A Critical Assessment of Child Custody Evaluations

Psychological
Science
in the
PUBLIC
INTEREST
A Critical Assessment of Child Custody Evaluations:
Limited Science and a Flawed System
Robert E. Emery, Randy I<. Otto, William
T. 0 7 D o n o h u e
2
Introduction
3
Demographics of Family Structure?Custody Disputes, and Custody Arrallgements
The Deer-Doe Case
T i
----.-....
:..I n
L-
U I l I l l i l I I I C U I ill TlIL3
Child Custody Disputes
Child Custody Decisions
Child Custody Arrangements Following Separation and Divorre
5
Custody Law and Child Custody Evaluations in Practice
The "Best Interests of the Child" Standard
A Psychologqcal Evaluation for the Deer-Doe Family
Practices Reported by Cuatnrly Evaluators
7
The (Limited) Science of Custody Evaluations
Forensic Assessment Instruments: No Scientific Support
Clinical Assessment Tnstruinents: Some Cautions in the Custody Context
Controversial Topics Requiring Further Investigation
The Deer-Doe Case: Dr. Bagan's Custody Report
A Bigger Problem: The Legal and Emotional Context of C~lstoclyDisputes
12
Average Effects anti Variation in the Well-Being of Children Froni Divorced Falllilies
Nonrandom Selection Into Divorce
Different Rislts for Different Outcomes
13
Predictors of Children's Psychological Adjustrnellt to Divorce
P a r m t a l Conflict
Conflict and the Deer-Doe Case
Parent-Child Relationships
A Referral for the Deer-Does
19
Children's Best Interests: A Standard With No Standards
Historical Pel-spec~ive
Custody- Eval~iatioils:A Solution to Judges5 Dilemma?
20
Recommendation 1: Encourage Alternative Dispute Resolution a n d Private Settlement
The E x a ~ n p l eof Divorce Mediation
T h e Urel--Does in Mediation
22
23
Recommendation 2: Adopt a Clear Custody Standard Like the Approximation Rule
Recommendation 3: Limit Expert Testimony and Clarify Standards of Practice
Rules of Evidence
Professiolial Stantlards a n d Guidelines
24
C o n c l ~ x d i nComment:
~
A Question of Values
A JOURNAL O F T H E
About the Authors
Robert E. Emery is Professor of' Psychology and Director of the Center for Children, Families, and the Law at the
and several I~ooksincluding 1Vhr-r-ingr,Di?,orce,
University of Virginia. He is the author of over 100 scientific I)~~l)lications
~1r1(1C'hildren'sAdjustment
(Sage, 2nd ed.. 1999). Rcr~egotiatingFa7~~il~-Relationsh~s:Divorce,ChildC'usto(~~;ar~dMediation (C;uilforc[, 1994), and TheEuth about C,'l~ilclrenandDivorce:
Dealingwith theEmotions SoYour~ndX~zrr-Children Can
Thrice (VikingIPenguin, 2004). His awards inclutle a "Citation Classic" designation from the lnstitu~efor Scientific
Information, an Outstanding Researcll Publication Award from the American Association of Marriage and Family Therapy,
and a Distinguished Research Awartl from the Association olFaniily and Conciliation Courts.
Randy I<. Otto is an Associate Professor in the Department of Mental Health Lair and Policy at the Florida Mental Health
Institute, University of South Florida, and he also senres as adjunct faculty at Stetson Lrniversity College of Law. Dr. Otto
has sewed as President of the American Academy of Forensic Psychology, President of the American Psychology-Law
Society (AP-LS), and President ofthe American Board of Forensic Psychology. He serves on the editorial boards of a i ~ u m h e r
of journals including Law &HumanBehaz'ior, Behaz~iorc~lSciences
& thelato; Assessm'ent,Journnl qf T111-c.at
zl.ssessment,
and Journal of Child Custo(ty. He c.ui1-ently chairs [he AP-LS committee revising the Spec7;altyGzritJeline.r,forForensic
Psychologists.
William T. O'Donohr~e,a licensetl clinical psychologist, is widely recogniztd in the field for his proposetl innovations in
mental health senrice deliwi-y and in treatment design and evaluation, and for his lcnowledge of empirically supported
cognitive behavioral therapies. He is also a nationally recognized expert in he area of sexual disorders and sexual
uictin~ization.O'Donohue is a member of the _4ssociationfor the Advancelllent for Behavior Therapy and has servrd on the
board of directors of this organization. He has an exemplary histor)- of successful grant funding and government contracts:
Since 1996. lie has received over $1,500.000 in federal grant monies from sources including the National Institute of
Mental Health and the National Institute of Justice. In addition, O'Donohue has published his work prolifically: He has
edited over 20 books, written 35 book chapters on various topics, published reviews for 7 books. and l)ublished rriore than
75 articles in scholarly journals. O'Uonohue is a natiollal expert in training clinicians in integrated care and developing
quality-improvement projects in integrated care. In addition, O'Donohue has provided expert testimony in forensics cases
across the nation.
Si:$$i
(3
Sign up to receive B l a c k ~ ~ eSynergy
ll
free e-mail alerls with complete Psychological Science in, the Public I ~ ~ t ~ r p y t
tai~lesof contents and quiclc links to a.i(i, abstracts from the most currenl ibsue. Simply go to 1rwr.blarkwel1-s>-orloi.
com, select the journal from the list of journals, and cliclc on "Sign-up" for FREE e-mail table of contents alert-.
PSYCHOLOGICdI, SCIENCE I h T H E PUBLIC I N T E R E S T
A Critical Assessment of Child
Custody Evaluations
Limited Science and a Flawed System
Robert E. ~ r n e r ~ Randy
,'
I<. ~ t t o , 'William T. 0 7 ~ o n o h u e 3
Unirel sity of Virginia, 2~nrz.pr-sity
of South Flol-ida. nnd ' U n i v e l s i ~of~Nevada-Reno
SUMMARY-Most parents who live ripart negotiate rltstody
arrangements on their own or w i t h the help of lazoyers,
mediators, or other professionals. However, psychologists
and other nzental health professionals increasingly have
berorne involved i n evaluating children and families i n
custody disputes, because of the large number of separated. divorced, and never-married parents and the substar~tir~l
conflict that often accompanies the breakup of a
family. Theoretically, the larc guides and controls child
custody evalziatior~s,but the prevailing custody standcird
(the "best interests of the child" test) is a vague rule that
directs judges to m a k e decisions unique to individual cases
according to w h c ~ twill be i n children's future (and unde,fined) best interests. Furthermore, state statutes typically
ofler only vague guidelines as to how judges (and evaluators) are to assess parents and the merits of their cases,
u,rzd how they should ~tltimatelydecide what custody arrangements u)i.ll be i n a child's best interests. In this vacu u m , custody evaluators typically administer to parents
and children a n array of tests and assess them through less
formal rnean,.sincluding interviews and observation. Sadly,
Ire jind that ( a ) tests specijiccilly developed to assess
questions relevant to custody are completely inadequate on
scientijic grounds; (b) the claims of some anointed experts
c~bouttheir favorite constructs (e.g., '>parent rilienation
;~.ndl-omo")
are eqltally hollow w h e n subjected to scientific
>i-rzltir~?.:
(c) evaluators should question the use even of
1crl1-estc~bli.shedpsychological
measures (e.g., rneasures of
intellig-erlce, personality, psychopathology, and academic
c~chie~.ernent)
because of their often limited relevance to
the cjz~estionsbefore the court: and ( d ) little empirical data
exist regurding other important and controversial issltes
(e.p. whether evaluators should solicit children's wishes
about custodj-; 7clhether infants and totldlers are harmed
Addrrss correspondence to Robert E. Emery. Department of Psychology, Giliner Hall, Box 4,00400. ITniversity of Virginia, Charlottesville, VA 22904-44,00; e-mail: [email protected].
Volume 6-Nurnhel-
1
or helped by overnight visits), suggesting a need~for
further
scientific investlgat1on.
We see the system for resolving custody disputes as
deePly,pawed, for reasons that go beyond the problem of
limited science. T h e cozzpling of the vague "best interests
of the child" test with the American adversary system of
justice pztts judges in the position of trying to perform a n
impossible t a s k , and it exacerbates parental conflict and
problems in parenting and coparenting, which psychological science clearly shows to be key r far tors predicting
children's psycl~ologicul dfjiculties i n response to their
parents'separation cind divorce.
Our analysis of t h e f i w e d system, together with our desire to sharply limit custody disputes and custody evaluations, leads us to propose three reforms. First, 1ue urge
continued eflorts to encourage parents to reach ctistody
agreements on their own-in divorce mediation, through
collaborative law, in good-faith attorney negotiations, in
therapy, and i n other forums. Some such eflorts have been
demonstrated to improve parent-parent and parent-child
relationships long after divorce, and they embrace the
philosophical position that, in the absence of abuse or neglect, parents themselves should determine their children 's
best interests after separation, just as they do i n nlarriage.
Second, w e urge state legislatures to move toward adopting
m o m r k a r and determinative custody rules. u step that
would greatly clargy the terms of the marriage contract,
limit the need for custody evaluations, and sharply narrow
the scope o f the evaluation process. We,findparticularmerit
in the proposed "approximation rule" (recently embraced
by the American L a ~ cInstitute), rn whirh postdivorce
parenting arrangements would approximate parenting involvement in marriage. Third andjinally. 1ue recommend
that custody etlaluators follow the law and only offer
opinions for which there is a n adequcrte scientz3c basis.
Related to this, w e urge professional bodies to enart more
specific standards of practice o n this and related issues.
Copyriph~$ 2005 American Psychological Society
1
Critical ilsscssn~nntof Chilcl Cnstc~tlpEv;~lnati<~ns
-
INTRODUCTION
reduce tlie number of custodj- disputes. Third, in disputes that
remain contested. we would limit mental health expert testimony
onlj- to opinic~nsclearly supported 1)y psychological science, a
circumstance that unfortunately does not characterize soille of
today's practice. This final point is not so much a call lor a reforrn
as a recommendation that expert witnesses in custody evaluations conform to existing standards for expert testirnonj-.
Child custody disputes (:an entail any number of emotionally
~vrenchingcircumstances. The prototypical case involves nlarried parents w~hoseparate and, in the heat of' divorce, cannot
reach an agreeinerit about where and how their children should
live. I11 other cases, u marital or cohabiting relatioilship clissolves before a child is born and parents must negotiate custody
~vithoutthe benefit of a shared history ot' parenting. Custody
dispu[es also can surface years after a brealt-up, for example T h e Deer-Doe Case
when a parent relocates, a n adolescent lraiits ~ I change
I
living We invite the reader to ljegin to consider the many enrotional,
arrangements, or parentsliave probleliis with a difficult child.
legal, empirical, and value conflicts involved in child custody
Child custody disputes also are not limited to conflicts be- dispute" with a hypothetical case. We revisit this case at
the monograpli to illustrate and anchor our
tween biological parents. Graridparen~smay dispute custotly of points th~.oughoi~t
their grandchildren with their own children, l ~ i r t hparents may discussion.
contest cllstody in the context of adoption. or sanie-sex couples
rnay dispute custody with each other or a I ~ i o l o ~ i c aparent.
l
Jane and John Deer-Doe, both 39 years old. have two chiltlren:
Filially. illfidelity and genetic testing, as well as technological
Isabella. a 10-year-old girl in the four~hgrade, and Carlos, a 3and social innovations in conception and childbearing, can
year-old boy ~vhoatlends preschool but spends most of the day al
create nightn~arish scenarios in which biological and social
home with his motlier. Jane continued to ~ \ ~ o f~111
r k time as a cerparents can end up disp~itingcustody (Sch~vartz,2003). Our
tilietl put~licac:countant after Isahella \\-asliorn, ~ U wit11
L , John's
focus here is on child custody disputes bet~veenparting ~ ~ a r e n l s , reluctant agreement, she quit work alier Carlos's birth. John, a
moderately succeasfnl computer e~igiiieer antl self-descrihed
whether married or not, hut many of the same issues and conhighly iri\~olvedfather. says that he had expected Jane to ~.eturrlLO
cerns appl) across these different circun~stances.
work
after n year or t~j-c~
at home with the children.
Our initial mission for this nlonograph was simply to critique
hat tlie!r had lo~igstandingconflirts ahout
Jane
and
John
agreed
the psychological science underpinning child custody evaluaparentilig, finances, and sexu;ility. John [lied repeatedly to get Jan?
tions. W;e malie such a critique in the section titlecl "The
to address their unhappirress by seeing a marriage therapist.
(Limited) Science of Custody Evaluations." H o ~ v e ~ ethe
r . subJane was open to therapy hut also accepting of an impel-fecl
ject olchild custody tlisputes is complicated I)y many e~notional,
marriage. Jane's acceptance ended. I~owe~er,
when she learned ol
practical, and legal issues that are of interest aiid relevance to
John's
psychologists. We therefore have broadened the scope of the
2-year-long aflair with a coworker. She innnediately contac~edan
report to consider these more general issues, particularly deattomex arid shortly tliel-eaI'trc John left tlie hause at lier request.
In their sul~sequentnegotia~ions,Jane inclicated her desire for a
velopments in child custody law. alternalive dispute resolution,
rlivorce, and Jolln agced. I-Ie hoped to remarry soon and wanted the
ethics, and societal values a b o u ~faniily life. Of course, psychildren with him half'of the time. Jane c,ountered that John should
chological science is our primary focus, and one of the strongest
have the children no more than eveiy other \veekeatI, consistent
findings of basic research in this area is that children fare hetter
with his "minimal" involvement during their marriage, and she
in separatiori a i d divorce if parental conflict is minimal or at
further insisted that their children have no vontact ~vithhis "friend."
least contained antl if children maintain a good relationship with
In the 3 months after her parents'srpai-atetl, Isahella refuseti to
at least one, and preferably both, of their parents (Emery, 1982,
see her father exrept on a couple ufoccasions. She colitinuetl to do
1999b, 2004). In other words, the process of'lamily dissolution
well in sc:l~oolhul was extremely 31ig1-y with her father for
and the nature of continuing farnily relationships are more im"cheating on my mother." Carlos aslted for his father repea~edlyin
portant to children's mental health than is the structure of any
thr (lays and weeks after he separation but did so less afier seeing
particular custody arrangealelit.
his father only sporadically during this time. His preschool
teachers co~nplainrtlthat Carlos had hecome very aggressive in
This finding, together with our analysis of the context of
sc:hool and had llegun to we1 and soil hiillself again.
custody disputes, leads us to call for three sets of reforiiis. First,
we encourage continued efforts to promote the private seitlement of child custody disputes through education, gootl-faith
negotiation, and alternative dispute resolution. Private settlement or custody disputes can reduce conflicl: it can encourage
rriore cooperative, ongoir~grelationships hetween cc~parents;and
it can facilitate positil-e relationships between cllildren and 110th
of their parents. Second, we support efforts to make chi111custody law more clear and tieternlinative, in order to substantially
How can psychological scientists help families like the DeerDoes? As we will review in this monograph, there is good research to help us hetter u~iderstancl children, divorce, and
custody confiicts, and there is sorile reasonablu strong evidence
l
Unfortunately, very little
on solrie s u c c e s s f ~ ~interventions.
research has been conclucted directly on legal issues in the
custody contest, including child custody evaluations.
Robcrt E J-ulrry, Randy 1C. Otto, Wilham T. O'Donolnl~
DEMOGRAPHICS OF FAMILY STRIJCTURE, CtISTODY
DISPUTES. AND CUSTODY ARRANGEMENTS
The structure oSL1nlericanfanlilies changetl draniatically in the
latter pari of the 20th century. As indicated in Figure 1, divorce
rates trended upward in the United States througllout the 1900s
and, follo~1-inga rapid rise in the late 1960s, pealied in 1981
before turning do~vnward(Eramlett & Rlosher, 2002).
Other key elements of the demographic story include an average risk of divorce ol somewhat less than SOVb,higher divorce
rates for African L1meric:ans. lower rates for Asian Americans,
and the declining risk for divorce as a fi~nctionof years in
niarriage (Eramlett 8; Masher, 2001; see Fig. 2). Ahorrt 6 0 % col
tlivorrrs involve children (Clarlte, 1995), and allout half take
place in the first 7 years of marriage (see Fig. 2), so that children
are liltely to be young when marriages end and custody is disputed (Furstenberg, Peterson, Nordi 8; Zill, 1983). As we discuss later, special concerns arise ahout custody for infants,
toddlers, and. to a lesser extent, preschoolers.
1
1 I-
Black
Asian
L 7"
One
~
Five
*..__.c
.',.
*..r..
.*.'
Ten
Years married
Fig. 2 . Risk of rlivul.cc over tlic first
(based on Bramlett & hloshcr, 2001).
20 years
Fifteen
of
Twenty
marriage, by rthnicity
iting u~lionsare more liltely to dissolvc than legal marriages are.
Forty-nine percent of cohabiting relationships end within 5
>-ears. whereas 20% of first marriages dissolve within 5 years
(Bramlett & Mosher, 20021. Ailthoughwe Itno\\. oS no data regarding ho~vmany d i s r u p ~ r dcohahitations involve parents and
their biological children; the disruption oSrelationships l-xtween
unmarried parents clearly is an important and growing area for
researc:ll on child c~rstodl-disputes.
Uninarried P a r e n t s
Over 40% of chiltlren born to married parents are expected to
rxperiencc the tlivorce of their parents (Eumpass, 1984; U.S.
Burcau of the Census, 1992), and the clualification "l~ornto
married parents" is an important one. In 2002. 34% of all
Child Custotly Disputes
children in the li'nited Stales were born outside of marriage
There is no good national data on how many custody disputes
(Martin et al., 2003). In lact. the apparent decline in divorce
arise ~ r h e ndivorcing, cohabiting, or unmarried parents part or on
sinre 1.981 may he attributable to at-risk individuals and C ~ L I how many s l ~ c hdisputes erupt years after the break-up (which
plrs self-selecting out of legal marriage ant1 childbearing. Rapid
may be a more cornuion circumstalice). What is clear is that courts
inctreases in nonmarital chil(1hirth did not stabilize until about
are ovei-whelmed hl- the huge number of fanlilies separating,
1990, and cohahitation (,which is more difficult to track) apclivorcing, and disputing custody. In 1995, domestic-relations
parently is continuing to increase in frequency.
disputes. which include but are not limited to child custody
The best estiiiiates suggest that ahout half of children Ijorn
litigation, accounted for one quarter of all legal filings, malting
ou\side of marriage actually are horn to u n m a ~ ~ i ebut
d cohabthis the largest category of court action (Ostrom & ICautler;
iting parents (Sigle-Rushton & McLanahan, 2002), and cohab1996). Other evidence intlicates that custody disputes forin
the largest percentage of tlomestic-relatiolis cases (&hepard,
2004).
1860
1880
1900
1920
1940
Year
1960
1980
2000
Fig. 1. -1nnual U.S. r l ~or
r cr rateifrom 186'7 to 2000 (l,,~srdonBralnlrttS
Moshcr, 2002, and Einer y, 1999b).
Child Custody Decisions
The Ixst evidence on how child custody is decided in the context
of divorce collies from Maccoby and Mnookin's (1992) study of
children in which the parc:nts filed for divorce in
1,124 families ~rit1-1
tw70 California counties in the middle 1980s. As illustrated in
Figure 3. mast of these cases were settled outside oS couit, as over
three quarters of custody mangements were negotiated either
by the parents tliemselves or through their lawyers. Since 1981.
Califo~llialaw has inalldated that mediation he attempted before
a custody hearing can be held hefore a judge: an additional 11% of
the cases were settled in mediation, while 5% of the cases
went the next step up in the hierarchy of legal conflict-a custody
evaluation-before reaching a settlement. Only 4% of cases went to
Critical Assessnlrn~of Child Custody Evaluations
Settled during trial
Settled after evaluation 1 5 . 2 ;
Settled but not uncontested
i
m
i
Settled with mediation 1
1
.
1
129.3
j
Uncontested divorce
50.4
0
10
30
40
Percent of families
20
50
Fig. 3. Percent of 1,124 fam~iliesin two California counties icttliilg divorce custody using various methods; during th,, rnirl-1080s (1,ased {rn
Maccoby & Bilnookin, 1992, p. 1.37).
trial, and most of these were settled during the trial process. A jutlgc
decided less than 2%, ofthr cases (Maccohy & Rlnooliin, 1992).
The generality of these findings is limited by the two-county
sampling. as well as by rapidly changing laws and societal expectations. Still, the data highlight several patterns observed
across the United States and much of the industrialized world
(Emery, 1999b; Pryor & Rodgem, 2001). First, inany parrr~ta
experience at least a inild degree of conflict surrour~rlingchild
custody, and conflict is substantial in a sig~~ificant
subset of cases.
Combining legal indicators arid self-repoited conflict measures,
Maccoby and Mnooliir~(199'2) estimated that 51% of divorces
involved negligible conflict over issues related to custody, rvhile
24% llacl mild conflict. 10%substantial conflict, and 15% ~ntense
'
conflict. Sccond, as is the case with other litigation, most custody~
disputes are decided outside of the courtroom. Third, alternative
dispute resolution metliods such as rriediation increasi~~gly
are
used, often successfully, in an attempt to settle disputed cases.
Fourth, mental health professionals often are involved in child
custody conflicts as mediators, custody evaluators, or therapists
(although the last role is not reflected in these data).
