Applied Psychology OPUS

A Meta-Analysis on Father Involvement and Early Childhood Social-Emotional Development

Robby D. Harris

Research on fathers has been burgeoning for the past three decades. However, results from this literature show conflicting findings and weaken any ability to draw generalizable conclusions on the unique contributions that fathers provide to their children. Therefore, a meta-analysis was conducted with 13 articles published between the years of 1998-2008 to better understand the relationship between father involvement and early childhood social-emotional development. Tests for the moderating effects of SES, race/ethnicity, and father residential status were also performed. Father involvement was positively associated with positive social-emotional abilities and negatively associated with behavior problems. In addition none of the three moderating variables were able to explain any of the variability in the relationship between father involvement and child outcomes.

Traditionally, researchers have focused on the role of mothers in child development, and the influences that this relationship has on the family. However, with about twenty percent of all new marriages ending in divorce or separation within the first five years of marriage, it is of great value to examine the distinct influences that each parent has on a child (National Center for Health Statistics, 2002). When considering the individual influences that a mother and a father have on a child, it is possible that each parent offers very discrete influences that somehow fit together (Amato, 1994; Black, Dubowitz, & Starr, 1999; Marsiglio, Amato, Day, & Lamb, 2000). However, it is also possible that each parent is separately capable of transmitting the necessary skills, behaviors, and values to his/her child. Historically, United States culture has regarded the mother as the primary caregiver and nurturer, and the father as the enforcer of authority and provider of sustenance (Campos, 2008). Yet with changing family structures and so much ambiguity as to how parents function in the 21st century, a new and encompassing approach to understanding modern parenting that truly fits the contemporary needs of parents and children alike is necessary. Thus, this study hopes to add to the literature, which is already serving to break down such stereotyped and possibly antiquated notions, and inspire new pathways of thought that are culturally relevant and based in understanding of our modern family structures.

The importance of parent involvement in the development of children has been studied extensively, illustrating that high levels of parent involvement are associated with high and positive levels of academic achievement (e.g., Flouri & Buchanan, 2004) and social-emotional competence (e.g., Overbeek et al., 2007) throughout childhood and adolescence. However, it is of great importance to consider who is actually being studied in such work on parenting. Most research that examines the role of parents or parent involvement has included mostly mothers or mother figures, rarely including fathers. Yet, these findings are generalized and discussed in terms of broader "parent" involvement. In our times of rapidly changing family structures, studying mothers alone is no longer enough. It becomes imperative, now more than ever, to better understand the role of fathers in child development and child outcomes. Generalizing the effects of mother involvement as representative of the effects of overall parent involvement threatens the veracity of the conclusions that can be drawn from such work. This tendency also disregards the growing body of research that demonstrates that fathers uniquely contribute to their children’s development as distinct from the contributions of mothers.

Over the past three decades, research on fathering has gained great momentum (Downer, Campos, McWayne, & Gartner, 2008; Marsiglio et al., 2000). Father involvement across the development of a child has been linked to children’s psychological adjustment (Flouri, 2008) fewer behavioral problems (Carlson, 2006), higher educational attainment (Flouri & Buchanan, 2004), and overall mental health (Boyce et al., 2006; Dubowitz et al., 2001). The heightened interest in, and great strides toward learning more about fathers and father involvement are exemplified by the creation of a peer-reviewed journal dedicated to the exploration of this field called Fathering in 2003. Still, most research on fathers and children’s outcomes has been limited by its focus on the periods of adolescence, middle childhood, and infancy. There is much less work that highlights fathers in relation to children during the early childhood period who are preschool aged (ages 3-6).

Early childhood is a critical period of development when children experience a variety of changes. One of the most significant changes is the transitioning into the formal schooling environment (i.e., kindergarten and first grade) from preschool settings or in-home care. Studying this transition is important because, traditionally, this is when children begin to develop independence from their parents as they spend significantly more time away from their parents in school with their teachers and peers (Entwisle & Alexander, 1998). Children take on the new role of student and are faced with assessment, comparison, and competition with their peers on an everyday basis. These new roles are paired with the rapid development of cognitive and social abilities (Bates et al., 2006; Entwisle & Alexander, 1998). Furthermore, children are no longer able to receive as much one-on-one care, contact, and encouragement as they might be used to, and are forced to become more self-reliant in this larger group setting. The concept of being equipped or prepared to successfully navigate this transition into formal schooling has been termed "school readiness." Measuring school readiness assists in our understanding of the necessary tools, abilities, and skills that young children need to be successful when they first enter school. School readiness encompasses children’s development across multiple domains, including: behavioral, social, cognitive, language, and physical development.

