Family Process Models

Bronfenbrenner's (1995) bioecological theory and Belsky's (1984) process model of parenting have provided useful frameworks for guiding research on parenting and child development. During the 1990s, several scholars have provided additional framing for those interested in more fully delineating how the resources available to families (or lack thereof) are implicated in parenting and child functioning (Conger, Wallace, Simons, McLoyd, & Brody, 2002; McLoyd, 1990). Generically, these frameworks can be organized under the rubric, family process models of parenting. They attempt to explicate how the resources available to parents affect parental mood, expectations, and mental health, which in turn affect quality of parenting (i.e., a cycle of exchanges between parent and child) and how that helps shape the course of behavioral development. In these models, very specific mechanisms linking the availability of resources to child adaptive functioning are stipulated. This has the advantage not only of offering something closer to an explanation for observed relations but also in designing interventions directed at forestalling or enhancing key processes. These models have been particularly useful in helping to explicate the relation between poverty and maladaptive behavior. An interesting recent use of family process models is a study by Yeung and colleagues (2002). These researchers found evidence for distinctly different paths linking family income to child outcomes depending on the outcome. The link to child achievement was primarily mediated by allowing the child greater access to stimulating experiences; the link to behavior problems was mediated through parental distress and negative parenting. Our analyses likewise implicate learning stimulation as a mediator of the link between family income and achievement test performance (Bradley & Corwyn, 2003). However, our analyses simultaneously controlled for and tested the effects of family financial assets (wealth), maternal education, and occupational prestige.

Learning stimulation was a significant mediator for all four types of socio-economic resources. However, its role as a significant mediator depended upon the age of the child and ethnicity. In our analyses we also simultaneously controlled for and tested the effects of maternal responsiveness as a potential mediator. When child behavior problems was used as the outcome variable, we found evidence that learning stimulation mediated relations between SES variables and behavior problems to at least as great an extent as did maternal responsiveness (typically more so). Moreover, we allowed for testing of direct effects as well; that is, effects unmediated through learning stimulation and maternal responsiveness. Our findings showed that there were often significant direct effects even when testing for these two mediators, which suggests that there are additional unmeasured mediators for both child achievement and behavior. Our findings attest to the value of examining what might be termed "full assets" models even if one 's principal interest is in one type of asset (family assets tend to covary and they tend to covary in somewhat different ways depending on culture). It also attests to the value of investigating models with direct as well as indirect (mediated) effects.


In a recent paper, Tanfer and Mott (1997) argued that changes in family patterns, brought on by changes in social and economic conditions, "signal a weaker commitment of women to men and of men to women; a weaker commitment by the partners to their relationship, and very possibly a weaker commitment to their children" (p. 2). Trends from the past half century seem clear and compelling: smaller families, fewer and delayed marriages, increased divorce, delayed childbearing, advancing numbers of children growing up in single-parent households. Growth in materialism, stresses from work, and the rising cost of rearing children almost certainly conspire to reduce the value of children and adults' commitment to them. More and more, status (and fulfillment) is derived from outside the home, outside the role of parenting.

According to parental investment theory, "the responsibilities of caring for young, often unruly children would seem to require putting one's needs and desires on the back burner so that one can attend to the needs of one's offspring" (Bjorklund & Kipp, 1996, p. 181). The ability to inhibit one's own impulses to leave one's offspring to their own resorts, by extension, requires that one derive pleasure in caring for the child and that one identify one's own well-being with the well-being of the child (Marsiglio et al., 1997).

Concerns about parental investment in children span many decades and many disciplines, from evolutionary biology (Trivers, 1972) to economics (Becker, 1991) to psychology (Hertwig, Davis, & Sulloway, 2002). Each decade and each discipline has put its own stamp on the issue, but there is consensus on one belief: investing in children entails costs. As the trend toward disinvestment in family and child rearing has accelerated (Popenoe, 1993), our research has focused more intensively on parental involvement in the lives of children. We constructed a measure called Parents' Socioemotional Investment in Children (PIC; Bradley, Whiteside-Mansell, Brisby, & Caldwell, 1997). The PIC assesses four components of investment: acceptance of the parenting role, delight in the child, knowledge/sensitivity, and separation anxiety. A study done on 137 mothers of 15-month-old children revealed that PIC was related to the quality of caregiving, maternal depression, neuroticism, agreeableness, social support, the quality of the marital relationship, parenting stress, and perceived child difficultness.

