Race/Ethnic Differences in Father Involvement in Two -Parent Families: Culture, Context, or Economy

Sandra L. Hofferth Institute for Social Research University of Michigan Ann Arbor, MI 48106-1248 Hofferth@umich.edu

February 24, 2001

Revised version of paper presented at the Urban Seminar on Fatherhood, April 23-24, 1999, Harvard University, Cambridge, MA. Helpful comments from Joan Aldous and David Harris are gratefully acknowledged as is the research assistance of Julia Naudain and Patricia Hall. This research was made possible by funding from the NICHD Family and Child Well-being Research Network and from the Alfred P. Sloan Foundation through the University of Michigan Center for the Ethnography of Everyday Life.

ABSTRACT Race/Ethnic Differences in Father Involvement in Two -Parent Families: Culture, Context, or Economy This paper examines the contribution of economic circumstances, neighborhood context, and cultural factors to explaining race/ethnic differences in fathering in two-parent families. Data come from the 1997 Child Development Supplement to the Panel Study of Income Dynamics, a nationally representative sample of children under age 13. Black childrens fathers exhibit less warmth but monitor their children more, Hispanic fathers monitor their children less, and both minority groups exhibit more responsibility for childrearing than white fathers. Economic circumstances contribute to differences in paternal engagement and control, while neighborhood factors contribute to differences in warmth and responsibility. Cultural factors such as intergenerational fathering and fathering and gender-role attitudes contribute to explaining differences from whites in control and responsibility on the part of both blacks and Hispanics.

INTRODUCTION The lag in academic achievement of minority compared with nonminority students is a critical national problem (Jencks & Phillips, 1998). While the substantial socioeconomic difference between their families is a key factor in achievement differentials, the gap between minority and white students is present or even greater among middle-income than low-income families (Jencks & Phillips, 1998). Contributing factors other than income may involve parenting practices. While at least one study has examined maternal behavior (Phillips, Brooks-Gunn, Duncan, Klebanov & Crane, 1998), few have examined the paternal side. A fathers emotional investment, attachment, and provision of resources are associated with childrens cognitive development and social competence (Lamb, 1997; Pleck, 1997), and such investment varies widely. Differential achievement among various groups may result from differential father involvement (McLanahan & Sandefur, 1994). Yet, in spite of the growing body of research examining fathers and fathering, research focuses primarily on majority, not minority, families (McAdoo, 1988). In addition, research often focuses on father involvement by fathers outside the home rather than by coresidential fathers (Mott, 1990; Seltzer & Bianchi, 1988). Minority children are especially likely to grow up without a father present (Hernandez, 1993). While the part played by absent fathers is important, it is also important to assess fathering behavior in two-parent families; research has found pre-divorce behavior to explain much of the relationship between divorce and childrens outcomes, particularly for boys (Cherlin & et al., 1991). Fathers who are more involved with children predivorce may remain more involved postdivorce. Including remarriage families is important; little research exists on fathering after divorce and remarriage. Fathering patterns may vary by race/ethnicity for several reasons. Different fathering patterns by race/ethnicity may be due to different economic circumstances and neighborhood environments. Differential paternal involvement may also result from attitudes and values which are cultural in origin, 3

since parenting occurs in a social context (Bronfenbrenner, 1979). Given that fathering plays an important part in childrens development, research is needed to determine whether economic conditions, cultural beliefs, or other factors are more important in explaining majority-minority fathering differences. Each has very different implications for public policy efforts to increase father involvement. This paper examines factors associated with differential involvement with fathers among black, white, and Hispanic children in two-parent families, using data drawn from a new study of children and their parents conducted in 1997. This paper takes advantage of a rich set of parenting measures compared to those used in past studies, a new set of measures of the time parents spend with individual children, and attitude and behavior data obtained directly from both mothers and fathers. While many studies of fathers and fathering are based upon small selective samples, this study is based upon a large nationally representative data base which can be generalized to all children living in two-parent families in the United States.

THEORETICAL PERSPECTIVES Key Parenting Practices Children learn through interacting with and observing parents (Bandura, 1969), and such learning is affected by the level of interaction or engagement (Lamb, Pleck, Charnov & Levine, 1985). While mothers engagement is generally high, fathers vary in the level of engagement with children and research has attempted to quantify this variation by examining the time fathers spend with them. Recent research finds fathers to spend about 1.5 hours on a weekday and about 3 hours on a weekend day with children under age 13 (Yeung, Sandberg, Davis-Kean & Hofferth, 2001). Other studies have shown levels of engagement ranging from 2 to 2.8 hours per day (Pleck, 1997). Fathers also vary in the degree of responsibility they assume for their children. Responsibility encompasses a fathers contribution to his family in his role as economic provider (McAdoo, 1988; Pleck, 1997), and 4

employment and work hours are key indicators of this dimension. However, responsibility also encompasses a degree of management of the childs welfaremaking sure that the child is fed, clothed, housed, monitored, managed, examined by physicians, and cared for when needed, a dimension which has not been examined in the context of race/ethnic differences. Few men take primary responsibility but many share it with their partners (Sandberg, 2000; Hofferth, Pleck, Stueve, Bianchi & Sayer, 2001). Additionally, most developmental psychologists argue that the quality of parenting and of the parent-child relationship are crucial to developing competent children. According to Baumrind's typology (1967; 1978; 1991), parents are characterized as using one of four disciplinary styles: authoritarian, authoritative, permissive, and uninvolved, based upon the cross-classification of demandingness and responsiveness (Maccoby & Martin, 1983). The authoritarian style, a combination of warmth and control, is believed to create the best environment for child development. In the present study, warmth measures responsiveness by providing information on the emotional content of the interaction between parent and child. Parental monitoring and control, which includes setting rules and enforcing them, measures demandingness.

Cultural Differences in Parenting Since parenting occurs within and interacts with that context, socialization practices in race/ethnic minority families may differ from those of white families (Garcia-Coll, Crnic & et al., 1996). Ogbu (1981), Kohn (1977) and others have proposed that childrearing is oriented toward the development of instrumental competencies that are adaptive for the settings in which children are expected to develop and participate. Minority parents may teach skills to assist their children navigate ethnic and racial barriers. As suggested by previous research (Bartz & Levine, 1978; Baumrind, 1968; Garcia-Coll, 1990), black parents may exhibit more control and less warmth than white parents, 5

perhaps as a reaction to a more dangerous and hostile environment. Alternatively, values and attitudes may vary across different ethnic groups as they have been shown to vary by socioeconomic status (Alwin, 2001; Kohn, 1977). Latino families have consistently been shown to endorse warm parenting practices, and to emphasize collective forms of obligation over individualism (Fuller, Holloway & Liang, 1996; Delgado-Gaitan, 1994). Latino fathers may exert less control over their children in the context of shared extended-kin responsibility for childrearing (Baca Zinn, 1994). The greater familism of minority groups may increase the responsibility of fathers for children. While research has examined interactions and control attempts of fathers, little research has examined the time minority fathers spend with children. A recent study shows a puzzling picture of black fathers spending less time in meals but more time talking, reading, and doing homework with their children than white fathers (Cooksey & Fondell, 1996). However, time expenditures to date are based upon single item reports (see (Yeung, et al., 2001) for an exception) rather than diaries and may be inaccurate. The greater familistic orientation in many Hispanic communities could increase the value to fathers of spending time with children (Baca Zinn, 1994; Vega, 1990), but, again, little is known about the value to fathers of raising children and the time they spend with them. Differences in Intergenerational Experiences. Learning theory posits that motivation, enjoyment of raising children, and appropriate skills are learned and sustained through a variety of mechanisms (Pleck, 1997; Lamb, et al., 1985). For example, growing up with a father who was very involved could be related to exhibiting more involved fathering behaviors as an adult. Taking parenting classes could be linked to more involvement. Finally, learning parenting from ones father could be linked to fathering behaviors in later life. To the extent that blacks, whites and Hispanics differ in exposure to father involvement while growing up, their own fathering behaviors will differ. The smaller proportion of childhood black children spend living with a father could contribute to differential involvement of black men with children. 6

