Abstract
Although many fathers today spend more time with children than was the case in the past, physical care of young children remains primarily mothers’ work. Yet some fathers claim that they do work traditionally seen as the “mother’s job” every day. Using subsample data from the male respondent file of the National Survey of Family Growth 2002 (n = 613), this study examines factors associated with married or cohabiting fathers’ daily involvement in physical care of children under age 5 years. Logistic regression results show that daily involvement is more likely if fathers were raised by their biological fathers, received more education, have employed wives or partners, have a young male child, or receive public assistance; it is less likely if they have school-age children. This study suggests that paternal involvement in physical care of young children is shaped by multiple factors including childhood experiences, education, economic conditions, and current family context.
Social expectations for fathers have changed, and today, the new father—a father who cares for and is emotionally close to his children—is the ideal (Furstenberg, 1988; Griswold, 1993; Messner, 1993). An increasing number of men consider participation in their children’s lives to be important: Many men express a desire to spend more time with their children (Gerson, 1993; Russell, 1999) and believe that both parents should share equally the various responsibilities of childrearing (e.g., Milkie, Bianchi, Mattingly, & Robinson, 2002). In the United States, the amount of time married fathers spend with children has increased (Bianchi, 2000; Bianchi, Robinson, & Milkie, 2006; Bryant & Zick, 1996; Casper & Bianchi, 2002; Gray & Anderson, 2010; Pleck, 1997; Sandberg & Hofferth, 2001). Yet, compared with mothers, fathers continue to spend considerably less time with children (Casper & Bianchi, 2002; Pleck, 1997), and this is the case even when mothers are employed (Bianchi et al., 2006; Craig, 2006). Furthermore, the kind of parenting activities provided by fathers is often gendered: Fathers are more likely to take part in interactive or recreational activities (such as play), leaving physical care (such as bathing and diapering) to mothers (Combs-Orme & Renkert, 2009; Gray & Anderson, 2010; Lamb, 1997; McBride & Mills, 1993; Robinson & Godbey, 1997; Starrels, 1994; Tichenor, 2005). Additionally, fathers in intact families typically spend time childrearing on weekends, with less time spent on weekdays (Yeung, Sandberg, Davis-Kean, & Hofferth, 2001).
These findings indicate that many fathers play the role of secondary caretaker at most (Wall, 2007) and that mothers continue to be burdened with the responsibility of meeting the daily needs of children. Still a growing number of fathers share, or at least claim that they share, child care tasks equally with their wives or partners (Milkie et al., 2002). What kind of father claims to perform tasks typically perceived by other fathers to be the “mother’s job?” What are the differences between fathers who change diapers every day and those who regularly play with their children but say no to daily diapering?
This study uses the National Survey of Family Growth, Cycle 6, 2002 (NSFG 2002) to examine factors associated with daily physical care of young children among married or cohabiting fathers. The NSFG 2002 is the first cycle that included a male respondent file (N = 4,928) and asked fathers to report how often they take part in child care activities for their biological or adopted children. The present study attempts to sort out factors relevant to paternal involvement by making use of this new data set. This study uses a subsample (n = 613) of married or cohabiting fathers who live with at least one biological or adopted child under age 5 years, the age group for which physical care is intensive and required every day. Drawing from prior research, three broad areas that might be associated with men’s level of involvement are examined: (a) men’s socialization, (b) men’s socioeconomic status, and (c) men’s household characteristics or present family context. Using logistic regression, this study first analyzes how these factors relate to men’s self-reports of daily involvement in physical care of young children and then compares the results to an analysis of factors associated with men’s daily involvement in play.
Paternal involvement in physical care is found to be beneficial to children (Bronte-Tinkew, Carrano, Horowitz, & Kinukawa, 2008; Marsiglio, Amato, Day, & Lamb, 2000) and associated with greater gender equality, better marital relationships (e.g., Erickson, 1993; Hewlett, 2000), and psychological well-being of fathers (Schindler, 2010). Thus, it is important to elucidate differences between fathers who identify themselves as regular caretakers and those who assume a secondary caregiver role or provide no child care. Many past studies on father involvement focus on nonresident fathers, and paternal involvement is often assessed by mothers’ or children’s reports. Mothers’ reports, however, tend to underestimate levels of father involvement, particularly for the case of resident fathers (Coley & Morris, 2002; Mikelson, 2008). Using fathers’ self-reports of their involvement, this study contributes to the literature on fatherhood by examining resident fathers’ participation in what are understood conventionally to be motherly tasks (i.e., physical care of young children).