Because of increased concerns about parental involvement in the lives of children, we examined factors that predicted three component scores from PIC (acceptance, delight, and knowledge/sensitivity) when children were 15 months old (Corwyn & Bradley, 1999). Predictors were chosen based on Belsky's process model of parenting (Belsky, 1984) and conceptions about fathering discussed by Doherty, Kouneski, and Erickson (1998). They consisted of income to needs, marital quality, mother's employment, mother's work strain, father's work strain, child's compliant behavior (adaptive social behavior inventory), and child's developmental competence (Bayley scores). Results were somewhat different for mothers and fathers. For fathers, the child's Bayley score, and father's work strain had negative influences on both acceptance while marital quality had a positive influence. Maternal employment and father's work strain were negatively correlated with paternal knowledge/sensitivity. Child temperament and being employed had negative impacts on maternal acceptance. Marital quality was positively associated with maternal knowledge/sensitivity and father's work strain had a positive impact on maternal delight. Overall, such findings are reminiscent of the findings of Woodworth, Belsky, and Crnic (1996).

A major reason for being concerned with parents' level of investment in their children is the belief that high levels of investment will redound to the benefit of children. However, the likelihood that parental investment matters for children at least partly depends on whether the parent's level of investment remains reasonably constant through time (Bloom, 1964). Accordingly, we were concerned with the stability of parental investment (Corwyn & Bradley, 2002a). To examine this issue, it was necessary that we first establish factorial invariance of PIC over time in that evidence for rank order stability through time would have equivocal meaning in the absence of factorial invariance. In a sample of 102 mothers who completed PIC when their children were 15 and 36 months of age, we found strong evidence for factorial invariance for both Knowledge/Sensitivity and Delight. However, we found evidence that one item from the Acceptance scale was less valid at 36 months. Given these results, we found evidence for substantial rank order stability from 15 to 36 months on all three components. To an extent, this was both good news and bad news. The good news: Because parental attitudes concerning investment tend to be stable, it is more likely that parents whose initial attitudes regarding investment are positive will persist in doing things in behalf of their children to the extent that investment attitudes shape their behavior. The bad news: It may be more difficult to change negative attitudes given that they tend to remain fixed. More recently, we have examined the stability of responses to PIC subscales for fathers as well as mothers (Corwyn & Bradley, 2002b). We investigated the stability of five constructs (i.e., protection, sensitivity, delight, proximity, and acceptance of the parenting role) of mothers' and fathers' socio-emotional investment in the child during early childhood (from 15 months of age to 36 months of age). Three types of stability were assessed: factor structure stability, mean level stability and rank order stability. All five constructs showed acceptable factor structure stability for both parents. The only significant mean level changes were decreases in proximity and protection among mothers. All constructs showed significant rank order stability for both parents with mothers consistently showing higher rank order stability than fathers.

The relation between parental attitudes and parental behavior remains murky (Holden & Buck, 2002). In an effort to more fully understand how attitudes regarding parental investment may help to shape parenting behaviors, we examined factors connected with paternal efforts to help prepare children to get ready to go to kindergarten. The participants were fathers of children in Head Start. Using Belsky's process model of parenting as a broad frame and Palkovitz's (2002) notions about generative fathering as a more specific frame, we selected the following variables to predict fathering behaviors such as reading to the child, sharing stories with the child, involving the child in daily activities, and involvement in Head Start: family conflict, family cohesion, marital quality, parenting stress, delight (from PIC), biological relationship to child, residency status, parenting efficacy, father's relationship with his parents. Although significant predictors varied as a function of what fathering behavior was used as the dependent variable, the father's sense of efficacy, the quality of his relationship with the child's mother, and his level of contact with his own mother tended to be the most consistent predictors. All and all they corroborate the assertion that situation/context plays a significant role in these relations (Holden & Buck, 2002).


For Americans, the twenty-first century began with a bang—literally. Two words conjure scorched images of life in the here and now, life as we did not realize it was: Columbine and September 11th. Both focus the mind on what family means and the awesome responsibility of parenting in an uncertain world. For those of us interested in the welfare of children, these events made immediately clear that the science of parenting has yet to equip parents with all the tools needed to assure that children thrive. In their aftermath, the American Psychological Association put out a book that addresses issues about how children respond to violence, terrorism, and natural disasters (La Greca et al., 2002). In that volume there is some useful advice, based on clinical experience, on how to help children cope with the aftermath of terrifying events. But, from the book, it is also clear that there is little in the way of science regarding what parents actually do to assist their children in such times or whether it is effective. Moreover, the book does not even address a myriad of other circumstances that may severely disrupt a child's sense of emotional security and move the child on a less adaptive personal life trajectory. Such research is a major need for the future. Parents need more than a solid scientific base for rearing children under more or less benign circumstances. They need a solid scientific base for helping their children cope with trouble, be it bullies, acts of terror, natural disasters, predators—whatever.

