Individual Parental Adjustment Moderates the Relationship Between Marital and Coparenting Quality

Abstract

Contemporary family research studies have devoted surprisingly little effort to elucidating the interplay between adults’ individual adjustment and the dynamics of their coparental relationship. In this study, we assessed two particularly relevant “trait” variables, parental flexibility and self-control, and traced links between these characteristics and the nature of the coparents’ interactions together with their infants. It was hypothesized that parental flexibility and self-control would not only explain significant variance in coparenting quality, but also act as moderators attenuating anticipated relationships between marital functioning and coparental process. Participants were 50 heterosexual, married couples and their 12-month-old infants. Multiple regression analyses indicated that even after controlling for marital quality, paternal flexibility and maternal self-control continued to make independent contributions to coparenting harmony. As anticipated, paternal flexibility attenuated the association between marital quality and coparenting negativity. Contrary to predictions, maternal flexibility and self-control did not dampen, but actually heightened the extent to which coparenting harmony declined in the face of lower marital quality.

Over the past two and a half decades, family theory and research have expanded the focus of their inquiries beyond mother–child dyads as the primary contexts for childhood socialization. The work of Belsky (e.g., Belsky, 1979, 1981; Belsky, Youngblade, Rovine, & Volling, 1991), Brody (e.g., Brody, Pellegrini, & Sigel, 1986), Cummings (e.g., Cummings, 1994), Easterbrooks (e.g., Easterbrooks & Emde, 1988; Goldberg & Easterbrooks, 1984), and other prominent family researchers (e.g., Cowan & Cowan, 1987, 1990, 1992; Cox, Tresch-Owen, Lewis, & Henderson, 1989; Dickstein & Parke, 1988; Emde, 1991; Engfer, 1988; Katz & Gottman, 1996, 1997) helped promote not only an enhanced awareness of family triads, but also an important reconceptualization of such triads as systems composed of qualitatively distinct units—parent–child dyads, spousal dyads, and individuals—all linked together through patterns of mutual influence.

In reflecting on these developments, workers in the field of family therapy, most notably Patricia Minuchin (1985, 1988; see also Nichols & Schwartz, 1998) welcomed what they saw as a more systemic approach to the study of the family. At the same time, however, they urged researchers to make further conceptual strides. Minuchin, in particular, pointed out that in the family triad, each participant was simultaneously in some form of contact with each of the two others. She also contended that this reality was not adequately captured by a model of triadic dynamics as an amalgam of discrete, interlocking dyads within the family, and that triadic processes should themselves be considered as units of analysis.

Responding to Minuchin's provocations, several researchers sought to operationalize this more holistic view of the family triad by developing and elaborating the construct of coparenting (see e.g., Gable, Belsky, & Crnic, 1992; McHale, 1995; McHale, Kuersten, & Lauretti, 1996; Weissman & Cohen, 1985). Coparenting has been conceptualized as an enterprise undertaken by two or more adults working together to raise a child for whom they share responsibility (McHale, Lauretti, Talbot, & Pouquette, 2002). It can involve any set of adults participating jointly in childrearing: for example, coparenting partnerships may be formed by children's two biological parents, whether married (e.g., Lewis, Beavers, Gossett, & Phillips, 1976; McHale, 1995; Katz & Gottman, 1996), divorced (e.g., Buchanan, Maccoby, & Dornbusch, 1991), or never married; by grandmother–mother dyads (e.g., Chase-Lansdale, Brooks-Gunn, & Zamsky, 1994; Wakschlag, Chase-Lansdale, & Brooks-Gunn, 1996); by same-sex couples (e.g., Patterson et al., 2004; Patterson & Chan, 1999) or by any other set of adults who make a commitment to joint child-rearing.

Although coparenting relationships are often enacted by members of marital dyads, the construct is implicitly triadic (or polyadic; see McHale et al., 2004), in that all the interactions it comprises pertain to the partners’ child and their shared connection to that child. Coparenting transactions, in other words, can be seen as a specific form of triadic-level family process. In the first decade of research on coparenting dynamics, several overall patterns were discerned, all of which help to establish the construct validity of the coparenting concept: First, coparenting quality, whether assessed via self-report (e.g., Abidin & Brunner, 1995; Block, Block, & Morrison, 1981; Brody & Flor, 1996; Floyd & Zmich, 1991; Jouriles et al., 1991; McHale & Rasmussen, 1998) or observational methodologies (e.g., Belsky, Putnam, & Crnic, 1996; Brody & Flor, 1996; Fivaz-Depeursinge, Frascarolo, & Corboz-Warnery, 1996; Lindahl, 1998; McConnell & Kerig, 1999; McHale, Johnson, & Sinclair, 1999; McHale & Rasmussen, 1998) has reliably shown systematic and lawful associations with a number of important indicators of children's socioemotional adjustment. For example, cooperative, warm, and balanced coparenting has been shown to predict children's positive socioemotional adjustment (e.g., Block et al., 1981; Jouriles et al., 1991; McHale, Lauretti, & Talbot, 1998), whereas conflicted or otherwise negative coparenting predicts a variety of internalizing and externalizing child behavior problems (e.g., McConnell & Kerig, 1999; McHale & Rasmussen, 1998; Stright & Nietzel, 2003).

A second important line of investigation has provided at least preliminary indications that coparenting dynamics are empirically distinguishable from dimensions of dyadic parent–child relationship quality (e.g., Buhrmester et al., 1992; Gjerde, 1986; McHale, Kuersten-Hogan, Lauretti, & Rasmussen, 2000). Third, coparenting and marital processes also seem to be distinct from one another in terms of their respective impacts on children's adjustment (Jouriles et al., 1991; McHale & Rasmussen, 1998; Snyder, Klein, Gdowski, Faulstich, & LaCombe, 1988).