According to the National Center for Education Statistics (2000) many children in the United States, especially those from a lower socioeconomic status (SES) are entering kindergarten and the first grade lacking the "requisite academic, social, emotional, and language skills to make use of classroom resources and successfully adjust to school" (Downer et al., 2008, p. 68). It is clear that more work must be completed in order to understand what children need to thrive when they enter this new stage, and how families, educators, and policy-makers can help assist the transition. Although cognitive ability is the area of school readiness that receives the most attention, it is equally important that the child’s social-emotional development be taken into consideration (Raver & Zigler, 1997). Social-emotional development in early childhood is described in terms of self-control, assertion, and cooperation, social competence, self-concept, self-esteem, empathy, and emotion and behavior regulation. The proper acquisition and development of these abilities plays a major role in a child’s ability to thrive in kindergarten and the first grade as an autonomous, social being. Children must learn how to navigate themselves in this new environment that is overwhelmed with many other children who are all seeking the same attention and care. They must learn how to interact and cooperate with their peers, but also must learn how to care for and sustain themselves when necessary. If children do not gain these necessary skills in early childhood, they will not be able to effectively communicate and interact with others, and are likely to have a difficult time functioning academically in this new environment. It is clear that social-emotional development in this period is necessary for children to continue healthy development, and that this development has a large influence on children’s ability to effectively transition to formal schooling.

Research has shown that parent involvement helps to facilitate social-emotional competence in this age (e.g., Overbeek et al., 2007). However, with the changing structure of families in the United States and because mother involvement is commonly generalized to represent overall parent involvement, it becomes increasingly important to understand social-emotional development of children in this age in relation to their fathers. Research in the field has begun to investigate the role that fathers play in children’s development, as separate from the effects of mothering. Father involvement has been associated with the same social-emotional development of preschoolers that research had previously associated with parent involvement. Fathers help their children to develop positive self-concept, self-esteem, social competence, empathetic abilities, self-confidence, and emotion regulation (Amato, 1994; Biller, 1993; Culp, Schadle, Robinson, & Culp, 2000; Downer & Mendez, 2005; Fagan & Iglesias, 2000). Although minimal research finds results to the contrary, there is still much work to be completed in this area as studies note that the role of fathers in children’s social-emotional development often varies across specific groups and contexts. Specifically, the influences of SES, race/ethnicity, and father residential status have contributed to the contradictory results in this emergent field. Thus, this study will examine the moderating effects of these variables on the relationship between father involvement and social-emotional abilities in early childhood.

The rationale behind investigating these three moderators in the meta-analysis was to work toward establishing a context for the trends that were reported in the literature. It is extremely important to understand how each of these moderators influences the relationship between father involvement and children’s early social-emotional development so to draw more accurate, directive, and consistent conclusions about the state of the literature. The importance of including these moderating factors is best explained by Coley (2001) who notes the "significant need for further clarification on the range of normative roles that low-income, minority, and unmarried fathers fulfill, as well as on the precursors and effects of their fathering behaviors" (p.743). In this regard, this meta-analysis hopes to add context to the outcomes that emerge within each of these subpopulations, beyond the overwhelming majority of research on fathers that has studied white, middle-class, and married men (Coley, 2001).

In terms of SES, comparison studies find that fathers of higher SES are more likely to have children with higher academic competence (e.g., Bowey, 1995). Yet there has been limited research on the implications of SES of fathers on social-emotional development of children in this age, a critical component of overall school success (Mitchell, 2008). With an increased focus on the importance of development in this domain, it is crucial that more work is completed so to better understand the unique roles that fathers play in this process (Raver & Zigler, 1997). Previous research has found that family income level is actually not a significant predictor of variability in the behavioral-emotional scores of children (Dooley & Stewart, 2007). However, research has also shown that higher family income was associated with greater father involvement with child socialization (Ahmeduzzaman & Roopnarine, 1992), and fewer child behavior problems (Black et al., 1999; Jackson, 1999). Given these mixed findings, there is a need for a systematic and empirical investigation of the overall trends that are actually occurring. Therefore, this review will explore if SES influences the relationship between father involvement and early childhood social-emotional development across multiple studies. It was hypothesized that father involvement from lower-income families would be associated with less social-emotional competence among preschoolers.

Similarly, there has been insufficient research on the influences of race and ethnicity on father involvement and social-emotional development in early childhood. Although emergent literature has focused on fathers of minority status, such as Latino and African American fathers, this construct of minority status is almost always conflated with being of low SES (Downer et al., 2008). And in this regard, it is impossible to parse out the different effects that being a father of either minority status or low SES may have on children’s social-emotional abilities. In fact, it is questionable whether the limited research that has attempted to focus on race/ethnicity as a separate construct from SES was actually successful in doing so. For example, Fagan (2000) found that Head Start teachers rated Puerto Rican American children, whose fathers were more involved in childcare, as lower in social competence than children whose fathers were not involved in childcare. Yet, this finding was not true for the African American fathers. The current literature on father involvement provides no indication that increased father involvement would be associated with lower social-emotional competence (e.g., Roopnarine, Krishnakumar, Metindogan, & Evans, 2006). Therefore, it is likely that the findings reported by Fagan are confounded by outside variables that are not accounted for in the measures of father involvement that were employed. It is possible that these Puerto Rican American fathers who were involved were fathers of children who were in greater need, and therefore were of lower social competence to begin with (Dodge, Pettit, & Bates, 1994). In addition, it is also possible that this lower child social competence was an effect of limited resources that would be associated with low SES and not necessarily with the father’s race/ethnicity.

