The Father I Knew: Early Paternal Engagement Moderates the Long-term Relationship between Paternal Accessibility and Childhood Behavioral Difficulties

Abstract

The long-term impact of a father’s involvement in offspring development is understudied. The current study investigated the relations between early paternal engagement (i.e., the amount of time fathers engaged in one-to-one activity with their child), paternal accessibility (i.e., fathers’ physical proximity to their children) and later child behavioral difficulties. Data were obtained from five phases of the Auckland Birthweight Collaborative (ABC) longitudinal cohort study: at the study children’s birth; at 12 months; 3.5 years; 7 years; and 11 years of age. Moderated linear regression analyses revealed that there was a negative, long-term effect of paternal departure from the family household (i.e., reduced paternal accessibility) by 3.5 years of age on children’s total behavioral difficulties and conduct problems scores, but only if departed fathers had been highly engaged during the child’s first year of life. These findings suggest that the relationship between paternal accessibility and paternal engagement is potentially more nuanced than previously thought.

Keywords

child behavior behavioral difficulties paternal engagement paternal accessibility fathers moderation

Introduction

Fathers play a much more prominent role in the psychosocial development of their children than was once thought. Findings from longitudinal studies are broadly consistent in observing that the more involved fathers are, the more positive the emotional and behavioral outcomes for their children over time. Indeed, the long-term positive effects of paternal involvement have been observed throughout childhood, adolescence, and into young adulthood (Sarkadi et al., 2008). In accordance with a proposed tripartite understanding of “paternal involvement” (Lamb et al., 1987; Lamb, 2000), fathering research typically parses paternal involvement into paternal accessibility (e.g., physical proximity), paternal responsibility (e.g., control over children’s health and well-being), and paternal engagement (e.g., direct, one-on-one involvement with play or child rearing). The concept of paternal engagement has been further refined to draw focus to the concept of positive engagement practices, or those which will directly foster adaptive psychosocial development in children, as opposed to routine care (Pleck & Masciadrelli, 2007; Sayer et al., 2004).

Paternal Accessibility vs Paternal Engagement

To date, a small number of longitudinal studies have addressed the effect of either a father’s accessibility in a child’s life or their engagement in child rearing. Paternal accessibility is defined as the father’s physical proximity and availability to the child, in the sense that he can be called upon at short notice (Lamb, 2000; Sarkadi et al., 2008). As such, father–child cohabitation is commonly used as a suitable proxy for paternal accessibility. In a short prospective cohort study, Vaden-Kiernan et al. (1995) observed that 10-year-old boys who were cohabitating with their fathers exhibited reduced aggressive behaviors at 12 years of age when compared to boys who were not cohabitating with their fathers. Children who cohabitated with their fathers from birth up to 14 years of age exhibited significantly better behavioral conduct than children whose fathers had not been present since birth (Carlson, 2006); however, other evidence suggests that father–child cohabitation may only exert a positive longitudinal effect on behavioral conduct for girls (Flouri & Buchanan, 2002). More recently, we observed paternal departure from the child’s household (i.e., cessation of father–child cohabitation) by the age of 3.5 years to have significant, negative effect on child behavioral outcomes at 11 years of age (Craig et al., 2018).

Paternal engagement is described as direct contact and joint activity between father and child that is positive and prolonged (Lamb, 2000; Pleck & Masciadrelli, 2007). As a predictor of later child behavioral outcomes, engagement has been much more extensively studied than paternal accessibility. A lack of paternal engagement with children as early as 2–6 months of age is a significant predictor of increased behavioral difficulties at 1 year of age (Ramchandani et al., 2013). Others have observed that only boys show an association between early paternal engagement prior to 6 months of age and behavior problems at 18 months of age (Keizer et al., 2014). Greater paternal engagement (reading with children) during infancy predicted fewer behavioral difficulties by 3 years of age (McMunn et al., 2017). Increased paternal engagement in play activities at 18 months of age predicted greater prosociality and less behavioral problems in children at approximately 7 years of age (Ramchandani et al., 2010). Data from the US National Survey of Families and Households (NSFH) indicated that fathers’ involvement in catering to children’s physical needs (e.g., feeding, bathing) and engagement in one-on-one play predicted fewer problem behaviors between the ages of 5 and 9 years (Aldous & Mulligan, 2002). Paternal engagement (i.e., feeding, playing, or reading with the child) at 5 years of age was shown to have limited short-term positive effects in reducing peer interaction problems in children at 7 years of age (Flouri et al., 2016).

Paternal Accessibility and Engagement: The Present Study

While some earlier longitudinal studies have merged the constructs of paternal presence and paternal engagement into general measures of “involvement” (e.g., Chang et al., 2007), none to date have examined the interplay between paternal accessibility and paternal engagement. The current study sought to determine the long-term effects between paternal accessibility and paternal engagement on child behavior as separate constructs. There is evidence to suggest that fathers who are not cohabitating with their children display lower levels of engagement (Castillo et al., 2011). Meta-analytic findings regarding noncohabitating fathers suggest that visitation frequency (paternal accessibility) does not significantly influence child behavior outcomes over time. However, a sense of closeness between father and child is related to better outcomes (Wilson, 2006). Based on adolescent self-report, there is some evidence that greater perceived paternal engagement predicts a greater sense of closeness (Armsden et al., 1990; Kenny & Gallagher, 2002).

