The exploration of the shared work of childrearing has now become a central focus for both scholars and professionals. Most commonly—but not exclusively—referring to parents, coparenting has been defined as two or more adults engaging in the shared activities and responsibilities of raising a child (McHale and Lindahl 2011). Coparenting theory centers around the notion that children are nurtured and socialized within a family relationship system that is often comprised of multiple primary caregivers (McHale and Lindahl 2011). Across numerous studies and several meta-analyses, there is consistent agreement that coparenting is distinct from the marital-relationship, the parent-child relationship, and the whole family system (e.g., Margolin et al. 2001; McHale 1997; Teubert and Pinquart 2010; Van Egeren and Hawkins 2004) and that it particularly affects parenting, severing the link between conflict – for example- and harsh, less effective parenting by creating a buffer zone by parental collaboration for purposes of child rearing (Pruett et al. 2017, 2019).

Coparenting is a distinct and an irreducible part of the parents’ relationship, practiced across the entire spectrum of relational possibilities. The family form in which two or more adults become parents to a child without ever sharing a household or a romantic relationship presents a family configuration in which coparenting is the preferred arrangement for children’s development and well-being (Luxton 2011), beyond a specific arrangement negotiated in the context of parental separation. At the same time, new research has been responding to the social trend of increased father involvement with their children to support or improve fathers’ parenting skills and to increase their involvement with their children in ways that promote children’s positive development (Pruett et al. 2017; Well and Exner-Cortens 2016). In addition, the emergence of the modern family has been growing increasingly incompatible with previous notions of distinct and rigid gender-based ideals of fatherhood and motherhood (Luxton 2011). Within most western cultures, parents can be single, married or separated; sexually and gender diverse; an adoptive or step-parent; and can be influenced by multicultural influences in how they promote children’s emotional development and well-being. In this context, we argue that coparenting should be conceptualized and measured in a way that is decoupled from gender-role and family-structure assumptions. Coparenting is then conceptualized and measured as specifically referring to the negotiated activities and relational aspects of two or more parents or caregivers.

The importance of coparenting for child development is well documented in social science literature. Research consistently indicates that ongoing parenting discord is associated with poorer coparenting, poorer quality parent-child relationships, and puts children at risk for behavioral and emotional issues (Cabrera et al. 2012; Cummings et al. 2010). In two-parent families, lower marital satisfaction is related to poorer coparenting (Christopher et al. 2015), coparenting predicts parenting quality—especially for mothers (Le et al. 2016), and healthier coparenting is related to better child outcomes (McConnell and Kerig 2002; Stright and Neitzel 2003). Similarly, in separated families, developing cooperative coparenting is considered an optimal goal to facilitate post-divorce/separation parenting and child adjustment (Pruett and DiFonzo 2014). Although there is growing awareness of the coparenting differences between separated and intact families, studies have not compared coparenting across family structures, nor evaluated how differences related to structure mediate associations between coparenting aspects and child well-being.

Unlike the construct of parenting styles, which often describe maternal or parental parenting practices in dyadic relation to the child, coparenting aims to assess the (at least) triadic parental interplay in childrearing (Teubert and Pinquart 2010), without presumptions about maternal and paternal (gender-specific) roles and styles of parenting. Conceptually, specific coparenting processes undoubtedly intersect with broader marital relations when romantic, financial, emotional, and companionate relational aspects are considered (McConnell and Kerig 2002), but theorists treat coparenting and its underlying dimensions as if they can be isolated and studied in specificity.

Margolin et al. (2001) outlined three broad domains of coparenting that appear in the literature with varying degrees of overlap: Conflict surrounding parenting issues; Cooperation involving the extent to which parents respect and value each other’s choices as parents and their support for the other’s parenting, and Triangulation, specifically in its negative, conflict-oriented reference to parent-child coalitions that undermine the other parent or blur parent-child boundaries. This last domain appears particularly salient in regard to coparenting in separated/divorced families (Madden-Derdich, Leonard and Christopher 1999). In an attempt to further detail the components of coparenting, Feinberg (2003) postulated four areas comprising a multi-domain model: childrearing agreement/disagreement, division of labor in childrearing, support and undermining actions between co-parents, and joint family management of interactions. Additional dimensions of coparenting suggested in the literature include coparenting support (Van Egeren and Hawkins 2004), teamwork (Pruett and Pruett 2009), conflict (Cabrera et al. 2012; Dorsey et al. 2007) and coparenting alliance (Hock and Mooradian 2012).

