Journal of Family Psycholog Copyright 2000 by the American Psychological Association, Inc.

2000, Vol. 14, No. 2, 220-2 0893-320O/00/$5.O0 DOI: 10.1037//0893-3200.14.2.220

Parental Reports of Coparenting and Observed

Coparenting Behavior During the Toddler Period
James P. McHale Regina Kuersten-Hogan
Clark University Brown University

Allison Lauretti Jeffrey L. Rasmussen

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Clark University Indiana University-Purdue University at

This document is copyrighted by the American Psychological Association or one of its allied publishers.

Indianapolis

Fifty-two married partners played with their 30-month-olds in both dyadic (parent-
child) and whole family contexts and reported on their own coparenting activities
(family integrity-promoting behavior, conflict, disparagement, and reprimand).
Coparenting behavior observed in the whole family context was evaluated for
antagonism, warmth and cooperation, child-adult centeredness, balance of positive
involvement, and management of toddler behavior. Parallel balance and manage-
ment scores were also formed using dyadic session data. Men's reported family
integrity-promoting activities and women's reported conflict and reprimand activ-
ities were reliable correlates of family group process in both bivariate and discrimi-
nant analyses, with links enduring even after controlling for marital quality. Whole
family- and dyad-based estimates of coparenting were altogether unrelated, and
reported coparenting was tied only to behavior in family context, not to family
measures created from dyad-based data.

Recently, the family literature has revealed a gan, & Gilliom, 1998; Katz & Gottman, 1996;
growing interest in the study of coparenting McHale, 1992, 1995). Broadly defined, copa-
dynamics (Abidin & Brunner, 1995; Belsky, renting refers to the quality of coordination
Crnic, & Gable, 1995; Brody, Flor, & Neu- between adults in their roles as parents (Lidz,
baum, 1998; Coiro & Emery, 1998; Cowan, 1963; S. Minuchin, 1974). Observational stud-
Cowan, Schulz, & Heming, 1994; Floyd, Costi- ies have typically assessed coparenting within
the context of family group interaction. This has
proven a sound strategy; coparenting varies
James P. McHale and Allison Lauretti, Frances L. along fundamental dimensions of cooperation
Hiatt School of Psychology, Clark University; Re- and antagonism (Belsky et aL, 1995; McHale,
gina Kuersten-Hogan, Bradley Research Center, 1995), and family group interactions provide
Brown University; Jeffrey L. Rasmussen, Depart- ample opportunity to observe whether parents
ment of Psychology, Indiana University-Purdue Uni- support or oppose one another's interventions
versity at Indianapolis.
with children. Group interactions also reveal
Preparation of this article and the work described other aspects of coparenting, including levels of
herein were supported by National Institutes of
Health Grant MH54250 and grants from the Frances involvement by the two parents, and of structure
L. Hiatt School of Psychology. A preliminary version and limit-setting provided for the child. To date,
of this article was presented at the Society for Re- however, studies have not examined whether
search in Child Development in Atlanta, Georgia, the patterning of mothers' and fathers' behavior
April 1997. within family group contexts deviates from the
Correspondence concerning this article should be patterning of their behavior within dyadic,
addressed to James P. McHale, Frances L. Hiatt parent-child contexts. Surprisingly few data ex-
School of Psychology, Clark University, 950 Main
Street, Worcester, Massachusetts 01610. Electronic ist to test the widely accepted family systems
mail may be sent to jmchale@clarku.edu. premise that behavior patterns within the family

220
 REPORTED AND OBSERVED COPARENTING 221

group are not readily estimated from parallel this study, highlighting key conceptual and
information about the family's component sub- methodological concerns that shaped this work.
systems.
There are other gaps in this emerging litera- Assessing Coparenting During
ture on coparenting. Despite the fact that copa- Toddlerhood: Conceptual and
renting alliances rely largely on the individual Methodological Considerations
conduct of the two parents, little is known about
parents' perceptions of their own coparenting Why Study Coparenting During the
behavior. Parents are likely to be aware of the Toddler Years?
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

things they do to influence coparenting solidar-

This document is copyrighted by the American Psychological Association or one of its allied publishers.

ity and of their family* s typical group dynamics. The toddler years are an important develop-
Most parents report making critical, as well as mental period for both child and family. Father
supportive, remarks about their partners to their involvement deepens relative to the infancy
young children (McHale, 1997), and few would years, and coherent, stable family patterns begin
find it surprising to learn that how they parent to crystallize. Coparenting and limit-setting in-
when trying to coordinate with their partner crease in importance, shadowing children's
differs from how they parent when alone with growing independence and willfulness. For cou-
the child. But are there ties between what par- ples, postbaby declines in marital satisfaction
ents say they do to support or oppose their finally plateau for some while continuing for
coparenting partner and the actual patterns of others (Cowan, 1988). In short, coparental co-
supporting, interfering with, matching, or ac- ordination is particularly important during the
ceding witnessed when the family is together as toddler years to satisfactorily meet every family
a group? Establishing relative degrees of over- member's needs, with coparenting solidarity
lap between parents' views of their own and family cohesion being prominent issues.
coparenting-related activities and enacted copa- In previous coparenting and family group
renting process when children are small would research (e.g., Belsky et al., 1995; Brody et al.,
enhance our understanding of early family dy- 1998; Cowan et al., 1994; Gordis, Margolin, &
namics, itself an important endeavor given re- John, 1997; Lindahl & Malik, 1999; McHale,
cent evidence tying both cohesive and compet- 1995), family cohesion and interparental antag-
itive coparenting behavior during infancy and onism have been the major dimensions studied.
toddlerhood to subsequent child adjustment Both are important during the toddler years,
(Belsky, Putnam, & Crnic, 1996; McHale & along with other key dimensions. Three—the
Rasmussen, 1998). child-centeredness of family interactions, the
extent to which parents structure and control
In this study we addressed three issues. First, their child's behavior, and the relative balance
we examined whether information about moth- of positive engagement by the two parents with
ers' and fathers* parenting conduct when alone the child—are examined in this article. Two-
with their toddlers leads to similar inferences year-olds are good at vocalizing their wishes,
about coparental coordination in the family as and in some families, parents build family in-
does parallel data culled from observations of teractions around these wishes, whereas in oth-
family group interaction. Second, we traced ers, parents dictate the tempo and contour of
linkages between mothers' and fathers* reports family activity. Toddlers also tend to shun pa-
of their own coparenting behaviors and concep- rental directives for other pursuits that may con-
tually related coparenting processes in the fam- flict with such directives; in some families, par-
ily group context. Last, we asked whether moth- ents allow the toddler lots of freedom and
ers' and fathers' reports of their coparental "rope," whereas in others, parents keep a firm
activities can be used in tandem to distinguish reign and regularly counteract children's trans-
distinctive types of coparenting and family gressions. Finally, families of 2-year-olds vary
group patterns. Below, we present findings from in the extent to which both parents participate
a new cross-sectional study of coparenting in equally, positively, and fully in the family
two-parent families with a 30-month-old. Be- group process. Balanced participation, requiring
fore doing so, we provide an overview of the that both parents accommodate one another's
coparenting and family processes examined in involvement with the child, is not easily
 222 McHALE, KUERSTEN-HOGAN, LAURETTI, AND RASMUSSEN

achieved in families experiencing distress (Katz related aspects of mother-father-child com-

& Gottman, 1996; McHale, 1995). Given the merce during family group interaction. More-
challenges of coparenting toddlers, it is impor- over, assuming that parents are attuned to the
tant to evaluate multiple characteristics of the organizational structure of the family group dy-
family group process. namic itself, self-reported coparenting conduct
should show ties to behavior within the inter-
Attending to the Context of Family personal structure of the family group setting,
Assessment but not to constructed family measures based on
interactive styles in dyadic settings.
Although studying coparenting patterns in
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Of the many coparenting activities parents

This document is copyrighted by the American Psychological Association or one of its allied publishers.