Fourth, in addition to underscoring the distinctions between coparenting interactions and marital processes, coparenting researchers have closely examined the nature of the associations between these constructs. In their approach to this issue, researchers have focused on testing the premise, central to clinical family theories of the structural (e.g., Minuchin, 1974), strategic (e.g., Haley, 1988), and Bowen schools (e.g., Bowen, 1976, 1978; Kerr & Bowen, 1988), that marital and coparenting dynamics are closely interrelated. Proponents of these mainstream family theories maintain that happily married spouses feel enduring trust and affection toward one another, and that these attitudes, in turn, predispose them to work cooperatively in the coparenting domain; on the other hand, negativity between spouses, whether in the form of covert bitterness or frank animosity, is believed to damage coparenting quality.

Numerous reports have substantiated the predicted positive association between marital quality and coparenting, including several correlational studies (e.g., Abidin & Brunner, 1995; Frank, Jacobson, Hole, Justkowski, & Huyck, 1986; Floyd & Zmich, 1991; Katz & Gottman, 1996; Kerig, 1995; Lewis, 1989; Lewis, Beavers, Gossett, & Phillips, 1976; Lewis & Looney, 1983; Lewis, Tresch-Owen, & Cox, 1988; Lindahl, Clements, & Markman, 1997; McHale, 1995, 1997) and three prospective investigations that assessed marital quality before the transition to parenthood and coparenting in the family triad at later points after the birth of the first child (McHale et al., 2004—3 months postpartum; Lewis, Tresch-Owen, & Cox, 1988—1 year postpartum; Lindahl, Clements, & Markman, 1997—5 years postpartum). Reports have concurred that whole-family processes favorable to children's adjustment are more often found in families headed by adult partners whose marital interactions are harmonious, whereas dysfunctional coparenting typically co-occurs with forms of marital discord such as openly expressed conflict or patterns of dominance and submission between spouses.

These findings regarding main effects raise a question as to whether parents in marital distress can ever “beat the odds” by somehow putting aside the differences between them and forging effective coparenting relationships. This issue was initially raised by Lewis (Lewis et al., 1976; Lewis & Looney, 1983) and rekindled by McHale (1995; McHale, Kuersten-Hogan, Lauretti, & Rasmussen, 2000), each of whom reported that among a subgroup of their participating families, the expectable overlap between marital and coparenting quality did not hold. That is, couples in certain families were experiencing appreciable levels of marital distress, and yet their coparenting did not exhibit the markers of risk that have been linked to adverse child outcomes. Lewis described such families as “Competent but Pained.”

In describing these cases, Lewis (Lewis et al., 1976; Lewis & Looney, 1983) portrayed spouses in the typical Competent but Pained family as having not succeeded in attaining a level of intimacy that was satisfying to both of them. One member of the couple, usually the man, appeared detached and professed not to value interpersonal connections highly, whereas the other partner was open in voicing her dissatisfaction about the lack of closeness in the marriage. This unhappy wife projected a sense of resignation as well as of sadness and anger; she made no efforts to heighten her partner's engagement in their marital interactions. In addition to the lack of emotional intimacy in the marriage, disparities in the distribution of power might also be present.

In the whole-family context, however, the two members of the couple maintained comparable levels of engagement with their children, and cooperated efficiently when task performance was called for. They refrained from engaging in competitive interactions and shows of hostility in the presence of their children. Although neither the spouses nor their children displayed the high degree of warmth or affective spontaneity characteristic of Lewis’ “Highly Competent” or optimally functioning families, the family was able to maintain a tone of matter-of-fact politeness. They were attentive to one another and generally clear in their communications of their own views. As Lewis observed, the most telling measure of the Competent but Pained family's success in coparenting could be found in the regularity with which their adolescent children exhibited high levels of psychological health. Though these adolescents tended to be somewhat less emotionally open than their peers from Highly Competent families, they were free of the adjustment problems that often beset adolescents from family environments rated dysfunctional in Lewis’ research samples.

In a 1995 study on coparenting in families with infants, McHale (1995; see also McHale, Kuersten-Hogan, Lauretti, & Rasmussen, 2000a) discerned a subset of families whose dynamics resembled Lewis’ Competent but Pained pattern. In McHale's report, there was again a positive correlation between marital and coparenting quality in analyses including the sample as a whole. But a small percentage of the maritally distressed partners were exceptional in that they were relatively civil and cooperative in the triadic context. Unlike other couples showing similar levels of marital discord, the members of this subgroup did not manifest hostility, competitiveness, or discrepant levels of involvement in triadic interactions with their babies. Their coparenting interactions appeared disturbed only to the extent that their displays of shared positive emotion appeared restricted when compared with those of happily married coparenting teams.

In sum, although family theories assert and studies indicate that marital discord typically bodes ill for coparenting quality, Lewis’ and McHale's findings raise the possibility that in some families, protective factors operate to mitigate the deleterious effects of marital problems on partners’ ability to function as a coparenting team.

But what sets families such as the Competent but Pained ones apart from families in which both marital and coparenting subsystems show disturbance? In Lewis’ work, researchers drew upon individual interviews with participating parents as a means of gaining insight into this question. In reflecting on these interview responses, Lewis noted that Competent but Pained partners, unlike the adult members of more distressed families, emphasized their willingness to make personal sacrifices as a means of promoting their children's well-being. They saw this value as implying that for the sake of their children, they must maintain cohesive family relationships and suppress their resentment against their spouses to promote whole-family functioning.