All influence of paternal accessibility and engagement were considered in this study using the bioecological model of human development (Bronfenbrenner, 1979, 1994). In order of proximity, an individual’s microsystem (immediate environment), mesosystem (combinations of microsystems), exosystem (individual social institutions), and macrosystem (overall socioeconomic milieu), each contains interrelated developmental influences, while the chronosystem refers to the pattern of an individual’s experiences over time. During early childhood, the primary domain of influence in an individual’s microsystem (and, therefore, their meso and exosystems) is the immediate family. It is now universally understood that variables within the immediate family can exert developmental influence on children from the moment of conception (Nau & Heckert, 2013).

In contextualizing the immediate family within a child’s microsystem of developmental influence, a theoretical assumption within this thesis is that greater paternal involvement in children’s lives does have long-term effects throughout development. Pleck’s (2010) parental capital model amalgamates ideas from social capital theory (Coleman, 1988) with another of Bronfenbrenner’s contributions: “proximal process” (Bronfenbrenner, 1994). The concept of proximal process suggests that adaptive child development is the product of increasingly complex, reciprocal actions between a child and their environmental systems. Pleck (2010) argues that the immediate family’s role in this proximal process is twofold: (a) to directly provide complex interactions that challenge the child; (b) to facilitate the child’s acquisition of “community social capital” (Coleman, 1988), or, as Bronfenbrenner might describe, the expansion of the child’s microsystem, and the creation of a mesosystem.

We hypothesized that paternal engagement may interact with paternal accessibility on child behavioral outcomes. In examining paternal engagement as a moderator, we expected that if fathers were less accessible, their engagement would ensure that they remained an important positive factor of influenced within children’s developmental microsystem. We predicted, therefore, that children, whose fathers were more highly engaged, would experience less behavioral problems despite any likely effects of departure from the household (Carlson, 2006; Craig et al., 2018).

Method

Participants

The Auckland Birthweight Collaborative (ABC) study is described in detail elsewhere (Thompson et al., 2001). Briefly, it is a longitudinal cohort study comprised of New Zealand children selected at birth between October 1995 and November 1997. Children were selected throughout the entire period in the Auckland District Health Board region and from between October 1995 and August 1997 in the Waitemata District Health Board region. All small for gestational age infants (SGA; ≤10th percentile for gestation and sex) were included in the study alongside a randomly selected cohort of appropriate for gestational age (AGA) infants. The sample was limited to infants that were residents and born in the allocated study regions, were born to term (>37 weeks gestation) from single births, and who did not possess any congenital abnormalities. Regarding the data collected at birth, there were 1,714 infants, 871 (50.8%) of whom were born to NZ European mothers. As rates of participation in the early phases of the study from non-European participants were low, the study was limited to contain only the NZ European cohort.

Data were obtained from five data collection waves of the ABC (birth of child, 1 year, 3.5 years, 7, and 11 years of age). At birth, demographics and information about maternal health, and lifestyle factors during pregnancy were collected by maternal interview and obstetric records. As children approached 1 year of age, information about their home environment (including maternal stress levels and perceived social support) and physical development was collected via postal questionnaire. At 3.5 years of age, children were assessed on measures of cognitive and physical development. Mothers were interviewed regarding the child’s health and development during the early years of life. Demographic information was also collected regarding the child’s family environment, including socioeconomic status. When children were aged 7 years, information was collected from maternal interview about children’s physical activity, diet, behavior, and health. Similarly, at 11 years of age, demographic information, information about the health of the child was collected via maternal interview.

Ethics approval for each phase of the study was obtained from the New Zealand Northern Regional Ethics Committee. At each phase of the study, parents or caregivers provided signed consent for study participation, with accompanying assent from the children.

Measures

Paternal accessibility was assessed via the proxy of fathers’ presence in the same household as the child. Fathers’ presence was measured at birth through to the 11-year phase as a binary yes/no response item. At birth and 1 year of age, mothers of children were asked “Do you live with the baby’s father?”, while in the remaining phases of the study they were asked “Are you living with the child’s natural father?” These two questions are consistent with Lamb and colleagues’ (1987) definition of paternal accessibility. In an attempt to quantify the degree of paternal involvement throughout the study (i.e., the length of time fathers spent in the same household as their children), a variable was derived from the initial paternal household presence response items in a stepwise fashion (six-dose response levels) to indicate the time at which a father ceased living in the same household as the child. Subsequently, a three-category version of the variable was created. If the father was cohabitating with the child through to 11 years of age this was labelled “full childhood accessibility,” if he had departed the household by either 11 years or 7 years of age this was labelled “early childhood accessibility,” and if he had not been present since birth or departed by 1 year or 3.5 years of age this was labelled “minimal accessibility.” Paternal accessibility was used as an ordinal predictor variable.

