Parents’ Fathering Identity Standards and Later Father Involvement

Abstract

Based on propositions from identity theory, this study used a sample of 1,596 coresident couples from the Fragile Families and Child Well-Being Study to examine whether parents’ fathering role centrality standards and fathers’ status centrality at the time of their child’s birth were associated with mother and father reports of father involvement 1 and 3 years later. Mothers and fathers who rated fathering roles as more important reported greater father involvement at both Years 1 and 3; centrality of the father status was associated with father reports of involvement at both years, and mother-reported involvement at Year 3. Interactions between fathers’ and mothers’ role centrality standards, and between parents’ role centrality standards and father status centrality, were found for mother reports of involvement at Year 3. Implications for research, practice, and theory are discussed.

Keywords

father-child relationship dyadic relationship/quality/satisfaction family processes gender & family family theory

The ways in which involved fathers potentially benefit children are well documented (Cabrera, Shannon, & Tamis-LeMonda, 2007; Marsiglio, Amato, Day, & Lamb, 2000). Research also demonstrates that there is wide variation in involvement among fathers (Amato, Meyers, & Emery, 2009; Hofferth, Pleck, Stueve, Bianchi, & Sayer, 2002). As such, scholars turned their attention to examining the factors that foster or inhibit father involvement. Understanding which factors affect father involvement is important among both married and unmarried parents, as currently 40% of births are nonmarital (Hamilton, Martin, & Ventura, 2011), and unmarried fathers exhibit both greater variation and lower overall levels of involvement compared with married fathers (Coley & Chase-Lansdale, 1999; Lerman & Sorensen, 2000).

One potential influence on father involvement is parental beliefs about the importance of father involvement in particular parenting roles (role-level identity standards) and the importance assigned to being a parent (status-level identity standards). However, research findings are mixed regarding when, how, and to what extent mothers’ and fathers’ role and status-level identity standards influence involvement. Also, research only has examined mothers’ and fathers’ identities individually, rather than considering how their interaction might influence father involvement. Here, we examined the associations of mothers’ and fathers’ fathering identity standards at the time of their child’s birth with both parents’ reports of later father involvement. We also tested whether identity characteristics interacted to uniquely influence reports of father involvement.

Identity, Centrality, and Fathering Behavior

Identity theory (Stryker, 1968, 1980) suggests that one factor influencing father involvement is fathers’ sense of what it means to be a parent (father identity; e.g., Ihinger-Tallman, Pasley, & Buehler, 1993). A derivation of symbolic interactionism, identity theory suggests that individuals construct meaning from interactions with others, and this meaning provides clues regarding societal expectations of the social statuses (e.g., parent) they occupy. These expectations (roles; e.g., breadwinner, disciplinarian) are interpreted by the individual, resulting in the creation of identities (self-meanings in a role; Burke & Reitzes, 1981), which become evident in behavior. The enactment of an identity further is affected by the level of importance an individual assigns to that identity (centrality), as more important identities are more likely to be enacted (Stryker & Serpe, 1994). An important distinction is warranted here; although Burke and Reitzes (1981) and Stryker (1968, 1980) implied that identities occur primarily at the level of roles, other research suggests that statuses also have associated identities and importance (Rane & McBride, 2000). Theoretically, men who view the status of being a father as unimportant are less likely to exhibit fathering behaviors in any form. Similarly, men who assign little importance to particular fathering roles are less likely to exhibit behaviors reflecting those roles.

Research attempting to link fathering identities and behavior has resulted in mixed findings (McBride et al., 2005), largely due to the variety of ways in which identity is operationalized from study to study. For example, Ihinger-Tallman and colleagues (1993) found a positive relationship between a combined measure of status-level fathering satisfaction, competence, investment, and identity salience and father involvement in child-related activities. Rane and McBride (2000) assessed the centrality of both the status (father) and specific roles (e.g., nurturer, breadwinner) and found that status centrality was not associated with involvement, but the centrality of the nurturer role was. Maurer, Pleck, and Rane (2003) also found no associations between father caregiving behaviors (akin to nurturing behaviors) and fathering status centrality. The above findings from Rane and colleagues demonstrate the importance of asking about the centrality of specific fathering role identities, rather than just the centrality of the father status.

Another reason behind inconsistent findings is differences in the scope of identity measures and measures of fathering behaviors. For example, some studies have associated status level identities (overall competence at being a father) with role-level behaviors (changing diapers—caregiving role), rather than matching role identities with role behaviors (competence at caregiving with changing diapers). Other studies that assessed both identities and behaviors at the level of roles often tapped different roles (e.g., associating caregiving identity standards with financial provision behaviors). One of the few studies to carefully match their father identity and behavior measures, Mauer, Pleck and Rane (2001) assessed caregiving and breadwinning identity standards (importance/centrality) and caregiving and breadwinning behaviors. They found that neither fathers’ nor mothers’ caregiving identity standards for fathers predicted fathers’ caregiving behavior; rather, father caregiving behavior was predicted by fathers’ perceptions of mothers’ caregiving identity standards for fathers. Unfortunately, these studies did not include mother reports of father involvement or father reports of mother involvement, so influences on varying perceptions of involvement could not be assessed. In this study, we assess both role and status identity standards and their associations with involvement, and we also assessed similar roles for both the identity and involvement (direct interactions/caregiving, mentoring, and showing love/affection).

The above findings suggest the importance of examining the fathering identity standards of both parents, although support is mixed for the proposition that maternal expectations of fathers are associated with father involvement. Some studies find that fathers are less involved when mothers are unsupportive and act in ways that discourage father involvement (maternal gatekeeping; Allen & Hawkins, 1999). However, most such studies have used small convenience samples; studies using representative samples have not obtained similar results (Hofferth, 2003).

Maternal gatekeeping research is limited in other important ways. We know little regarding mothers’ beliefs about specific fathering roles (role-level identity standards), as studies have focused on the importance of overall involvement or father competence (status-level identity standards). Studies have compared mother and father reports of involvement, typically finding that fathers report higher levels than mothers, although differences lessen when assessments are corrected for times when mothers cannot observe father involvement (Wical & Doherty, 2005). However, no studies have examined parents’ perceptions of father involvement and how such perceived behavior might relate to identity. Finally, only mothers’ beliefs typically are examined in the gatekeeping literature, failing to consider mothers’ and fathers’ beliefs jointly. This ignores the reciprocal nature of family interactions and the dyadic processes suggested by identity theory (Burke, 1991, 1997). To address this gap, we obtain reports of identity standards and father involvement from both mothers and fathers, and we analyze the contributions of such variables individually and in interaction with one another.

