D. Wendy Manning, Susan D. Stewart, and Pamela J. Smock
The Complexity of Fathers' Parenting Responsibilities and Involvement with Nonresident Children
PSC Research Report
Report No. 01-478
June 2001
PSC
P OPULATION S TUDIES C ENTER
AT T H E
I NSTITUTE
FOR
S OCIAL R ESEARCH
OF
U NIVERSITY
M ICHIGAN
The Population Studies Center (PSC) at the University of Michigan is one of the oldest population centers in the United States. Established in 1961 with a grant from the Ford Foundation, the Center has a rich history as the main workplace for an interdisciplinary community of scholars in the field of population studies. Today the Center is supported by a Population Research Center Core Grant from the National Institute of Child Health and Human Development (NICHD) as well as by the University of Michigan, the National Institute on Aging, the Hewlett Foundation, and the Mellon Foundation. PSC Research Reports are prepublication working papers that report on current demographic research conducted by PSC associates and affiliates. These papers are written for timely dissemination and are often later submitted for publication in scholarly journals. The PSC Research Report Series was begun in 1981. Copyrights for all Reports are held by the authors. Readers may quote from, copy, and distribute this work as long as the copyright holder and PSC are properly acknowledged and the original work is not altered.
PSC Publications http://www.psc.isr.umich.edu/pubs/
Population Studies Center, University of Michigan PO Box 1248, Ann Arbor, MI 48106-1248 USA
The Complexity of Fathers Parenting Responsibilities and Involvement with Nonresident Children
Wendy D. Manninga Susan D. Stewarta Pamela J. Smockb
Department of Sociology Bowling Green State University Bowling Green, OH 43403 [e-mail] [e-mail] 419-372-2850
a
b
Department of Sociology and Population Studies Center The University of Michigan 426 Thompson Street P.O. Box 1248 Ann Arbor, MI 48106-1248 [e-mail]
This research was supported by grants HD34391 and HD34392 from the National Institute of Child Health and Human Development. Computer resources were provided by Bowling Green State University.
The Complexity of Fathers Parenting Responsibilities and Involvement with Nonresident Children
ABSTRACT Most policies that legislate father involvement with nonresident children treat men as if they have obligations to only one set of children. This paper describes the complexity of nonresident fathers parenting circumstances, and assesses whether and how parenting configurations are associated with their involvement with nonresident children. We find that nonresident fathers often have parenting obligations within and outside their current residences, and that the complexity of these obligations may result in less economic support and visitation with nonresident children. Our results suggest that new policy efforts need to recognize the complexity of nonresident fathers family ties.
Families in the United States are continuing to undergo rapid transformation, resulting in increasingly complex family relationships. Changes in nonmarital childbearing, marriage, divorce, cohabitation, and remarriage (e.g., Bumpass, Sweet, and Cherlin 1991; Bumpass and Raley 1995; Cherlin 1992; McLanahan and Casper 1995) create family ties that often extend across households. Most divorced men and women remarry or cohabit, and many go on to have new biological children (Bumpass et al. 1991; Sorensen 1997; Manning and Smock, 1998), leading to potentially intricate webs of parenting and family obligations across households. However, these shifts to more complicated family structures have not been fully incorporated into empirical research on child support and father involvement with nonresidential children. The child support literature generally treats nonresident fathers as individual actors, ignoring the complexity of their parenting circumstances. Even child support policies are often based on the premise that men have only one set of children to support, and some state-level child support policies implicitly favor one set of offspring (resident versus nonresident) over another (Takas 1991). Drawing on the National Survey of Families and Households (NSFH), this paper employs a family systems approach to assess the implications of nonresident fathers parenting within and across households. We have two central goals. First, we provide a descriptive portrait of nonresident fathers families that encompasses the full array of parenting obligations of both the father and his new spouse or partner. Second, we evaluate the associations between these parenting configurations and nonresident fathers financial and social involvement with nonresident children, an issue with critical importance for understanding nonresident father-child involvement. BACKGROUND Past research has identified factors associated with nonresident fathers child support payments and their social involvement with biological children (e.g., Seltzer 1994; Sorenson 1997; Teachman 1991; Veum 1993; Zill 1996). However, few studies have directly investigated the effects of competing parenting responsibilities. Some studies include an indicator for whether the father lives with new children, but do not specify the biological relationship between the father and children (Cooksey and Craig 1998; Hill 1992; Smock and Manning 1997; Veum 1993). Even those studies that are specifically interested in nonresident fathers parenting, have not accounted the possibility that some men may have nonresident children from several prior relationships or more than one set of nonresident children (e.g., Cooskey and Craig 1998; Manning and Smock 1999, 2000). Thus, this paper provides the most comprehensive examination of parenting by including the following categories of parenthood: nonresident children born to more than one mother, resident children from prior relationship, resident children from current relationship, resident step-children, and nonresident step-children. Moreover, this paper moves beyond prior work by examining nonresident fathers social and economic ties to children.
