Existing driving behavior standards
- Feedback by eco-feedback technologies compared with driving
standards
- Enablement to adapt existing driving behavior
- Set realistic goals without punishment or control from management
- Different personal goals and standards are defined hierarchically
- Saving time is most important for corporate drivers
- Sustainable driving is a subordinate factor
- Feedback-standard gap: Disparity between driving behavior and
eco-driving behavior
- Smartphone application and MIS directs attention towards disparity
- Increase awareness of eco-driving practices
- Driving eco-friendly is at the moderate level of drivers’ hierarchy
- Set realistic goals that are attainable
- Drivers must feel competent in order to be intrinsically motivated
- Feedback intervention changes drivers’ locus of attention
- Deviation away from saving time towards driving more sustainably
- Positive impact of more ecological driving behavior
116
Argument 1: Behavior is regulated by comparisons of feedback to goals or standards.
The first argument states that individuals have their own standards and might be
reluctant to change their behavior without having external goals and receiving feedback about their performance. This underpins the importance of influencing corporate car drivers’ driving behavior through feedback interventions. The participants’ extensive driving experience has caused them to develop driving habits, which then
constitute their standards. The feedback participants receive is generated by the
feedback technology, which compares their driving behavior to relevant eco-driving
practices. This helps raise their awareness by compelling them to consider accepting
the feedback given and enables them to choose to adapt their driving behavior. Improved driver awareness was attained amongst the corporate car drivers. Together
with improving their fuel efficiency, 79% discussed their driving behavior with
friends and/ or colleagues during the treatment phase, as revealed in the postexperiment online survey.
The semi-structured interviews conducted with participants revealed that they were
comfortable with management interventions in the form of goal setting, as long as
there were no associated punishments. These findings align with the FIT, which
suggests that employees can only be intrinsically motivated if they feel a degree of
autonomy and competence in their task (Deci and Ryan, 1985). Employees should
not feel controlled by external motivators (Amabile, 1993), and management must
set realistic fuel efficiency improvement goals and provide constructive feedback.
The study by Siero et al. (1989) supports these findings, concluding that it is important for management to provide task-orientated assignments and sufficient control (e.g., to set feasible goals) in order to successfully implement an eco-driving
program.
Argument 2: Goals or standards are organized hierarchically.
Individuals set their own goals and standards according to their intrinsic motivation
in the context of extrinsic goals or standards. In addition to setting these goals, individuals order them hierarchically. Corporate car drivers’ personal driving goals varied; for example, some aimed to save as much time as possible while driving, while
others aimed to drive safely or to drive more efficiently (see Table 20). The goal
given in the field test was to drive more sustainably, but consideration of partici-
117
pants’ personal goal hierarchies made it apparent that time took precedence for these drivers. One interviewee noted, “If you are on the road a lot, the limit is reached very fast and time becomes the most important factor” (Corporate car driver 3,
2012). These different goals and standards need to be understood in order to define
realistic goals that meet employees’ expectations.
Argument 3: Attention is limited, and therefore only feedback-standard gaps that
receive attention actively contribute to behavior regulation.
The feedback-standard gap in this research is the disparity between eco-driving habits and corporate car drivers’ driving behavior. For instance, the eco-driving
smartphone application offered eco-driving feedback during driving. This intervention was an important factor in improving fuel efficiency, indicating the need to actively direct drivers’ attention to the difference between their driving habits and eco-driving practices. Hence, increasing the driver’s awareness through ecofeedback technologies could shrink the gap between the actual driving style and a
more ecological style. As emphasized by De Young (1985) in the household energy
context, raising residents’ awareness of energy consumption by providing them with these details can stimulate their intrinsic motivation.
Drivers also expressed that the eco-driving smartphone application or monthly
feedback email raised their awareness of their own driving behavior; they advised
that, when considering future guidelines, it is important to increase awareness of the
practices that support eco-friendly driving. The following preferences for receiving
such feedback were highlighted in responses from the post-experiment survey: via
visualization only (6 out of 7) and from a weekly email (4.88 out of 7). Providing
feedback via e-mail may be a promising way to supplement and enhance the feedback provided by an MIS.
Argument 4: Attention is normally directed to a moderate level of the hierarchy.
The case study company’s corporate car drivers spend most of their time on the road, driving hundreds of kilometers almost every day. Ninety-six percent of participants used their car daily or 5-6 times per week, and 92% drove on average more
than 24,000 kilometers per year; this conforms to the general criteria for corporate
car drivers. They also had the added pressure of time, with a need to be punctual for
customer meetings. This explains why time was the most important factor for these
corporate car drivers, as highlighted in the semi-structured interviews (see Table
118
20). Participants had a strong desire to reach customers on time and to reach home
as soon as possible at the end of their workday. Understandably, saving time is allocated the top spot in drivers’ goal hierarchy. The FIT states that attention is not always actively directed to the highest goal but rather at a moderate level (Kluger and
DeNisi, 1996b). Driving more eco-friendly was not the highest goal; it appeared at a
moderate level of the hierarchy, and so attention could have been redirected towards
reaching this goal. The post-experiment survey indicated that participants had moderate environmental attitudes (5.01 out of 7 on average), supporting the argument
that driving more eco-friendly was at a moderate level of their hierarchy.
With respect to management interventions, Argument 4 provides the criteria for setting realistic goals. For instance, an improvement of 10% within two months would
have been unrealistic as studies have shown that an improvement of up to 15% is
only possible by attending eco-driving training (Onoda, 2009). Setting an unrealistic
goal could have lowered drivers’ commitment to achieving it;; they instead may have placed it at the bottom of their hierarchy, where it would not have received any
attention. This could have also lessened participants’ feelings of competence when
they were unable to reach the goal (Amabile, 1993; Deci and Ryan, 1985). Consequently, the extrinsic motivation behind setting the goal is diminished, and the potential to positively influence the drivers’ intrinsic motivation is lost. This supports the notion that it is highly important to make an employee feel competent by setting
realistic standards that intrinsically motivate them (Amabile, 1993).
Argument 5: FIs change the locus of attention and therefore affect behavior.
Feedback provided by the eco-driving smartphone application and the MIS are
CPM-mediated feedback interventions, which offer eco-driving feedback or fuel
consumption details according to the car type. As explained, feedback directed drivers’ attention away from their preoccupation with time towards driving more sustainably;; hence, it could have changed the drivers’ locus of attention. By doing so, participants’ driving behavior was positively impacted, and they shifted gears earlier (5.45 out of 7) and accelerated more smoothly (5.13 out of 7). These aspects of
driving behavior are part of the eco-driving concept, supporting the goal of enabling
participants to drive more sustainably. However, a reduction in the driving speed on
motorways was not achieved (2.83 out of 7), as most drivers took advantage of the
119
unlimited speed limit on German motorways in order to reach their destination faster.
VI.5 Conclusion
This chapter introduced recommendations through which corporate car drivers can
be motivated to reduce their fuel consumption with the support of eco-feedback
technologies and without the benefit of any direct financial reward. Appraising the
findings using the FIT of Kluger and DeNisi (1996) and an evaluation of extrinsic
and intrinsic motivations revealed that it is imperative to stimulate drivers’ awareness of their fuel consumption. Feedback from the smartphone application and the
MIS enhanced participants’ awareness of their personal performance and enabled them to modify their behavior when necessary. Actively directing drivers’ attention to the difference between their driving style and recommended sustainable driving
habits was essential to the process and reduced the gap between actual driving habits and eco-driving practices (shrinking the feedback-standard gap defined in the
FIT).
Intrinsic motivation was motivated by extrinsic motivations when selfdetermination and feelings of autonomy and competence were enhanced without
making the participants feel controlled by the management. Research by Amabile
(1993) supports these findings. Corporate car drivers were concerned that managerial monitoring would lead to loss of autonomy and punishment if their fuel efficiency did not improve. This highlights the need to ensure that an organizational
roll-out of an eco-driving program is not associated with punishments, instead using
positive motivation and developing drivers’ behavior via realistic goal setting and constructive feedback (e.g., providing fuel efficiency figures from the same vehicle
model).
VI.6 Limitations and Future Research
Due to the relatively small sample size of the post-experiment survey (n = 24) and
post-survey semi-structured interviews (n = 15), it may not be possible to generalize
these findings to other companies. However, these findings can provide preliminary
insight into the level of influence that eco-feedback technologies can have on corpo-
120
rate car drivers’ fuel efficiency. Furthermore, due to the long-term involvement of
the researcher in this research domain, his interpretation of the interview data during
the coding process may have been less objective than would be ideal. Findings from
this study’s survey should be combined with theories and methods from the socialpsychological research field in order to gain a more objective perspective by evaluating the data using statistically proven methodologies. For the purpose of further
research, a field test with corporate car drivers from the same company in another
region is planned. That region currently has 11,000 corporate car drivers, enabling
access to a much larger sample size.
Development of an Eco-driving Dashboard System
121
VII Development of an Eco-driving Dashboard System
Research findings were received well by the case study company’s different business areas. A follow-up project will focus on developing a prototype portal solution
that informs each driver of their individual fuel consumption, and a follow-up field
test with 11,000 corporate car drivers based in Germany is planned for end of 2013
or beginning of 2014. The goal is to offer the solution as a standard service in
SAP´s fleet-management software and thus available to external customers. Business departments supporting this project include the fleet management department,
the fleet-management software managers, the sustainability department, and the
company’s workers’ council.
The prototype portal solution, which is currently in development, will utilize concepts and findings from the eco-driving field test presented in Chapter V and VI.
The goal is to not only provide each corporate car driver with fuel consumption figures (including fuel costs, fuel consumption, fuel efficiency, total distance driven,
and average / total CO2 emissions) but also build a management dashboard enabling
fleet managers to review the data via an online interface. As of now the data is being received as a .csv flat file and evaluated and validated within MS Excel.
VII.1 System Architecture, Data Structure, and Data Processing
The architecture comprises three back-end systems (see Figure 51): a) internal service provider system (ISP) in which each driver’s master data is stored, b) systems from SAP´s partner petrol companies (Aral, Esso, Total) that store fuel consumption
data from each filling, and c) SAP Business Intelligence (BI). The BI system uses
the SAP HANA database (DB), which stores fuel consumption data according to
each driver´s master data, to calculate consumption figures. The front end is SAP´s
Portal solution, which allows users to view the fuel consumption data in via an internet browser.
Drivers´ master data (see Figure 52) include the driver´s ID, car type, and the DIN
consumption figures for the car. The fuel consumption data details each petrol filling, including the date, city, country, petrol type, total mileage, and total costs. The
fuel consumption data are sent monthly as a flat file (.csv) to SAP´s fleet management department and figures needed for the fuel consumption calculations (such as
total kilometers, total amount of liters filled per filling, and total costs) are stored in
122
the HANA DB according to each driver´s ID, and consumption figures are calculated by the BI system.
Figure 51: Eco-driving Dashboard Architecture
Figure 52: Driver´s Master Data Types
The HANA DB is an In-Memory database technology used by the BI system for
primary data storage. The In-Memory technology allows faster data loading of
123
DataStore Objects (DSO), i.e. master data and fuel consumption data, and the BI
models and calculates fuel consumption figures using defined mathematical equations in InfoCubes. The BI system is connected to the Portal via an application programming interface, and the fuel consumption data is transferred via an .xml file to
the front end (SAP´s corporate portal). The fuel consumption data are then displayed to each driver under the ‘employee’ section of the portal and can also accessed from a mobile device.
Figure 53: Eco-driving Dashboard Architecture
Since drivers have a unique single sign on token to SAP´s corporate portal, the data
can only be accessed in the employee´s section from each individual driver.
Through role management concepts assigned to each employee´s ID, the management cockpit can only be accessed from the car fleet management team and assigned users.
VII.2 Mock-ups
Mock-ups were validated in the second quarter of 2013 using two user groups, corporate car drivers and fleet managers. Corporate car drivers receive individual fuel
consumption figures, social comparison to other corporate car drivers with the same
car type, information on used filling stations, historical fuel consumption, and eco-
124
driving tips (see Figure 54). A social-normative comparison with others’ data will be given if the corporate fleet has at least five cars of the same model. This threshold was defined by the company´s workers’ council and is necessary to maintain
each driver’s privacy. A social comparison chart, color-coded from green to red,
displays how fuel efficient the driver is when compared to other drivers. Fuel consumption is also shown on a timeline, allowing the driver to compare figures historically. Finally, eco-driving tips are available through the portal, providing each
driver with specific recommendations for improving fuel efficiency.