The in1portanr:e of each of these patterns is multiplied by high
rates oi separation and divorce, cuslody disputes between cohabiting and never-married parents, antl the poterilial for conflict throughout the duration of the cl~ilclren's childhood. This
rneans that (a) even if they represent a minority of cases, large
numbers of children al-e exposcd to sul~staiitialor intense parental arid legal conflict in the midst of their p;irmts' separation;
(b) judges facc the prospect of spending a great deal of theirtime
lieariilg custody cases; ( c ) alternative dispute resolution and
custody evaluations have become important parts of the process;
and (d) tnental health lxofessionala are becoming increasingly
involved in the child custody arena in a variety of ways.
Child. Custody _4rrangements Following Separatioti and
Divorce
Although laws. definitions, and terms vary from state to slate,
most of the key aspects of child custodj- ar~,ar~gcnients
arc
captured by the following concepts:
Legal custody refers to parental aulhority or tiecision malting.
In cases of sole legal cus~udy,one parent has the right to n l a l i ~
major decisior~s about thc children's lives, especially
schoolirlg, elective mrdical care, and religious training.
Whenjoirzt legal cwtody is in effect; 110th parents share these
major decisioiis, while each parent makes day-to-day decisions autonomously when the children are with her or him. In
some cases, the court rvill assign more sperific decision
making over dav-to-day matters to one or both parents.
Phpical czlstody refrrs to the time children actuall!~spend
with then- i~arents.In cases of 11rimnry
, p h ..~ s i c n lcustod?.. tlie
children spend the niajority of their time with one parent antl
generally "visit" (a term many find pejorative) with the
"nonresidential parent" on some agreed-to schedule (e.g..
one evening during the week and every olher weekend). In
cases ofjoint physical cusborl~clliltlren spend close to equal
amounts of time with hot11 parents. Although there is no
uniform dcfinitiol~of joint physicnl custody, Inany consider
it to be a ll~iilimurnof an average of two overnights per week
(Maccoby & B'Inookin, 1992). This definition is consistent
with child-support laws in 28 states that lower support ohligations for joint p1lysical cus~odparrangements and often
define joint physical custody at about 100 overnights peryear
(Elrod (% Spector, 2004).
Split cwtody refers to circumstances in which each parent
has sole physical custody of at least one child-that is, when
siblings are split up between their parents.
Single Mothers nn'rl Siizgle Fat11,ers
The Ur~itedStates Census Bureau generall!. does not track joint
physical custody, but instead lists children as li\.ir~gwith two
married parents, a single mother, a single lalher, or in solnc other
arrangenlent. In 2002, of all c:l~iltlrenliving with a single parent.
just oler 82% lived wilh a single inother while approximately
18%lived w i ~ ha single fathcr (Fields, 2003). This percentage of
c.ll;ld~enliving with a single father represents an inr:rease over
he historical level of about 10% (Meyer & Carasky, 1993).
Intc3rpretation of these c:msus data, however, is clouded by
several factors includiiig (a) remarriage, as children who live
with remarried parents are counted as living in a two-parent
household; (h) cohabitation, as many "singlex parents live with a
partner, including 11% of single mothers and 33% of single
fathers in 200:2 (Fields, 2003); and (c) reason for single-parent
status, as the category includes separated, divorced, nevermarried, and widowed parents. Another limitation is that joint
physic-a1 (or legal) custody is not routinely docun~pnted.
joint Custody
Joint custody (a term that confounds legal and physical custody)
has heel1 a n~ucll-discussedand much-debated coparenting arrangemcnt since the 1980s (Folberg, 1991). Later, we discuss
cvidence about the well-being of children living in joint custody.
Our present task is to estimate its prevale~lce.
Robert E . Emery, Randy I<. O L L William
~.
We lcnow of three natiotial estimates of the lreyuency of joint
custody (Child Trends, 2002; Clarlie, 1995; Doniielly & Finkelhor, 1993), the best coming from special supplemental 1998
United States Census data (and also 1994 and 1996 data that
provide essentially the same results). In this ntialysis. 65% of
mothers had sole physical and legal custody. 10% had sole
physical and joint legal custody, 1 1 % of fathers hat1 sole phys~
either joint or sole legal cus~otly),9 % of
ical c u s ~ o d j(with
parents had joint physical and legal custody. and 5% had split
custody or some other arrangement (Child Trends, 2002). Thus,
about 75% of children not living with both parents lived primarily with their mothers, approxitilately 10% lived primarily
wit11 their fathers: about 10% lived in joint physical custody, and
another 5% lived either in split custody or i11 some other arrangement. Although some people argue that join1 physical
custody is becoming far more common, no trends for increased
prevalence hetween 1994 and 1998 were found in the census
data (Child Trends, 2002).
Historical Eerzd Euidence and Joint Custody
Historical data from Wisconsin demonstrate the importance of
distinguishing legal custody and physical custody, and also
malce us suspect that joint lepnl custody is becoming considerably more common than suggested 11y he census es~iinates.A
review of 9,500 Wisconsin divorce se~tleliientshet~veen1980
and 1992 revealed that sole physic.al custody to fathers remained stable during these years 15-hile sole physical custody to
mothers declined (see Fig. 4). Joint physical c:ustody rose from
2 % to 14% 01the Wisconsin cases. while joint legal custody
increased from 18%' to 81% (Melli. Bro~vn,& Caiician. 1997).
Our experience leads us to believe that this dramatic increase in
joint lrgal custody and more modest inc.rerrse in joint physical
custody have also occurred in many other stales. Estimates from
1990 data gathered by the National Center for Health Statistics
(Clarice, 1995) also support this suggestion, as different states
rrportecl widely varying rates ofjoiiit custody ( l ~ ~and
a l physical
cus~orlywere not distinguished)-for exan~ple,4% percelit in
Nehraslca compared with 44% in geographic:ally a r ~ dpoliticall!.
similar ICansas.
C
n
'E
80
40 -
g
n
"-,
/-
- Jo~ntlegal
60 - - Jo~ntphyslcal
--
20
0
1980
C - _ _
Mother pnmary
I'
Fatherprlmary - - - - - - - - - - - ' '
-----'
...-- -... ----......
.......... ........-....-...I
I
I
1982
1984
1986
Year
I
I
1988
1990
1992
Fig. 4. Prrcrntapr t11' i.ustod,- arrangements in Wisconsin d i ~ o r c e sfrom
1980 to 1'19%(data
collcctcd ar,.c,ds calendar
thus ,962
to 1961-62, etc.; based on hfelli. Crown, & Canelan, 1997).
T O'l)o~~ol~ae
Ch,unges in CztstorLj-Arrungrment.c
Custocly arrangetiients change over time, and legal agreements
often do not correspotid to tle facto residence. The best evidence
on these points also cotties from Maccoby and Mnooliiii's (1992)
longitudinal study. For 78.3 cases where complete data were
obtained during the 3-year study, initial legal agreements designated the following custody arrangements in the two California
counties: 66% sole mother c-uslody, 9 % sole father custody, 21%
joint physical custocly, and 4% split custody. Shortly alter the
divorce decree was filed, ho~revrr,only 52% of the cases with
designated joint physical custody actually had a de facto joint
physical custody. Among the 48% of the joint physical cases in
which the living situation was not consistent with the legal
agi-rement, mosL involved sole mother physical custody. Of cases
with designated mother custody. 8 7 % follo~vedthat arrangement
in practice, as did 82% of father custody agreettlellts, hut only
3 5 % of split-custody agreements actually cotifortried to that
arrangenieli t.
Three years later. only 45% of legally designated joint physical custody cases actually conformed to that arrangement,
cotriparerl LO 85% of cases with designaled mother custody, 71%
of cases of father custody, and 34% of split custody axvards
(Rlaccoby c ' Mnookin, 1992). The absolute percentages of the
four t y e s of custody arrangements 3 years after the divorce
decree were siulilar to tlie initial arrangements, but the longitudinal analysis demolistrated that many families shifted out ol
their original custody arrangements and into new ones.
CUSTODY J,Aw AND CHILD CUSTODY EVALUATIONS
IN PRACTICE
Later, me consider hroad conceptual issues related to child
custody law and custody evrrluations. V\re begin, however, with
a brief overvie~vof the currerit legal landscape and a minimal
critique.
T h e "Ekst Interests of t h e Child" Stanclard
Each state legislature in the United States controls its own child
custody law, and laws can vary considerably fiom state to state.
Still, every state law indicates that custody decisions are to be
made according to "best interests of the child" standard, the
principle that judicial determinations should he based on each
child's unique future best interests (Elrod & Spector. 2004).
Many mental health professionals applaud this "hest interests of
the child" standard as being responsive to individual children
and families. We differ. lndividualizerl decision making is a[)pealine on the surface, b u ~we are deeply concerned that a
standartl ~ ~ n g uenough
e
to be interpreted diffrrrtitly for each
family that comes beforr the court (a) encourages paretits to
enter into custody disputes liere re by increasing parental conflict), l~ecausethe outcome of a court hearing is difficult to
Critiral Ilssrssment oE Child Cnstodv
predict: and (1)) allows for bias to intrude in the exercise of judicial disc.retion.
For reasons we do not fully understand. the law apparently has
interpreted chilclrerl's Ijest interests to bc primarily their best
psychological i n t ~ r e s t s(as opposed to other possibilities such as
their economic; educational, or tnedical interests). This is evident in the various factors deemed relevant to children's llest
interests listed in rriost state laws, which typically are rooted in
the Uniform Marriage and Divorce Act (1979): which lists the
folloruing:
The wishes of the children's parent or parents as to their
custody
The wishes of the children regarding their custodian
The interaction and interrelationship of the children and
their parrnt or parents, their siblings, and any other person
who may significailtly affect the children's hest interests
The children's adjustment to their home, school, and community
The mental and physical health of all individuals involved
Because child custody laws differ from state to state. some
factors designed to be considered by judges are idiosyncratic to
one or only a handful of states. South Carolina, for example,
takes into account the religious beliefs and corrirrlitn~entof the
parents, whilc Alabama, Florida, Michigan, North Dakota, and
Utah consider parents' "moral character" to be relevant to
children's best interests. One of the goals of a child custody
evaluation-the overriding goal, accorcliilg to sotlle-is tu assess the child and parents relative to these state-specified hestinterest factors.
A Psychological Evaluation f o r the Deer-Doe Family
After several rlronths of separation and still no custotlvagreement,
Jane's attorney suggrs~eda child custody evaluation as a next step
in their negotiations, and, eager for some outside help, John
agreed. Several weeks later, a psychologist, Dr. Davicl Hagan, who
was rnutually agreed upon by bud^ parties, was appointed by the
court to assess Jane, John, his girlfriend, and their children.
O+er the course of 6 weeks, Dr. Hagan conducted a cornprehensive evalu~tioiiconsisting of interviews and psyc.hologiral
testing ~vithboth parents; tests inrluded the Minnesota Multiphasic Personality Invrntory-2 (RIMPI-2), the Rorschach Inltblot
Technique, and the Wechslrr Abbreviated Scale of Intelligence.
Both children were interviewed, ohsrrved interacting with each
other, and obsrnred interacting with each parent at Dr. Hagan's
office and at the respective parental homes. Dr. Hagan also adn~ir~is~rl-ed
a numbcr of psy~ho'lo~ical
tests to the children including thr Achenbach Child Behavior Checlilist (with parents
and teachers as informants), the Kober-ts Apperception 'Lest, the
Bricltlin Percep~ualScales, and thc Wcchsler Intelligence Scale
for Children-IV. In addition, Dr. Hagan obtained collateral information by interviewing the children's teachers and grandparents,
reviewing school and rnedical recurds, and reading all litigation-
Evaluations
d a t e d documents. Finally, Dr. E-Iaganevaluated John's girlfrienti
by way of extensive i~lterviewingand adrninistraliul~of the psychological tests mentioned earlier.
Dr. Hagan'h bill for $7,400 reflected that he speiit 37 hours
conducting he evaluation, reviewing records, and writing a 35page report sumrriarizing his observations, findings, ant1 opinions.
(We discuss the report later.)
Practices Reported by Custody Evaluators
Given their frequency, high cost, and social and personal importance, we niigllt expect to find a large body of research on
custody evaluationi arltl their scientific underpinnings. HOWever, only a few studies of custody evaluations have been cornpleted. One thing these studies show is that. in real life, many
evaluators use the instrurrle~itsemployed by our fictional Dr.
Hagaii. Another thing research shows is that most of these
measures are deeply flawed when used in the custody context.
With the exception of one study (Bow & Quinnell, 2002j all
research examining child cuslody evaluation practices has been
bascd on he self-report of examiners. Although these data
provide some helpful information, we must keep in mind that
professionals' reports of their behavior may not accurately de, Long. 2003).
pict their actual practices (Greenberg, O t ~ oSr
Keilin and Bloom (1986) describer! the practices reported by
82 custody evaluators !78% psychologists) who responded to an
anonymous survey Respondents devoted an average of 19 hours
each
to each evaluation and almost always repoi-tell ir~~er~riewing
parent and the children. Most used psychological tests with
adults (76%) and children (74%); most observed parent-chiltl
illteractio~ls(69%);half said they obsemed interactions between
the two parents; and ahout one third reported visiling the children's homes or schools. Approximately one half interviewed
third parties (e.g., friends and relatives) in an attempt to gain a
better understanding of the children and their parents.
No one particular psychological test was used by a majurity of
the respo~ldentswhen assessing children. Iiltelligence tests
were used most frequently, with almost half of the evaluators
using them in the majority of their cases. The next most frequently used instruments with children were the Therualic
-4pperception Test or thc Children's Apperception Test (39%),
followed by miscellaneous projective drawings, the Rorschach
Inlchlot Technique, and the Bender-Gestalt Visual Motor Test. In
assessments of parents. the MMPI was the mosl comnlonly used
assessmerlt ~ e c l ~ n i (70%).
q ~ ~ e followed by the Rorschach Inkhlot
Teclinirlue (42%). and the Thematic Apperception Test (38%).
ICeilin and Bloom (1986) also asked the evaluators to rank
order 2 1 different factors with respect to their importance when
considerir~gcustody. In descending order of significance, the ten
most important were (1)the stated preferences of a 15-year-old
(or older) child, (2) parental attempts at alienation (i.e., attempting to turn a child against the other parent), (3) the nature
and quality of the child's emotional relationship with each
parent, (4) the emotional or psychological stability of each
Robert E . Emrrv. K a n d v I<. Otto. V i l l i a m T. O'llonohur
parent, (5) each parent's parenting skills, (6) each parent's central focus in child cus~odycases. Still, evaluators who assess
openness towards the child's contact with the other parent, (7) such l'actors are following explicit legal guidelines. More diffithe parents' preseparation caretaking and parenting roles, (8) cult to explain and more problematic? however, are other aspects
the parents' expressed anger and bitterness regarding the di- of evaluation practices including the widespread use of wellvorce. (9) the parents' sexual orienlation, and (10) the stated established measures with no clear relevance to tlie custody
preferences of a 5-year-old chiltl.
context (e.g., nleasures of intelligence), attempts to measure
Ten years later, Aclterman and Aclterman (1997) surveyed constructs created to apply to child custody decision making
800 doctoral-level psychologists who conducted child custody (e.g., "parent alienation syndrome"), efforts to identify "parent
evaluations and obtained usable responses from 201 (25%). of choice" (e.g., the Rricklin Perceptual Scales), and the use of
Respondents spent 2 1 hours per evaluation-similar
Lo the n~easuresthat a significant number of psychologists view with
earlier survey-but these respondents reported devoting more sltepticlsrn (e.g., the Rorschach Inkblot Technique).
time to reviewing collateral materials and report writing. IntelWe are dubious about many child custody evaluation pracligence tests and projective measures continued to be the in- tices, because of' tlie absence of solid psychological science
struments most frequently etnployed with children, and the and of clear criteria to he predicted by psyrhological science.
MMPIIMMPI-2 remained the most frequently used assessment We also hold two much nrore fundamental questions about
ins~rumentfor parents, followed by the Rorschach Tnkblot child custody evaluations: Why has society and the law placed
such importance on a prediction about psychological factors
Technique.
Many custody evaluators also reported using assessment in- in determining custody? And if the goal is to minimize chilstruments with children that were developed specifically for use dren's psychological risk, might there be better roles for psyas practitioners and as scientists-in
in custody contexts (Ackerman gL Aclterman. 1997). Over one chologists to play-both
third used the Biicltlin Perceptual Scales (Bricklin, 1990a) while custody disputes? For now, however, we focus on the lack of
1.6% used tile Perception of Relationships Test (Bricklin, 1989). scientific evidence to support many of the instruments and
Fewer respondeilts (11%) used the A c k e r m a t - h o e n d o Scales practices of mental health professionals wlio serve as custody
for Parent Evaluation of Custody (Ackennan & Schoendorf, evaluators.
1992). the one custody-assessment measure designed for entire
Heilbrun, Rogers, and Otto (2002) described a three-category
families and adults. Fewer than 10% used other custody- typology of assessment techniques used in forensic contexts,
assessment measures, specifically, the Parent Awareness of Skills including custody evaluations. Clinical assessmen,t in,strurnen.ts
Survey (Bricklin, 1990b) and the Custody Quotient (Gordon & are those developed to assess psychological constructs, typically
Peek, 1989). Other investigators (e.g., Bow & Quinnell, 2001; for intervention purposes (e.g., measures of intelligence, psyGourley Sr Stolberg, 2000) have reported findings regarding test chopathology, academic achievement). Forensicctlly relevant
usage l ~ ycustody evaluators similar to those detailed by ICeilin instrl~merltsassess constructs that are psychological 111 nature
and Hloom (1986) and Ackerman and Aclterman.
but may be of particular lelevancr in forensic contexts (e.g.,
Like ICeilin and Bloom (1986) hefore them, Ackerman and nleasures of response style, risk for criminal offending). Finally,
Ackerman (1997) also asked custody evaluators to rate the im- ji~rensicassessnzen,t instrumen.t.s are specifically designed to asportance of various factors to issues of child custody. According sess psycho-legal constructs. Here we review evidence in regard
to the custody evaluators, the ten most important, in descending to the third and first categolies of dssessment techniques. We do
order of significance, were (1)the substailce abuse status of each not consider forensically relevant instruments because none
parent, (2) the parents' parenting skills, (3) parental attempts at have been used widely by custody evaluators. although that may
alienation, (4) the nature and quality of the child's emotional change (Posthuma, 2003). We also raise concerns about "parent
relationship with each parent, (5) the emotional or psychological alienation syndrome" and other constructs that have been crestability of each parent, (6) each parent's openness toward the ated for, and asserted to have scientific standing in. the context
child's contact with the other parent. (7) the parents' history of of custody evaluations.
compliance with the court during the separation, (8) the parents'
preseparation caretaking and parenting roles, (9) the stated Forensic Assessment Instruments: No Scientific S u p p o r t
preferences of a lS-yea~.-oldor older child, and (10) the parents' In the past 1.5 years, psychologists have developed a number of
expressed anger and bitterness regarding the divorce.
forensic assessment instruments purporting to assess children's
best interests in custody disputes (see Grisso, 2003). Our hottorn-line
evaluation of these measures is a harsh one: These
THE (LIMITED) SCIENCE OF CUSTODY EVALUATIONS
nleasures assess ill-defined constructs, and they do so poorly,
State statutes regarding children's best interests help us un- leaving no scientific justification for their use in child custody
derstand at least some 01 the practices of custody evaluators. We evaluations.
could (and later do) question, f o ~example, whether (01 \\lien)
The most widely used forensic assessment instrument ( A d a parent's mental health or the wishes of a child should be a erman & Ackerman, 1997) is the Biicl<lin Perceptual Scales
- ----
Critical s s ~ s s o ~ cof
n tChild Custody
(BPS), described as a p~ojectivenieasure of parents' competence, suppot-tireness, follo~v-upconsistenc*y, and possessioli of
adrnira1)le traits (Bricklin, 1990a). Using a stylus and rating
card, children rate each parent on 32 difi'erent activities considered to be relevant to these four capacities. The parent rvho
receives the greater number of positive ratings is identitied as
the "Parent of Choice." Rricl<liri asserts that the ~lonverbalnature of the task (using a styllis rather tliari a verbal response)
allows for the assesfment of the child's "unconscious prefkrences." which are less liltely to be subject to distortiori due to
social desirability or p r e n t a l persuasion. Hon ever, the RP5 has
heen criticized on numerous grountls: There is no support for
claims tlrat it assesses children\ unconsrio~lspreferences or
that responses are not subject to external influence; the de~:eloper perniits varia~ionfrom standard lest administi-ation: the
measure samples a relati~relynarrow r a ~ ~ gofe pare~lting(lomains; the developer has 11ot prnvided basic norms anti psyc:hon~etric: properties of [lit: nleasurc; and data regarding
co~~current
and predictive validily are either absent or unconvincing (Heinze 8: Grisso, 1996; melt or^, 199.5; Mclton, Petrila,
Poythress, Pi Slobogin, 1997: Otto & Edens, 2003; Otto, Edens,
B Barcus, 2000; Shaffer, 1992).
Another measure used fairly ii.equentl!r is the Perceptio~~
of
Relationships Test (PORT; Briclclin. 1989),a projective drawing
that is described as measuring the "whole organis111 or gut-level
responses a child has toward a parent [that] are much more reflective of what the child's actual interactions or experiences
with that parent have been" (Bricklin, 1993, p. 1). Seven
tlrawing tasks conipletecl by the child are scored to identify the
"Pri~naryCaretaking Parent." Like the BPS,the PORT has been
widely rriticized. Oliec~ionsinclude the incomplete ant1 confusing m:~nnal,unclear administration and scoring guidelines,
minimal reliwhility data. ~nissingnonns,anti lack of ~laliditydata
(Carlson, 1095; Conger, 1995; Heinze & Grisso, 1996; Melton
ct al., 1997; Otto & Rdel~s,2003; Otto et al., 2000).
