The Role of Sociograms in Social Network Analysis
Transcription
The Role of Sociograms in Social Network Analysis
THE ROLE OF SOCIOGRAMS IN SOCIAL NETWORK ANALYSIS Mar yann Durland Ph.D. EERS Conference 201 2 Monday April 20, 10:30 -12:00 FORMAT OF PRESENTATION Part Part Part Part Part I SNA overview – 10 minutes II Sociograms Example 1 - 25 minutes III History & Development – 10 minutes IV Sociograms Example 2 – 25 minutes V Questions &Answers – 20 minutes PART I - WHAT IS SNA? Methodology for analyzing relational data Uses matrix data for analysis (graph theory, matrix algebra) Uses sociograms or maps for visualization and data (graph theory, visualization of data) Uses both simple & complex algorithms (matrix algebra, graph theory) Is focused on the whole -structure, patterns, systems (history) Looks at both the whole & parts within context of whole (history) Has multiple applications aligned with theories (support, structural holes, leadership, diffusion) 3 RELATIONAL DATA Relational data means that in some form or another the data measures a connection between two points. The two points might be cities and the connection could be railroad lines, or one way streets, or stop lights, or friendships, co-memberships, support, etc. SNA begins with the connection or absence of a connection between two points Points might be towns, street corners, staff, a classroom, members of a group, a team, an organization, etc. T YPES OF RELATIONSHIPS People to people, but also: Activities Actions Events Materials & Resources Ideas etc. 5 RELATIONAL DATA COMES FROM RELATIONSHIPS coaching Communities of practice exchange kinship Diffusion of methods power groups Teams gossip trust mentoring Matrix org Sharing info communication Working Relationships innovation Knowledge networks Leadership Communities of Practice advice conflict friendship influence Club membership Research 6 RELATIONSHIPS Defining Relationships Based on theory – support, leadership Based on behavior – defined by program or project Combination of both Measuring Relationships Align measure to behavior Use multiple measures Look at all levels (as results indicate, or aligned to theory) Network Sub-groups Individual 7 IMPORTANCE OF SOCIOGRAMS IN SNA Foundation for SNA - History Has features or characteristics that can and cannot be measured (graph theory) Adds context to measures (graph theory, program theory, social science theory) Locates the position within a network, based on a measure, which is different from a rank or score. PART II READING SOCIOGRAMS Nodes & lines Direction/or not Coded by attributes or measurement This is a tree This line is a bridge Person 2, 3, and 1 form a clique Person 2 is a cutpoint • • • • • Square Nodes Non-directional lines No attribute or measurement info One network (person 1-5) Person 2 has an Indegree of 3. Persons 3, and 1 have an Indegree of 2. Who is in a better or more critical position? 9 EXAMPLE 1: COLLABORATION AND SPREAD OF TEAMS WITHIN NETWORK Frequency of connections How are teams connected What is the structure of connections (pairs, cliques, ego network, etc.) How is the overall structure organized? (grade levels, peers, across grades) 10 SOCIOGRAMS Sociograms are created by software programs that use algorithms to plot the nodes and ties or lines. Among other elements in the formulae, these algorithms include four basic “rules” for creating a sociogram; push unconnected nodes away from each other, pull connected nodes closer together, adjust for line length, so that lines are not too long or too short, and position nodes and connections to minimize line crossings. SCHOOL 1 – THE BIG PICTURE Team members are yellow Highly connected to each other, and central to the network (connected to others as well) Edges of sociogram– nonrespondents, few choices, etc. 12 RECIPROCAL VIEW – CODED BY AREAS The bold lines indicate mutual connections. Square nodes are team members, within their grade levels. This map is by grade levels (Primary, elementary and middle). Many reciprocals are by grade level, but not all. 13 60-61 Pre-k, K 98 Other 99 AP 100 Principal 14 CLIQUE DATA 1:00 ORG1 ORG2 ORG5 ORG3 ORG4 2:00 ORG1 ORG2 ORG3 ORG4 ORG6 3:00 ORG1 ORG4 ORG6 ORG7 4:00 ORG1 ORG4 ORG8 5:00 ORG1 ORG11 ORG12 6:00 ORG1 ORG12 ORG13 7:00 ORG1 ORG14 ORG15 8:00 ORG1 ORG14 ORG16 ORG15 CLIQUE OVERLAP Cliques are at least three people, and almost all of them talk to each other. School 1 has 94 cliques from size 3 to 7 members 16 EGO NETWORKS FOR TEAM- RECIPROCATED BY FORMAL GROUP 18 PART III. HISTORY Jacob Moreno developed the sociogram. In the beginning it was a piece of paper, with people as the points, and their connections to each other as lines. (1930’s) Additions to the field were from many fields including anthropology, sociology, matrix algebra New fields developed such as graph theory During the middle 1980’s computer technology and speed prompted interest 2012, now at a heightened level of interest, from data analysis to visualization GRAPH THEORY The mathematical study of the properties of the formal mathematical structures called graphs. Simple Graphs Trees Gear Graphs DIFFERENCES – NET WORK ANALYSIS VS. STUDY Network Analysis Whole structure Structural features, Structural characteristics, Subsets of the whole and/or Individuals within the whole Sociograms Data Analysis Some statistical analysis Network Studies A network measure is used as an attribute Generally only one measures, as an individual “score”, but not connected to the whole 21 ATTRIBUTE DATA Case Gender School age Club1 Teacher Suzy F GW 11 1 Scott Casey M GW 11 1 Scott 1 JT M IP 10 1 Scott 1 Tommy M IP 9 Smith 1 Isabel F IP 10 1 Jones Rose F GW 11 1 Jones Spring F GW 9 Maple Club2 Club3 Grade Bus A 47 C 47 1 B 47 1 A 47 A 47 D 47 A 47 1 1 MATRIX DATA Outdegre Read left to right; 1st Row is the chooser and their choices Simplest Measures 1. Freeman’s Indegree- who choose me is the column total. 