Theories, Prior Research, and Hypotheses
Drawing on prior research, this study examines three broad areas as factors that affect father involvement in physical care. These areas are (a) men’s socialization, (b) men’s socioeconomic status, and (c) men’s household characteristics or present family context.
Men’s Socialization
Although many studies on father involvement examine the effects of men’s characteristics and present circumstances, men’s past family relationships also influence their conception of the father role (Forste, Bartkowski, & Jackson, 2009; Gray & Anderson, 2010; Roy, 2006; Snarey, 1993). Through socialization, individuals shape their attitudes and behaviors and learn to take on roles. Socialization during childhood is particularly powerful because children are usually not aware they are being socialized. Individuals often take what they have learned in childhood for granted, perceiving this worldview as reality even though it is constructed by society (Berger & Luckmann, 1967). Though parents are by no means children’s sole agents of socialization, the household is the primary setting for children to observe how family members interact in terms of gender role taking (Goffman, 1977).
Two aspects of men’s childhood experience—men’s parents’ housework allocation and the level of men’s fathers’ involvement—are examined in this study. According to Chodorow (1974), children who were raised by parents who shared housework are likely to hold egalitarian attitudes as young adults, and research by Cunningham (2001) supports this theory. In the present study, men’s mothers’ employment is used as an indicator of parents’ shared housework. Although wives’ employment by no means indicates equal sharing of housework by married couples (e.g., Hochschild, 1989), wives’ employment is likely to make necessary some level of housework participation by husbands. Thus, men who grew up with employed mothers are likely to have been exposed to parents sharing housework. Therefore, this study hypothesizes the following:
Hypothesis 1: Men’s mothers’ employment increases the odds of men’s daily involvement in physical care of their young children.
The nature of men’s relationships with their fathers is also expected to have an impact on men’s involvement with their children. Forste et al. (2009) found that among nonresident low-income fathers, the nurturing role was an important aspect of fatherhood for those who had close relationships with their own fathers. Biological fathers are theorized to invest more in children compared with “social fathers,” who have no biological ties to children (e.g., Cherlin & Furstenberg, 1994; Daly & Wilson, 2000). Empirical evidence suggests that this is indeed the case (e.g., Amato, 1987; Combs-Orme & Renkert, 2009; Hofferth & Anderson, 2003; Marsiglio, 2004), though some recent research found mixed or contrary evidence for married social fathers (Berger, Carlson, Bzostek, & Osborne, 2008; Gorvine, 2010). Social fathers may be becoming more involved with children as, for instance, blended families become more institutionalized. However, no clear norms yet exist for step-parenting (Cherlin & Furstenberg, 1994; Coontz, 1997) and this was especially the case during the time the respondents in this survey were growing up. Men who were raised by their biological fathers are also likely to have received more consistent paternal involvement, compared with those who were not. Thus, this study assumes that biological fathers were more involved parents.
Hypothesis 2: Men raised (mostly) by their biological fathers are more likely to provide physical care to their young children every day.
Men’s Socioeconomic Status
As briefly discussed above, our cultural ideal of fatherhood has changed to include child care (Furstenberg, 1988; Griswold, 1993; Messner, 1993). Many studies have found that, compared with the past, more men believe in the equal sharing of various childrearing responsibilities (Bittman & Pixley, 1997; Burgess, 1997; Casper & Bianchi, 2002; Colemen & Ganong, 2004; Gerson, 2002; Milkie et al., 2002; Pleck & Pleck, 1997). However, as LaRossa (1988) succinctly argued, this change appears to be largely ideological: Most fathers believe in equal sharing, but many do not live up to these beliefs (Griswold, 1993; LaRossa, 1988; Rustia & Abbott, 1993).