Nonetheless, research has found that more father involvement was associated with higher levels of children’s emotion regulation among African American families (Downer & Mendez, 2005), just as we would expect based upon the father involvement literature as a whole (Downer et al., 2008). Clearly, there is a demonstrated need to study the two constructs separately to truly understand how each relates to father involvement and child outcomes (Downer et al.). Effective conclusions cannot be drawn from these studies, as there is no way to completely detangle the two constructs of race/ethnicity and SES. That is, there are no means to distinguish whether an effect size was indicative of low SES, and not of race/ethnicity, or vice versa. Therefore, this analysis explored if race/ethnicity influenced the relationship between father involvement and early childhood social-emotional development, across the literature, net of SES characteristics of the samples. It was hypothesized that race and ethnicity, as distinct from income level, would have no moderating effect on the relationship between father involvement and early childhood social-emotional development.

Finally, research suggests differences in the relationship between father involvement and early childhood social-emotional development may exist depending on the residential status of the father (e.g., Black et al., 1999). Typically, father residential status is measured by whether the father lives at home with the child or not. Most commonly, nonresidential father involvement is considered in terms of monetary support. However, this is only one dimension of father involvement. Research has expressed a challenge in measuring nonresidential father involvement in any other terms, such as closeness, quality of interactions, and emotional commitment (Coley, 2001). Yet, researchers have argued that it is these other dimensions of involvement that measure quality in addition to quantity, that may be more influential for, and predictive of, child outcomes (Campos, 2008). Although involvement by nonresidential fathers has been associated with higher levels of social and emotional adjustment and fewer behavior problems (Greene & Moore, 2000), there has been no research which explicitly compares how children’s social-emotional development in families with resident and nonresident fathers differ.

Even though this area of research is limited, the findings that do exist are mixed, making unclear what, if any, effect father residential status has on the relation between father involvement and social-emotional development. Some research has shown that children of nonresident fathers do not display any differences in behavior outcomes from children whose fathers did live in residence (Black et al., 1999). Yet, other research demonstrates that children of nonresident fathers were less sociable, and expressed less interest in social relationships (Peretti & di Vitorrio, 1993). Therefore, in including this moderating variable, the influences of father residential status on the relationship between father involvement and early childhood social-emotional development will be explored. It was hypothesized that children of residential fathers will display higher levels of social and emotional adjustment than those of nonresidential fathers, across equitable levels of involvement (Peretti & di Vitorrio, 1993).

It is clear that the literature on all three of these potentially moderating variables is both limited and inconclusive. A meta-analysis is ideal for addressing these contradictory findings and clarifying the relationships that may exist. The current analysis hopes to elucidate the work that has been completed on father involvement and early childhood social-emotional development, and to also provide guidance for the work that should be completed in the future. The results of this meta-analysis will help to clarify previous findings and provide innovative and meaningful directions. It was hypothesized that there would be a positive association between father involvement and early childhood social-emotional outcomes. With the growing variability of what families in the United States look like, it is increasingly critical that research on the importance and influences of fathers continues to strengthen and grow. In addition, results from this analysis have practical implications for policy development related to promoting positive child outcomes by identifying populations in need (those who display inadequate social-emotional adjustment) and targeting the fathers of these children as important contributors to this development as it relates to school readiness.


Data Collection Method

To collect the studies that were included in this meta-analysis, a broad keyword search was performed online, utilizing all of the major social science databases, totaling 26 databases (e.g., PsycINFO, Sociological Abstracts, International Bibliography of the Social Sciences, ERIC). The keywords used in the search included: father, father involvement, paternal involvement, and male involvement. Furthermore, these keywords were also searched in varying combinations with the keywords: Head Start, Preschool, and Early Childhood in order to collect articles pertaining to the age range in investigation. 3,775 titles were collected from this comprehensive search process and were compiled into an online bibliographic program. After eliminating all duplicate citations, the remaining titles were sorted based upon several primary inclusion criteria (see below). An advanced doctoral student and an advanced undergraduate student completed this sorting procedure with the ultimate goal of retaining only the articles that would hold the potential for analysis.

Articles that were targeted for the current review had to fulfill the following requirements: (1) include a measure of father involvement and school readiness, (2) utilize quantitative measurement, (3) be published in a professional journal between the years 1998-2008, and (4) include data specific to children in early childhood (ages 3-8). Books, book chapters, literature reviews, meta-analyses, and strictly qualitative works were not included in this study. However, dissertation research was kept in order to include the greatest breadth of work in the analyses, which was especially important being that the study of father involvement is still relatively young (Downer, Campos, McWayne, & Gartner, 2008). After sorting the articles based upon the inclusion criteria described above, 120 articles were then passed along for additional sorting to determine whether or not it was possible and/or appropriate to code each article for analysis.