the family group context has been productive, could be asked about, behavior conveying in-
we do not know whether the same impression of formation about coparental and family solidar-
"joint parenting" in families would result were ity may be most salient in families with young
we to consider mothers1 and fathers* parenting children, affecting the child's sense of security
behavior when alone with their child (Parke, in the family (Cummings & Davies, 1996;
1988). Adults often behave very differently in McHale, 1997). Such information is transmitted
dyadic and family group contexts (Buhrmeister, by parent-parent exchanges that include and
Camparo, Christensen, Gonzalez, & Hinshaw, involve the child and in remarks the parents
1992; Clarke-Stewart, 1978; Deal, Hagan, Bass, make about one another to the child. The value
Hetherington, & Clingempeel, 1999; Gjerde, of asking parents about their coparenting con-
1986; Lauretti & McHale, 1997). For example, duct is seen in studies of preschoolers' socio-
parents who track children's behavior closely emotional development (Jouriles et al., 1991;
when parenting together may characteristically McHale & Rasmussen, 1998; Vaughn, Block, &
be more lax and permissive when alone with the Block, 1989), with hints in the literature mat
child. As another example, a parent who ac- men's and women's self-reports may reveal dif-
cedes to a partner during family group interac- ferent facets of the family process (Feldman,
tion, creating an impression of imbalance in Wentzel, & Gehring, 1989; McHale & Rasmus-
mother-father engagement with the child, may sen, 1998). For example, in McHale and Ras-
engage fully during time alone with the child— mussen1 s work, men reporting greater attention
leading to a completely different impression of to family integrity-promoting activities (attend-
family balance. To date, however, no published ing to both partners and children, talking posi-
study has examined the comparability of "fam- tively about the mother when she is away) and
ily" data gathered across both dyadic and family women reporting that they do not denigrate the
contexts. child's father (undo his disciplinary efforts,
speak disparagingly about him to the child) had
Linking Parental Reports of Coparenting children rated as better adjusted than other pre-
to Observed Coparenting Process schoolers. The authors proposed that fathers'
proactive family engagement and mothers'
If indicators of coparenting "constructed" by stance in sanctioning father's participation in
pairing mother-child (dyadic) data with father- family life may be particularly important indi-
child (dyadic) data are not reconcilable with ces of coparental and family process (cf. Allen
indicators of coparenting estimated from family & Hawkins, 1999; Westerman & Massoff,
group interaction, how should parents' reports 1995).
of their own coparenting conduct be expected to
fit into this puzzle? If, as is argued here, family These findings substantiate that there are in-
group process does indeed represent a unique deed other coparenting processes, beyond child-
"reality" (P. Minuchin, 1985) cocreated by both rearing conflict, worthy of study. Ttie finding
coparenting partners and possessing a reason- that fathers' investment in promoting family
ably robust and stable set of defining features, integrity and mothers* propensity to denigrate
then parents' reports of their own coparenting the father are particularly strong correlates of
activities (behaviors that simultaneously take preschooler adjustment seems important. De-
into account both the coparenting partner and spite changes in family life over the past 30
the child) should show at least modest ties to years, most studies still place mothers at the
 REPORTED AND OBSERVED COPARENTING 223

family's "hub," with fathers' patterns of partic- Summary and Prospectus

ipation and involvement in family life trigger-
ing various patterns of family adaptation and The considerations summarized above
distress when children are young. Hence, it may prompted three research questions. First, is
not be surprising that children's behavior prob- there concordance between family indices
lems are greatest when their fathers report rel- based on data from whole family and from
atively fewer behaviors that support family in- parent-child dyadic contexts? Second, are par-
tegrity and their mothers are denigrating of their ents' perceptions of their own coparenting con-
fathers. To date, however, no study has asked duct related to observed coparenting and
family-level dynamics? Informed by prior
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

whether parents* self-reported coparenting ac-

work, we predicted links tying men's reports of
This document is copyrighted by the American Psychological Association or one of its allied publishers.

tivities can be linked to conceptually related

family integrity-promoting behavior to obser-
patterns of coparental and family behavior. The
vational measures of cohesiveness and tying
current study addresses this issue.
women's reports of interparental antagonism
(disparagement, conflict) to observational mea-
Exploring Family Profiles sures of divisiveness. We expected self-reports
to correlate with behavior in whole family con-
Studies of coparenting and family process texts, but not parent-child dyadic contexts. And
typically take univariate views of the family, third, can distinctive family types, varying
describing certain coparenting behaviors and along dimensions of directedness-permissive-
ness, warmth and cooperation, and antagonism,
then seeking predictors or correlates of these
be discerned—and, if so, are they meaningfully
indices. Although this approach is useful in
distinguished by either marital quality or by the
establishing the validity of certain family con-
combined reports of mothers and fathers regard-
structs, it misrepresents the complexity of fam- ing coparental conduct?
ily interaction patterns, best described along
multiple dimensions simultaneously (Epstein &
Bishop, 1981). Filsinger (1990) advocated seek- Method
ing out meaningful similarities among subsets
of families within samples, grouping those shar- Participants
ing a common profile across a number of vari- Participants were 52 families (with two parents, a
ables. Such typologies can be constructed from 30-month-old son or daughter, and any siblings). Of
cross-tabulations of small numbers of variables 31 toddler boys, 20 were firstborn. Of 21 toddler
(e.g., Olson, Sprenkle, & Russell, 1979) or girls, 13 were firstborn. Families were recruited
through some form of cluster analysis (e.g., through media announcements and mailings to day-
Moos & Moos, 1976). This strategy enables care centers. The sample was largely middle to upper
determinations of whether distinctive family middle class; 70% of mothers and 73% of fathers had
attained some postsecondary degree, with a mean
types can be distinguished by individual "pre-
1997 family income of $51,020. Eighteen percent of
dictor" variables (such as marital quality) or by participants were African American, Asian Ameri-
predictor sets containing data from multiple in- can, or Hispanic, and 82% were Caucasian. Partici-
formants or sources. Typological approaches pant families received a small payment for their
have proven useful in the parenting literature, in contribution to the study.
which Baumrind's (1971) parenting dimensions
have had considerable heuristic value. Though Procedure
it is premature to posit with conviction parallel
superordinate, organizing dimensions of copa- Each family contributed both self-report and ob-
renting and family process, it seems that the servational data. Self-report measures were com-
group dynamics of families with toddlers could pleted at home. Observational data were drawn from
meaningfully be characterized by how restric- (a) mother-child dyadic play sessions, (b) father-
tive they are (i.e., degree of parental structure, child dyadic play sessions, (c) a whole family inter-
action session, and (d) a marital interaction task. All
control, and limit-setting) in tandem with infor- four of these assessments took place in a university
mation about cohesion (warmth-cooperation) family study center during two separate visits sched-
and divisiveness (interparental antagonism). uled about a week apart.
 224 McHALE, KUERSTEN-HOGAN, LAURETTI, AND RASMUSSEN

During each parent-child session, the parent spent Family Integrity is a seven-item scale describing
approximately 30 min with the toddler, participating both overt behaviors—for example, "How often
in a series of structured and unstructured play tasks. when you and your partner are together with this
The structured tasks involved interacting together child do you say or do something to promote an
with one of two computer games that posed questions affectionate interchange between your partner and
for the toddler to answer (by pressing the correct this child (e.g., 'show mommy* or 'let datfdy
button on a game board), solving a matching game or play')?"—and covert behaviors—for example, "How
puzzle, and reading together a storybook containing often when you are alone with this child do you make
pictures but no words. These tasks were followed by a comment to enhance his or her mental image of
an unstructured free play period and then by a your absent partner ('daddy loves you'; 'mommy is
cleanup period. proud of you')?"—that help to cultivate a sense of the
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

coparenting team and family unit as a cohesive unit.

This document is copyrighted by the American Psychological Association or one of its allied publishers.