In their review of fatherhood literature from the 1990s, Marsiglio et al. (2000) recommend that future research assess the driving forces behind the change in the father role—whether this change was due to adoption of a new cultural ideal or to other external forces, such as changing economic conditions (Marsiglio et al., 2000). It is important to point out that the new father ideology contradicts hegemonic masculinity, the dominant image of masculinity. Hegemonic masculinity is defined in opposition to femininity, with men’s ability to (economically) provide the central component of manhood/fatherhood (Connell, 1993; Kimmel, 1994; Townsend, 2002). Physical care of young children is traditionally feminine work, and therefore, engagement in this work could undermine masculine identity. However, provision of such care is encouraged in the new father ideology. When trying to understand what types of men adopt the new father role at the expense of masculine identity, it is important to consider men’s location in terms of socioeconomic status, especially under the current economic context in which a growing number of men are unable to fulfill the primary provider role.
Empirical studies on the association between social class and father involvement are, however, scarce (Shows & Gerstel, 2009). In terms of the effect of men’s education, much of the prior research on housework allocation (e.g., Bianchi, Milkie, Sayer, & Robinson, 2000; Kamo, 1988; Presser, 1994) shows that the more educated men are, the more time they spend on housework. This implies that higher education facilitates an egalitarian ideology, and therefore, more educated men are likely to adopt the new father ideology. Thus, this study hypothesizes the following:
Hypothesis 3: Men’s education increases the odds of daily paternal involvement in physical care of young children.
However, empirical findings on the relationship between men’s economic conditions and father involvement are mixed. Some studies (e.g., Brayfield, 1995; Marsiglio, 1991; Yeung et al., 2001) found that men’s income and occupational prestige have little effect on time spent by fathers on child care. Other studies indicate that fathers with low-income jobs are less involved (Goodman, Creuter, Lanza, & Cox, 2008) or that the association is curvilinear—both high- and low-income fathers are less involved compared with middle-income fathers (Presser, 1986). In contrast, still other studies show that fathers with lower income are more involved in the care of young children (Casper & O’Connell, 1998; Gaunt, 2005, Yeung et al., 2001).
These contradictory findings may suggest that effects are difficult to measure based simply on men’s income levels. Gerson (1993) observed that, as the primary breadwinning role became increasingly unattainable for them, working-class men’s masculine identity became threatened. This may have caused these men to resist performing traditionally feminine tasks or to downplay the fact that they are highly involved in such tasks (e.g., Brines, 1994; Hochschild, 1989). However, recent qualitative studies indicate the opposite may be the case. Economically disadvantaged men, such as working-class men or low-income nonresident fathers, were observed to be putting strong emphasis on care work for the family, deemphasizing the importance of the provision role (which they could not play adequately; Hamer, 2001; Lamont, 2000) and participating actively in daily child care (Shows & Gerstel, 2009). These men did not appear to be ashamed of their engagement in family care work. These qualitative studies suggest that economically disadvantaged men may be reconstructing the image of manhood, or “undoing gender,” by taking, or being compelled to take, a more active role in parenting (Deutsch, 2007; Griswold, 1993; Shows & Gerstel, 2009).
In light of these findings from prior research, this study expects that the important variable is men’s ability to provide for the family rather than men’s income. Thus, the following is hypothesized:
Hypothesis 4: Men’s inability to provide for the family is positively associated with men’s claim that they provide physical care for young children every day.
Men’s Household Characteristics/Present Family Context
The time availability perspective and the relative resources perspective suggest that wives’ or partners’ employment has a positive association with men’s level of involvement in domestic tasks (i.e., traditionally female tasks; Bianchi et al., 2000). Men are expected to participate more in such tasks as wives spend more time in the labor market (Coverman, 1985; Hiller, 1984) or gain power to negotiate by making substantive contributions to household income (Brines, 1994; Lundberg & Pollak, 1996). These perspectives imply that, compared with men who are sole breadwinners, men with employed wives or partners are more likely to take on physical care of young children. On the other hand, gender scholars argue that wives’ employment and/or income contributions alone do not determine the division of labor by couples; gender continues to structure role allocation (e.g., Hochschild, 1989; Tichenor, 2005).