Figure 54: Corporate Car Driver´s Eco-driving Dashboard
The management´s cockpit (see Figure 55) provides an overview of the total corporate car fleet, including total fuel costs, consumed fuel, total distance, and the petrol
stations at which the most fuel was consumed. The social comparison shows average fuel consumption by vehicle types across the corporate car fleet in meaningful
comparisons across the same car brands and also car types. It is planned to develop
further data useful for fleet management, such as vehicles’ maintenance cycles, individual contract data, and drivers who are misusing their petrol cards (e.g., refilling
the wrong petrol type or having average fuel consumption far above or below the
average fuel consumption for their car type). Including a map displaying the mostused petrol stations has not been approved due to privacy issues.
125
Figure 55: Management´s Corporate Fleet Dashboard
VII.3 Mock-up incorporated in SAP´s Portal
Figure 56 shows the latest version (as of June 2013) of the driver´s dashboard embedded in the Corporate Portal. This version shows all feedback types except for the
filling stations. The interface is embedded in the portal with the mark-up language
HTML5 and enables an interactive navigation experience by hovering over the consumption figures. By rolling-over a consumption figure (top left), the social comparison figures are displayed in the top right. The ‘own consumption’ figure is shown in green, yellow, or red, depending on how ‘own comparison’ compares to drivers with the same car type. Historical monthly data are displayed in the timeline
at the bottom of the screen. The historical data also show the actual fuel price in the
market and compares the market price with the average price of the total number of
fillings per month. This version for corporate car drivers is currently being tested
with end-users to receive further feedback about its usability.
126
Figure 56: Management´s Corporate Fleet Dashboard
The fleet management version is still under development. It will apply the same
concepts and technological features as for the driver´s dashboard and enables the
management to get an understanding of the overall consumption data from the entire corporate fleet down to the driver´s and car types´ level.
VII.4 Drawbacks of the current eco-driving System
Findings from the research showed the impact real-time feedback has on sustainable
behavior. However, fuel data has a one-month time lag, as data is transmitted from
the petrol stations only once per month. Receiving the data more frequently from
the petrol station companies would be beneficial for providing more up-to-date and
frequent fuel consumption feedback to drivers.
The feedback is communicated via a monthly email to all drivers; afterwards, drivers can access the portal to review their fuel consumption data. This approach requires the users to proactively seek the data and might not be as effective for improving overall fuel consumption. Further feedback concepts, such as eco-driving
training and videos, should be used in combination with eco-driving campaigns
supported by the company´s management.
127
Another drawback involves the eco-driving tips. Findings showed that providing
eco-driving tips solely by email has no effect. Therefore, the plan is to provide individualized eco-driving feedback by evaluating each corporate car driver’s historical
data.
Conclusion, Recommendations, Outlook and Future Research
129
VIII Conclusion, Recommendations, Outlook and Future
Research
This chapter summarizes key findings from the previous chapters for practitioners
and academics, reflecting on the main research question:
How can eco-feedback technologies influence the average fuel efficiency of corporate car drivers when monetary incentives are not given?
Theoretical and practical motivations underpin the potential impact that ecofeedback technologies can have on promoting eco-friendly driving habits in corporate car drivers. Within the research domain, only a limited number of studies focus
on this driver group. This population of drivers remains unique, as their employing
company reimburses their vehicle’s fuel and maintenance costs. This has allowed for a direct investigation of how use of eco-feedback technologies influenced drivers’ fuel efficiency in the absence of financial reward, the factor found to have the greatest impact in motivating private car drivers to modify their driving habits.
The next section, Key Findings, provides answers to the research sub-questions that
were evaluated and summarized throughout this thesis. This demonstrates that theoretical and practical contributions have been made by filling the research gaps that
were outlined in the State of the Art and Related Work chapters. The Recommendations section provides a novel conceptual framework, including guidelines that need
to be considered when applying a fuel reduction program within an organization.
Finally, the Outlook section discusses future research areas and offers insights into
how a prototype eco-driving feedback information system is currently being developed by the case study company.
VIII.1 Key Findings
Key findings are structured to align with the research sub-questions, highlighting
important findings from each chapter.
130
Two online pre-experiment surveys completed by private and corporate car drivers
compared each group’s preferences for receiving feedback from eco-feedback technologies to improve their driving behavior. Key findings show that only 13% of
private car drivers were aware of eco-driving smartphone applications and that navigation systems and in-vehicle on-board technologies were the best known. The lack
of awareness of these smartphone technologies could have been due to the limited
acceptance rate of smartphones in 2010. According to Gartner (2011), the
smartphone penetration rate was only 19% in 2010 yet is expected to rise to 40% by
2014. A surge in popularity in eco-driving applications can be expected to directly
correlate with this greater smartphone use.
Both driver groups had a strong preference for receiving undisruptive real-time
feedback, as also identified by Meschtscherjakov et al. (2009). This real-time feedback should be non-obtrusive and easy to use; offline-feedback, in comparison, can
outline historical data about a person’s driving behavior. When comparing types of feedback information, such as personal fuel consumption, average fuel savings potential, and eco-driving tips, the feedback detailing the personal fuel consumption
was the most important type for both driver groups. Following this, private car drivers preferred average fuel savings potential to personalized eco-driving tips while
corporate car drivers preferred eco-driving tips to average fuel savings potential.
This is not surprising, given that corporate drivers’ fuel costs are reimbursed.
Feedback about CO2 emissions was not as important, as 54% of the private car drivers and 27% of the corporate car drivers did not understand the meaning of the
emissions unit, 100 grams of CO2 per kilometer. Further education and distinct
comparisons need to be made to help users grasp this value by, for example, noting
that 167 grams of CO2 per kilometer is equivalent to using a computer for nine
hours daily for eleven working days.
These insights from the online surveys helped define which feedback types were
given to corporate car drivers in the field tests.
131
The first field test provided direct/ real-time and indirect/ accumulated eco-driving
feedback to twenty-five corporate car drivers who used the eco-driving smartphone
application for eight weeks. In total more than 7,000 fuel fillings and 800 journeys
by both treatment and control driver groups were recorded. An average improvement of 3.23% in fuel efficiency was found when comparing the treatment group to
the control group, supporting the argument that mobile feedback technologies can
positively impact corporate car drivers’ driving behavior. This conforms to the findings of Boriboonsomsin et al. (2010) and Siero et al. (1989), who provided real-time
feedback using an on-board eco-driving system. Boriboonsomsin et al. (2010)
showed an average fuel consumption improvement of 1% on highways and 6% on
city streets with twenty participants, and Siero et al. (1989) demonstrated an improvement of 7.3%. This large improvement in fuel efficiency was partly due to
additional management guidelines and classroom-based training programs for the
drivers.
SQ3: Which feedback type, direct or indirect, do corporate car drivers prefer
when using an eco-driving smartphone application?
The second part of the first field test evaluated driver’s preferences for direct/ realtime feedback as opposed to indirect/ accumulated feedback or offline feedback.
Participants who used the eco-driving application for eight weeks were able to select and change the feedback type when the application was in use. Results showed
a strong preference for direct/ real-time feedback. These findings correlate with the
literature on in-household energy savings, in which participants also favored realtime feedback (Darby, 2006; Graml et al., 2010). Besides the quantitative findings
obtained from the data generated by the eco-driving smartphone application, qualitative data was collected via the post-experiment survey and interviews. Evaluation
of this data explained that real-time feedback was favored because it immediately
notified the experienced corporate car drivers of when during their journey they
could use specific eco-driving techniques. This was more effective than receiving
feedback in an accumulated form or in three-minute intervals.
132
When considering offline feedback, an online driving portal (ODP) can play a role
in increasing awareness of driving-related data (i.e. fuel usage, CO2 emissions, or
routes driven) and eco-friendly driving concepts. However, showing a summary of
driving-related figures in isolation does not initiate a change in driving behavior.
The ODP must provide further feedback types, such as social comparisons, in order
to motivate drivers. This was further evaluated to answer sub-question four.
SQ4: How does social-normative feedback about a driver’s fuel consumption via email influence corporate car drivers’ average fuel efficiency? The second field test provided 240 corporate car drivers, assigned into three treatment groups, with driving-related figures (e.g., fuel efficiency, CO2 emissions, total
petrol costs, and total kilometers driven per month) via a monthly email generated
by a mobility information system (MIS). The feedback about the average fuel consumption was also shown using social-normative feedback by displaying the driver’s fuel efficiency in comparison to the average fuel efficiency of all 240 drivers participating in the field test. At the end of the email, eco-driving and fuel reduction
tips were provided.
Findings indicate that monthly feedback provided by an MIS email increased the
average fuel efficiency of the first treatment group by 1.24% and that of the second
group by 1.69%. The first and second groups received the social-normative feedback displayed on categorical and continuous scales, respectively. These findings
support literature in the fields of Green IS and social psychology that suggests that
Green IS stimulates behavioral modifications to reduce energy consumption (Graml
et al., 2010; Loock et al., 2011). However, providing eco-driving tips without any
social-normative feedback figures did not have an impact on overall fuel efficiency.
Showing the feedback on discrete/ categorical in comparison to continual scale formats showed a difference of 0.73%; however, statistical validation did not provide a
statistically significant difference in fuel efficiency between the two treatments
groups. Hence, it cannot be concluded that showing feedback on a discrete/ categorical in comparison to a continual scale format has a stronger impact.
In summary, offline feedback presented in the form of a monthly fuel consumption
email did have a positive impact on drivers’ fuel efficiency. It was important to rep-
133
resent fuel consumption figures in a social-normative feedback, only providing ecodriving recommendations did not have an impact.
The fifth sub-question derived guidelines and recommendations for how to stimulate drivers’ intrinsic motivation according to the five basic arguments identified by the Feedback Intervention Theory (FIT) (Kluger & DeNisi, 1996). According to the
FIT, eco-driving is at the moderate level of drivers’ hierarchy and therefore possible to direct drivers’ locus of attention to new driving habits. It is essential to actively direct drivers’ attention to eco-driving techniques in comparison to their existing
driving behavior. Feedback technologies can be the medium through which to address this disparity and to raise drivers’ intrinsic motivations so that eco-driving
becomes a natural driving habit. Drivers need to feel competent and autonomous in
order to be intrinsically motivated; managerial control or punishment when fuel reduction goals are not met can jeopardize this. Management should define realistic
fuel reduction goals and provides constructive feedback (e.g., social-normative
comparison with vehicles from the same car type).
VIII.2 Theoretical Contributions
VIII.2.1 Feedback Intervention Theory
This research contributes to the Feedback Intervention Theory (FIT) of Kluger and
DeNisi (1996). Specifically, it builds on findings by Alder (2007), who applied the
theory to motivate employees to improve their overall task performance by providing feedback through a computer performance monitoring (CPM) system. Furthermore, the research combines the FIT, originating in social psychology, with the organizational behavioral research domain’s consideration of how employees’ extrinsic and intrinsic motivational aspects can be triggered (Beswick, 2007). The FIT has
never before been applied to appraise how feedback interventions influence the
driving motivations and behavior of corporate car drivers.
134
According to the terms and definitions of the FIT, the external agents in this research were the eco-feedback technologies that provided different feedback types to
motivate employees (corporate car drivers) to drive more sustainably. These ecofeedback technologies influenced corporate car drivers’ task performance and improved their fuel efficiency. Drivers compared the feedback provided to their goals,
standards, and workplace norms. Through this evaluative process, drivers became
more aware of the importance of sustainable driving. Applying the five arguments
of the FIT created a structured approach to identifying motivations for corporate
care drivers. Extrinsic and intrinsic motivational factors were evaluated to identify
how drivers’ intrinsic motivations can be triggered in order to improve their fuel efficiency. These findings were evaluated in research sub-question five and thoroughly discussed in Chapter VI.
VIII.2.2 Social Normative Feedback
These findings also contribute to the social-psychological theories developed by
Cialdini et al. (1991), as this research evaluates how social-normative feedback influences behavioral modifications. Existing findings by Domnez (2007) and
Meschtscherjakov et al. (2009) showed that direct/ real-time feedback should not
distract the driver; this was highlighted in experiments and interviews throughout
this research. Indirect/ offline feedback needs to be presented in an enriched format
to be influential. Providing historical data about fuel consumption alone did not
have a strong influence on behavior. Instead, drivers requested a comparison between themselves their cohort. Therefore, individual fuel efficiency figures were
compared with the average fuel efficiency of all drivers participating in the field
test. Drivers requested a valid comparison: most favored a comparison with drivers
who had the same car model as theirs and drove similar average distances. Whether
drivers shared the same job role, driving style, or worked for the same organizations
was not relevant for the efficacy of social-normative feedback. These findings emphasize the importance of providing the correct feedback type when applying social-normative feedback to corporate car drivers.