During the family play session, all members of the

family participated in three structured tasks— Scale items measure such activities as demonstrating
interacting with a new computer game similar to the affection toward the partner and toward the child
first two, playing a game of horseshoes, and having a when the family is together as a unit, inviting the
"pretend" family meal involving kitchen and toy food partner to join in parent-child activities, and speak-
props. This was followed by a free play period, ing affirmatively about the partner and the family
during which the family explored a well-stocked toy when alone with the child (for fathers in this sample,
box containing both gender-neutral and gender-typed mean Family Integrity score = 37.85, SD = 5.01,
toys. As in the parent-child sessions, the final part of range = 26-47; for mothers, M = 40.83, SD = 3.78,
the session involved a cleanup period. range — 30-47).
In the marital session, the couple (without the Disparagement is a three-item scale describing
toddler present) engaged in two 5-min discussions of behavior that undermines the partner's credibility or
topics they had independently identified as areas of authority. It measures the frequency of negative com-
the marriage in which they desired change. One topic ments about the coparenting partner to the child in
was an issue for which the husband, but not the wife, private and overturning of the partner's disciplinary
desired change; the other was a topic for which the efforts (for fathers, M = 6.60, SD = 2.03, range =
wife, but not husband, desired change. Order of dis- 3-11; for mothers, M = 6.81, SD = 3.21, range =
cussion of these two topics was counterbalanced (for 3-12).
more details about this procedure, see Sagrestano, Conflict is a two-item scale measuring the fre-
Christensen, & Heavey, 1998). Among the issues quency of mild and intense interparental disputes
couples deliberated were handling family finances, aired in front of the child (for fathers, M — 6.54,
demonstrating affection, and dealing with in-laws; no SD - 2.09, range = 2-10; for mothers, U = 7.02,
topic presented to or discussed by the couple per- SD = 2.18, range = 3-14).
tained to children or parenting. Following each dis- Reprimand, a scale originally named Solidarity, is
cussion, each partner rated satisfaction with progress a four-item scale that measures both the frequency
made during the task. with which the respondent sets limits or seeks the
Parent-child, family, and marital tasks were all setting of limits by the partner and the emphasis the
videotaped, and independent teams of trained gradu- respondent places on the coparenting unit during
ate student coders, blind to all other information conversations with the child. This scale comes closest
about the families, later provided clinical ratings of to assessing the "work of parenting" traditionally
the sessions. emphasized in coparenting instruments (for fathers,
M = 18,35, SD = 2.32, range - 13-25; for mothers,
Measures M = 20.33, SD = 2.18, range - 15-25).
All four Coparenting Scale measures show marked
Self-reports of coparenting behavior. Parents cross-time stability during the preschool years
completed a 16-item Coparenting Scale (McHale, (McHale, Kuersten, & Loding, 1998). The Family
1997), on which respondents are asked to report how Integrity, Disparagement, and Conflict scales are sig-
frequently they engage in several different activities nificantly correlated with marital satisfaction in the
in the child's presence. Each item is rated on a scale expected direction; they are also associated with co-
ranging from 1 {absolutely never) to 7 {almost con- parental support and conflict (assessed by Ahrons,
stantly; 1-2 times an hour). Coparenting-related con- 1981, Quality of Coparenting Scale) and with the
duct in both the family triad ("overt" coparenting Cohesion and Conflict scales of Moos's (1974) Fam-
behaviors) and when alone with the child ("covert" ily Environment Scale (McHale, 1997). Maternal rat-
coparenting behaviors) is queried. The 16 items yield ings of Disparagement and paternal ratings of Family
four derived factors: Family Integrity (a — .82), Integrity have also been linked to early observational
Disparagement (a — .75), Conflict (a = .79), and ratings of family process assessed during infancy and
Reprimand (a = .62; see McHale, 1997). to concurrent measures of preschool adaptation
 REPORTED AND OBSERVED COPARENTTNG 225

among 4-year-olds (McHale & Rasmussen, 1998). In sitivity, which measured the timing and quality of
U.S. samples, scores on the Reprimand scale have not parents' interventions with the child (providing as-
reliably distinguished maritally satisfied from dissat- sistance when it was needed, refraining from inter-
isfied respondents, though higher Reprimand scores fering when it was not); and (e) Investment, which
have been associated with parental reports of greater measured the parent's psychological "presence" and
misbehavior by their preschooler in a sample of ur- engagement in the activities of the play session.
ban, middle-class families in Beijing, China Each dimension was rated along a behaviorally
(McHale, Rao, & Krasnow, 2000). No studies to date anchored scale ranging from 1 (low) to 7 (high). To
have tested for links between Reprimand scores and reduce carry-over bias, raters never evaluated the
observational indices of how parents administer to same parent in both dyadic and family contexts or
their young children. both parents within the whole family context. Reli-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Observed parenting, coparenting, and family be- abilities (intraclass correlations) across raters and
havior. Coparenting and family behavior were contexts were .88, .78, .85, .72, and .81 for Warmth,
evaluated using constructs from the Coparenting and Provision of Structure, Limit-Setting, Sensitivity, and
Family Rating System (CFRS; McHale, Kuersten- Investment, respectively. Means, standard deviations,
Hogan, & Lauretti, 2000). Working from videotaped and intercorrelations for these parenting variables, in
records of the family group sessions, trained coders both dyadic and family contexts, are provided in
rated aspects of parenting and family process known Table 1.
to predict child adjustment among older children. The These five parenting variables were used to con-
set of five CFRS parenting and five CFRS coparent- struct two of the family-level indices, and both com-
ing ratings (described below) was then used to form parison scores. Balance of positive parental engage-
the five major coparenting and family variables of ment during family interaction was operationalized
interest in this study—antagonism, warmth and co- as a parenting discrepancy score. Ratings of the
operation, child/adult-centeredness, balance of posi- mother's Warmth, Provision of Structure, Sensitivity,
tive parental involvement, and management of tod- and Investment during family group interaction were
dler behavior. added together to form a single positive engagement
CFRS parenting ratings were also generated for by mother score. A parallel score was formed for the
the dyadic sessions, enabling a "test" of whether father, and then the absolute difference between these
Coparenting Scale reports are uniquely tied to family two positive engagement scores was calculated. The
group-level dynamics. Two indicators of how parents resulting score reflected disparities in positive en-
behaved during their respective dyadic play sessions gagement by the two parents while interacting to-
with the child were formed. First, an index that gether with their toddler. For the comparison index,
paralleled the score for balance of positive parental balance of positive parental engagement during dy-
involvement during family interaction was formed by adic interaction, we summed the same four CFRS
comparing mothers' and fathers' engagement with ratings scored from dyadic mother-child sessions,
the toddler during their respective dyadic play ses- did the same for fathers, and computed the absolute
sions. Second, an index paralleling the score for difference between these two scores. This yielded the
management of toddler behavior during family inter- comparison index, reflecting discrepancies in posi-
action was formed by combining relevant maternal tive engagement during dyadic interaction. For both
and paternal ratings (see below) taken from the dy- discrepancy measures, larger values indicate more
adic sessions. Because these scores were based on dissimilarity in positive parental engagement.
parents' behavior in the absence of their spouse, they Greater discrepancies in parental engagement are
were not expected to show ties to parents' reports of more characteristic of the whole family interactions
Family Integrity-promoting behavior or of family of maritally distressed couples (Katz & Gottman,
Reprimand activities on the Coparenting Scale. The 1996; McHale, 1995) and of families of preschoolers
CFRS parenting variables used to form the five fam- who show high levels of internalizing symptoms
ily and two comparison scores were (a) Warmth, (McHale & Rasmussen, 1998; McHale, Johnson, &
capturing the parent's positive regard for and affect Sinclair, 1999). Correlates of discrepancies in par-
toward the child (voice tone, praise, facial expres- ents' engagement with children during dyadic
sion, and touch entered into evaluations of parental parent-child interactions have not been similarly
warmth); (b) Provision of Structure, which measured studied.
the extent to which parents organized the tasks for The second pair of indices, reflecting management
children, provided clear instructions, and helped the of toddler behavior in (a) the family and (b) the
child understand the task; (c) Limit-Setting, which respective parent-child dyads, involved the CFRS
measured the extent to which parents prevented the ratings Provision of Structure and Limit-Setting.
child from wandering away from assigned tasks, en- These ratings were summed separately for mothers
gaging in activities not sanctioned by the parent, or and fathers in the family group setting and then added
opposing the parent during cleanup periods; (d) Sen- together. A parallel comparison index was formed by
 226 McHALE, KUERSTEN-HOGAN, LAURETTI, AND RASMUSSEN