Prior research shows mixed results regarding the effects of wives’ employment and income on paternal involvement. Some studies found that wives’ relative income contributions have positive effects on husbands’ participation in child care (Bianchi et al., 2000; Casper & O’Connell, 1998; Marshall, 2006; Shows & Gerstel, 2009). But couples’ high earnings increase the use of paid child care resources and thus do not necessarily increase paternal involvement in child care provision (Bianchi et al., 2000; Brayfield, 1995; Casper, 1996; Marshall, 2006; Shows & Gerstel, 2009; Tichenor, 2005). Other studies found no significant effects of wives’ employment on child care time spent by fathers (Combs-Orme & Renkert, 2009; Marsiglio, 1991; Sandberg & Hofferth, 2001; Yeung et al., 2001) or that the association depends on the work schedules of couples (Brayfield, 1995; Coltrane, 2000; Presser, 1986, 1994). Furthermore, lower involvement of fathers is not necessarily the result of fathers’ reluctance—it may be mothers who resist relinquishing total control over family work and attempt to hold on to the primary caregiving role (Allen & Hawkins, 1999; Blair-Loy, 2003; Fox, Bruce, & Combs-Orme, 2000; Hays, 1996; Hochschild, 1989; Stone, 2007; Tichenor, 2005). Additionally, quality of the marital relationship positively affects the level of father involvement (Belsky, Gilstrap, & Rovine, 1984; Pleck, 1997).
The findings above suggest important effects of mothers’ income contributions, couples’ work schedules, gender ideology held by mothers, and couple relationships. In this study, however, only the effect of maternal employment is assessed because the NSFG 2002 does not contain items that allow measurement of the other variables just mentioned. With this limitation in mind, and taking the time availability perspective and relative resource perspective, maternal employment is expected to be positively associated with daily paternal involvement in child care. Thus, the following is hypothesized:
Hypothesis 5: Maternal employment increases fathers’ daily involvement in physical care of young children.
In addition to wives’ characteristics, empirical findings point to the relevance of characteristics of children on father involvement. In general, fathers are more likely to share residence with sons and feel that interaction with them is important, compared with daughters (Kane, 2006; Lundberg, McLanahan, & Rose, 2007; Raley & Bianchi, 2006; Starrels, 1994), though some studies found no difference in kinds of paternal interaction with, or time spent on, male and female children (Snarey, 1993), or few differences for children under age 5 years (Combs-Orme & Renkert, 2009; Marsiglio, 1991). Other studies have even found that fathers are more involved with daughters than with sons (Lamb et al., 1988). Children’s age also appears to have influences on paternal involvement. Fathers (and mothers) are more likely to spend time with infants and toddlers than with older children (Yeung et al., 2001), but first-born children are more likely than later-born children to be given infant care by fathers (Combs-Orme & Renkert, 2009; Pleck, 1997).
These findings suggest that fathers may change their level of involvement depending on the gender and age/birth order of children. Thus, the present study hypothesizes the following:
Hypothesis 6: Children’s gender is associated with fathers’ daily involvement in physical care of young children.
Hypothesis 7: Presence of older (school-age) children is associated with fathers’ daily involvement in physical care of young children.
Method
Data
This study uses the NSFG 2002, which is the first cycle that includes male respondents. Based on a national area probability sample of households in the United States, 4,928 men aged between 15 and 44 years were interviewed (the response rate was 78%). This survey asks men several questions about the frequency of their involvement with biological or adoptive children. Men are given the opportunity to answer separate sets of questions, according to residential status (whether the respondents live with children or not) and by age group of child(ren) (whether children are under age 5 years or aged between 5 and 18 years). The goal of this study is to find factors associated with regular involvement by fathers in physical care of children, using this new data set. Physical care is more intense and required by infants and preschoolers every day. Care of young children is typically viewed as the mother’s job. Therefore, the present study uses a subsample of fathers who are married or cohabiting and live with at least one biological or adopted child under age 5 years (n = 613).
Dependent Variables: Fathers’ Daily Involvement in Physical Care and Play
The focal variable of this study is the frequency of paternal involvement in physical care. An analysis will also be done on the frequency of involvement in play and the results will be compared to examine whether relevant factors differ for these two types of activities. Two dichotomous variables (1 = did every day, 0 = did occasionally or never) are constructed from two items on the NSFG 2002. The item used for the physical care variable is created from the question that asks fathers how often they bathed, diapered, or dressed their child(ren) or helped them to bathe, dress, or use the toilet in the last 4 weeks. The item used for the play variable is based on another question that asks fathers how often they played with their child(ren) in the last 4 weeks.