Two investigators further sorted the articles based upon secondary inclusion criteria regarding the specific data included in each article. Firstly, the mean age of the children included in each sample had to be between three and eight years old at the time of the father involvement data collection. For example, an article that collected data on father involvement when a child was two years old, but collected child outcome data at age four did not meet the coding criteria. And secondly, the relevant father involvement and child outcomes data had to expressed in terms of correlations or regression analyses, the forms utilized in these analyses. After sorting through the articles based upon the system described above, 30 articles were left to be included in the larger meta-analysis that will evaluate the association between father involvement and school readiness of children entering the formal school environment (i.e., first grade or kindergarten) from preschool.

The current study examines the moderating influences of father residential status, socio-economic status, and race and ethnicity on the association between father involvement and children’s social-emotional development during the preschool period. Social-emotional development is one construct of school readiness that was measured by the larger meta-analysis. Therefore, the 30 articles from the larger meta-analysis were further sorted based upon relevance to social-emotional development and child outcomes leaving 13 articles to be included in this smaller meta-analysis.

Coding Schemes

Two investigators created a coding manual that was used as a tool to extract and organize all of the pertinent information and data from each article. Included in this coding manual was demographic information about the children and fathers in the study (including age, race/ethnicity information, income level, location of study, residential status of father, relationship of father to child), the sample size, the measures of school readiness and father involvement included in the work, study quality indicators of reliability for each construct (e.g., Cronbach’s alpha, Cohen’s kappa coefficient, Spearman’s rho), and the associations between each set of constructs. The work of this coding process was split between two researchers who established inter-coder reliability on 20% of the articles at 90% agreement or higher. Once each article was coded, it was entered into an SPSS database for organizational purposes.

Organizing Data

Data across measures of father involvement were organized into two dimensions: active and passive father involvement. These dimensions arose from standards established by previous literature but were ultimately based upon the author’s discretion. Measures of father involvement that were defined as "active" included engagement or performing activities with the child, communication patterns, attachment styles, and caregiving roles that the father employed. Measures of father involvement that were deemed as "passive" included general presence of the father in his child’s life, and financial contributions to the child. Although there was an inherent value judgment in placing these labels on the measures of father involvement, this was not the intention of the author, as any involvement in a child’s life should be regarded as valuable.

Data across measures of children’s social-emotional abilities were also organized into two dimensions: positive and negative outcomes. Again, these dimensions arose from standards in the literature, but each measure was categorized based upon the author’s judgment. Positive social-emotional outcomes were regarded as those including social skills and emotion regulation. Negative social-emotional outcomes were equated with behavior problems (including measures of both internalizing and externalizing behavior problems). Each measure included in the 13 articles utilized in these analyses was organized into one dimension of father involvement and one dimension of child outcomes, and effect sizes were calculated at this point.

Calculating Effect Sizes

Each of the 13 articles included in this work reported data in terms of Pearson’s correlation coefficient r except for one article (Keown & Woodward, 2002) for which the data included were transformed into Pearson’s correlation coefficients. To avoid problems with overestimation and skewness that using correlations often induces, the correlations were converted into a standard unit of analysis, Fisher’s Z scores, and were "weighted by the inverse of the variance to give greater weight to larger samples than smaller samples" (Sirin, 2005, p. 423). Seventy-three associations were included in the 13 articles and the associations that were calculated across these 73 associations are called effect sizes. An effect size is the standardized measure of association that is employed in meta-analyses which allows researchers to compare effects across various studies (Borenstein, Hedges, Higgins, & Rothstein, 2009)

These effect sizes were separated in terms of the two dimensions of father involvement and child outcomes as outlined above. In theory, four separate meta-analyses were conducted: one on the relationship between active father involvement and positive social-emotional outcomes, one on the relationship between active father involvement and negative social-emotional outcomes, one on the relationship between passive involvement and positive outcomes, and one on the relationship between passive involvement and negative outcomes. However, there were no measures of association between active father involvement and negative
child outcomes, and consequently no data to
analyze in this regard.

A mean effect size was then calculated for each study included in these three meta-analyses. Although this approach allows for the possibility of overlooking legitimate differences that may exist across multiple correlations within a study (i.e., correlations based upon father involvement, as rated by mothers versus as rated by fathers) it does avoid giving too much power and significance to those studies that include multiple correlations (Lipsey & Wilson, 2001). At this point, each mean effect size was integrated to find one summary effect size to describe the direction and strength of the relationship between the variables of father involvement and social-emotional outcomes. These statistics were transformed back to Pearson’s correlation coefficient r through a z-to-r transformation with 95% confidence intervals to designate the range where the population mean were most likely to fall in the observed data (Hedges & Olkin, 1985).