2. Freeman’s Outdegree – who I chose is the row total Indegree 23 SEPARATED OUT BY AREA 24 EXAMPLE 2. HOW Defining relationships Based on theory – support, leadership Based on behavior – defined by program or project Combination of both Measuring relationships Align measure to behavior Use multiple measures Look at all levels (as results indicate, or aligned to theory) Network Sub-groups Individual 25 Sample from Internet Component 1: Develop a network of strategically placed and financially viable community health centers in Clark County for the uninsured/ underserved to access their primary health care needs. Descriptions Resources Activities Outputs Outcomes Impact Appropriate physical locations that provide ready access and are free of barriers Determine the best sites through research of target populations and through consensus of CCHAC members Secure funding to equip and staff facilities Maximize funding through an effective RFPbased selection of providers who will run the centers Secure funding to build an appropriate MIS system Maximize funding by building, testing and refining the MIS system among existing centers, and by hiring competent MIS staff. Target population will obtain quality primary health care at the centers. The perception that quality health care for the poor/ uninsured/ underserved is available only in hospital emergency rooms will be dispelled. The health and wellbeing of members of the target population will improve. CCHAC’s understanding of the target population will improve. CCHAC’s ability to effectively meet the primary health care needs of the target population will improve. Will eliminate the use of emergency room visits for primary care services in Southern Nevada. The overall health and well being of Southern Nevadans in general will improve. CCHAC’s mission will evolve to include the entire range of healthcarerelated and social services. CCHAC MIS committee Selected providers will Appropriately equipped and staffed facilities MIS that allows effective oversight, data gathering, tracking, reporting, and appropriate sharing of patient information among clinics Assumptions CCHAC continues Target population will utilize referrals from the centers to medical specialists and other needed health care and to needed social services. Client use of the centers and their satisfaction will be effectively tracked, recorded, reported on, and, as appropriate, shared among centers. 26 Quality primary health Once it ends, EXAMPLE FROM A LOGIC MODEL Target population will utilize referrals from the centers to medical specialists and other needed health care and to needed social services. Use referrals is an anticipated behavior that will occur as a result of this project. This implies a connection, or a relationship Center Medical Specialist Client This connection is the behavior 27 QUESTIONS TO ASK: GO BEYOND THE ASSUMPTIONS OF THE STATED What program components indicate a relationship? ( target population use referrals from health center s, to specialists, other ser vices ?) Can you define the relationship(s )? ( From direct relationship with health center s to new relationships; also the center s to the specialists ) Who are the actors, groups, etc in the relationship(s )? ( target population, health center s, specialists, other.) What behaviors, actions, and activities would you expect in this relationship and by whom? ( Target population initiates new relationships, what is role of center s? How does the relationship(s) contribute to the success of the project or to understanding the project 28 INFORMAL NETWORKS Defined by theory of relationships & questions Mentoring relationship Work with Do research and write with Consider important to career Respect and would ask for work related help from Would like to work on a committee with 29 FREEMAN’S INDEGREE N umber of times a person is chosen by others / n -1 (column total/n-1) 30 DEGREE CENTRALIT Y 31 NODES BY INDEGREE FREEMAN’S BETWEENNESS How much an individual is indirectly linked to others, and to what extent an individual is between two others C b(i) = SSb ijm , across all n 's. b ijm = g ijm/ jm; g ijm is equal to the number of geodesics containing i that are linked to both j and m; jm is equal to the number of geodesics linking j to m. 33 COMPLETE SOCIAL NETWORK Top 3 Choices Old Group and New Group are NOT integrated 34 PC2 - DECLINING Principal 36 PC1 - IMPROVING Principal 37 CLIQUE OVERLAP COMPONENTS WHAT TO MEASURE?