Another distinguishing characteristic of Competent but Pained partners was their capacity for objectivity in their appraisals of family dynamics. Typically, members of a Competent but Pained couple were able to articulate a distinction between their feelings about one another as spouses and their assessments of one another's parenting. Although these partners clearly voiced dissatisfaction with their marriages, they were open to acknowledging one another's strengths in the domain of child rearing. Furthermore, parents in Competent but Pained families made an effort to view family dynamics through the eyes of their children. Thus, each partner evinced respect for the children's attachment to the other parent. Competent but Pained partners also showed an ability to reflect on their children's responses to their coparenting behaviors. Thus, they perceived that coparenting cooperation could enhance their children's sense of security, whereas coparenting conflict could evoke feelings of tension and alarm in their children.

In sum, Lewis’ observations suggest that perhaps certain individual-level psychological strengths on the part of parents could in fact attenuate the relationship between marital quality and coparenting quality, and thus limit the damage done to coparenting by distressed marital functioning. In particular, his findings point toward the possibility that when maritally distressed parents endorse a belief in approaching family interactions with an attitude of altruistically motivated self-control, they may be relatively successful in minimizing their hostility and reactivity toward their partners and in maintaining some level of cooperation within the family domain. In addition, it appears that coparenting quality may be partially protected from the impact of marital conflict in cases where parents display a certain cognitive flexibility in their thinking about family dynamics. When maritally dissatisfied parents consider their children's perspectives on family relationships and strive to be objective in their judgments about their partners’ parenting, they may arrive at a heightened appreciation for the importance of their partners in their children's emotional lives. In consequence, they may have increased motivation both to inhibit their expressions of resentment toward their partners in the triad and to engage more actively in collaborative coparenting.

The current investigation built on the work of Lewis and his colleagues by considering whether linkages between marital quality and coparenting differ depending on the levels of self-control or flexibility shown by parents. Hence, we predicted that the personality traits of better-developed flexibility and self-control would not only be associated with more harmonious and less dissonant coparental dynamics in the family triad, but would also alter the expected relationship between marital processes and coparenting dynamics. That is, we anticipated that among parents reporting higher flexibility and self–control, marital strain would be less likely to have eroded coparenting harmony than it was in companion families where parents did not possess these same attributes. Second, it was predicted that high self-control or flexibility in mothers or fathers would also limit the escalation in coparenting negativity that generally accompanies declines in marital quality. This examination of potential moderators reflected a first attempt to move beyond the study of main effects that has thus far predominated in coparenting research.

METHOD

Participants

Participants were 50 heterosexual, married couples with 12-month-old infants (35 boys and 15 girls). Families were contributors to a larger longitudinal investigation on the early development of coparenting and other whole-family dynamics. All were native speakers of English, and 92% of the adult participants were Caucasian. Mothers’ mean age was 32, and the mean age for fathers was 33. About 50% of the sample earned incomes between $45,000 and $75,000. A total of 98% of fathers and 70% of mothers engaged in paid employment. Among the 12-month-olds, 64% were first-born and the remainder second-born with one sibling 18–30 months older. In these latter families, only the 12-month-olds participated in this study.

Recruitment

Participants were recruited through media announcements, advertisements placed in local newspapers and free parenting magazines, and a direct mailing targeting eligible families within a 19-mile radius of the research site. Recruitment was also conducted via contacts with local agencies including OB/GYN practices, birth preparation classes, family health centers, and day care centers.

Design and Procedures

Each participating family visited a university-based Family Study Center on three separate occasions approximately 1 week apart. At the time of the first evaluation, each adult participant was given the California Psychological Inventory (Gough, 1987) to complete at home. They were instructed not to discuss their responses with their partners, and to return their responses by the time of their third assessment session.

During the second evaluation, families took part in a 10-min play session involving the mother—father–infant triad. In the first 5 min of this session, parents were asked to help the infant perform, or at least attempt, three structured tasks: (a) placing pieces in an inset puzzle, (b) stacking three blocks, and (c) detaching and reattaching two interlocking objects. Members of the family triad were invited to play together for the remaining 5 min in whatever way they liked. The triadic interaction was videotaped.

Parents participated in the third session without the infant. At this time, the couple engaged jointly in a 20-min revealed-differences procedure, in which the couple discussed two issues in their marriage that they had identified as problematic. Partners also completed a Post-Discussion Questionnaire (PDS) that asked them to describe their own feelings and to indicate their perceptions of their partners’ feelings about the discussion. The session was videotaped in its entirety.

Measures

Marital Quality

Both observational and self-report measures were used to estimate the quality of marital relationships. First, videotapes of the revealed-difference interactions were coded by two trained graduate student judges, who received 20 hr of training in the Cox et al. (1989) adaptation of the Timberlawn rating system. The variables of interest in this report are Overt Conflict (M = 2.95, SD = .93, range = 1–5) and Problem Solving (M = 3.05, SD = .72, range = 1–5). The reliability and validity of the Cox et al. scales have previously been established in reports by McHale (1995) and Lauretti (1997); in the present study, interrater reliability was acceptable (intra-class correlations of .83 and .79 for the Conflict and Problem Solving scales, respectively). As the observational measures of Overt Conflict and Problem-Solving were significantly intercorrelated (r = –.45, p < .001), each measure was standardized and then the two were combined by subtracting Overt Conflict from Problem-Solving, yielding a single score for observed marital quality.

Second, the partners’ own ratings of their marital interactions were called upon as “insider” indicators. Following each of the two discussions in the revealed-differences task, each spouse completed a PDQ using a 9-point scale to indicate the extent to which they agreed with each of 24 statements about the interaction. Twelve PDQ items concerned the respondent's own affective reactions to the marital discussion in which s/he had just taken part (e.g., “during the discussion, I was: anxious and nervous; defensive and apologetic; withdrawn and silent”), whereas the other 12 asked the respondent to provide a parallel appraisal of the partner's emotional responses to the discussion (e.g., during the discussion, my partner was: anxious and nervous; defensive and apologetic; withdrawn and silent). Principal components analyses with varimax rotation were then conducted as a means of data reduction.