Random Effects Model

With consideration to the generalizability of meta-analysis effect sizes, there is often debate over whether a fixed effects model or a random effects model should be utilized (Cooper & Hedges, 1994; Hedges & Vevea, 1998; Sirin, 2005). A fixed effects model assumes that the effect sizes calculated are adequate in estimating the population effect sizes, and therefore are generalizable (Lipsey & Wilson, 2001). However, a random effects model assumes that the included studies do not embody any ‘identical’ true effect sizes that would necessarily be representative of the larger population (Borenstein et al., 2009). Under a random effects model, there is no standard for what should be expected as an outcome of the independent variable (i.e., father involvement), and therefore generalizability is low. A random effects model was utilized in this meta-analysis as there is no standard for social-emotional outcomes that would be associated with any level of father involvement, and under this model, more weight was given to studies that included more variance and larger sample sizes.

Tests for Heterogeneity among Correlations

Tests for heterogeneity were included to examine how much of the variation present in the included studies was real and representative of the larger population (Borenstein et al., 2009). In other words, these statistics ensured that the observed findings were representative of real effects and were not merely spurious, or based completely in error. In addition, the I2 statistic indicated what percentage of the observed variation was real. A large I2 statistic would warrant further analyses on the influence of possible moderating variables on the relationship between father involvement and early childhood social-emotional development (Borenstein, Hedges, Higgins, & Rothstein).

Publication Bias

Publication bias is another critical aspect of meta-analyses that was considered. Only articles that were published in peer-reviewed journals were utilized in this meta-analysis. However, it is possible that more work on father involvement and early childhood social-emotional development was conducted during the years 1998-2008, but was not published. This is especially possible if such work lacked statistical significance and included a small sample size. Although these features do not discredit the findings of such work, it certainly might have impeded the opportunity for publishing. Due to this fact, it is possible that the trends captured in this review are only representative of the work that was published, and not of all of the work that was actually conducted in this period. Therefore, tests for publication bias were also included to ensure that the results accurately captured all work conducted between the years 1998-2008.

Test for Moderator Effects

To check for the significance of each moderating variable included in this review (SES, race/ethnicity, and father residential status), the heterogeneity analysis delineated by Borenstein Hedges, Higgins, and Rothstein (2009) was followed. Q statistics from tests for heterogeneity were analyzed. A significant Q-between would have signified that the mean effect sizes across each level of the moderating variables differed by more than merely sampling error, and that the moderating variable did have an effect on the relationship between father involvement and early childhood social-emotional development.


Overall Effect Sizes

The seven articles (31 effect sizes) that were included in this work in relation to active father involvement and positive social-emotional abilities yielded a summary effect size of .22 (p < .05). The eight articles (32 effect sizes) of relevance to active father involvement and negative social-emotional outcomes yielded a summary effect size of -.13 (p = .069). And the three articles (10 effect sizes) of relevance to passive father involvement and negative social-emotional outcomes yielded a summary effect size of -.14 (p=.089).

Tests for Heterogeneity and Publication Bias

In interpreting the Q statistic, we found that it was significant, meaning that there was heterogeneity in the observed effects. When looking at the I2 statistic, we found that 60% of the observed variation was real (which is a medium-large percent for this statistic), meaning that there was enough variation in the overall sample to examine what was contributing to the association (Borenstein et al., 2009). In other words, we were justified to look at the moderating variables to see if they could explain any of the variability. In addition, there was no evidence of publication bias in the sample.

Moderating Variables

In these analyses, we combined both active and passive dimensions of father involvement into one variable, as there were not enough effect sizes from each dimension to run the analyses separately. Also, there were only enough data to run moderator analyses on SES in relation to positive outcomes. There were, however, enough data to run the moderator analyses across all three moderators in relation to negative outcomes. After running the analyses, we found that, with nonsignificant Q-between statistics, neither SES, race/ethnicity, nor father residential status was able to explain any of the variability in relation to both positive and negative social-emotional outcomes. In other words, none of these three variables were found to be significant in moderating the relationship between father involvement and social-emotional outcomes.


Summary Effect Sizes

There was a significant positive effect between active father involvement and the development of positive social-emotional abilities in early childhood, meaning that more father involvement was associated with more positive child outcomes. With an effect size of .22 (p < .05), this relationship was of a medium size in terms of the standards for interpreting meta-analyses (Borenstein et al., 2009). In addition, there were significant trends between father involvement (both active and passive dimensions) and fewer social-emotional problems. In other words, more father involvement was associated with less negative child outcomes. However, it is important to keep in mind that although these effects were not, in fact, significant, when working with such a limited number of studies, significant trends are still impressive. In addition, when interpreting effect sizes in meta-analyses, the standard is to examine the strength of the effect primarily, and to not rely solely on the significance of the effect, as the significance will vary based upon many factors and does not necessarily discredit the presence and strength of an effect (Borenstein et al., 2009). Therefore, father involvement was associated with child outcomes in the expected directions, and supported the hypothesis that there would be a positive association between the two variables (after considering the desired direction of the effect between father involvement and negative social-emotional outcomes).