Bricltlin (,I 990b) tlesrl.ihes another measure, the Parelit
Awareness Skills Survey (PASS),as a '"clinical tool designed to
illuminate the strengths and wealtriesses in awareness sl<ills a
parent accesses in reaction to typical r:\iilti care situations" (p.
4). The PASS consists of 13 childcare sc.rrlarios selected to
reprewnL caretaking of children of various ages. The parent's
responses al-e followed up aitll questioning by the examiner as
needed, and scorir~gis based on giidelines in the test manual.
' f i e PASS also has Lea11 criticized for basic shortcomings: the
absence of norms, relial~ili~y
and validitv data, arrd clear scoring
guidelines (Otto 8: Edens, 2003; Otto et al., 2000). Of particular
concern is the developer's sugges~iunthat "thc evaluator, by
virtue of' appropriate training in psychology and/or child devrlopment. can apply his or her own star~dardsin assigning the
suggested scores. 'l'he PASS allows for wide la~itudein scoring
since its main purpose is to discover the relative (rather than
absolute) strengths and weal<nesses any individual or cvnlparrd
set of respondents manifest" (Bricklill, 1990b, p. 11).
-
Evaloatitms
The Parent Perception of Child Profile (PPCP; Briclclin &
Elliol~,1991)is dcscribed as a measure of parents' ~lnderstanding
of a cluldk development and needs across eigrht areas: interpel-sonal relaliur~s,daily routine, health history, tle~,elopinentalhiston, scl~oolhis~oly,fears, pcrsonal hygiene, and rommunication
style. Because parerils who nlore a c c ~ r r a t eassess
l~
their chiltl are
assumed to be hetter parent., the PPCP requires the exaniirler to
assess the accuracy of eacli parent's report, using vaguely ciefined
criteria that include the exanlir~er'sand third-pasty informants'
opinions. According to the manlral, clata need not be gathered in
all eight categories, and the examiner can decide rrhich issues are
most critical for a particular cbilti and parenl. The PPCP has been
criticized for its incomplete manual, lac:l< of sco~.irlgdirections.
and a13sence of'reliahilitg and validit>-data (Otto et al., 2000; Otto
& Edens, 2003).
Anothpr instrument used by evaluators with sollie frequency,
the Aclcerman-Schoendur~or Scales for Parent Evaluation of
Custody (.4SPECT), is purportetl to be "a c:linic,al tool desigrietl
to aid mental health professionals in making child custody
recommendations" (Ackerman & Schoendorf, 1992, p. 1).Tlie
ASPECT is not a test. but an assessment approach that aggregates data from the parent (an open-ended "Parenting Questionnaire," tile MMPI-2, the Rorschach Inkblot Technique, and
an intelligence test) and froni the child (the Korscl~achInkblot
Technique, an intelligenre test, an academic achievement test,
allti a p r o j c c t i ~ story).
:~
Measures were selected based on the
developers' r c v i e ~of
\ the literature? ancl test scores are used to
calculale a "Parental Custody Index" (PC[) for each parent. The
PC1 is co~lsidered to indicate parenting elfectiveness, and
jutlgmrnts aboul the par:nts are based on their relative PC1
values. Tiith rare exc:rptions (e.g.. Brodzinsl<y, 1993), reviews of
the ASPECT have been uriiforrrlly negative. Criticisn~sinc:lurle
the absence of a clear rrlaiionship hct\veen many of'the measures and bellavior relevalit Lo custody; tho failure to assess
factors c:learly deemed relevant LO custociy decisions; and an
ahscnce of iniportant data regarding l~usicps!-chornctric properties, including predictive validity (Ardi~li,1995; Hcinze &
Grisso, 1996: R4elton, 1995; Melton et al.. 1997; Otto Sr Edcns,
2003; Otto et al., 2000; Wellman, 1994).
In summaq, all measures that purport to assess cor~stmcts
directly relevant to chiltl custody determinations s d e r from
significant limitations. In fact, no study examining the properties of these measures has ever been pul~lishedin a peer-revie\\-ed journal-an
essential criterion for science and, in
theory, for the courts. In our rierr, the ahsence of scientific
support should preclude the use of arry of these ibrensic assessrnent instruments for any purpose other than research. We
ever1 have doubts about the value of research using these
measures, hecause it is hard to conceive of an!; psychological
test that could measure all the factors t h a ~inight be relevant to
child custody (Shnman, 2002) or that rnight assess the hest
custody arrangenlents fbr cliilclren rvhen the criteria for fulfilling
cl~ildren'sbest interests are so poorly defined (Emerb 1999b).
Rolre1.t E. Emcry, Randy Ii. Otto, William T. O'l)oiiohl~r
Clinical Assessment Ilistruments: Some Cautions i n t h e
Custody Context
Heilhrun el al. (2002) describe measures of intelligence, personality, psychopathology. and academic acliievemen~as cliiiical assessment instruinents. In contrast to forensic assessment
instruments, we believe use of many ol these measures is warranted in forensic assessment contexts to the degree that they
ofler reliable arid valid assessments of relevant constructs
itlentified in the law. We (lo, howe%er,wonder about the routine
use of measures such as IQ tests, ~ i l i i c hcan add to the time and
expense of a custody evaluation without holding a clear relevance to the issue hefore the c:ourt.
A greater concern is the validity of c1init:al assessnlent instruments in the custody contextl as a number of considerations
suggest the need for caution. For one thing, as iii other forensic:
contexts. examinees may he less than candid in their responses,
including on psychological tests. Tests ~liatdo not include
measures of response style are particularly vulnerable to dissimulation, while tests will1 embedded nleasures of response
style are not necessarily impervious to false repolfing.
LVliether the constructs assessed by the instrument are,
broadly conceived, "states" or "traits" is another inlportant issue. Assessrneiits of characteristics that cornmonly change over
time (e.g., parental depression) provide a \veal< basis for an
evaluato~.to make claims ahout how a parent functioned ill the
past or will function i11 the future. Rrc:ause families are evaluated during a period of high stress, moreover, evaluators also
must be cautious about drawing inferences ahout functioning at
some later, hopefully less stressful; point in time. Given the very
nature of custody disputes ant1 the context in ~vhichmost custody
evaluations occur, il is particularly inlportant that the evaluator
not assume that instrui~~ents
assessing more enduring styles will
rlot change in response to situational factors. The Standards for
Educational and Psycho1ogic:al Testing (American Educational
Reseul.ch Association, American Psychological Association, &National Council on Measurement in Education, 1999) direct
that "a test taker's score should not be accepted as art.flec~ionof
lack of ability with respect to the characteristic being tested for
without coiisideratioii of alternate explanations for the test
taker's inabi1it)- to perforni on that test at that time" (p. 43). The
upheaval of divorce c o n s t i ~ u ~ ea sreasonable "alternative explanation" that should certainly he considered when interpreting a test score.
We do not want to throw out the baby with the bathwater.. There
may be a role for clinical assessment instruments in some custody evaluation contexts. More specifically, to the degree that
there is a psychological construct that is relevant to the issues at
the heart of a custody matter and there are valid psychological
measures of that construct available, use of such measures can
he ol some value. Examples of relevant things that may need to
he detem~inedin a custody rase might include whether a child
has a learning disorder that needs special attention, whether a
motlirr suffers fi-om depression that affects her ability to meet
her children's eriiotional needs, or whether a lather has a suhstanc,e-abuse disortlei- 111atresults in him placing the children in
a ~ - r i s ksituations when in his (.are.
Prqjectil:e Measures
Our concerns about clinical assessment iusLruments apply to
highly structured, well-validaterl. and well accepled measures ol
intelligence, academic achievemc:nt, and psychopathology. These
issues present the greatest concerns. howevel; for unstructured,
projective measures, given questions that ha\-e been raised about
even ljasic psychon~etricproperties of such tests, including their
i~eliahilityand validity. There is a considerable difference of
opinion and ongoing, active debate regarding the general utili~yof
projective measures such as the Rorschach Inkblot Technique
(compare Wood, Nezworski, & Stejskal, 1997; Wood, Nezwoi-ski,
Lilienfeld, (92 Carb, 2003 and Weiner, 1996; Meyer, 1997,2001),
Draw a Person, and Human Figure Dra~jings.The veiy existence
of this debate, in combination with some of the specific cri~icisms
and potential dangers in the custody context, lead us to suggest
that such iiieasures not he used in child custody evaluation
contexts, or any other evaluation c:oatexts for that matter.
We do not have the space. exper~ise,or the inclination to review
literature on projective tests in thib
the broad and
monograph. Thus we only point to the extensive and serious
controversy, arid nole this: Questions about tlir value of projectives
or any other assessment technique need to be tlrbated and answered by psycliological scientists outside of the courtroom. It is
naive to expect judges to make intormed judgments ahout the
psy(:hometric adequacy of projertive measures iii the context of a
custodj- hearing. We also are concemetl about he potential lor
evaluators to assert that projective measures have scientific authority while the underlying empirical. legal, and values questions
remain unanscvered, precisely because the "test" is mysterious to
lay obsemers and therefore potentially misleading or difficult to
challenge. A nonexpert might feel competent challenging the
relevance or the didi it^ of a relatively straiglitfolward measure
like an IQ test or an MMPI-2. Yet, despite more significant concerns about its psychometric prope~-ties,results of a Rorschach
may be more difficult to challenge precisely because of its more
obscure source of material ancl scoring (Shuman, 2002j.
The cliiiical interview is another assessnlenl technique that
requires considerahle caution when use11 as a rneasurenlent
technique in custody evaluations. Iri~erviewers may yield
inferences that are reliable or unreliable, valid or. i n d i d . hut
there are no structured intervielvs with well-established psychometric properties specifically developed for use in the child
custody context, and survey data regaiding psychologists' custody evaluation practices indicate that use of any structured
interview approach is virtually unheard of (Ackerman & Ackerman, 1997; ICeilin & .Bloom, 1C186).Thus, differences between
interviewers may result from variance in the family's responses
or from the contrasting structure, content, or interpretation ofthe
interview. We urge psychological scientists to work to develop
structured interviews for the custody contcxt. In the meantin~e,
we expect custody evaluators to continue to inteiview families.
Although we are dubious about the psychometrics of urlsll.uctured interviews, we find some comfort in the fact that, unlike
prqjective measures, interviews are more straightforward and
understandable and hopefi~llpare not presented as providing
data as scientific-sounding as that of a test.
syndroille (PAS),(b) children's wishes regarding custody. and (c)
overnight visitation for very young children.
PAS: Asseriin,g S'cir>n,ceWhere There Is 12'0ne
"Parental alienation" is a construct ranlted high on the list of
factors evaluators consider to be directly relevant 1.0c:ustody
decision making. There is no test instrumelit designed to
measure parent alienatiori. Ra~hel;it is a "diagnosis" reached
through clinical interviews. Sorrle esperts have testified to
malting the diagnosis of parental alienation syndrome. ancl ~ l ~ e i r
testimony is claimed to be an important influence on judicial
decision making (Cardner, 2004).
Direct Obseruation
"Parental Alienation Syndrome" is a term created by psyDirect observation of parent-child interactions is another
chiatrist Richard Cardnel. (2001) based on his clinical expericomplex and generally unstandardized assessment strategy.
ence with custody disputes. Cardner asserts that PAS, which he
Threats to validity itlclude reactivity, unreliable coding systems,
says develops almost exclusively in the context of custody disunrepresentative samples of behavior, and problerriatic data
putes, is characterized by onc parent "programming" a child
conlpilation and analysis. As with inter\-iews, we urge the deagainst the other parent (Cardner, 2001). The assumption is that
veloprrleri~of standardized obscrvation measures for use in the
a child's disdain for one parent is gt.lir~allyunjustified and
custody context, and urge evaluators to describe their ohservasolely attributahle to denigration on the part of the other, altions clearly and to identify the inferences ~lieytlra~vfrom obienating parent. Gardner i2004) also claims that PAS can be
servational assessments.
"diagnosed" reliably and validly by expert evaluators, although
he offers no explicit criteria for doing so or objective evidence to
Contbinirzg .4ssrssnre1~~
Resz~ltsC L I Z Drazoing.
~
Inferences
s u ~ ~ p ohis
r t claim (Emery, 2005).
This last point raises a broader and very i~nportantissue. A11
Rie recognize that parents often undermine each other3 relaclinical assessment instniments assess constructs that, a1 rnosl, are tionships with their children following separation (F,mery, 2005;
only indirectly relevant to custody; thus their use in custody I<elly &- Johnston, 2001). We also note that many state statutes
evaluatioiis typically requires inferences to be made. Once a i~lclutlea "friendly par en^" rule, a prefrrencc for awarding cusparent's depression or a child's academic abilities are assessed. for. tody to the parent who will be more likely to promote the chilexample, the examiner may draw some inference regarding how dren's relationship with the other parent (Elmd & Spector; 2004).
that factor is relevant to the hest interests of a child. The question However, the scientific stalus of PAS is, to be blunt, nil. As
is: How is the examiner to dlaw conclusions from a single measure Gardner (2004) himself noted in a recent posthun~ouspublicaor, even morc importantly, cnrnhine data from several sources to tion, only one study of parent alienation ever attempted a statisfo1-m a conclusion ahout the best interests of the child? For ex- tical analysis: his own. Ve17 recently, Johnston conducted two
ample, how does an evalualul weigh the results of a hypothetically studies of' case records designed to identify the sources of alaccurate (i~npossiblein practice) evaluation where all data indi- ienation; she found many contributing factors leading to a child
cate that the mother is an effective disciplinarian but not tei~ibly aligning with one parent against the othei; including high-conflict
warm and that the father is warmly supportive hut not good at custody litigation and poor parenting on the part of the "aliensetting limits? Thus. our concerns with clinical assessnlent in- ated" parent (cited in Johnston & lielly. 2004).
struments are not only how to measure relevant constn~ctsreliably
We believe that it is blatantly misleading to call parental aland validly in a difficult context, but also how to synthesize rnul- ienation a scientifically based "syndrome" (Emery, 2005). Caretiple rlleasurements in a manner relevant to the ultimate issue of a ful assessn~entsof each parent's willingness to suppor-t the other
custody determination. Perhaps ideally. the law would provide a coparent clearly may be relevant to custody, but there is no esformula for making such decisions, but the factors to be considered tablished way of measuring "alienation." Evaluators therefore
in the law are rarely even ranked relative to one another.
must carefully identify the sources of their information concerning a more or less "friendly" parent, as well as the inferences
they draw from these assessments. Certainly, these assessments
are best conducted by an evaluator who interviews both parents.
Colltroversial Topics Kequiring F u r t h e r Investigation
Surveys of practicing custody evaluators indicate that, in ad- something Gardner (2001) did not do in many cases.
dition to using clinical assessment instruments and dubious
forensic assessment instruments, they also frequently assess Children S Frzshes
certain q ~ l i t econtroversia'l constructs. We illustrate our con- Surveys lnd~catethat custody ex aluator s place c o n c ~ d able
e~
cerns by focusing on three in particular: (a) parental alienation importance on children's stated preferences regarding custo-
Robert E. Enwry, Randy I<.Otto, William T. O'Donohuc
dy-particularly
the preferences of adolescents, but also of
children as young as 5 years old (Ackerman & Ackerman, 1997;
I<eilin Br Bloom, 1986). This surely reflects the fact that children's wishes regarding custody typically are included in state
laws as a factor to be considered when detrrminirig children's
best interests. In fact. some statutes explicitly direct that thr
wishes expressed by a child of a given age-for example, 12
years old-should determine custody if there is no reason why
those wishes should not be followed (Elrod & Spector. 2004).
Although all agree that the wishes of teenagers can be influenced
by unfortunate circumstances (e.g.. a parent's greater material
resources or permissiveness), laws regarding the expressed
wishes of children of a certain age both respect the increasing
autonomy of adolescents and recognize the realistic difficulty of
~ryingto keep children in an arrangement to which both they and
one parent object.
A policy of acting on the freely expressed wishes of' an adolescent is not without
but far bigger prohlems (and
controversies) arise in regarcl to wishes of children who (a) are
school aged or even younger andlor (b) do not come folward with a
freely expressed preference. Some psychologists have offered
that, even in these circumstances, children should be encouraged
to express a preference regarding custody a.; a means of empowering them (see Weithorn: 1987). Others express concern that.
instead of giving children the right to have input, such policies
give children the responsibility for making adult decisionsdecisions that the aclults have failed to make themselves (Emery,
2003). Still others say that children's preferences should be assessed only sensitively and indirectly and that this information
should be used as feedback to facilita~eindependent parental
decision making (McIntosh, Long, & Moloney, 2004).
One of us has taken a strong position against attempting to
assess children's unexpressed wishes (Emery. 2003), but our
present concern is more basic.. The freely offered preferences of
children-particularly
older children-are
important considerations in custody evaluations for both practical and legal
reasons, but there is no direct evidence on how or indeed
whether evaluators should assess the wishes of children who, for
whatever reason, do not express them.
0uernight.s With Infar~tsand Toddlers
A final controversy we will discuss is whether or to what extent
infants and toddlers should have overnigh1 visits with their
nonresidential parents. Children's age in relation to overnights
is not a consideration mentioned often in surveys of custody
evaluators. but it stands as an example of the sorts of controversial issues that evaluators often are asked to address. Other
such issues include the question of whether a residential parent
with primary physical custody should be allowed (if there are
good reasons) to move with the child away from a nonresidential
parent, or under what circumstances parental conflict is so intense that joint physical custody is unworkable.
Using differing interpretations of attachment theory, leading
psychological scientists have talten strong and very different
positions on the issue of overnights involving young children. A
document prepared for the Spoltane (X'ashington) Bar Association, and endorsed by many leading attachment researchers,
called attention to the psychological inlportance of young children's secure attachment with a primary attach~nentfigure.
Based on research and theory on the primary attachment, the
report recommended against overnight visits with the nonresidential parent until children are 4 years old (Spokane County
Bar Association. 1996). In contrast, in a paper published in a
major family-court journal, other leading psychological scientists highlighted the importance of children's attachments to
multiple caregivers. Focusing on the value of developing multiple attachments, the authors recommended that infants should
have regular overnight visits with nonresidential parents in the
first year of life (I<elly & Lamb, 2000). Both interpretations offered various caveats about the quality of children's relationship
with the nonresidential parent, parental cooperation, and similar issues. hut they clearly came to ~ e r ydifferent substantive
conclusions about what psycholog~calscience indicates regarding whether, when, or how often infants and toddlers should
have overnight visits with nonresidential parents.
Therr is only meager direct evidence on the harm (Solomon &
George, 1999) or absence of harm (Pruett, Williams, Insabella,
& Little, 2003) associated x-ith overnight visits for very young
children. As with the issue of children's wishes, the psychological scientists debating the question of overnight visits apparently come to logical conclusions based on their own,
theoretical premises, yet the limited state of knowledge allows
reasonable scientists to come to opposing conclusions. Such
differences of opinion are of great value in science, but when
translated into policy recommendations, they can confuse and
confound judges, lawyers, evaluators, and parents. For example,
we have had distraught mothers approach us in shock after being
court-ordered to stop breast-feeding their infants to allow for
smoother overnight visits, yet we also lcnow of judges who claim
to overturn consensual parenting plans if they include overnight
visits for children 3 years of age or younger because of worries
about disrupting attachments.
One of us has developed a set of guidelines for parents about
overnights and other arrangements for young children that
represents what we believe to be a balanced position (Emery,
2004). Howex er, our point h e ~ is
e that. whatever conclusion one
reaches, it is based on limited evidence. Psychological scientists need to recognize and acknowledge their limited data base.
Our bigger point, to which we turn shortly, is this: Custody decision malting and custody evaluations have an impossible task in
attempting to determine children's future "best interests" in cases
where parents cannot agree. Neither the wisest judge nor the most
insightful evaluator has good answers to impossible questions.
The custody report completed by Dr. Hagan in our fictional
case illustrates our various concern5 with the limited science of
Critical Assessmr~ltof Child Custo~lyEvalu;~tions
c u s ~ u devaluations
~
a d , more i~nportanlly.s h o ~ v tsh e p r ~ ~ h l e n i s
that [:an arise u n d e r t h e reginie of' vague c:ustody laws an11vague
professional a n d ethical standards for custody evaluators.
urged Ms. Deer-Doe LO cheer up, c:ontinue to he a wondelful
mother, and LO be on her very hrst behavior so a? not 10 give her
soort-lu-be-ex-husband any anmrunilion in his canlpaigr~against
her and mothel-hood.
T h e D e e r - D o e Case: Dr. H n g a n ' s C u s t o d y R e p 3 . t
A Bigger P r o h l e r n : T h e Legal a n d E m o t i o n a l C o n t e x t of
l)r. Hagan wrote a 35-page report sulnmarizing his evaluation of
111eDrer-Doe family. The reporl containrd precise details of the
results of the various standardized tests, hut the lawrpers were only
really interested in the final paragraphs under the heading,
Sr~n~rnary
a r ~ dR~comrr~end(~tions.
''In sumlnary, substantial evidence poink to Ms. Dee1.-Doe'.;
longstanding depression, her irllt.rls+ repressetl hostility lo~vilrd
Mr. Deer-Doe, ancl PI- alienation of the chilclren against their
father. In contrast. Mr. Deer-l>oe appears Lo be \\,ell adjusted. is
eager LO promote thc children's relalionship with theirmothrL and
is able and interested in being a full-time lather. It therefore is
recommended that; in order lo promole his k~estintet.ests, Carlos
Deer-Doe h~ sliilted imnlediately to his lather's cus~ody xith
regular visits with his mother, provided that she enters inlo intlividual psychotherapy.