Three distinct factors were replicated in the analyses of mothers’ and fathers’ data. These three factors, respectively, appeared to represent the extent to which respondents felt that their interactions had positive outcomes (e.g., “After we stopped the discussion, I felt: understood by my partner; satisfied with our talk”), the extent to which they perceived themselves as critical of their partners (e.g., “During the discussion, I was critical and blaming”), and the degree to which they felt disengaged from the discussions (e.g., “During the discussion, I avoided talk about the issue”). The factors were therefore named Good Outcome (self-report), Criticism (self-report), and Disengagement (self-report). Items loading high on each factor were standardized and combined to create a score representing that factor. Kindred factors also replicated for respondents’ ratings of their partners, and hence they were named Good Outcome (report on Partner), Criticism (report on Partner), and Disengagement (report on Partner). Once again scores for the three factors were combined following the same approach applied to the respondents’ ratings of themselves.

For both mothers and fathers, the three self-report and three report on Partner scores were significantly and highly intercorrelated with one another. Hence, two summary variables were formed (one for mother's data and one for father's data). The Criticism and Disengagement scales were reverse weighted (multiplied by −1) and added to the Good Outcome scales. Hence, high scores on the final two summary variables (Mother's overall appraisal of marital interaction, α = .79; Father's overall appraisal of marital interaction, α = .83) reflected views by the respondent that both they and their partner had engaged productively in the problem-solving exchanges (neither disengaging nor engaging in hostile criticism) and worked toward a good outcome, whereas low scores signified perspectives that one or both partners had behaved defensively and been unproductive in their efforts.

Correlations among the composite indicators of observed marital quality, mother's overall appraisal of the marital interaction, and father's overall appraisal of the marital interaction are presented in Table I. As shown in Table I, these three composite measures of marital functioning were significantly related to one another. Therefore, they were normalized and added together to create a final, global index signifying overall marital quality. This measure of overall marital quality is the one that will be used in the major analyses of the study.

Table I.

Correlations Among Marital Quality Composites

	Observed marital quality composite	Mother's overall appraisal of marital interaction	Father's overall appraisal of marital interaction
Observed marital quality composite	1.00	.48***	.48***
Mother's overall appraisal of marital interaction	—	1.00	.72***
Father's overall appraisal of marital interaction	—	—	1.00

Note. N = 50.

^***p < .001.

Individual Adjustment

Each parent also completed the California Psychological Inventory (CPI; Gough, 1987), a 472-item self-report instrument measuring individual psychological and interpersonal adaptation. The CPI consists of 20 subscales designed to represent “folk concepts” commonly used by lay people to interpret and predict their own behavior and that of others. The validity of the CPI has been well- documented (Gough, 1987; Megargee, 1972).

The individual parental adjustment composites employed in this study were based on Gough's (1987) factor analyses of the California Psychological Inventory. The CPI Flexibility factor is defined by high loadings on subscales for Tolerance, Achievement via Independence, and Flexibility, with moderate loadings on Intellectual Efficiency and Psychological-Mindedness. Those who score high on the Flexibility factor portray themselves as “open-minded, perceptive, ingenious, and independent” (Gough, 1987, p. 34). They are interested in understanding why others behave as they do, and are open to considering viewpoints different from their own. High scorers may also display creative approaches to problem solving (Megargee, 1972). A measure of maternal flexibility was created for this study by normalizing and combining mothers’ scores on the five CPI sub-scales that define the Flexibility factor. An analogous procedure was carried out using fathers’ CPI responses to devise a measure for paternal flexibility. Descriptive data for these five subscales are presented in Table II.

Table II.

Descriptive Statistics for CPI Subscales

	Mothers			Fathers
	M	SD	Range	M	SD	Range
Flexibility Scale
Tolerance	23.45	3.97	13–30	22.39	4.28	6–30
Achievement via Independence	25.26	4.07	13–34	25.00	4.56	7–33
Flexibility	13.81	3.87	5–23	14.43	4.10	3–23
Intellectual Efficiency	30.97	3.95	20–40	30.95	4.43	18–38
Psychological Mindedness	16.84	3.23	9–24	17.42	3.27	10–24
Self-Control Scale
Responsibility	26.91	4.10	14–34	25.53	4.53	9–33
Socialization	32.20	5.70	13–42	29.88	4.87	13–39
Self-Control	23.15	5.29	9–34	21.94	5.45	2–33
Good Impression	19.02	5.31	6–32	19.18	4.70	6–30
Well-Being	31.08	4.58	15–38	32.37	3.65	21–38
Achievement via Conformance	29.70	4.10	16–37	28.40	3.72	17–35

The second CPI factor of interest is Self-Control. This factor has high loadings on the sub-scales for Responsibility, Socialization, Self-Control, Good Impression, Well-being, Tolerance, and Achievement via Conformance. These subscales are thematically related, in that they all assess the extent of the individual's identification with social norms and his/her capacity for self-control. People who rank high on this factor present themselves as “rule-favoring, rule-following, conscientious, and self-disciplined” (Gough, 1987 p. 33). They are perceived by others as “calm, mature, dependable people who are warm and responsive to others but in good control of their own feelings” (Megargee, 1972, p. 125). To construct a measure for maternal self-control for use in the present study, the seven above-mentioned CPI subscales composing the Self-Control factor were normalized and summed; a paternal self-control measure was created by similar means. Scores for these variables can also be found in Table II.