1
This gap in knowledge is problematic because it precludes recognition of potentially critical influences on nonresident father-child relationships as well as incorporation of these influences into child support policy. Three-quarters of divorced men eventually remarry, with 70% forming new unions (either remarriages or cohabitations) within five years of marital separation (Bumpass et al. 1991; Sweet and Bumpass 1987). Estimates also suggest that about two-fifths of nonresident fathers live with new children (Manning and Smock 1999; Sorensen 1997). Moreover, the handful of recent studies utilizing more complete information on nonresident fathers parenting responsibilities provides suggestive evidence that both child support and visitation are reduced when fathers reside with new biological children (Cooksey and Craig 1998; Manning and Smock 1999, 2000). Most past research on nonresident father-child involvement also shares the underlying assumption that nonresident fathers are individual actors when, in fact, nonresident fathers may be incorporating the well-being of new partners and/or children when making decisions about child support payments and visitation. Indeed, there is some indirect evidence that this is the case. Drawing on a sample a sample of Missouri residents Ganong and Coleman (1999) report that the majority believe that divorced fathers financial responsibility toward nonresidential children depends on remarriage, that men who marry women with children should support those children, and that financial responsibility for children should, to some extent, be based on whether men share a residence with them. We thus employ a family systems approach to investigate how nonresident fathers parenting obligations is associated with involvement with nonresident children. This approach assumes that a family is a system of relationships that subsumes any and all varieties of living arrangements, household patterns, legal and residential structures, and so on, as well as the ongoing decision-making processes that occur within these patterns and arrangements (Scanzoni et al. 1989: 52). Central to this perspective is recognition that adults can and do make commitments -- and thus be family -- to children who are not their biological offspring. Commitment to family members is characterized by univocal reciprocity, a term that refers to exchanges based on a sense of duty or obligation, where immediate or direct reciprocation is not expected or required (Scanzoni and Marsiglio 1991). While an example of this type of commitment is that of a biological parent to his or her children, univocal reciprocity is not limited to genetic ties; relationships between children and unrelated adults (i.e., social parents) formed as a result of remarriage, cohabitation, or nonmarital childbearing may take on varying degrees of univocality as well. CURRENT INVESTIGATION Our work extends knowledge about nonresident fathers by incorporating a fuller definition of nonresident fathers parenting responsibilities. Like a few prior studies, we identify whether nonresident fathers have entered a new union and whether his new coresident children are biological or step-children (Cooksey and Craig 1998; Manning and Smock 1999, 2000). But we also include new measures that capture potentially important aspects of nonresident fathers circumstances. The first is whether nonresident fathers have more than one set of nonresident offspring, acknowledging that some men have fathered children with more than one mother, and 2
thus may have more than one set of nonresident children to support and to visit. To our knowledge, a similar measure has not been included in any prior study. Second, we take account of what might be called nonresident stepchildren. These are children of the current spouse or partner who reside elsewhere. Manning and Smock (1999,2000) did not include this category of parenting in their work on changes in child support and visitation. Using the second wave of the NFSH and children as the unit of analysis, Cooksey and Craig (1998) report that children living with stepmothers who are nonresident parents are as likely to visit their nonresident fathers monthly as other children and spoke to their fathers less frequently. Their analytic sample includes fathers not living in a union which may confound the effect of nonresident stepchildren - only those fathers in a union can have nonresident or resident stepchildren. Their work focuses on social ties and does not address economic ties to children. Third, we take account of biological children who are currently living with the nonresident father, but are unrelated to his current spouse or partner. This measure recognizes that not all resident biological children are offspring of the couple. Cooksey and Craig (1998) did not distinguish between those resident children who were biologically related to only the father and those related to both the father and his spouse/partner. Manning and Smock (1999, 2000) model change in nonresident fathersinvolvement and not levels of involvement suggesting that the effects of parenting configurations may differ somewhat depending on the conceputalization of involvement. Both sets of authors include men not in unions in their analysis which may influence how the new biological child coefficient is interpreted, (men not living in unions by definition cannot have a new biological child with their current spouse or partner). Our analysis consists of two parts. First, we describe the extent to which nonresident fathers experience these various parenting roles, and summarize this through descriptive statistics. We start with presenting the proportion of nonresident fathers with