Gender Role Attitudes. Differences among black, white, and Hispanic parents in gender role and parenting attitudes may explain differences in fathering behaviors. Traditional gender role theory argues that mothers' and fathers' attitudes and values will determine how much time they spend with their children. Traditional fathers provide instrumental support, including financial support and discipline, but less emotional support. Consequently, they are likely to exhibit less warmth with children than nontraditional fathers. From a role perspective, fathers who hold traditional values with regards to marriage and parenthood are likely to be less involved, whereas fathers who endorse gender equity are likely to be more involved with their children. To the extent that fathers endorse the importance of fathers in childrens lives, they should be warmer in their relationships with their children and take more responsibility for them. Based upon the strong financial role of the black mother, black childrens fathers are likely to have less traditional attitudes towards marriage and motherhood than do white childrens fathers. Black fathers may espouse more gender equitable and more individualistic attitudes than white fathers as well. Hispanic childrens fathers are likely to have more traditional attitudes towards marriage and mothering. However, the research on Latino families suggests that machismo is more a stereotype than an accurate depiction of the Hispanic male (Baca Zinn, 1994), and maternal employment has increased equity in Mexican families. Less traditional attitudes may be associated with more warmth and responsibility for children. Different beliefs about the appropriate roles of men and women may also alter the parenting of various groups, but, again, little is known about gender-role attitudes of minority fathers.

Economic Differences among Race/Ethnic Groups. Differences in parenting among different race/ethnic groups may result from such economic differences as lower labor earnings, lesser employment, and fewer hours of work. Differences in educational levels between blacks, Hispanics and whites may help explain differences in parenting 7

practices as education affects the human capital of parents and, therefore, their wage rates. There are several ways income and work hours could influence parenting practices. First, fathers who contribute to the family economically may feel that they have made their contribution and that they do not need to contribute in other ways. The more hours they work, the less time they have to participate, as well. Fathers who do not work and who do not contribute economically to the family may attempt to make up for this lack of responsibility in the traditional sense by becoming involved in other ways. On the other hand, viewed more as resources, earnings could be positively related to behaviors such as parental monitoring. According to Goode (1982) and others, middle- and upper-income parents have monitored and controlled their childrens behavior more than lower-income parents, because they have control over valuable resources, both money and opportunities. However, not only are there differences in socioeconomic status vis a vis the outside world, but minority and majority families also differ in the relative status of spouses within the marital relationship. Division of labor theory (Becker, 1991) suggests that earning more is associated with doing less household work, including caregiving. Bargaining theory attends to the spouses earnings as well (Blau, Ferber & Winkler, 1998). Earnings of husbands and wives are more equal in black families than in either white or Hispanic families. As a result, we would expect black fathers to spend more time caring for children than white fathers. Black husbands have lower rates of employment and work hours than white husbands, while black wives have higher rates of employment and work hours than white wives. According to bargaining theory, in families in which their mothers work more, children should spend more hours with their fathers.

Differences in Residential Neighborhood Characteristics. Previous research suggests that parenting varies by context. Since blacks, whites and Hispanics live in very different neighborhoods, these contexts may explain race/ethnic differences in fathering. 8

Social characteristics of neighborhoods have been shown to be at least as important as family economic characteristics in containing negative youth behavior (Sampson, Raudenbush & Earls, 1997). We anticipate that the extent of control fathers exert over their childrens behavior will be a function of two aspects of neighborhoodthe overall quality of the neighborhood and its race/ethnic composition. Parents may parent their children differently in an ethnically homogeneous community than one in which they are a minority of the population (Duncan, Brooks-Gunn & Klebanov, 1994). In particular, an ethnically homogeneous community may support ethnic traditions which maintain a sense of community cohesion and pride, leading to greater familial involvement by fathers and less control. In contrast, fathers are expected to exert more control in neighborhoods that are less than ideal for raising children.

Other Explanations for Race/Ethnic Differences in Fathering Differences between blacks and whites or between Hispanics and whites may be due to a number of factors other than economic, neighborhood, and cultural differences. It is important to control for these possible confounding factors in examining race/ethnic differences in parenting. Child Characteristics The present analysis adjusts for differences between race/ethnic groups

in age and gender of the child. Fathers spend less time with children as they grow older (Yeung, et al., 2001). Since their higher fertility levels imply more young children, black and Hispanic fathers may appear more involved if we do not control for child age. Family Structure and Composition Differences. Adjustment for differences in family types between race/ethnic groups is important. The most important distinction we make is whether the childs residential father is the biological or nonbiological parent. The latter tend to be less involved with children (McLanahan & Sandefur, 1994). Among two-parent families, black children are more likely than white children to live with a nonbiological father, while Hispanic children are similar to whites in the proportion living with a biological father. 9

The age of the father and number of children are also controlled. The age of the parents may influence their interaction and parenting style. In addition, controlling for the number of children in the family is important; with more children, parents are expected to spend less time with each one (Zajonc & Markus, 1975).

DATA AND METHODS Data: The Child Development Supplement to the Panel Study of Income The study sample comes from the Panel Study of Income Dynamics (PSID), a nationally representative sample of U.S. men, women, children, and the families in which they reside which has been followed for more than 30 years. Until 1997, measures were collected annually from interviews with one adult respondent about all family members, but only limited information was available on children and parental interaction. During the spring and fall of 1997, information on up to two randomly selected 0-12-year-old children of PSID respondents was collected from the primary caregivers, from other caregivers, and from the children themselves (Hofferth, Davis-Kean, Davis & Finkelstein, 1999). The Child Development Supplement (CDS) completed interviews with 2,394 child households and about 3,600 children. The response rate was 90 percent for those families regularly interviewed in the core PSID and 84 percent for those contacted the first time in 1997 for an immigrant refresher to the sample, with a combined response rate for both groups of 88 percent. When weights are used, as is done throughout this paper, the results have been found to be representative of U.S. individuals and their families (Fitzgerald, Gottschalk & Moffitt, 1998a). Weights are also applied to adjust for differential nonresponse across instruments. Case counts represent actual sample sizes. The sample used in this paper consists of 1,229 children living with two parents (stepfather as well as biological father) and for whom information reported by each parent was available. The sample for analysis of parental time with children consists of 1,172 children, since the majority but not all 10

families completed time diaries. The sample size including those with complete information on the control variables ranges from 937 to 1,077 children, depending upon the number of complete responses to the parenting measure and on the ages of the children measured.

Measurement of Race/Ethnicity While ethnic groups exhibit different cultural traditions and practices, groups defined by racial or physical characteristics may differ in cultural characteristics as well (Taylor, 1994a). For this reason, both black and Hispanic fathers may differ from white fathers in parenting practices. In this paper race/ethnicity is categorized as non-Hispanic white, non-Hispanic black, Hispanic, and other based upon the report of the household head on a set of Census Bureau questions about race and about Hispanic origin that were standard up until 1999. The reports were compared with the report of the race/ethnicity of the child; in cases of inconsistency, the heads report takes precedence. In 1997, black children represented about 15 percent of the U.S. population and Hispanic children another 15 percent. As a result of the selection of children living with two parents, in our sample 8 percent were black, 11 percent were Hispanic, 5 percent were of other race/ethnicities, and 76 percent were white. In the United States, about 62 percent of Hispanics in the United States are of Mexican descent. While a considerable proportion of persons of Mexican descent have lived in the United States for generations, recent immigration increased the Mexican-origin population by one-third since the 1980s and, thus, first and second generation groups in the United States (Taylor, 1994c). Our sample reflects this reality. Eighty percent of the 98 Latino families in our sample arrived post-1968. Of these 75 percent are of Mexican descent, with most of the remainder from Central America and a few from South America. More than 80 percent of the immigrant children are second generation. Thus our Hispanic sample describes a primarily Mexican population, in which most families arrived within the past

11

30 years. The other category includes children of Asian, Native American, and other backgrounds; there are too few to analyze separately.