Quality of Coparenting Interactions Between Parents in Triadic Context

The quality of families’ coparenting interactions was assessed using the Coparenting and Family Rating System (CFRS), an observational coding system designed by McHale (1995; McHale, Kuersten-Hogan, & Lauretti, 2000). The CFRS, which is described in detail by McHale, Kuersten-Hogan, and Lauretti (2000), comprises a series of behaviorally anchored Likert-type rating scales with values ranging from 1 to 5 or 1 to 7. Seven of the scales in the coding system focus on the following aspects of mothers’ and fathers’ coparenting interactions: Cooperation, Competition, Verbal Sparring, Child- (vs. Adult-) centeredness, Couple Warmth (i.e., warmth within the marital dyad during whole-family interaction), Mother Endorses Father, and Father Endorses Mother. McHale, Kuersten-Hogan, Lauretti, and Rasmusseen (2000) report extensive reliability and validity data for the instrument. In the present study, satisfactory interjudge reliability was likewise obtained for the standard CFRS scales Cooperation, Competition, Verbal Sparring, Child-Centeredness, and Couple Warmth scales (with intraclass correlations ranging from .66 to .89). Intra-class correlations for the Mother Endorses Father and Father Endorses Mother scales, used for the first time in this study, were .94 and .73, respectively.

The CFRS also includes 7-point Likert-type scales for rating individual parent-to-child behavior during coparenting interactions. Behaviors assessed in this study were Mother–child and Father-Child Warmth, Mother's Involvement with Child, and Father's Involvement with Child. Intraclass correlations for these four scales ranged from .71 to .82. In earlier work, McHale (1995; McHale & Rasmussen, 1998) employed these indicators to construct scores for Total Family Warmth and Mutuality of Involvement (which was defined in past studies on a negative dimension, operationalized as the extent of the Parenting Discrepancy witnessed by the two adults. In this study, we created a Total Family Warmth variable by standardizing and summing scores for Mother–child Warmth, Father–Child Warmth, and Warmth between Parents. A Mutuality of Involvement variable was also created by calculating the absolute difference between mother's and father's Involvement scores and then reverse weighting this value (multiplying it by −1). Hence, higher scores on the resulting variable reflected greater mutuality in involvement (or smaller disparities in the two adults’ levels of involvement), whereas lower scores reflected less mutuality in involvement.

Descriptive data for the CFRS measures can be found in Table III To reduce these coparenting data, a factor analysis involving principal components extraction and varimax rotation was conducted with seven of the eight variables (Competition was dropped because it was not correlated significantly with any other variable that was rated). A two-factor description of the coparenting variables emerged, accounting for 57.7% of the variance in the scores. Factors were composed of variables loading .45 or above. Factor 1, which included high factor loadings for Cooperation (.81), Total Family Warmth (.77), Child-centeredness (.63), and Mutuality of Involvement (.48), was named Coparenting Harmony as it reflected greater of warmth and more cooperative, child-centered engagement during coparenting exchanges. A Coparenting Harmony composite (α = .80) was formed by standardizing and summing scores for each variable loading highly on that factor. Factor 2 included high factor loadings for Verbal Sparring (.83), Father Endorses Mother (−.78), and Mother Endorses Father (−.52). In that this factor implied verbal undermining and an absence of affirmation or mutual support between coparents, it was named Coparenting Negativity. A Coparenting Negativity composite (α = .78) was constructed by taking the inverse of the two Endorsement scores (multiplying them by −1), then standardizing and summing scores for the three variable loading highly on the Negativity factor.

Table III.

Descriptive Statistics for CFRS Variables

	M	SD	Range
Coparenting variables
Competition	2.99	1.31	1–5
Cooperation	3.27	1.10	1–5
Verbal sparring	1.77	1.30	1–5
Coparental warmth	2.92	0.92	1–5
Child-centeredness	3.80	1.22	1–5
Mother endorses father	3.71	1.00	1–5
Father endorses mother	3.98	0.53	2–5
Parenting variables
Mother–child warmth	4.88	1.08	3–7
Father–child warmth	4.50	1.22	1–7
Mother–child involvement	4.82	0.87	3–7
Father–child involvement	4.71	1.00	2–7

RESULTS

Infant Gender and Birth Order

To establish whether gender or birth order of participating infants exerted any main effect in this study, families of boys and girls were compared along all major constructs of interest (overall marital quality, mothers’ and fathers’ self-control, mothers’ and fathers’ flexibility, coparenting harmony, and coparenting negativity). Families with first- and second-born infants were also compared on this variable set. No significant differences as a function of child gender or birth order were obtained for any of these comparisons. Therefore, to maximize subjects-to-measures ratios, data were collapsed across categories of infant gender and birth order in the principal analyses of the study.

Marital Quality, Parental Self-Control and Flexibility, and Coparental Process

A series of hierarchical multiple regression analyses were conducted to test the hypotheses (a) that marital quality would explain a significant proportion of the variance in coparenting harmony and coparenting negativity and (b) that with the effects of marital quality taken into account, parental flexibility and self-control would explain additional unique variance in the coparenting indices of interest. For heuristic purposes, separate regressions were conducted for each of the four parental indicators, and for both of the coparenting indicators under consideration.

In each regression, the marital quality variable was entered on Step 1 and the individual parental adjustment variable on Step 2. Results of these analyses can be found in Table IV. As Table IV indicates, overall marital quality explained a statistically significant 12% of the variance in Coparenting Harmony and a statistically significant 21% of the variance in Coparenting Negativity. With variance owing to Overall Marital Quality accounted for, three of the four individual parent measures explained additional variance in Coparenting Harmony, but not Coparenting Negativity (Table IV). More specifically, Maternal Self-Control and Paternal Flexibility each explained an additional, statistically significant 9% of the variance in Coparenting Harmony, whereas Paternal Self-Control explained an additional 6% (p < .10).

Table IV.