"Although Isabella's intense anger at her father is largely a
pl.oduc;t of alienation, no change in c,ustody is recon~n~ended
hr
hel- at ilris point in time, 1,rvause she is clo.;rly allied wit11 her
mother and is likely to continue LO reject and rebpl against her
father's care. Instead, indi\,idual psychotherapy and family therapy rriih hn- lather is recommended for Isiibella, wilh further
evalualion in 3 to 6 months depending upon the rrcommenda~ions
of Isabella's therapists and he1 mother's t l ~ e r a ~ ~ iiIs trelevant.
,
h
key consideration at that time xvill be whether Isahella's stated
wish to live with her nlother, if she continues to voice this prefrrence, is a rcsult of alienation."
Eihen he read the evaluation, John Deer-Iloe was juhilarlt. H(>
felt vindicated. ragrr to k~ea full-time f~ltheragain. and eucilecl
almllt the prospect of' startinghis new famil!; He V I I H - he
~ ~ T P ~ n01r
S
going to get remarried " t h ~day after nl!. divorce is final." His
lawyer, w~hoalso \\a> e ~ i c o u r a ~ eI -dJ ~ Dr. Hagan's sepor-l and serommendations, told John that the evaluation was not only a victory
for him but for all fathers. '-Sometimes the system really does
worlc," she offered.
Jane Deer-Doe's reactions were understandably quite different.
Shocked dnd l~anic:ked,she hrca~ileemotionally t l i s ~ r a u ~ in
h t her
la~r~!.er'soffice. He eventually Irelped Jane calm down by telling
her that he had learlied on1)- recently that Dr. Hagan, who used to
he fail. aucl evenhanderl. had lxconir notol.iously 1)iasecl in favor of
fathers as a result at losing custody in his own. bitter cjivorre. If he
had kno~enthis a Cew months ago, Jane's laxvyer told her. he never
would have agreed to Dr. Hagan as the c:our-t-appointed evaluator.
M s . Deer-Doe's attorney wen1 on to offer that he would postpone
the pending hearing in order to get a sacorld evalua~ionby anoihcr
mental health profjssional and have Dr. Hagarl's evaluation revier~edhy a third professional so as to identilp any important
lilnitations 01. rurilknessrs. If' the c:ourt rehseil to appoint a more
ol,jective, neutral evaluato~;then he ruoulrl hire an expert rvho
would do the job right. In any case, the postponrment meant that,
at a minimum, no changes in custody would take place for 6 to 9
months given the congested court calendar. [n the meantime, lle
Custody Disputcs
Wie could conclude our monograph here with this summalp:
There is essentially no psychological science to support the
measures a n d i:onstructs designed specifically for the assessmen1 of child custody arrangements for individual children.
Moreover, established measures of clinical constructs must h e
u s e d with caution d u e t o threats t o thcir validity a n d questions
about the relevance i n t h e custody context o f t h e constructs they
assess. We also could conclude t h a t t h e state of psychological
science i s too l i n ~ i t e dto reat:lr clear conclusiorls about controversial issues such a s c-hildren's wishes, overnight visits, or even
PAS, a n d remind t h e reader that t h e burden of proof falls o n
proponents of a particular hypothesis or reconimendation. To
these three p i n t s , we could a d d questions about ethics arid
professional practice-for
example, potential concerns about
s y s t e ~ n a i i cbias o n the part of evaluators, q u e s t i o r ~ sabout
whether evaluators should a r l r 1 r . r ~t~h e "ultimate issue" (i.e.,
recommend specific custody arrangenients), a n d worries about
a battle of experts when e a c h side hires its own evaluator.
Howevc:r, xvc believe there are bigger problems i n custody
evaluations than shoddy science, a n d we also believe that
consitleration of these broader issues points t h e way to some
prorr~isir~
svlutiorls
~
for custody evaluations; children, a n d
families. Thus, we turn now to examine t h e more general literature on children's adjustment to their parents' separation a n d
divorce. After this, h e outline three gcneral recommendatioris
that xve consider ill light of psychological research, legal analysis, a n d professional responsibilities including various issues
we raised about Dr. Hagair's custody evaluation.
-4VERAGE EFFECTS AND VARIATION IN THE WELLBEING OF CHlLDREN FROM DIVORCED FAMILIES
There is a large, sophisticated, multidisciplinary research literature on 1101~childrerl are affected by parental separation a n d
divorce. Wie cannot review marly original sourc:es from this l i ~ e r aturc in his limited space, illthoupll we have done s o elsewhere
(Emer): 1999b). In the following section, w e offer an ovcl-vieru of
t h e nitljor r ~ n c l u s i o n sresearchers have drawn. After this, \re
consider what factor:;
children's niore or less adequate
adjustment. For prrsent purposes, researc.11 u11 the a\-erage wcllbeing of c h i l ~ l r r nfrom divorced families is of interest p r i ~ r ~ a r ial ys
a starting point fol- examining predictions of individual differtly
ences i n outrome, one of the main goals of a c ~ ~ s ~ uevaluation.
Thus. we review this extensive literature only briefly.
On average, parental divorce i s associatetl with a n increased
risk fbr a variety of' pspchological pl.oblems among c:Ilild~-en
Robert E. Erncr,, Rand\ I\. Otto. Bilham T. O'Uonohur
(Emery, 1999b; Hetherington & Icelly, 2002; McLanalian &
Sandeiur, 1994). In a meta-analysis of 9 2 studies, Arnato and
Keith (1991) found an average effect size of .14 standard deviation units when comparing children from divorced versus
married families across all child outcomes. Anolher metnanalysis of studies in the 1990s found that the average effect size
was somewhat larger than this earlier estimate, ranging horn a
low of .12 standard deviation units for measures of self-concept
to a high of .22 standard deviation units for conduct problems
(Amato, 2001).
While the effect sizes suggest a modest, average increase in
psychological problems, it is important to underscore the variability in the psychological adjustment of children whose parents separate and divorce. Most children are resilient despite
their parents' divorce, as indexed by measures of psychological
inaladjustment that do not diSSerenliale them from children
\\hoar parrnls rrmain continuously marrietl (Emrl?, 1999a;
Emery & Forehand, 19'94). Still, depending on the outcome,
parental sepnrrrtion or ilivorce is linked will1 a 25% to 100% (a
d~uI)ling)
increase in the rislt for psychological difficulties at the
extremes of the distribution (Hetherington Sr l<elly, 2002; Zill,
Morrison. & Coiro. 1993). Given the high prevalence of sepaI-ation and divorce. ?\;en a inodest increase in risk translates into
an important societal concern.
Noni-ancloni Selection I n t o Divorre
Still. at least some of the putative "effects" of parental divorce on
children, perhaps as much as 50% of lhe variance, are due lo
nonrandom seleclion into divorc:e. IHany of the problems found
among children from d i v o r c d families actually are present
before the parents separate (Cherlin et al., 1991) and therefore
cannot he consequences of parental divorce, although this selection effect seems to be stronger in accounting for the psychological difficulties of children than for those of young adults
(Cherlin, Chase-Lansdale, & McRae, 1998). Behavior geneticists have raised the strongest selection argument, suggesting
that children's risk in divorce may be fully or partially attribtable to the passive gene-environment correlation, because
genetic factors inllluence divorce and may also affect children's
behavior (McGue & Lyklten, 1992). Despite this importan1
concern, in one adoption study (O'Connor, Caspi. DeFries, &
Plomin, 2000) and one twin study (D'Onofrio et al., in press),
tlivorce still was associated with a diniiriished but increased risk
lor psychological problems, particularly externalizing problems,
among children.
Different Risks for Different Outcomes
Externalizing difficulties are the child emotional problems most
strongly linked to parental separation and divorce (Amato,
2001; .Illlato Sr Keith, 1991; Emery, 1982, 1999b). Other
emotional difficulties less strongly tied to parental marital status
include depression; anxiety; poor school behavior and per-
fornlance; and dilficulties in romantic relationships, including
an inrl.eased risk for divorce among offspring (e.g., McLanahan
& Bumpass, 1988). A significantly increased risk lor trouk~led
falnily relationships, especially belween chiltlren and their fathers, also accompanies divorce. One national study found that
fully 65% of young adults between the ages of 18 and 22 ~vhose
parents were divorced had poor relationships with their fathers;
only 29% of those whose parents were married had poor relationships with their fathers (Zill el al., 1993).
Scientific research notwithstanding, some clinical investigators point to case studies indicating that the adverse consequences of divorce for children are unexpectedly large (e.g.,
Wallerstein, Lewis, & Blakeslee, 2000). We believe that this
conclusion, and much of the debate about it, is due to confusion
of psychopathology with what one of us has termed psychological distress or "pain" (Laumann-Billings & Emery, 2000). Even
resilient, well-functioning young people whose parents divorce
report considerable distress in regard to their memories of their
childhood ("I had a harder childhood than most people"), feelings about their current family relationships ("Sometimes I
wonder if my father even loves me"), and concern over events
where both of their parents will be present ("1 worsy about big
events lilie graduations or weddings where both of my parents
will have to come"; Laumann-Billings & Emery, 2000). Thus,
even if resilience-as defined by the absence of mental health
problems-is
the normative outcome of divorce for children,
children's resilience often is colored by painful memories of the
past, difficult ongoing feelings about family members, and
concerns about future family interactions. There is increasing
agreement that malting this distress-versus-disorder distinction
may help clear up much of the controversy about the consequences of divorce for children (Kelly & Emery, 2003; Wallerstein, 2003).
PREDICTORS OF CHILDREN'S PSYCHOLOGICAL
ADJUSTMENT TO DIVORCE
Average outcomes are an important baclidrop to our discussion, but the prediction of individual differences in children's
psychological well-heing is more directly relevant to custody
evalualions. In the follo~+-ing
sections. w-e review research on
different risk factors. relying primarily on secondary versus
original sources because of space limitations and the large
number of studies.
P a r e n t a l Conflict
A large body of research den~onstratesthat conflict between parents is associated with an increased rislt for psychological problems
among children in all families, whether the parents are married,
separated, or divorced (Ahrons & Miller; 1993; Ahrons & Enner;
2003; Amato & Keith, 1991; Emery, 1982; Johnston & Roseby,
1997; Otto, B~Lffington-Vollum,& Edens, 2003). Although non-
Critical A S S I ~ S S I IoIl IChild
: ~ ~ Custody Evaluations
random selection cannot he completely ruled out, man!; analogue diator, and not asking children to make decisions that the parents
experiments demonstrale that conflict simulated in the laboratory themselves cannot make (Emery, 2004). Another biief excerpt
or recorded systematically at home directly causes some adverse from the Deer-Doe case illustrates the sort of' conflicts that can he
reactions among children (Cummings &- Davies, 1994; Davies, all too familiar in separation and divorce.
Harold, Goelte-More); & Cunlmings, 2002).
Parental conflict often precedes a separation or divorce, and
Conflict a n d t h e Deer-Doe Case
various studies demonstrate that children fare better psychologically if they live in a harmonious divorced family than in a
As conlinued legal maneuvering delayed what lie though^ would
conflict-ridden two-parent family (Emery. 1982). Because sepbe the speedy implemenlation of the reco~nmendationsmade in
aration can bring relief from the struggles of living with parents
Dr.. Hagan's cus~odyevaluation, John Deer-Doe grew extremely
in a conflict-ridden marriage, we therefore must add improved
fiustrated with his children's mother. wit11 the legal system, and
psychological adjustment to the range of variability found i n
especially with not ]wing uhle to see his cllildren regularly. As a
children's
outcomes following their parents' diresult ol several letters Com his lawyer and angry e-mails with
vorce. This "relief hypothesis" is supporled by research findings
Jane, for the first time since the separation, he had the children
wilh him for a long, 3-day holiday weekend. John had a great time
that show childrrn's improved adjustment a l ~ e rseparation in
with C:arlos on his Friday off and on Saturday, hut he was deeply
high-conflict marriages. However, a new and important twist is
disappointed by Iaal~ella'spersistent distance and moodiness. Ilis
what happens to children from low-conflict marriages: Sex-era1
frustration erupted on Saturday evening when he asked Isabella
recent studies have found that children fare better follolving
why she didn't spend more time with him and answered his own
separation from a high-conflict marriage but worse when their
question by blaming her mother's interference. Reforr Isabella
low-conflict parents separate (Amato, Loomis, & Booth. 1995;
could even react, he asked, "Wouldn't you like to live with me half
Peris & Emery, in press). In I'act. Ainato (2001) argues that. in
of' the time?" At lhis point, Isabella pxploded. ''I told Mom a
close to half of divorces, the marriage had been "good enough"
hundred times. I want to li\-e with her! I don't want Lo see you! I
from the children's perspective. That is, parental conflict had
wanl LO go home!"
been sufficiently well contained that the children do more poorly
Huit anti angry. John S C I - ~ L I I I Iback.
~~
"Fine!" He threw Isafollowing their parents' separation than they would have done
hrlla's things into her hackpack, and relurned her to her mother's
house. They drove in silence, hut as Isahella opened the car door,
had their parents stayed together.
John tnld her, "You can tell your inother lhat 1'11 bring Carlos back
Whether or not more parents could stay together for their
101no1-row..
. inavLe." Isahella burst into tears, slammed the cur
children's sake, these data point LOthe psychological importance
door
shut.
and
ran to her mother's front door. Jolln drox-e away
of coliflict and to the fact that parental separation does not
hefore the door opened, not knowing whether Jane was even home
necessarily decrease it. Conflict can, i11 fact, increase following
or not.
separation, continue for years, and come to focus more squarely
on children who are a point of connection between former
As this vignette illustrates. hurt, anger, and conflict between
partners (Emery, Laumann-Billings, Waldl.on, Sbarra. SI Dillon,
separated parents can take many forms. and can erupt even in
2001; Johnston, 1994).
the absence of the other parent. The vignette also shows how the
Parental conflict can affect children directly by creating stress
conflicts that may undermine relationships between separated
and anxiety (Kelly, 1998) and indirectly by u n d e r m i n i i ~ ~
parents can lead to conflicts between parents and children that
parenting quality and the children2 relationship with one or 1,oth
undermine crucial parent-child relatioilships as well.
parents (Otto et al., 2003). As with divorce itself, conflict after
divorce is linked with a variety of short- and long-telm psychological problems among cliildrer~.ranging Irom conduct prohleins Parent-Child Relationships
to depression (Emery, 1999b; Schmidtgall, IGng, Zarski. & In most studies of children Iron1 dix-orced families, the quality of
Cooper, 2000). However; not all conflict is equally disruptive to the relationship between a child and his or her primary resichildren's emotional well-being. The results of systematic ana- dential parent is the strongest predictor of that child's psychologue studies (Cummings & Davies, 1994), together with field logical well being (e.g., Ruchanan. Maccoby, &- Dornbusch,
research (Grych & Fincham, 1990) ant1 clinical experience 1996; tIrtherington & ICelly, 2002: Mar~inez&- Forgatch, 2002).
(Emery. 2004). suggest that conflict is least destructive when it (a) The most xridely accepted classification of parenting groups
is contained between parents; (b) is relatively infrequent: (c) is caretakers into four categories based on the degree of warmth
less inlense emotionally or physically; (d) resolves; (e) is not about and control they offer to their children (Lamborn, Mounts,
the children or chiltlrearing; and (f,) does not involve the chil- Steinberg, & Dornbusch, 1991; Maccoby & Martin. 1983; Nodren-which
includes not arguing in front of or around the vak, 1996; Steinberg, 2001). Authoritative parents are warm and
children, not asking children to carry messages hetwern parents. involved, and they consistently and democratically enforce
not deriding the other parent to the children, not expec~ingthe developmentally appropriate rules and discipline. iluthoritarian
childlreil to take sides, not malting a child a scapegoat or a me- parents offer their children low ~varmthand high control, using
-
Robert E. Emrrv. R a n d ~I<.Otto. William T. O.Ljonoh11~
more frequent and autocratic punishment (Novak, 1996). Perm.issi.ue parents are loving but indulgent, and they offer children
e discipline about controlling their behavior.
little g ~ ~ i d a n cand
Finally, neglectJiL1 parents provide children with little affection
or discipline.
Research on two-parent families consistently indicates that
children ofpreschool age through adolescence who are raised by
authoritative parents fare best on indicators of psychological
and behaviol.al health, while the children of neglectful parents
fare worst (Lamborn et al., 1991; Maccol~y& Martin, 1933;
Novak, 1996: Steinberg, 2001). Research on children in divorced lamilies also shows that authoritative parenting hy the
primary residential parent is linked w,ith better postdivorce
adjustment (Buchanan et al., 1996; Fauber, Forehand, Thomas,
& Wierson. 1990; fietherington, Cox, & Cox; 1982; Hetheringtori & Kelly, 2002; Thotnson. Hanson, 8; McLanahan, 1994).
Wc should note, however, that more authoritarian parenting
styles are found to he equally or inore effective in certain contexts, for example among minority families living in potentially
dangerous en~ironments(where increased parental vigilance
and authority rnay be needed; Deater-Decltard, Dodge, & Bates,
1996). Authoritarian parenting also predicts lower levels of
substance use among adolescents living with divorced parents
(,Buchananet al., 1996).
Ilother,, lelauc Fathers
1 s noted al~ove,most children live primarily with one parent
iollo~vingseparation and divorce-approximately 75%' live with
their motl-~ersand 10% live with their fathers. Although some
early, small-scale studies intlicated that children who lived with
their same-gender parents were better adjusted than their
c30unterpa~-ts
living with opposite-sex parents (e.g., Santrock &
'iiirshak, 1979). these findings have not been replicated in more
recent research employing large samples (Buchanan et al.,
1996: Downey & Powell, 1993). In general, researchers find that
c,lrildren of both genders function equally well living priirlarily
either with their mothers or fathers (Downey, Ainswor-th-Darnell,
B Dufu~; 1998); however, a few investigators have fount1 that
children do somewhat better in sole-mother residence than the!.
(:lo in sole-father residence (Buchanan et al., 1996). Still, differences between primary-mother versus primaiy-lather resiidential arrangements, if they are found at all, are not large in
i~iagnitude.Thus, neither p r e n t a l gender, nor the interaction
!:letween parent and chilcl gender, has been found to inoderate
,hildren's well-being in an important way.
The extent to whicl~children's relationships with their "other"
i7arentspredicts their psychological well-l~eing,particularly
:)hen there is parental conflict, is one of the most controversial
i s u e s in custody law (e.g., favoring or opposing joint physical
.~,ustody)and in custody evaluations. Data are not coriclusive,
liut there is research relevant LO these issues. Given that the
;.sue is so pressing, we believe it is important to draw some
:lear, if qualified, conclusions from the available research.
Contact Between Children a d Nonresident Parents
An iniportant demographic issue that we have not addressed, but
that bears in a very important way on parent-child relationships
following a separation, is the extent of contact between children
and their nonresidential parents. Seltzer's (1991) analysis of the
1937-88 round cd the National Survey of Fanilies and Honseholds data provides detailed and higli-quality, if soniewhat dated,
evidence on this issue. especially on the frequency c~fcontact
between children and nonresidential fathers. Three broad trrnds
characterized the findings from this national survey. First, contact
between nonresident, separated, or divorced fathers and their
children was not terribly frequent, even immediately after the
sepamtion. For example, onlp43% of fathers separated for 2 years
or less saw their children on a weelcly basis or Inore frequeutly.
while 30% of fathers separated for less than 2 years saw their
vhildren several times a year or less. Second, contact droppecl off
substantiall!. over time, such that 6 to 1 0 years follo~vingseparation, only 19% of nonresident fathers saw their children weekly
or more. while 62% had face-to-face contact with their children
several times a year or less. Third, higher contact levels Tirere
predicted by a variety of factors including less geographic Jistance between the parcnts' hc~useholds,a shorter length of time
since separation, al~senceof remarriages, the child having been
born into a legal marriage instead of out ofwedlock, and the child
being older rather than younger (Seltzer; 1991). Other evidence
from national samples shows that nonresidential mothers maintain somewhat more frequent contact with their children than
nonresidential fathers do (Zill. 1988).
Some commentators believe that father contact has increased
dramatically in the last 15 years, but the relatively modest increases in sole father custody and joint physical custody (reviewed earlier) make us skeptical that there have really been
any dramatic changes. In the most recent national data we could
locate, an analysis of 1998 U.S. Census data, 40% of nonresident fathers and 22% of nonresident mothers had had no contact
with their children in the previous year. Among the 60% of
nonresident fathers who had seen their children, contact
occurred on an average of 6 9 days per year. The 78% of nonresident mothers who saw their children did so more often, an
average of 3 6 days per year (Child Trends, 2002). These data
were not disaggregated by levels of contact, o\,ernight visits, or
time since separation, and they included parents who did not
live with their children for a variety of reasons (e.g., divorced,
never married). Still; the evidence indicates that. even in a recent cohort? 3 subs~antial nutnber of nonresident parents
maintain little contact with their children, and contact in the
range considered to be joint physical custody (aboul 100 overnights per year') is nclt the norm.
Nonresident F n t h ~ r i n gur~rlCh.ildran:r Psychological Well-Being
'Tlre normalive baclcdrop is important in considering the question
of whether more frequent contact with nonresident parents
predicts 1)etter psychological adjustment among children. A
Critical A s s e s s m ~ n of
t
Child Custody Evaluations
~neta-analysiso f 6 3 studies examining the relationship between
children's psychologic:il ~vell-beingor academic success and
different dimensions of the relationship between a child ant1 his
or. her nonresident father (i.r., payment of child support, amount
of contact, feeling close to the father, and authoritative parentingj
indicated that the amount of contact is a poor predictor of children's psychological well-being (Amato & Gilbreth, 1999). As
shown in Table 1, the weighted effect sizes (product-moment
correlationsj between contact levels and three indices of children's psychological well-being were uniformly very small. For
externalizingprol-~letns
and academic success. afathers' payment
of child suppol-t was a better predictor of his children's adjustment than was a father's contact with his children. In contrast,
authoritative parenting and, to a lesser extent, closeiiess to the
father consistently accounted for a significant if statistically
small proportion of variance in all three measured outcon~es.
