Commentary: Early father–infant interaction and externalizing behaviors – a response to Ramchandani et al. (2013)

Authors


Abstract

Many researchers have attempted to uncover the precise contribution of fathers to childrearing in relation to both young and older children’s development during the past five decades (Lamb, 1975), including during the infancy period (Parke & O’Leary. S, 1975). However, few have been able to isolate precise mechanisms by which specific types of paternal childrearing practices may be linked to specific types of prosocial and problem behavior. The current paper by Ramchandani et al. (2013) breaks new ground in identifying a precise dimension of paternal parenting during early infancy - engagement - and linking it to maternal reports of infant externalizing problem behavior. Importantly, this association was found after accounting for the influence of several child, paternal, and maternal characteristics, including observed maternal sensitivity. Specifically, the authors found that fathers observed to be less engaging with their 3-month olds were reported by mothers to show fewer disruptive problems 9 months later.

Many researchers have attempted to uncover the precise contribution of fathers to childrearing in relation to both young and older children’s development during the past five decades (Lamb, 1975), including during the infancy period (Parke & O’Leary, 1975). However, few have been able to isolate precise mechanisms by which specific types of paternal childrearing practices may be linked to specific types of prosocial and problem behavior. The current article by Ramchandani and colleagues breaks new ground in identifying a precise dimension of paternal parenting during early infancy – engagement – and linking it to maternal reports of infant externalizing problem behavior. Importantly, this association was found after accounting for the influence of several child, paternal, and maternal characteristics, including observed maternal sensitivity. Specifically, the authors found that fathers observed to be less engaging with their 3-month olds were reported by mothers to show fewer disruptive problems 9 months later. Furthermore, these associations appeared to be more strongly evident for male infants versus female infants. Although the precise meaning and implications of elevated externalizing problems at age 1 remain unclear (see discussion below), the authors’ achievement in identifying a paternal-driven correlate of early problem behavior is still notable. Several substantive and methodological features of the study are worth highlighting. These attributes may have increased the authors’ ability to successfully identify a paternal childrearing dimension predictive of early problem behavior, whereas others have failed to do so in the past.

The first two issues involve the authors’ decision to focus on remoteness and low engagement as a key construct for fathers during early infancy and relatedly, the creativity the authors demonstrated in measuring this factor. On the basis of feedback from piloting the parent–infant interaction procedure, the authors found that using a traditional alignment in which infants are placed in a car seat directly across from parents (i.e., typically mothers) inhibited fathers’ normative use of physical and more active play strategies. Instead, they opted to set infants on a mat, providing fathers with more opportunities for physical and active play. Note associations between low engagement and later problem behavior were only found for father–infant interactions in the floor mat setting, not with the car seat. Thus, the importance of this innovation in measurement cannot be underestimated and should be taken into account in assessing dimensions of paternal caregiving with young and older children by other researchers. An analogous finding was identified with fathers of older children in applying the commonly used family discussion task, in which parents/youth identify areas of disagreement and are then instructed to discuss this issue for 8–10 min. Fathers were found to be much more expressive when these discussions were convened in the home (i.e., on their own turf) than in the lab.

Another methodological issue that might have influenced the authors’ ability to identify an association between paternal engagement and later problem behavior was the sample’s composition, which was somewhat atypical in multiple ways. Fathers are not typically eager to engage in parent–child interaction studies, particularly during infancy when they traditionally spend disproportionately less time with children relative to mothers than during later developmental periods. As the authors highlight, the sample included an older (M age = 35) and highly educated sample, making their engagement in the project more likely than younger, lower SES fathers. An issue less discussed in the article is father’s level of psychopathology, which might have affected the variability in father–infant interaction and pattern of associations with later problem behavior. Approximately 3/4 of fathers had scores above 10 on the Edinburgh Postnatal Depression Scale when recruited. Assuming some stability in depressive symptoms during the first 3 months, the sample may have had atypically high variability in levels of father engagement by oversampling of fathers likely to show unusually low levels of engagement because of their atypically high levels of depressive symptoms. As a result, at the 3-month assessment, fathers who showed more normative levels of engagement during the mat setting play task likely would have stood out more than in a sample of fathers not initially screened on the basis of depression. By recruiting fairly extreme groups of participating fathers (but not mothers), the authors may have increased their probability of linking a paternal caregiving dimension to subsequent child behavior. Follow-up data on this issue using the current sample or an independent sample could provide important information about these speculations, including the degree of association between initial levels of paternal depressive symptoms and paternal engagement in the 3-month interaction task, and how stability in engagement and/or paternal depression during the first 12 months may have contributed to maternal perceptions of externalizing at 12 months (e.g., independent contributions of depression and engagement, depression x engagement interaction).

The dependent variable in the current study, mothers’ perceptions of externalizing problems at 1 year of age, also merits discussion. Although an acceptable internal consistency for the Externalizing factor was evident, reasonable correlations between maternal and paternal reports were found, and the 1.5- to 5-year version of the CBCL has been used by other researchers with 1-year olds, from a cognitive perspective many of the disruptive behaviors demonstrated by 1-year olds have different meanings for infants than they would for slightly older children (i.e., 2-year olds). Thus, although 1-year olds clearly engage in aggressive-like behavior, their reasons for doing so might differ significantly from 2- to 3-year olds who have a more advanced cognitive appreciation of the effects of their physical assaults on others than infants. This issue is important because across informant, longitudinal associations between maternal perceptions of externalizing and later disruptive behavior only begin to show modest predictive validity between the ages of 2 and 3 (see Shaw & Gross, 2008). Despite this important caveat, the authors have clearly identified a phenomenon (i.e., infant externalizing problems) predicted by earlier paternal engagement whose relationship cannot be attributed to informant or method bias. Theoretically, because of the association with paternal engagement, I tend to agree with the authors that higher levels of problem behavior at age 1 may reflect a history of less than optimal involvement by one or both parents, as infants strive to get their attentional needs met by showing relatively higher levels of disruptive behavior.

Increasing the ‘mystique’ around the meaning of age-1 externalizing problems even more was the dearth of associations with several risk factors traditionally linked to child externalizing problems during the toddler and preschool period, including parental depression, maternal sensitivity, paternal antisociality, parental employment, and infant irritability. This consistent pattern of null findings for traditional correlates of early child problem behavior may reflect the uncertainty about the precise meaning of externalizing symptoms at age 1. Alternatively, that paternal engagement was one of only two factors (i.e., the other being paternal age) found to distinguish disruptive infants suggests the salience of fathers’ contributions to the development of children’s early self-regulation abilities, particularly for boys.

In summary, the authors should be commended for making this novel and innovative contribution to our understanding of paternal caregiving factors related to children’s early adjustment. As with other important contributions, the findings raise as many questions as they address. For example, although no direct effects were found for maternal sensitivity, are there interactive effects between paternal engagement and maternal caregiving factors that amplify or attenuate associations between paternal engagement and child externalizing? Similarly and as noted above, it would be helpful to know more about bidirectional patterns of association between paternal depressive symptoms and paternal remoteness over time, particularly within the current sample for which the majority of fathers reported elevated depressive symptoms at recruitment. With much anticipation, I look forward to hearing about follow-up studies on these issues and other related topics in the authors’ future research efforts.

Acknowledgement

This commentary article was invited by the Editors of JCPP; while it has gone through a process of editorial review, it has not been formally peer reviewed. The author has declared that he has no competing or potential conflicts of interest.

Correspondence

Daniel Shaw, University of Pittsburgh, 210 South Bouquet Street, 4101 Sennott Square, Pittsburgh, PA 15260-0001, USA; Email: casey@pitt.edu

Ancillary