Measures of Parental Involvement and Parenting Time Children Spend Engaged with their Fathers. The CDS collected a complete time diary for one weekday and one weekend day for each child age 0-12 in the family. The data obtained are generally considered superior to those obtained using standard questions asked in most surveys because of internal consistency and reduced social desirability (Juster & Stafford, 1985; Hofferth, 1999). The time diary, which was answered primarily by the mother or by the mother and the child, asked several questions about the childs flow of activities over a 24-hour period beginning at midnight of the randomly-assigned designated day. These questions ask the primary activity that was going on at that time, when it began and ended, and whether any other activity was taking place. An additional questionWho was doing the activity with child?when linked to activity codes such as playing or being read to, provides unbiased details on the extent of one-on-one interactions of others with the child. For this analysis, times in which the father (biological father, stepfather, foster or adoptive father) was engaged in activities with a child were coded as father engaged. Times engaged were summed over all activities for weekdays and weekends for each child. Weekly time was computed by multiplying weekday time by 5 and weekend day time by 2. Variables indicate whether time was spent with that parent and the total number of hours, with 0 for none. To test results from previous research, we also examined the time children spend with their fathers in two specific activitiesreading with the father and eating meals together. Responsibility. The 8 responsibility items used here focus upon the care of childrenbathing children and changing diapers, disciplining children, choosing childrens activities, buying childrens clothes, driving children to activities, selecting a pediatrician and making appointments, selecting a child 12

care program, preschool, or school, and playing with children. 1 Response categories are: 1) I do this, 2) another household member does this, 3) I share this task, and 4) someone else does this task. If the respondent did the task, the response was coded 2, if the respondent shared it, it was coded 1; otherwise the task was coded 0. Scores were summed over all items. Overall scale reliability using Cronbachs alpha was 0.73. Parental Warmth. Parental warmth is a 6-item scale developed by Child Trends for use in measuring the warmth of the relationship between child and parent. The questions, asked of each parent about all children, ask how often the parent hugged the child, told the child they love him/her, spent time with child, joked or played with child, talked with child, and told child they appreciated what he/she did. The response categories range from 1) not in the past month to 5) every day. A scale was created by summing the number of behaviors that the parent said they did with the child in the past month. Scale reliability was 0.77. Parental Monitoring and Control. Parental monitoring is measured by a set of 9 items asking each parent of children 3 and older whether they have rules setting limits on their childrens activities, their schedules, their food, their whereabouts, and their homework, and whether they discuss these rules with their children. In contrast to the other scales, this scale measures control across all children. 2 The response categories (reverse coded) range from 1) never to 5) very often, with 45 the highest possible score. The reliability coefficient was 0.73. Measurement of Key Variables Economic Characteristics. Economic variables are measured, first, by the employment of

These measures are self-reports; they give the respondents perception. While subjective perception is not the same as what would be measured were time diaries of parental activities collected, this scale has been used in a variety of previous studies, such as the National Survey of Families and Households. The crucial factor is that there is no reason to expect minority-majority differences in reporting on this scale. 2 It is unfortunate that this measure is not child-specific, since control may vary from child to child. However, lacking any alternative, we take this as a measure of the overall degree of control a father exerts over his children.
1

13

mother and father. Employment status is a 4-category variable describing the employment of both parentsmale breadwinner-female homemaker family; dual earner family; female breadwinner-male non-employed family; and two parent-neither employed family. White and black children are most likely to live in a dual earner family; Hispanic children (and other races) are most likely to live in a male breadwinner-female homemaker family. Black children are also more likely than white children to live in a female breadwinner-unemployed male family. Work hours vary, with black women working more hours than any of the others. In contrast, black men work the fewest hours, resulting in only a 5hour difference between black husbands and wives. Consistent with the small hourly difference in time, mothers and fathers earnings are the most similar in black families, with black mothers earning $17,000 on average and black fathers $20,000. Hispanic mothers earn only $5,400, compared with $18,000 for Hispanic fathers. White fathers earn the most, $41,000, with $14,000 for mothers. Neighborhood. In the Child Development Supplement, the respondent defines his own neighborhood; the CDS then asks several questions about this neighborhood. The quality of the neighborhood is measured by how the respondent would rate the neighborhood the family lives in as a place to raise children. Of responses ranging from excellent to poor, excellent was coded 1, the rest 0. Neighborhood racial composition was obtained from questions asking whether none, less than half, about half, more than half and almost all the residents were black, Hispanic, and White. From the means, it is clear that each group is highly likely to live in close proximity to others of the same race/ethnicity. Gender Role Attitudes. Attitudes towards gender roles are measured by 20 standard items included in national surveys since the early 1960s (Mason, Czajka & Arber, 1976).3 Four response categories range from strongly agree to strongly disagree, with items (re)coded so that a high score indicates agreement with the item. Mean values were substituted for a small number of missing items.

14

Using principal components analysis, these 20 items were factored into four main factors separately for primary and secondary caregivers and then rotated using varimax rotation into four orthogonal factors. These represent 1) traditional marriage values Most of the important decisions in the life of the family should be made by the man of the house, 2) traditional mothering valuesPreschool children are likely to suffer if their mother is employed, 3) equityIf a husband and wife both work full-time they should share household tasks equally, and 4) individualismPersonal happiness is the primary goal in marriage. Factor scores assigned to each individual childs father have a mean of 0 and a standard deviation of 1. Scale reliabilities calculated using Cronbachs alpha were 0.75, 0.77, 0.63 and 0.47 for fathers attitudes towards marriage, traditional mothering, gender equity, and individualism. Fathering Attitudes and Skills. Questions on fathering were drawn from the Being a Father scale (Pleck 1997b) and from the Role of the Father questionnaire (Palkowitz, 1984), tapping the belief that the father role is important in child development. Items include: A father should be as heavily involved in the care of his child as the mother. and In general, fathers and mothers are equally good at meeting their childrens needs. The responses were scored from 1 (strongly disagree) to 4 (strongly agree). These eight items were factor analyzed using principal components analysis, but one factor was sufficient to describe these data. Consequently, after recoding so that a high score indicates a positive attitude towards fathering and substituting mean values for individual missing items, the total score on the 8 items was then obtained by summing. The reliability coefficient was 0.70 on this scale. Intergenerational Learning. The sources of knowledge of the father and mother about parenting were also explored, including the extent to which the respondents own father was involved in raising him/her, and whether they use their own father as a role model in their own parenting. The parent was asked whether he/she had ever taken a parenting class. Parenting classes include childbirth preparation classes; such classes often include a parenting component either before or shortly after birth and many
3

A complete list of the items in the attitude scales used in this paper is available from the author.

15

also have a support function. Finally, questions asked from whom the father most learned to parent their mother or father or someone else. Control Variables Variables controlled in the analysis include the age and gender of the child, number of children, fathers age, and whether the father is a biological or nonbiological father to the child.

RESULTS Mean Differences in Paternal Involvement In two-parent families, children average 14.93 hours per week engaged with their fathers (Table 1). This result is consistent with other research (Yeung, et al. 2001} which shows that children spend, on average 13 hours per week engaged with their (biological) fathers. Black childrens fathers are significantly less engaged than are white childrens, 12.76 hours compared with 15.35 for white childrens fathers (p<.01), a difference of 2.59 hours per week. There is no significant difference between Hispanic and white childrens fathers engagement. Consistent with previous research, black childrens fathers spend marginally less time eating meals with them; however, in contrast to that research, they also spend less time reading to children. This difference may be due to differences in the methodology used to collect data on time with children. Because of the lack of significance of meal time and reading time in both bivariate and multivariate models, we do not pursue these measures further. (Table 1 about here) Consistent with the literature, black childrens fathers rate themselves lower on warmth than do white fathers, while Hispanic childrens fathers do not differ significantly from whites. Again, consistent with expectations, black childrens fathers are more controlling than white childrens fathers, while Hispanic childrens fathers are less controlling. On responsibility for caring for children, both black and 16