The a ~ ~ t l l o rtested
s
for a number of variables that might
moderate these relationships, iuclutling cliild gender; age, race;
divorce versus nonmarital birth, and remarriage of the parents,
but none of these variables moderated the effect sizes in ail).
meaningful way. Although the ineta-analysis did not test for the
inoderating effects of parental conflict, Amato and Gilbreth
(1999) discussed the critical role ol' parental conflict, iiicluding
research indicating a positive effect of contact when parents
cooperate and a negative efkct when parents are in conflict
(,Amat0 & Rezac, 1994; Hetherington, Cox, & Cox, 1982).
Joir~tCustody nnd Child7-en%Psych~ologicalWell-Being
Whether joint physical custody is linked with better psychological adjustment among children is an iii~portantquestion ill
its own right, and if rhildren fare notably better under joint
physical custody than in other arrangements, a nonlinear relationship also might explain the weak association between nonresident-father contact and child outcome. Children may benefit
from spending more time with their fathers only when contact
reaches some high threshold (see, e.g., Cabrera, Tarnis-LeMonda. Bradley, Hofferth, & Lamb, 2000; Lamb, 1999; Lewis &
Lamb, 2003). Surprisingly, relatively few investigators have
examined how joint physical custody is assoc:iated with children's well-being. A recent meta-analysis (Bauserman, 2002)
located only 11 published studies and 22 unpublished studies
(21 of which were unpublished dissertations) with a combined
sample size of 814 joint-c:ustody children and 1,846 sole-custody children. Combining the results across measures, Bauserman reported a study-level overall effect size of 2.3 standard
deviation units, slightly above what is traditionally considered
to be a small effect. This analysis included hotli joint physical
a i d joint legal c ~ s t o c l ybut
~ surprisingly these arrangements clid
not differ significantly in their effects when compared to sole
custody (joint physical, d = .29 for 2 0 studies; joint legal, rl =
.22 for. 15 studies).
Importailtly, neither presence of past parental conflict (5
studies) nor that of current parental conflict (14 studies) accounted fhr significant variance in the joint-custody effect sizes;
perhaps of' more importance, ho~vever.joint-cuslody groups had
lower levels of both past and present conflict than sole-custody
groups did (Bausermali, 2002). As Bauserman noted, this suggests the very important possibility that self-selection into joint
custody may acc:ount for part or all of the results. We cannot
extrapolate frorii voluntary joint physical custody to circumstances when joint physical custody is imposed upon parents by
laws favoring joint physical custody, by evaluators who recom-
TABLE 1
WIeta,-Analysis o f 6 3 Stu'dies Shomfiing Ho~roStrongly Different Rela,tionships Between Children und Nonresitlent
Fathers Predicted Children's.-4cadem'ic S~rccessa,nd Psychologica,l Well-Being (Adapted From Am'uto and
Gilbretl~,1999)
-
-
Dimension oP ion resident fathei--child relationshil,
Child well-being index
Academic success
Effect size (weighted r )
Number of effect sizes
Sum of sample sizes
Externalizin~problems
Effect size (weighted T )
Number of effect sizes
Sum of sample sizes
Internalizing problems
Effect size (weighted r)
Number of effect sizes
Sum of sample sizes
Payment of
child support
by father
Amount of'cc~ntact
I~etweenchild
a ~ father
~ d
09"""
17
7,156
,03::
17
< .05,%&p<,01,**$ < ,(),jI,
Authoritative
parenting
h y father
4,918
OF"*
8
2.917
.02
6,308
-.01
8
1,916
03"
4.3
4',841
-
37
-
-
*p
Child feeling close
to father
-.07*
14
1.61 7
-
12:"'**
13
545
-
Kobrrt
E. Emery, Randy I<. Otto, William T. O'Donohur
nlend that arrangement, or by judges who order it. Finally, it is
importil~~t
to note that, although conflict difirences did not
accouni h r the advantage trf joint over sole custody in the metaanalysis, this analysis does not address the possibility that joint
physical custody may be the right solution for the wrong people
in contested-custody or other high-ccrnflict situations (Emerv,
1999b). At least some research sho~vsthat high conflict predicts
worse child adjustment within joinl-physical-custody groups
(Johnston, Icline, Si Tschann, 1989).
Thus, our conclusion al~outthe potential I~enefitsof joint
l~hysicalcustody is a cautious one because of (a) he important
couples selfand unanswererl question of whether 1.o~~-conflict
e1ec:t into that arrangement; (I,) concerns about the potential
tlamage to children caused by likely greater exposure to parental
conflict in such an arrangement; (c) the null results for father
contact lound in a more extensive body of research xvllere self,election is less of a concern; ant1 (d) the continued low prevalence rates of joint physical custody despite two decades of
expel.irnrntation. We believe that joint physical custody benefits
children when parental conflict is contained. Therefore, more
parents T V ~ Owant t(r attempt joint physical custody (and therefore are likely to be fairly cooperative) should be encouraged to
ti? it. However, joint physical custody seems to be a worlcable
arrangement only for a minority of parents and should not be
encourag~das the fair solution for parents who dispute custody
or othel~viseare in high conflict. Finally, B-e note that there is no
clear line defining when joint physical custody is potentially
lheneficial or potentiallj- harmful for children. The field woultl
l~enefitgreatly from research on what kinds antl levels of parenlal conflict and cooperation distinguish "good" from "had"
joint physical c:ustody.
Parents' Ililental Health
The Uniform Marriage and Divorce Act explicitly indictates thal
tl-rr mental liealth of all parties should be a consideration in deteimining cliildrenk hest interests. Statutes offer. little Illore than
this general guidance, however, thereby leaving much room ior
i~lterpretation.Thus, although mental health professionals can
assess mental health I$-ithadequate reliability and validity, cluestioris arise about the specific relevance of parents' mental health
prohlems for children, parenting, and custody anangemenis.
Einery 11999b) suggested that four mental health problems
ainong parents are of special concern to understanding the potential consequences of divorce for children: (a) depression, (b)
antisocial behavior, (c) major mental illness (e.g., schizophrenia
Y I I ~I~ipolardisorder), and id) personality disorders. Substance
s l ~ u s eshould also 11e added to this list. Parental depression is
associated with negative child outcomes in a number of studies
'Otto et al.. 2003), but the efferts are likely to be mediated
tllrougli parental conflict and inadequate parenting (Emery, Weintraub, & Neale, 1982). In their revie~vofthelilerature, Otto and
colleagues (Otto et al., 2003) reported that onc- of the rrlost con-istent findings is that parents who engage in antisocial behaviors
tend to have children ~ ~ exhibit
h o a number of behavior problems,
particularly aggression, delinquency, and other externalizing
problems. Children whose parents suffer from schizophrenia also
are at a significantly elevated risk both for schizophrenia and for a
range of serious emotional problems, although the increased risk
appears to result primarily fiom genetic effects as opposed to
childrearing (Gottesnian, 1991). Perhaps the greatest concern in
regard to schizophrenia antl other major mental illness is whether
the parent with the disorder is, with treatment, functioning sufficiently well to care fhr his or her children. A similar concern
arises regarding the well-1,eing of child re^^ who have a substanceabusing parent. Evidence shows that both genetic and environmental liabilities contribute to the increased risk for psychological prol~lemsamong such children (Walden, McGue, Iacono,
Burt. & Elkins, 2004), but the most pressing issue is the parrnt's
imi~iediatefunctioning and whether or not this impairs the parent's ability to care for or protect the safety of his or her children.
Finally, little research is available on how children are affected by
parental personality disorders, although experts in custody disputes increasingly recognize that personality disorders often are
an important concern, particularly in cases characterized by
chronic high conflict (Ehrenberg, Hunter; & Elterman, 1996:
Johnston & Roseby, 1997).
The literatures on parents' mental health, parent-child relationships, genetic tra~lsmission,and children's psychological
well-l~eiiigare too vast and complicated for us to consicler in any
detail here. Still, several broad conclusions seem clear. First.
some evidence shows that children are adversely affected when
their parents have emotional, behavioral, or substance-abuse
problems, but the children's problems might be caused not by
their invariably troubled relationships with their parents but by
genetic risk or life hardships associated with heir parents'
I)sychological problems (Jenuxvine & Cohler; 1999). Second,
whether or not a parent is engaged in treatment is a major
consideration for serious emotional p r o b l e ~ ~like
~ s severe depression, substance al~use,or schizophrenia, since appropriate
treatment can do much to mitigate symptoms and improve
parents' functioning. Tliird, although the assessment of parents'
mental health is of critical and obvious impoi-tance when a
parent's emotional difficulties are serious enough to necessitate
the involvemen1 of child protective services, in other circumstances (i.e., when a parent's emotional difficulties would not
lead to unwanted legal intervention in a two-parent family) such
assessment seems to us to be mrrely a search for a "tie breaker"
under a vague custody rule fraught with problems. Once again,
our view is that it is better to change an impossible rule than to
do one's best to follow it.
In summary we conclude that, as others have suggested
(Herman et al., 1997; Otto et al., 2003), a parental diagnosis is
not, in and o l itself, the prima~yconcern when deciding custody;
rather, what is of utmost impor-tance is the impact of parental
psychological functioning on the child's development and hehavior. When a parent's emotional prol~lemsare sufficiently
Critical Assessmeilt of (.hild Custodv Evaluations
severe that they would warrant legal intervention independent of
a custody dispute, we have no doubt that parental mental health
shoultl be a central consideration in custody cases. In more
ordinary c,ircumstances. however, we see 110 ol~viousreason why
a history of parental depression, for example, should be a determinative factor in a custody dispute unless it clearly and
substantially interferes with parenting.
Econ,omic Well-Being
A family's standxd of living falls after separation and divorce, if
for 110 other reason than it is more expensive to live in two
households than to live in one. We should note. however, that the
average decline is greater for divorced women than for divorced
men, as women typically have lower incomes aiid the extra expense of chiltirearing (Duncan & Hoffman, 1985). Ecotlomic
strains can set into lnotion a number of cliaiiges for children,
including possibly movirlg from the family home, changing
schools, losing contact with old friends, and spending more time
in childcare and having less contact with pareuts as the parents
work to make ends meet. Not surprisingly, research shows that
economic stability is an important predictor of postdivorce child
functioning (Dunn, 2004; Lamb, Stelllberg, &- 'rhompson, 1997).
The differences found between thr adjustment of children in
married and single-parent families are reduced by about half for
academic measures like school attainment and by a lesser amount
for internalizing and externalizing problems when incoine is
statistically controlled for (Brooks-Gunn & Duncan, 1997; King,
1994; McLanahan. 1997; McLanahan & Sandefur; 1994).
But while family income no doubt is important, much of the
variance in children's psychological adjustment in divorced and
married families is not explained by economics. Moreover, income may exert its effects indirectly, for exalnple by influencing
parenting and other aspects 01family fi~nctioning,rather than
directly, for example by affecting living conditions and opporlunities available to children. Results of one study indicatrd. for
example, that divorced working mothers, but not married
working mothers, pro\-ided less cognitive and social stimulation
to their children than married nonworking mothers did (MacKinnon, Brody, &- Stonenian, 1982). and other research indicates that parents under economic: stress are less likely to be
supportive (Thomson et al., 1994).
These findings suggest that caution should he exercised when
using parents' incotlies as a predictor of children's well-being
following divorce but we par-titularly call attention to a more
basic issue. The suggestion that custody should go to the parent
with the higher income soulids outlandish and biased; but we
could, if we chose, muster ~lrgumentsthat living with the higherincome parent might be in a child's best interests in terms both of
the correlates of greater wealth (e.g., health, well-being) and the
direct benefits of greater wealth (e.g., living conditions, opportunities). We would not want to make such arguments too seriously, but we do believe h e y illustrate an important point: Why
should parents' relative mental health, parenting skills. or any
other factor determined on a case-by-case basis determine
custody? We believe that the essential problem of determining
children's "best iliterests" based on criteria that are only vaguely
specified is the same whether evaluators consider children's
economic or psychological best interests. In the latter case, the
difference is tlial he core problem is more efkctively disguised.
Ron,kin,g Pretiictors
Based on an extensive review of the literature, one of us (Emery,
1999h) concluded that the following four factors were the most
consistent pretlictors of children's positive psychological adjustment following separation and divorce:
A good relationship with an authoritative residential parent
Minimal or controlled parental conflict that does no1 involve
the children
Economic security
A good relationship with an authoritative nonr~siderltialparent
Our present review is consistent with this earlier conclusion,
and also with the suggestion (Emery, 1999h) that the four factors
are ranked in their order of importance (defined as proportion of
variance explained) for various measures of children's psychological well-being. Given this co~lclusion,we urge any profrssional intervening with separating and divorcing families to
attempt to promote all four goals. Since this cannot always be
accomplished, howevel; our rank ordering indicates that factors
ranked higher should take precedence over factors ranked
lower-if, that is, the objective is to minimizr childrell's risk for
developing psychological problems. This means, for example,
that if parrntal conflict is high, and if thenature of that conflict is
such that it harins children (e.g., revolves around issues of
childrearing, involves the children in the parelits' disputes) then
frequent contact with bolh parents is likely to be more harmful
than beneficial to children. In the face of high conflict, therefore,
children would do better living primarily in one household with
an authoritative mother or father and having more limited contact H ith the other parent. Even as we reach this conclusion, we
recognize that philosophiral or legal considerations might place
a higher value on goals other than maximizing chiltiren's niental
health-for example. [he value that children should have frequent contact with both of their parents despite the presence of
damaging conflict. Wk recognize that a degree of conflict between former partners, sollietimes intense conflict, can be expected in divorce, but that c.onflict also can be contained,
diminished, and hopefully resolvetl over time.
A Referral f o r t h e Deer-Does
Jane Deer-Doe was frightened and infuriated when she unexpectedly found Isahella knocking on her door a day early, after her father
had returned her in a fu17.Jane was more angry than ~volriedaljout
Isabellak flood of bitter tears. Jn the face of Dr. Hagan's adverse
cus~odyrecornlnendation, she thought this was her opportunity to
Ro1)rr.tE. Emery, Rand! K. Otto, Will~am'l'.O.L)onoh~~c
turn he tables on John. Secretly, she also hoped for vindication riot
only lor all of her actions since her separation but also for the
choicrs, mistakes, and sacrifices she had made in nlar~iage.Enraged ant1 not wanting to waste a moment-and wit11 Isabclla lise rhis home and tried to tell
tening ill-Jane telephoned her l a ~ v ~at
him about what happened and about her outrage. But he was ahrupt
wit11 Janp and suggested that she instead come by his office on
Monday morning. There was nothing to be done on a Satuvtlay night.
To Jane's surprise, her lawyer did not launch into a case against
John, even when she finally relaled all of the details in his office.
Ile listened patiently, bla told Jane he needed to give hera "reality
check" about what the courts could and could not do. He talked
about the cost of extendir~gthe liiigation process, delays in hearing
dates, legal counter-tactics like bringing up any and all of her
vulnerabilities as a parent ar~das a person (and her husband
certainly knew her weak spots), and hol~children can get caught in
the mitldle of such contests. He pointed out that nu court was going
to deny John all of' his rights as a father, so she was going to have to
deal with him one way or mother. He also noted that local court
rules miandated that parents attempt mediation before a custotly
hearing cvuld be held.
Jane's lawyer told her that he wanted her to tiy medialion to see
if' she ar~dJohn niigh~worli oul at least some issues a1)out their
chilrlren without going to couit. He described ho~vmediatiun
works and offered that, even if it failed, her effort would look good
if the case did go to court. Jane's lawyer eventually told her lhat he
had, in fact, already spoken with John's lawyer and that she agreed
that they should try mediation. John's laxyer had proniised she
would convince John to ~ r it.
y After raising a nuntber of objections
to the idea, Jane eventually accepted he]. lawyer's advice-hut
only wilh great reluctance ant1 trepidation.
litigation by making judges' decisions unpredictable; (h) increases aclimon>, because virtuall> any evidence \hat makes
one parent loolc bad may be deemed relevant (lecdll the morality
statutes found in some state laws); (c) increases the potential for
bias in the exercise of judicial discretion; and id) limits appellate review, because the guidelines governing judicial decision
making are unc:lear (Garrison, 1996; Mnookin, 1975). In fact,
the problems with the best-interests standard have led at least
one distinguished legal commentator to propose a fair and
simple alternative: Flip a coin (Chambers, 1984). This flip
suggestion highlights the extent of the problems that lay hidden
underneath the best-interests standard's superficial appeal.
Flistorical Perspective
Until the middle of the 19th century. custody laws wel-e perfectly
clear: Fathers were autoniaiically granted custody of their
children, who were viewed, like a wife, as a man's property
(W!-er, Gaylord, 8; G r o ~ e ,1987). L.aws began to change in the
late ISOOs with the emergence of' the "tender years" doclrine,
which held that mothers are uniquely suited to rear children (Ex
Parte Denine, 1981; Lyman &- Roherts, 1985; Mason, 1994;
Wyer et al., 1987). The tender-years doctrine came to control
custody decision making during much of'the 20th century, but in
the 1970s he presumption was challenged as sexist (Hall,
Pulver, 8; Cooley, 1996; Mason, 1994). The subsequent decline
of the tender-years presumption left courts without clear guidance in following the best-interests standard, a principle that
had been place since the beginning of the 20th century
(Mnookin, 1975). For decades, children were auton~atically
Research shows what the Deer-Does' lawyers intuitively recogplaced with their mothers in their best interests (unless the
nized: The process of change, the quality of family relationships,
mother was "unfit"), but the desire to avoid sexism left courts
and the management of coilHict are more ilnpoi-tant to children's
without a dominant guiding principle.
psychological acljustment to divorce than are the structure of
As we noted earlier; some states today list factors that they deem
custody armngements or, indeed, the structure of the fanlily
relevant to children's best interests. at least in general terms, hut
(Ahrons, 1998; Amato & Booth, 1997; Buchanan uiet al., 1996;
the ultimate goal is never defined (Mnoolcin, 1975). This presents
Emery, 1999b; Hetheringtoil &- Kelly. 2002). This conclusion
judges with a n impossible practical, legal, and ethical dilemma.
creates a problem for lawyers in traditional practice? ho~vever,
As noted family law professor Robert Mnookin (1975) put it:
l~ecausethe adversary system on which our legal procedures are
based can exacerbate rather than help to contain parental conflict
Deciding what is hest for a child poses a questior~no less ultimate
and can further undermine rather than promote coordinated
than the purposes and values uf life itself. Should thc judge be
primarily concerned with the child's happiness? Or with the
coparenting. The dilemma for l a ~ e r and
s other professionals who
child's spiritual and religious training? Should the judge be coriwork with custody disputes k particularly vexing under the regime
cemed
with the econornic .'productivityn of thr child I\-hen he
of the vague children's-best-interests standard. We bllrfly evaluate
grows
LIP?
Are the primary values of life in warm interpersonal
this custody standard in historical context before turning to our
relationships, or in discipline and self-sacrifice? 1s stability arid
specific recommendations for refo~m.
security fur a child more desirable than intellectual stimulation?
These questio~iscould be elaborated endlessly. And yet, where is
CHILDREN'S BEST INTERESTS: A STANDARD WITH
the judge to look for the set of values t h a ~should infolni the choice
NO STANDARDS
of whal is best for the child? (pp. 260-261)
In theory, the "best interests of'the child7'standard gives judges
the flexibility to craft custody decisions that are uniquely appropriate for each individual family. In practice, however, the
standard has been widely criticizecl because it (a) encourages
Custody Evaluations: A Solution t o Judges' Dilemma?
Without clear guidance from the law, judges have turned to
mental health professionals and custody evaluations for help in
Critical Assessincnt of Child Ciistodv Evahiaiions
discerning chiltlren's hest interests (I'eller, rlavidson, Hardin,
& Horo~vitz,1992; Mnooliin, 1975; Shun~an,2002; Qild, 19T6j.
By doing so, the courts have implicitly enlhracetl the value
happinessthat children's psychological well-being-their
comes first and foremost on the list of their best interests. hlternative cxpcrts the courts instead might employ inclllde accountants who have evaluated each parent's al~ilityto provide for
their children economically, Educators who can comn~cilton the
parents' relative commitment to promoting success in school,
religious leadels or philosophers who have assessed the quality of
each parenti moral values and training, or perhaps dieticians
who have evaluated each parent's preference for healthy versus
convenience food. These suggestions map seem outrageous, ].jut
so is the idea that custody should be awarded to a parent who has
an edge over another parent in pl.nrnoting children's psychological
well-being, particularly when the construct is ill defined or undefirled.