Hierarchical Regression Mode Is Predicting Coparenting Harmony and Coparenting Negativity From Marital Quality and From Parent Adjustment

Models and predictors	R	R2	Change in R2	F change	Sig. change
Coparenting harmony
Step 1: Overall marital quality	.35	.12	.12	6.50	.01
Step 2: Maternal self-control	.46	.21	.09	5.20	.03
Coparenting harmony
Step 1: Overall marital quality	.35	.12	.12	6.50	.01
Step 2: Paternal self-control	.42	.18	.06	3.28	.08
Coparenting harmony
Step 1: Overall marital quality	.35	.12	.12	6.50	.01
Step 2: Maternal flexibility	.37	.14	.02	1.08	ns
Coparenting harmony
Step 1: Overall marital quality	.35	.12	.12	6.50	.01
Step 2: Paternal flexibility	.45	.21	.09	5.13	.03
Coparenting negativity
Step 1: Overall marital quality	.45	.21	.21	12.10	.00
Step 2: Maternal self-control	.46	.23	.01	0.75	ns
Coparenting negativity
Step 1: Overall marital quality	.45	.21	.21	12.10	.00
Step 2: Paternal self-control	.47	.22	.01	1.01	ns
Coparenting negativity
Step 1: Overall marital quality	.45	.21	.21	12.10	.00
Step 2: Maternal flexibility	.46	.21	.00	0.40	ns
Coparenting negativity
Step 1: Overall marital quality	.46	.21	.21	11.86	.00
Step 2: Paternal flexibility	.48	.23	.02	1.30	ns

Do Individual Parental Traits Moderate Marital-Coparenting Linkages?

To determine whether self-control or flexibility in mothers or fathers altered the association between marital and coparenting quality, interactions between marital quality and these individual parental adjustment indices were also examined (see Holmbeck, 1997, for relevant discussion of testing moderator effects). Separate regression analyses tested contributions of each of the four marriage-by-parent trait interactions (i.e., Overall Marital Quality × Paternal Self-Control, Overall Marital Quality × Paternal Flexibility, Overall Marital Quality × Maternal Self-Control, Overall Marital Quality × Maternal Flexibility) toward explaining Coparenting Negativity and Coparenting Harmony. Three of the eight regression analyses indicated significant interaction effects. Significant results are reported in Table V.

Table V.

Predicting Coparenting Harmony and Negativity From Marriage × Parent Interactions

Models and predictors	R	R2	Change in R2	F change	Sig. change
Coparenting negativity
Step 1: Marital quality	.46	.21	.21	12.10	.00
Step 2: Paternal flexibility	.48	.23	.02	1.30	ns
Step 3: MQ × PF	.56	.31	.08	5.33	.03
Coparenting harmony
Step 1: Marital quality	.35	.12	.12	6.50	.01
Step 2: Maternal flexibility	.37	.14	.02	1.08	ns
Step 3: MQ × MF	.48	.23	.09	5.48	.02
Coparenting harmony
Step 1: Marital quality	.35	.12	.12	6.50	.01
Step 2: Maternal self-control	.46	.21	.09	5.20	.03
Step 3: MQ × MSC	.54	.29	.08	5.25	.03

As Table V indicates, one of the significant interaction terms helped qualify the marriage-Coparenting Negativity association, and two helped qualify the marriage-Coparenting Harmony linkage. The nature of these findings is outlined next.

Paternal Flexibility Moderates the Marital-Coparenting Negativity Linkage

Figure 1 depicts the effect of marital quality on coparenting negativity at high (i.e., mean plus 1 SD) and low (i.e., mean minus 1 SD) levels of paternal flexibility. Slopes of these two lines were compared using procedures outlined by Aiken & West, 1991). Among the group of fathers low in flexibility, there was a significant, negative association between marital quality and coparenting negativity (b = −2.03, t = −3.87, p < .001). That is, among families in which fathers were relatively inflexible, coparenting negativity increased significantly as marital quality decreased. By contrast, among the fathers highest in flexibility, the association between coparenting negativity and marital quality was not significant (b = −.24, t = −0.48, ns.). That is, high paternal flexibility attenuated the relationship between marital quality and coparenting negativity in the manner anticipated. In families where fathers were highly flexible, lower marital quality did not bespeak a concurrent elevation in coparenting negativity.

Fig. 1.

Effect of overall marital quality on coparenting negativity as a function of paternal flexibility.

Maternal Flexibility Moderates the Marital-Coparenting Harmony Linkage

The maternal flexibility analyses followed the same procedure outlined above, with results depicted in Fig. 2. As Fig. 2 indicates, a significant, positive association between marital quality and coparenting harmony emerged only when maternal flexibility was high (b = 2.20, t = 3.28, p < .01). In other words, it was only when mothers were highly flexible that coparenting harmony suffered in the face of low marital quality, and appeared strong in the face of high marital quality. Among mothers low in flexibility, coparenting harmony did not vary significantly as a function of concurrent marital quality (b = .34, t = 0.71, ns). This set of findings ran counter to the expectation that high maternal flexibility would serve an ameliorative function; on the contrary, it was only low maternal flexibility that was consistent with the attenuation model.

Fig. 2.

Effect of overall marital quality on coparenting harmony as a function of maternal flexibility.

Maternal Self-Control Moderates the Marriage-Coparenting Harmony Linkage

Maternal self-control had a moderating effect analogous to that of maternal flexibility (Fig. 3). That is, marital and coparenting harmony were positively correlated only for families in which maternal self-control was high (b = 1.80, t = 2.83, p < .01). In contrast, among families in which maternal self-control was low, coparenting harmony did not decline systematically as a function of declines in marital quality (b = .12, t = 0.26, ns). As with maternal flexibility, and contrary to the hypothesis that high self-control by mothers would render coparenting harmony less vulnerable to disruption in the face of lower marital quality, it was only among mothers lower on the self-control dimension that the marital-coparenting link did not emerge.

Fig. 3.

Effect of overall marital quality on coparenting harmony as a function of maternal self-control.