simple versus complex parenting roles. By simple, we mean that the nonresident father has only one set of nonresident children and no other parenting obligations either within or outside his current household and complex are those in all other situations. This measure is useful for a basic understanding of parenting for our sample. Our second group of measures specifically describes the parenting obligations of those with complex roles. These measures are restricted to those nonresident fathers who have formed new unions. These fathers face the most potentially complex parenting. The following categories are considered: more than one set of nonresident children, resident biological child from a prior union, resident biological child from current union, resident step-child, and nonresident step-child. The final categorization essentially counts the number of parenting situations. This measure provides a sense of the accumulation of paternal responsibility to children. Second, we evaluate how simple versus complex configurations are related to whether child support is paid and frequency of visits to nonresident children. We consider the specific types of complexity as well as the accumulation of complexity. We expect that nonresident fathers with complex parenting obligations will be less likely to be involved with their nonresident children than fathers who have obligations to only one set of nonresident children. Nonresident fathers with the greatest accumulation of responsibilities are expected to visit their nonresident children less often.
3
We expect that the effects of types of parenting circumstances may operate in two ways: biological relationship and residence. Based on a biological imperative model nonresident fathers own fertility is expected to have a greater impact on their involvement with nonresident children than their spouse/partners prior fertility. In other words the resident or nonresident stepchildren will have a weaker effect than the fatherss own biological offspring on their ties to nonresident children. Specifically, the new offspring are expected to have the greatest impact because they are biologically related to both members of the couple. The residence of the children may also be an important factor predicting involvement. Nonresident fathers resident children (new biological, biological from prior relationship, and resident stepchildren) may have the most negative influence on ties to nonresident children because their residence makes their needs and requests for support and attention immediate. Taking these factors together suggests that nonresident fathers nonresident stepchildren will have the weakest influence on social and economic ties to children and the resident biological children belonging to the couple will have the greatest impact on father involvement. There are several possible underlying factors which are unmeasured and/or causally ambiguous. It may be that families with parenting obligations extending across at least two households simply have less time and less money because of greater demands on those resources. It may also be that the needs of the resident family receive priority. While we attempt to take account of economic constraints by controlling for the income of the nonresident fathers as well as the income of his spouse or partner, we may still find that nonresident fathers in complex families are less involved with their children. This may be because the emotional resources required to maintain contact are too high, and that a complex web of family responsibilities creates uncertainty about obligations to family members. In addition, the role expectations for nonresident fathers are generally ambiguous, and nonresident fathers feel they are less competent as parents and are less satisfied with their role than resident fathers (Minton and Pasley 1996). Moreover, fathers inundated with complex parenting demands may feel it is easiest to sacrifice time spent with nonresident children (Arditti 1995). Selection may be operating such that men who take on new parenting roles do not have the capacity or interest in being involved with and financially supporting their nonresident children. While we think it highly implausible that this selection could entirely account for any association, it remains possible. DATA AND METHODS Data We use the first wave of the National Survey of Families and Households (NSFH). The NSFH was collected in 1987 and 1988, and is a national probability sample of over 13,000 individuals with a 74% response rate (Sweet et al. 1989). The NSFH offers many advantages over other data sources. Its primary advantage is that it is the only nationally representative survey that includes information about the spouses or partners children outside of the household, allowing us to construct a portrait of family ties that extends beyond household boundaries. In addition, unlike other data sources, the NSFH includes data on payment of child support and visitation for both the main respondent and the spouse or partner of the main respondent. This allows us to maintain a large enough sample because we can identify families 4