11 summing the same two ratings, using data from

dyadic sessions. These two indices reflected overall
amounts of structure and limit-setting provided by
the child's parents.
The remaining three family variables—coparental
antagonism, family warmth and cooperation, and
child- (vs. adult-) centeredness—were all derived
from the five CFRS coparenting and family ratings.
These ratings included the following: (a) Behavioral
Competition, capturing active parental interference in
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

the partner's child-directed activities (e.g., distracting

This document is copyrighted by the American Psychological Association or one of its allied publishers.

the toddler with a novel toy while she was engaged in

a different activity with the other parent; M — 2.19,
SD = 1.21, range = 1-5); (b) Verbal Sparring, mea-
suring sarcasm and hostile remarks; higher scores
entrusts IN reflected frequent and clear-cut use of such remarks
(M = 2.46, SD = 1.36, range - 1-5); (c) Coopera-
tion Between Partners, measuring propensity to work
P-O oo >n O as a team and support and build on one another's
© 1-H 1-^ ^H I-H
interventions with the toddler (M - 3.40, SD - .87,
range = 2-5); (d) Couple Warmth, capturing the
degree of enjoyment and positive affect shared be-
tween partners during the play session (Af = 2.94,
SD = 1.07, range = 1-5); and (e) Child-
Centeredness, measuring the extent to which the ac-
tivities of the session allowed for the interests and
initiatives of the child. Child-Centeredness scores of
1 indicated that the pace and flow of the session was
dictated exclusively by the parents; scores of 5 indi-
cated that the session was regularly guided by the
initiatives of the child. Midrange scores reflected an
equal balance between the interests and initiatives of
parents and child (M = 3.02, SD = 1.02, range =

I
.9
1-5).
Variables were rated along behaviorally anchored
scales ranging from 1 (low) to 5 (high). Behavioral
Competition and Verbal Sparring (r = .30, p < .05)
were standardized and summed to form the variable
I interparental antagonism. A family warmth and co-
2 s <s c* 1 ^ en operation index was created by standardizing and
summing ratings of couple, mother-child, and
father-child Warmth in the family (all of which
ov © o r- correlated .45 or above) to form a variable called total
family warmth and then adding this score to stan-
dardized Cooperation Between Partners scores (r =
»ri oo © m •«*• .58, p < .01). The final family-level variable was
Child-Adult Centeredness.

l! Several specific hypotheses were advanced. Copa-

renting Scale indices reflecting commitment to pro-
moting a strong sense of family (Family Integrity)
were expected to correlate with observational mea-
sures reflecting cohesiveness, mutuality, and inter-
personal attunement (high warmth and cooperation,
similar levels of positive engagement with the child,
and a child-focused family orientation). Coparenting
Scale indices reflecting interparental antagonism
(Conflict, Disparagement) were expected to correlate
with observational measures signifying divisiveness
 REPORTED AND OBSERVED COPARENTING 227

(high interpaiental antagonism, low family warmth cussions, both partners completed a series of Likert-
and cooperation). And parents* reports of more fre- type ratings indicating on a scale ranging from 1-9
quent limit-setting and disciplinary activities by the how they thought the discussion had gone. Included
coparental team (higher Reprimand scores) were ex- in these ratings was a global score indicating satis-
pected to correlate with more noteworthy manage- faction with the discussion. Husband and wife satis-
ment of toddler behavior (high structure during fam- faction scores for each of the two discussions were
ily tasks, quicker intervention when the toddler was standardized and summed to form a second compos-
defiant or strayed off task) during the family sessions. ite.
Marital quality. Because we wanted to provide a 3. Partner ratings of marital satisfaction. Both part-
strong test of linkages between the four Coparenting ners completed the Locke-Wallace (1959) Marital
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Scale indices and their hypothesized behavioral ref- Adjustment Test, a widely used instrument that has
This document is copyrighted by the American Psychological Association or one of its allied publishers.

erents in family interaction, we controlled for the been validated in the literature (e.g., Gottman, Mark-
effects of marital quality. The tenor of the marital man, & Notarius, 1977). In this sample, the mean
relationship has been linked to both reported and
marital satisfaction score for men was 110.1 (SD —
observed coparenting process in prior work (e.g.,
22.2, range - 65-161), and for women it was 112.9
Belsky et al, 1995; Dickstein & Parke, 1988; Katz &
(SD = 24.4, range - 32-146). Eighteen of the 52
Gottman, 1996; Lindahl, Clements, & Markman,
1997; McHale, 1995,1997). However, in most stud- families (35%) contained at least one partner whose
ies, properties of the marital and coparental relation- total satisfaction score fell below 100, placing them
ship are not delineated separately. To devise as pure in the clinically distressed range. Because husband
a measure as possible of marital quality, we used and wife scores on mis instrument were significantly
three separate measures to form a marital composite: correlated with one another (r • .42, p < .01), they
1. Observational ratings of marital discussions. were standardized and summed to form a third com-
Trained graduate student coders watched both posite.
problem-solving discussions and rated marital pro- Principal-components factor analysis of these
cess using the system of Cox, Tresch-Owen, and three variable sets indicated a single factor (eigen-
Lewis (1989). Six variables—warmth, intimacy of value = 1.86) accounting for 62% of the variance.
communication, problem-solving, conflict power, Factor loadings for the clinical observations, partner
and autonomy—were rated along behaviorally an- ratings of the laboratory task, and marital satisfaction
chored scales ranging from 1 (low) to 9 (high). In- indices were .83, .77, and .76, respectively. We then
terrater reliabilities for the last two variables were formed a composite marital quality index by sum-
unacceptably low (.56 and .59), and so these vari- ming standardized scores for the three indices; this
ables were omitted from subsequent analyses. The index showed good internal consistency (a = .75).
remaining four variables were defined as follows: (a) This summary score was used in the partial correla-
warmth referred to the degree of positive affect ex- tions reported below.
pressed between partners during the problem-solving
discussion (M = 5.77, SD •= 2.11, range = 1-9); (b)
intimacy of communication referred to the willing- Gender Difference Analyses
ness of partners to divulge their feelings, to share
sensitive and personal information during the discus- Before testing the hypotheses outlined above, we
sions, and to listen sensitively and empathically to the
conducted two multivariate analyses of variance to
partner (M = 5.40, SD = 2.33, range = 1-9); (c)
examine gender and birth order effects on the pattern
problem-solving described the couple's success in
of relationships among the self-report and observa-
making headway on the issues under discussion,
rather than avoiding the topic or rushing to premature tional measures of coparenting. Of particular interest
closure (M » 4.96, SD = 2.17, range « 2-9); and (d) was the statistic Box's Af, which indicates the degree
conflict ratings entailed an estimation of the degree to of similarity in the variance-covariance matrices for
which hostile, cutting, or sarcastic remarks interfered boys and girls and for firstborns and later boras
with the couple's progress in attending to the goals of (Tabachnik & Fidell, 1996). A statistically significant
the task (M = 5.14, SD = 2.57, range = 1-9). Box's M suggests that the variance-covariance ma-
Reliabilities for these four variables were acceptable trices are different and hence that correlations should
(K = .87, .76, .81, and .83, for warmth, intimacy of be examined separately for boys and girls or for
communication, problem-solving, and conflict, re- firstborns and later borns. However, findings for both
spectively). Scores were standardized (with conflict child gender, Box's U - 119.46, F(91, 6432) =
reverse coded) and then summed to form a composite 0.94, /w, and for birth order, Box's M = 131.84,
score. F(9\, 6432) = 1.02, ns, fell far short of statistical
2. Partner ratings of satisfaction with marital dis- significance, indicating that data for the different
cussions. After each of the two problem-solving dis- groups could be examined together.
 228 McHALE, KUERSTEN-HOGAN, LAURETTI, AND RASMUSSEN