Individual and Joint Identity Standards and Behavior

Identity theory assumes that identities are constructed through social interactions. Burke’s (1991, 1997) self-verification model provides one explanation of the mechanisms through which identities are interactively negotiated and maintained. Building on earlier work by Swann and colleagues (e.g., Swann, 1987; Swann & Read, 1981), Burke suggested that individuals hold identity standards (ideals regarding the enactment of particular roles), and act in ways that they believe reflect these standards. In response to such behaviors, individuals receive feedback from others, which they perceive reflects that their behaviors are consistent with or discrepant from their identity standards. Theoretically, consistent feedback results in identity confirmation, whereas discrepancy creates distress. In response to short-term distress, individuals typically alter their behavior in an attempt to regain consistency (i.e., convince the feedback source that they are acting in appropriate ways). However, when faced with continued discrepant feedback and chronic distress, individuals alter their identity standards, dismiss or ignore the feedback, or withdraw physically or psychologically from the disconfirming relationship (Burke, 1991, 1997; Swann, De La Ronde, & Hixon, 1994).

Feedback received in response to identity-relevant behaviors is particularly influential when it is provided by someone holding a counteridentity. Counteridentities stem from the fact that identities are assumed to exist in pairs (e.g., one is a husband because someone is a wife; one is an employee because someone is an employer; Burke & Cast, 1997). Although not explicitly stated as part of Burke’s model, Adamsons (2010) suggested that, for each status held, a person implicitly holds two sets of identity standards: one delineating a person’s expectations for him/herself in that identity, and one delineating that person’s expectations for the other holding the counteridentity. For example, when a woman holds the status of mother, she might believe that daily caretaking and discipline are highly central to her identity as a mother, but involvement in financial provision has low importance. This set of identity standards implies a second set of complementary role-centrality standards for the father—she holds his involvement in daily caretaking and discipline to be of low importance, but she views his financial provision as very important. As such, each set of identity standards implies an additional set of beliefs regarding what it means to be the counteridentity.

Theoretically, when individuals holding counteridentities interact, feedback to identity-relevant behavior originates from these counteridentity standards (Adamsons, 2010). For example, if a father and mother both value the breadwinner role for the father, then the father would enact identity-relevant behaviors (working long hours in paid employment), and the mother would react positively to such behaviors, providing congruent feedback. Thus, two sets of identity standards held by two individuals are involved in the enactment and calibration of each one’s individual identity standards and associated behaviors. Rather than being a simple example of identity confirmation, as proposed by Burke (1991, 1997), this process then becomes one that we label as “identity negotiation” between the two individuals. Although multiple others likely influence father identity negotiation (e.g., children, extended kin), here we focused on the influence of mothers, as her expectations and feedback likely have the most direct implications, particularly for coresident couples with young children, the focus of the present study.

Empirical support exists for the importance of considering both parents’ views. Pasley, Futris, and Skinner (2002) examined fathers’ appraisals (self-perceptions) and reflected appraisals (his perceptions of mother’s view of himself as a father) and found that the latter were associated strongly with fathering behavior, even more so than his own appraisals. This supports the hypothesis that a father’s perceptions of counteridentity standards (mothers’ expectations) will affect his behavior. However, their study was cross-sectional and did not directly assess mothers’ expectations. Maurer (2007) examined the influence of both spousal reflected appraisals and appraisals of the caregiving behavior of other same-sex parents (i.e., what fathers believe other fathers do for their children). He found that beliefs about other parents were more important than reflected appraisals. Although he assessed mothers’ beliefs, the study also was cross-sectional and did not consider mothers’ and fathers’ beliefs jointly.

Based on the foregoing, the present study focused on the associations between mothers’ and fathers’ role centrality identity standards (the importance assigned to particular roles) and father ratings of status centrality (the importance of being a father), individually and in interaction with one another, and father involvement 1 and 3 years later. Our study fills several theoretical and empirical gaps. First, identity and father involvement measures are matched to be similar in scope (i.e., both role centrality standards and measures of fathers’ behaviors tap the same roles—caretaking, teaching/mentoring, and showing affection). Second, both centrality of the father status and centrality of fathering roles are included, so the importance of role vs. status-level identity standards can be assessed and compared. Empirically, our study takes a longitudinal view of the identity-behavior link and the processes underlying this association. (Recall that most previous work is cross-sectional.) Finally, we examine such processes from a dyadic perspective, whereas prior studies focused primarily on either mothers or fathers, but not both jointly.

We hypothesized that parents’ assigning greater centrality to the fathering roles of caregiving, teaching about life, and showing love and affection, both individually and jointly, at the child’s birth, and fathers assigning higher centrality to the father status would be associated with mothers and fathers reporting more father involvement later (measured as caregiving, teaching, and affectionate behaviors). We further expected fathers’ identity standards to be more strongly related to mother and father reports of father involvement than mothers’ standards. Given the lack of prior research or theoretical propositions about how identities at different levels (role vs. status) might interact, we were unable to hypothesize about whether role centrality and status centrality would interact to affect later involvement.

Demographic Characteristics

Although our primary purpose was to test the theoretical relationship between identity and behavior among biological coresident parents, we included characteristics of fathers and children as control variables to permit a clearer test of the theoretical propositions. Fathers’ age, education, and income level are known to affect fathering; older fathers, fathers with more education, and fathers with higher incomes tend to be more involved (Pleck, 1997). Fathers’ race/ethnicity also has been associated with levels of involvement, with some research suggesting that African American and Hispanic fathers tend to be more involved with childrearing than are European American fathers (Hofferth, 2003). Finally, some research suggests that immigrant status is associated with father involvement (D’Angelo, Palacios, & Chase-Lansdale, 2012). Collectively, these characteristics were included as control variables.

Maternal characteristics often are associated with father involvement as well (Pleck, 1997), but typically they are correlated highly with paternal characteristics due to the tendency of individuals to partner with people similar to themselves. As such, we included measures representing age or education discrepancies between parents, so mothers’ unique information could be retained when discrepant from fathers’ characteristics without risking collinearity.