Moderator Variables

Our next finding was that neither SES, race/ethnicity, or father residential status served as significant moderators of the relationship between father involvement and social-emotional outcomes. These findings ultimately mean this relationship may be the same across all levels of the moderators. In other words, it is possible that the relationship between father involvement and early childhood social-emotional abilities may not differ whether the father were of low SES, minority status, or nonresidential nor whether the father were of mid/high SES, non-minority status, or residential. These findings are consistent with the work by Dooley and Stewart (2007), which also found that family income level was not a significant predictor of variability in the behavioral-emotional scores of children. Therefore, our hypotheses in relation to SES and father residential status were incorrect. However, our hypothesis regarding race/ethnicity was correct in that we predicted that the minority status of the father would not have any influence on the relationship between father involvement and child outcomes.

However, caution should be used in interpreting these results with the great deal of data missing from the moderator analyses. Many of the studies did not include the data necessary to run the moderator analyses, and in fact, data from only five out of the 13 studies were included in these analyses. Some articles did not report data on the three variables whereas others reported mixed samples. With the minimal number of studies that were able to be included in the moderator analyses, it is possible that our results do not fully represent the sample. Therefore, it is necessary for researchers to study context more purposefully in relation to father involvement and child outcomes. Although the relationship did not vary as a function of the three variables, it would be injudicious to assume that father involvement across all of the moderator variable levels was identical. Recent work on parenting demonstrates that involvement across different cultures cannot be measured or interpreted in one distinct manner (McWayne, Owsianik, Green, & Fantuzzo, 2008). Therefore, although we can safely argue that, based upon the findings of this analysis, father involvement is important, it will be important for more work to be completed in which culture and context are explicitly taken into account.

Directionality of Results

Another conclusion to be drawn from this meta-analysis is that the field presently lacks an indication of the directionality of the relationship between father involvement and child outcomes. In other words, we do not know if father involvement affects child outcomes, if child behavior affects father involvement, or if the relationship is transactional, meaning that it works in both directions. This point is important to consider so
that future research can identify where possible interventions and additional supports will be most beneficial. Recent work suggests that the direction of this relationship in relation to parent involvement is transactional in nature (Hoglund, Jones, Brown, & Aber, 2009). Therefore, it is possible that both father involvement and child behavior affect each other. However, only longitudinal data on father involvement and child outcomes will help us to address this properly and test for causality. Nonetheless, engagement of fathers in their children’s lives is an exciting finding, regardless of the direction of the relationship, and should be further supported in various ways.

Conceptions of Father Involvement

This meta-analysis has made clear that the field needs to develop a more explicit and unified conception of what father involvement truly is. Is father involvement how much time a father spends with his child, how the father communicates with his child, or how much money the father contributes to his child’s life? When the field measures one concept in so many diverse ways, the true meaning of the variable can get lost, and therefore it will be important for future research to create measures that more concretely measure father involvement. Before research on fathers involvement can be employed in the development of programs and policies that focus on supporting fathers in raising their children, it will be necessary for the field to synthesize what father involvement includes, how to measure this construct, and what the implications of such involvement are in the lives of children.

Similarly, measures that directly sample fathers will be important to develop. Most of the measures on father involvement included in this project were either completed by teachers or mothers, and only a few studies included data on father involvement that were reported by fathers. There are many extraneous variables that may make reports by mothers and teachers inaccurate in obtaining a true measure of a father’s involvement. For example, the quality of a relationship that a father has with the mother of his child may certainly influence how she reports his involvement regardless of his actual involvement with the child. A divorced custodial mother who does not get along with her divorced husband might report lower father involvement than a divorced custodial mother who does get along with her divorced husband, even if both fathers are involved in the same exact manner. In addition, it is possible that a teacher who does not witness much father presence during the school day would rate such a father as low in involvement. However, it is possible that this father works during the school day, but is very involved with his child outside of school, and therefore such a report might not accurately capture the involvement of this father. The importance of studying context and the intersectionality of variables should be taken into consideration when developing measures of father involvement and study designs in the future.

It is also important to consider what has been addressed in the field since Amato and Gilbreth (1999) conducted the last meta-analysis on father involvement. One of the most significant future directions that these authors suggested was that father involvement should preferably be measured in an engaged manner rather than in a more superficial manner (i.e., physical presence of the father in the child’s life). From examining the articles included in this meta-analysis, it would appear that the field of father involvement has, in fact, taken the directives laid out by Amato and Gilbreth. The overwhelming majority of father involvement measures were "active" in nature (63 effect sizes) with only ten measuring father involvement in "passive" terms. However, what is important to keep in mind is that the meta-analysis by Amato and Gilbreth was conducted on nonresidential fathers, who, as research demonstrates, are less likely to be actively engaged with their children due to limited access (Black et al., 1999). Nonetheless, it is possible the great deal of active father involvement measures that were included this analysis are indicative of a shift in how the field conceives of fathers and their importance.