Paternal engagement was assessed via tracking the amount of direct contact/play between father and child. At 1 year, and 3.5 years of age, mothers of children were asked how many hours and minutes per week resident fathers spent “playing or doing things” with their children. Separate hours and minutes variables were combined to produce a total of engagement time per week. This measure of engagement is consistent with the Lamb et al. (1987) conceptualization of paternal engagement.

At 11 years of age, behavioral difficulties were measured using the Strengths and Difficulties Questionnaire (SDQ; Goodman et al., 1998). The SDQ is a brief, 25-item scale addressing positive and negative emotional and behavioral characteristics of the child. The SDQ captures whether the identified reporter believes the child has a problem in a behavioral area, and any resulting distress or social impairment. It is comprised of five subscales of five items each: emotional problems; conduct problems; hyperactivity; peer relationship problems; and prosocial behavior. Each individual item is scored on a 3-point Likert scale (0 = not true, 1 = somewhat true, 2 = certainly true). The scores on the first four subscales are summed together to create a Total Difficulties score that can range from 0 to 40. The SDQ has a test–retest validity of 0.62 after a period of 6 months, and subscale internal consistencies range between 0.62 and 0.75 (Goodman, 2001). At 11 years of age, maternal caregivers completed the parent-report version of the questionnaire.

Child birthweight information obtained from obstetric records has been used in conjunction with sex and gestational age to ascertain whether a participant in the study was SGA or AGA. The officially recorded biological sex of children was coded dichotomously as male or female. Given the association between maternal antenatal stress and offspring behavioral difficulties (Slykerman et al., 2015), a measure of mothers’ perceived stress at birth was included as a covariate. Perceived stress was measured using a 10-item version of the Perceived Stress Scale (PSS; Cohen & Williamson, 1988). The scale assesses the degree to which mothers’ thought their lives to be uncontrollable, unpredictable, and overwhelming in the four weeks preceding assessment. Each item was scored on a 5-point Likert scale ranging from 0 (never) to 4 (very often), allowing total scores of 0 to 40.

A measure of family income was obtained at birth. It reflected the total income of the household in which the child resided and was categorized based on whether a household earned either above or below the median household income in the cohort. This was to ensure that the variable remained relevant over time, as wage inflation caused the median NZ household income to fluctuate. If a household earned up to NZ$35,000 a year this was labelled “lower income” and if a household earned greater than NZ$35,000 a year this was labelled “higher income.” At 3.5 years of age, a measure of family SES was derived from parental occupations. Making use of the Elley–Irving occupation classification (Elley, 1985; Elley & Irving, 2003), occupations of both parents were coded on a 1–6 scale from most professional occupations to unskilled labor. Any parents receiving government benefits were coded as a 7. The highest of either parent’s occupation was used to code SES. Codes 1 and 2 were combined to form the “high” SES group, codes 3 and 4 were combined to form a “middle” group and the remaining codes comprised the “low” group.

The age at which both parents left school was obtained at birth. Both maternal and paternal school-leaving ages were dichotomized into those who left before the age of 16 (the age that students acquire basic secondary school certification and are legally allowed to leave formal education in New Zealand) and those who left at the age of 16 or over. The Family Support Scale (FSS; Dunst & Trivette, 1986) measures different sources of support for families rearing young children. The FSS contains 18 items that are scored on Likert scale ranging from 1 (not at all helpful) to 5 (extremely helpful). Dunst and Trivette (1986) reported a five-factor structure of the FSS (formal kinship, informal kinship, social groups, professionals, professional groups). The informal family support score (IFSS) quantifies the presence of informal social support networks across the subscales. Finally, maternal age at birth of child was obtained through obstetric records at birth. For the purposes of these analyses, it was included as a continuous variable. All variables (predictor, outcome, and covariate) are shown in Figure 1, contextualized within Bronfenbrenner’s ecological model of development.

Figure 1.

Position of predictors/covariates of interest within Bronfenbrenner’s (1979; 1994) bioecological model, with individual outcomes of interest shown.

Data Analysis

SDQ scores were initially subject to logarithmic transformation due to non-normal distribution. Given similarity in results obtained from analyses of transformed and untransformed SDQ scores (data not shown), analyses using the untransformed scores are reported here. Next, we screened data for use in moderated linear regression analysis. Variance among groups (i.e., among different levels of categorical predictors) were suitably heterogeneous, and there was no presence of heteroscedasticity across predictor variables. Each variable in final models related to SDQ scores in a linear manner. The ABC cohort is selectively sampled with regard to birthweight such that there are approximately equal number of SGA and AGA participants. To address this, we initially explored whether any relationships between paternal engagement, paternal accessibility, and behavioral difficulties scores differed among birthweight groups. We retained birthweight as a covariate to account for any direct effects of birthweight on behavior (Pryor et al., 1995).