Several child-related factors also were included as control variables. Most research supports the effects of sex of child on father involvement—fathers tend to be more involved with sons than with daughters and demonstrate higher overall levels of involvement when sons are present (Wood & Repetti, 2004; also, see Pleck, 1997, for a review). Another important factor is the number of children fathers have, particularly if they have biological children by other mothers. Such children can compete for his time, attention, and resources, and fathers with other children are less involved than are fathers with no outside childrearing obligations (Manning, Stewart, & Smock, 2003). Presence of other biological children is important for theoretical reasons as well. If fathers have previous children, then their fathering identities likely are more established and might undergo less adaptation than if the child is his firstborn. Each of these characteristics were included as control variables.

Finally, relationship characteristics are associated with father involvement. Here, we included parents’ marital status as a control variable, because married fathers exhibit higher levels of involvement than cohabiting fathers (e.g., Landale & Oropesa, 2001).

Method

Participants

We used data from the Fragile Families and Child Well-Being Study (FFCWB; Reichman, Teitler, Garfinkel, & McLanahan, 2001), a longitudinal study of married and unmarried mothers, fathers, and their newborn children. Data were collected over 9 years from a birth cohort of approximately 5,000 families. Baseline interviews occurred between 1998 and 2000 at the time of the child’s birth, with follow-up interviews conducted when the children were 1, 3, 5, and 9 years old. We used the first three waves of data and restricted our sample to include only fathers who were resident at Years 1 and 3. Resident fathers were chosen as the focus because different processes and factors likely influence the involvement of nonresident fathers due to the additional barriers and constraints they face (Adamsons & Pasley, 2006). Also, the measures of father involvement used (frequency of involvement in activities in an average week) are most appropriate for assessing the involvement of resident fathers who have the opportunity to be engaged in such activities on a daily basis. The sample also was restricted to 18 of the 20 cities sampled, as different questions were asked in two cities. After these restrictions were imposed, our sample included 1,596 sets of parents.

The sample was relatively diverse, with 34% of mothers and 32% of fathers being White, non-Hispanic; 32% of mothers and 35% of fathers being Black, non-Hispanic; 28% of mothers/fathers being Hispanic; and 5% other races/ethnicities. The average household income at Year 1 was just under $47,000, but there was wide variation (SD = 42.98), and half of the sample had a high school education or less. At the time these children were born, just under half of the parents were married to one another, and 60% of fathers had other biological children. See Table 1 for descriptive characteristics on all demographic and key variables.

Table 1.

Means and Standard Deviations for Demographic and Key Variables for Mothers and Fathers (N = 1,596).

Variables	Mothers		Fathers
Variables	M	(SD)	M	(SD)
FRCS	8.80	(.47)	8.88	(.41)
Status centrality	----	----	3.77	(.40)
Father involvement—Y1	42.25	(12.64)	35.52	(9.10)
Father involvement—Y3	46.02	(14.08)	49.90	(12.83)
Age at Y1 (years)	26.93	(6.16)	29.40	(6.98)
Couple-level variables
Household income (Y1; thous $)	46.77	(42.98)
Household income (Y3; thous $)	53.31	(56.35)
	n	%	n	%
Education
Less than high school	370	23.2	372	23.4
High school or equiv.	446	27.9	481	30.3
Some college	442	27.7	416	26.2
College or higher	337	21.1	319	20.1
Ethnicity
White, non-Hispanic	542	34.0	508	32.0
Black, non-Hispanic	513	32.1	550	34.5
Hispanic	457	28.6	451	28.3
Other	80	5.0	83	5.2
Father has other bio children	----	----	882	60.0
Born in the U.S.	1267	79.4	1181	80.0
Couple level variables
Focal baby gender
Boy	841	52.7
Girl	755	47.3
Relat status - birth (% married)	768	48.1

Note: ---- signifies the variable was not assessed/included for that individual.

Measures

Father involvement

Father involvement was assessed using mother and father reports on seven items for fathers and nine items for mothers at Year 1, and 10 items for both parents at Year 3. Items were designed to measure the average amount of fathers’ direct involvement in various activities with their children, asking “For each activity, please tell me how many days a week you [he] do[es] this in a typical week.” To assure that our father involvement measure assessed the same roles as those assessed in the role centrality measures, items were selected that tapped the domains of direct caregiving, showing love and affection, and teaching about life, with sample items for Year 1 including “Read stories to child,” and “Hug or show physical affection to child.” Here, we included recreational involvement items with direct caregiving, as they reflected direct engagement with the child. Mothers were asked two additional items (“Change child’s diaper” and “Feed or give a bottle to child”). Sample items for Year 3 include some of the same items and some new items (e.g., “Let child help you with simple household chores”), reflecting changes in developmentally appropriate parenting. Sum scores were created for mother and father reports of father involvement for Year 1 and Year 3; alphas were .81, .70, .84, and .81, respectively. Somewhat lower alphas were expected, because many of the items reflected recreational forms of involvement (particularly at Year 1), so items relating to teaching and showing affection appeared not to fit and decreased the reliability of the measures. We chose to retain the nonrecreational items here, as we were interested in assessing a broad spectrum of father involvement behaviors, and an overall picture of involvement and alphas were acceptable.

Father role centrality standards (FRCS)

Mothers’ and fathers’ father identity standards about the centrality of various roles were assessed with six items at the time of the child’s birth using the prompt: “Fathers do many things for their children. Please tell me how important each of the following activities is to you.” Although responses addressed several roles (e.g., financial provision, being an authority figure, protecting), our goal was to match the roles tapped by the role centrality standard items with the roles tapped by the father involvement items, so only three roles with corresponding involvement items (teach child about life, provide direct care, and show love and affection) were used here. Responses ranged from 1 = very important to 3 = not important; codes were reversed so higher values reflected greater importance. Item responses were summed separately for mothers and fathers to create individual Father Role Centrality Standards (FRCS) scores. Alpha were .57 for fathers and .37 for mothers; however, we did not expect parents to rate all roles similarly and for them to be highly correlated. In fact, we fully expected different assessments of importance across roles. Rather, the FRCS variables were constructed to reflect the total importance assigned to fathering by mothers and fathers across multiple roles.

The joint influence of parents’ FRCS was assessed by centering mothers’ and fathers’ FRCS scores and creating an interaction term for each mother-father FRCS pair. Using an interaction term has been recommended in place of discrepancy or difference scores (Birman, 2006), even when including mothers’ and fathers’ original ratings of the importance of each role (Kenny, Kashy, & Cook, 2006). By using this interaction term, potential multiplicative effects of having, for example, fathers and mothers both rate roles highly rather than unimportant could be captured. If significant interactions emerged, post-hoc analyses dividing the mother high–father low and father high–mother low categories were used to explore differences in involvement across different joint-FRCS combinations.