135
VIII.2.3 Eco-Driving Concept
This thesis defined an initial concept for encouraging eco-driving within an organization through eco-feedback technologies. Chapter VIII.4 defines the concept,
which relates to relevant findings from the thesis. Clear recommendations were given from the perspectives of both management and drivers. The concept defines realistic fuel efficiency improvement goals, which feedback types should be used, and
how feedback should be provided to best raise the awareness of corporate car drivers. Implementations of eco-driving in a corporate context should not only trigger
short-term changes through extrinsic motivational factors like financial rewards.
Long-term sustainable options for inspiring intrinsic motivations that make ecodriving a habit for employees are also highlighted. This concept can be used as a
foundation for further research combining social-psychological, HCI, and information systems research streams.
VIII.3 Practical Implications
As stated in the introduction, the research context focused on three areas: the technological, behavioral, and business value perspectives. Practical implications are
outlined in each of these three areas.
VIII.3.1 Technological Perspective
Multiple technologies that communicate eco-feedback were tested. Outcomes of
both field tests support Fogg’s (2002) argument that persuasive feedback technologies can change an individual’s attitude or behavior. The applied eco-feedback
technologies enabled corporate car drivers to improve their fuel efficiency, reducing
the gap between each participant’s environmental awareness and their actual daily car usage. Froehlich et al. (2010) described this as closing the environmental literacy gap through feedback technologies. The smartphone eco-driving application provided two different feedback types (direct and indirect) and revealed eco-driving
techniques while an individual drove, whereas the MIS and the monthly fuel consumption email increased awareness of actual car usage by providing socialnormative feedback. The fuel efficiency calculations and algorithms used to provide
the social-normative feedback and findings from the second field test provided the
136
foundation for developing the fleet management system prototype. Development of
the prototype is explained in detail in Chapter VII.
VIII.3.2 Behavioral Perspective
The pre-experiment surveys with private and corporate car drivers helped establish
an understanding of both driver groups’ preferences for feedback technologies and types. They clearly identified that smartphone-based eco-driving technologies remain underutilized; this is predicted to change in tandem with increasing
smartphone sales figures (Gartner, 2011). Furthermore, both groups preferred to
receive feedback in terms of fuel efficiency;; the meaning of ‘liters per kilometer’ was well understood. The meaning of CO2 emissions savings is unclear to most
drivers, so further education is required to improve their level of understanding.
This thesis’ findings have also enhanced our understanding of corporate car drivers’ preferences and motivations for ecological driving. This research expanded upon
Geller et al.’s (1982) examination of how to increase seat belt use and Graham et
al.’s (2012) study into reducing car usage. It was found that it is important for corporate car drivers to maintain their autonomy and competence and not to be controlled or punished if they are unable to attain fuel reduction goals defined by management. As Greengard (1996) stated, computer-mediated feedback can be seen as
an invasion of employees’ privacy and lead to decreased job satisfaction;; it is important to motivate drivers to improve their driving style themselves instead of forcing such changes upon them.
VIII.3.3 Business Value Perspective
The business value of this research includes financial savings for companies as well
as environmental and conceptual rewards. Since the ratio of corporate to private
cars in Europe is increasing annually (e.g., one quarter of cars in Germany and one
third of those in Switzerland are corporate cars) potential fuel savings can be immense for companies with a large corporate fleet. Both field tests showed a potential fuel savings as high as 3.23% without management following any further guidelines. If management implements further recommendations, these saving could be
even higher. Siero et al. (1989) demonstrated this, achieving long-term fuel savings
of up to 7.3%. With total fuel expenditures of 42.5 million Euros in 2011, the case
137
study company could save up to 3.1 million Euros per year on fuel costs. Lower
eventual deterioration of the car and fewer accidents due to improved driving styles
are not included in this figure.
The environmental reward is a reduction in overall CO2 emissions without large
investments by companies. Due to the ongoing environmental debate, fleet managers have a strong interest in reducing overall fleet emissions as well as costs. Regulations and financial incentives that encourage employees to buy cars with lower
CO2 emissions are already in place, but a change in vehicle alone will not lead to
lower emissions. The driver must also adapt a more eco-friendly driving style, as up
to 30% of fuel efficiency depends on driving style (Romm and Frank, 2006). Increasing awareness and providing eco-driving tips via feedback technologies can
sensitize drivers, as shown in both field tests.
Since the case study company’s core business is selling business software solutions, they have a strong interest in developing a software solution that can be affordable
to a wide range of customers. Current fleet management software solutions offer
functions used only by fleet managers, such as cars’ contracting details, car maintenance schedules, tank refill data, management of corporate fuel credit cards, and
vehicle accident management. Providing this information to the individual driver is
not current practice but can be implemented as an extra function within the fleet
management software. The outcomes of the field tests were used as the foundation
for designing and implementing a prototype solution to reduce overall CO 2 emissions and fuel costs.
VIII.4 Recommendations and High Level Concept34
Analyzing the findings with respect to the FIT (Kluger and DeNisi, 1996b) and the
concept of extrinsic and intrinsic motivations revealed important insights. These
included ways to enable corporate car drivers to explore their intrinsic motivation to
drive more sustainably by providing extrinsic motivation in the correct manner. The
three core recommendations derived from the analysis are: a) to set realistic goals,
b) to define feedback, and c) to provide feedback. The long-term process of enhanc34
138
ing intrinsic motivation by applying these recommendations is illustrated in Table
22.
Table 22: Recommendations and High-Level Concept
Recommendation One: Set realistic goals
Behavioral change should be goal directed, as defining a realistic and contextspecific goal is essential for stimulating change (Kluger and DeNisi, 1996b). Results of the experiments showed an average fuel efficiency improvement of 3.23%
when using the eco-driving smartphone application for the duration of eight weeks
and 1.69% and 1.24% when providing an individual with a monthly fuel consumption email for three months. Siero et al (1989) found a 7.3% reduction in the fuel
consumption of postal truck drivers when extrinsic incentives were given, realistic
139
goals defined, and management directions provided. Considering the results of these
field tests and other studies, the case study company’s management is advised to set a realistic overall fuel efficiency improvement goal of 3-5% per driver. This could
lead to potential fuel savings of up to 2.1 million Euros per year for the case study
company.
Recommendation Two: Define feedback
The corporate car drivers from the case study company drive an average of 24,000
kilometers per year without any knowledge of their average fuel consumption and
fuel costs. Fuel and other costs (e.g., car maintenance) are reimbursed by the company and are not known by the drivers. By exposing drivers to these details, their
fuel efficiency could be positively influenced, as their attention is directed to the
feedback-standard gap (i.e. actual vs. expected fuel consumption) (Kluger and
DeNisi, 1996b). It is recommended that the management provide drivers with these
details in order to increase their overall awareness of their fuel consumption and
petrol costs.
Once fuel consumption and cost figures are publicized and realistic goals set, drivers require constructive advice about how to improve fuel efficiency. For instance,
eco-driving techniques should be promoted and training courses offered. Regular
feedback should indicate drivers’ progress towards meeting their goal, enhancing the driver’s understanding and extrinsic motivation. The driver must be able to recognize the positive impact they are having by reducing CO2 emissions. The management should provide information on fuel costs, consumption, and efficiency,
offer personalized eco-driving tips, and provide a valid comparison between drivers
driving the same vehicle model for similar annual distances. Feedback leveraging
such social comparisons has already been successfully applied in the literature on
household energy reduction (Loock et al., 2011).
Recommendation Three: Provide feedback
How should the feedback be provided? The findings revealed that an eco-driving
smartphone application is a cost-effective way to modify drivers’ driving behavior by providing positive real-time feedback. It is recommended that the application be
incorporated into an overall organizational energy reduction plan that focuses on not
140
only feedback technologies but also governmental aspects. Providing details
through a monthly fuel consumption e-mail or a mobility information system (e.g.,
an online portal) raises drivers’ awareness of their own fuel consumption.
Raising awareness about drivers’ fuel consumption is an important step towards enhancing corporate car drivers’ intrinsic motivation to drive in a more eco-friendly
manner. Management should not punish or control drivers if they are unable to
reach the fuel reduction goals given, as this would result in employees feeling controlled, with a loss of autonomy and a diminished sense of motivation. Efforts to
implement these recommendations in the long term should support employees’ autonomy, especially in situations where drivers must use their car for their daily
work.
VIII.5 Limitations35
One limitation is the generalizability of these findings to other organizations or regions, as the field tests were conducted with only one company within one region,
Switzerland. In social science research, it is common to apply findings from a case
study to a broader geography or to a wider social phenomenon (Flick, 2006; Williams, 2000). In this particular case, employees of the Swiss organization frequently
travel to Germany, and a follow-up field test in Germany is planned in 2013.
Participants in the eco-driving smartphone study were selected from one company’s pool of drivers. As it was an opt-in field test, participants who chose to take part in
the experiment may have already had a pro-environmental attitude or a higher technological affinity, creating a degree of bias in sample selection. It was possible to
control this aspect by analyzing their environmental beliefs and technological affinity, as expressed in the post-experiment survey, using scales validated in the energy
savings literature. Findings indicated that drivers only had moderate (i.e. ‘partly agree’) pro-environmental attitudes (5.02 out of 7.0) and technological affinity (5.1
out of 7.0); this suggests that the results of this study can be applied to other corporate car drivers with similar environmental attitudes and technological affinity.
35
141
The data obtained in the second field test (using monthly feedback provided by an
MIS) could have been evaluated further. The rich datasets obtained from the MIS
enable making comparisons based on many potential variables, including petrol and
car type, gender, the employees’ position in the company, and the total monthly distance driven.
The layout and design of the social-normative feedback provided via the monthly
fuel consumption email needs to be evaluated further. The data showed no statistically significant difference between the continual and categorical social-normative
feedback scales, so it was unclear whether one of these feedback designs has a
stronger influence when compared to the other. Nevertheless, providing feedback
using social-normative comparisons had a positive influence and was preferred by
participants. The follow-up field test with 11,000 corporate car drivers will focus on
this aspect.
VIII.6 Future Research
Future academic field tests could evaluate if the social-normative feedback should
be displayed in color or using grey bar charts and could consider which layout has a
stronger impact on fuel efficiency. Whether the feedback scale displaying fuel efficiency in comparison to other drivers with the same car type should be continual or
categorical needs to be further evaluated, as findings from this research did not
show a statistically significant difference between the two feedback scales.
Furthermore, the impact of different feedback types (i.e. only fuel costs vs. CO2
emissions vs. both feedback types) could be evaluated. Findings from Graham et al.
(2011), examining private car drivers, already identified that financial savings have
a stronger influence in motivating individuals to drive less than does feedback stating CO2 emissions savings. This information would help practitioners choose the
feedback that has the strongest impact.
In addition to behavioral modification elements, further research should evaluate
how conceptual and managerial recommendations can be further applied in a realworld setting. It is important to gain a better understanding of how and if the stated
recommendations trigger corporate car drivers’ intrinsic motivation. If so, does sustainable driving become habitual even when the employees are not financially reim-
142
bursed? This question could not be answered fully through this research and requires long-term field studies to be able to measure the changes in fuel consumption
over a longer period of time.
Appendix
143
Appendix
LIST OF APPENDIXES
Appendix A: Pre-survey Transportation Habits of Private Car Drivers ............ 144
Appendix B: Pre-survey Transportation Habits of Corporate Car Drivers ....... 152
Appendix C: Screenshots of EcoDriving Application (EcoDrive, 2012) .............. 154
Appendix D: Eco-Driving Smartphone Application Stata Output and............... 157
Appendix E: Data set Example from DriveGain Application Data Logger ........ 159
Appendix F: Eco-Driving Smartphone Application Post Survey ........................ 160
Appendix G: Eco-Driving Smartphone Application Post Interviews .................. 166
Appendix H: Monthly Personal Car Consumption Email with discrete Scale ... 169
Appendix I: Environmental Behavior Survey ...................................................... 169
144
Appendix
Appendix A: Pre-survey Transportation Habits of Private Car Drivers
Conducted in August 2010
Welcome and thank you in advance for your time!
This survey is part of a study about feedback systems, how they could influence
your driving style and what kind of features you would expect from such a system.
The survey will have 24 questions and will not take more than 10 minutes.
Please try to answer all questions, as it is very important for the validity of results.
However, there is not a right or wrong answer; what is more important is your opinion and true preferences.
Should you have any questions or want to leave some feedback, please do not hesitate to contact me on [email protected].
Please note that your answers will be treated confidentially.
Introduction
Feedback systems can measure your driving style and provide relevant information.
This can help you to understand e.g. how much fuel you consume or if you accelerate aggressively. By making you aware of these aspects, such systems can help you
to achieve a more efficient driving style, with reduced fuel consumption and consequently less CO2 emissions.
Here are some photos of systems that currently exist in the market. The first one is
based on a navigation system, the second on an onboard gauge 36 and the third on an
iPhone application.