The most striking thing we have learned from our long years of studying parents and children, is how diverse life is "in the moment" for children. This "fact" has consumed us as we have struggled to develop measures that are useful representations of that broad phenomenon called parenting (a.k.a., the home environment). What, among the myriad acts, objects, conditions, and events that constitute the home environment, should be included in measures of parenting? What indicators are telling enough across the broad diversity of families in America and throughout the world that they capture what is most salient about parenting? With family life so bound up with history, culture, community, and place, it seems impossible that any one measure (given that it contains only a limited census of indicators) could sufficiently capture what is most salient about parenting for all groups and all situations. But, without "marker" measures or indicators, how can we construct an integrated science of parenting, a theory of environment-development relationships that is anything more than parochial? This dilemma would seem to lead to a kind of "Sophie's Choice" regarding which strategy to keep and which to leave behind; but should not. It has led us to recommend a dual strategy for studying parenting: It is what we call an inside and outside strategy with regard to marker measures of key parenting domains. The inside strategy means looking inside the measures themselves and, more specifically, looking at the issue of measurement invariance across groups. Using such techniques as confirmatory factor analysis, structural equation modeling, and multitrait-multimethod analysis, we are looking at whether the internal structure of measures is similar across groups and whether a measure links with similar constructs across groups (Whiteside-Mansell, Bradley, Owen, Randolph, & Cauce, 2003). Initial results suggest similarities in some of the behaviors commonly coded in mother-child interaction paradigms for European American and African American families, but not complete isomorphism. The outside strategy is to supplement marker measures with outside indicators that may be particularly revealing of a parenting domain in a particular group. Perhaps, the best example of the outside strategy in our own research has been our work with children with disabilities. In those studies, we added items to the Infant-Toddler, Early Childhood, and Middle Childhood versions of HOME that were designed for families having children with various types of disabilities. The thought was that these children may require additional things from their environments in order to do well. Our findings showed that the items on the HOME generally worked well in predicting how children would do on standard measures of social and cognitive functioning but, as the severity of the disability increased, the additional items became more useful in predicting the course of development (Bradley et al., 1989). When approached by researchers from other cultures, we now routinely recommend gathering information on parenting indicators of local value in addition to those found on the HOME, then analyzing data from the standard HOME and from the additional items in tandem. In addition to recommending the outside strategy with regard to the measurement of parenting, we also have begun examining the relation between parenting and children's development within cultures (e.g., our studies using the Early Adolescent HOME and our recent studies involving NLSY). There is convincing evidence for the value of within-group analyses for building a comprehensive science of human development (Garcia Coll et al., 1996).

If the past is prologue yet only dimly foreshadows what is to come, then one may wonder what the future of parenting science holds. The realities of child rearing in the twenty-first century present extraordinary obstacles to definitive research, with complexity and diversity in arrangements compounding the ordinary problems of linking particular parenting practices to the trajectory of development in highly evolved, phylogenetically advanced organisms living in rich, multilayered contexts. The challenge requires longitudinal designs with repeated measures of all the caregiving environments in which a child spends a meaningful amount of time (Friedman & Haywood, 1994). In the NICHD Study of Child Care and Youth Development, we have constantly been concerned with the combined effects of child care and home or school and home (NICHD Early Child Care Research Network, 2003). As a result we often include measures of both contexts in a single analysis aimed at understanding children's development. Constructing a useful knowledge base for parenting in the twenty-first century also requires a more integrated, holistic view of children. The evidence is compelling that children are not just people with behavior problems or people who regulate their emotions or people who possess a certain level of physical or academic competence (Zaff et al., 2003). The goals of parents and of cultures recognize the integrity of development and how developmental systems feed offone another. Yet most studies of parenting fail to take account of the dynamic interplay among these systems. Just as we have argued that one cannot usefully examine aspects of parenting in isolation from one another (children don't just experience one aspect of parenting separately from the rest), so it is important that we construct studies that do not consistently isolate one component of development from the others.

One way to find the whole child again in studies of development may be to use person-centered approaches to research to complement our current almost exclusive reliance on variable-centered approaches. The person-centered approach has made significant contributions to research on resiliency, personality and adolescent behaviors, but has not been equally utilized in parenting research. We recently used a person-centered strategy as a way of better understanding how socio economic status influences child behavior (Corwyn & Bradley, 2005). We found support for all three propositions of the person-centered approach outlined by Bergman (2000); that models likely do not apply to everyone, relations are frequently not linear and patterns of values often have more meaning than variables considered individually (Bergman, 2000).

As the future intrudes into daily life with ever increasing speed and persistence, the job of parenting will rely less on prescriptions and more on flexible problem-solving strategies aimed at finding the best fit between long-term goals for a child, the child's current needs and capabilities, and what the environment affords by way of demands and opportunities. Parenting science must move quickly to develop new theories and methodologies just to keep up.

Defeating Divorce

Defeating Divorce

Get All The Support And Guidance You Need To Be A Success At Saving Your Marriage. This Book Is One Of The Most Valuable Resources In The World When It Comes To Understanding The Opposite Sex And Saving Your Marriage.

Get My Free Ebook

Post a comment