Importantly, as currently phrased (“Fathers do many things for their children”), the measure could assess parents’ standards about fathers in general, not necessarily their standards for themselves and their partners specifically and at this specific time (with a newborn infant). However, because the prompt also asks how important each of the following activities is “to you,” we believe that this is a reasonable proxy for tapping the identity standards that mothers and fathers hold for themselves at the time a child is born, and in what roles they would like themselves and their partners involved over the children’s life course.

Father status centrality

Centrality of the father status was assessed via three questions reflecting the importance fathers placed upon occupying the status of father: “Being a father and raising children is one of the most fulfilling experiences a man can have,” “I want people to know I have a new child,” and “Not being a part of my child’s life would be one of the worst things that could happen to me.” Responses ranged from 1 = strongly disagree to 4 = strongly agree; higher scores represent higher status centrality. Responses to the three questions were summed to create a single status centrality score with α = .72.

The importance fathers assign to parenting overall (status centrality) could interact with the importance mothers and/or fathers assign to various roles (role centrality). As such, interaction terms were created for role and status centrality measures. Specifically, status centrality was centered, and then this term was multiplied by the centered fathers’ FRCS, mothers’ FRCS, and joint mother-father FRCS terms to create three interaction terms.

Demographic characteristics

Fathers reported their age in years, whether they had been born in the United States (1 = yes, 2 = no), and whether they had any other biological children at the time of the child’s birth (1 = yes, 2 = no). Constructed variables were used to assess fathers’ education level, race/ethnicity, and household income (created by FFCWB by combining mother and father reports across waves; see http://www.fragilefamilies.princeton.edu/documentation.asp for more information). Parents’ marital status at baseline also was included via a variable constructed by FFCWB from several different items, which was used to designate couples as married or cohabiting. Discrepancies between mother and father age were assessed by subtracting mothers’ age from fathers’ age; differences in either direction of more than 5 years were coded as being discrepant. Similarly, fathers’ and mothers’ education were coded as discrepant if education levels differed by more than one category. Response sets are provided in Table 1.

Analyses

Hierarchical regression using SPSS tested the ways in which parents’ identities at birth (status centrality, FRCS) were associated with later father involvement as reported by mothers and fathers, controlling for demographic characteristics. We used regression rather than structural equation modeling because most constructs in the analyses were represented by single observed variables rather than latent variables. Data missing on specific items (no more than 10% on any item and usually less than 1%) were estimated through the use of maximum likelihood estimation.

Results

The influence of status centrality and individual and joint FRCS on later father involvement was tested using a series of four hierarchical regression models for Years 1 and 3, and for mother and father reports of father involvement. Regression was used rather than structural equation modeling because most variables were represented by individual observed variables rather than latent variables. An initial step assessed the direct relationships of the control variables with father involvement. Each identity variable then was added in its own step, to assess the relative contribution of each identity variable (status centrality, individual FRCS measures, joint FRCS, and then interactions among status and role centrality measures) to mother and father reports of father involvement. Results for each model are discussed below.

Year 1—father-reported involvement

In the initial step, only two control variables were associated with father reports of father involvement at Year 1 (see Table 2 for coefficients of the full model). Hispanic fathers reported being less involved with their children (relative to White fathers), and fathers reported greater involvement when they did not have other biological children at the time of this child’s birth.

Table 2.

Regression Estimates for Full Models Predicting Mother and Father Reports of Father Involvement at Years 1 and 3 (N = 1,596).

Path	Year 1		Year 3
Path	B (S.E.)	β	B (S.E.)	β
Outcome—FR Father Involvement
Status centrality	2.31 (.62)	.10^**	2.28 (.90)	.07^*
Father FRCS	1.93 (.71)	.08^**	2.04 (.98)	.07^*
Mother FRCS	2.36 (.53)	.12^**	1.85 (.76)	.07^*
Mother FRCS × Father FRCS	−.03 (1.37)	−.00	−1.00 (1.62)	−.02
Father FRCS × Centrality	−.43 (1.44)	−.01	−.50 (1.99)	−.01
Mother FRCS × Centrality	.86 (1.31)	.02	−1.19 (1.92)	−.02
M FRCS × F FRCS × Centrality	−2.03 (2.47)	−.03	−1.84 (3.20)	−.02
Child gender	−.54 (.48)	−.03	−.17 (.69)	−.01
Father age	.09 (.05)	.07^*	.05 (.07)	.03
Fathers’ other bio children	2.21 (.52)	.12^**	2.27 (.75)	.09^**
Father immigrant status	−1.14 (.69)	−.05	−2.41 (1.00)	−.08^*
Father education	.46 (.30)	.05	.14 (.43)	.01
Father race/ethnicity
Black	.27 (.64)	.01	−1.34 (.92)	−.05
Hispanic	−1.80 (.73)	−.09^*	−1.33 (1.06)	−.05
Other	−.24 (1.21)	−.01	.43 (1.74)	.01
Household income Y1	−.04 (.05)	−.02	.03 (.06)	−.01
Age discrepancy	−.46 (.60)	−.02	−.24 (.86)	−.01
Education discrepancy	1.18 (.79)	.04	1.92 (1.13)	.05
Marital status (birth)	−.89 (.58)	−.05	−1.19 (.84)	−.05
Outcome—MR Father Involvement
Status centrality	1.38 (.85)	.05	1.56 (.93)	.04
Father FRCS	3.69 (.97)	.12^**	2.46 (1.02)	.07^*
Mother FRCS	2.33 (.72)	.09^**	2.77 (.79)	.09^**
Mother FRCS × Father FRCS	.03 (1.88)	.00	4.22 (1.73)	.08^*
Father FRCS × Centrality	−1.79 (1.96)	−.03	.54 (2.02)	.01
Mother FRCS × Centrality	2.57 (1.72)	.04	3.98 (1.92)	.06^*
M FRCS × F FRCS × Centrality	.39 (3.36)	.01	9.05 (3.31)	.10^**
Child gender	−1.29 (.65)	−.05^*	−.87 (.72)	−.03
Father age	−.03 (.06)	−.02	.02 (.07)	.01
Fathers’ other bio children	1.93 (.71)	.08^**	3.54 (.79)	.12^**
Father immigrant status	−2.96 (.94)	−.10^**	−4.44 (1.02)	−.13^**
Father education	1.36 (.40)	.12^**	1.06 (.44)	.08^*
Father race/ethnicity
Black	.06 (.87)	.00	−1.72 (.96)	−.06
Hispanic	−1.67 (1.00)	−.06	−.79 (1.10)	−.03
Other	.87 (1.66)	.02	1.35 (1.81)	.02
Household income Y1	−.07 (.06)	−.03	−.05 (.07)	−.02
Age discrepancy	.87 (.82)	.03	−.37 (.90)	−.01
Education discrepancy	1.35 (1.08)	.03	1.44 (1.17)	.03
Marital status (birth)	−1.55 (.80)	−.06	−1.38 (.87)	−.05

Note: FR = Father report. MR = Mother report.

p < .05. **p < .01.