Hispanic childrens fathers rank higher than white childrens. Could these differences be due to differences in attitudes? As expected, black childrens fathers hold less traditional mothering attitudes compared with white fathers, while Hispanic and other race fathers mothering attitudes are more traditional. While Hispanic fathers and fathers of other race/ethnicities hold more traditional marriage attitudes than white fathers, as expected, black fathers are unexpectedly also more traditional. Other research has pointed to discrepancies between the marriage expectations of black males and females (Taylor, 1994b). Compared with whites, and in contrast to the stereotype, Hispanic fathers are the stronger believers in gender equity. Finally, black, Hispanic and other fathers all hold more individualistic attitudes than whites. Overall, only 27 percent of childrens fathers said they ever had a parenting class (including childbirth classes). This is lowest for black fathers of whom only 17 percent ever had such a class. The proportion who said they learned to parent from their mother is high for all groups 7 out of 10. In contrast, the proportion of fathers who learned from their own father varies by race/ethnicity, with all minority fathers less likely that white fathers to learn to parent from their own father. While 63 percent of white fathers learned from their father, 45 percent of black, 44 percent of other and 38 percent of Hispanic fathers did so. Fathering attitudes are generally positive, but black, Hispanic, and other race fathers consistently hold less positive attitudes than whites. This analysis shows considerable race/ethnic differences in father involvement with their children. Some, such as the greater control and less warmth expressed by black fathers, were expected. Others, such as the greater responsibility shown by black and Hispanic childrens fathers, were not. Some of these differences in parenting may be explained by differences in the economic situations and living arrangements of these different groups. Others could be explained by attitudinal differences. Some differences in attitudes, such as the espousal of more traditional mothering among Hispanics and less traditional mothering attitudes among blacks, were expected. Unexpected were traditional marriage 17

attitudes among blacks, high levels of equity among Hispanic fathers and high levels of individualism among all three minority groups.

Hierarchical Regressions To examine the extent to which race/ethnic differences in fathering remain after controlling for family and neighborhood factors and gender role or fathering attitudes, we use a set of hierarchical models (Table 2). In the first step, measures of parenting are regressed on race variables only. In the second step our control variablesage, and gender of the child, fathers age, nonbiological father, and number of childrenare added. The third step adds economic characteristics of the familyeach parents earnings, education of father, employment status, and hours of mother and father. Measures of neighborhood are added in step 4 and measures of fathering attitudes are added in step 5. Finally, fathering attitudes are dropped and gender role attitudes substituted in step 6, since one of the gender role attitudes (equity) is strongly associated with positive fathering attitudes. We test the significance of the change in the effect of being in a particular racial or ethnic group on parenting due to the addition of each set of variables (Clogg, Petkova & Haritou, 1995)(Table 3). (Table 2 about here) (Table 3 about here) Childrens Time Engaged with Fathers Table 2, panel A, column 1, shows that the coefficient for childrens time engaged with their father is negative, reflecting a 2.4-hour lower weekly father engagement time for black compared with white children, but this effect is only marginally statistically significant (p<.10). Even though the blackwhite difference in engagement is relatively small, we examine what factors explain this important component of parenting. With controls for childrens age and gender of child, nonbiological father, fathers age and family 18

size (column 2), the effect of black race declines significantly and is no longer associated with childrens time with their father. Black families in our sample tend to be older, to have slightly older children and to have a larger number of children. Fathers spend less time with children in these circumstances. Black children are twice as likely as white children to be living with a nonbiological father, and children receive less paternal attention in such families. Controlling for factors that reduce time with children lowers the negative association of black race/ethnicity and time with the father. We do not examine further changes in the coefficient on black, since it is not significantly associated with paternal engagement in later models. Across all children, the more hours the father works and the more he makes, and the better the neighborhood (perhaps suggesting wealth), the lower his time with his children. Similarly, the more hours the mother works, the more hours the father spends. These findings are consistent with division of labor and bargaining perspectives. Positive attitudes toward fathering and grandfather involvement are associated with increased paternal time with children as are more equitable sex role attitudes.

Parental Warmth Consistent with the unadjusted means, black childrens fathers report significantly fewer warm behaviors with them than white childrens fathers (Table 2, Panel B, Column 1). This finding is consistent with previous research and suggestive of more traditional fathering behavior in black than white families. This result holds adjusting for all variables, though it declines or increases slightly at each step. About of the variance is explained in the full model. The effect of being black declines significantly with control for family structure and other background factors (column 2). This is because black childrens fathers are more likely to be stepfathers or cohabiting, and such fathers tend to be less warm with their stepchildren. Controlling for these differences in family structure reduces the differential in warmth between white and black fathers. 19

Controls for economic status reduce the black-white warmth differential, but the reduction is not statistically significant. Black childrens fathers and mothers wages are lower, and lower wages are associated with reporting fewer warm behaviors. Why higher wages are linked to warm behaviors may reflect differential parenting values by class, as suggested by Kohn (1977), with middle-class fathers more likely to provide a supportive, warm environment than lower-class fathers. We shall see later on that they also exert more control, thus suggesting the dominance of an authoritative model of parenting (warmth with control) among middle class-fathers. Controlling for neighborhood increases the black-white differential in warmth substantially and the increase is significant at p<.10. Black children tend to live in black neighborhoods, in which locations the data suggest that fathers tend to be warmer. Not adjusting for living in these neighborhoods masks the fact that black fathers tend to show less warmth than white fathers. Adding intergenerational fathering and fathering attitudes (column 5) increases the effect of being black on warmth slightly, but not significantly, even though there are significant relationships between fathering attitudes and warm behaviors. Black childrens fathers have a less positive attitude towards fathering than white childrens and having a less positive fathering attitude is also associated with fewer warm behaviors. While having a grandfather involved in raising ones own father is associated with warm behaviors, black childrens fathers do not differ from those of whites on grandfather involvement.

Adding gender role attitudes is associated with no change in the black-white difference in warmth (column 6 vs. 4), which remains significant. Black fathers espouse more individualistic attitudes, which are associated with less warmth. Believing in gender equity is associated with greater warmth, but on this attitude blacks and whites do not differ.

20

Parental Monitoring and Control Race/ethnic differences in control are shown to be highly significant in Table 2, with 2.4 points greater control by black fathers and 2.4 points less control by Hispanic fathers compared with white fathers (Panel C, column 1). The black-white difference in control/monitoring declines significantly when we adjust for background characteristics of the child and family. Blacks have more 6-8 year olds than whites, and parents control such children more. Blacks also have more children, and having more children increases the amount of control exerted. The Hispanic-white difference, in contrast, increases significantly. Hispanic children are more likely to live with a biological father, which is associated with more control. Adjusting for differences in living arrangements increases the negative relationship between Hispanic origin and monitoring. Controlling for economic status results in increasing the black-white difference in monitoring by a small and nonsignificant amount while decreasing the Hispanic-white monitoring difference by a large and significant amount. Black families are more likely to have no breadwinner, and that is associated with less monitoring. Black fathers also have lower wages, which are associated with less monitoring. Adjusting for these variables increases the black-white difference over what it was in the previous step. Economic differences also explain some of the Hispanic-white monitoring difference. Hispanic men earn less money, which is associated with lower monitoring. When this is taken into account, the monitoring difference is reduced. Our data show that black and Hispanic families report living in less desirable areas for raising children than white families. Under these circumstances, we would expect parents to monitor and control childrens behavior more, as black fathers do. However, controlling for neighborhood reduces only slightly and insignificantly the black-white difference in monitoring because the neighborhood rating is not linked to monitoring. The Hispanic-white monitoring difference increases when neighborhood is controlled, though the increase is not statistically significant. Hispanics live in majority Hispanic 21

neighborhoods; monitoring is more common in such neighborhoods. Hispanics monitor less, given the types of neighborhoods they live in. Adding attitudes does not significantly change the effect of black race on monitoring while it significantly increases the effect of Hispanic origin. Hispanic fathers are less likely to have learned parenting from their father than white fathers. Not having learned from their father is associated with more monitoring. In addition, Hispanic fathers are more likely to have had an involved father. Having had an involved father is associated with more monitoring. Adjusting for these two factors, Hispanic fathers monitor even less than white fathers. Finally, adding gender role attitudes significantly increases the strength of black race/ethnicity and significantly reduces the effect of Hispanic origin on monitoring. Hispanic and black childrens fathers are more likely to express individualistic attitudes than white childrens, and greater individualism is associated with lesser monitoring and control. When these differences are adjusted, the effect of being Hispanic on monitoring and control is reduced and the effect of being black is increased. Hispanics also express more gender-equitable attitudes, and gender-equitable attitudes are associated with greater monitoring.