As per ABC study protocol, initially all predictor variables and covariates were initially regressed against outcome variables of interest. If covariates met a relaxed significance threshold (p = .10) in individual regression analyses, they were considered for inclusion in multivariate models. Among these shortlisted variables, bivariate correlations were conducted to assess for multicollinearity effects. For this study, multiple linear regressions were conducted to ensure that effects of paternal accessibility as the primary predictor held in the presence of covariates. This was true only for the total difficulties, conduct, and hyperactivity SDQ scores. All of these analyses are provided in appendices (A, B, and C).

Finally, hierarchical, moderated regression analyses were conducted to explore possible interaction effects between paternal accessibility (measured as fathers’ presence in same household as the child) and paternal engagement (at 1 year, and 3.5 years of age) on child behavioral difficulties scores (total, conduct, hyperactivity) at 11 years of age. Included covariates were birthweight, child sex, mothers’ perceived stress, family income, parental schooling, and mother’s age at child’s birth. The conditional effects of paternal accessibility were assessed at 1 standard deviation (SD) below the mean, the mean, and 1 SD above the mean level of paternal engagement. Given the use of continuous moderator, the Johnson–Neyman technique was applied post hoc to highlight turning points, where the effects of paternal accessibility on behavioral difficulties became statistically significant.

Statistical significance for all multivariable analyses was given at an alpha level of 0.05. All analyses were conducted with IBM SPSS statistics software, version 23.0.

Results

Sample sizes and descriptive statistics for all moderated regression models are shown in Table 1. Significant moderations are reported here. Full description of all moderation analyses is provided in Appendix C.

Table 1.

Sample Sizes, Mean Paternal Engagement, and SDQ Behavioral Difficulties Scores for Multiple, Moderated Regression Models.

Model	Total Difficulties			Conduct Problems			Hyperactivity
Model	N	Total DifficultiesM (SD)	Engagement(Hrs/Wk)M (SD)	N	ConductM (SD)	Engagement(Hrs/Wk)M (SD)	N	HyperactivityM (SD)	Engagement(Hrs/Wk)M (SD)
Paternal accessibility and engagement (1 year)	503	6.95 (5.03)	8.92 (6.57)	514	1.24 (1.49)	8.90 (6.55)	509	2.82 (2.29)	8.98 (6.66)
Paternal accessibility and engagement (3.5 years)	420	7.14 (5.19)	10.57 (9.66)	434	1.24 (1.44)	10.52 (9.57)	427	2.86 (2.31)	10.60 (9.69)

Note. N refers to sample of cases with responses to all predictor, outcome, and covariate items. SDQ = Strengths and Difficulties Questionnaire (Goodman, 1997).

Described fully in Table 2, moderated regression analyses of paternal engagement at 1 year of age showed that the model as a whole explained a significant proportion of the variance in children’s total behavioral difficulties scores (R² = 0 .32, F_{(10, 492)} = 5.84, p < .001). There was no direct effect of either paternal accessibility, or paternal engagement, on total behavioral difficulties. There was, however, an interaction effect whereby paternal engagement moderated the effect of paternal accessibility on behavioral difficulties (p = .02). Post hoc simple slope analyses examined the conditional effects of paternal accessibility at three levels of paternal engagement (1 SD below the mean, at the mean, and 1 SD above the mean). As illustrated in Figure 2, paternal accessibility was significantly related to behavioral difficulties at 11 years of age when paternal engagement was 1 SD above the mean (B = 1.53, p = .03, 95% CI = 0.19, 2.87) but not when at the mean value of paternal engagement (B = 0.39, p = .36, 95% CI = −0.44. 1.22) or 1 SD below the mean (B = −0.75, p = .22, 95% CI = −1.95, 0.45).

Table 2.

Hierarchical, Moderated Linear Regression Analysis on the Effects of Paternal Accessibility and Paternal Engagement (1 Year of Age) on Total Behavioral Difficulties Scores.

	Model 1				Model 2
Variable	B (SE)	95% CI		p	B (SE)	95% CI		p
Constant	7.42 (1.86)				7.53 (1.90)
Birthweight	0.51 (0.45)	−0.37	1.38	.25	0.47 (0.46)	−0.44	1.37	.31
Child sex	1.99 (0.44)	1.13	2.85	<.001**	2.04 (0.45)	1.17	2.92	<.001**
Perceived stress	0.08 (0.03)	0.01	0.14	.02*	0.07 (0.03)	0.01	0.13	.03*
Family income	1.65 (0.62)	0.42	2.87	.01*	1.61 (0.68)	0.26	2.97	.02*
Maternal schooling	0.63 (0.92)	−1.17	2.43	.49	0.67 (0.78)	−0.87	2.20	.39
Paternal schooling	0.21 (0.71)	−1.19	1.60	.77	0.31 (0.64)	−0.95	1.56	.63
Mother’s age	−0.09 (0.05)	−0.19	0.02	.10	−0.09 (0.06)	−0.20	0.02	.10
Paternal accessibility	0.31 (0.41)	−0.49	1.11	.45	0.39 (0.42)	−0.44	1.22	.36
Paternal engagement (1 year)	−0.03 (0.03)	−0.10	0.04	.40	−0.02 (0.03)	−0.09	0.05	.56
Accessibility × engagement					0.17 (0.07)	0.03	0.32	.02*
R ²		0.08				0.10

Note. Model 1 is the initial multiple linear regression; Model 2 includes the added accessibility engagement interaction variable.