Status centrality, fathers’ FRCS, and mothers’ FRCS next were entered into the model (in separate steps), and all three were significantly associated with father reports of involvement. Relative contributions of each variable did not change with the addition of other identity variables, so results are presented together. When both parents rated fathering roles as more important (β_F = .08, p < .01; β_M = .11, p < .001) and fathers rated the father status as more central (β = .10, p < .001), fathers reported being more involved with their 1-year-old children. None of the identity interaction terms were significantly associated with father reports of involvement. This model explained approximately 6% of the variance in father reports of Year 1 father involvement.

Year 1—mother-reported involvement

In the initial step, a few different control variables were associated with mother reports of father involvement at Year 1 (see Table 2). When parents were married at the time of the child’s birth, mothers reported greater father involvement at Year 1. Mothers also reported higher father involvement when fathers had more education, were born in the United States, and did not have other biological children at the time of this child’s birth.

When status centrality was added to the model, it was significantly associated with mother-reported father involvement; when fathers rated fathering as more important, mothers reported fathers were more involved at Year 1 (β = .08, p < .01). However, once father FRCS was added to the model, status centrality was reduced to marginal significance (β = .05, p = .056). The addition of mothers’ FRCS also was significant (β = .08, p < .01) and did not affect father FRCS, but further lowered the significance of status centrality (β = .05, p = .078). When both parents rated fathering roles as more important, mothers reported higher father involvement at Year 1. None of the identity interaction terms were significantly associated with father reports of involvement, and this model explained approximately 6% of the variance in mother reports of Year 1 father involvement.

Year 3—father-reported involvement

Again, we first tested an initial model of the direct associations between the control variables and father involvement (see Table 2). As was the case at Year 1, only two control variables were associated with father-reported involvement. Fathers who were born in the United States (β = –.08, p < .01) and who did not have other biological children (β = .09, p < .01) were more involved with their 3-year-old children.

Status centrality, fathers’ FRCS, and mothers’ FRCS next were entered into the model (in separate steps), and again all three were significantly associated with father reports of involvement. Relative contributions of each variable again did not change with the addition of other identity variables, so results are presented together. When parents rated fathering roles as more important (β_F = .08, p < .01; β_M = .07, p < .05) and fathers rated the father status as more central (β = .07, p < .05), fathers reported being more involved with their 3-year-old children. Again, none of the identity interaction terms were significantly associated with father reports of involvement. This model explained approximately 4% of the variance in father reports of Year 3 father involvement.

Year 3—mother-reported involvement

In the initial model, three variables were associated with mother-reported involvement. Mothers reported greater father involvement at Year 3 when fathers were born in the United States (β = –.12, p < .001), did not have other biological children (β = .13, p < .001), and were more educated (β = .08, p < .05).

When status centrality was added to the model, it also was significantly associated with mother-reported father involvement (β = .07, p < .01), such that fathers who held fathering to be more important were reported by mothers as being more involved. This association remained when father FRCS was added to the model, which also was significant (β = .06, p < .05). However, once mother FRCS was entered into the model, which was significant (β = .09, p < .01), father FRCS dropped to marginal significance (β = .05, p = .059).

Unlike previous models, three interactions were significant in this model. First, joint FRCS scores (mother FRCS × father FRCS) were associated with mother reports of father involvement. Interestingly, if fathers assigned little importance to various fathering roles, mothers’ FRCS were negatively associated with their reports of father involvement (the more she valued various fathering roles, the less involved she reported him being). However, if fathers valued multiple roles, mothers’ FRCS were not associated with her reports of fathers’ Year 3 involvement. Second, the interaction between mothers’ FRCS and father status centrality also was associated significantly with mother-reported involvement. When mothers assigned high importance to multiple fathering roles, fathers who assigned more importance to being a parent tended to be more involved with their children. However, when mothers assigned low importance to fathering roles, the more importance fathers assigned to parenting, the less involved mothers reported fathers being. Third, the three-way interaction between joint FRCS and centrality also was associated with mother reports of father involvement. When joint FRCS were low (neither parent valued fathering roles), status centrality was not associated with mother reports of father involvement. However, when joint FRCS were high (both parents valued fathering roles), high status centrality was associated with mothers reporting substantially higher levels of father involvement. Overall, this model explained 6% of the variation in mother reports of Year 3 involvement.

Discussion

Using longitudinal data from the FFCWB study, we tested the associations of fathers’ status centrality and parents’ father involvement identity standards, individually and jointly, on later father involvement. We further tested whether identity measures interacted with one another in their associations with involvement. We found partial support for our hypotheses and notable differences depending on whether fathers or mothers were the reporters of father involvement.

Role Centrality Identity Standards and Status Centrality

It is notable that the addition of mother FRCS to models diminished the associations between father identity variables and mother reports of involvement (status centrality at Year 1, father FRCS at Year 3), whereas all identity variables were influential for father reports of involvement. Although father identity variables retained marginal significance in mother report models, it appears that mothers’ identity perspectives influence their perspectives of father behavior more strongly than fathers’ own identity standards. This runs contrary to our hypothesis that fathers’ identity standards would be more influential, but is consistent with previous literature suggesting that fathers’ perceptions of mothers’ identity standards more strongly predicted his behavior than did his own identity; here, we assessed mothers’ actual standards, rather than fathers’ perceptions of mother standards. This finding also indicates some spillover of perceptions, such that mothers’ expectations and standards color their subjective view of fathers’ behavior, leading her expectations to be strongly associated with her perceptions of his behaviors, more so than fathers’ own identity standards, and more so than her standards influence his reports of his own behavior.