School Readiness Meta-Analysis

Finally, this meta-analysis examined one construct of school readiness. The larger meta-analysis from which this project is based will examine all aspects of school readiness in relation to father involvement. In addition, the school readiness meta-analysis will include more studies across all of the dimensions of school readiness, including social-emotional development, so to increase the variability of the data and allow us to draw more generalizable conclusions.


*References marked with an asterisk indicate studies included in the meta-analysis.

Ahmeduzzaman, M. & Roopnarine, J. L. (1992). Sociodemographic factors, functioning style, social support, and fathers' involvement with preschoolers in African-American families. Journal of Marriage & the Family, 54, 699-707.

Amato, P. R. (1994). Father child relations, mother-child relations, and offspring psychological well-being in early adulthood. Journal of Marriage and the Family, 56, 1031–1042.

Amato, P. R., & Gilbreth, J. G. (1999). Nonresident fathers and children’s well-being: A meta-analysis. Journal of Marriage and the Family, 61(3), 557-573.

Bates, M. P., Mastrianni, A., Mintzer, C., Nicholas, W., Furlong, M. J., Simental, J., et al. (2006). Bridging the transition to kindergarten: School readiness case studies from California's first 5 initiative. California School Psychologist, 11, 41-56.

Biller, H. B. (1993). Fathers and families: Paternal factors in child development. Boston: Auburn House.

Black, M. M., Dubowitz, H., & Starr, R. H. (1999). African American fathers in low income, urban families: Development, behavior, and home environment of their three-year-old children. Child Development, 70, 967-978.

Borenstein, M., Hedges, L.V., Higgins, J. P.T., & Rothstein, H.R. (2009). Introduction to meta-analysis. Hoboken, NJ: Wiley.

Bowey, J. A. (1995). Socioeconomic status differences in preschool phonological sensitivity and first-grade reading achievement. Journal of Educational Psychology, 87, 476-487.

Boyce, W. T., Essex, M. J., Alkon, A., Goldsmith, H. H., Kraemer, H. C., & Kupfer, D. J. (2006). Early father involvement moderates biobehavioral susceptibility to mental health problems in middle childhood. Journal of the American Academy of Child and Adolescent Psychiatry, 45(12), 1510-1520.

*Burgos, L. (2003). The effect of the father-child relationship on the social conduct of 2 1/2 year old children in preschool. European Journal of Psychology of Education, 18(2), 135-155.

Campos, R. (2008). Considerations for studying father involvement in early childhood among Latino families. Hispanic Journal of Behavioral Sciences, 30(2), 133-160.

Carlson, M. J. (2006). Family structure, father involvement, and adolescent behavioral outcomes. Journal of Marriage and Family, 68,137-154.

Coley, R. L. (2001) (In)visible men: Emerging research on low-income, unmarried, and minority fathers. American Psychologist, 56(9),743-753.

Cooper, H. M., & Hedges, L. V. (Eds.). (1994). Handbook of research synthesis. New York, NY: Russell Sage.

*Cugmas, Z. (1998). The correlation between children's personal behavioural characteristics and indicators of children's attachment to their mother or father, respectively. Early Child Development &Care, 143, 65-78.

*Culp, R. E., Schadle, S., Robinson, L., & Culp. A. M. (2000). Relationships among paternal involvement and young children’s perceived self-competence and behavioral problems. Journal of Child and Family Studies, 9, 27-38.

*Denham, S. A., Workman, E., Cole, P. M., Weissbrod, C., Kendziora, K. T., & Zahn-Waxler, C. (2000). Prediction of externalizing behavior problems from early to middle childhood: The role of parental socialization and emotion expression. Development & Psychopathology, 12(1), 23-45.

Department of Health and Human Services. (2002). Cohabitation, Marriage, Divorce, and Remarriage in the United States (DHHS Publication No. PHS 2002-1998). Washington, DC: U.S. Government.

Dodge, K. A., Pettit, G. S., & Bates, J. E. (1994). Socialization mediators of the relations between socioeconomic status and child conduct problems. Child Development, 65, 649-665.

Dooley, M., & Steward, J. (2007). Family income, parenting style, and child behavioural-emotional outcomes. Health Economics, 16(2), 145-162.

Downer, J. T., Campos, R., McWayne, C.M., & Gartner, T. (2008). Father involvement and children’s early learning: A critical review of published empirical work from the past 15 years. Marriage & Family Review, 43(1/2), 67-108).

*Downer, J. T., & Mendez, J. L. (2005). African American father involvement and preschool children’s school readiness. Early Education and Development, 16, 317-340.

Dubowitz, H., Black, M. M., Cox, C. E., Kerr, M. A., Litrownik, A. J., Radhakrishna, A., English, D. J., Schneider, M. W., & Runyan, D. K. (2001). Father involvement and children’s functioning at age 6 years: A multisite study. Child Maltreatment: Journal of the American Professional Society on the Abuse of Children, 6, 300-309.

Entwisle, D. R., & Alexander, K. L. (1998). Facilitating the transition to first grade: The nature of transition and research on factors affecting it. Elementary School Journal, 98, 351-364.