Figure 57: Audi navigation system (under development)
36
Gauge = any kind of instrument used for measurement. In cars, a gauge may show information
about fuel level, temperature, etc.
Appendix
145
This navigation system provides calculations for three different types of routes,
namely optimal fuel efficiency, time and emissions. Drivers can also see the amount
of different types of emissions they save if they choose the one with lowest emissions, compared to the rest of the routes37.
Figure 58: Ford Gauge
This gauge, placed in the car's dashboard, provides different levels of information to
choose from and displays either instant or long-time feedback regarding fuel consumption, depending on what the driver is most interested in finding out. 38
Figure 59: greenMeter iPhone application
This iPhone application displays fuel usage and carbon emissions amount, in order
to improve efficiency by providing instantaneous feedback. 39
37
Source: Audi ""Clean Air, a Viable Planet"" Initiative, Eco-Driving Systems: Now Your Car Can
Gently Nag You Into Being More Fuel-Wise, Green Car Advisor, 2009
38
Source: Ford, Eco-Driving Systems: Now Your Car Can Gently Nag You Into Being More FuelWise, Green Car Advisor, 2009
39
Source: greenMeter Application, Hunter Research & Technology, 2009
146
Appendix
I. Introductory questions
1. Do you know any feedback devices that aim to improve one's driving behavior?
Yes (135; 68%)
No (63; 32%)
2. If yes, what kind of feedback devices do you know? Multiple answers are
possible
Navigation System (89; 35%)
Smartphone Application (33; 13%)
Trip Computer (64; 25%)
Other (9; 4%)
On-board Gauge ( 57; 23%)
3. Have you ever used such a device while driving?
Yes (103; 51%)
No (99; 49%)
4. If yes, did you notice any change in the way you drove?
Yes (65; 68%)
No (30; 32%)
5. Could you describe these changes briefly?
Answers see Chapter IV.
6. If no, do you think that it could affect your driving behavior towards a more
efficient driving?
Yes (83; 85%)
No (15; 15%)
II. Attribute Importance
1. What type of information would you prefer to see in a feedback device?
1 Most preferable … 5 Least preferable
 = 4.47
 = 4.02
 = 3.86
Alternative eco-efficient route
 = 3.79
Goal setting
 = 3.46
 = 3.39
Social comparison
 = 3.16
Rewards
 = 3.03
2. When would you prefer to see information relevant to your driving?
During driving
 = 3.97
After driving
 = 3.70
Aggregated
 = 3.54
Before driving
 = 3.23
Appendix
147
3. How would you prefer feedback information to be provided?
Visually (93; 52%)
Audibly (10; 5%)
Both (11; 43%)
4. Many feedback systems display the amount (in grams) of CO 2 emissions per
mile or km produced by driving. Are you able to translate the meaning of
e.g. 100 g/km to a feasible dimension?
Yes (83; 46%)
No (97; 54%)
4a. If Yes: Do you believe that displaying carbon emissions is relevant to
improve your driving efficiency?
Yes (54; 67%)
No (27; 33%)
An average car in the UK emits 167 CO2 g/km driven, whereas for each liter of petrol consumed 2.3 kg CO2 are emitted.
In order for you to understand, 2.3 kg of CO2 emissions result from:
- showering for 35 minutes
- cooking on the electric stove for 2.3 hours
- operating a lift for 92 minutes
- using a computer for 11 working days of 9 hours each
- having the fridge on for 14 days
- driving an average car for 9.2 km
- traveling by bus, train or tube for 37km
4b. If No: Do you believe that displaying carbon emissions is relevant to
improve your driving efficiency once you have a better understanding of
the meaning of CO2 emissions?
Yes (69; 71%)
No (28; 29%)
III. Intro systems description
You will now see pictures of feedback systems, which aim at helping you to improve your driving behavior. After reading their description carefully, you will have
to rate them according to your preference order.
Figure 60: Gauge in cars dashboard
148
Appendix
The leaves are part of the gauge in the car's dashboard 40. They react instantly to
your driving behavior by flourishing when you drive efficiently and disappearing
when you don't.
Figure 61: Screen on top of the car dashboard
This screen is on top of the car's dashboard.41 It changes colors according to your
driving behavior; it turns dark blue when you consume more fuel and bright green
when you drive more efficiently.
Figure 62: iPhone Application
This is an iPhone application.42 While you drive efficiently the silver ball remains
into the green area. If you start consuming more fuel or accelerate aggressively, the
ball moves to the yellow and red areas. In addition to this you get a driving score,
marking your driving behavior at each trip.
40
Source: First Drive: 2010 Ford Fusion Hybrid, autoinsane, 2009
Source: The Game (Changer), motive, 2009
42
Source: GreenGasSaver, The Apple Seed Story, 2009
41
Appendix
149
Figure 63: PC Application
This is an application for your PC, which collects data of your driving behavior.43
The amount of CO2 emissions that is saved because of efficient driving is translated
to the energy consumed for certain everyday activities, such as lighting, heating or
cooking.
Figure 64: iPhone Application
This is an iPhone application.44 It shows the amount of CO2 emissions produced for
each speed range, and which proportion is produced by acceleration, aerodynamic
drag or rolling resistance. With this information, you can adjust your driving style
towards producing less emissions.
43
44
Source: ecoDrive, Fiat
Source: greenMeter Application, Hunter Research & Technology, 2009
150
Appendix
Figure 65: Audi navigation system (under development)
This is an on-board navigation systems.45 It provides three alternative routes, optimized for time, fuel efficiency or emissions, based or real-time data it receives for
traffic conditions and road topology. Plus, you can see the amount of emissions savings, depending on which route you choose.
1. Now, please rank the previous systems according to the order of your
preference
Color-changing dashboard
 = 4.40
Leaves shown in dashboard
 = 3.72
Navigation system
 = 3.71
Silver Ball iPhone app.
 = 3.41
CO2 emissions iPhone app.
 = 3.24
PC application
 = 2.60
2. What made you choose the #l_3# (your most preferred one)?
3. What didn't you like in #l_4# (your least preferred one)?
4. Do you have any further ideas to display information about eco-efficient
driving?
IV. Demographic Questions
1. What is your gender?
Male (88; 64%)
45
Female (50; 36%)
Source: Audi "Clean Air, a Viable Planet" Initiative, Eco-Driving Systems: Now Your Car Can
Gently Nag You Into Being More Fuel-Wise, Green Car Advisor, 2009
Appendix
2. What is your age?
18 - 24 (28; 28%)
25 - 29 (67; 49%)
30 - 34 (23; 17%)
35 - 39 (4, 3%)
151
40 - 44 (2; 1%)
45 - 49 (2; 1%)
50+ (2; 1%)
3. How many years of driving experience do you have?
0 - 4 (39; 28%)
15 - 19 (10; 7%)
5 - 9 (60, 43%)
20+ (7; 5%)
10 - 14 (23; 17%)
4. How often do you drive?
Less than once a month (22; 16%)
1 - 3x per month (11; 8%)
Weekly (9; 6%)
2 - 3x per week (17; 12%)
Every day (80; 58%)
5. An average European driver drives about 40 km (roughly 25 miles) per
day. How much do you drive per day?
Less than that (85; 61%)
More than that (54; 39%)
6. Which of the following driving styles do you think best describes
yours?
Economic (32; 23%)
Aggressive (17; 12%)
Careful (45; 32%)
Risky (4; 3%)
Sportive (41; 30%)
152
Appendix
Appendix B: Pre-survey Transportation Habits of Corporate Car Drivers
Conducted in August 2011
I. Questions about Transportation
1. How often do you go to work by public transport?
1 - 2x per week (21; 14%)
Every day (7; 5%)
3 - 4x per week (14; 10%)
Not applicable (105; 71%)
2. How often do you drive to work by your car?
1 - 2x per week (22; 15%)
Every day (71; 48%)
3 - 4x per week (43; 30%)
Not applicable (11; 7%)
3. Please estimate your average monthly amount of kilometers driven
Not known (10; 7%)
Between 2011 – 3000 (39; 21%)
Between 0 – 1000 (29; 20%)
Above 3000 (24; 16%)
Between 1001 – 2000 (45; 31%)
4. Please estimate your monthly average fuel emission (in l/ 100km)
Less than 4 l (1; 1%)
10 - 11 l (8; 5%)
4 - 5 l (15; 10%)
11 - 12 l (0; 0%)
6 - 7 l (71; 49%)
More than 12 l (0; 0%)
8 - 9 l (36; 32%)
Not known (5; 3%)
5. Please estimate and indicate your average monthly cost of petrol? (in CHF)
0 - 200 (37; 25%)
Above 600 (7; 5%)
201 - 400 (54; 37%)
Not known (38; 26%)
401 - 600 (11; 7%)
6. Please estimate and indicate your average monthly CO2 emission (in kg)
0 - 200 (8; 5%)
Above 600 (3; 2%)
201 - 400 (10; 7%)
Not known (114; 78%)
401 - 600 (11; 8%)
7. Do you understand the meaning of 100gram of CO2 per km?
Yes, but I could not
Yes (54; 37%)
No (39; 26%)
relate it (54; 37%)
If the answer to the previous question is 'No': In your opinion, would it be easier to understand the meaning of '100 grams of CO2 per km' if a comparison
would be given. For instance: 100g of CO2 per km produced is equal to the usage of a computer for 4.5 hours?