A recent literature review and meta-analysis of coparenting research by Teubert and Pinquart (2010) proposed a comprehensive assessment of coparenting based largely on the models of both Margolin et al. (2001) and Feinberg (2003) and found that major cross-cutting categories included parental cooperation, child rearing agreement, conflict, and triangulation.

McHale and Irace (2011) further proposed that the domains crucial to positive and effective coparenting structures include factors relating to the coparents’ mutual understanding, communication, and coordination about the child; level of trust, each parent’s support for the other parent’s efforts and the parents’ ability to successfully resolve disagreements about the child’s best interests. Thus, there remains no agreement about the primary domains that are considered fundamental to the definition of coparenting and which of these domains actually influence child well-being and adjustment (Baril et al. 2007), which suggests further research is needed to establish the core domains of coparenting to further distinguish it from other family influences on child development.

Given the shifting family structures that separating families face, the coparenting relationship often contains additional challenges for families involved in couple relationship breakdown. Conceptual efforts have emphasized the need to consider both parents, their relationship post separation and the sharing of responsibility for the raising of children (Austin 2012; Ganong, Coleman and McCalle 2012; McBride and Rane 1998; Pruett and Pruett 2009), and as a bidirectional process of influence and control over the other parent’s involvement (Adamson 2010; Trinder 2008). Changing boundaries within the parental relationship can lead to ambiguity involving roles and communication and can undermine parents’ and children’s perceptions about who is included and excluded in the family, further complicating roles and expectations (Pruett and Donsky 2011; Saini 2012). Kelly and Emery (2003) found that children of divorced parents whose parents involved them in the conflict showed higher levels of depression and anxiety compared to parents who did not involve their children in such interactions. In contrast, Teubert and Pinquart (2010) found no stronger associations between coparenting conflict and child adjustment than did intact families, emphasizing the need to consider both similarities and differences among intact and separated coparenting.

The increase in sexually and gender diverse families further argues for considerations of coparenting that are non-binary, including planned gay and lesbian families; gay and lesbian stepfamilies, which include a heterosexual parent from a prior relationship; families with three or more identified parents; and families where parental figures have no legal ties to the children (Johnson, O’Connor and Tornello 2016). Psychological research has lagged behind social changes regarding intact and separated sexually and gender diverse coparents (Johnson, O’Connor and Tornello 2016) and there remains no questionnaire that assesses coparenting relationships within these various parenting constellations Johnson and O’Connor (2002) found that lesbian couples who were co-parenting perform as well or better on measures of family relationship quality and parental attitude variables than heterosexual couple co-parenting counterparts. But further research is needed to consider implications of coparenting among intact and separated sexually and gender diverse parents.

While researchers are learning more about the impact of coparenting on child well-being, in order to advance understanding in this field, there is a need for valid and reliable coparenting measures. There are psychometrically robust measures available for either intact or separated coparenting (Feinberg et al. 2012; Feinberg and Kan 2008; McHale 1997; McHale et al. 2008; Teubert and Pinquart 2011), yet there remains no coparenting instrument that can measure the general coparenting construct across intact and separated multi-domain coparenting relationships, to assess if there are consistent coparenting qualities that are universal, at least in some contexts.

McHale’s (1997) 16-item Coparenting Scale was designed to capture both overt and covert coparenting processes in married couples. It yielded four factors relating to family integrity, disparagement communications, conflict in the child’s presence and co-parental disciplinary activities. Husbands’ and wives’ reports of their coparenting behavior were significantly and positively correlated for each factor (McHale 1997). The focus of the instrument is parental behaviors that support or undermine children’s sense of family; two factors are related directly to the children, and two are more parentally attuned. In later research, McHale focused on conflict and cohesion, as well as observational indices collected when parents are in the presence of their infant (McHale et al. 2008).