Figure 2.

The conditional effect of paternal accessibility on children’s total behavioral difficulties with varying levels of paternal engagement.

A Johnson–Neyman analysis revealed that the relationship between paternal accessibility became significant at the value of 3.8 hours of paternal engagement per week (p = .03) and remained significant for all values above this. Children in families earning under the median wage at the time of their birth scored 1.6 points higher SDQ scores on average (p = .02). There was a significant, but small positive association between mothers’ antenatal stress on SDQ scores (p = .03). Boys obtained SDQ scores approximately 2 points higher than girls across the sample (p = <.001).

Described fully in Table 3, moderated regression analyses of paternal engagement at 1 year of age showed that the model as a whole explained a significant proportion of the variance in children’s conduct problems scores (R² = 0.08, F_{(11, 503)} = 4.20, p < .001). There was no direct effect of either paternal accessibility or paternal engagement on conduct problems scores. There was, however, an interaction effect whereby paternal engagement moderated the effect of paternal accessibility on conduct problems (p = .002). Post hoc simple slope analyses examined the conditional effects of paternal accessibility at three levels of paternal engagement (1 SD below the mean, at the mean, and 1 SD above the mean). As illustrated in Figure 3, paternal accessibility was significantly related to conduct problems when paternal engagement was 1 SD above the mean (B = 0.54, p = .003, 95% CI = 0.19, 0.90) but not when at the mean value of paternal engagement (B = 0.15, p = .20, 95% CI = −0.08, 0.38) or one standard deviation below the mean (B = −0.24, p = .15, 95% CI = −0.57, 0.09).

Table 3.

Hierarchical, Moderated Linear Regression Analysis on the Effects of Paternal Accessibility and Paternal Engagement (1 Year of Age) on Conduct Problems Scores.

	Model 1				Model 2
Variable	B (SE)	95% CI		p	B (SE)	95% CI		p
Constant	2.34 (0.61)				2.35 (0.58)
Birthweight	0.14 (0.13)	−0.13	0.40	.31	0.13 (0.13)	−0.13	0.39	.32
Child sex	0.41 (0.13)	0.15	0.66	.002**	0.42 (0.13)	0.17	0.67	.001*
Perceived stress	0.02 (0.01)	−0.01	0.03	.14	0.01 (0.01)	−0.01	0.03	.22
Family income	0.25 (0.19)	−0.11	0.62	.18	0.24 (0.19)	−0.13	0.60	.20
Socioeconomic status	0.39 (0.38)	−0.35	1.12	.30	0.48 (0.37)	−0.25	1.21	.20
Maternal schooling	−0.40 (0.27)	−0.56	0.48	.88	−0.03 (0.26)	−0.55	0.49	.91
Informal social support	−0.01 (0.01)	−0.03	0.01	.16	−0.01 (0.01)	−0.03	0.01	.20
Mother’s age	−0.04 (0.15)	−0.07	−0.01	.01*	−0.04 (0.02)	−0.07	−0.01	.001*
Paternal accessibility	0.12 (0.12)	−0.11	0.35	.31	0.15 (0.12)	−0.08	0.38	.20
Paternal engagement (1 year)	−0.01 (0.01)	−0.03	0.01	.35	−0.01 (0.01)	−0.03	0.01	.54
Accessibility × Engagement					0.06 (0.02)	0.02	0.10	.002**
R ²		0.09				0.10

Note. Model 1 is the initial multiple linear regression; Model 2 includes the added accessibility engagement interaction variable.

Figure 3.

The conditional effect of paternal accessibility on children’s conduct problems with varying levels of paternal engagement.

Johnson–Neyman analysis revealed that the relationship between paternal accessibility and conduct problems became significant at the value of 3.4 hours of paternal engagement per week (p = .01) and remained significant for all values above this. Boys obtained SDQ scores approximately 0.4 points higher than girls across the sample (p = .001). There was a significant, but small negative association between mothers’ age and conduct problems scores (p = .001).