We appreciate the terrible dilemma that the vague hestinterests standard creates fhr judges, custody evaluators, and, of
course, parents and children. We also believe that a mental
health professional or other neutral third party or parties may I)e
in a better position than a judge hound by rules of legal procedure to make recommendations ahout custody. However, we
h e l i ~ vi t ~is legally, niorally, and scientifically wrong to make
custody evaluators de facto decision mal<er.~
in custody cases,
~ v h i c lis~ often what l~appensbecause jl~tlges often accept
evaluators' recommenclations. As law professor Daniel Shuman (2002) I-ecerl~lysummarized, "thc rolc of mental health
from
professionals in custody- litigation is k~eingtra~isforn~ed
experl as expert to expert as judge" (p. 160). Sliulnan Ivellt or1 lo
point out:
RECOMMENDATION 1: ENCOURAGE ALTERNATIVE
DISPUTE RESOLUTION AND PRIVATE SETTLEMENT
We believe that the best solutions to the p~oblernspose11I-~yc h ~ l d
custodj disputes and unscientlflc custody evaluations involve
changing the systen~of dispule resolution in ways that encourage
parents to reach their olvn decisions about rearing their children
fnllnwing a separatio~~
(Emery, 199911, 2004). C~ln~iously,
there
will be fewer c:ustotlv evaluations, and fewer cases that judges
must decide, if more parents resolve their differences hv deciding custody arrangements on their own. We also believe that
encouragi~lgp ~ i v a t esettlement is the best way to promote
children's mental health in sepi~rationand divorce. If tlle reEncourage
search-based goals are to contain parental col~flic~,
cooperative coparenting, support 110th parents' authoritative
relationships with the cliildrell, and prescrve econorriic resources. then it seems ~.easonahleto steer clear of sometlling
called .'the adversary system," the method of dispute resolution
embi-aced by-t l ~ eAmerican system of justice (Emery &- Wyer,
1987b). "Going to war'' is not the way to promote peace, certainly not in a dlr~orcetlfamily.
Over the laft two decades, many legal and mental health
professionals, and many divorced ~)arcllts,have come to this
same conclusion. As ail alternative, they urge separated parents
to de~erminetheir own childre~i'sbest interests by grappling
with and working out the difficult issues oS residence and
childrearing themselves. One important reason to (lo this from
the outset of a separation is that parents ultimately must deal
with custody decisions, parenting, and each other on their 01~11.
If a tiegree of cooperation in roparenting is the ultimate goal for
promoting children's best interests, then it seems reasonable lo
hypothesize tllat a Illore coopel-ativc approach like mediation,
for exa~llple,will help parents achieve this outcome better. than
If society ~vielicsto use mental liealth prac~itionersas caaper.tsin
aciversarial negotiatiolls or litigation in the courtroolr~will.
child custody cases, the law dntl science denland rigorous
More cooperative approaches to dispute settlement-those in
so hat c.nl11-i.;
threshold scrutiny of their methods and pr.ocrcl~~r.ri
wllich
parents exercise a greater degree of control ovel- both the
are iriformed consumers of [his evidence. If society ~vishesto use
process and the outcome than they tlo in the adversary systenllr~entalIlealth practi~ioneraas judges in child custody cases, theri
social policy demancis n puhlic dehate arid legislatire appr.o~~al
of
inclutle a range oS options sut.h as (a) pro .se divorce, in which
this change . . . (p. 162)
parents manage all legal matters on their own wilhout the use of
lawyers; (h) divorce education, usually involving court-manWe agree. Establishing panels of mental health professior~ali datcd classes on parenting in tlivorce that encnul-ape cooperawho would decide custody disputes would be a major procedural tive coparenting, even during settlement negotiations; (6)more
negotiatio~lsbctwecn parents and their
change in the law, perhaps an important onr. Ho~vever.we 11e- informal, cooperati~~e
attorneys,
an
approach
that includes but is not limited to collieve that there are simpler and likely more effective changes in
policy that would improve custody decision making for children labora~ivelaw,a nem- option invented by family lawyers ill which
and divorcing families and simultaneously solve Inany of the both attorneys agree to represent their clients only so long as
problems faced by custody evaluators, lawyers, judges, and they negotiate in good faith and settle their tlisputes outside of'
otllei- professionals who now work with custod!. disputes. Our court (Tesler, 2001'); (d) family therapy and parent training,
recommendations include (a) promoting parental self-determi- ~vllich,while not focused on resolving custody disputes, do focus
nation through alternative tlispute ~esolutionand other mcans, on the impnrtance of authoritative parenting anti cooperation in
(h) working to develop and implement clear custody s ~ a n d a r d s ~ coparenting for separated and divorced parents (Mar~inez&
and (c') altering the practicr of current custody evaluations un- Forgatch, 2001; WTolchik et al., 2000); (e) divorrr mediation, tlie
most firmly established of the new approaches, in which parents
der the best-interests standard.
Ilohert E. Emery, Randy Ii. Otto, William T. O'Uonohue
~legotiatea srttlen~entwith the help of a neutral expert, usually a
mental health professional or a lawyer (Emel?;. 1994); and (f) use
i~~
family
f
coordinators, for that subset of Iligh-cclnflict families
that cannot participate in or henefit from any of the previous
options ( e . ~ .Coates
,
et al., 2003).
The Example of Divorce Mediation
Importantly, research shows that sonie of these new approaches
1 1 0 help encourage private settlement, cooperative coparentirig,
and a long-term perspective on childrearing follo~vingseparation and tlivorce. The evidence is strongest for divorce rnediation, which has been stutlied more thoroughly than other legal
interventions in divorce, although there undoubtedly is a need
for more research on all. types of custody-dispute-resol.utit111
ixocedures-perhaps
especially on the adversary settlement
1"-~cessitself (Eeclc & Sales, 2001). 4 few randomized trials and
.
a number of evaluations of large-scale programs have shown the
following: Relative to traditional adversary settlement (attorney
negotiations and formal courtroom litigation), mediation (a)
settles a large percentage of cases otherwise headed for court: (b)
possibly speeds the time involved in reaching a settlement,
saves moiiey, and increases compliance with agreements; (c)
rlearly isicreases party satisfaction with the process of dispute
resolution; and, most importaritly, (d) leads to improved relationsliips belrueen nonresidential parents all(! chil(lrei1. as well
as between the separated or divorced parents themselves (Emer): Sbarra. & Grover, 2005).
One of us has conducted a randomized trial of custody mei 71
diation ~ l n dlitigation, including a 12-year follow-up c ~ the
families in tlie study (Emery et al., 2001). The stutly included
primaril!; young, low-income parents, all of whom co111d be
considered high conflict because they failed to reach a settlement on their own and were recruited into the study at the time
thal they filed a petition for a contested-custody hearing. Participants were randomly assigned at this time to participate either in mediation or in an evaluation t ~ ythe ~ o u r t(adversary
conlrol group). and various tests were conducted to examine
;ell-selection and attiition oler time (neither of M hirh proved to
bias the study's results in any detectable manner). Amon5 the
major finciings of an initial study arid replication (Emery et al.:
2001: Emei?~,M a t t h e ~ ~&
s , ICitzmann, 1994; Emei-y. Matthews,
&- F y e r , 1991; Ealel-y & Wyer, 1987aj were the following:
~~-
Only 1 1 % of cases randomly assigned to mediation appeared
in front of a judge, compared with 72% of cases randonllp
assigned to the adversary-settlement group.
On average, parents repol-ted greater satisfaction with mediation than with aclversary settlement on items assessing
both tlle presunied slrengths of n~rdiation(e.g.; "your feelings were understood") and the presunied strengths of adversary settlement (e.g., "your rights were protecled").
Reports of greater satisfaction were notably stronger for fathers than for mothers, apparenlly as a result of a ceiling
effect: Mothers almost always won in court and therefore
senerally were quite satisfied following adversar!. settlement.
The pattern of results held not only immediately after the
dispute resalutio~lshut also in a 1.5-year follow-up and even
1 2 years later.
Nonresidential parerits who mediated were far more likely to
maintain contact with their children. Thirty percent of nonresidential parents who mediated saw their cliiltlren once a
week or more 1 2 vears aftrr the initial disx~ute.in comuarison
to only 9% ofparents in the adversary group. In the media~ion
group, Lully 54% of nonresidential parents also spolce to their
children on the telephone once a week or morr 12 years later.
in contrast to 13% in the adversary group.
The iricreased contact between parents nec:rssitatetl by
greater non1,esidential parent-child contact did not iricrease
parent conflict; ~.atlier,c:onflict was somewhat lower in the
mediation erouo.
u
Among parents who mediated rather than contirluing with the
legal action over the custody dispute, 1 2 years later the
residential parents reported that lie nonresidential parents
were significantly Inore likely to discuss prohlems with them;
had a greater influence on cliildrearing decisions; and were
more involved in the children's discipline, grooming, moral
training, errands, holidays, significant events, school or
church functions, recreational activities, and vacations.
1
,
L
These studies proviile strong evidence ahout the potential for
tnediation to hriiig about irllproved farnily relationships after
separation and tlivorce: even many years later. Still, while the
study's internal validity is strong. its external validity can he
questioned. The results of various other evaluations of mediation
and adx~ersa1-ysettlement help to support tlie generality of the
findings, but an appropriate degree of caution is suggestetl hy
variation in the quality of mediation in different settings. the
push in s o ~ n ecourt-based mediation programs to "gel agreements'' rather tlian focus on fostering positive pos tdivorce famil!;
relationships, and the general nred for more research (Emery,
Sbarra, &- Grover. 2005).
The limited evidence on other legal and inerital health proce(1u1.e~(e.g., divorce education, parent training) also suggests
that encouraging parents to take the long view ant1 work together
a s parents even in the middle of separation and divorce can
benefit cliildreii, parent--child relationships, and .coparents
(Emery, Valdron, &- ICitzniann, 1999). This is not to suggest that
people should not feel hurt, angry, and bitter in tlie midst of
separation and divorce, but instead that, i i they have children:
former partners who remain parents need to find a way not to act
on their understandably painful rmotions as they renegotiate
their family relationships (Emel?, 1994,2004). Also, despite the
proven henefits, it is important to aclcnowledge that mediation is
not a panacea, and there may he a subsrt of parents for whom
lnetliation is not indicatrd (e.g., families with a history of significant domestic violence).
Critical Assrssn~mtof Child Cuslorly Evaluations
T h e Deer-Does i n Mediation
Neither John nor Jane Deer-Doe entered mediation with a sense o l
optimism, let alone a desire to he: in the same 1.oori1with one anolher. Their mediator. 111. Cyrithia Barnes, who also was a clinical
psychologist and farnil! therapist. was pleasant, calm, and clearly
in control of the meeting. but she ror~ltlnot prevent the Deer-Does
from erupting into an angry argument after onl) ahout 20 minutes.
A tense discussion concerning their disagreeriients about the
children exploded when John accused Jane of using the children to
meet her own, limitless need for attention. Jane shot baclc, "It
wasn't me who hat1 an af'Cair." In an angly, loud voic:e, John was
retorting, "I never \<-oul(lhave had to go outside the marriage if you
. . ."when Dr. Barnes interrupted to ask to speak with each parent
alone.
At first, .lane fumed during her caucus alone with Dr.. Barnes.
hut she found herself in tears within a few minutes. "I just can7
he:lieve 1'111losing my mar~iage,"shesaid, "and now he wants me to
lose nly kids too." She talked about her feelings olloss, grief, fear,
h~rrt.and angel; not about problems with the custody an-angement.
At one point. Jane even confessed that at times she longed to get
her mall-iage back; John had, after all, been a good lather and
husl~antl.Bul this revelation quicltly led Jane baclc Lo John's affair
and the 1)ain it caused her; she was heconling angry again when Dr.:
Barnei: interrupted her.
Dr. Barnes olferetl hat she recognized thal Jane was in great
pain in response to losing h o marly things, ancl that she iieedeci to
grieve. In fact, Dr. Barnes recommentletl a therapist for Jane to
consult in ortier to discuss tliese issues. Yet, Dr. Barnes also
pointed out that the goal of medialion was to preserve and protect
the hest part of Jane's relationship with John-their chiltlret~.She
wanted Jane to think about w;~ysthe!- might l ~ able
e to t l to~ (lo that.
John was far less emotional \\-lien lie met with Dr. Barnes alone.
H e clearly was very fi.ustrated, hut kept saying that all he wanted
was to have time with his cliildren and gel on with his life.
Dr. Barnes ackno~sledgedJohn's feelings, but suggested that
not as
maybe Jane-and
maybe Isahella and Carlos too-were
ready to move on as he was, especially in regard to his new relationship. She also obliquely suggested that John niight want to slow
down his current romantic relationship a bit for his own sake, as
well. Her strong advice to John was to work on taking small but
positive steps forwa~.dnit11 tlie kids, and to focus on first rebuilding
his relationship with tlierii alone before including his new girlfriend in his time ~viththeni.
FTlien the Deer-Dues and the mediator got back together toward
the end of their two hours, Dr. Barnes again acknowledged eve17;one's difficult emotions. but pointed out how mediation was focused on trying to solve problems. She repeated her theme ahout
taking small but positi\.e steps, and to rhe parents' surprise, they
took one by arranging a plan for Carlos and Isabella to spend time
with John for an overnight during the coming weekend. They
agreed on very explicit details, not only lor timing and transportation but also on what to ell the children allout the plan and what
to do if one of thern grew distraught.
Jane and John did not work everythinp out in one medialion
session; hut they did discover aforum where the!; conld ],ring their
conflicts and t1-y to sort them out. Mediation offered them an environment that accepted their p i n f u l emotions ]jut simultaneously
encouraged them to put their clwn feelings on hold aud focus on a
plan lor lheir children. Jaue anri John did not realize it, but (his is
exactly what they needed to (lo in a rnuch higger \ray. i n c~rtlerto
move lorwartl as parents and also as people in the conling nlonths
and years.
RECCIMMENDLTION2: -4DCIPTA CLEAR CUSTCIDY
STANDARD LIKE THE APPROXIMLTIONRULE
Our primary recommentlation is to continue to develop practices
and policies lhat encourage parents to reach their own, hopefully
reasonably amicable decisions about residence and parenting,
even when they are in the midst of separation and divorce. We
1-iew mediation as only one of a range of options designed to
facilitate that goal. Our second recommendation is that state
legislatures move to enact clear guidelines for determilling
custody in cases ~vlierethe parents cannot reach an agreement.
Alfair standaril that results in more prediclahle oulcomes should
reduce the number. of i:ontested custody cases, alter the need for
and nature of custody evaluations, and as a result, we believe,
help to reduce or at least nut exacerbate conflict between separating parents. In short, a clear, determinative custody rule is
likely to serve the children's hest interests in separation and
divorce.
lhere is one proposal for a clear c.ustod>r guideline whose
potential we find particularly hopeful. The "approximation rule"
e
apsuggests that parenting arrangements after d i ~ ~ o r cshould
proximate, as much as is possible, the respecti~ieinvolvement of
the parents in childrearing during marriage (Scoll, 1992). Parents who had equal or near-equal in~iol\-ementduring the marcustody after
riage ~vouldmaintain some form of joint
separation. Parents who divided their chil.drearing roles disproportionately during the marriage also would continue that
arrangement. Parents cvho had agreed to change their roles over
time, or who wanled a different postdivorce custody arrangemenl
for whatever reason, would be encouraged to negotiate their own
arrangements according to the primary, private settlement recommelidation of those who have advocated for the approximation rule.
In our view. the lnost important advantage of the approximation
rule is that it is a clear. drtenninative standard. Parents and their
lawyers would know ~cllatto expect of the courts, and this
l<nowledge ~vouldpl.omote settlement. I n custody disputes that
are nevertheless litigated, the approxiniation rule would sharply
limit the scope of' the legal inquiry, as well as any custody
evaluations that might occur. Rather than assessing children's
future best interests, under the approximation rule jutlges and
custody evaluators would focus on the far clearer and far narrower q ~ ~ e s t i oofneach parent's past involvement in childreali~ig.
No state has implemented the proposed approximation i-tlle,
so tllerr is no evidence on its effectiveness. We note, howel-el;
that tlie American Law Institute (2002), whose model statutes
r 7
Robert E. Emerv. Randv I<. O t t ~ William
.
T. O'Donoliuc
often become the basis for state law, has endorsrd the approximation rule in its proposed reforms of divorce and custody law
(along witli the princ:iple of parent self-detern~ination,consistent
with our iirst recommentlation).
WTr also shoul(1 be clear that our suppoit of the approximation
rule is motivated more by the proljlems created by the ill-defined
nature of the current best-interests standard than by the approximation rule itself. Wi: wuuld he open to any clear and deternlinative rule for c1ec:iding children's best interests, but favor
the approximation rule over its two major rivals: (a) a primarycaretaker parent standard, which would award sole legal and
physical cus~odyto the parent who did most of the childrearinp;
and ihj apresumption in favor ofjoint physical custotly. We view
the approximation rule as a pluralistic hybrid of these two alternatives.
We find the approximation rule appealing because it is a clear
and determinative alternative but not a "one size Gts all" solution. At the same time, vie are aware that he approximation rule
is riot without problems. Parents' invol\~enientchanges over.time
and as children grow older (lor example, fathers' involvement in
childrearing tends to increase). In addition, parents and their
lawyers cel-taiiily would debate circumstances lilie the DeerDoes, in which parents agree that one parent will tenlporarily
become Inore involved in childrearing. We also would not expect
the approximation rule to eiid strategic maneuvering. For example, an unhappily married parent might quit worl; or even get
fired in order to be home with he children-and
have an advantage in a future custody dispute.
WJe (lo not propose solutions to these possible diffivulties, but
again note that the best-interests standard is itself fraught with
problems-some similar, and sorne much bigger, in our view. We
believe that the benefits of a clear rule potentially far outweigh
the costs and that implementing the rule is a social experiment
undertaking. In f a c ~ ?divorce po1ii:y already has
well TV~I-th
witnessed the success of moving from vague to specific guidelines. In the early 1C)80s, the rules governing child support were
unclear, and this uncertainty encouraged conflict and poor enfhrcement. Fecleral legislation used financial incentives to ellcourage states to adopt clear child support guidelines by 1986
(National Institute for Child Support Enforcement, 1986). Despite struggles with initial implementation-and
man!. continuing problems with child support-two decades later, all agree
that the clear guidelines are a vast improvement for families.
legal professionals, and the child-welfare system. We expect the
same outcome ~i-llrnlegislatures finally move to adopt a clear
child custody rule.
RECOMMENDATION 3: LIMIT EXPERT TESTIMONY
AND CLARIFY STANDARDS OF PRACTICE
As long as the best-interests principle remains in place as an
ill-defined standard, our third and final recorninendation is to
utilize existing evidence law, professional ethics cocles, and
practice stantlards to limit the expert testimony of nlenlal health
in child custody cases to the presentation of scientifically supported evidence. Until far stronger scientific
support is forthcoming, this recommendation specifically includes the suggestions (reviewed earlier) to (a) abandon use of
all custotly-specific "tes~s" that purport to nieasure children's
t
about
best interests directly or indirectly. (h) ~ r o h i b i testimony
PAS or any other "syndrome" that laclcs scientific: support, (c)
identify the specific nature and sources of inference based on
unstructured interview and observational assessments, and (d)
apply appropria~ecaution in interpreting established measures
and integrating infornintion across different areas of assessment.
Kules of Evidence
Our recommendation to limit expelt testimony may seem radical. but our proposal simply urges the application of established
rules lor expert testimony to such tes~irnonyin custody cases
(Shuman, 2002). Expert testimony in all legal proceetlings is
guidetl by rules of evidence that identify the circumstances
under which such testimony is appropriate (Ewing, 2003; Shuman. 2000: Shuman 8: Sales, 1998). A liey problem for courts
and legislatures is determining exactly what makes testimony
scientific and expert. Histoi~cally,the testimony of experts was
admitted if it passed a legal test tleveloped by a United States
district court. In F r ~ r!.
e Unitrci States (1923) the ctrurt wrote that
Just when a scientific principle or discovery crosses the line between the experimental and demonstrable stages is diffic~lltto determine. Some\\-herein this twilight zone the evidential fo~.ceof the
principle must he recognized, ancl while courts will go a long way in
admitting expert testimony daducecl from a well recognized scientific principle fur discover!; the thing from which he tieduction is
rnacle must 11e sdficiently estahlislled to have gained general acceptance in the particular Gelcl in which it l~elongs(p 1014,).
The F r y test, however, ha.; been crl~icizedon a number o l
grountls (Shuman, 2000). Some have argued that iL is too conservative and may result in exc~lusionof testimony based on
novel-yet-valid techniques and approaches; others say it is too
liberal and allows for testimony based on techniques that have
gained general acceptance despite being invalid. In Dnubert .u.
M ~ r r r l lDow Pharmaceuticals. Inc. (1993), the United States
Supreme Court ruled that the general-acceptance test developed
in Frye "is not a necessary precondition LO the admissibility of
scientific evidence under the Fetleral Rules of Evidence" (p.
2799). The Court luleti that the trial judge should ensure that the
opinion is based on an "inference or. assertion . . . derived hy the
scientific method" and determine "whether the reasoning or
methodology underlying the testimony is scientifically valid and
. . . whether. that reasoning or methodology can be applied LO the
farts in issuc" (p. 2796).
The Court went on to identify four factors that judges could
enlploy when considering specific testimony, including la) he
Critical 4ssrsnment of Child Cuatodv Evaliiations
"testability" of the theoretical basis for tlie opinion; (I,) the error
rates associated with the approach, if kno~vn;(c) whether the
technique or approach on which the opinion is based has beer1
subjected to peer review; and (d) whether the technique or approach is generally accepted in the relevant scientific comniunity.
The guidance provided by Daubel-t could he usetl to examine
whether the expert opinions offered by mental health prolessionals in custody disputes are science hased, but there is no
evidence indicating that trial judges have actively done this.
Those offering anecdotal accounts or personal impressions,
however, are essentially unanimous in their impression that
evidence offered by experts in custody cases is rare1.y objected to
and even less frequently excluded (Sliuman. 2002). Simila~ly,a
review of appellate cases also suggests that the opinions of
mental health experts are rarely excluded on tlie grounds that
the hasis for tlie expert opinions oflered does riot meet required
scientific standards. Our view is that the l o ~ vscientific standards
for expert testimony again can he traced to the vague best-interests principle and tlie impossible tiilemlna it creales for
judges. For this reason, and because individual trial judges
rarely have the time or the expertise to evaluate the scientific
status of psychological measures, we believe that it is incumbent
upon the mental health professions to develop clear professiollal
standards regarding expert testimony in child custody cases.