Yes (79; 60%)
No (40; 31%)
Other comparison (12; 9%)
Appendix
8. Do you believe that knowing your personal CO2 emission is relevant to im
prove your average fuel efficiency
Yes (77; 53%)
No (69; 47%)
9. Please evaluate the importance of receiving following information types,
which could influence you to improve your fuel efficiency
1 Strongly disagree … 7 Strongly Agree
 = 3.83
 = 3.77
 = 3.11
AVG km driven (monthly)
 = 3.06
 = 2.89
10. If you are an iPhone user, can I contact you in order to provide more in
formation about the EcoDriving
Yes (45; 35%)
No (85; 65%)
153
154
Appendix
Appendix C: Screenshots of EcoDriving Application (EcoDrive, 2012)
Figure 66: Description in iTunes Store
Figure 67: Main Screen
Figure 68: Setting Screen
Figure 69: Add Vehicle
Appendix
155
Figure 70: Enable Upload of Data
Figure 71: Shift to N Gear
Figure 72: Shift to 1st Gear
Figure 73: Shift to 2nd Gear
156
Appendix
Figure 74: Shift to 3rd Gear
Appendix
Appendix D: Eco-Driving Smartphone Application Stata Output and
Description of Variables
A. Stata Output
diff_
id_
avg_
diff_ oem_
diff_
trtmt_ f_
km_
fe_ fe_ oem_ trtmt_ avg_
trtmt_
grp type km_fil* l_fil* date_fil* drvn
n id
fe rtmt_t oem trtmt_t t_perc fe_y avg_fe_y
1 23
0
D 95185 63.11 27.10.11 586 10.77
8.29 7.10
1.19 16.80 8.38
-1.04
2 24
0
D 67051 54.02 28.10.11 727 7.43
7.44 7.00
0.44
6.34 7.47
-0.39
3 26
0
D 10300 37.62 01.11.11 500 7.52
7.82 5.20
2.62 50.44 6.82
14.71
4 27
0
D 21876 46.45 08.11.11 823 5.64
8.08 7.10
0.98 13.81 7.99
1.11
5 58
0
D 12070 66.9 29.10.11 578 11.57 12.42 10.60
1.82 17.21 9.40
32.16
6 29
0
D 143895 51.76 26.10.11 1431 3.62
8.20 8.10
0.10
1.21 8.00
2.42
7 30
0
D 97630 46.28 26.10.11 720 6.43
6.57 6.20
0.37
5.90 6.62
-0.76
8 31
0
D 45585 57.1 25.10.11 735 7.77
8.34 5.70
2.64 46.36 8.09
3.09
9 32
0
D 162157 67.4 26.10.11 2077 3.25
6.84 6.20
0.64 10.26 6.67
2.50
10 33
0
U 97371
58 28.10.11 500 11.60 12.49 10.10
2.39 23.68 12.89
-3.12
11 36
0
D 27800 63.56 03.11.11 900 7.06
7.67 6.20
1.47 23.72 7.39
3.84
12 38
0
D 34865 60.72 31.10.11 788 7.71
7.64 7.00
0.64
9.21 7.69
-0.65
13 39
0
U 17181 43.58 01.11.11 489 8.91
8.43 7.20
1.23 17.05 7.77
8.40
14 40
0
D 53200 52.62 25.10.11 1100 4.78
8.82 6.70
2.12 31.70 8.50
3.84
15 41
0
D 63317 53.09 07.11.11 790 6.72
6.80 6.00
0.80 13.34 6.95
-2.15
16 42
0
U 43425 50.35 24.10.11 607 8.29
7.85 7.20
0.65
9.00 7.87
-0.30
17 43
0
D 130770 72.38 31.10.11 820 8.83
8.22 8.70 -0.48 -5.50 8.13
1.11
18 45
0
D 61882 44.3 26.10.11 527 8.41
8.70 7.50
1.20 15.96 8.88
-2.04
19 47
0
D 68010 59.4 04.11.11 690 8.61
8.81 6.94
1.87 27.01 8.75
0.78
20 49
0
U 56387 47.7 04.11.11 -392 -12.17 11.33 8.10
3.23 39.92 9.42
20.36
21 53
0
D 80747 35.73 24.10.11 623 5.74
6.31 5.70
0.61 10.71 6.64
-4.94
22 54
0
D
5034 51.85 12.11.11 872 5.95
5.92 5.90
0.02
0.31 6.05
-2.13
23 55
0
U 118777 24.46 24.10.11 6252 0.39 10.48 8.00
2.48 30.98 9.73
7.67
24 56
0
D 25380 96.54 12.11.11 880 10.97 11.10 9.30
1.80 19.37 10.92
1.66
25 57
0
D 28100 63.28 25.10.11 950 6.66
7.66 5.90
1.76 29.83 7.38
3.86
26 1
1
D 53336 43.99 25.10.11 681 6.46
6.41 5.90
0.51
8.68 6.49
-1.25
27 2
1
D 30500 50.78 25.10.11 700 7.25
6.76 5.80
0.96 16.58 6.63
1.92
28 3
1
D 104306 58.16 24.10.11 1176 4.95
5.13 5.30 -0.17 -3.21 4.64
10.67
29 4
1
D 38400 45.3 25.10.11 750 6.04
6.25 4.80
1.45 30.19 6.24
0.09
30 5
1
D 67500 59.2 30.10.11 900 6.58
6.25 6.60 -0.35 -5.26 6.84
-8.56
31 6
1
U 22054 26.49 25.10.11 306 8.66
8.95 7.60
1.35 17.81 9.09
-1.50
32 7
1
D 71712 45.07 24.10.11 212 21.26
7.59 6.70
0.89 13.34 7.87
-3.54
33 8
1
D 28794 64.91 24.10.11 729 8.90
8.32 7.00
1.32 18.80 8.31
0.02
34 9
1
D 14567 58.87 18.11.11 948 6.21
6.72 6.10
0.62 10.21 7.65
-12.15
35 10
1
D 120720 54.3 25.10.11 659 8.24
7.62 6.00
1.62 26.96 7.55
0.92
36 11
1
D
724 61.39 04.12.11
0
6.86 5.20
1.66 31.99
37 12
1
D 118573 16.02 25.10.11 798 2.01
6.82 6.00
0.82 13.71 6.78
0.60
38 13
1
D 66666 64.33 30.10.11 1666 3.86
8.58 7.50
1.08 14.42 9.37
-8.44
39 14
1
D 44432 61.5 31.10.11 886 6.94
6.71 6.40
0.31
4.80 6.72
-0.12
40 15
1
D 95675 44.33 26.10.11 658 6.74
6.90 6.00
0.90 15.05 7.17
-3.69
41 16
1
U 107200 49.53 24.10.11 700 7.08
7.36 6.60
0.76 11.49 7.61
-3.28
42 17
1
D 51000 59.4 30.10.11 700 8.49
8.77 6.20
2.57 41.48 8.74
0.41
43 18
1
D 15428 57.92 27.10.11 1041 5.56
5.60 5.20
0.40
7.68 5.97
-6.26
44 19
1
D 65700 78.29 25.10.11 950 8.24
8.57 8.20
0.37
4.51 8.46
1.28
45 20
1
U 56305 37.86 26.10.11 405 9.35
9.35 8.60
0.75
8.70 9.42
-0.72
46 21
1
U 43159 45.4 06.11.11 551 8.24
7.63 7.00
0.63
9.07 7.72
-1.13
47 22
1
D 165441 64.65 25.10.11 786 8.23
8.58 7.40
1.18 15.96 8.46
1.38
*Data example shows only last filling. All tank filling data were evaluated for the fuel efficiency calculations.
157
158
Appendix
B. Explanation of Variables
Variable
n
id
fe_oem
Explanation
Number of participants
The id number of the driver
Sorts the drivers into
0 = Control Group (No scale, just eco-driving tips)
1 = Treatment Group 1 (Discrete scale)
The fuel type of the filling
Total amount of km shown on the board computer at the
time of the filling
Amount of liters filled up (in liters)
Date of the filling
Calculated difference of the value of km_fil of two consecutive fillings
The fuel efficiency for each filling is calculated as
l/100km
Average fuel efficiency in the treatment time
(24.10.2011 - 16.12.2011) (in l/100km) per driver
Average Fuel Efficiency defined by car manufacturer
diff_oem_trtmt_t
Difference between avg_fe_trtmt_t and fe_oem
diff_oem_trtmt_t_perc
Percentile difference between avg_fe_trtmt_t and
fe_oem
Average Fuel Efficiency for the whole year until
23.10.2011 (in l/100km) per driver
Difference of the average fuel efficiency between the
treatment time and the whole year until 23.10.2011 (in
l/100km)
id_trtmt_grp
f_type
km_fil
l_fil
date_fil
km_drvn
fe
avg_fe_trtmt_t
avg_fe_y
diff_trtmt_avg_fe_y
Appendix
159
Appendix E: Data set Example from DriveGain Application Data Logger
Name of driver: Driver 1
Trip 1
25.10.2011
05:38
Trip 2
26.10.2011
04:47
Trip 3
26.10.2011
15:34
Trip 4
26.10.2011
15:54
Bahnhofstr.,
Lengnau (BE)
Fabrikstr.,
Lengnau (BE)
Weiningen
(ZH)
Badenerstr.,
Dittwil
Vauffelin
Birrhard
Wettingen
Kestenholz
Journey
Score
65.32
60.24
70.9
80.37
Distance (m)
6680
80443
11506
49272
Duration (s)
459
3249
900
2498
Fuel (l)
0.48
7.44
1.69
4.7
Vehicle
Type
Audi A3
Sportback 2.0
Audi A3
Sportback 2.0
Audi A3
Sportback 2.0
Audi A3
Sportback 2.0
Fuel accel
0
0.4
0.54
0.01
Fuel brake
0.04
0.21
0.12
0.16
Fuel speed
0.05
2.34
0.23
1.43
100
0
802
498
359
3249
98
2000
Time
From
To
Fuel savings
meter (s)
Advanced
savings (s)
160
Appendix
Appendix F: Eco-Driving Smartphone Application Post Survey
Dear Colleague,
The goal of this questionnaire is to receive feedback from you about the usage of
the eco-driving DriveGain application, its functionalities and your preferences regarding different feedback forms (section I and II). In addition, questions relate to
your perception on sustainability, your technological affinity and demographic aspects (section III, IV, V).
Duration: Five sections. Around 12 minutes (pre-tests have been conducted).
Answer Options: Multiple Choice (1 or more options)
Scale (Select one option: Strongly disagree, Disagree, Partly disagree, Neutral, Partly agree, Agree, or Strongly agree)
Evaluation and Confidentiality: No intention to relate any answer to you as an
individual. All answers are confidential and only an anonymzed summary of the
answers will be presented. I would be grateful though if you would provide your
email address at the end of the survey.
Many thanks for your support.
I. Questions about the DriveGain Application
1. How often did you use the application per week during the field test on
average?
Never (0; 0%) 3-4x (6; 25%) 7-8x (9; 38%) more than 10x per week (1; 4%)
1-2x (2; 8%)
5-6x (4; 17%) 9-10x (2; 8%)
2. For which routes and travel purpose did you use the application?
(Multiple selections possible)
Motorway (23; 96%)
Business related travel (23; 96%)
Country/ rural road (18; 75%)
For long distances, more than 40 km (22; 92%)
Inner cities (10; 42%)
For short distances, less than 40 km (11; 46%)
Private related travel (12; 50%)
3. The position of the application at the window negatively influenced my
attention level to the road
 = 2.67
4. How often did you check your journey score in the DriveGain online portal
per week during the 8 weeks of the field test?
Never (2; 8%) 3-4x (2; 13%) 7-8x (1; 4%)
more than 10x per week (0; 0%)
1-2x (17; 71%) 5-6x (1; 4%)
9-10x (0; 0%)
Appendix
161
5. How often did you discuss your driving behavior with your colleagues/
friends/ family during the 8 weeks of the field test?
Never (5; 21%) 3-4x (11; 46%) 7-8x (0; 0%)
more than 10x per week (0; 0%)
1-2x (6; 25%)
5-6x (2; 8%)
9-10x (0; 0%)
II. Questions about Feedback Types/ Meters
II A: Feedback by DriveGain Application
1. The feedback about my eco-driving behavior was presented clear and
understandable.
 = 5.29
2. I am more aware of my driving behavior after receiving eco-driving
feedback by the smartphone application.
 = 4.79
3. The feedback given by the application helped to change my driving
behavior.
 = 4.63
4. The feedback given by the application influenced me in…
… a smoother acceleration.
 = 5.13
…smoother braking.
 = 4.83
...reducing my average speed on motorways.
 = 2.83
…shifting gears earlier during acceleration. (please do not answer if you have an automatic
 = 5.45
gear-box.)
…driving in the highest gear as long as possible. (please do not answer if you have an automatic
 = 4.36
gear-box.)
5. The feedback I received from the application distracted me during driving.
 = 3.08
6. Below you can see the feedback meters of the application:
162
Appendix
Figure 75: Fuel Savings Meter
Figure 76: Advanced Savings Meter
Have you been aware of the option to change the feedback
Yes (16, 67%)
No (8, 33%)
Items
I mostly chose to receive feedback via
the “Fuel Savings” (feedback every three
 = 2.41
minutes; left figure).
I mostly chose to receive feedback via
the “Advanced savings” (real-time feedback; right figure).
 = 6.06
I used both feedback meters evenly.
 = 2.22
Receiving feedback about the last three
minutes of my driving behavior was use = 2.00
ful.
Receiving real-time feedback about my
 = 5.06
driving behavior was useful.
7. I received enough feedback about how I drive from the DriveGain
application.
 = 4.71
8. I would like to receive more feedback about how I drive.
 = 4.13
II B: Form and frequency of feedback
Appendix
163
1. I would like to receive feedback from the application about how I
drive …
 = 3.33
…by visualization
 = 6.00
…by a computerized voice and visualization
 = 3.33
…by a weekly email
 = 4.88
 = 3.21
2. I would prefer to receive feedback from the application when I request it
(=pull) rather than receiving feedback automatically (=push).
 = 3.21
1. I would like to receive feedback from the application about how I
drive in comparison to…
…drivers with the same car model.
 = 6.00
…drivers with the same driving style.
 = 3.42
…drivers who drive on average a similar amount of km.
 = 5.63
…drivers from my organization.
 = 2.92
...drivers with the same job role in my organization.
 = 3.83
4. I would like to receive feedback from the application about how I drive
(multiple selections possible):
Real-time during driving (21; 88%)
Summarized once a month (11; 46%)
Immediately after driving (12; 50%) Summarized once every 3 months (3; 13%)
Summarized once a day (2; 8%)
Summarized once a year (3; 13%)
Summarized once a week (15; 63%)
III. Environmental Behavior Scale
The validated scale evaluates your environmental behavior. I would appreciate if
you answer all questions (including 5 to 9), since it will provide a complete picture
of your ecological perception.
Items
1. To drive sustainable is an important part of my
 = 5.00
personality.
2. Through sustainable driving, I would like to show to
 = 3.50
others that I am an ecologically-aware person.
3. I like to talk about my fuel efficiency.
 = 4.60
4. I discuss sustainable driving tips and new
 = 4.30
technologies (e.g. electric vehicles) with others
5. In hotels I let my towels to be changed every day.
 = 6.20
164
Appendix
6. If I wear a shirt for one day, it will always be
washed.
7. At home when I leave the room, I switch off the
light.
8. At home I leave electrical equipment, such as TV,
printer, coffee machine, etc., running on standby mode.
9. In winter it is warm enough in my flat/ house to wear a
T-Shirt.
Total Average
 = 5.00
 = 6.10
 = 5.15
 = 5.40
 = 5.0
IV. Technological Affinity
Items
1. My friends and colleagues often ask my opinion
 = 5.00
about new technologies (e.g. smartphones).
2. My friends and colleagues usually know more
 = 4.60
about new technologies than I do.
3. I am always up to date with new technological
 = 5.00
improvements which interest me.