Discussion

The aim of the present study was to determine if paternal engagement moderated the relationship between early paternal accessibility (i.e., a father’s presence in the same household as the child) and later behavioral difficulties. We observed, contrary to our primary hypothesis, that if a father had been highly engaged with the child in the first year of life (more than 4 hours per week), later total behavioral difficulties, and conduct problems scores were significantly, negatively correlated with paternal accessibility. If fathers were engaged at a low or average level, there was no significant effect of paternal accessibility. Given that higher levels of paternal engagement could represent fathers’ playing a greater role in the proximal process of children’s development (Bronfenbrenner, 1994; Pleck, 2010), it is possible that that subsequent paternal household departure compromises an established relationship. This subsequently may have a negative, long-term effect on children’s behavior. This result suggests that fathers, by being highly engaged during the first year of their children’s lives, establish themselves as an integral part of children’s microsystem of development (Bronfenbrenner, 1994). It is plausible that their subsequent departure represents a vacancy, or a significant reduction in the quality of developmental proximal process that must then be accounted for within the wider network of children’s developmental ecology.

Our findings are somewhat consistent with prior longitudinal investigations that have observed effects of both paternal accessibility (Carlson, 2006; Craig et al., 2018; Vaden-Kiernan et al., 1995) and paternal engagement (Aldous & Mulligan, 2002; McMunn et al., 2017; Ramchandani et al., 2010) as predictors of children’s behavioral conduct. Contrary to recent findings, we did not observe any relationship between the dimensions of paternal involvement and peer interaction problems (Flouri et al., 2016; Ramchandani et al., 2010).

Our findings suggest that the long-term effects of paternal accessibility and paternal engagement on children’s behavior may not be additive. Recent investigations are consistent in their broad findings that the greater the amount of paternal accessibility or engagement in a child’s life, the more positive their behavioral outcomes over time (McMunn et al., 2017; Ramchandani et al., 2013; Sarkadi et al., 2008). Alternatively, our findings introduce the possibility that it may not be a matter of a greater quantity of positive paternal engagement, and that the timing of this engagement could also be an important factor of consideration.

In this regard, we observed no interactions between paternal accessibility and paternal engagement when children were 3.5 years of age. This suggests that paternal engagement during the child’s first year of their life may be a more important predictor of long-term behavioral outcomes. It may be that father’s steadily become a smaller part of children’s microsystem of developmental influence (i.e., as their overall developmental ecology expands), and therefore paternal involvement becomes less significant over time. In any case, these findings draw focus to the generally vital first year of life with regard to the engagement of fathers and potential long-term effects. Generally, we furthered recent findings regarding paternal accessibility and total behavioral difficulties in observing similar relationships between paternal accessibility and children’s conduct, and hyperactivity problems (Craig et al., 2018).

It is noteworthy that the significant interactions between paternal accessibility and engagement on child behavior were present in the context of a number of influential covariates. While sex, family income, maternal stress, and mothers’ age when giving birth all exhibited significant direct effects on behavioral difficulties, they did not confound the interaction effect between paternal accessibility and engagement. Particularly with regard to family income, as a measure of socioeconomic status, it is noteworthy that our results were significant despite the discussed confounding financial impact that a father’s departure from the household can have on family income (Sarkadi et al., 2008).

Our study had several limitations. First, the ABC study cohort is selectively sampled for birthweight, which ensures that children born SGA are overrepresented with regard to population norms. Although we controlled for any birthweight effects in this investigation, it may be that the unique nature of the sample prevents these findings from being generalizable to larger populations. Likewise, this investigation was conducted with an exclusively NZ/European sample. As such, we encourage the replication of our results with larger and more diverse population cohorts.

Our measures of paternal engagement and accessibility were limited by the scope of the ABC, and consisted of only one item each. Although they were both appropriately suited to Lamb et al.’s (1987) tripartite conceptualization of paternal involvement, they were not more thoroughly able to capture a father’s proximity and availability to his child, nor the specific type and quality of his engagement interactions with them. Such measures are comparable to those used throughout the existing paternal involvement literature (Carlson, 2006; Flouri et al., 2016; Sarkadi et al., 2008); however, added specificity with regard to the nature of the engagement between fathers and their children would have added more depth to our results. Our assessment of children’s behavior was limited to parent-report SDQ data. Obtaining data from other reports (e.g., children or teachers) would have added further to our findings.

Finally, our investigation was unable to explore further the finding that the observed interaction effect only held for paternal engagement in the child’s first year of life. Subsequent investigations should attempt to explain the mechanism by which engagement at 1 year moderated paternal accessibility, but not engagement at 3.5 years. Taken as a whole, our findings suggest that the relationship between paternal accessibility, paternal engagement, and child behavior is complex; therefore, it is important to consider both the quantity and timing of fathers’ involvement with their children.