The categories of answers (for both variables) are 1 = not at all, 2 = less than once a week, 3 = about once a week, 4 = several times a week, and 5 = every day (at least once a day). There are three reasons that this study compares fathers who gave answer 5 with those who gave answers 1 to 4. First, the present study attempts to examine what kinds of fathers assume child care responsibilities as part of their daily routine. Because physical care is required every day for young children, occasional involvement is likely to indicate a secondary caretaker role. Second, most fathers in this survey claim that they are involved either every day or several times a week. Because of this clustering of data, this study sets apart those who answered “every day” from others and examines whether fathers who give physical care every day differ from the rest. Last, the term several can be interpreted as meaning anywhere from twice a week to six times a week. These fathers may or may not be highly involved. This study interprets the answer “every day” to be an indicator of fathers’ strong daily commitment to the child care role.
Independent Variables
Socialization variables
Two variables are included as indicators of men’s socialization: (a) men’s mother’s employment status and (b) raised mostly by biological father. Mother’s employment status is based on the question that asks the respondents about their mothers’ employment status when they were aged 5 to 15 years. Two dummy variables—(mother) employed full-time and employed part-time—are created from this question and the reference category is not employed. The answer “mother worked both full-time and part-time” is coded as full-time employment; responses indicating the man did not know about his mother’s employment were coded as nonemployment. The second variable, raised mostly by biological father, is a dichotomous variable constructed from a question that asked respondents “Who was the man who mostly raised you when you were growing up?” Among all respondents, 78.6% indicated biological fathers, 6.4% said step-fathers, 9.3% reported “others” (e.g., grandfathers, foster fathers, mother’s boyfriends), and 5.7% answered that they had no father figure. Those who were not raised by biological fathers were combined and coded 0 (= no).
Men’s socioeconomic status
Three variables are created for men’s socioeconomic status: (a) years of education, (b) employment status in the previous year, and (c) use of public assistance. The employment status and use of public assistance variables are used as proxies that measure men’s ability to provide for their family. Years of education is a continuous variable, the value of which indicates number of years of education received. Employment status is a dichotomous variable (1 = employed, 0 = not employed), created from a question that asks whether the respondents received wages or salary in the previous year. The use of public assistance variable is constructed from five items. The NSFG 2002 asks the respondents whether the family received (a) any welfare or public assistance, (b) food stamps, (c) WIC, (d) child care services or assistance, and (e) job training or job search help from social services. Those who received any one of these services are categorized as using public assistance. This variable is a dichotomous variable (1 = yes, 0 = no).
Men’s household characteristics
There are three variables used as indicators of men’s household characteristics: (a) wife or partner’s employment status, (b) gender of children under age 5 years, and (c) presence of school age children (ages 5 to 18 years). Two dummy variables are created for wife or partner’s employment—employed full-time and employed part-time—from the question regarding their employment status in the previous week. The reference category is not employed. Those who answered that their wife or partner worked both full-time and part-time are categorized as full-time employment. Two other variables are dichotomous variables, for which “at least one male child under age 5 years” and “presence of school age (5-18 years) child(ren)” are coded 1.
Control Variables
Household income, respondents’ marital status, age, race and ethnicity, foreign born status, religion, and number of children under age 5 years are controlled for. The NSFG 2002 provides income data only in ordinal categories, and therefore, household income is an ordinal variable (1 = less than $10,000; 2 = $10,000-19,999; 3 = $20,000-29,999; 4 = $30,000-39,999; 5 = $40,000-49,999; 6 = $50,000-59,999; and 7 = $60,000 and more). The variable marital status is a dichotomous variable (1 = married, 0 = cohabiting). Age is a discrete variable.
For the race and ethnicity variable, respondents are classified as Hispanics regardless of race if they were screened as Hispanics in the NSFG 2002 data. Non-Hispanic American Indian, Alaska Native, Asian, Native Hawaiian, and other Pacific Islander are combined as other races due to the small number of study subjects. The race and ethnicity variables are three dummy variables: Non-Hispanic Black, Hispanics, and Other races, with non-Hispanic White as the reference category. Fathers who were born outside the United States are coded 1 for the foreign born status variable and those born in the United States are coded 0.
To control for the effect of conservative beliefs (Gaunt, 2005), the variable religious fundamentalism is created from a question that asks fathers “Which of these do you consider yourself to be, if any?” Those who identified themselves as born again Christian, charismatic, evangelical, or fundamentalist are categorized as fundamentalist (coded 1), and those who answered “none of the above” as nonfundamentalist (coded 0). Number of preschool children is a dichotomous variable, for which “having more than one child under age 5 years” is coded 1. Table 1 provides descriptive statistics on all independent and control variables.