4. I have fun to test new technologies.
 = 5.80
5. For me technology is only a means to an end
 = 5.10
(= dient nur als Mittel zum Zweck).
Total Average
 = 5.10
V. Demographic Questions
1. What is your gender?
Male (18; 75%)
2. What is your age?
Below 30 (1; 4%)
30 - 34 (4; 18%)
Female (6; 25%)
40 - 44 (6; 25%)
45 - 49 (7; 28%)
50 - 54 (3; 13%)
55 - 60 (3; 13%)
3. How many years of driving experience do you have?
0 - 9 (0; 0%)
15 - 19 (1; 4%)
25 - 30 (8; 33%)
10 - 14 (5; 21%)
20 - 24 (6; 25%)
More than 30 years (4; 17%)
4. How often do you drive?
Less than once a month (0; 0%)
1 - 2x per month (0; 0%)
Every week (0; 0%)
5 - 6x per week (5; 21%)
1 - 2x per week (0; 0%)
3 - 4x per week (1; 4%)
Every day (18; 75%)
Appendix
5. How many km on average do you drive per year?
Less than 6000 km (0; 0%)
Between 24.000 - 30.000 km (4; 17%)
Between 6000 - 12.000 km (0; 0%)
Between 30.000 - 36.000 km (10; 42%)
Between 12.000 - 18.000 km (0; 0%)
More than 36.000 km per year (8; 33%)
Between 18.000 - 24.000 km (2; 8%)
6. Which of the following driving styles do you think best describes
yours?
Ecological (14; 58%)
Aggressive (1; 4%)
Conservative (6; 25%)
Risky (1; 4%)
Sportive (14; 58%)
Depends on the situation (14; 58%)
7. Are you available for a short follow up interview (max. 20 Min) to gain
more ‘qualitative’ understanding?
Yes (17; 71%)
No (7; 29%)
8. What is your E-Mail address?
All data will be anonymzed, but for a better evaluation it would be beneficial to
relate the data collected from the DriveGain application to the answers from this
survey.
Thank you for your time and support!
165
166
Appendix
Appendix G: Eco-Driving Smartphone Application Post Interviews
Interview Dates
Corporate Car Driver
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Date
March 22nd, 2012
March 26th, 2012
March 28th, 2012
April 03rd, 2012
April 12th, 2012
April 13th, 2012
April 16th, 2012
April 18th, 2012
April 20th, 2012
April 23rd, 2012
April 25th, 2012
April 28th, 2012
April 29th, 2012
April 30th, 2012
April 30th, 2012
Interview Questions
A. DriveGain Application
1. What do you think was the greatest benefit of the application?
2. Besides the feedback you have received, which further feedback criteria are relevant for you?
3. Do you think a monthly e-mail with a comparison of your FE to another driver
would be useful?
i. If yes, what kind of comparison would you prefer?
4. Which other aspects would be relevant for you to use the application regularly in
your company car?
B. Conceptual Aspects
5. How should the management roll out a fuel improvement/ CO2 reduction initiative within the organization?
i. What steps do you feel are crucial for implementation?
ii. Where do you see problems?
6. Any further remarks/ questions?
Appendix
All categorized answers from interviews:
Questions
1. Driver Level
2. Management Level
1. What do you
Awareness:
think was the
- Less about the feedback as
greatest benefit of
such, more about using the
the application?
application (15)
- Raise awareness (13)
- Sensitize (11)
- Learn/ familiarize ecodriving practices (5)
- Testing a new mobile
technology (4)
2. Besides the
Feedback:
feedback you
- Real time feedback (15)
have received,
- Should not distract from
which further
driving (12)
feedback criteria
- Transparent feedback (12)
are relevant for
- If per voice, it should
you?
provide various driving tips (9)
- Comparison with other
colleagues (9)
- More feedback related to
eco-driving techniques (8)
- Weekly feedback per email
with summarized fuel
consumption information (8)
- Feedback has to show
progress (8)
- Individual feedback
according to the
driving type (7)
3. Do you think a - Yes (15)
monthly e-mail
- Comparison with the same
with a comparicar model type (15)
son of your FE to
- Same amount of total km per
another driver
year (14)
would be useful?
- Transparent comparison (10)
- Similar driver type (8)
If yes, what kind
- The same job profile is not so
of comparison
important (3)
would you prefer? - Same route (2)
167
168
Appendix
4. Which other
aspects would be
relevant for you
to use the application regularly in
your corporate
car?
Time importance:
Provide Information:
- Everyday on the road (15)
- Provide personalized
- Important to be home fast (15)
information (14)
- Many customer visits per
- Expose fuel costs (13)
day (10)
- Expose fuel efficiency (9)
- Drive faster if no speed
- CO2 Emission (6)
regulations (8)
Incentives:
Rewards:
- High influence financial
- Realistic/ fair rewards (12)
incentives (12)
- Bonus point program (9)
- Incentives through deciding
- Bonus Points difficult (4)
to get a car with lower CO2
emission (8)
- Becoming the eco-driver of
the month (4)
5. How should
Punishment/ Control vs.
the management
Awareness:
roll out a fuel im- No punishment (14)
provement/ CO2
- Unfair to punish drivers
reduction initiawho need their car for daily
tive within the
business (13)
organization?
- No control from
management (12)
i) What steps do
- Sensitize corporate car
you feel are crudrivers, since no
cial for impleunderstanding of any fuel
mentation?
consumption (comparison
with mobile phone) (9)
ii) Where do you
Goal setting
see problems?
- Goal setting necessary (12)
- Realistic goals related to
corporate car drivers (11)
- Goals enhance motivation (7)
- Goals not important (3)
Driving/ Routes
- Consider different routes (9)
- Depends on drivers driving
behavior, sometimes no
further reductions are
possible (7)
6. Any further
remarks/ questions?
- What are the next steps (12)
- Application updated to our
needs (8)
Appendix
169
Appendix H: Monthly Personal Car Consumption Email with discrete Scale
FOR INTERNAL USE ONLY
Your personal Car Consumption and
EcoDriving Tips
Dear Name of Driver,
Please review your personal fuel emissions. Your average fuel consumption per month
(liters per 100 km) until end of October 2011, compared to the average consumption recommended by your car manufacturer, is in the middle area:
- Average fuel consumption: 8.5 Liters per 100 km (difference of -0.2 Liters compared to
the previous month)
- Difference of 19% between recommended and actual fuel consumption
- Average kilometers driven: 3025 km per month
- Average CO2 emissions: 168 gram CO2 per km
Recommendations, how your fuel efficiency can be improved through eco-driving:
1. Accelerate quickly and upshift to next gear by latest 2500 rpm (petrol) or 2000 rpm
(diesel) (applicable for manual gear only);
2. Downshift as late as possible (applicable for manual gear only);
3. Drive with the highest gear as possible (applicable for manual gear only);
4. Drive anticipatory and smooth: do not accelerate aggressively and maintain constant
speed (applicable for manual and automatic gears);
5. Use cruise mode when possible (applicable for automatic gear only);
6. Check your tire pressure monthly (applicable for manual and automatic gears).
Best regards,
Johannes
SAP Research
Note: Due to the low average fuel consumption figures stated by car manufacturers and in
order to improve the comparison between the drivers, the categories in the graphic above
are based on average fuel consumption values from eligible SAP (Switzerland) AG corporate car and car allowance drivers. These drivers were randomly selected for the field test.
The consumption rates are taken from the data received from the petrol cards, are treated
highly confidential and are calculated individually for each driver. Please contact Johannes Tulusan for any further questions or recommendations.
Unsubscribe | Subscribe | Copyright/Trademark | Privacy | Impressum
Appendix I: Environmental Behavior Survey
170
Appendix
I. Environmental Behavior / Umweltbewusstsein
Items
1. To drive sustainable is an important part of my
 = 5.50
personality.
2. Through sustainable driving, I would like to show to
 = 4.66
others that I am an ecologically-aware person.
3. I like to talk about my fuel efficiency.
 = 3.65
4. I discuss sustainable driving tips and new technologies
 = 4.48
(e.g. electric vehicles) with others
5. In hotels I let my towels to be changed every day.
 = 5.99
6. If I wear a shirt for one day, it will always be washed.
 = 4.95
7. At home when I leave the room, I switch off the light.
 = 6.14
8. At home I leave electrical equipment, such as TV,
 = 5.26
printer, coffee machine, etc., running on standby mode.
9. In winter it is warm enough in my flat / house to wear a
 = 5.33
T-Shirt.
Total Average
 = 5.11
II. Technological Affinity (= Technologieaffinität)
Please evaluate your perception of statements 1 to 5 by choosing one of the answer
options:
Items
1. My friends and colleagues often ask my opinion about
 = 4.62
new technologies (e.g. smartphones).
2. My friends and colleagues usually know more about
 = 3.42
new technologies than I do.
3. I am always up to date with new technological
 = 4.92
improvements which interest me.
4. I have fun to test new technologies.
 = 5.44
5. For me technology is only a means to an end
 = 4.19
(= dient nur als Mittel zum Zweck).
Total Average
 = 4.52
References
171
References
1.
Agostinelli, G., Brown, J.M. and Miller, W.R. (1995), “Effects of normative feedback on consumption among heavy drinking college students,” Journal of
drug education, Vol. 25 No. 1, pp. 31–40.
2.
Aitken, R.C. (1969), “Measurement of feelings using visual analogue scales.,” Proceedings of the royal society of medicine, Vol. 62 No. 10, p. 989.
3.
Ajzen, I. (1991), “The theory of planned behavior,” Organizational behavior
and human decision processes, Vol. 50 No. 2, pp. 179–211.
4.
Ajzen, I. and Fishbein, M. (1977), “Attitude-behavior relations: A theoretical
analysis and review of empirical research.,” Psychological bulletin, Vol. 84
No. 5, p. 888.
5.
Alder, G.S. (2007), “Examining the relationship between feedback and performance in a monitored environment: A clarification and extension of feedback intervention theory,” The Journal of High Technology Management Research, Vol. 17 No. 2, pp. 157–174.
6.
Alder, G.S. and Tompkins, P.K. (1997), “Electronic performance monitoring:
An organizational justice and concertive control perspective.,” Management
Communication Quarterly, Vol. 10, pp. 259–288.
7.
Amabile, T.M. (1993), “Motivational synergy: Toward new conceptualizations of intrinsic and extrinsic motivation in the workplace,” Human Resource
Management Review, Vol. 3 No. 3, pp. 185–201.
8.
Angel, N.F. (1998), “Evaluating employees by computer,” Personnel Administrator, Vol. 34 No. 11, pp. 67–72.
9.
Arroyo, E., Sullivan, S. and Selker, T. (2006), “CarCoach: a polite and effective driving coach.,” Proceedings of the CHI, Presented at the Conference on
Human Factors in Computing Systems, pp. 357–362.
10. Baeriswyl, M., Przepiorka, W. and Staake, Thorsten. (2011), “Identifying individuals’ preferences using games: A field experiment in promoting sustainable energy consumption,” Proceedings of International Conference on Information Systems. from https://edit.ethz.ch/im/people/bmichi/BaeriswylIdentifying_Preferences_using_Games_2011.pdf
11. Bahnmüller, K. (2012), “Ein Jahr der Freude,” Schweizer Fahrzeugmarkt, Vol.
12 No. 1, p. 40.
12. Balcazar, F., Hopkins, B.L. and Suarez, Y. (1986), “A critical, objective review of performance feedback,” Applied Psychology, Vol. 66, pp. 314–323.
13. Bamberg, S., Ajzen, I. and Schmidt, P. (2003), “Choice of travel mode in the
theory of planned behavior: The roles of past behavior, habit, and reasoned action,” Basic and applied social psychology, Vol. 25 No. 3, pp. 175–187.
14. Bamberg, S. and Schmidt, P. (2003), “Incentives, morality, or habit? Predicting students’ car use for university routes with the models of Ajzen, Schwartz,
and Triandis,” Environment and behavior, Vol. 35 No. 2, pp. 264–285.
172
References
15. Baron, J. (2008), “Normative Models of Judgment and Decision Making,” pp. 19–36.
16. Belz, A. and Kow, E. (2011), “Discrete vs. continuous rating scales for language evaluation in nlp,” Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics (ACL-HLT’11). Retrievedfrom
http://www.newdesign.aclweb.org/anthology-new/P/P11/P11-2040.pdf
17. Bernstein, I.H. and Nunnally, J.C. (1994), Psychometric theory, New York:
McGraw-Hill.
18. Beswick, D. (2007), “Intrinsic Motivation and Extrinsic Rewards.” Retrieved September
18,
2012,
from
http://www.beswick.info/psychres/management.htm
19. Black, T.R. (1999), Doing Quantitative Research in the Social Sciences: An
Integrated Approach to Research Design, Measurement and Statistics, Sage
Publications Ltd, 1sted.