Mothers’ and fathers’ individual ratings of role importance were associated with later levels of father involvement at both Year 1 and Year 3, but the interaction of their FRCS was significant only for mother-reported father involvement at Year 3, providing support for Kenny and colleagues’ (2006) assertion regarding the importance of including original scores when assessing the effects of joint influences. Identity theory suggested that dissimilar identity standards (lower joint FRCS) would generate incongruent feedback, which would lead to shifts in identity-relevant behavior or identity standards (Adamsons, 2010). However, our findings suggest that fathers generally continue to act in accordance with their own and mothers’ identity standards individually, rather than such standards having multiplicative or interactive influences. This likely is due at least in part to the generally high ratings most parents assigned to the majority of the fathering roles. Specifically, with most mothers and fathers in agreement that most of the listed roles were highly important, there was relatively little variability in joint standards, leaving little variability to predict other outcomes.

The one exception to the lack of interactions occurred for mother reports of involvement at Year 3, where joint FRCS scores were associated with mother-reported father involvement. This finding suggests a contrast effect. If fathers’ FRCS were high, mothers reported fathers being uniformly highly involved, regardless of her own FRCS. However, if fathers assigned little importance to various fathering roles, then the more she valued various fathering roles, the less involved she reported him being. This might suggest that high FRCS mothers are uniquely disappointed by their low FRCS partners, perhaps reflecting unmet fathering expectations that led such mothers to report uniquely low levels of father involvement. As such, it appears that when mothers have low expectations, fathers can rely on their own identity standards to maintain high involvement with their children; however, when parents’ expectations do not match, and it is fathers who hold father involvement in low esteem, this interacts with mothers’ high expectations to cast a shadow on mothers’ views of father involvement. It is further interesting that such dynamics did not appear until Year 3. Given the suggestions of Burke (1991, 1997) that incongruence must be long-standing before it causes chronic distress, our finding suggests that in the first year of parenting a new child, parents might still be in a “honeymoon” period of identity negotiations. However, two years later, it has become clear that differences in expectations remain unresolved.

Father status centrality also was associated with involvement, but only for father reports. These findings suggest that fathers who value fathering are more likely to view themselves as being more involved with their children, although such higher involvement is not reported by others who do not necessarily share the same subjective view (here, mothers). Given that mother and father reports of father involvement were correlated only at r = .30, leaving plenty of room for variation between reports, centrality appears to be a source of cognitive bias for fathers when rating their own levels of involvement.

Mothers’ and fathers’ FRCS also remained significant even after accounting for centrality of the father status. Theoretically, these findings confirm those of Rane and McBride (2000) and Maurer (2007) that the importance assigned to particular fathering roles plays a unique part in explaining father involvement, above and beyond that explained by the importance of being a parent overall. Theoretically, Burke’s (1991, 1997) identity verification model also operates at the level of role identities (breadwinner, nurturer) rather than at the status level (being a father), suggesting that identity enactment garners feedback specific to those roles. However, in the present study, father involvement (i.e., father identity relevant behavior) was assessed at a relatively broad level. Although the roles were matched for both role centrality standards and role-related behaviors, we cannot parse out relationships between specific identity standards and particular forms of involvement. This will be important to address in future studies.

Interactions Between Role Centrality and Status Centrality

At Year 3, mothers’ FRCS and parents’ joint FRCS each interacted with status centrality to predict mother-reported father involvement. Regarding mothers’ FRCS, when fathers’ status centrality was low, mothers reported less involvement when they valued fathering roles more. Because the status centrality measure asked fathers about the importance they personally assigned to being a father, this measure might better capture their overall intent regarding fathering than the questions about his involvement across different roles. As such, this measure interacted uniquely with mothers’ expectations to produce a contrast effect more so than father FRCS. When mothers placed high value on roles but fathers placed low value on the overall status of being a father, mothers’ presumably unmet expectations led them to view fathers as being particularly uninvolved, beyond the low levels of involvement that would be predicted just by fathers having low status centrality. Mothers with high FRCS appear to be uniquely disappointed in fathers who saw fathering as less central, perhaps biasing them to perceive these fathers as less involved. In contrast, mother-reported involvement of high centrality fathers appeared to be affected little by mothers’ FRCS; such high centrality fathers were uniformly highly involved, whether mothers’ FRCS were high or low. These findings suggest that low centrality fathers are uniquely vulnerable. That is, mothers will view them as being particularly uninvolved, if mothers’ high role expectations presumably were unmet by fathers who place less value on being a parent, regardless of fathers’ own views of their involvement. It is noteworthy that such findings arose in a sample of resident fathers, where mothers could be more aware of fathers’ activities with children (compared with nonresident fathers). Thus, such findings cannot be explained simply by mothers’ assumptions about what fathers do (or do not do) when mothers are not present (Wical & Doherty, 2005). Instead, these findings may suggest that mothers’ subjective perceptions/reports of what they “objectively” see fathers doing might shift, depending on the interactive context of what mothers wanted fathers to do and what fathers themselves value doing.

Regarding the three-way interaction among joint FRCS and centrality, status centrality was not associated with mother reports of involvement when neither parent valued fathering roles. However, when both parents valued fathering roles, higher status centrality was associated with more frequent involvement. Once again, father status centrality appears to interact with mothers’ views of fathers’ behaviors. When both parents hold high expectations of fathers across roles, fathers having high status centrality is linked with mothers viewing them as being more involved with their children, beyond what mothers’ or fathers’ individual FRCS would predict by themselves. As no previous research has examined possible interactions between status centrality and role centrality standards, this finding extends our empirical knowledge about the ways in which identity is associated with behavior. It also might account for previous findings that status centrality was not associated with father involvement (Maurer, Pleck, & Rane, 2003; Rane & McBride, 2000), as the present findings suggest that status centrality might be associated with involvement only under certain conditions (high joint FRCS) and not under others (low joint FRCS). These findings demonstrate a need for fathering research using identity theory to move beyond individual models of prediction to dyadic and interactive models.

Limitations, Implications, and Future Directions

Several important limitations exist that require consideration when interpreting the findings reported here, and many questions remain unanswered. First, although the sample was fairly representative in terms of race/ethnicity and socioeconomic status, the focus on resident fathers leaves the unanswered question of how these processes work for nonresident fathers. Given that father involvement was our focus and that nonromantic/nonresident couples are the least likely to have such involvement, our findings offer little insight into fathers who are completely or partially disengaged from their children. Also, given hypotheses about disengagement from disconfirming relationships, disengaged fathers and nonromantically involved couples are of particular theoretical importance. As such, future research should focus on these groups but with involvement measures that are more relevant to nonresident fathers.