Descriptive Statistics of Independent and Control Variables (n = 613)
Analytical Approach
Logistic regression models are used to examine what variables increase the odds of fathers’ daily child care and play. For the dependent variable, physical care, the first model examines the effects of socialization, the second model is to analyze the effects of men’s socioeconomic status, and the third model is for the impact of men’s current family characteristics, controlling for household income, marital status, age, race and ethnicity, foreign born status, religious fundamentalism, and number of preschool children. The fourth model includes all the variables. For the dependent variable, play, an analysis is done for all variables (i.e., the model equivalent to the fourth model for physical care), and the results will be compared with those of physical care.
Results
Fathers’ Involvement: A Descriptive Account
Table 2 shows the frequencies of fathers’ involvement in two types of activities. In this sample, approximately half of fathers (49.8%) claimed they did physical care every day and another half (50.2%) said they performed such care less than once a day or not at all. On the other hand, 80.4% of fathers claimed they played with their young child(ren) every day and 19.6% played less than once a day. These results show that there is a significant difference between the percentage of fathers who provide physical care every day and the percentage of those who play every day.
Frequencies of Father Involvement in Physical Care and Play
Factors Associated With Fathers’ Daily Involvement With Children Under Age 5 Years
Physical care
Table 3 shows the results of logistic regression analysis on fathers’ daily involvement in physical care of children under age 5 years. Among the explanatory variables, robust effects are found in six variables—being raised by one’s biological father, years of education, use of public assistance, wife or partner’s full-time employment, presence of male child under age 5 years, and presence of school age child(ren). In Model 4, which includes all the variables, the odds ratios of these variables are 1.618 (p < .05), 1.206 (p < .001), 1.850 (p < .01), 2.167 (p < .01), 1.453 (p < .05), and 0.628 (p < .05), respectively. This means that the odds of fathers’ daily involvement in physical care increase by 61.8% if men are raised by their biological fathers, by 20.6% for each additional year of men’s education, by 85.0% if the couple used public assistance, by 116.7% if respondents’ wives or partners are employed full-time, and by 45.3% if the couple has a young male child. Having school-age children, on the other hand, decreases the odds of paternal daily involvement in physical care by 37.2%. Wife or partner’s part-time employment has a positive association with father involvement, but its odds ratio is only marginally statistically significant (p < .10). Respondent’s mother’s employment and respondent’s employment status have no significant effects on fathers’ daily involvement in physical care.
Logistic Regression Analysis for Variables Predicting Father’s Daily Involvement in Physical Care of Children under Age 5 (n = 613)
Note: R = reference category.
p < .10. *p < .05. **p < .01. ***p < .001.
Play
How do the above findings compare with fathers’ daily involvement in play? Table 4 shows the results of logistic regression analysis on the dependent variable, play. The same independent and control variables as the above analysis in Model 4 are included in the analysis, but the variables that had no significant effects on physical care had no effects on play either, and therefore, these nonsignificant variables are omitted from the table. Among the variables that had impacts on physical care, only two variables—being raised by biological father and years of education—have statistically significant effects on play (at p < .05). Three other variables that have significant effects on physical care—use of public assistance, wife or partner’s full-time employment, and presence of school age children—have only marginally statistically significant effects on father involvement in play (p < .10), and child’s gender has no discerning effect at all.
Logistic Regression Analysis for Selected Variables Predicting Fathers’ Daily Involvement in Play With Children Under Age 5 Years (n = 613)
Note: R = reference category.
p < .10. *p < .05. **p < .01. ***p < .001.
Among the control variables, only Hispanic status had statistically significant and robust effects on both types of paternal involvement. Hispanic fathers are less likely to provide physical care for, and play with, their young children every day compared with non-Hispanic White fathers. The “other races” category of fathers also reduces the odds of play, but this variable has no statistically significant association with physical care of young children.