that include nonresident fathers who are not the main respondent. Finally, the NSFH includes measures commonly used to predict the frequency of visitation and child support payments (i.e., sociodemographic characteristics of both the father and the child.) Our analytical sample consists of families in which one member is a nonresident father i.e., either the main respondent or a male spouse/partner has a nonresident child less than age 18. In other words, all of the families included in our analytical sample contain a father with a nonresident biological child. The NSFH includes 850 such families. Of these, we limit the sample to nonresident fathers who have children living with their biological mother (N=759) and those with complete focal child, child support, and visitation information. We limit our analyses to fathers of children living with their biological mother because fathers child support obligations and visitation rights may differ when children live with other relatives or foster families. Also not all questions are asked about children who do not live with mothers. In the NSFH questions about child support and visitation refer to one randomly selected nonresident child or focal child. Our final sample consists of 649 families. Men who have formed unions have potentially more complex parenting obligations so we conduct separate analyses on the 424 nonresident fathers who are coupled, cohabiting or married. Dependent Variables We use the following two measures of father-child involvement as our dependent variables: frequency of visitation and whether any economic support was paid in the last year. These questions are asked of a focal child; this refers to a randomly-selected child from among all nonresident children. If the main respondent is female, then we ascertain this information from a separate spouse/partner interview. In cases of missing data in that interview, we use the main respondents proxy report of her spouse or partners involvement. We had to rely on proxy reports from the main respondent when the spouse was missing data for only 3 cases. Frequency of visitation contains six categories: none, about once a year, several times per year, one to three times per month, about once a week, several times a week. Table 1 shows the distribution of this variable, and the median level of visitation is several times per year. Child support is a simple dichotomous variable indicating whether any economic contribution (formal or informal payments) was made to the focal childs household; data limitations prevent us from using a measure indicating the amount of child support paid. Table 1 shows that, on average, fathers pay child support, but a substantial minority (22%) make no financial payments to their focal child. Independent Variables We present the distribution of our variables for the total sample and for the coupled sample because we estimate models for both samples. Note that some variables only apply to the coupled sample, for example partners fertility. Our core independent variables measure nonresident fathers parenting circumstances. Recognizing that fathers can have more than one set of nonresident children, we first construct a dichotomous variable differentiating between fathers with one and those with more than one set of nonresident children. Table 1 shows that eight percent of nonresident fathers have more than one set of nonresident children. This estimate is probably conservative because some men might be hesitant to admit to the number of 5
children they fathered or be unaware of some biological offspring. A second variable indicates whether the nonresident father is currently living with a child (or children) he fathered prior to his current union (if he is in one). Approximately 8% of nonresident fathers report that they are doing so. We also include two measures tapping the prior fertility of the fathers spouse or partner. One indicates whether a nonresident father is residing with step-children (i.e., his spouse or partners biological children from a prior union). Table 1 shows that roughly one-quarter (26%) of nonresident fathers live with step-children. Another measures indicates whether the spouse or partner is also a nonresident parent (i.e., she has children who live in another residence). About 14% of nonresident fathers are living with someone who is also a nonresident parent. Our next measure of parenting complexity indicates whether or not nonresident fathers have biological children with their current spouse or partner; Table 1 shows that about 42% do. We combined these five measures of parenting circumstances to create two general measures of complexity. First, we created a simple dichotomous measure indicating whether they was any complexity. Nonresident fathers are classified as simple if they only have one set of nonresident children and complex if they are connected to more than just one group of nonresident children. Half of all nonresident fathers reported having complex parenting circumstances and three-quarters of nonresident fathers who were married or cohabiting had complex parenting roles. The second measure refers only to nonresident fathers who were cohabiting or married because single fathers can only have two levels of complexity based on their prior fertility (more than one set of nonresident children and resident biological children from a prior union). We measure the level of complexity by counting the different types of children nonresident fathers have. For example, a value of two shows that a nonresident fathers has two types of complexity - perhaps more than one set of nonresident children and a new biological child. This measure provides an indication of the accumulation of potential fathering responsibility. Almost half of nonresident fathers have just one other type of parenting configuration but one-quarter have two or more different types of children in their lives. The other independent variables act as control variables, and are commonly used in research on nonresident father involvement.1 Many of these variables are related to whether nonresident fathers from new unions and have subsequent children (e.g., Clarkberg et al. 1995; Nock 1998; Sweeney 1997). Table 1 presents their distributions for the total sample and