Supplemental Material

sj-pdf-1-jfi-10.1177_0192513X20980128 – Supplemental material for The Father I Knew: Early Paternal Engagement Moderates the Long-term Relationship between Paternal Accessibility and Childhood Behavioral Difficulties

Supplemental material, sj-pdf-1-jfi-10.1177_0192513X20980128 for The Father I Knew: Early Paternal Engagement Moderates the Long-term Relationship between Paternal Accessibility and Childhood Behavioral Difficulties by Angus G. Craig, John M. D. Thompson, Rebecca Slykerman, Clare Wall, Rinki Murphy, Edwin A. Mitchell and Karen E. Waldie in Journal of Family Issues

Footnotes

Acknowledgements

We sincerely thank the parents and children for participating in this study. This is an original article and there are no copyright issues. There is no material that has appeared in other publications. We acknowledge the assistance of Gail Gillies, Barbara Rotherham, and Helen Nagels for contacting or assessing the participants.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The initial study was funded by the Health Research Council of New Zealand. The 12 month postal questionnaire was funded by Hawke’s Bay Medical Research Foundation. The 3.5-year follow-up study was funded by Child Health Research Foundation, Becroft Foundation, and Auckland Medical Research Foundation. The 7-year follow-up study was funded by Child Health Research Foundation. The 11-year follow-up was funded by Child Health Research foundation and the Heart Foundation. The genetic component of this study was funded by Child Health Research Foundation. E. A. Mitchell and J. M. D. Thompson were supported by Cure Kids. The 7-year follow-up study was conducted in the Children’s Research Centre which is supported in part by the Starship Foundation and the Auckland District Health Board.

ORCID iD

Angus G. Craig

Supplemental Material

Supplemental material for this article is available online.

References

Aldous

Mulligan

G. M.

(2002). Fathers’ child care and children’s behavior problems. Journal of Family Issues, 23(5), 624–647. https://doi.org/10.1177/0192513X02023005003

Armsden

G. C.

McCauley

Greenberg

M. T.

Burke

P. M.

Mitchell

J. R.

(1990). Parent and peer attachment in early adolescent depression. Journal of Abnormal Child Psychology, 18(6), 683–697. https://doi.org/10.1007/BF01342754

Bronfenbrenner

(1979). Contexts of child rearing: Problems and prospects. American Psychologist, 34(10), 844.

Bronfenbrenner

(1994). Ecological models of human development. International Encyclopedia of Education, 3(2), 37–43.

Carlson

M. J.

(2006). Family structure, father involvement, and adolescent behavioral outcomes. Journal of Marriage and Family, 68(1), 137–154. https://doi.org/10.1111/j.1741-3737.2006.00239.x

Castillo

Welch

Sarver

(2011). Fathering: The relationship between fathers’ residence, fathers’ sociodemographic characteristics, and father involvement. Maternal and Child Health Journal, 15(8), 1342–1349. https://doi.org/10.1007/s10995-010-0684-6

Chang

J. J.

Halpern

C. T.

Kaufman

J. S.

(2007). Maternal depressive symptoms, father’s involvement, and the trajectories of child problem behaviors in a US national sample. Archives of Pediatrics & Adolescent Medicine, 161(7), 697–703. https://doi.org/10.1001/archpedi.161.7.697

Cohen

Williamson

(1988). Perceived stress in a probability sample of the US. In Spacapam

Oskamp

(Eds.), The social psychology of health: Claremont symposium on applied social psychology (pp. 31–67). SAGE Publications.

Coleman

J. S.

(1988). Social capital in the creation of human capital. American Journal of Sociology, 94, S95–S120.

10.

Craig

A. G.

Thompson

J. M.

Slykerman

Wall

Murphy

Mitchell

E. A.

Waldie

K. E.

(2018). The long-term effects of early paternal presence on children’s behavior. Journal of Child and Family Studies, 27(11), 3544–3553. https://doi.org/10.1007/s10826-018-1206-1

11.

Dunst

C. J.

Trivette

C. M.

(1986). Looking beyond the parent–child dyad for the determinants of maternal styles of interaction. Infant Mental Health Journal, 7(1), 69–80.

12.

Elley

W. B.

(1985). The Elley–Irving socio-economic index 1981 census revision. New Zealand Journal of Educational Studies, 20(2), 115–128.

13.

Elley

W. B.

Irving

J. C.

(2003). The Elley–Irving socio-economic index: 2001 census revision. New Zealand Journal of Educational Studies, 38(1), 3–17.

14.

Flouri

Buchanan

(2002). Father involvement in childhood and trouble with the police in adolescence. Journal of Interpersonal Violence, 17(6), 689–701. https://doi.org/10.1177/0886260502017006006

15.

Flouri

Midouhas

Narayanan

(2016). The relationship between father involvement and child problem behaviour in intact families: A 7-year cross-lagged study. Journal of Abnormal Child Psychology, 44(5), 1011–1021. https://doi.org/10.1007/s10802-015-0077-9

16.

Goodman

(1997) The strengths and difficulties questionnaire: A research note. Journal of Child Psychology and Psychiatry, 38, 581–586.

17.

Goodman

(2001). Psychometric properties of the strengths and difficulties questionnaire. Journal of the American Academy of Child and Adolescent Psychiatry, 40(11), 1337–1345. https://doi.org/S0890-8567(09)60543-8

18.