Feinberg and colleagues developed scales to support interventions with couples (Feinberg and Kan 2008; Feinberg et al. 2010, 2012). Three scales assess Coparental Support, Parenting-Based Closeness (feeling close to the other parent as a result of his/her parenting), and Coparental Undermining in his Family Foundations program (Feinberg and Kan 2008). Sub-domains of Support and Undermining emerged as major tenets of his subsequent research, and his Coparenting Relationships Scale (CRS) provides an overall score for quality coparenting as well as scores for seven subscales based on four domains (childrearing agreement, support/undermining, satisfaction with the division of labor, and family management) (Feinberg et al. 2012). CRS demonstrated excellent internal consistency, ranging from .91 to .94 across gender and data collection time points. Internal consistency for the seven subscales were more varied, but generally strong, ranging from .75 to .90 for both men and women for the Coparenting Closeness, Exposure to Conflict, Coparenting Support, and Coparenting Undermining subscales. The internal consistency of Endorsement of Partner Parenting was strong for women (.83–.88), but weaker for men (.61–.71). The internal consistency of the Coparenting Agreement was also weaker but still acceptable (.66–.74). The Division of Labor subscale contained only two items, thus internal consistency was not reported (Feinberg et al. 2012). Although CRS may provide useful information about coparenting across diverse sociodemographic backgrounds, levels of risk, and stages of family development, the measure is designed for and germane to intact families.

Teubert and Pinquart’s (2011) Coparenting Inventory for Parents and Adolescents (CI-PA) is a comprehensive measure that effectively integrates and advances previous coparenting measures and utilizes both parents’ and adolescent self-reports, but it too is limited to intact families. Convergent validity of the CI-PA showed high correlations between mothers’ and fathers’ reports about their coparenting dyad (r = .56) and weaker correlations between adolescent and parental ratings about the coparenting dyad (r = .35). CI-PA diverges from most of the existing measures, which focus on younger children. Cooperation, conflict, and triangulation rose to the forefront of their inquiry, with their analyses confirming that all three aspects were salient dimensions of coparenting with acceptable to high internal consistency (α > .70).

For couples who are not living together, Conflict is the central theme in a number of conceptually overlapping instruments. Other primary indices in these various measures include communication, hostility, triangulation, cooperation, respect for the other parent, gatekeeping, support, and cohesion (e.g.s., Ahrons 1981; Amato and Rezac; 1994; Austin et al. 2013; Kitzmann et al. 2003; Maccoby, Depner, and Mnookin 1990; Madden-Derdich and Leonard 2000; Mullett and Stolberg 1999). Many of the instruments focus on both positive and negative parental behaviors; Mullett and Stolberg (1999) investigate from the child’s perspective, as well.

There has been little consistency in the field concerning the variation within and across contexts of coparenting, and assessments remain unintegrated across studies. Global terms (e.g., positive vs. negative coparenting) are not used consistently among researchers. Thus, there is a need to clarify the nature and measurement of coparenting across family configurations.

In sum, although various coparenting measures exist, none serve as a measure applicable across family structures and free of gender-role assumptions for use by researchers as well as agencies and programs. We are trying to develop a measure that is adequately multi-dimensional to capture the most important underlying factors of this complex construct. To move away from siloed assessment tools for married and separated coparenting partners and across family structures, we aim to develop an instrument that could be valid with families in which the parents live together as well as those that live separated, for whatever reason and in any configuration. A salient reason for this is that in longitudinal research, couples move between these structures over time. Also, we aim to be more inclusive than previous research has tested for in terms of sexually and gender diverse families, but in this first iteration of instrument development, we focus on family structure due to sampling limitations. The objective of this pilot study was to test the construct validity as well as the internal consistency of the newly constructed Coparenting Across Family Structures in a sample of parents with minor children still living in the home. No additional information was garnered about the children for this pilot.

Method

Participants

Participants were 252 parents (81.7% female; 18.3% male) who completed the online coparenting survey on SurveyMonkey (https://www.surveymonkey.com). Of these, 219 participants (87%) completed all items. Although we attempted to recruit a diverse sample, the majority of the participants self-identified as Caucasian (71.8%), were highly educated (64.3% completed schooling beyond college), employed full-time (70.2%), and reported annual incomes over $80,000 (73%) (see limitations section).