Discussion and Conclusion
This study hypothesizes that men’s socialization, men’s socioeconomic status, and household characteristics or present family context have impacts on paternal involvement in physical care—tasks typically considered to be the mother’s job. The analyses show that among married and cohabiting fathers who live with their biological or adopted children under age 5 years, these three general areas have important influences, though not all variables included in this study do. Fathers are more likely to claim they give physical care daily if they were raised mostly by their biological father, they received higher education, the couples used at least one social assistance program, their wives or partners are employed (especially full-time), and at least one of the young children is male. On the other hand, the odds that fathers will take part in physical care are reduced when couples have school-age child(ren) in the same household. Among these factors, only upbringing by biological fathers and years of education had statistically significant effects on play. The most striking difference was observed in the effects of the gender of young children. Sons encourage daily paternal involvement in physical care, but child’s gender makes no difference in father involvement in play.
In this study, the important aspect of men’s socialization was whether they were raised by a biological father, not men’s observation of parental housework sharing. Men who were raised by their biological fathers were more likely to claim that they play with and provide physical care to their young children every day. This study cannot assess exactly what aspect of being raised by a biological father impacts paternal involvement for men. It may be because biological fathers invest more in their children (Amato, 1987; Hofferth & Anderson, 2003; Marsiglio, 2004) due to biological ties (e.g., Daly & Wilson, 2000), or that their role is more institutionalized compared with, for instance, fathers in remarriage or cohabitation (e.g., Cherlin & Furstenberg, 1994). Being raised by biological fathers during the time respondents grew up may have meant a close father–son relationships and/or consistent, greater paternal involvement. Families in our society are, however, increasingly diversified, and today, high-level paternal involvement is observed among social fathers as well (Berger et al., 2008; Gorvine, 2010; Marsiglio, 2004). When more boys from diverse family backgrounds grow up and become fathers, we might be able to assess how their role-taking is affected by biological and social ties with their fathers, quality of father–son relationships, and so on.
One may argue that an upbringing by biological fathers indicates men’s parents’ continuous marriage because biological mothers are more likely to have custody of their children in the case of divorce. Interestingly, however, a separate analysis (the results of which are not shown) using another item in the NSFG 2002 that asks whether respondents always lived with both biological parents showed no significant impact of this variable on paternal involvement. This seems to suggest that the important factor is not an intact family, but perhaps positive childhood experiences with fathers or the strong presence of a father figure in men’s childhoods.
A higher level of daily paternal involvement by highly educated men and by men who are unable to fulfill the primary provider role may suggest that the change in the paternal role is realized through two parallel channels: egalitarian value orientation acquired through higher education and men’s declined economic conditions that disallow the breadwinning role. As the sole-provider role becomes increasingly unachievable for many men, fathers who are in economically disadvantaged positions might take a more active role in child care and derive their masculine identity from it—possibly rejecting hegemonic masculinity or “undoing gender” (Brines, 1994; Deutsch, 2007; Griswold, 1993; Shows & Gerstel, 2009). This study, however, used only two items as measures of men’s ability to provide: employment in the previous year (which has no effects) and the use of public assistance. These measures are somewhat limited. The mechanism behind cultural and behavioral changes observed among men could be delineated through future research, which should investigate further how social class and changing economic conditions for men relate to paternal involvement and men’s (re)construction of masculine identity.
This study found a strong, significant effect of maternal employment (especially full-time) on paternal involvement in daily physical care of young children. At a glance, this finding seems to support the time availability perspective and relative resources perspective, and contradicts many past studies that found no effect of maternal employment (e.g., Combs-Orme & Renkert, 2009). As many gender scholars point out, equality in the division of household labor has not yet been achieved (e.g., Hays, 1996; Hochschild, 1989; Tichenor, 2005). It is important to remind readers that this study uses men’s self-report on frequencies of their involvement. The claim that they do diapers, and so on, every day does not necessarily indicate that these fathers are equal sharers. Physical care of young children entails much more than changing diapers, dressing, and bathing children. Fathers (and mothers) tend to rate their own contributions higher than their spouses do (Mikelson, 2008), which may be because fathers often compare themselves to men of past generations (e.g., Hochschild, 1989) or take no notice of the “invisible” work mothers do (e.g., Tichenor, 2005).