Further, only mothers’ and fathers’ identity standards were examined and in a limited area (role importance). Future research should explore the importance of FRCS held by other individuals (e.g., extended family, employers, friends, children), use a broader conceptualization of identity standards than role centrality (e.g., how much time should be spent in various roles, evaluations of parental competence in roles), and mothers should be asked about centrality of the father status. Also, responses regarding status and role centrality were closed-ended, directive, and limited in possible variability (ranged from 1 to 3). We suggest that fathers should be asked whose feedback they value regarding fathering, and scholars should account for such information when assessing the influence of feedback (e.g., weighting feedback by relative importance). For example, if fathers are no longer romantically involved with their children’s mothers, fathers might not value mothers’ feedback; thus, having dissimilar sets of standards would not cause distress. In addition, asking parents to suggest and define the roles that are important to them, as well as whose advice and feedback they seek, would allow for a broader conceptualization of fathering and greater variability to emerge than was possible with the measures used here. Too, given differences found between influences on mother and father reports of father involvement, future research should investigate other reporters of fathering behaviors, such as child reports or observations.

Although identity is important for behavior, various contextual influences also encourage or inhibit this association (Stryker, 1968). For example, recent economic downturns have resulted in job loss by numerous fathers, leaving them unable to fulfill the provider role regardless of their beliefs or those of their spouse. Exploration of additional contextual influences that moderate the associations between identity and behavior will give greater insight into the potential challenges faced by fathers who wish to be involved with their children as well as factors that might increase the involvement of fathers who do not personally assign great importance to parenting.

Future research also should tease apart the relationships among specific sets of role identity standards and specific forms of fathering behaviors. For example, when fathers and mothers agree that nurturing is an important fathering role, are fathers more likely to be highly involved in nurturing their children? If mothers believe that a particular role is unimportant, are they more likely to gatekeep father involvement in that arena? Further, measures of mothers’ and fathers’ views of these particular fathers, rather than fathers in general as was the case here, would provide more specific and potentially more accurate assessments of mothers’ encouragement or discouragement of father involvement. A mother might believe that fathers generally should be involved in direct care but also might think that her partner is not a competent caregiver or should not be involved with her child for other reasons (e.g., his infidelity). In the present sample, post hoc analyses reveal that 96% of mothers stated that they could always trust the father to take good care of the child, with an additional 3.5% stating that they could sometimes trust the father to do so; only 0.4% stated that they rarely could trust him. As such, gatekeeping due to presumed father incompetence does not appear to be a common issue in the present sample. Presumably this would be more problematic for nonresident than resident fathers (i.e., couples experiencing infidelity or other personal issues would be less likely to be cohabiting/married), and so should be investigated in future studies focusing on families with nonresident fathers.

Important to scholars interested in identity processes, measures of identity should be asked at all data collection points, rather than limited to being asked at birth, especially for first-time fathers. This would provide a better test of Burke’s (1991, 1997) identity verification model and the proposed identity negotiation process by demonstrating how and whether identity standards shift over time; whether dissimilar identity standards and disconfirming feedback continued over time; and whether, when, and how dissimilarity was resolved over the crucial first few years of parenthood. More frequent points of data collection might be necessary during the first year to accurately capture the identity negotiation process occurring immediately following birth.

Our findings generally confirm Burke’s (1991, 1997) model, but they also suggest some refinements. Individual identity standards and the standards of counteridentities appear to remain strongly influential to fathers’ behavior. Burke’s model could be expanded to include more than just dyadic interactions and incorporate interactive influences among various aspects of identity occurring at multiple levels (e.g., role centrality, status centrality), rather than viewing “identity standards” as a single, independent predictor of behavior. Identities often are enacted in the presence of multiple individuals, and feedback is received simultaneously from multiple sources. Lastly, exploration of the ways in which individuals manage to “self-verify” their identities (continuing, as many of these fathers seem to have done, to believe they are doing what is “right” even in the face of people telling them otherwise), would be a fascinating area of study.

Our findings have several practical implications. Mothers’ and fathers’ identity standards are noteworthy. Practitioners and policymakers would be well served to continue developing programmatic efforts and policies that are designed to support parents’ sense of the value of father involvement in various roles, even among the relatively “low risk” population of married or coresident parents. Programs and policies that emphasize the importance of fathers and their numerous roles in children’s lives (beyond breadwinning) could increase father involvement in childrearing both early on and later. This is a fairly simple and inexpensive yet apparently powerful step to take to prevent fathers from disengaging from their children. The present findings also suggest that mothers’ views of father behaviors are particularly affected by how important fathers view the status of parent and their standards for involvement in various fathering roles. Providing support for fathers’ views of themselves as important influences in their children’s lives might increase “objective” reports of involvement and also substantially increase mothers’ perceptions of their involvement. Given that maternal satisfaction with father involvement is an important predictor of relationship satisfaction and maternal well-being (Coley & Schindler, 2008; Rafferty, Griffin, & Robokos, 2010), such goals are worthy of promotion.

Overall, our findings confirm that individual identity characteristics (role centrality standards and status centrality) are linked with involvement, particularly during the early years of parenting and among this unique population. In addition, individual perspectives of parenting behavior can be influenced by one another’s parenting identities. Given the direct influence of identity characteristics and the variety of factors that could be influenced indirectly by identity, our findings support continued investigation of the ways in which parents negotiate their parenting roles and responsibilities early in childrearing, and how parental identities and expectations ultimately shape fathers’ behavior with their children.

Footnotes

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) received no financial support for the research, authorship, and/or publication of this article.

References

Adamsons

(2010). Using identity theory to develop a mid-range model of parental gatekeeping and parenting behavior. Journal of Family Theory and Review, 2, 137–148.

Adamsons

Pasley

(2006). Coparenting following divorce and relationship dissolution. In Fine

Harvey

(Eds.). Handbook of divorce and relationship dissolution (pp. 241–262). New York, NY: Erlbaum.

Allen

S. M.

Hawkins

A. J.

(1999). Maternal gatekeeping: Mothers’ beliefs and behaviors that inhibit greater father involvement in family work. Journal of Marriage and the Family, 61, 199–212.