Most of the parents identified as their youngest child’s biological parents (88.5%), while the remaining reported being a step-parent, adoptive parent, or legal parent. Three-quarters of the parents reported having either one child (40%) or two children (36.1%) under the age of 18, with the other participants having three children (10.3%), four children (2.4%), or five children (1.2%).

Over half of the participants were living together with the other parent, either married or in common law relationship (57.9%), a third was separated or divorced (33.3%), and the rest were living together but not married or in a common-law relationship (2.8%). On average, participants had been in a relationship with the other parent for eight or more years (76.6%), whether or not they were separated or together at the time of completing the survey.

Procedures

This research protocol was approved by the Smith College’s IRB Committee. Parents were recruited through multiple websites and parenting blogs, as well as the membership list of the Association of Family Conciliation and the Courts, a multidisciplinary and international organization with over five thousand legal and mental health professionals. An online survey was offered to parents or professionals who were parents themselves, but also was forwarded or shared with their client and personal contacts. Decision to participate was private and anonymous. The inclusion criteria included: (1) a parent with a child under the age of 18 years of age at the time of completing the survey; (2) a parent who shared parenting in some capacity with at least one other parent; and (3) a parent able to read English in order to complete the survey.

Measure

Based on the review of the literature and current measures for coparenting (Feinberg et al. 2012; Feinberg and Kan 2008; McHale 1997; McHale et al. 2008; Teubert and Pinquart 2011), we initially created a list of all items from previous measures and new items developed based on a thorough coparenting literature review. We then deleted repetitive items and edited the remaining list to ensure the language was gender neutral (e.g. “the other parent” instead of “mother or father”) so the items could be applied across family structures (e.g. not including items that were specific to marital relationships). This process created a 56-item scale that captured nine dimensions of coparenting identified in the literature, each with sub-categories that further assessed the main dimension: (1) Communication (2) Sharing (3) Anger (4) Restrictive coparenting; (5) Facilitative coparenting (6) Respect; (7) Trust; (8) Conflict and (9) Valuing. To explore face validity, we provided the tool to colleagues in the fields of psychology, social work and family law for their initial feedback and we made changes to the wording of items based on this feedback.

For each item, respondents were asked to rate from 1 (strongly agree) to 5 (strongly disagree) with a coparenting statement. Examples include: I usually just give in to the other parent so we do not argue; We can usually find solutions about parenting that we are both happy with; I get annoyed easily about the mistakes that the other parent makes with our child; The other parent undercuts my decisions; We have similar hopes and dreams for our child; We generally agree on how to discipline our child; or Although we don’t always agree, we respect each other’s differences as parents.

The overall score for each subscale was calculated by a simple non-weighted addition of the score on each of the composing item and the scale overall score was calculated by a simple non-weighted addition of the scores on each of the composing subscales.

Data Analyses

The psychometric properties of the CoPAFS scale and the 9 subscales composing the measure were calculated on the basis of the total sample. The internal consistency of each of the 9 subscales and the CoPAFS scale as a whole, expressed as a Cronbach’s alpha coefficient, was measured as an indicator of the internal consistency of each subscale and the total CoPAFS scale.

Confirmatory factor analyses (CFA) were conducted with a maximal likelihood estimation method for each of the 9 subscales. Confirmatory factor analysis assesses how well the measurement model underpinning the scale captures the covariance between all of the items that compose each of the 9 subscales, or in the case of the whole CoPAFS scale, how well the measurement model captures the covariance between the 9 subscales. The analysis also estimates the regression coefficients for each item or subscale and the proportion of the variation of each item or subscale predicted by the model. The CFA hence provides information about the construct validity – to what extent does the subscale or scale actually measure what it was intended to measure. Following Kline (2016), the model fit indices calculated and reported were first a chi-squared test indicating the difference between observed and expected covariance metrics. The P value of the Chi-squared test should be above .005 (not significant). However, as this is strongly influenced by sample size, this may be misleading in either small samples (leading to accept an inappropriate model) or large samples (leading to reject appropriate models). Kline (2016) suggests that sample size above 200 cases may result in non-significance even when the model is appropriate. The second type of model fit indices calculated and reported was the root mean square error of approximation (RMSEA), which measures the discrepancy between the hypothesized model, with optimally chosen parameter estimates, and the population covariance matrix. Thirdly, the root mean square residual (RMR) shows the square root of the discrepancy between the sample covariance matrix and the model covariance matrix. The Goodness of fit index (GFI) is a measure of fit between the hypothesized model and the observed covariance matrix. The normed fit index (NFI) analyzes the discrepancy between the chi-squared value of the hypothesized model and the chi-squared value of a null of baseline model in which all the variables are assumed to be uncorrelated. Comparative fit index (CFI) analyzes the model fit by examining the discrepancy between the data and the hypothesized model, while adjusting for the issues of sample size inherent in the chi-squared test of model fit.