20. Boriboonsomsin, K., Vu, A. and Barth, M. (2010), “Eco-Driving: Pilot Evaluation of Driving Behavior Changes among US Drivers,” Faculty Research,
University of California Transportation Center, UC Berkeley, pp. 17–26.
21. Bostrom, R. and Heinen, J.S. (1977), “Problems and Failures: A SocioTechnical Perspective,” MIS quarterly, Vol. 1 No. 3.
22. Burton-Jones and Gallivan. (2008), “Toward a Deeper Understanding of System Usage in Organizations: A Multilevel Perspective,” MIS quarterly, Vol.
31 No. 4, pp. 657–679.
23. Cengel, Y.A. and Boles, M.A. (2002), Thermodynamics: An Engineering Approach, McGraw-Hill, New York.
24. Chapnick, N. (2007), “Hypermiling: Quest for Ultimate Fuel Economy.” Retrieved
January
26,
2012,
from
http://www.edmunds.com/cars/fueleconomy/articles/120880/article.html
25. Cialdini, R.B. (2003), “Crafting normative messages to protect the environment,” Current directions in psychological science, Vol. 12 No. 4, pp. 105–
109.
26. Cialdini, R.B., Kallgren, C.A. and Reno, R.R. (1991), “A focus theory of normative conduct: A theoretical refinement and reevaluation of the role of norms
in human behavior,” Advances in experimental social psychology, Vol. 24 No.
20, pp. 1–243.
27. Climate Company. (2011), “Environmental zones in Germany.” Retrieved January
26,
2012,
from
http://www.umweltplakette.de/int_england.php?SID=js9eb7lu4kg3kcssqi8vq1ijq4
28. Corbett, A.T. and Anderson, J.R. (2001), “Locus of feedback control in computer-based tutoring: Impact on learning rate, achievement and attitudes,” Proceedings of the SIGCHI conference on Human factors in computing systems, pp. 245–252.
References
173
29. Cornford, T. and Smithson, S. (2006), Project Research in Information Systems: A Student’s Guide, Palgrave Macmillan.
30. Darby, S. (2006), The effectiveness of feedback on energy consumption: a review for DEFRA of the literature on metering, billing and direct displays., pp.
1–21.
Retrievedfrom
http://www.eci.ox.ac.uk/research/energy/electricmetering.php
31. Dataforce. (2011), “Fuhrparkmanagement in deutschen Unternehmen.” Retrieved January 26, 2012, from http://www.dataforce.de/web/guest/infocenter
32. Deci, E.L. and Ryan, R.M. (1985), Intrinsic Motivation and SelfDetermination in Human Behavior, Springer, 1sted.
33. Department for Transport, G.M.H. (2011), “Transport Statistics Great Britain 2011,” Web page, . Retrieved September 19, 2012, from http://www.dft.gov.uk/statistics/releases/transport-statistics-great-britain-2011/
34. Deslauriers, B.C. and Everett, P.B. (1977), “Effects of intermittent and continuous token reinforcement on bus ridership.,” Journal of Applied Psychology,
Vol. 62 No. 4, p. 369.
35. Donmez, B. (2007), “Safety implications of providing real-time feedback to
distracted drivers,” Accident analysis and prevention, Vol. 39 No. 3, p. 581.
36. Doppelt, B., Markowitz, E.M. and Initiative, C.L. (2009), “Reducing Greenhouse Gas Emissions Through Behavioral Change: An Assessment of Past Research On Energy Use, Transportation and Water Consumption.”
37. DriveGain Ltd. (2012), “DriveGain.” Retrieved September 19, 2012, from http://drivegain.com/about
38. DTLR. (2001), “Department for Transport: Transport statistics 2000.” Retrieved
January
26,
2012,
from
http://dftsearch.dft.gov.uk/client=default_frontend&proxyreload=1&proxystyl
esheet=default_frontend&output=xml_no_dtd&q=transport+statistics+2000&
pg=1&Search.x=0&Search.y=0
39. Dula, C.S. and Geller, E.S. (2003), “Risky, aggressive, or emotional driving: Addressing the need for consistent communication in research,” Journal of
Safety Research, Vol. 34 No. 5, pp. 559–566.
40. EcoDrive. (2012), “The Golden Rules of EcoDriving.” Retrieved January 26, 2012,
from
http://www.ecodrive.org/en/what_is_ecodriving/benefits_of_ecodriving/
41. Eisenhardt, K.M. (1989), “Building theories from case study research,” Academy of management review, Vol. 14 No. 4, pp. 532–550.
42. Enerdata. (2011), “World Energy Statistics | World Energy Consumption & Stats| Enerdata.” Retrieved January 26, 2012, from http://yearbook.enerdata.net/
43. European Commission. (2010), “Energy: Energy Efficiency Directive - European commission.” Retrieved January 26, 2012, from http://ec.europa.eu/energy/efficiency/eed/eed_en.htm
174
References
44. Flick, U. (2006), An Introduction to Qualitative Research, Sage Publications
Ltd, Third Edition.
45. Fogg, B.J. (2002), “Persuasive technology: using computers to change what we think and do,” Ubiquity, Vol. December, p. 5.
46. Formosa, D. (2009), “Ford SmartGauge: Designing an Extra 9 MPG?,” Smart
Design.,from
www.smartdesignworldwide.com/pdf/SmartDesign_SmartGauge.pdf
47. Forward, S. (2004), “The prediction of travel behaviour using the Theory of Planned Behaviour,” International Conference of Traffic and Transport Psychology.
48. Foxx, R. and Hake, D. (1977), “Gasoline conservation: A procedure for measuring and reducing the driving of college students,” Journal of Applied Behavior Analysis, Vol. 10 No. 1, p. 61.
49. Foxx, R. and Schaeffer, M.H. (1981), “A company-based lottery to reduce the
personal driving of employees,” Journal of Applied Behavior Analysis, Vol.
14 No. 3, p. 273.
50. Freyd, M. (1923), “A graphic rating scale for teachers,” The Journal of Educational Research, Vol. 8 No. 5, pp. 433–439.
51. Froberg, D.G. and Kane, R.L. (1989), “Methodology for measuring healthstate preferences—II: Scaling methods,” Journal of clinical epidemiology,
Vol. 42 No. 5, pp. 459–471.
52. Froehlich, J., Dillahunt, T., Klasnja, P., Mankoff, J., Consolvo, S., Harrison,
B. and Landay, J.A. (2009), “UbiGreen: investigating a mobile tool for tracking and supporting green transportation habits,” Proceedings of the 27th international conference on Human factors in computing systems, pp. 1043–1052.
53. Froehlich, J., Findlater, L. and Landay, J. (2010), “The design of eco-feedback
technology,” Proceedings of the 28th international conference on Human factors in computing systems, pp. 1999–2008.
54. Frost & Sullivan. (2009), 2010 Outlook & Forecast: Mobile & Wireless
Communications.
55. Fujii, S. and Gärling, T. (2003), “Application of attitude theory for improved predictive accuracy of stated preference methods in travel demand analysis,” Transportation Research Part A: Policy and Practice, Vol. 37 No. 4, pp. 389–
402.
56. Gardner, G.T. and Stern, P.C. (2008), “The short list: The most effective actions US households can take to curb climate change,” Environment: Science
and Policy for Sustainable Development, Vol. 50 No. 5, pp. 12–25.
57. Gartner. (2010), Forecast: Mobile Devices, Worldwide 2003-2014, 2Q10 Update.
58. Gartner. (2011), “Smartphone Sales Increased 42 Percent.” Retrieved January
26, 2012, from http://www.gartner.com/it/page.jsp?id=1848514
References
175
59. Geller, E.S., Winett, R.A., Everett, P.B. and Winkler, R.C. (1982), Preserving
the environment: New strategies for behavior change, Pergamon Press Elmsford, New York.
60. German Federal Environmental Agency. (2009), Ratgeber: Sprit sparen und
mobil sein ( No. 2), p. 19.
61. Goldenbeld, C., Twisk, D. and Houwing, S. (2008), “Effects of persuasive communication and group discussions on acceptability of anti-speeding policies for male and female drivers,” Transportation Research Part F: Traffic
Psychology and Behaviour, Vol. 11 No. 3, pp. 207–220.
62. Goldstein, N.J., Cialdini, R.B. and Griskevicius, V. (2008), “A room with a viewpoint: Using social norms to motivate environmental conservation in hotels,” Journal of Consumer Research, Vol. 35 No. 3, pp. 472–482.
63. Graham, J., Koo, M. and Wilson, T.D. (2011), “Conserving energy by inducing people to drive less,” Journal of Applied Social Psychology, Vol. 41 No. 1,
pp. 106–118.
64. Graml, T., Loock, C., Baeriswyl, M. and Staake, T. (2011), “Improving Residential Energy Consumption at Large Using Persuasive Systems.,” Presented at the . 19th European Conference on Information Systems (ECIS), Helsinki,
Finland.
65. Graml, T., Loock, C., Staake, T. and Fleisch, E. (2010), “Online Energy Behavior Change Systems: What makes people stick? pdf free download,” Presented at the Persuasive Conference, Copenhagen.
66. Grant, R. and Higgins, C. (1989), “Monitoring service workers via computer: The effect on employees, productivity, and service,” National Productivity
Review, Vol. 8 No. 2, pp. 101–112.
67. Greengard, S. (1996), “Privacy: Entitlement or illusion?,” Personnel Journal,
Vol. 75 No. 5, pp. 74–76.
68. GreenRoad. (2008), “Is Safe Driving More Economical? Driver Safety and Fuel
Consumption.” Retrieved May 29, 2011, from
http://www.greenroadtech.com/documents/driver_safety_and_fuel_consumpti
on_us.pdf
69. Henriques, V.E. (1996), “In defence of computer monitoring,” Training, Vol.
75 No. 5, pp. 74–76.
70. Holmes, T.G. (2007), “Eco-visualization: combining art and technology to
reduce energy consumption,” Proceedings of the 6th ACM SIGCHI conference
on Creativity & cognition, C&;C ’07, New York, NY, USA, ACM, pp. 153–162. doi:10.1145/1254960.1254982
71. Van Houwelingen, J.H. and Van Raaij, W.F. (1989), “The effect of goalsetting and daily electronic feedback on in-home energy use,” Journal of consumer research, pp. 98–105.
72. Hussey, J. and Hussey, R. (1997), Business Research: A Practical Guide for
Undergraduate and Postgraduate Students, Macmillan.
176
References
73. Hutton, K.A., Sibley, C.G., Harper, D.N. and Hunt, M. (2001), “Modifying driver behaviour with passenger feedback,” Transportation Research Part F:
Traffic Psychology and Behaviour, Vol. 4 No. 4, pp. 257–269.
74. Johansson, H., Gustafsson, P., Henke, M. and Rosengren, M. (2003), “Impact of EcoDriving on emissions,” Transport and Air Pollution. Proceedings from
the 12th Symposium, pp. 97–105.
75. Keen, P.G.W. (1981), “Information systems and organizational change,” Commun. ACM, Vol. 24 No. 1, pp. 24–33.
76. King, G., Keohane, R.O. and Verba, S. (1994), Designing social inquiry: scientific inference in qualitative research, Princeton: Princeton University Press.
77. Kluger, A.N. and DeNisi, A. (1996a), “The effects of feedback interventions on performance: A historical review, a meta-analysis, and a preliminary feedback intervention theory.,” Psychological bulletin, Vol. 119 No. 2, p. 254.
78. Kluger, A.N. and DeNisi, A. (1996b), “The effects of feedback interventions on performance: A historical review, a meta-analysis, and a preliminary feedback intervention theory.,” Psychological bulletin, Vol. 119 No. 2, p. 254.
79. Lawson, T. (2008), Introduction to Critical Realism and Social Science: Context and Contribution., Presented at the MP0-2 Lecture, Judge Business
School, Cambridge University.
80. Lee, H., Lee, W. and Lim, Y.K. (2010), “The effect of eco-driving system towards sustainable driving behavior,” Proceedings of the 28th of the international conference extended abstracts on Human factors in computing systems,
pp. 4255–4260.
81. Loock, C., Graml, T., Baeriswyl, M. and Staake, T. (2011), “How to Motivate Energy Efficiency Online,” Presented at the Proceedings of 20 th International Conference on Management of Technology, Florida, USA.
82. Loock, C., Staake, T and Landwehr, J. (2011), “Green IS Design and Energy Conservation: An Empirical Investigation of Social Normative Feedback,” Presented at the International Conference on Information Systems (ICIS),
Shanghai, China.