DISCUSSION

Several findings from this investigation supported the contention that it is important for family researchers to take stock of individual parents’ adaptive resources when attempting to fully account for intrafamilial relational patterns. Among the families in this sample we found, consistent with both prior empirical evidence (e.g., Belsky, Crnic, & Gable, 1995; Floyd & Zmich, 1991; Lewis, Tresch-Owen, & Cox, 1988; Lindahl, Clements, & Markman, 1997; McHale, 1995) and clinical family theory (e.g., Haley, 1988; Kerr & Bowen, 1988; Minuchin, 1974), that overall marital quality was positively associated with concurrent coparenting harmony and negatively associated with concurrent coparenting negativity. But beyond this now well-replicated finding, we also found evidence that the individual adaptive traits of mothers and fathers helped to further clarify which families were struggling and which were faring relatively well.

First, we established that paternal flexibility made a significant contribution toward explaining variance in coparenting harmony, over and above the effects of marital quality. When fathers were highly flexible (i.e., highly perceptive, interested in entertaining others’ viewpoints, and skilled in adjusting their behavior to changing or unfamiliar interpersonal demands), the coparenting interactions in their families were more likely to be cooperative, child-centered, and characterized by warmth. One interpretation consistent with this finding is that highly flexible fathers were better attuned to their partners’ interpersonal cues and more adept at generating parenting behaviors compatible with partners’ initiatives with respect to the child. In addition, given their openness to alternative perspectives, they may have been more willing, as well as more able, to adapt themselves to their partners’ visions.

Such adaptability in fathers would be regarded as especially conducive to coparenting harmony when viewed from the perspective of the maternal gate-keeping hypothesis (e.g., Allen & Hawkins, 1999; DeLuccie, 1995; Haley, 1988; Wamboldt & Reiss, 1989): According to this premise, mothers are typically more influential than fathers in structuring or orchestrating family processes, especially when their children are very young. This pattern emerges, so the argument runs, because mothers tend more often to be the primary caregivers of their children. Therefore, they have more intimate ties with their children and attain greater skill in childcare than do fathers. In the present sample, mothers did devote significantly more hours per week to care of the infant than did fathers (28 for mothers as opposed to 11 for fathers, F(1, 98) = 5.86, p < .001). If as the maternal gate-keeping hypothesis proposes, this disparity rendered mothers more likely to take the lead in family interactions, then it becomes clear how fathers’ readiness to accept their partners’ guidance would facilitate cooperation and mutual engagement in coparenting.

Second, we found that maternal self-control, like paternal flexibility, also made contributions to coparenting harmony over and above the effects of marital quality: Data indicated that mothers who were highly conscientious, warmly responsive to others, and disciplined in their expression of their feelings were more likely to achieve high levels of collaboration, to fostering whole-family warmth, and to attain a balanced and shared focus with their partners when parenting the infant together. We speculate that mothers with high self-control may have found it easier to accommodate fathers’ preferred roles in the family triad. That is, when fathers displayed interest in being more involved, mothers high in self-control may have afforded them the space to do so, reinforcing the father's participation. In addition, mothers with high self-control may have been more likely to show skill in inhibiting impulses to exclude the father when his initiatives with the child did not correspond to their standards or preferences.

Concerning our questions about parental characteristics as moderators of the marital-coparenting connection, we found that paternal flexibility did moderate the association between marital quality and coparenting negativity in the manner predicted. We had hypothesized that high levels of flexibility, affording parents insight into the impact of overt conflict on children, might provide parents with incentives for curbing their verbal expressions of hostility in the family triad and thus limit the extent to which coparenting negativity escalated in conjunction with declining marital quality. In this sample, coparenting negativity did not vary significantly with changes in marital functioning among those families in which fathers were highly flexible. That is, for this subgroup of families, deterioration in partners’ marriages was not accompanied by increases in negative, sarcastic verbal exchanges in the coparenting context. On the other hand, among families where fathers were less able to adopt other points of view or to adjust their behavior to accommodate changes in interpersonal context, decrements in marital quality were linked with increases in coparenting negativity.

In contrast with these father data, the story told by the maternal data was not the one we had forecast. Although maternal flexibility moderated the link between marital quality and coparenting harmony, it did so in a manner counter to predictions. We found the correspondence between marital quality and coparenting harmony to be more pronounced when mothers were high in flexibility than when they manifested this trait to a lesser degree. That is, lower marital functioning accompanied lower levels of interadult cooperation, warmth, child-centeredness, and mutuality during triadic interactions with the infants more reliably among mothers high in flexibility. In families where mothers were less flexible—that is, less perceptive, less open to considering alternative viewpoints, and less fluid in adapting their behavior to changing interpersonal demands—coparenting harmony did not show the anticipated systematic declines as a function of decreasing marital quality.

To account for this surprising finding, we draw upon two other themes that have been developed elsewhere: the maternal gate-keeping hypothesis outlined above, and the notion that, other things being equal, fathers make themselves emotionally available in the triad to the extent that their marriages are functioning well. When men are maritally dissatisfied, they may signal affective disengagement from family interaction; when they are content with the marriage, they may communicate greater accessibility and warmth. If these premises have some merit, then the seemingly paradoxical implications of maternal flexibility are rendered somewhat more interpretable.

By definition, a mother high in flexibility is adept at reading interpersonal cues, including those of her partner; moreover, she smoothly adjusts her own behavior in response to these cues. In her case, then, the extent to which she recruits the father to join her in harmonious coparenting will depend crucially on her assessment of the father's willingness to work with her in this capacity. A highly flexible mother in a troubled marriage may be acutely sensitive to signs that her partner's demeanor toward her in the family triad is being influenced by hostile or distressed feelings originating in the marital relationship. If she detects such signs, she may anticipate that her efforts to promote coparenting cooperation may evoke the father's irritation and precipitate conflict. For this reason, she may leave the father to his own devices in the family triad, with the result that his involvement in coparenting is diminished, and coparenting harmony necessarily declines. On the other hand, a flexible mother in a high-functioning marriage may perceive that her partner's marital satisfaction renders him emotionally available to her. She may therefore see him as open to following her lead in the domain of coparenting. Acting upon this perception, the flexible mother may rapidly and effectively adjust her own behavior in the triad so as to capitalize on the father's motivation, integrating him fully into coparenting interactions.