Professional S t a n d a r d s a n d Gnidelines
The American Psycliologic:al Association (.4P14; 1994'), Association of Family and Conciliation Courts (AFCC; 1995), and
American Arademj- of Child and Adolesc~nt Psychiatq
iL4ACAP;1997) all have developed guidelines i o professionals
~
condut:~ingcustody evaluations. All of these guidelines recornlnend an assessment of chiltlrens' needs, parents' abilities to
meet these needs, and parents' ahilities to provide for future
needs. The APA and A-ACAP guidelines also identify a number
of factors considered to be integral to child custotlv evaluations,
including assessment of parenting ahilities, assessment of capacity to provide a stable loving Iion~e,identification of inappropriate behavior that negatively influences the cliild (e.g.,
suhstance uselabuse), consideration of parental psycliopathology as it affects parenting ability or the child directly. and
consideration of the child's wishes.
Despite broad agreement about factors that should be assessed. there is little agreement about how to assess them. For
example, the AFCC guidelines (which are currently undergoing
revision) do not provide assessment guidelines, while APA and
AACAP bolh generally advocate a multimethod approach
combining clinical intervie~vs,direct observation, and psychological tests. Guidelines promulgated by AACAP question the
value of psychological testing, while suggesting that collateral
inforniatiorl be obtained from school personnel, healthcare
providers. childcare providers, family, frimds, and other incli-
viduals who may provide information germane to cliild custody
placement. The lack of consensus begs the question: What accounts for the variability in recommendations? We conclude that
much of the variability is the result of a lack of cequisite
Itnowledge. There is not enough scientific evidence (and legal
guidance) about how evaluations should be conducted and about
lvhat type of evaluatiori is most helpful. Accordingly, we urge
professional organizations to develop very clear guidelines
concerning acceptable, scientifically hased practices and what
inferences can appropriately be drawn from them. W k have offerrcl our review of the literature on these measures as a starting
point to these discussions and negotiations.
We also urge professional o~:ganizationsto adopt clear ethical
standards for mental health professionals to follow in custody
evaluations. For example. profkssional organizations have failed
to take a clear stand on principles of practice that are widely
embraced by those with extensive professional experience in the
custody context. We suggest three such ~ r i n c i p l c sare worthy of
becoming standards of practice: Evaluators should
Show preference for evaluations conducted by mutually
agreed-upon or court-appointed experts
Promote settlement and other steps that will facilitate a degree of parenla1 cooperation in childrearing and authoritative
par~nt-child relationships-for example, by providing concrete, private feedback to the parties about the evaluator's
opinion before submitting a final report
Ackno~vledgethat custody is ultimately a legal decision and
tlius avoid offering "expert opinion" on legal matters-such
as who should enjoy primary legal or physical. custody and
considerahie pressure to do
under what conditions-despite
so within the legal system
CONCLUDING COMMENT: A QUESTION OF lTALUES
A clear custody rule-whether
the approximation rule, the
primary-caretaker-parent staildartl, a presumption in favor of
joint physical custotly, or some other law-~vould necessarily
talte a stand on values concerning family life, values that often
are contested in our changing, pluralistic society. Custody laws
once did take a clear and strong stand favoring fathers as
property holders, and latrl; mothers as nurturers. Today, there is
no social consensus about the appropriate falriily roles for men
and womenj and we believe this is one reason why legislatures
ha\-e failed to adopt a clearer and more determinant custody
standard. The "childl.en's best interests" standard seems to
ellibrace a laudable value, tlie well-being of children; yet as we
have seen, the standard actually encourages uncertainty and
parental conflict that is contrary to children's interests.
No matter what the goals or actual effects of tlie best-interests
standard, it is inipossible to sidestep tlie values issue. Reaber
(1982) provides some illustrative examples of key value questions raised by child custotly disputes:
Robert E. Emery. Railtlr I<. Otto, \VilIiam T. O'Donohur
I. Should brothers and sisters he in the custody of the same
2.
3.
4.
5.
6.
7.
parent?
Should ail older child, over age 12, have veto po~verin a
custody dispute between two parents?
Should boys be placed with fathers and daughters with
mothers?
Should young rhildren, under age five, be placed with
mothers?
Should coiltinuitv of residence and school district control
placement?
Sliould children be placed with the parent who does not worb
outside the home or who works the fewest hours and/or the
most convenient hours?
Should children h e placed in the h o n ~ ethat does not have/
will not have a stepparent? (p. 319)
Science cannot answer such value questions. Philosopher of
science Carl Hempel (1965) has argued for the demarcation
between factual issues that science in principle can settle and
value issues tliat it cannot, and it is perhaps nowhere more
iniportant to make tliis distinction than in matters of child
custody. Hempel makes this point using a thought experirnent
involving Laplace's demon-a hypothetical entity who knows all
scientific laws and all initial coi~ditionsand who can perfectly
and instantaneously make all rele\;ant calculations needed to
make an empirical tlecision:
Let us assume, then that facet1 with a moral decision we are able to
call upon the I,aplacean denion as a consultant. What help might
we get from him! Suppose that wr have to choose one of several
alternative courses of action to use, and that \re want to know
ivhirll of these \ye ouUght1 0follolv. The tiernon would thrn he ahlr to
tell us, for any conienlplateti choice. what its consequences would
be for the luture course of the univn.de, down to the most minute
detail, hou~everremote in spare and time. Rut having done this for
each of the alternative courses of action under consideration, the
demon would have completed his task: he ~vouldhave given us all
the information that an ideal science might provide l~nclerthe
circumstances. And yet he would not have resolved o u ~moral
prohlem, for this requires a decision as to which of the several
alternatib~esets of consrcliiences mapped out by he demon as
attainable to us is the best; which of thein we ouglit to bring about.
And the burden of this decisinn ~vouldstill fall upon our slioulders;
i t is we who would have to c.ommit ourselves to an i~nconditional.
judgment of value singling out one of the sets of consequences
as superior to its alternatives. Even Laplace's demon, or the ideal
science he stands for, cannot relieve us of this responsibility. ipp.
88-89)
Iii short, even if all of the relevant empirical relations regarding various child custody options were known, we would
still be left with the value questions of' what outcomes are the
hest. This conclusion gives us a final perspective on our three
sets ol recommendations. Our recommendation favoring alternative dispute resolution and parent self-determination not only
recognizes the psyctlological importance of renegotiating family
relationships for children but embraces the value that, except in
cases of abuse or neglect. parents theinselves should have the
option of determining their children's best interests. Our call for
the enactment of a custody standard such as the approximation
rule that has the potential to produce more predictable outcomes
urges a clear articulation of "family values" a s embodied in the
law. Finally, our ret:onimenclation that mental health professionals limit their role in providing expert testimony i n custody
cases places the value of science above all others in professional
practice.
REFERENCES
Ackernian, M.J.. Pi Ackerman, M.C. j19c17). Custody evalualion practice: A survey of experienced profebsionals (revisited). Profes.sionnl Psycholog3: Re.uearch (17x2'Practice, 28, 137-145.
Aclterman, M.; & Schoendorf, K. (1992).S P E C 1 : Ackerman-Schoelzd o ~ Scales
f
for Porent Eruluation of C ~ ~ s t o dLos
y . ilngeles, C,4:
Western Psychological Services.
Ahrons, C. (1.998).The good cii~~orce.
Next- York: Quill.
-4hrons: C.R., & Miller: R.B. (1993). The effect of postdivorce relationships oil parent involvement: .4 longitudinal analysis. m e r i can .loll,rnnl of Orthopsychintry. 63; 441-450.
Ahrons; C.R.. & Tanner. J.L. (2003). Adult childrerl and their fathers:
Relationship changes 20 years after parental divorce. Frrrnily
Relations: InterrLiscipLinory Journc~,lof Applied Fanlily Studies, 52,
340-351.
Amato. P.K. (2001). Children of divorce in the 1990s: An update of the
A~natoand Keith (1991) meta-analysis. Jolol~rnaLof Family Psyc h o l o ~ )1
)1,5, 355-370.
Anlato, P.R., & Booth. A. (1997). A generation nL risk. Cambridge, MA:
Harvard.
Amato. P.R., & Gilhreth, J. (1909). Nonresident fathers and children's
\\-ell-heing:A nieta-analysis. Jou~.nnlof M(~rriageond the Family,
61,557-573.
ilmato. P.R., 8 Keith, G . (1991). Parental divorce and well-heing
of children: A n~eta-analysis.Ps3-cl~olo~icc~l
BulLetirf. 110, 2646.
Amato, P.R., Loomis, L.S., Pi Booth, A. (1995).Parental divorce, marital
conflict, and offspring well-l,eillg during early adulthood. Social
Forces, 73, 895-91 5.
Amato, P.R., & Rezac. S.J. (1994). Contacl with nonresident parents,
interparental conflict. and children's behavior. Journal ofFamiL).
Issues, 15, 191-207.
.American Academy of C:hild and Adolescent Psychiatr?: (1997). Practice parameters for child custody evaluation. Journal qf the
American Acnrlrmy of Child and Adolescent Psychiatry, 36, 57s68s.
American Educational Research Association, American Psychological
Association, & National Council on hleasuremrnt in Education
(1999). Standards for rducntionc~land psych,ologicnl ~psting(3rd
ed.). Washington, DC: American Eclucational Research Association.
American Law Institute (2002). Princij~lesof th,e law oSfamiI3 dissolu,tion: An,aL?:~isand recommendations. Bender: Nrwark, N J .
;\merican Psychological Association (1994). Guidelines for child
custody rvalualions in divorce proceedings. American Psvchologist. 49. 677-680.
(hitical Assessment of Child Custo(1yEvaluations
Xrditti, J.A. (199.5). Acltel.man-Schoendoif Scale.; for Parent Evaluation of Custody. In J.C. Conole!, & J.C. Irnp;lrt~(Eds.), The twelfth
mental rr~~(~surernent.s
yearbook (pp. 20-22). l>incc~ln.NE: Buros
Iristitute of Mental Meahuremrnls.
Associatiori of Family and Concilialion Ctr~u-ts(1095). !l.locic~lstandards
,for child C I L . F ~ O ~~ J~ , c z I ~ a Matli~on.
i i ~ n . RiI: u t h o r .
Bauserman, R. (2002). Child adjustment in joint-custody \eraus solecuslody arrangements: .I meta-analytic review. .Iorlrrlai c!J'Ftz7nily
Psychology, 16, 91-102.
Beaber, R.J. (1982). Custody quagmire: Some psvchological dilemm~is.
.Journal of Ps!.chic~tr, & ILazo,10, 309-326.
Beck, C.J.A., & Sales. B.D. (2001). Family mediation: Fa,cts, myths,
and future prospf7ct.s. Washington, DC: American Psychological
Association.
Bow, J.N., & Quinnell. F.A. (2001). Psychologists' currenl practices and
procedures in chili1 c:ustody evaluations: Five years after American Psychological Association guidelines. Professional Psychology: Research and Pra,ctice, 32, 261-268.
Bow, J.N., & Quinnell, F.A. (2002). A critical review of child custody
evaluation reports. Family Court Review, 40, 164-1 76.
Bramlett, M.D., Pr hlosher; W.D. (2001). First ma,rriage dissolution,
divorce, and remarriage: United States (Advance Data From Vital
and Health Statistics No. 323). Hyattsville, MU: National Center
for Health Statistics.
Bramlett, M.D., & Masher, W.D. (2002). Cohabitation, ma,rria,ge, divurcc., clnd rerntrrriage in the United States (Vital and Health Statistics. Series 23. No. 22). Hyattsville, MD: National Center Sor
Health Statistics.
Brirklin. B. (1?8?). Perception of Relatior~shi~s
Test. Furlong, PA: Village Puhlishing.
Bricklin, B. (1990a). BricXlin Perceptual Sca,les. Furlong, PA: Village
Publishing.
Bricklin, B. (1990b). Parent Aw~~rf,rless
of Skills Survey. Furlong, PA:
Village Publishing.
Brirklin, B. (1993). Test /lilanuals Supplement #9. Furlong, PA: Village
Publishing.
Bricklin, B., & Elliott, G. (1991). Parent Perception c?f Cl~ildProfile.
Furlong, PA: Village Publishing.
Brodzinsky. D. (1993). On the use and misuse of psyrhological testing
Resrctrch
in child custody evaluations. Professional P.syc~holog~:
and PI-actice,24, 213-219.
Erooks-Gunn, J., & Duncan, G.J. (1997). The elfects of povei.ty on
children. The Future of Child7-en, 7, 55-71.
Buchanan, C.M., Maccohy, E.E., & Dornbusch, S.M. (1990). .-ltlolrscents 4 e 7 - divo7-ce.Cambridge, MA: Harvard.
Bumpass, L., (1984). Children and marital disruption: A replication and
update. Demog.7-aphy,21, 71-82.
Cabrera, IV., 'I'amis-LeMonda, C.S., Bradley, K.H., Hofferth. S.: Lamh,
M.E. (2000). Fatherhood in the twenty-first century. Clzild Development, 71, 127-136.
Carlson, J.N. (1995).Perception of Relationships Test. In J.C. Conoley 8;
J.C. Impara (Eds.), The ttvelJth mental measu7-ements yearhook (pp.
746-747). Iincoln, NE: Buros Institute of Mental Measurements.
Chambers, D.L. (1984). liethinlcing the substantive rules for custody
disputes in divorce. University ofNIichigczn Law Re~,if,~t,.
83, 480569.
Cherlin, A.J., Chase-Lansdale, P.I,., & McKac, C. (1C)?8). Effects of
t life course.
parental divorce on mental health t h n ~ u ~ h o uthe
American Sociological Review, 6.7, 239-249.
Cherlin, A.J., Furstenherg. F.F., Jr., Chase-Lansdale, P.L.. ICiernan,
ICE., Robinrs. P.K.. Ruane Morrison, TI.. 'l'eitler. J.O. (19?1).
Longitudinal studies of effects of divorce on children in Great
Britain and the United States. Science, 252, 1306-1309.
Child Treritis (2002). Charting parenthootl: A stati.stica1 portrait of
fathel-s and motlzel-s i n Americu. Washington, DC:: Child Trends.
Clarke, S.C. (1995). Advance report offinal rlivorce stati.~tics,1989 and
1990. (Monthly Vital Statistics Report, Vol. 43, No. 9, Supplement). Hyattsville, MD: National Center for Health Statis~ics.
Coates, C.A., Jones, W., Bushard, P.; Deutsch. K., Hicks. B., Stahl, P.,
Sullivan, M., Sydlik, B.. R- Wistner, K. (2003). Parenting coordination: lmple~nentationixsurs. Family Court Review, 41, 533564.
Conger. J. (199.5). Percrption-of-Kelation~l~ips
l'est. I r r J.C. Conoley &
J.C. Impara (Etls.)., Tile tlcrlJth rnerlttrl rneclsl~rements yea,rbook
(PI'. 747-71.8). Lincoln, NE: Buros Institute of' 1LIental MeasureinrnLs.
Cummings. E.RI.. & Davirs. F'. (1994). Cllildren crnd mtzri/crl conjict.
New I-ork: Guilford.
Davies, P.T., Harold. G.T.. Goelte-hlorev. h1.C.. & C~unrriings.E.hJl.
(2002). Child el~lotional security and interparental eonllict.
~ W o n o ~ r aof
~ hthe
s ,Yocic>t~
for Rescczrch on Child Developmer~t,67.
.. ...
v11-V111.
Dauhert v. Merrell DOIVPharmaceuticals, Inc., 509 U.S. 579, 113 S. Ct.
2786 (1993).
Deater-Ueclcard, I<., Dodge. Ii..I.; & Bat?<: J.E. (1996). Physical discipline among African .Imerican and European American mothers: Linlcs to children's externalizing behaviors. Developmental
Psychology, .32, 1065-1072.
Donnelly, D., & Finkelhor. D. (1993). Who has joint custody? Class
differences in the drtermination of custody arrangeinen~s.Family
Relations, 42, 57-60.
D'Onofrio, B., T~~rkheimer,
E.; Emery, R., Slutske, W:, Heath, A.,
Maddrn. P., & Martin, N. (in press). 11genetically informed study
of marital instability and its association with offspring psychopathology. .[ourr~trlc!/'Abnormnl Psychology.
Downej-. D.B: Ainslvorth-Darnell,
& Dufur, M.J. (1998). Sex of
parent and childrelis well-being in single-parent households.
.Iourrzal of' Wtzrri(rgetlnd the Famil?; 60, 878-893.
Downey7 D.R.. cP Powell. R. (1993). Do children in single-parent
householtls fare better living with same-sea parents? .Journal of
!Marriage t1ntY the Farnil!. .55,
55-71.
Dnnran. G.J., cCr Hollman. S.D. (1985). Economic consequences of
marital instability. In hl. David & T. Smeeding (Eds.), Horizontal
eyuit); unce7-tainty and ~t'ell-beirlg(pp. 427-469). Chicago: Univerrity of Chicago Press.
Dunn, J. (2004). Annotation: Children's relationships with their noiiresident lathers. .Iou7-nc~lof Child Psyclzolog? and Psjchiatry, 45,
659-671.
Ehrenberg, M.F., Hunter, M.A., & Elterman, M.F. (1C196). Shared
parenting agreements after marital separation: The roles of en]pathy and narcissism. .Journal of Consulting and Clirlical P.71.cholog?; 64, 808-818.
Elrod, L.D., & Speclor, R.G. (2004). A review of the year in family law:
Children's issues remain the focus. Family Law Quar/er-1); 31.
527-575.
Emery, R.E. (1982). Interparental conflict and the children of discol-d
and divorce. Psychological Bulletin, 9.2, 310-330.
Emery, R.E. (1994). Reneg.otiating,family relationships: Divorce, (,hilt/
c u s t o d ~and mediation. New York: Guilford.
Emery, R.E. (1999a). Changing the rules for determining child custody
in divorce cases. Cli7ricnl f's's)chology: Science and Practicr, b.
323-327.
Rohert E. Emrr.y, Randy I<. Otto, William T. O'Donohue
Eme~y,R.E. (1999bj. ~lforriage,tdi~,orce,c111dchildren: arljwtment (2nd Gardner, R.A. (2004.). Co~nmeritar~
on Kelly and John~tori's "The
ed.). 'Thousand Oal;s, CA: Sage.
Alienated Child: A Rrfo~.mula~ion
of Parental Alienation SynEme~y.R.E. (2003). Childe11's voices: Listeni~ig-and deciding-is an
dro~ne."Family Courr R~l'iew,42. 611-621.
adult responsibility. Arizona La~uReview, 4.5, 621-627.
(;arrison, M. (1996). How do judges decide divorce cases'? An empirical
Emery, R.E. (2004). Thr ti-uth obout children and divorce: Ilmling wit11
analysi~of discretionary decision making. North Carolina Laiu
the e~notionsso VIIclnd you,,-c~hildrrncan thrive. New York: Vilting.
Xruiero, 74, 401-552.
Emery, 1i.E. (2005). Parental alienation syndrome: Propo~lentshear the Gordon, R., K Peek, T..A. (1989). T/rr Custody Quotient: Krseclrch
1 ~ 1 r d ~ol'proof.
r1
Family Court Recirr~.43, 8-13.
nlanual. Dallas, TX: Filiningtori Institute.
Gottesman, 1.1. (1991). Schizophr~niagenesis: The origins vj'madness.
Emeiy, R.E., & Forehand, R. (1994). Parenlal tiivorcr and children's
well-l~eing:A roc.us on ~.esilience.Tn K.J. Haggrrty, L. Sher~otl.N.
Nr~vYork: Freeman.
Garrnez!; & M. Rutter (Eds.). Risk anrl Resilie17,cein Childrm (pp. Gourley, E.V., & Stoll~erg.A.L. (2000). An empirical inves~igationol
psychologists' custody evaluation practices. Journc~lqf Diuorcr
64-99). I,ondon: Caml~ridgeU~liversityP~.ess.
and Remarriage. 3.3, 1-29.
Ernel?: R.E., Laumann-Billings, L., %dldmn. M., Sljar~a,D.A., and
I)illori, P. (2001j. Child custody tnediation and litigation: Custoc-15 Creenbrrg, S.. Otto, K.IC., & Long, A. (2003). The utility of psychological
testing in personal injun? evaluation. Assessnient, 10, 4,lO-419.
contact. and co-parenting 12 years after initial dispute resolution.
Journal of Coruultirlg and Cli7zicol Psj-rholog?; 69. 323-332.
Grisso; T. (2003). Evalnc~tingcompeten,cies: Forensic. as.irssmerits and
instrl~mrr7,ts(2nd ed). New York: ICluwerll'lenum.
Emer3~,R.E., Matthelts, S.G.. & l<itzniann, 1C.M. (1904). Chiltl custody
nlediiition and litigation: Parents' satisfactiori and Cu~~ctioning
a (Aych. J.H., & Fincham. F.D. (1990). Marital conflict arid children's
adjustment: A cognitive-co~~textualframework. Psychological
year after settlement. JVZZI-nr~l
of Consulti~zga7zd Clirricczl J's!-Eulbctin, 100, 267-290.
cliologJ, 62, 124-129.
Eniery, R.E., Matlh~ws,S., & Wyer. M.M. (1991). Child custody me- Ilall, A.S., Pulver, C.A., &- Cooley, M.J. (1996) Psychology of the best
interests standard: Fifty state statutes and their theoretical antediaticrn and litigation: Further evitlence of the diffeling views of
cedents. Americtzn .Journal of Furnil) Therap): 24, 171-180.
mothers and fathers. Jour-nal oj' Consulting cind Clinicol PsyHeilbrun, I<., Rogers, R., & Ottu. R.K. (2002j. Forensic assessment:
rholog?: 59, 410-418.