Results three observational measures expected to corre-

late with them. Men reporting more frequent
Three sets of findings are reported below. Family Integrity-promoting efforts belonged to
First, correlations between family indices based families that were more child-centered and
on data obtained from (a) the whole family marked by parenting styles that were more sim-
setting and (b) observations of mother- and ilar than dissimilar. These links persisted even
father-child dyadic interactions are reported. after controlling for effects of marital quality.
Then, associations between parental reports of
Women reporting more Family Integrity-
coparenting behavior and observed family-level
promoting efforts also belonged to more child-
dynamics are described, both before and after
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

centered families, with this link reduced to a

controlling for the effects of marital quality.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Correlations between the parental reports of co- trend once marital quality was taken into ac-
parenting and family measures constructed count. Surprisingly, neither parents' Family In-
from dyadic data are also reported. Finally, a tegrity scores correlated with observed family
cluster analysis of the five observational mea- warmth and cooperation.
sures describing the family group interaction Disparagement. Men's reports of covert
data is described. Three clusters emerging from Disparagement were not associated with either
this analysis are compared along the dimension family warmth and cooperation or with interpa-
of marital quality, and the results of a discrimi- rental antagonism. Women's reports of covert
nant analysis involving the eight maternal and Disparagement were associated with higher lev-
paternal self-report measures as predictors of els of interparental antagonism during family
these clusters are presented. group interaction, though this link dissipated
once marital quality had been taken into ac-
count.
Comparability of Family Measures Based
on Data From Dyadic and Family Group Conflict. Men's reports of Conflict in the
child's presence were not significantly related to
Contexts
The indices of balance of positive parental 1
Because men who reported low levels of Family
engagement based on data from dyadic and Integrity-promoting behavior tended to belong to
from family interactions were clearly not com- families characterized by greater disparities in the
parable. In fact, the two sets of scores were balance of positive parental engagement with the
completely unrelated (r = .00, ns). For the two child, we wondered whether this was a reflection of
composite management of toddler behavior relatively uninvolved fathers who had negative rela-
scores, a modest but significant relationship tionships with their children and so did little to pro-
across contexts did exist (r = .41, p < .01). mote family solidarity. When we looked closely at
the family interactions of the 10 men in the sample
whose Family Integrity scores fell at least one stan-
Reports of Coparenting Behavior and dard deviation below the group mean (scores of 32 or
Observed Family Interactions less), we did not find that it was inevitably the fathers
who were the ones showing less positive engagement
Recall that for each of the four Coparenting with the toddler during family interaction. In fact, in
Scale measures, significant correlations were 6 of the 10 families, it was the fathers who were more
anticipated only with those family variables positively engaged with the child during family in-
conceptually related to the underlying construct teraction than were the mothers. Though this finding
being measured (detailed above). Given the spe- is a little puzzling, recall that Family Integrity scores
cific, a priori hypotheses, and out of concern for capture fathers' active efforts to promote a sense of
alpha inflation, we report here only these spec- family solidarity and are composed of several items
ified associations. Table 2 contains the relevant querying fathers' inclusion and support of mothers. It
bivariate correlations between pertinent self- may be that in some families, fathers do not work
actively to promote a sense of family solidarity and
report and observational measures of coparent- integrity when mothers do not engage positively with
ing process, as well as the same correlations children—they just step in and engage positively
controlling for the effects of marital quality. with the children themselves. Unfortunately, it is not
Family Integrity. Self-reported efforts to possible to ascertain direction of effects from the
promote Family Integrity were tied to two of the current, contemporaneous data.
 REPORTED AND OBSERVED COPARENTTNG 229

Table 2
Correlations Between the Self-Report and Observational Measures of Coparenting Process
Family process measures (observed)
Child- Parenting Warmth- Management of
Self-report centeredness discrepancy cooperation Antagonism child behavior
Fathers
Integrity .40** -.42** .17 NE NE
33** -36** .17 NE NE
Disparagement NE NE .01 .18 NE
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

NE NE .02 .09 NE
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Conflict NE NE -.13 -24f NE

NE NE -.14 .02 NE
Reprimand NE NE NE NE .09
NE NE NE NE .09

Mothers
Integrity .28* -.10 .12 NE NE
-20f -.02 .12 NE NE
Disparagement NE NE -.06 .28* NE
NE NE -.05 .13 NE
Conflict NE NE -.32* -.52** NE
NE NE -37** 36** NE
Reprimand NE NE NE NE .29*
NE NE NE NE .29*
Note. Values in boldface reflect the same correlations, controlling for marital quality (partial correlations).
NE = not examined.
•fp < .10 (marginally significant). *p < .05. **p < .01.

observed antagonism or to warmth and cooper- Reports of Coparenting and

ation. However, women's reports of interparen- Observational Indices Based on Parenting
tal Conflict were associated with both (high) Data From Dyadic Contexts
observed antagonism and with low warmth and
cooperation during family group interaction. Family Integrity. Although men's Family
These significant associations remained even Integrity scores were associated with a more
after controlling for marital quality. mutual balance of positive parental engagement
during family interaction (i.e., smaller parenting
Reprimand. Men's Reprimand scores were
discrepancies), a similar association did not ob-
not associated with management of toddler be-
tain between Family Integrity scores and dis-
havior in the family. Women who reported crepancies in positive mother-father engage-
more frequent Reprimand activities did belong ment with the toddler during dyadic interaction
to families that demonstrated greater manage- for either men (r = .10, ns) or women (r = .04,
ment of the toddler's behavior, and this connec- ns).
tion remained even after controlling for overall Reprimand. Though women's Reprimand
marital quality. Further, this link did not simply scores (i.e., their reports of limits set within the
reflect mothers* reporting accurately on their family triad) were associated with limit-setting
own behavior in the family group; the relation- by the parenting partners observed during fam-
ship between maternal Reprimand scores and ily group interactions, they were completely
the management by both parents score (r = .29, unrelated to management of the toddlers1 be-
p < .05) exceeded that of both the maternal havior within dyadic contexts. This remained
(r = .26, ns) and paternal (r = .23, ns) man- true regardless of whether the dyad-based index
agement of toddler behavior in the family scores reflected maternal, paternal, or combined man-
considered independently. agement by both parents (the latter index paral-
 230 McHALE, KUERSTEN-HOGAN, LAURETTI, AND RASMUSSEN

leling the one formed for the family group data). child. As in oppositional families, parents in
Correlations between men's Reprimand scores cohesive families provided moderate levels of
and the maternal, paternal, and combined dy- control and structure. Nonrestrictive families
adic scores were .02, .00, and .00. Correlations (Cluster 3) were so named because their most
between women's scores and these three mea- striking feature was the low level of structure
sures were .00, .00, and .00. and limit-setting afforded by parents. This fea-
ture of the family process did not appear to
Coparenting Profiles Within Families reflect a lack of parental incentive, in that both
The observational data reflecting coparenting parents showed similar levels of positive en-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

within the family context were subjected to a gagement with the toddler and an overall orien-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

k-means cluster analysis, computed on stan- tation that was more child- than adult-centered.
dardized scores for interparental antagonism, Indeed, these families appeared to have been
warmth and cooperation, child-centeredness, centered around accommodating to the child
balance of parental involvement, and coparental rather than asking the child to accommodate,
management of the toddler's behavior. This and the connection between parents was not as
analysis revealed three primary clusters (see strong as the one found in cohesive families
Table 3). Oppositional families (Cluster 1) were (levels of antagonism, and of warmth and co-
so named because of very low warmth and operation, were more like those of oppositional
cooperation scores combined with high interpa- coparenting families, though scores were cer-
rental antagonism scores. In these families, par- tainly not as extreme). Of note, 10 of the 11
ents were also more discrepant in their stances
families in this last cluster had toddler sons, not
toward the child (more noteworthy imbalances
in levels of positive engagement) than was true daughters (see Table 3).
of parents in the other two groups. Children's The value of empirical typologies rests both
interests and initiatives were frequently over- on their accurate approximation of underlying
looked in these families (with sessions much data structures and on their reflection of mean-
more adult- than child-centered), while parents ingful, basic differences in family patterns. It
provided moderate levels of control and struc- was hence important to examine the utility of
ture. Cohesive families (Cluster 2) were so the clustering results. We conducted two such
named because they were the most warm and tests. First, we asked whether oppositional, co-
cooperative, and the most child-centered, of any hesive, and nonrestrictive coparenting families
group. Interactions were characterized by very differed from one another in overall levels of
low levels of interparental antagonism and by marital quality. Then, we conducted a discrimi-
balanced levels of positive engagement with the nant analysis to examine the collective ability of