Parental Responsibility Consistent with mean differences, both Hispanic and black childrens fathers report themselves significantly more likely than white childrens fathers to take responsibility for their care (Table 2, Panel D, column 1). However, while the coefficient for Hispanic origin remains significant over all models, the coefficient for black declines to insignificance as variables are added. The sizes of the coefficients in column 1 are consistent with the sizes of effects in Table 1; the effect of Hispanic origin is considerably larger than the effect of black race. In column 2, controls added for background characteristics reduce both effects, significantly for 22

Hispanics but not for blacks. Black children are much more likely to live with a nonbiological father, which is associated with fathers taking less responsibility. Hispanic and black children tend to have more siblings, which is associated with greater father responsibility. Controlling for these variables reduces the effect of black race and Hispanic ethnicity. Controlling for economic status significantly (p<.10) reduces the black coefficient and it is longer significant. This is because Black families are characterized by lower fathers earnings and work hours, higher mothers earnings, and by being more likely to have no breadwinner, which are all associated with fathers taking greater responsibility for children. The Hispanic coefficient is not significantly reduced but retains a significant association with paternal responsibility. Mothers wages are lower, leading to less responsibility taken by fathers, but fathers low wages are associated with fathers taking more responsibility. These results are consistent with a division of labor or bargaining explanation for differential involvement. Controlling for neighborhood (column 4) significantly reduces the effect of Hispanic origin. This is because Hispanic children are highly likely to live in Hispanic neighborhoods, and fathers in majority Hispanic areas are more likely to take responsibility for children than families in other neighborhoods. Adjusting for neighborhood reduces the Hispanic effect. Controlling for attitudes further reduces the effect of Hispanic ethnicity. Though they are less likely to have learned to parent from their father, Hispanic fathers are more likely to have had a father involved with them, and that is associated with greater responsibility for children. Additionally, having a positive attitude towards parenting is associated with greater responsibility. Controlling for these Hispanic-white differences reduces the effect of Hispanic ethnicity on responsibility, though it is still statistically significant. Finally, adjusting for gender roles significantly reduces the effect of Hispanic origin on parental responsibility over column 4. Hispanics exhibit more gender-equitable values, which are associated with 23

fathers taking more responsibility for their children.

24

DISCUSSION AND CONCLUSIONS This paper has focused upon black-white and Hispanic-white differences in fathering behavior in two-parent families. Several differences were identified. Black childrens fathers exhibit less engagement and less warmth but exert more control and take more responsibility than white childrens fathers, thus exhibiting a more authoritarian model of parenting. Hispanic childrens fathers exert less control but also take more responsibility than white childrens fathers, while they do not differ from white fathers in warmth. Their parenting style appears more permissive. The results suggest that parenting attitudes are a major factor in race differences in control and responsibility whereas economic factors strongly affect engagement and responsibility. Differences in warmth are due to background differences, neighborhood, and family structure. Demographic and economic factors, particularly family size, relationship to father, family structure, and employment explain most of the fact that black fathers appear to spend less time engaged with their children than do white fathers. In particular, black fathers have larger families, are less likely to be the biological father, and are more likely to be unemployed living with a female breadwinner, all of which, by themselves, would result in less time spent with children. When these factors are controlled, black-white differences in engagement are no longer statistically significant. Differences in warmth between black and white fathers are also partially explained by the fact that fewer black fathers are biological fathers, and nonbiological fathers are less warm. Neighborhood characteristics, however, appear to mask a larger black-white differences in warmth. Living in a good neighborhood is associated with less time engaged with children. Black families tend to live in black neighborhoods and, according to these data, fathers tend to be warmer in such neighborhoods. This fact appears to mask their lower overall levels of warmth. Perhaps living in a racially homogeneous community is less stressful than living in a heterogeneous one (controlling for desirability), and this contributes to greater warmth in the former. Alternatively, black fathers with interpersonal styles that 25

are less warm may select heterogeneous neighborhoods. More research is needed on the interaction between ethnicity and neighborhood in influencing parenting styles. Differences in degree of control and monitoring between blacks and whites are partially explained by family size. Black families tend to be larger and fathers in large families exert more control than those in smaller families. This explains why black childrens fathers are higher on control than white childrens fathers. Hispanic childrens fathers monitor and control their children less than white childrens. This difference does not disappear, but declines as economic conditions are added. Hispanic fathers incomes are lower, and lower incomes are associated with less monitoring, thus explaining some of their lesser monitoring and control. The effects of fathers income are consistent with a resource rather than division of labor model. Finally, black, Hispanic, and other minority fathers all show more responsibility for their children than do white fathers. Black-white differences in responsibility are partially explained by neighborhood factors. In bad neighborhoods, which is where black families are more likely to reside, fathers take more responsibility for their children. Differences in neighborhood composition and in culture between Hispanics and whites also explain much of the variation in responsibility between the two groups of fathers. Fathers take much more responsibility for their children in Hispanic neighborhoods, and (in analyses not shown) this effect is greater for Hispanic fathers. Hispanic fathers also live in less desirable neighborhoods, and living in such neighborhoods is associated with more responsibility. While traditional economic and demographic factors are linked to the attention fathers pay to their children, this study has found evidence that attitudes, values and motivational factors, what we call cultural factors, are closely tied to the father's engagement with children. Fathers with more positive attitudes towards parenting, whose own father was involved with him as a child, who have more gender-equitable responses, and who had taken a parenting class spend more time with the child, exercise more control and take more responsibility for their children. Fathering attitudes and behavior 26

explain significant variation in control for Hispanics, and in responsibility for both blacks and Hispanics. Gender role attitudes explain some of the difference in control between black vs. white and Hispanic vs. white fathers and some of the difference in responsibility for Hispanic fathers compared with white fathers. For example, although they have more gender-equitable attitudes, which are associated with more control, Hispanic fathers are also highly individualistic in attitudes towards marriage and have less positive fathering attitudes, both of which are associated with less control. Adjusting for these factors explains some of the lesser monitoring of Hispanic fathers. The fact that Hispanic fathers are more individualistic is inconsistent with the expected greater familism of Hispanic fathers. However, if responsibility is shared by a variety of extended kin, individual control may not be necessary. Differences in responsibility between Hispanic childrens fathers (and those of other races) and white childrens remain strong even with controls for other factors. Hispanic fathers report their own fathers were more involved in rearing them, and that increases their involvement with their own children. In addition, more gender equitable attitudes, which are held by Hispanic men, are associated with more responsibility towards children. The potential strength of minority families regarding fathering should be examined more thoroughly. Though it examined only a small set of possible parenting behaviors, this paper showed that, of the behaviors we studied, minority fathering in two-parent families varies in only a matter of degree from that of majority families. Black fathers tend to be more authoritarian, with less warmth and greater control while Hispanic fathers are more permissive, with less control than white fathers. The stereotype of a macho, noninvolved minority father is not typically true; black and Hispanic, primarily Mexican, males are as involved as are majority fathers, and they share greater responsibility for childrearing with their partners. Unfortunately, our study does not describe the fathering experiences of all black, white, and Hispanic children. While it included a variety of relationships, our study of children in coresidential two-parent families represents only about half of black, 75 percent of Hispanic, and 80 percent of white 27

children in the United States. Nonresidential fathers are typically less involved. Research which examines the involvement of noncoresidential fathers in the lives of children is an important next step for fully evaluating the strengths and weaknesses of majority and minority families for the United States of the 21st century.