Goodman

Meltzer

Bailey

(1998). The strengths and difficulties questionnaire: A pilot study on the validity of the self-report version. European Child & Adolescent Psychiatry, 7(3), 125–130. https://doi.org/10.1007/s007870050057

19.

Keizer

Lucassen

Jaddoe

Tiemeier

(2014). A prospective study on father involvement and toddlers’ behavioral and emotional problems: Are sons and daughters differentially affected? Fathering, 12(1), 38. https://doi.org/10.3149/fth.1201.38

20.

Kenny

M. E.

Gallagher

L. A.

(2002). Instrumental and social/relational correlates of perceived maternal and paternal attachment in adolescence. Journal of Adolescence, 25(2), 203–219.

21.

Lamb

M. E.

(2000). The history of research on father involvement. Marriage & Family Review, 29(2–3), 23–42. https://doi.org/10.1300/J002v29n02_03

22.

Lamb

M. E.

Pleck

J. H.

Charnov

E. L.

Levine

J. A.

(1987). A biosocial perspective on paternal behavior and involvement. In Lancaster

J. B.

Altmann

Rossi

A. S.

Sherrod

L. R.

(Eds.), Parenting across the life span: Biosocial dimensions (pp. 111–142). Aldine Publishing.

23.

McMunn

Martin

Kelly

Sacker

(2017). Fathers’ involvement: Correlates and consequences for child socioemotional behavior in the united kingdom. Journal of Family Issues, 38(8), 1109–1131. https://doi.org/10.1177/0192513X15622415

24.

Nau

Heckert

(2013). Integrating perspectives on child health: Where do we go now? In Landale

McHale

Booth

(Eds.), Families and child health (pp. 213–226). Springer.

25.

Pleck

J. H.

(2010). Paternal involvement: Revised conceptualization and theoretical linkages with child outcomes. In Lamb

M. E.

(Ed.), The role of the father in child development (5th ed., pp. 459–485). Wiley.

26.

Pleck

J. H.

Masciadrelli

B. P.

(2007). Paternal involvement by U.S residential fathers: Levels, sources, and consequences. In Lamb

M. E.

(Ed.), The role of the father in child development (4th ed., pp. 222–271). Wiley.

27.

Pryor

Silva

P. A.

Brooke

(1995). Growth, development and behaviour in adolescents born small-for-gestational-age. Journal of Paediatrics and Child Health, 31(5), 403–407. https://doi.org/10.1111/j.1440-1754.1995.tb00847.x

28.

Ramchandani

P. G.

Domoney

Sethna

Psychogiou

Vlachos

Murray

(2013). Do early father–infant interactions predict the onset of externalising behaviours in young children? findings from a longitudinal cohort study. Journal of Child Psychology and Psychiatry, 54(1), 56–64. https://doi.org/10.1111/j.1469-7610.2012.02583.x

29.

Ramchandani

P. G.

IJzendoorn

M. V.

Bakermans-Kranenburg

M. J.

(2010). Differential susceptibility to fathers’ care and involvement: The moderating effect of infant reactivity. Family Science, 1(2), 93–101. https://doi.org/10.1080/19424621003599835

30.

Sarkadi

Kristiansson

Oberklaid

Bremberg

(2008). Fathers’ involvement and children’s developmental outcomes: A systematic review of longitudinal studies. Acta Paediatrica, 97(2), 153–158. https://doi.org/10.1111/j.1651-2227.2007.00572.x

31.

Sayer

L. C.

Bianchi

S. M.

Robinson

J. P.

(2004). Are parents investing less in children? trends in mothers’ and fathers’ time with children. American Journal of Sociology, 110(1), 1–43. https://doi.org/10.1086/386270

32.

Slykerman

R. F.

Thompson

Waldie

Murphy

Wall

Mitchell

E. A.

(2015). Maternal stress during pregnancy is associated with moderate to severe depression in 11-year-old children. Acta Paediatrica, 104(1), 68–74.

33.

Thompson

J. M.

Clark

P. M.

Robinson

Becroft

D. M.

Pattison

N. S.

Glavish

Pryor

J. E.

Wild

C. J.

Rees

Mitchell

E. A.

(2001). Risk factors for small-for-gestational-age babies: The Auckland birthweight collaborative study. Journal of Paediatrics and Child Health, 37(4), 369–375. https://doi.org/10.1046/j.1440-1754.2001.00684.x

34.

Vaden-Kiernan

Ialongo

N. S.

Pearson

Kellam

(1995). Household family structure and children’s aggressive behavior: A longitudinal study of urban elementary school children. Journal of Abnormal Child Psychology, 23(5), 553–568. https://doi.org/10.1007/BF01447661

35.

Wilson

G. B.

(2006). The non-resident parental role for separated fathers: A review. International Journal of Law, Policy and the Family, 20(3), 286–316.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.16 MB

0.00 MB

0.16 MB