The variable “What is your current relationship status with your youngest child’s other parent”, originally coded as a nominal scale of 4 options (living together married or common law/living together not married or common law/separated or divorced/no relationship) was recoded into a dichotomous variable (separated/cohabiting) and the data file split accordingly to test for the stability of the psychometric properties across family structures while accounting for sample size limitations.

Although it was our intention to explore gender differences and gender orientation differences, the lack of diversity in the sample collected restricted that possibility for this pilot; hence we focus on family structure subsamples only.

Results

Internal consistency of full scale and its composing subscales

Table 1 provides the Cronbach’s Alpha Coefficients for both the total measure and each of the nine dimensions; only the Facilitative Coparenting dimension was below the .70 cutoff. Internal consistency was stable across separated parents and parents who cohabited. The internal consistency of the Sharing and the Conflict subscales was slightly more sensitive to whether the parents were separated or cohabiting, though remaining above the cutoff of .7.

Table 1 Cronbach’s coefficients for the CoPAFS scale and subscales
Full size table

Scale and subscale intercorrelations

Table 2 depicts the scale and subscale intercorrelations for the total scale and each of the nine dimensions. All were significant and ranged between .641 (Facilitative coparenting and Conflict) and .952 (Respect and the CoPAFS as a whole).

Table 2 Scale and subscales intercorrelations
Full size table

Factor structure

Figure 1 presents the confirmatory factor analysis of the measurement model underlying the CoPAFS scale. As can be seen in Table 3, the Chi-Squared test indicating the difference between the observed and predicted covariance metrics was significant; for a good model fit, the P value should be above .05, representing non-significance and the test result itself as close to 0 as possible for a sample size > than 200 (Kline 2016). Excellent model fit was indicated by the Goodness of Fit Index (GFI), the Normal Fit Index (NFI) and the Comparative Fit Index (NFI), which were all above their respective cutoff points. At the same time, the Root Mean Square Error of Approximation Measure (RMSEA) which should be below .06 in order to indicate a good model fit, and the Root Mean Square Error Measure (RMR), which should be below .08 in order to indicate a good model fit, were both slightly above the required cutoff point. Taken together, and given that all estimates for the 9 subscales were significant and ranged between .951 and .747, explaining between 55.9% and 90.5% of the variation on the CoPAFS scale, the measurement model underlying the CoPAFS scale displayed an acceptable and even good - albeit not outstanding - model fit on the basis of this pilot survey sample. The model fit indices were stable across family structure (separated/cohabiting). The model fit for all subscales except Anger was acceptable to excellent, and the model fit indices were generally stable across family structure. Model fits for each subscale with statistical details are presented in tables 4-12 as supplementary information.

Fig. 1
figure 1

CoPAFS measurement model confirmatory factor analysis. Note: Numbers over the arrows indicate the regression coefficient and the number above the rectangles indicate the proportion of the variation of the subscale predicted by the underlying construct

Full size image
Table 3 Model fit indices and subscale estimates for the CoPAFS scale measurement model
Full size table

Discussion

Despite the importance of coparenting relationships for child development (Cowan and Cowan 2000; Feinberg 2003; McHale 2007; Pruett et al. 2017), reliable and valid measures for assessing coparenting relationships across relationship dynamics have not been identified. The Coparenting Across Family Structures is based on a comprehensive review of the social science literature and an expansion on earlier works that have explored coparenting dimensions (e.g., Ahrons 1981; Feinberg 2003; Margolin et al. 2001; McHale 1997; Teubert and Pinquart 2010). The goal of the Coparenting Across Family Structures was to create a valid measure with good internal consistency that would be applicable across family structures and gender. Beginning with concepts derived from a review of the empirical literature, a 56-item instrument was developed and tested.