28

References Alwin, D. F. (2001). Parental Values, Beliefs, and Behavior: A Rview and Promulga for Research into the New Century. In T. Owens & S. Hofferth (Eds.), Children at the Millennium: Where did we come from, where are we going? New York: Elsevier Science. Baca Zinn, M. (1994). Adaptation and Continuity in Mexican-Origin Families. In R. L. Taylor (Ed.), Minority Families in the United States: A Multicultural Perspective (pp. 64-81). Englewood Cliffs, NJ: Prentice-Hall. Bandura, A. (1969). Principles of Behavior Modification. New York: Holt, Rinehart and Winston. Bartz, K., & Levine, B. (1978). Child rearing by Black parents: A description and comparison to Anglo and Chicano parents. Journal of Marriage and the Family, 40, 709-720. Baumrind, D. (1968). An Exploratory Study of socialization effects on black children: Blackwhite comparisons. Child Development, 43, 261-267. Baumrind, D . (1978). Parental disciplinary patterns and social competence in children. Youth and Society, 9, 239-276. Baumrind, D. (1991). The influence of parenting style on adolescent competence and substance abuse. Journal of Early Adolescence, 11, 56-94. Baumrind, D., & Black, A. (1967). Socialization practices associated with dimensions of competence in preschool boys and girls. Child Development, 38, 291-327. Becker, G. S. (1991). A treatise on the family (rev. ed.). Cambridge, MA: Harvard University Press. Blau, F., Ferber, M., & Winkler, A. (1998). The Economics of Women, Men, and Work. Upper Saddle River, NJ: Prentice Hall. Bronfenbrenner, U. (1979). The Ecology of Human Development. Cambridge, MA: Harvard University Press. Cherlin, A., & et al. (1991). Longitudinal Studies of the Effects of Divorce on Children in Great Britain and the United States. Science, 252, 1386-1389. Clogg, C. C., Petkova, E., & Haritou, A. (1995). Statistical Methods for Comparing Regression Coefficients between Models. American Journal of Sociology, 100(5), 1261-93. Cooksey, E. C., & Fondell, M. (1996). Spending Time with His Kids: Effects of Family Structure on Fathers' and Children's Lives. Journal of Marriage and the Family, 58, 693-707. Delgado-Gaitan, C. (1994). Socializing Young Children in Mexican-American Families. In P. Greenfield & R. Cocking (Eds.), Cross-Cultural Roots of Minority Child Development: An Intergenerational perspective (pp. 55-86). Hillsdale, NJ: Erlbaum. Duncan, G. J., Brooks-Gunn, J., & Klebanov, P. K. (1994). Economic deprivation and early childhood development. Child Development, 65(2), 296-318. Fitzgerald, J., Gottschalk, P., & Moffitt, R. (1998a). An Analysis of Sample Attrition in Panel Data: The Michigan Panel Study of Income Dynamics. Journal of Human Resources, 33(2), 251-299. Fuller, B., Holloway, S., & Liang, X. (1996). Family Selection of Child-Care Centers: The Influence of Household Support, Ethnicity, and Parental Practices. Child Development, 67, 33203337. Garcia-Coll, C. (1990). Developmental Outcome of Minority Infants: A process-oriented look into our beginnings. Child Development, 61, 270-289. Garcia-Coll, C., Crnic, K., & et al. (1996). An Integrative Model for the Study of Developmental Competencies in Minority Children. Child Development, 67, 1891-1914. Goode, W. J. (1982). The Family. New York: Free Press. 29

Hernandez, D. J. (1993). America's Children: Resources from Family, Government and the Economy. New York: Russell Sage. Hofferth, S. L., Pleck, J., Stueve, J., Bianchi, S., & Sayer, L. (2001). The Demography of Fathers: What Fathers Do. In K. Tamis-Lemonda & N. Cabrera (Eds.), Handbook of Father Involvement . Hofferth, S. (1999, May 27-28). Family Reading to Young Children: Social Desirability and Cultural Biases in Reporting. Paper presented at. Workshop on Measurement of and Research on Time Use, Committee on National Statistics, Washington, DC: National Research Council. Hofferth, S., Davis-Kean, P., Davis, J., & Finkelstein, J. (1999). 1997 User Guide: The Child Development Supplement to the Panel Study of Income Dynamics. Ann Arbor, MI:: Institute for Social Research, The University of Michigan. Jencks, C., & Phillips, M. (1998). The Black-White Test Score Gap. Washington, DC: Brookings Institution. Juster, F., & Stafford, F. P. (1985). Time, Goods, and Well-Being. Ann Arbor, MI: Institute for Social Research. Kohn, M. L. (1977). Class and Conformity. Chicago, IL: University of Chicago Press. Lamb, M. E., Pleck, J. H., Charnov, E. L., & Levine, J. A. (1985). Paternal Behavior in Humans. American Zoologist, 25, 883-894. Lamb, M. (1997). Nonparental Child Care: Context, Quality, Correlates, and Consequences. In I. Sigel & K. Renninger (Eds.), Handbook of Child Psychology (Fifth edition) (pp. 73-134). New York: Wiley. Maccoby, E., & Martin, J. (1983). Socialization in the context of the family: Parent-child interaction. In E. M. Hetherington (Ed.), Handbook of Child Psychology, Vol. IV (pp. 1-101). New York: John Wiley & Sons. Mason, K. O., Czajka, J. L., & Arber, S. (1976). Changes in U.S. Women's Sex-Role Attitudes, 1964-1974. American Sociological Review, 41, 573-596. McAdoo, J. L. (1988). The Roles of Black Fathers in the Socialization of Black ChildrenH. P. McAdoo (Ed.), (pp. 257-269). Newbury Park, CA: Sage. McLanahan, S., & Sandefur, G. (1994). Growing Up with a Single Parent: What Hurts, What Helps. Cambridge, MA: Harvard University. Mott, F. L. (1990). When is a Father really Gone? Paternal Child Contact in Father-Absent Homes. Demography, 27, 499-517. Ogbu, J. (1981). Origins of Human Competence: A cultural-ecological perspective. Child Development, 52, 413-429. Palkowitz, R. (1984). Parental attitudes and father's interactions with their 5-month-old infants. Developmental Psychology, 20, 1054-1060. Phillips, M., Brooks-Gunn, J., Duncan, G., Klebanov, P., & Crane, J. (1998). Family Background, Parenting Practices, and the Black-White Test Score Gap. In C. Jencks & M. Phillips (Eds.), The Black-White Test Score Gap (pp. 103-145). Washington, DC: Brookings Institution. Pleck, J. H. (1997). Paternal Involvement: Levels, Sources, and Consequences. In M. E. Lamb (Ed.), The Role of the Father in Child Development (pp. 66-103). New York: John Wiley & Sons. Sampson, R. J., Raudenbush, S. W., & Earls, F. (1997, 15 August). Neighborhoods and violent crime: A multilevel study of collective efficacy. Science, 277, 918-924. Sandberg, J. F. (2000, March). Modeling Multi-dimensionality of Involvement in Two Parent Families: A Latent Class Analysis of Paternal Responsibility. Paper presented at. Annual Meeting of the Population Association of America, Los Angeles, CA. 30

Seltzer, J., & Bianchi, S. (1988). Children's Contact with Absent Parents. Journal of Marriage and the Family, 50, 663-677. Taylor, R. L. (1994a). Minority Families in America: An Introduction. In R. L. Taylor (Ed.), Minority Families in the United States: A Multicultural Perspective (pp. 1-16). Englewood Cliffs, NJ: Prentice Hall. Taylor, R. L. (1994b). Black American Families. In R. L. Taylor (Ed.), Minority Families in the United States: A Multicultural Perspective (pp. 19-46). Englewood Cliffs, NJ: Prentice Hall. Taylor, R. L. (1994c). Minority Families and Social Change. In R. L. Taylor (Ed.), Minority Families in the United States: A Multicultural Perspective (pp. 204-248). Englewood Cliffs, NJ: Prentice Hall. Vega, W. A. (1990). Hispanic Families. Journal of Marriage and the Family, 52, 10151024. Yeung, W. J., Sandberg, J., Davis-Kean, P. E., & Hofferth, S. L. (2001). Children's Time with Fathers in Intact Families. Journal of Marriage and the Family, 63(1), 136-154. Zajonc, R. B., & Markus, G. B. (1975). Birth order and intellectual development. Psychological Review, 82, 74-88.