Contrast these patterns with those of mothers low in flexibility. Such individuals may be relatively insensitive to their partners’ affective displays in the triad, failing either to perceive them accurately or to alter behavior in response to them. As a result, mothers low in flexibility may not take the father's level of emotional engagement into account as they attempt to structure triadic process. Instead, the way in which they try to define the partner's role may depend primarily on their own sense of how fathers ought to be taking part in coparenting interactions. Thus, even when a maritally dissatisfied father is signaling emotional distance or lack of interest, the relatively inflexible mother may sometimes seek, and occasionally even elicit from him, a higher level of behavioral involvement than a more flexible mother would attempt to evoke. On the other hand, when a father content with his marriage communicates emotional availability in the triad, the inflexible mother may have difficulty accommodating her behavior to the father's participation in such a way as to optimize the quality of their coparenting cooperation.

We propose a similar explanation for the related finding concerning the moderating effect of maternal self-control on the relationship between marital quality and coparenting harmony: When a maritally distressed mother is high in self-control, her attitude of altruism and self-sacrifice may lead her to attach great importance to her partner's preferences, and her conscientiousness may lead her to regard parenting as primarily her responsibility. Thus, she may readily acquiesce to the father's lack of involvement in the family triad. Also, in comparison to mothers lower in self-control, she may be more intent on compensating her child for what she sees as the father's deficient investment in the family. In so doing, she may focus on the child in such a way that she unintentionally blocks the father's participation. On the other hand, when marital quality is high and the father is ready to engage, the mother with high self-control may be better equipped to respond positively to her partner by integrating him into coparenting interactions, because she may have more patience with his relative inexperience with the child.

By contrast, the mother low in self-control may be less willing to cater to her partner's preferences in the triad, and so more or less deliberately disregard his signals relating to his level of emotional availability. Thus, even a maritally dissatisfied father whose demeanor conveys disengagement may find himself recruited into more active coparenting by a partner low in self-control. On the other hand, a mother low in self-control may be just as unwilling to alter her agenda for the happily married father who wishes to be involved in closer coparenting collaboration. In consequence, the level of coparenting harmony achieved by some such couples may be lower than that which would be predicted solely on the basis of their marital functioning.

We recognize that these interpretations are ad hoc and highly speculative, and that the findings themselves demand replication. We also call attention to a number of study features that further qualify the results obtained. A major drawback of the data in this study was in how we estimated the moderating individual adjustment variables. Recall our proposition that parental flexibility and self-control would give rise to cognitions favorable to coparenting. A question raised by the investigation's counterintuitive findings is the extent to which the traits measured actually did promote such positive cognitions about coparenting among our study participants. Because we did not assess ideas about coparenting directly, the relationship of such ideas to parental flexibility and self-control could not be explicitly examined. A logical next step in future work on this topic would be pursuit of such an assessment. Second, we relied solely on a self-report personality inventory as the means of evaluating parental flexibility and self-control. Notwithstanding the CPI's distinguished track record, a potentially more valid approach would be to determine whether parents actually manifest signs of these traits in their patterns of thinking as they reflect on significant relationships.

The current investigation was further limited by its use of global, clinical measures to assess dimensions of coparenting. Although we assessed coparenting dimensions globally, we subsequently proposed several possible stances and maneuvers that parents with different personality profiles may call upon to different degrees when coparenting. To make further progress in clarifying the mechanisms through which marital quality is linked to coparenting quality, microanalytic approaches are needed to document key behavioral exchanges between partners as they coconstruct their coparenting interactions (see, e.g., Fivaz-Depeursinge, Frascarolo, & Corboz-Warnery, 1999; McHale & Alberts, 2003; van Egeren et al., 2004).

Finally, we underscore that the data from this study are correlational, and as such, do not permit decisive statements to be made regarding the direction of effects. In the foregoing discussion, we drew heavily upon the language of influence in proposing how individual parental adjustment might have influenced coparenting, and here we emphasize that we did so for heuristic purposes only. The results obtained are also consistent with the conclusion that personality traits in parents may shift in response to developments within the coparenting alliance (see McHale et al., 2004).

In conclusion, this study provides evidence that the individual personality traits of paternal flexibility and maternal self-control may be important contributors to a harmonious coparental alliance, that the contributions of these personality features are not simply redundant with variance captured by indicators of marital quality, and that in some circumstances parent personality traits may help alter connections between marital and coparenting systems. Especially noteworthy was the finding that high paternal flexibility neutralized the negative impact of declining marital quality on coparenting negativity, providing some support for the contention that some parents’ individual strengths may help them to “beat the odds” and limit the extent to which marital difficulties create disturbances in the realm of coparenting. This finding, along with the intriguing findings concerning maternal flexibility and self-control as moderators, call for replication studies. Taken together, they underscore the point that when models of family dynamics fail to include consideration of family members’ psychological characteristics, they may prove to be incomplete and even misleading.

The family theorists’ axiom that “the whole is greater than the sum of its parts” has undoubtedly had a salutary effect on empirical family studies, stimulating investigators to conceptualize triadic family interactions in more complex, comprehensive, and useful ways; the growing literature on the coparenting construct attests to this fact. Notwithstanding the importance of holism to family research, the results of this study suggest that it may now be appropriate and useful to focus more intensively on the study of individual family members and the contributions of their psychological traits to the creation of coparental and whole-family process.