We argue that coparenting should be conceptualized and measured without any implicit assumptions about gender or family structure, so as to crystalize and capture the bare practice of two or more adults parenting a child. While our sample did not have sufficient power for testing stability across gender, we were able to test that the internal consistency and factor structure of the CoPAFS scale and composing subscales were stable across two types of family structures, one in which the parents live together and one in which the parents are separated. Our exploration of gender neutrality or specificity awaits the next step in our research.

Coparenting has been suggested to reflect the importance of actively and deliberately encouraging the relationship of the other parent with the child. However, these results show that facilitative coparenting (e.g. encouraging the relationship with the other child) predicted the least amount of variance in coparenting (56%), which suggests that the facilitative coparenting role and function needs further delineation and exploration. It also is possible that the general construct is less useful than specific behaviors such as sharing or valuing of the other partner.

The subscale Anger, though itself predicting 74.7% of the variation on the coparenting construct, did not have good model fit and though all its composing items significantly predicted its variation, the proportion predicted varied considerably between the items. The item “I feel out of control when I am speaking to the other parent” predicted 80.3% of the variation. Further testing is needed in order to see where this subscale needs be fine-tuned, although its importance emerges here as it has across family research areas.

The only item that was not significantly predicting variance on its subscale (P = .068), only predicting 1.9% of the variation, was the item “it is important that our child does not hear us talking negatively about each other” which is part of the Valuing coparenting subscale. Though further testing is needed, this item indeed may not be indicative of an adherence to coparenting. Negative interaction between the parents may be too broad and vague a category, and a more specific phrasing may prove better.

Coparenting as a construct reflects an expanded theoretical understanding of parenting partnerships as both bi-directional and multi-dimensional. They are bi-directional because they both affect the parents and are affected by it (see McHale 2007). Sharing parenting is predicated on the specifics of positive support and cooperation, whether parents that are together or trying to coparent without further involvement in each other’s lives, and the value that is placed on the importance of the other parent in the lives of the children.

The difficulty of capturing the dimensions of coparenting and replicating them across samples speaks to the need for larger and more representative studies. Although several domains of coparenting were captured, the findings point to the possibility of capturing coparenting by a singular dimension. This may be particularly useful for agencies trying to capture a rough estimate of parents’ values towards and potential to work together in childcare. That is not to say that the nuances inherent in the items and factors in our own and others’ research are not an important part of the assessment process that enables us to best understand and intervene in coparenting dynamics when they are strained. Being able to label less optimal coparenting relationships will enable the development of more successful and appropriate intervention programs that can resolve maladaptive coparenting functioning as early in family formation or separation as possible. Further understanding of the concept of coparenting can lead to more targeted interventions and more positive outcomes for children and families by assessing the contribution of coparenting relationships to overall adjustment.

Limitations

The obvious limitation to this study is that despite efforts to obtain a representative sample, the initial iteration of this instrument utilized a convenience sample that scored high on coparenting (mean 223/280) and was populated by middle class professionals and more women than men. The strong participation by Association of Family and Conciliation Court (AFCC) professionals and clients, an organization dedicated to reducing family conflict, may have biased the sample due to their professional involvement in questions of coparenting in the context of divorce and separation. The scale was designed to be inclusive of sexually and gender diverse family structure. However, the stability of face value and factor structure across family structures is yet to be tested with a sufficiently diverse sample. Given that the majority of the sample was from the United States, difference in coparenting across national as well as cultural borders remains unknown, and so caution is needed when generalizing the findings to other samples. Another limitation is the distinct groups of marital status used in the study. There may be similarities between the ways couples coparent and separated partners coparents once the break-up period has passed. Our subsequent research will consider coparenting across couple relationships over time.