31

Table 1. Means on Fathering, Gender Role Attitudes, Economic Status and Neighborhood, by Race/Ethnicity All Races Variable Engaged? Hours Engaged Time Eating Meals Time Reading Responsibility for Child Warmth Control Traditional Marriage Traditional Mothering Equity Individualism Grandfather Involved No Parenting Class Learned from Mother Learned from Father Positive Attitude Mother's Work Hours Father's Work Hours Father's Education (years of school) Mother's Earnings (in 0000s) Father's Earnings (in 0000s) Male Breadwinner, Female Homemaker Dual Earners Female Breadwinner, Male Unemployed No Breadwinner Good Neighborhood to Raise Children Majority Hispanic Majority Black Majority White Number of Cases Note: Significance tested relative to whites +p<.10, *p<.05, **p<.01, ***p<.001 Mean 0.94 14.93 4.79 0.19 6.03 5.02 31.87 0.03 0.08 0.03 -0.08 3.11 0.73 0.70 0.58 26.01 24.98 44.35 13.21 1.30 3.67 0.34 0.55 0.03 0.03 0.31 0.13 0.13 0.80 White Black Hispanic Other

Std. Dev. Mean Std. Dev. 0.25 0.95 0.23 10.93 15.35 11.47 4.62 4.74 4.67 0.72 0.22 0.81 2.49 5.59 2.05 1.46 5.10 1.43 5.60 31.96 5.53 1.00 -0.11 0.98 1.02 0.07 1.11 0.99 0.01 0.98 0.98 -0.22 0.93 0.89 3.06 0.91 0.44 0.73 0.47 0.46 0.72 0.47 0.49 0.63 0.51 2.90 26.19 3.06 18.87 25.04 19.05 12.87 44.45 11.95 3.04 13.88 2.30 1.67 1.40 1.78 3.52 4.12 3.94 0.47 0.31 0.48 0.50 0.61 0.51 0.18 0.02 0.15 0.17 0.02 0.13 0.46 0.36 0.50 0.33 0.03 0.19 0.34 0.06 0.25 0.40 0.92 0.28 1172 819

St. Dev. Mean Std. Dev. Mean Std. Dev. Mean 0.86 *** 0.23 0.91 0.33 0.92 0.32 12.76 ** 7.49 14.16 12.20 13.62 11.32 4.13 + 3.29 5.26 6.05 5.42 5.72 0.12 + 0.45 0.10 0.47 0.05 0.36 6.27 *** 1.70 8.03 *** 4.27 7.86 *** 3.88 4.40 *** 1.24 4.84 + 1.96 4.98 1.62 34.08 *** 3.70 29.48 *** 7.55 32.52 7.28 0.24 *** 0.59 0.51 *** 1.28 0.81 *** 1.35 -0.32 *** 0.51 0.25 + 0.88 0.42 * 1.02 -0.04 0.63 0.27 * 1.41 0.03 1.39 0.12 *** 0.66 0.60 *** 1.33 0.20 ** 1.23 3.14 0.61 3.45 *** 1.00 3.08 1.23 0.83 *** 0.24 0.75 0.52 0.60 + 0.57 0.68 0.30 0.71 0.54 0.53 ** 0.57 0.45 *** 0.32 0.38 *** 0.58 0.44 ** 0.57 25.73 * 1.74 25.32 ** 3.24 25.39 + 3.13 31.52 *** 12.94 24.66 24.35 14.30 *** 21.83 36.47 *** 11.14 47.21 * 16.03 48.02 * 20.90 13.06 *** 1.34 9.22 *** 4.97 11.86 *** 5.42 1.71 * 1.38 0.54 *** 0.94 0.66 ** 1.15 2.01 *** 1.14 1.79 *** 1.24 3.50 3.02 0.20 ** 0.26 0.51 *** 0.60 0.58 *** 0.56 0.48 *** 0.32 0.33 *** 0.56 0.20 *** 0.46 0.15 *** 0.23 0.04 0.23 0.02 0.14 0.06 *** 0.15 0.06 ** 0.28 0.15 *** 0.41 0.19 *** 0.25 0.09 *** 0.34 0.20 * 0.46 0.15 *** 0.23 0.71 *** 0.54 0.17 * 0.42 0.78 *** 0.27 0.17 0.44 0.10 0.34 0.44 *** 0.32 0.31 *** 0.55 0.62 *** 0.56 213 94 46

Table 2. OLS Regressions of Parenting Behavior on Race/Ethnicity and Controls (continued) Panel A Variables 1 2 3 Engagement 4 5 6 1 2 3 Panel B Warmth 4 5 6

White (omitted) Black Hispanic Other Race Age of Child Gender of Child Age of Father Nonbiological Father Number of Children Father's Education Mother's Hours Father's Hours Male Breadwinner (omitted) Two-Earner Family Female Breadwinner No Breadwinner Mother's Earnings Father's Earnings Good Neighborhood Majority Hispanic Majority Black Grandfather Involved No Parenting Class Learned From Mother Learned From Father Positive Attitude Traditional Marriage Traditional Mothering Equity Individualism Intercept R-squared Number of Cases +p<.10, *p<.05, **p<.01, ***p<.001

-2.45 + -0.56 -2.54 +

-1.31 -0.50 -2.85 -0.59 0.15 -0.11 -5.59 -0.76

+ *** + *** ***

-2.02 0.14 -1.87 -0.58 0.21 -0.11 -5.69 -0.57 0.19 0.04 -0.06

*** * *** + + *

-2.00 1.12 -1.87 -0.57 0.26 -0.12 -5.77 -0.55 0.22 0.05 -0.07

*** * *** + + *

-2.14 0.68 -1.72 -0.52 0.15 -0.11 -5.36 -0.59 0.11 0.05 -0.07 -0.45 -4.06 -0.55 0.16 -0.22 -1.53 -1.80 -0.12 0.68 -0.48 0.42 -0.22 0.36

*** * *** * + *

-1.88 1.24 -1.52 -0.55 0.20 -0.11 -5.34 -0.64 0.12 0.04 -0.06

-0.86 *** 0.00 -0.10 *** + *** *

-0.71 *** 0.01 -0.19 -0.14 *** -0.05 0.00 -0.75 *** -0.05

-0.65 *** 0.18 -0.07 -0.14 *** -0.05 -0.01 -0.68 *** -0.02 0.01 0.00 -0.01 0.02 -0.18 -0.38 0.06 + 0.03 *

-0.87 *** 0.06 -0.08 -0.14 *** -0.05 -0.01 -0.69 *** -0.02 0.01 0.00 -0.01 0.04 -0.15 -0.36 0.06 + 0.03 * 0.14 0.21 0.34 +

-0.91 -0.01 -0.02 -0.13 -0.08 0.00 -0.58 -0.03 -0.02 0.00 -0.01 0.06 -0.11 -0.19 0.04 0.04 0.14 0.19 0.43 0.14 -0.13 0.14 0.04 0.10

***

***

**

-0.87 *** 0.00 -0.08 -0.13 *** -0.08 0.00 -0.57 ** -0.04 -0.01 0.00 0.00 0.12 -0.10 -0.16 0.06 + 0.03 * 0.09 0.16 0.38 *

-0.40 -3.83 + -1.10 0.17 -0.25 *

-0.49 -4.18 * -1.11 0.21 -0.25 * -1.59 * -1.78 -0.27

* *

-0.37 -4.00 + -0.30 0.20 -0.25 * -1.71 * -1.91 -0.24

**

* **

** -0.27 -0.15 0.69 * -0.51 26.20 0.13 1027

*** -0.01 -0.02 0.33 *** -0.07 6.22 0.23 1077

15.50 0.01 1027

25.14 0.09 1027

24.63 0.11 1027

25.18 0.12 1027

15.16 0.13 1027

5.10 0.02 1077

6.09 0.17 1077

6.22 0.18 1077

6.10 0.18 1077

3.31 0.24 1077

Table 2. OLS Regressions of Parenting Behavior on Race/Ethnicity and Controls (concluded) Panel C Variables 1 2 3 Control 4 5 6 1 2 Panel D Responsibility 3 4 5 6