Table 3
Summary of Cluster Analyses
Cluster 1: Cluster 2: Cluster 3:
Index Oppositional Cohesive Nonrestrictive
No. of families 12 29 11
No. of boys/girls 8/4 13/16 10/l a
Firstborn/second born 676 17/12 9/2b
Observational index
Child-centeredness -1.00 .42 -.02
Parenting discrepancy .92 -.27 -.31
Warmth-cooperation -.88 .48 -.31
Antagonism .76 -.50 .49
Management of child behavior .29 .36 -1.27
a
F(2, 49) = 4.04, p < .05. SchefK analyses identified Cluster 3 (91% boys, 9%
girls) as significantly different from Cluster 2 (45% boys, 55% girls); mean
difference = -0.46, F(2, 49) = 3.87, p < .05.
b
F(2, 49) = 1.34, ns.
 REPORTED AND OBSERVED COPARENTING 231

the self-report measures of coparenting to dis- Discussion

tinguish the three family types.
Analyses of the (standardized) marital data With recent advances in the assessment of
revealed that the means for marital quality were family-level behavior, researchers examining
— 1.32 for the oppositional coparenting fami- the family context of early child development
lies, 0.21 for the nonrestrictive families, and are now able to consider whether family-level
0.56 for the cohesive families, F(2, 49) = 3.25, dynamics reflect more than a simple recapitula-
tion of the interpersonal dynamics characteristic
p < .05. Scheffie analysis established that mar-
of dyadic parent-child interaction. In this study,
ital quality among the oppositional coparenting
we tested this notion in two ways. First, we
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

families was significantly lower than it was

asked whether family measures constructed
This document is copyrighted by the American Psychological Association or one of its allied publishers.

among the cohesive families (mean differ- from dyad-based data tell the same story about
ence = 1.88), F(2, 49) = 3.22, p < .05. coparenting and family dynamics as do direct
We then asked whether this means of classi- measures of family group process. Then, we
fying families would distinguish them along asked whether parents* perceptions of their own
dimensions of reported coparenting conduct. A coparenting activities correlate with the types of
multivariate discriminant function analysis as- family group dynamics that might be expected
certained how well maternal and paternal re- to coincide with such activities. As part of this
ports of coparenting collectively distinguished analysis, we asked whether reported coparent-
the groups. The overall Wilks's lambda for this ing behavior would show the same ties to con-
analysis (df — 8) was 0.66 {p < .01). Univariate structed measures of family process (reflecting a
results are shown in Table 4. Self-report vari- synthesis of parental behaviors exhibited within
ables distinguishing the groups were men's re- dyadic parent-child settings). On the basis of
ports of Family Integrity-promoting activities the proposition that the family whole is indeed
(lowest in the oppositional families), women* s distinct from the sum of its constituent parts,
reports of interparental Conflict (highest in the such linkages were not expected.
oppositional and lowest in the cohesive fami- The news from this study can be summarized
lies), and women* s reports of the frequency of as follows. First, researchers do need to be
limit-setting and Reprimand activities (highest sensitive to the assessment contexts from which
in the cohesive families and lowest in the non- family data are drawn. Attempting to character-
restrictive families). These results indicate that ize the relative balance of positive parental in-
parents' reports of the behaviors they engage in volvement with toddlers by comparing maternal
to support or undercut coparenting solidarity do and paternal behavior drawn from dyadic,
indeed differ among families exhibiting diver- parent-child interaction sessions gives a com-
gent patterns of coparental activity during pletely different impression of the family than
laboratory-based family assessments. The self- the one formed by comparing maternal and pa-
reported measures most useful in distinguishing ternal behavior within the context of family
groups, at least in this study, differed for men group interaction. The degree of concordance in
and women. parents' propensity to set limits for their toddler

Table 4
Significant Findings From Discriminant Analyses
Coparenting Scale index
Discriminant analysis F-Integrity M-Conflict M-Reprimand
Cluster 1: Oppositional 35.00 8.25 19.75
Cluster 2: Cohesive 38.72 6.28 20.97
Cluster 3: Nonrestrictive 38.64 7.64 19.27
Univariate F(2, 49) 2.69t 4.60* 3.19*
Note. F-Integrity = fathers' Family Integrity scores; M-Conflict = mothers'
Conflict scores; M-Reprimand = mothers' Reprimand scores,
tp < .10 (marginally significant). *p < .05.
 232 McHALE, KUERSTEN-HOGAN, LAURETTI, AND RASMUSSEN

across dyadic and family settings is somewhat marital quality, which itself then influences co-
greater, but at the same time, parents* reports of parenting process (see, e.g., Belsky & Hsieh,
limit-setting within the family group context are 1998). Concluding that coparenting conduct
completely unrelated to their limit-setting be- makes no difference because direct effects dis-
havior when alone with the child. In short, ev- appear when marital quality is taken into ac-
idence from this study indicates that important count would hence be an erroneous conclusion.
behaviors by the two parents may not be orga- In our view, the persistence of several cross-
nized in the same way across dyadic and family method, domain-specific ties, all along pre-
group contexts. dicted gender lines, even after accounting for
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

The second piece of news is that self-reports overall marital quality, calls for further research
This document is copyrighted by the American Psychological Association or one of its allied publishers.

from parents about their own behavior in sup- on how parents' perceptions of their coparent-
port of coparenting solidarity are a useful sup- ing activities can contribute to our understand-
plement to information about the marriage in ing of family process. In future work, we rec-
predicting family interaction patterns. In partic- ommend at least two lines of inquiry. First, we
ular, fathers' reports of family integrity- see value in the formulation and testing of hy-
promoting behavior and mothers' estimates of potheses guided by specific predictions about
both interparental conflict and limit-setting ac- the potentially unique contributions of mothers
tivities explain significant variance in family and fathers to the family group process. This
group-level dynamics. The general pattern con- study began with the notion that active efforts to
cerning fathers' efforts to promote family integ- promote family integrity would be a particularly
rity and mothers' awareness of family conflict salient index of men's commitment to family
behaviors accords well with predicted gender solidarity, with correlates in observational mea-
patterns, though no family-based correlates of sures of family cohesion, and that women's
mothers' self-reported disparagement of fathers attunement to conflict would be a telling indi-
turned up in this study. We note also that, col- cator of dissonance in the family group process.
lectively, fathers' Family Integrity scores and This latter prediction derived from an emerging
mothers' Conflict and Reprimand scores helped literature on maternal gatekeeping, in which
to distinguish oppositional, cohesive, and non- women's feelings about their partners' involve-
restrictive family group patterns. ment (positive or negative) influence and are
influenced by family group dynamics. Though
These results indicate that it is important to patterns in our data set were largely consistent
attend to assessment context in coparenting and with these general hypotheses, matters are cer-
family research. They also provide beginning tainly more complex than this in many families,
evidence for the validity of parental reports of and longitudinal work is needed to disentangle
coparenting behavior. However, it is important direction of effects between individual parents'
not to overinterpret these data. Our hunch that behavior and family group process. Our point
we would find little overlap between family here is that such advances are unlikely to occur
indices based on parents' behavior in dyadic in the absence of specific predictions about
contexts and parallel indices based on data from whether different dimensions of contribution by
family group interaction was only partially con- mothers and fathers really affect families in
firmed. Findings concerning the relative balance different ways.
of positive parental engagement with the child
underscored this point dramatically, whereas This dovetails with our second recommenda-
data on parent-parent management of toddler tion, that researchers consider querying parents
behavior revealed at least some between- about their own coparenting-related conduct
settings consistency. Similarly, the fact that (allying themselves verbally with the partner
only 5 of 18 anticipated correlations between rather than the child after the partner has repri-
self-reported and observed coparenting behav- manded the child, speaking on the child's behalf
iors persisted after marital quality was consid- when the child protests a course of action pro-
ered might be taken as evidence that parental posed by the partner, rebuking a request by the
reports of coparenting are not very potent indi- partner to engage in a nondesired activity with
ces. In this latter case, however, it could also be the child) and then studying the individual's
argued that coparenting behaviors influence behavior during subsequent family commerce
 REPORTED AND OBSERVED COPARENTING 233

