Research That Matters (January 17 - 20, 2008)

Purpose. Co-parenting, a relatively new construct focused on dyadic processes specific to parenting, may help us better understand the relationship among marital dynamics, individual parenting, and child outcomes in family foster care. To date, co-parenting has not been studied in foster parents and no measure of foster co-parenting exists. Further, almost no research has been conducted on foster couples, although they care for the majority of children in family foster care. Studies of other mother-father constructs suggest that women and men often have different conceptualizations of constructs. Therefore, it is particularly important that measurement equivalence be addressed as it is seen as important for substantive interpretations and is considered a logical and necessary precursor to tests of construct validity. This presentation will report foster couples' measurement invariance of the Casey Foster Applicant Inventory-Applicant – Co-parenting Scale (CFAI-CP), a new scale developed to measure foster parents' co-parenting potential. Methods. In the largest sample of foster fathers ever collected, licensed foster couples (N = 111) from 29 states completed the CFAI-CP as part of a larger study of foster families. The CFAI-CP has 10 items rated on a 4-point scale ranging from strongly disagree (1) to strongly agree (4). Measurement invariance was evaluated statistically in CFA on nested model differences using dyads as the unit of analysis. A series of increasingly restricted models assessed (1) the basic CFA model, (2) configural invariance, (3) metric invariance, (4) scalar invariance, (5) invariance of unique variances, (6) invariance of factor variances, (7) invariance of factor co-variances, and (8) invariance of factor means. Invariance was tested using Mplus (Muthén & Muthén, 2006). Mplus was used because it can accommodate ordinal-level data. Further, Mplus is the only program that utilizes the robust weighted least squares (WLSMV) estimator. This estimator has produced accurate test statistics, parameter estimates, and standard errors of CFA models under a variety of conditions and was more appropriate than weighted least squares (WLS) for the characteristics of the variables and sample in this study. Results. CFA strongly supported a 10-item, one-factor Co-parenting construct for foster mothers and fathers. Internal consistency reliability analyses results were excellent (α = .89 for mothers; α = .88 for fathers) and the dyadic CFA model had excellent overall model fit: CFI = .97 and WRMR = 1.08. Measurement invariance testing indicated loading invariance across 8 of 10 items, supporting partial configural invariance. Implications. This first-known study to explore the construct of co-parenting in foster care demonstrated that the co-parenting construct exists for fathers and mothers and can be assessed with the CFAI-CP. However, measurement invariance results indicated that couples do not view the construct in exactly the same way. This reinforces the need to include foster fathers in practice and research as these similarities and differences may relate to recruitment and retention outcomes in foster families. As well, just as research in the general population has consistently shown co-parenting as uniquely related to child outcomes, its importance for the care and well-being of children in foster care should be investigated.