White (omitted) Black Hispanic Other Race Age of Child Gender of Child Age of Father Nonbiological Father Number of Children Father's Education Mother's Hours Father's Hours Male Breadwinner (omitted) Two-Earner Family Female Breadwinner No Breadwinner Mother's Earnings Father's Earnings Good Neighborhood Majority Hispanic Majority Black Grandfather Involved No Parenting Class Learned From Mother Learned From Father Positive Attitude Traditional Marriage Traditional Mothering Equity Individualism Intercept R-squared Number of Cases +p<.10, *p<.05, **p<.01, ***p<.001

2.39 *** -2.39 *** 0.92

1.97 ** -3.06 *** 0.52 -0.02 0.42 -0.03 -0.23 0.56 ***

2.21 -2.21 1.13 -0.01 0.44 -0.07 0.24 0.63 0.11 -0.02 -0.02

** **

* ***

1.87 -3.09 1.02 -0.01 0.43 -0.07 0.21 0.65 0.11 -0.02 -0.02

* ***

* ***

1.77 -3.56 1.16 0.04 0.29 -0.06 0.59 0.59 0.01 -0.02 -0.01 -0.35 0.62 -3.44 0.03 0.12 0.50 1.34 0.63 0.53 -0.73 0.80 -1.09 0.41

* ***

+ ***

2.24 -2.53 1.53 0.01 0.28 -0.05 0.92 0.44 -0.04 -0.03 -0.01

* ** +

0.65 * 2.68 *** 2.12 ***

+ ** *

0.60 2.43 2.00 -0.01 -0.21 -0.04 -0.67 0.11

* *** ***

** * +

0.45 2.34 1.66 0.00 -0.17 -0.04 -0.95 0.12 -0.01 -0.01 -0.01 -0.41 -1.55 1.84 0.25 -0.05

*** ***

** ** +

* * *** *** *** *

0.07 1.72 1.55 0.00 -0.15 -0.04 -0.98 0.14 0.00 -0.01 -0.01 -0.36 -1.53 1.89 0.27 -0.05 -0.25 0.85 0.45

*** ***

** ** *

* + *** *** *** * + **

-0.09 1.41 1.53 0.02 -0.23 -0.03 -0.90 0.14 -0.02 -0.01 -0.01 -0.33 -1.57 2.09 0.24 -0.03 -0.23 0.87 0.49 0.37 -0.06 0.27 -0.78 0.13

*** *** + ** ** *

0.08 1.61 1.54 0.01 -0.19 -0.03 -0.90 0.13 -0.02 -0.01 -0.01 -0.27 -1.54 2.05 0.27 -0.05 -0.29 0.81 0.49

*** ***

* ** *

-0.52 0.48 -4.24 *** 0.09 0.10 +

-0.44 0.62 -4.16 *** 0.10 0.10 + 0.40 1.35 0.49

** * + * + + * ***

-0.04 1.23 -2.81 ** 0.12 0.10 + 0.13 0.92 0.77

+ *** *** ***

** + ***

*** *** *** * + **

*** *** -0.04 0.03 0.33 *** 0.00 7.50 0.23 1077

32.04 0.03 937

31.92 0.05 937

32.60 0.08 937

32.09 0.08 937

20.92 0.14 937

-0.13 0.08 1.16 *** -1.03 *** 33.46 0.15 937

5.54 0.13 1077

6.83 0.14 1077

7.82 0.20 1077

7.49 0.21 1077

3.18 0.26 1077

Table 3. Changes in Race/Ethnic Coefficients Across Models from Table 2, with Significance Tests
Background Earnings Neighborhood Fathering Gender Roles

1-2 Engagement Black d S(d)(a) t Hispanic d S(d) t Warmth

Black

2-3

3-4

4-5

4-6

-1.137 0.148 7.68 -- (b) ---

0.711 0.281 2.53 ----

-0.028 0.997 N/S ----

0.141 0.02 7.04 ----

-0.114 0.154 N/S ----

d S(d) t Hispanic d S(d) t Control Black d S(d) t Hispanic d S(d) t Responsibility Black d S(d) t Hispanic d S(d) t
(a) (b)

-0.154 0.036 4.28 ----

-0.60 0.048 1.23 ----

0.221 0.122 1.81 ----

0.04134 0.033 1.24 ----

0.005 0.030 N/S ----

-0.426 0.141 3.02 0.668 0.205 3.25

0.242 0.165

-0.336 0.519

1.47
-0.849 0.338 2.51

N/S
0.880 0.578 1.52

-0.107 0.130 N/S

0.476 0.062 7.67

0.367 0.145 2.53 -0.554 0.063 8.79

-0.056 0.043 1.30

-0.151 0.135 1.12 -0.089

-0.377 0.202 1.86 -0.613 0.199 3.11

-0.162 0.051 3.18 -0.314 0.022 14.27

0.015 0.014 N/S -0.114 0.053 2.15

-0.252 0.071 3.55

0.115 N/S

s(d) = square root [(se2)2-(se1)2 (Root MSE2/Root MSE1)] = not calculated because of nonsignificance of the coefficient

Appendix Table 1. Background Characteristics of Fathers in Two-Parent Families, by Race/Ethnicity Total Sample Variable White Black Hispanic Other Race Age of Child Sex of Child (1=female) Mother's Age Father's Age Nonbiological Father Number of Children Number of Cases Mean Std. Dev. 0.76 0.43 0.08 0.27 0.11 0.32 0.05 0.22 5.87 3.72 0.52 0.50 34.77 6.57 36.84 7.12 0.07 0.26 2.34 1.22 1229 Whites Mean Std. Dev. 1.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 5.87 3.85 0.52 0.52 34.97 7.01 37.26 7.50 0.65 0.26 2.19 0.96 854 Blacks Mean Std. Dev. 0.00 0.00 1.00 0.00 0.00 0.00 0.00 0.00 6.38 2.40 0.43 0.32 35.20 4.59 38.10 5.11 0.14 0.22 2.85 1.35 230 Hispanics Mean Std. Dev. 0.00 0.00 0.00 0.00 1.00 0.00 0.00 0.00 5.93 4.68 0.52 0.60 32.92 6.28 33.60 6.56 0.07 0.30 2.88 1.82 98 Other Mean Std. Dev. 0.00 0.00 0.00 0.00 0.00 0.00 1.00 0.00 4.85 4.09 0.59 0.56 35.32 6.46 35.91 6.71 0.04 0.21 2.69 2.07 47

Availability 4

1.470 2.053 + -1.073

1.264 2.233 + -1.025

0.306 2.682 + -1.625

0.384 1.751 -1.730

0.358 1.540 -2.281

-2.322 * -1.491 -1.512 -1.151

-2.691 * -1.876 + -2.208 * -0.991 0.064 -0.103 6.110 -5.816 + 0.028

-2.790 ** -1.973 + -2.144 + -0.900 0.044 -1.348 5.919 -5.764 + 0.023 0.013 -0.035 -0.032 -0.022 3.689 6.715 ** 0.923 ** -0.017

-2.711 ** -1.843 + -1.904 + -0.824 0.041 -1.460 5.765 -6.196 + -0.030 0.036 -0.035 -0.031 -0.083 2.779 6.857 ** 0.987 *** -0.013 -1.222 1.173 -0.351 -3.141

-2.637 * -1.715 -1.765 -0.824 0.037 -1.386 6.179 -6.585 * -0.043 0.055 -0.028 -0.031 -0.321 2.533 6.713 ** 0.994 *** 0.002 -1.137 1.009 -0.607 -3.371 + 0.460 -1.073 -1.309 -0.658 -0.078

14.925 *** 0.013

12.813 *** 0.019

14.266 *** 0.048

17.452 *** 0.054

20.216 *** 0.060

1027

1027

1027

1027

1027

0.243 1.613 -1.236

-2.581 * -1.706 -1.730 -0.802 0.044 -1.473 5.408 -6.092 + 0.003 0.008 -0.042 -0.034 -0.220 2.569 7.004 ** 0.902 ** -0.025 -1.057 1.322 -0.345 -3.525 +

-0.610 -0.589 0.065 0.185 18.393 *** 0.059

1027