(rather than seeking corollaries in the group family, fathers of toddlers who say that they
dynamic as we did in this study). Linking par- work to promote a cohesive sense of family
ents* reports of their own behavior to measures belong to family systems that (a) are sensitive in
of the same parents' behavior in the family responding to the child and (b) involve more
group might be expected to reveal even greater balanced levels of positive parenting by both
concordance between reported and observed co- adults. It remains possible that a link between
parenting behavior than was found in this in- Family Integrity-promoting behavior and fam-
vestigation. Such work could also extend the ily warmth and cooperation exists in nature but
focus of this investigation by asking whether that it went undetected in this study either be-
parents' reports of their partner's behavior sim- cause of a lack of statistical power related to the
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

ilarly relate to family group dynamics. It would relatively small sample size or because the
be important to know whether the same sets of range of participant fathers in the sample re-
constructs, when estimated by the partner, flected a relatively narrow subset of dads,
would likewise emerge as the most important skewed toward the warm and involved end of
predictors of the family group process. the fathering continuum. Further research is
We now turn to the "nonfindings" from this needed to determine whether these findings can
study. First, self-reports of Disparagement did be replicated with a more diverse group of fam-
not show reliable connections with observed ilies and to establish whether the family con-
family process. Though maternal reports of Dis- comitants of parents' self-described coparenting
paragement, an index associated with preschool activities can be replicated at other stages of
behavior problems, were significantly associ- child development beyond the challenging tod-
ated with interparental antagonism and greater dler years.
parenting discrepancies, these ties disappeared The limited sample size in this study made it
once marital quality was taken into account. impractical to conduct separate analyses for
This may be an instance in which an observa- boys and girls, or for firstborn and later-born
tional measure directly paralleling the self- toddlers. Yet family process is definitely sensi-
report construct (i.e., critical remarks about the tive to child gender (Ablow, 1997; Kerig,
partner to the child) would have been useful. Cowan, & Cowan, 1993; McHale, 1995), re-
Alternatively, this nonfinding may have more to flected in the unbalanced gender distributions
do with the context-dependence issue raised within the three types of coparenting families
earlier in this article. Disparagement as concep- that emerged from empirical clustering of the
tualized in the Coparenting Scale emphasizes family process data. Moreover, there are very
communications between parents and children different demands faced by parents for whom
when they are alone (in a dyadic context). This the toddler is an only child, a younger sibling, or
emphasis on the dyad may be why no behav- an older brother or sister, and different sibling
ioral referents were found in the family context constellations undoubtedly call forth and give
Because parents rarely criticize their partner to different shades of meaning to parenting and
their child in front of researchers* cameras coparenting behavior within the family context.
(McHale & King, 1996), validating this scale Given these complexities, we advocate that the
against observed disparagement of the partner results of this study be interpreted primarily for
promises to be a challenging task. However, their heuristic value. Systematic studies of dif-
because disparagement connotes disapproval, ferent family configurations with large enough
future studies might examine whether parents samples to test specific gender-related hypoth-
who say they frequently disparage the coparen- eses are clearly in order.
tal partner to the child are aware of, and disap-
prove of, the other parent's parenting style We also emphasize that the variables exam-
when alone with that child. ined in this study sample only a small subset of
what researchers have traditionally referred to
A second puzzling nonfinding was the null as coparenting. For example, the findings re-
relationship between fathers' reports of Family ported do not address perceived or actual divi-
Integrity-supporting behavior and observed sion of labor in families. In addition, the find-
family warmth and cooperation. Rather than ings should be interpreted only with reference to
having an extraordinarily warm and supportive the particular assessment contexts in which they
 234 McHALE, KUERSTEN-HOGAN, LAURETTL AND RASMUSSEN

were studied. More comprehensive assessments ily contexts: Does the presence of children make a
of family behavior should be (a) gathered in difference? Paper presented at the meeting of the
naturalistic as well as laboratory contexts and Society for Research in Child Development, Wash-
(b) sampled both during "down times" and at ington, DC.
Ahrons, C. (1981). The continuing coparental rela-
times of normative stress (meal-, bed-, and bath
tionship between divorced spouses. American Jour-
times; transitions to day care). These caveats in nal of Orthopsychiatry, 51, 415-428.
mind, it is noteworthy that the significant and Allen, S., & Hawkins, A. (1999). Mothers' beliefs
domain-specific associations found in this study and behaviors that inhibit greater father involve-
were all established using a relatively brief (30- ment in family work. Journal of Marriage and the
min) sample of playful family behavior ob-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Family, 61, 199-212.

This document is copyrighted by the American Psychological Association or one of its allied publishers.

tained in a laboratory context from participants Baumrind, D. (1971). Current patterns of parental
not in clinical distress. It seems safe to conclude authority. Developmental Psychology, 4, 1-103.
that there is value in asking parents to report on Belsky, I , Cmic, K., & Gable, S. (1995). The deter-
their own coparental behavior and in tracing the minants of coparenting in families with toddler
linkages between these reports and family-level boys: Spousal differences and daily hassles. Child
dynamics. Development, 66, 629-642.
Belsky, J., & Hsieh, K. (1998). Patterns of marital
change during the early childhood years: Parent
Implications for Application and personality, coparenting, and division of labor cor-
Public Policy relates. Journal of Family Psychology, 12, 511-
528.
This study represents foundational research Belsky, J., Putnam, S., & Crnic, K. (1996). Coparent-
in a brand new area, and as such replication and ing, parenting, and early emotional development. In
extension studies are needed before we can as- J. McHale & P. Cowan (Eds.), Understanding how
family-level dynamics affect children's develop-
sert with confidence that self-reported coparent-
ment: Studies of two-parent families. New Direc-
ing data should invariably be a component of tions for Child Development (Vol. 74, pp. 45-55).
any clinical assessments of children and fami- San Francisco: Jossey-Bass.
lies. This said, we certainly can imagine a time Brody, G., Flor, D., & Neubaum, E. (1998). Copa-
a few years down the road when practitioners renting process and child competence among rural
might consider targeting coparental practices as African-American families. In M. Lewis & C. Feir-
the focus of their interventions rather than try- ing (Eds.), Families, risk and competence (pp. 227-
ing to intervene directly in couples' marital 243). Mahwah, NJ: Erlbaum.
relationships. Further studies carefully delineat- Buhrmeister, D., Camparo, L., Christensen, A.,
ing coparental from marital domains in the fam- Gonzalez, L., & Hinshaw, S. (1992). Mothers and
ily, substantiating the interplay between these fathers interacting in dyads and triads with normal
two systems, and evaluating their differential and hyperactive sons. Developmental Psychology,
28, 500-509.
and combined impact on child adjustment will
Clarke-Stewart, K. (1978). And daddy makes three:
eventually enable us to offer confident and spe- The father's impact on mother and young child.
cific suggestions about how best to conceptual- Child Development, 49, 466-478.
ize and work clinically with coparenting dy- Coiro, M., & Emery, R. (1998). Do marriage prob-
namics in a manner that benefits children and lems affect fathering more than mothering? A quan-
families. At present, we do recommend that titative and qualitative review. Clinical Child and
professionals working with families observe Family Psychology Review, 1, 23-40.
parents in both dyadic and family group con- Cowan, P. (1988). Becoming a father: A time of
texts, because impressions of the coparenting change, an opportunity for development. In P.
team inferred from these two settings may pro- Bronstein & C. Cowan (Eds.), Fatherhood today:
vide different pictures. Men's changing role in the family (pp. 13-35). New
York: Wiley.
Cowan, P., Cowan, C , Schulz, M., & Heming, T.
References (1994). Prebirth to preschool family factors predict-
ing children's adaptation to kindergarten. In R.
Abidin, R., & Brunner, J. (1995). Development of a Parke & S. Kellam (Eds.), Advances in family re-
Parenting Alliance Inventory. Journal of Clinical search (Vol. 4, pp. 75-114). Hillsdale, NJ: Erl-
Child Psychology, 24, 31-40. baum.
Ablow, J. (1997, April). Marital conflict across fam- Cox, M., Tresch-Owen, M., & Lewis, J. (1989,
 REPORTED AND OBSERVED COPARENTING 235