Fagan, J. (2000). African American and Puerto Rican American parenting styles, paternal involvement, and Head Start children’s social competence. Merrill-Palmer Quarterly, 46, 592-612.

Fagan, J., & Iglesias, A. (2000). The relation between fathers’ and children’s communication skills and children’s behavior problems: A study of Head Start children. Early Education and Development, 11, 307-320.

Flouri, E. (2008). Fathering and adolescents’ psychological adjustment: The role of fathers’ involvement, residence and biology status. Child: Care, Health and Development, 34(2), 152-161.

Flouri, E., & Buchanan, A. (2004). Early father’s and mother’s involvement and child’s later educational outcomes. British Journal of Educational Psychology, 74, 141-153.

Greene, A. D. & Moore, K. A. (2000). Nonresident father involvement and child well-being among young children in families on welfare. Marriage and Family Review, 29, 159-180.

*Hagan, L. K., & Kuebli, J. (2007). Mothers' and fathers' socialization of preschoolers' physical risk taking. Journal of Applied Developmental Psychology, 28(1), 2-14.

Hedges, L.V. & Olkin I. (1985). Statistical methods for meta-analysis. Orlando, FL: Academic Press, Inc.

Hedges, L. V. & Vevea, J. L. (1998). Fixed and random-effects models in meta-analysis. Psychological Methods, 3, 486–504.

Hoglund, W. L. G., Jones, S. M., Brown, J. L., & Aber, J. L. (2009, April). Family home and school- based involvement in the inner-city context: Transactions with child school adjustment. Paper symposium presented at the 2009 biennial meeting of the Society for Research in Child Development, Denver, CO.

*Jackson, A. P. (1999). The effects of nonresident father involvement on single Black mothers and their young children. Social Work, 44, 156 166.

*Keown, L. J., & Woodward, L. J. (2002). Early parent-child relations and family functioning of preschool boys with pervasive hyperactivity. Journal of Abnormal Child Psychology, 30(6), 541-553.

Lipsey, M. W., & Wilson, D. B. (2001). Practical meta-analysis. Thousand Oaks, CA: Sage.

Marsiglio, W., Amato, P., Day, R., & Lamb, M. E. (2000). Scholarship on fatherhood in the 1990s and beyond. Journal of Marriage and Family, 62, 1173-1191.

McWayne, C.M., Owsianik, M., Green, L.E., Fantuzzo, J. W. (2008). Parenting behaviors and preschool children’s social and emotional skills: A question of the consequential validity of traditional parenting constructs for low-income African Americans. Early Childhood Research Quarterly, 23, 173-192.
Mitchell, S. J. (2008). Low-income African American fathers’ contributions to toddlers’ social and emotional development (Doctoral dissertation, University of Maryland, College Park, 2007). Dissertation Abstracts International Section A: Humanities and Social Sciences 69(2-A).

National Center for Educational Statistics. (2000). The kindergarten year. Washington, D.C.: National Center for Educational Statistics.

Overbeek, G., Stattin, H., Vermulst, A., Ha, T., Engels, & Rutger C. M. E. (2007). Parent–child relationships, partner relationships, and emotional adjustment: A birth-to-maturity prospective study. Developmental Psychology, 43(2), 429-437.

*Page, T., & Bretherton, I. (2001). Mother- and father-child attachment themes in the story completions of pre-schoolers from post-divorce families: Do they predict relationships with peers and teachers? Attachment & Human Development, 3(1), 1-29.

Peretti, P. O., & di Vitorrio, A. (1993). Effect of loss of father through divorce on personality of a preschool child. Social Behavior & Personality, 21, 33-38.

*Pettit, G. S., Brown, E. G., Mize, J., & Lindsey, E. (1998). Mothers' and fathers' socializing behaviors in three contexts: Links with children's peer competence. Merrill-Palmer Quarterly, 44(2), 173-193.

Raver, C. C., & Zigler, E. F. (1997). Social competence: An untapped dimension in evaluating Head Start’s success. Early Childhood Research Quarterly, 12, 363-385.

*Roopnarine, J. L., Krishnakumar, A., Metindogan, A., & Evans, M. (2006). Links between parenting styles, parent-child academic interaction, parent-school interaction, and early academic skills and social behaviors in young children of english-speaking Caribbean immigrants. Early Childhood Research Quarterly, 21(2), 238-252.

Sirin, S. R. (2005). Socioeconomic status and academic achievement: A meta-analytic review of research. Review of Educational Research, 75(3), 417-453.

*Stover, C. S., Van Horn, P., Turner, R., Cooper, B., & Lieberman, A. F. (2003). The effects of father visitation on preschool-aged witnesses of domestic violence. Journal of Interpersonal Violence, 18(10), 1149-1166.

*Verschueren, K., & Marcoen, A. (1999). Representation of self and socioemotional competence in kindergartners: Differential and combined effects of attachment to mother and to father. Child Development, 70(1), 183-201.