Emery, R.E., Shal-ra, D., c9r Grovel; 1:' (2005). Divorce mediation: ReCurrent status and future directions. Tn J.R.P. Ogloff (Ed.). Taking
search and reflections. Fornily Cortrt Review, 43, 22-37.
psycho log^. and lr~loirzto the truent?.-first ceniury (pp. 119-146).
Emery, R E . , Waldron, M.; 8: I<itz~nann,M. (1999). Delinquent beNc~vYork: ICluwe~~IPlenum.
havior, future clivorce or nonn~aritalc~l~ildl~eariilg,
and extet-nal- Heinze, bl.. 8. Grisbo. T. (1996). Review oI instn~rt~ents
assessing
parenting cornpete~lciesused in child custotly evaluations. Beizing behavior among offspring: A 14-year prospective studv.
Journal vf Famil?- Psychology~1.3, .568-579.
hariural Scirnces and the k ~ r o 14.
, 293-313.
Enie13; K.E., Weintraub; S.. & Neale, J.M. (1982). Ef1ec:ts of marital Hempel; C. (1965). ilsprcts of scie~~tzj'ic
eq)lanct/ion and o/,he7- essay.^ in
the p h i l o ~ o p hqfscirnce.
discortl on 111e school he11avio1-of children with schizophrenic.
~
New J'ork: Collier-Mac:Millan 1,imited.
alfectively disordered, and ~ ~ o r m payrnts.
al
Joi~rrraIoJ'Atiiorma1 Herman, S.P.. Duanne,J.E., Ayrrs, \Ii., Arnold, V.,Benetlek, E., Bellson,
Cllilrl Ps).chology, 10. 215-228.
R.S., Bel.net, KC, Eernstoin, G.A., Bryant, E., Licamele; W..
Eniery$ K.E.. & ITJ~er.M.M. (19873). Chiltl cus~odymetliatioii and litk1r:Clellan. J., &- Shaw. I<. (1997). Practice parameters for child
igation: An experimental evaluatio~iol the experiet~croPparents.
c ~ ~ s t o devaluation.
y
.[onn~al($the Anzer-ican Ac.ad~m?of Child &
Journal oJ'C01~su11Ing~ I I Clinical
I ~
PSI-cholopj;5.5, 179-186.
Atlolrscent Psj.chiatq, 36, 57s-68s.
E m e r ~ :R.E.. 8: h e r ; M.M. (19Z7h). Divorce metliation. An~ericr~~rHe~herington,E.M., & C l i ~ ~ g e n ~ p eW.G.
e l , (1992). Coping with marital
Prj.chologist. $2, 47'2-480.
transitions: A fainily systems perspeclive. Morsographs of thr SoEwing, (2.P. (2003). Expert testi~noriy:Law and practice. In A. (loltlstein
ciet~,/ill-Rescurch i71 Chilci Droelop~nrnt,5712-3), 1-242.
(Etl.), Fol.msic Psycholog)- (pp. 55-66). New Jbrk: Wiley.
Hetherington. E.M., Cox. M., & Cnx, R. (1982). Effects of divorce on
Ex IP~rteDevine. 398 So.2~1686 (Ala.1981).
parents and children. In M. La~nli(Ed.), Nontr-r~rlitionc11,j'amilies
Fauber; R.I,.. Forehand. It., Tho~tlas;A.M.. & Wiersou, M. (1990). A
(pp. 233-288). Hillsdale, N.J.: Erlhautn.
mediational model of the impact of marital conflicl on atlolesc,rnt Hetherington, E.M., & Icelly, J. (2002). For better orfor u)orse: Divor-cr
adjustn~entin intacl and divorced families: The role of disrupted
reconsiderrd. New York: Vv:W. Norton.
parenting. Child I)n~elopm~n/~.
61, 11 12-1 123.
Je~iu~vine,
M.J.. & Cohler. B.J. (1999). Major parenlal pspchopathology
Feller. J.N., Dabidson. H.A.. Hardin, M., & Horowitz, R.M. (1992).
antl child custody. In R.M. Galatzer-Levy & L. ICraus (Eds.), The
Working ruith tllr courts iiz cl~ilrlprotec,tion. Washington, DC: U.S.
scient6c basis of cllild custorly decisions (pp. 285-318). New York:
Ilepaltment of Health and Hunian Services.
Tiley-.
Fields, J. (2003). Chil(lren5 li~!ingarrorigem~~lts
ond characteristics:
Johnston. J . R (1993,).High conflict divorces. The Future oj'Children, 4,
March2UM. (Currenl Population Reports, P20-547). Washington,
165-182.
DC: U.S. Census Flureau.
Johnston, J.R., & ICelly, J.E. (2004). Rejoinder LO Gardner's "CornFolberg, J. (1991). Joint custodj- and s h a r ~ dparrrrting. New York:
menlar) on lCelly antl Johnston's 'The AilienatedChild: A ReforC;uill'ord.
lr~ulatio~lof Parental Alienation Syridrott~e.'" Fornil?- Court
Frye v. United States, 293 F. 1013 (U.C. Cir. 1923).
Reuirlo, 42; 6 2 2 4 2 8 .
Furstenherg, F.F., peter so^^. J.I,., Nord, C.w., 8: Zill, N., (1983). The life Johnston, J.R., ICline, M.. &- Tschann, J. (1989). Ongoing postdivorce
course of children of divorce: Marital disruption antl parental
conflict in families coritesting custody: Effects on chiltlrm of joint
contact. .4merica11Sociologicr~lRevie~r,,.48, 656-(268.
custudy ant1 frequent accesi. tlmerican Journal ofOrthopsychiatry,
Gardner, R.A. (2001). Should c:ourts order PAS children to visit/reside
59.576-592.
with the alienated parent'! A% follow-up study. ilnrericar~Journal
Johnston, J.R., & Roseby, T-. (1997). In the name vfthe child. New 'iorlc:
o/'Fore~~sic
Psycholog); 19; 61-106.
Free Press.
Critical ilssrssment of Child C~~stody
Evaluations
Iceilin, W.G., R- Bloom, L.J. (1986). Custody evaluation practice: A
survey of experienced professionals. Professional Psjcholog3-:
Research and Practice, 17. 338-346.
Icell!: J. (1998). Marital conflict, divorce, ant1 chiltlren'h adjustment. Child
c~ndAdolescent Psychiatric Clinics c?fNorth America, 7, 259-271.
Kelly, J.B.. & Emery, R.E. (2003). Cl~ildren's adjustment folloxving
divorce: Risk and resilience perspectives. Famil!- Relations:
Interrli~ci~~linary
Journul oj"App1ied F a n ~ i 1Strrdies,
~52, 3.52-362.
Kelly, J.B. & Johnston, J.R. (2001). The alienaled child: A reformulation of Parental Alienation Svndrome. Family Court Revieul. 39,
249-266.
Icelly, J.. & Larnh, M.E. (2000). Using child development research to
make appropriate custody and access decisions for young children. Family and Conciliation Cour~sRec'ielti, 38, 297-311.
Icing, V. (1994). Nonresiclent father involvement ant1 child well-being:
Can dads make a difference? Journul oj'Fumil?~lssues,15. 78-96.
Lamb, M.E. (1999). Noncustodial lalhers and their impact on the
children of divorce. In R.A. Thompson & P.R. .inlato (Eds.), The
pu.rtdivorcefa mil^: CI~ildren,parenting,anrl.societj-(pp. 105-125).
Thousand Oaks, CA: Sage.
Lamb, M.E.. Sternberg. I<.J., & Thompsorl. R.A. (1997). The eilects of
divorce and custody arrangements on children's behavior, development. and adjustment. Family and Corlciliation Courts Re7;iew:
35,393404.
Lamk~orn,S., Mounts. N.; Steinberg, L., & Dornbusch, S.M. (1991).
Patterns of competence and adjustment among adolesceilts from
authoritative; authoritarian, indulgent, and neglectful homes.
Child Development; 62, 1049-100.5.
Laumann-Billings, L., 8 Emery, R.E. (2000). Distress among young
adults from divorced faniilies. Journal oj?f't.'r~milyPsycholog?. 14.
671-687.
Lewis, C., & Lamh, M.E. (2003). Fathers' influences on children's development: The evidence froin two-parenl Ianiilies. Europec~n,
Jrjuneal of Psych,ology o f Educu~ion,18; 211-228.
Lyman. R.D., &Roberts, M.C., (1985). Mental health testimony in child
cuslocly litigation. Lez~o& Psych,ology Reuieco, 9 , 15-34.
Maccoby, E.E..& Martin, J.A. (1983). Socialization in the context of
the family: Parent-child interaction. In E.M. Hrtheringlon
(Vol. Ed.), Handbook of child psychology: kjl. 4. Socialization,
personal it^ and Social Decelopmen~ (4th. ed. pp. 1-102). New;
York: Wiley.
Maccoby, E.E., & Mnookin, R.H. (1992). Dividing the chi1d:Social ctnd
legal dilf~rnrnusoj'ct~stody.Camhridgr, MA: Harvard University
Press.
MacKinnon. C.E., Brotly, G.H., & Sloneman, Z. (1982). The effects of
divorce and maternal employment on the home environments of
preschool children. Child Development, 5.3. 1392-1399.
Mason, M.A. (1994). From father's property righ,ts to ch,ildreni rights:
Thc7 histriry oJ'cl~ildcustody in the United States. New York: Columllia University Press.
Martin, J..A., Hamilton, B.E., Sutton, P.D., \;ntur+ S.J., Menaclter. F.,
& Munson. M.L. (2003). Birth,s: Final data for 2002. (National
Vital Statis~icsReports, M)I. 52., No. 10). Hyattsrille, MD: National Center for Health Statistics.
Martinez, C.R., & Forgatch, M.S. (2001). Preventing problems with
boys' noncompliance: ESfects of a parent training intervention for
dirorcing mothers. .lour.nal of Cons~rlting& Clinicr~lPsycholog).
69,416428.
Martinez, C.R.. & Forgatch, M.S. (2002). -1djusting to change: Linking
family structure transitions with parenting and boys' adjustment.
Journal oJ'Fa,mily Psychology, 16, 107-1 17.
McGue, M.,R- Lykken, 11.T. (1002). Genetic influence on risk of divorce. Psychologicnl Science, 6,368-373.
McIntosh, J., Long, C., & Moloney, L. (2004). Child-focus~dand chiltlinclusive mediation: .4 coniparative s ~ u d yof outcomes. .lournal
oJ'Fr~milyStudies, 10, 87-95.
McLanahan. S.S. (1997). Parent ahsence or poverty: R~'hit:limatters
more? In G.J. Duncan & J. Brooks-Gunn (Eds.), Consequences ?f
growing up poor (pp. 35-48). New York: Russell Sage.
McLanahan, S.S., R- Bumpass, L.L. (1988). Intergenerational consequences of family tlisruption. American Journal of Sociology, 94,
130-1.52.
McLanahan. S., & Sandefur, G. (1994). Growing rrp ~citlcu singleparerzt:
What hu7ts; u ~ h ahelps.
~
Cambridge, MA: Harvard University
Press.
Melli, M.S., Brown, P.R., & Cancian. M. (1997). Child custody in a
changing world: A study of postdivorce ar~angementsin Wisconsin. University of Illinois Law Reuiriw, 3, 773-800.
Melton, G.B. (1995). Aclcerman-Schoendolf Scales for Parent Evaluation of Custody. 111J.C. Conoley & J.C. Impara (Eds.), Tl~c,~cc!e(ftlc
mental measurements yeurljook (pp. 22-23). Lincoln, NE: Buros
Institute of Mental Measure~nents.
Meltun. G.B., Petrila, J., Poylhress, N., & Slobogin, C. (1997). P~qychological eualur~~ions
oj'th,e courts: A h,andbooX.for mental health
prujessionals and lr~wyers(2nd ed.). Ne\r York: Cuilford.
Meyer, G. (1997). Thinking clearly about reliability: Mure critical
corrections regarding [he Rorschach Comprehensive System.
Psychological assess men^, 9, 4 9 5 4 9 8 .
Meyel; G. (2001). Introduction to the final special section in he special
series on the u~ilityof the Rorschach for clinical assessment.
Psych,ological Assc..ssment, 13,419-422.
Meyer. D.R., & Garasky, S. (1993). Custodial fathers: Mylhs. realities,
and rhild support policy. .louneal oJ'!Marriage c~ndthe Fc~mil3:55,
7:1-89.
Mnookin, R. (1975). Chiltl custody adjudication: Judicial functions in
the face of indeterminacy. Law c~ndConternpo7-aryProblems. .?Y,
226-293.
National Institute for Child Support Enforcement (1986). History and
,fundamentals of child .~rrpportenjurcernent (2nd etl.). Washington,
DC: U.S. Goveiilment Prinling Office.
Novak, G. (1996). L)e7.elopmental psjrhology: Dynamic s?:ytrms and
behauior c~nalysis.Reno, NV: Context Press.
O'Coniior, T.G., Caspi, A., DeFries, J.C., Plomin, R. (2000). Are associations k~etween parental divorce and childrenk atljustment
genelically mediated? An adopliotl study. Developmental Psycholog3.; 36, 4 2 9 4 3 7 .
Ostrom; B.J., & ICauder, N.B. (1996). E.~arnireingthe ccorkofstc~tecolrrts,
1995: .4 natiorzal perspectiue from the court stc~tis~ics
projec~.F i l liarnsl~urg,VA: National Center for State Courts.
Otto, R.I<.. Buffington-Volluni; J., & Edens, J.F. (2003). Child custody
evaluation: Research and practice. In A . Goltlstein (Ed.), Forensic
psych,ology (pp. 179-208). New Yorlt: Wilev.
Otto, R.K., & Edens, J.F. (2003). Parenting capacity. In T. Grisro (Ed.),
Et.nluating cor~zpetencies: Fo7-ensic assessments and irwtrunler~ts
(pp. 229-308). New York: Iilu~ve~~/Plenum.
Otto. R.l<., Edens, J.F., & Barcus, E. (2000). The use of psychological
testing in child cus~odyevaluations. Family and Con,ciliation
Courts Kel:iru, 38, 312-340.
Peris, T., & Emery, R. (in press). .+Iprospective study ol" the c:onsequences of' marital disruption for adolescents: Pre-disluption
family dynamics ant1 post-disruption adolescent adjustment.
Journal oJ'Clinicul Child c~ndAdolescent Ps).chology.
Ro1)rr.tE. Emery, Rand? I<. Otto. Witliam T. O'Donohur
Posthuma. A. (2003). A new MMPl-2 scale for custody disputes. Uniform Marl.iage and Divorce Act; 9A Unifbrln Laws Annotated, Sec.
Arrrei-ican Journal of Forensic Psycholog).; 21, 51-G4.
316 (1979).
Pruett, M.K., Williams; T.Y., lnsabella, G., & Little. T.D. (2003). U.S. Bureau of (he Census (1992). Marriage, divorce, and remarriage in
Family and legal indicators of child adjustlneni io divorce arriorlg
the 19901. (Current Population Reports, P23-1.80). Kashington,
fanlilies with young children. Jourrzal of Fam,ily Ps),chology, 17,
DC: U.S. Government Printing Office.
169-180.
Wald, M. (1976). State intervention on behalf of "neglected" children: h
Pryor, J. & Rodgers, B. (2001). Children in changing,fi~m,ilies:
Li&e after
search for realistic standards. In kI. Rosenheim (Ed.), P~~rsuirtg
Bos~on,MA: Blackwell Publishers.
,justice.for the child (pp. 246-278'). Chicago: University of Chicago
parental separr~tior~.
Press.
Santrock, J.W[, & W~~rslialt,
R.A. (1979). Father custody and social development in boys and girls. Journal qff'Sor.ia2Issttes, 35,112-13.5.
Walden, B., McGue, M., Iacono, w.G., Burt, S.A.. & Elltins, I. (2004).
Identifying shared environmental coniribu~ioiisio early substar~ce
Schcpard, A.I. (2004). Children, courts, arzd custody: l ~ ~ t e r d i s c i p l i n a ~
models fbrrlivurcingjamilies. New York: Cambridge University Press.
use: The respective roles of peers and parents. Jourrlc~lqf'AbnorSchmidtgall, I<., Icing, A,, Zarslti. J.J., & Cooper, J.E. (2000). The efm,al Psycholog?; 113 , 4~1,O-450.
f e c ~ sof parent conflict on later child development. Jol~rnnl oj' Wallerstein, J.S. (2003). Children of divorce: A society in search
of policy. In M.A. Mason, A. Skolniclt, 6.S.D. Sugarman (Ede.).AIl
L)il?orce a ~ z dRemarriage, 33, 149-157.
our families: 1Vewpolicies for n new c e ~ ~ t u (2nd
r y ed., pp. 66-95).
Schwal-tz, L.L. (2003). A nightmare for Icing Solomon: The new reproductive ~echnologies.Journal of Familg Psychology7 17,229-237.
New York: Oxford University Press.
Scott; E. (1992). Plui-alisrn, paternal preferences, and child custody. E'allers~ein,J.S., Lewis; J., &- Blaltrslee, S. (2000). Unexpected legacy of
divorce: The 25 year landmarli s ~ u d y New
.
York: Hyperiolr Press.
Calijornin ~ L Rez,iew,
W
80, 615-672.
Seltzer; J.A. (1991). Relationship between fathers and children who live Weiner, I. (1996). Some observations on the validity of the Rorschach
Inkblot Method. Psychological Assessment, 8 , 20(>-213.
apart: The father's role after separation. J o ~ ~ r n of
a l Marriage and
R'ellman, M. (1994). Acker-man-Schoendorf Scales for Parent Evaluathe Famil?; 53,79-101.
tion of Custody. In D. Iceysrr & R. Sweetland (Eds.), Test critiques
Shaffer, M.B. (1992). Bricklin Perceptual Scales. In J.C. Conoley S- J.C.
(Vol. 10, pp. 13-19). Austin, TX: Pro-Ed.
1rnpal.a (Eds.), Tlre ~ w e y ~nrerc~rrl
h
rrlrasuremenls yearbook (pp.
118-119). Lincoln, NE: Buros Instilute of Mental Mrasurements. Weiihorn; L. (1987). Psychology andchildcustod\ evalua~ions.Lincoln.
NE: University of Nebraska Press.
Shuman, D.\Y. (2000). Psychiatric and ps~chological ecidence. Minneapolis: West Group.
Wolchik. S.X.. \Yest, S.G., Sandler; I.N.. Tein, J.. Coatsworth, D., LeShuman, D.W. (2002). The role of mental health experts in custody
ngua, I>.,Feiss, L., Anderson, E.R., Greene, S.M.; &Griffin, W.A.
decisions: Science, psycl~ologicaltests, and clinical judgment.
(2000). An esperimental evaluation of theory-based mothel. and
mother-child progranis for children of divorce. Journal ?f ConFamily Law Qirc~rierly,36, 135-162.
Shuman, D.W., & Sales, B. (1998). The impact of Daubert and its
sulling and Cliniccll Ps3-chology. 68, 8483-856.
progeny on the admissibility ol behavioral and social science Rioclod, J.M., Nezrvorski; M.T.. Liliel~feld,S.O., & Garb, H.N. (2003).
evidence. Psychology, Public Policy, r ~ ~ Law,
z d 5, 3-15.
What's wrong with he Rorschnch? Sail Francisco, CA: Jossey-Bass.
Siglr-Kushton, Ti:, & McLanahan, S. (2002).The living arrangements of Wood, J.M., Nezworski, M.T., & Stejsltal, W.J. (1997). The reliability of
new unniarried mothers. Demogrnph3; 39. 4 1 5 4 3 3 .
the Comprehensi~eSystem for the Rorschach: A comnlent on
Solonion. J., S- George, C. (1999). The effects o11 attachment of overMeyer (1997). Ps~chologicalAssessrnrni. Y, 490-494.
night visitation in divorced and separated families. In ]. Soloinon Wyer, M., C:avlord3S.. &- Grove, E. (1987). The legal context of child
& C. George (Eds.), Attachrtlent disorganizatiotz (pp. 243-264).
cus~odyevaluations. In L. Weilllorn (Ed.), Psychology and child
New York: Guilford.
custo&. de~erminc~tions
(pp. 4-22). Lincoln, NE: University of
Spokane Couniy Bar Aissociation (1996). Child-centered residerrtial
Nebraska Press.
schedules. Spokane: WA: Aulhor.
Zill, N.D. (1~988).Behavior. achievement, and health problrn~saniong
Steinberg, L. (2001). We know sonic things: Parent-adolescent relarhildren in stepfamilies: Findings kom a national survey of' child
tionships in retrospect and prospect. Journnl of Research on Adhealth. In E.M. Hetherington S- J.D. Arasteh (Eds.), Impact of
ole~cence,11. 1-19.
dicorce, .~ingleparenting;and stepparenting on children (pp. 325Tesler; P.H. (2001). Collabora~i~ue
law: Achiruing <#icti~~e
resolution i n
368). Hillsdale, NJ: Erlbaum.
divan.? ~ t i i t h olitigation.
~~i
Chicago, 11,: American Bar Association.
Zill, N., Moivison, D.R.. Si Coiro, M.J. (1993). Long-term elfects of
Thornson: E., Hanso~l.H.L., & McLanahan, S.S. (1994). Family st1-u~:parental divorce on parent-child relationships, atljustmcnt, and
ture and child well-being: Ecol~omicresources vs. parental beachievement in young adulthood. Journal of'Family Psychology, 7,
haviors. Social Forces, 7.7: 221-2012,
91-103.