April). Pre-birth marital relationships and child quality and gender differences in parent-child in-
development at age 2 1/2. Paper presented at the teraction. Developmental Psychology, 29, 931-939.
meeting of the Society for Research in Child De- Lauretti, A., & McHale, J. (1997, April). Shifting
velopment, Washington, DC. patterns of parenting styles between dyadic and
Cummings, E. M., & Davies, P. (1996). Emotional family settings: The role of marital quality. Paper
security as a regulatory process in normal develop- presented at the meeting of the Society for Research
ment and developmental psychopathology. Devel- in Child Development, Washington, DC.
opment and Psychopathology, 8, 123-139. Lidz, T. (1963). The family and human adaptation.
Deal, J., Hagan, M., Bass, B., He then ng ton, E. M., & New York: International Universities Press.
Clingempeel, G. (1999). Marital interaction in dy- Lindahl, K., Clements, M., & Markman, H. (1997).
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

adic and triadic contexts: Continuities and discon- Predicting marital and parent functioning in dyads
This document is copyrighted by the American Psychological Association or one of its allied publishers.

tinuities. Family Process, 38, 105-115. and triads: A longitudinal investigation of marital
Dickstein, S., & Parke, R. (1988). Social referencing processes. Journal of Family Psychology, 11, 139—
in infancy: A glance at fathers and marriage. Child 151.
Development, 59, 506-511. Lindahl, K., & Malik, N. (1999). Marital conflict,
Epstein, N., & Bishop, D. (1981). Problem-centered family process, and boys' externalizing behavior in
systems therapy of the family. Journal of Marital Hispanic and European-American families. Journal
and Family Therapy, 7, 23-31. of Clinical Child Psychology, 28, 12-24.
Feldman, S., Wentzel, K., & Gehring, T. (1989). A Locke, H., & Wallace, K. (1959). Short marital-
comparison of the views of mothers, fathers, and adjustment and prediction tests: Their reliability
pre-adolescents about family cohesion and power. and validity. Marriage and Family Living, 21, 251-
Journal of Family Psychology, 3, 39-60. 255.
Filsinger, E. (1990). Empirical typology, cluster anal- McHale, J. (1992, April). Cross-contextual coheren-
ysis, and family-level measurement. In T. Draper & cies between couple and coparenting dynamics dur-
A. Marcos (Eds.), Family variables: Conceptual- ing infancy. Paper presented at the International
ization, measurement, and use (pp. 90—104). New- Conference on Infant Studies, Miami, FL.
bury Park, CA: Sage. McHale, J. (1995). Coparenting and triadic interac-
Floyd, F., Costigan, L., & Gilliom, K. (1998). Mar- tions during infancy: The roles of marital distress
riage and the parenting alliance: Longitudinal pre- and child gender. Developmental Psychology, 31,
diction of change in parenting perceptions and be- 985-996.
haviors. Child Development, 69, 1461-1479.
McHale, J. (1997). Overt and covert coparenting
Gjerde, P. (1986). The interpersonal structure of fam-
processes in the family. Family Process, 36, 183—
ily interactional settings: Parent-adolescent rela-
210.
tions in dyads and triads. Developmental Psychol-
McHale, J., Johnson, D., & Sinclair, R. (1999). Fam-
ogy, 48, 711-717.
ily dynamics, preschoolers* family representations,
Gordis, EM Margolin, G., & John, R. (1997). Marital
and preschool peer relationships. Early Education
aggression, observed parental hostility, and child
and Development, 10, 373-401.
behavior during triadic family interaction. Journal
of Family Psychology, 11, 76-89. McHale, J., & King, A. (1996, August). Supportive
Gottman, J., Markman, H., & Notarius, C. (1977). and critical remarks about the coparent during
The topography of marital conflict: A sequential parent-child play. Paper presented at the 104th
analysis of verbal and nonverbal behavior. Journal Annual Convention of the American Psychological
of Marriage and the Family, 9, 461-477. Association, Toronto, Ontario, Canada.
Jouriles, E., Murphy, G, Farris, A., Smith, D., Rich- McHale, J., Kuersten, R., & Loding, B. (1998, Au-
ters, J., & Waters, E. (1991). Marital adjustment, gust). Stability of coparenting process from the
parental disagreements about child-rearing, and be- toddler to the preschool years. Paper presented at
havior problems in boys: Increasing the specificity the 106th Annual Convention of the American Psy-
of the marital assessment. Child Development, 62, chological Association, San Francisco, CA.
1424-1433. McHale, J., Kuersten-Hogan, R., & Lauretti, A.
Katz, L., & Gottman, J. (1996). Spillover effects of (2000). Evaluating coparenting and family-level
marital conflict: In search of parenting and copa- dynamics during infancy and early childhood: The
renting mechanisms. In J. McHale & P. Cowan Coparenting and Family Rating System. In P. Kerig
(Eds.), Understanding how family-level dynamics & K. Lindahl (Eds.), Family observational coding
affect children's development: Studies of two- systems: Resources for systemic research (pp. 151-
parent families. New Directions for Child Develop- 170). Mahwah, NJ: Erlbaum.
ment (Vol. 74, pp. 57-76). San Francisco: Jossey- McHale, J., Rao, N., & Krasnow, A. (2000). Con-
Bass. structing family climates: Chinese mothers' reports
Kerig, P., Cowan, P., & Cowan, C. (1993). Marital of coparenting behavior and preschool behavior
 236 McHALE, KUERSTEN-HOGAN, LAURETTI, AND RASMUSSEN

problems. International Journal of Behavioural tive: A multi-level developmental approach. In M.

Development, 24, 111-118. Hetherington, R. Lerner, & M. Perl mutter (Eds.),
McHale, J., & Rasmussen, J. (1998). Coparental and Child development in life-span perspective. Hills-
family group-level dynamics during infancy: Early dale, NJ: Erlbaum.
family predictors of child and family functioning Sagrestano, L., Christensen, A., & Heavey, C.
during preschool. Development and Psychopathol- (1998). Social influence techniques during marital
ogy, 10, 39-58. conflict Personal Relationships, 5, 75-89,
Minuchin, P. (1985). Families and individual devel- Tabachnik, B., & Fidell, L. (1996). Using multivar-
opment: Provocations from the field of family ther- iate statistics. New York: Harper & Row.
apy. Child Development, 56, 289-302. Vaughn, B., Block, J., & Block, J. (1989). Parental
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Minuchin, S. (1974). Families and family therapy. agreement on child-rearing during early childhood
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Cambridge, MA: Harvard University Press. and the psychological characteristics of adoles-
Moos, R. (1974). Combined preliminary manual for
cents. Child Development, 59, 1020-1033.
the family, work, and group environment scales.
Palo Alto, CA: Consulting Psychologists Press. Westerman, M., & Massoff, M. (1995, August). Tri-
Moos, R., & Moos, B. (1976). A typology of family adic coordination, marital adjustment, and child
social environments. Family Process, 15, 357-371. behavior problems. Paper presented at the 103rd
Olson, D., Sprenkle, D.( & Russell, C. (1979). Cir- Annual Convention of the American Psychological
cumplex model of marital and family systems: I. Association, New York, NY.
Cohesion and adaptability dimensions, family
types, and clinical applications. Family Process, 18, Received January 6, 1998
3-28. Revision received November 18, 1999
Parke, R. D. (1988). Families in life-span perspec- Accepted December 13, 1999 •