83. Mankoff, Jennifer, Matthews, D., Fussell, S. and Johnson, M. (2007), “Leveraging Social Networks To Motivate Individuals to Reduce their Ecological
Footprints,” IEEE, pp. 87–87. doi:10.1109/HICSS.2007.325
84. Marshall, C. and Rossman, G.B. (1989), Designing qualitative research, Sage
Publications.
85. Meschtscherjakov, A., Wilfinger, D., Scherndl, T. and Tscheligi, M. (2009),
“Acceptance of future persuasive in-car interfaces towards a more economic
driving behaviour,” Proceedings of the 1st International Conference on Automotive User Interfaces and Interactive Vehicular Applications, pp. 81–88.
86. Van Mierlo, J., Maggetto, G., Van de Burgwal, E. and Gense, R. (2004),
“Driving style and traffic measures-influence on vehicle emissions and fuel
References
177
consumption,” Proceedings of the Institution of Mechanical Engineers, Part
D: Journal of Automobile Engineering, Vol. 218 No. 1, pp. 43–50.
87. Mingers, J. (2001), “Combining IS research methods: towards a pluralist methodology,” Information systems research, Vol. 12 No. 3, pp. 240–259.
88. Morgan Stanley. (2010), “Internet Trends,” Presented at the CM Summit, New
York City.
89. Mountain, D. (2008), Real-Time Feedback and Residential Electricity Consumption: British Columbia and Newfoundland and Labrador Pilots, Ontario,
Canada: Mountain Economic Consulting and Associates Inc.
90. National Research Council. (2010), Assessment of Technologies for Improving
Light Duty Vehicle Fuel Economy ( No. Traffic Safety).
91. Newnam, S., Watson, R.T. and Murray, W. (2004), “Factors predicting intentions to speed in a work and personal vehicle,” Transportation Research Part
F: Traffic Psychology and Behaviour, Vol. 7 No. 4-5, pp. 287–300.
92. Nussbaum, K. and duRivage, V. (1986), “Computer monitoring: Mismanagement by remote control,” Business and Society Review, Vol. 56, pp. 16–20.
93. Onoda, T. (2009), “IEA policies–G8 recommendations and an afterwards,” Energy Policy, Vol. 37 No. 10, pp. 3823–3831.
94. Orlikowski, W.J. and Baroudi, J.J. (1991), “Studying information technology in organizations: Research approaches and assumptions,” Information systems
research, Vol. 2 No. 1, pp. 1–28.
95. Pace, T., Ramalingam, S. and Roedl, D. (2007), “Celerometer and idling reminder: persuasive technology for school bus eco-driving,” CHI’07 extended abstracts on Human factors in computing systems, pp. 2085–2090.
96. Polk, M. (2003), “Are women potentially more accommodating than men to a
sustainable transportation system in Sweden?,” Transportation Research Part
D: Transport and Environment, Vol. 8 No. 2, pp. 75–95.
97. Reno, R.R., Cialdini, R.B. and Kallgren, C.A. (1993), “The transsituational
influence of social norms.,” Journal of personality and social psychology, Vol.
64 No. 1, p. 104.
98. Romm, J.J. and Frank, A.A. (2006), “Hybrid vehicles gain traction,” Scientific
American, Vol. 294 No. 4, pp. 72–79.
99. Rotter, J.B. (1966), “Generalized expectancies for internal versus external
control of reinforcement,” Psychological monographs, Vol. 80 No. 1, pp. 1–
28.
100. Schultz, P., Nolan, J.M., Cialdini, R.B., Goldstein, N.J. and Griskevicius, V.
(2007), “The constructive, destructive, and reconstructive power of social
norms,” Psychological Science, Vol. 18 No. 5, p. 429.
101. Schwartz, B., Ward, A., Monterosso, J., Lyubomirsky, S., White, K. and Lehman, D.R. (2002), “Maximizing versus satisficing: Happiness is a matter of 178
References
choice.,” Journal of Personality and Social Psychology, Vol. 83 No. 5, p.
1178.
102. Scott, W.R. and Huskisson, E.C. (1977), “Measurement of functional capacity with visual analogue scales,” Rheumatology, Vol. 16 No. 4, pp. 257–259.
103. Shipworth, M. (2000), Motivating Home Energy Action: A handbook of what
works, Canberra, Australia: Australian Greenhouse Office.
104. Siero, S., Boon, M., Kok, G. and Siero, F. (1989), “Modification of driving behavior in a large transport organization: A field experiment.,” Journal of
Applied Psychology, Vol. 74 No. 3, pp. 417–423.
105. Simon, H.A. (1955), “A behavioral model of rational choice,” The quarterly
journal of economics, Vol. 69 No. 1, p. 99.
106. Smiley, A. and Michon, J.A. (1989), “Conceptual Framework for generic intelligent driver support.,” Traffic Research Centre, Vol. University of Groningen Haren, Netherlands, p. 37.
107. Stead, M., Tagg, S., MacKintosh, A.M. and Eadie, D. (2005), “Development and evaluation of a mass media Theory of Planned Behaviour intervention to
reduce speeding,” Health Education Research, Vol. 20 No. 1, p. 36.
108. Steg, L. and Sievers, I. (2000), “Cultural theory and individual perceptions of environmental risks,” Environment and Behavior, Vol. 32 No. 2, p. 250.
109. Stenner, F. (2009), Handbuch Automobilbanken: Finanzdienstleistungen für
Mobilität, Gabler Wissenschaftsverlage.
110. Stoppani, S. (2009), “Bliss TrekTM,” Overview of the BlissTrek Application.
Retrieved August 5, 2012, from http://blisstrek.com/
111. Strauss, A.C. and Corbin, J. (1998), Basics of Qualitative Research: Second
Edition: Techniques and Procedures for Developing Grounded Theory, Sage
Publications, Inc, 2nded.
112. Svensson, E. (2000), “Comparison of the quality of assessments using continuous and discrete ordinal rating scales,” Biometrical Journal, Vol. 42 No. 4,
pp. 417–434.
113. Tasca, L. (2000), “A review of the literature on aggressive driving research,” Presented at the Aggressive driving issues conference, p. 25.
114. Tester, J., Fogg, B.J. and Maile, M. (2000), “CommuterNews: a prototype of persuasive in-car entertainment,” CHI’00 extended abstracts on Human factors in computing systems, pp. 24–25.
115. Thambusamy, R. and Salam, A.F. (2010), “Corporate Ecological Responsiveness, environmental ambidexterity and IT-enabled environmental sustainability strategy,” Presented at the International Conference on Information Systems, Lima.
116. Thogersen, J. and Moller, B. (2008), “Breaking car use habits: The effectiveness of a free one-month travelcard,” Transportation, Vol. 35 No. 3, pp. 329–
345.
References
179
117. Toulmin, S.E. (1960), The Philosophy of Science: An Introduction, Harper &
Row, Harper torchbooks.
118. Tversky, A. and Kahneman, D. (1986), “Rational choice and the framing of decisions,” Journal of business, pp. 251–278.
119. Ullmenstein, T. (2012,September13), “Ökologie wie Ökonmie ins Spiel bringen,” Neu Zürcher Zeitung, p. 11.
120. Van Velsor, E., Leslie, J.B. and Fleenor, J.W. (1997), Choosing 360. A Guide
to Evaluating Multi-Rater Feedback Instruments for Management Development., Center for Creative Leadership, Greensboro,.
121. Van der Voort, M., Dougherty, M.S. and Van Maarseveen, M. (2001), “A prototype fuel-efficiency support tool,” Transportation Research Part C: Emerging Technologies, Vol. 9 No. 4, pp. 279–296.
122. Wahlberg, A. (2007), “Aggregation of driver celeration behavior data: Effects on stability and accident prediction,” Safety science, Vol. 45 No. 4, pp. 487–
500.
123. Watson, R.T., Boudreau, M.C. and Chen, A.J. (2010), “Information systems and environmentally sustainable development: energy informatics and new directions for the IS community,” MIS quarterly, Vol. 34 No. 1, p. 23.
124. Williams, M. (2000), “Interpretivism and Generalisation,” Sociology, Vol. 34
No. 2, pp. 209–224.
125. Winkler, R.C. and Winett, R.A. (1982), “Behavioral interventions in resource conservation: A systems approach based on behavioral economics.,” American
Psychologist, Vol. 37 No. 4, p. 421.
126. Worsnop, R.L. (1993), Privacy in the workplace, CQ Press.
127. Yin, R.K. (2008), Case Study Research: Design and Methods, SAGE Publications.
128. De Young, R. (1985), “Encouraging environmentally appropriate behavior: The role of intrinsic motivation,” Journal of Environmental Systems, Vol. 15
No. 4, pp. 281–292.
129. Zahorsky, D. (2010), “Ignite the Two Fires of Employee Motivation.” Retrieved
September
18,
2012,
from
http://sbinformation.about.com/cs/benefits/a/033003.htm
180
References
Publications
Tulusan, J., Steggers, H., Staake, T., Fleisch, E., Supporting eco-driving with ecofeedback technologies: Recommendations targeted at improving corporate car drivers’ intrinsic motivation to drive more sustainable, Energy Informatics 2012 (EI
2012), Atlanta, Georgia, United States, October 2012.
Tulusan, J., Staake, T., Fleisch, E., Direct or indirect sensor enabled eco-driving
feedback: Which preference do corporate car drivers have?, Internet of Things 2012
– Third International Conference on the Internet of Things (IoT 2012), Wuxi, P.R.
China, October 2012.
Tulusan, J., Staake, T., Fleisch, E., Providing eco-driving feedback to corporate car
drivers: what impact does a smartphone application have on their fuel efficiency,
14th ACM International Conference on Ubiquitous Computing (UbiComp), Pittsburgh, Pennsylvania, United States, September 2012.
Tulusan, J., Soi, L., Paefgen, J., Staake, T., Eco-efficient feedback technologies:
Which eco-feedback types prefer drivers most?. World of Wireless, Mobile and
Multimedia Networks IEEE Conference, Lucca, Italy, June 2011
Tulusan, J., Brogle, M., Staake, T., Fleisch, E., Becoming a Sustainable Driver: The
Impact of Mobile Feedback Devices. 4th ERCIM eMobility WWIC Conference,
Lulea, Sweden, June 2010
Curriculum Vitae
181
Curriculum Vitae
Personal Data
Johannes Tulusan
Born 29th of October 1979 in Erlangen (Germany)
Education
01/2010 – 12/2012
University of St. Gallen
Institute of Technology Management, St.Gallen (CH)
PhD Student in Management
10/2008 – 12/2009
Cambridge University
Judge Business School, Cambridge (UK)
PhD Candidate in Management (First year)
09/2006 – 09/2007
London School of Economics & Political Science
IS and Innovation Department, London (UK)
MSc ADMIS (Analysis, Design and Management of
Information Systems)
09/2000 – 06/2003
Brunel University
Management Department, London (UK)
BSc (Hons) Business and Computing
09/1994 – 06/1999
Work Experience
01/2013 – now
01/2010 – 12/2012
Richard-Wagner-Secondary School
Bayreuth (Germany)
Graduated with German Abitur
SAP AG, Munich / Walldorf (Germany)
Business Development Manager (full time)
Strategic Business Development Department
SAP AG, Zurich (Switzerland)
Research Associate (full time)
SAP Research Switzerland
182
10/2007 – 09/2008
Curriculum Vitae
Viewmy.TV Ltd., London (UK)
Head of Business Development (full time)
Headquarter London
03/2004 – 03/2006
SAP AG, Walldorf (Germany)
Business Development Specialist (full time)
Advisory Office EMEA
01/2004 – 02/2004
Siemens Medical Services, Philadelphia (USA)
Healthcare IT Consulting Internship
‘Sorian’ Consulting Department
08/2003 – 12/2003
SAP AG, Walldorf (Germany)
Technology Consulting Internship
Technology Consulting II Department
06/2002 – 09/2002
Siemens Business Services, Singapore
Business and IT Consulting Internship
Application Hosting Department
07/2001 – 09/2001
Suse Linux Ltd., Borehamwood (United Kingdom)
Sales Internship – Sales and Marketing Office
07/1998 – 10/2000
Courissima Ltd., Bayreuth (Germany)
Co-founder and Business Development Manager (full
time)
Awards and Scholarships
01/2010 – 12/2012
PhD funded from the cooperation bursary of SAP
and canton St. Gallen.
10/2008 – 12/2009
Economic and Social Research Council Full Scholarship.
10/2008 – 12/2009
Cambridge European Trust Scholarship.
10/2007
MSc dissertation result award.
Author: Mr. Johannes Tulusan
Title:
Combining ICT with Eco-driving Concepts to Improve Corporate Car
Drivers’ Fuel Efficiency

Combining ICT with Eco-driving Concepts to Improve Corporate Car

Transcription

Similar documents

DP 09-016

speaker business expert