Amato

P. R.

Meyers

C. E.

Emery

R. E.

(2009). Changes in nonresident father-child contact from 1976 to 2002. Family Relations, 58, 41–53.

Birman

(2006). Acculturation gap and family adjustment: Findings with Soviet Jewish refugees in the United States and implications for measurement. Journal of Cross-Cultural Psychology, 37, 568–589.

Burke

P. J.

(1991). Identity processes and social stress. American Sociological Review, 56, 836–849.

Burke

P. J.

(1997). An identity model for network exchange. American Sociological Review, 62, 134–150.

Burke

P. J.

Cast

A. D.

(1997). Stability and change in the gender identities of newly married couples. Social Psychology Quarterly, 60, 277–290.

Burke

P. J.

Reitzes

D. C.

(1981). The link between identity and role performance. Social Psychology Quarterly, 44, 83–92.

10.

Cabrera

N. J.

Shannon

J. D.

Tamis-LeMonda

(2007). Fathers’ influence on their children’s cognitive and emotional development: From toddlers to Pre-K. Applied Development Science, 11, 208–213.

11.

Coley

R. L.

Chase-Lansdale

P. L.

(1999). Stability and change in paternal involvement among urban African-American fathers. Journal of Family Psychology, 13, 416–436.

12.

Coley

R. L.

Schindler

H. S.

(2008). Biological fathers’ contributions to maternal and family functioning. Parenting, 8, 294–318.

13.

D’Angelo

A. V.

Palacios

N. A.

Chase-Lansdale

(2012). Latino immigrant differences in father involvement with infants. Fathering, 10, 178–212.

14.

Hamilton

B. E.

Martin

J. A.

Ventura

S. J.

(2011). Births: Preliminary data for 2010. National Vital Statistics Reports, 60(2), 1–26.

15.

Hofferth

S. L.

(2003). Race/ethnic differences in father involvement in two-parent families: Culture, context, or economy? Journal of Family Issues, 24, 185–216.

16.

Hofferth

Pleck

Stueve

Bianchi

Sayer

(2002). The demography of fathers: What fathers do. In Tamis-LeMonda

Cabrera

(Eds.), Handbook of father involvement: Multidisciplinary perspectives (pp. 63–90). Mahwah, NJ: Lawrence Erlbaum.

17.

Ihinger-Tallman

Pasley

Buehler

(1993). Developing a middle-range theory of father involvement postdivorce. Journal of Family Issues, 14, 550–571.

18.

Kenny

D. A.

Kashy

D. A.

Cook

W. L.

(2006). Dyadic data analysis. New York, NY: Guilford Press.

19.

Landale

N. S.

Oropesa

R. S.

(2001). Father involvement in the lives of mainland Puerto Rican children: Contributions of nonresident, cohabiting and married fathers. Social Forces, 79, 945–969.

20.

Lerman

Sorensen

(2000). Father involvement with their nonmarital children: Patterns, determinants, and effects on their earnings. Marriage and Family Review, 29, 137–158.

21.

Manning

Stewart

Smock

(2003). The complexity of fathers’ parenting responsibilities and involvement with nonresident children. Journal of Family Issues, 24, 645–667.

22.

Marsiglio

Amato

Day

R. D.

Lamb

M. E.

(2000). Scholarship on fatherhood in the 1990s and beyond. Journal of Marriage and the Family, 62, 1173–1191.

23.

Maurer

T. W.

(2007). Gender congruence and social mediation as influences on fathers’ caregiving. Fathering, 5, 220–235.

24.

Maurer

T. W.

Pleck

J. H.

Rane

T. R.

(2003). Methodological considerations in measuring paternal identity, Fathering, 1, 117–129.

25.

Maurer

T. W.

Pleck

J. H.

Rane

T. R.

(2001). Parental identity and reflected-appraisals: Measurement and gender dynamics. Journal of Marriage and the Family, 63, 309–321.

26.

McBride

B. A.

Brown

G. L.

Bost

K. K.

Shin

Vaughn

Korth

(2005). Paternal identity, maternal gatekeeping, and father involvement. Family Relations, 54, 360–372.

27.

Pasley

Futris

Skinner

(2002). Effects of commitment and psychological centrality on fathering. Journal of Marriage and Family, 64, 130–139.

28.

Pleck

(1997). Paternal involvement: Levels, sources and consequences. In Lamb

(Ed.), The role of the father in child development (3rd ed., pp. 66–103). New York, NY: John Wiley & Sons.

29.

Rafferty

Griffin

K. W.

Robokos

(2010). Maternal depression and parental distress among families in the Early Head Start Research and Evaluation Project: Risk factors within the family setting. Infant Mental Health Journal, 31, 543–569.

30.

Rane

T. R.

McBride

B. A.

(2000). Identity theory as a guide to understanding fathers’ involvement with their children. Journal of Family Issues, 21, 347–366.

31.

Reichman

N.E.

Teitler

J. O.

Garfinkel

McLanahan

(2001). Fragile families: Sample and design. Children and Youth Services Review, 23, 303–326.

32.

Stryker

(1968). Identity salience and role performance: The relevance of symbolic interaction theory for family research. Journal of Marriage and the Family, 30, 558–564.

33.

Stryker

(1980). Symbolic interactionism: A social structural version. Menlo Park, CA: Benjamin Cummings.

34.

Stryker

Serpe

R. T.

(1994). Identity salience and psychological centrality: Equivalent, overlapping, or complementary concepts? Social Psychology Quarterly, 57, 16–36.

35.

Swann

W. B.

Jr. (1987). Identity negotiation: Where two roads meet. Journal of Personality and Social Psychology, 53, 1038–1051.

36.

Swann

W. B.

Jr. De La Ronde

Hixon

J. G.

(1994). Authenticity and positivity strivings in marriage and courtship. Journal of Personality and Social Psychology, 66, 857–969.

37.

Swann

W. B.

Jr. Read

S. J.

(1981). Self-verification processes: How we sustain our self-conceptions. Journal of Experimental Social Psychology, 17, 351–372.

38.

Wical

K. A.

Doherty

W. J.

(2005). How reliable are fathers’ reports of involvement with their children? A methodological report. Fathering, 3, 81–91.

39.

Wood

J. J.

Repetti

R. L.

(2004). What gets dad involved? A longitudinal study of change in parental child caregiving involvement. Journal of Family Psychology, 18, 237–249.