Coley and Morris (2002) and Mikelson (2008), however, stress the importance of using fathers’ self-reports, particularly for the case of resident fathers, because mothers’ reports (that typically underestimate paternal involvement) are not necessarily more reliable than fathers’ reports. This is not to say that the present study is more accurate than studies that use mothers’ reports. But it is intriguing that, by using fathers’ self-reports, the effect of maternal employment is strong and significant, which is not necessarily the case in studies that use mothers’ reports. It is possible that the gap in perceptions of paternal involvement is larger between employed mothers and fathers who are married or cohabiting (i.e., sharing the household). For instance, men who live with employed wives or partners may perceive that they make great contributions to child care tasks because they focus on what they do when their wives or partners are absent (due to employment). Employed mothers, on the other hand, may perceive that their husbands or partners are not doing much because these women focus more on what is not done by fathers. Employed mothers of young children are probably some of the most overworked people in our society (if we combine both paid and domestic work), and this may cause their perceptions to differ from those of their husbands and partners. Though there is no way to be sure, the use of different genders’ self-reports—men’s in this study versus women’s in earlier research—may have caused the contradiction in findings regarding the impact of maternal employment, and thus the use of residential fathers’ self-reports in this study makes an important contribution to the fatherhood and gender literature.
The effects of age and gender of children imply the persistent gendered division of labor by couples. When a couple has school-age children, physical care of infants and preschoolers seems to be assigned to mothers. In terms of the gender of young children, the present study found no effects on play but a strong, significant effect on physical care in favor of sons. The different levels of father involvement in these two activities may suggest that it is not fathers’ preference for sons, but perhaps their hesitation to care physically for female children due to, for example, cultural expectations such as Christian sexual modesty or fear that they will be accused of sexual abuse.
Several limitations of this study need to be discussed. First, the NSFG 2002 gives a limited range of questions regarding men’s involvement in child care, and the two dependent variables in this study are based on single items. The question used to create the physical care variable is constructed from one question that asks how often the respondents bathed, diapered, or dressed their children, or helped them bathe, use the toilet, or dress. Unfortunately, fathers who answered “every day” to this question may not necessarily do all of these activities every day or spend the same amount of time on these activities as their wives or partners. There may be fathers who do other types of physical care (e.g., feeding) every day, but not the listed activities. Ideally, several questions should be asked on each child care activity, but the question used in this study was the best available question on physical care involvement in the NSFG 2002. The answer category “several times a week” allows respondents to make different interpretations, and therefore the level of involvement of fathers who chose this answer can range from very high to very low. A better way to measure level of involvement would be of great use to researchers.
Second, as discussed earlier, the NSFG 2002 data do not allow the inclusion of certain important variables such as men’s relationship with their children (i.e., biological or adopted), their wives or partners’ relationships with their children (i.e., biological, adopted, or step), wives’ or partners’ relative income contributions, men’s occupation and work hours, beliefs in gender ideology held by couples, and couples’ marital quality. Male ability to provide economically is measured by the couple’s use of public assistance and men’s employment status in this study, but other measures of men’s socioeconomic status are needed.
Last, this study is limited to married or cohabiting fathers who live with their biological or adopted child(ren) under age 5 years. The results cannot be applied to single fathers, step (or social) fathers, fathers who live away from their children, or fathers who have no young children under age 5 years.
Despite these limitations, this study contributes to the fatherhood literature by demonstrating that fathers’ roles are shaped through men’s socialization, education, and economic circumstances, that these roles are negotiated in particular family contexts, and that the factors relevant to involvement in children’s physical care (i.e., the mother’s traditional job) differ from those relevant to play. It is important to understand what encourages men to take part in physical care of children because such involvement is found to be beneficial not only to child development (Bronte-Tinkew et al., 2008; Marsiglio et al., 2000), but also to better marital quality and greater gender equality (e.g., Erickson, 1993; Hewlett, 2000), and to fathers’ psychological well-being (Schindler, 2010). This study also makes an important contribution by using residential fathers’ self-reports on their involvement. Although this study does not identify men who share all child care tasks equally, it clarifies those factors associated with men who claim the daily provision of physical care as their routine work.
Footnotes
Acknowledgements
I would like to thank Loretta Bass, Brian Bentel, and the anonymous reviewers at the Journal of Family Issues. I thank the participants at the 2008 NSFG Research Conference for their comments during the meeting.
An earlier version of this article was presented at the 2008 NSFG Research Conference.
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
The author(s) received no financial support for the research, authorship, and/or publication of this article.