the coupled sample. Our discussion below refers to the distributions for the total sample. First, we include several characteristics of the focal child: age, sex, distance from the fathers residence, and number of siblings. The childs age is coded as a continuous variable, with the mean in this sample being 10 years old. Past research has yielded mixed results about the relationship between childrens ages and father involvement (Furstenberg and Harris 1992; Maccoby and Mnookin 1992; Seltzer 1991; Veum 1993). The relationship between gender and involvement is also unclear (e.g., Maccoby and Mnookin 1992; Paasch and Teachman 1991; Seltzer and Bianchi 1988; Seltzer 1991), but we include it as a control variable. The number of miles between the childs and fathers residence is coded as a continuous variable. We tested for nonlinearity with a Unfortunately the spouse and partner data prevent us from obtaining information about their marital status at the time of the childs birth. 6
1
squared term but found the effect was linear. The average distance between father and focal child is 384.5 miles, although the median value is a substantially lower 35 miles. Past research indicates that greater physical distance is linked to lower child support payments and less frequent visitation (Braver et al. 1993; Maccoby and Mnookin 1992; Seltzer 1991; Sonenstein and Calhoun 1990; Veum 1993). Finally, we include the number of full siblings less than age 18 in the focal childs home because some research indicates that fathers visit more often when there are more children to visit (Seltzer and Bianchi 1988; Veum 1993). The mean number of siblings is 0.5, with values ranging from 0 to 4. Several sociodemographic characteristics of the father are also included. The fathers age is measured in years with the mean age of 36 years old. Race is divided into three categories (other race-ethnicity, white, Black) because sample size doesnt permit further distinctions. Some prior work shows that Black nonresident fathers visit more frequently than white fathers (Mott 1990), but other studies indicate that there is no association between race and visitation (Cooksey and Craig 1998). The majority (70%) of the sample is white, 20% is Black, and 10% of the sample belongs to another other race or ethnic group. We code fathers completed education into the following categories: less than 12 years, 12 years, 13-15 years, and 16 or more years of schooling. Education tends to be positively associated with paying child support (Braver et al. 1991; Veum 1993) and visitation (Maccoby and Mnookin 1992; Seltzer et al. 1989). The mean level of education in the sample is 12 years. Fathers income is represented by a logged value. We code cases with missing income data to the mean value of $25,275. There were 45 cases missing on fathers earnings (7% of the total sample). In addition, we include the income of his spouse or partner (if any), a variable rarely included in analyses of father involvement. This variable is not included in the analyses of the total sample, just the coupled sample. This variable is measured identically as the fathers, except that we exclude any child support received in the last year. Average spouse/partner income is $13,830. There were 20 missing on spouses income (5% of coupled sample). We use two variables to measure union status. One variable indicates whether or not the nonresident father is in a union, two-thirds are either cohabiting or married. The second variable is used in the couple analysis and differentiates between married and cohabiting nonresident fathers. One-quarter of the nonresident fathers living in a union are cohabiting. Finally, we include a measure of conflict with the focal childs mother, because higher conflict is associated with less frequent visits (Amato and Rezac 1994). We use the mean of the fathers response to six questions about areas of conflict (where child lives, how child is raised, how mother spends money on child, how father spends money on child, fathers visits with child, and fathers contribution to economic support). With three possible response categories (none, some, a great deal), the average sample value is about halfway between none and some (1.4). Analysis We use different estimatation strategies for our two dependent variables. For the analysis of whether or not the father made any economic contribution to the focal child, we use logistic regression (DeMaris 1995). We present the ordinary least squares regression models for our analysis of the frequency of father-child visits. We also estimate the same models using ordered logits and find our key independent variables operate in a similar manner. 7
We first estimate the log odds of making payments and visiting often with only the parenting complexity variables in the models. We include variables that have been found to influence nonresident father involvement as well as their new union formation and determine the effects of parenting complexity net of the control variables. Our next two sets of analyses are restricted to only the coupled sample because men who are married or cohabiting confront the most potential parenting complexity. We estimate whether payments made are and frequency of visitation using the components of the parenting complexity variable as well as the level of complexity variable. RESULTS Descriptive Results Table 1 includes the distributions of the parenting responsibilities of nonresident fathers. These distributions were mentioned briefly in the data section. In all, only about one-half (49%) have simple arrangements (i.e., only one set of absent children, no other resident children, and spouse/partner, if present, has no children living elsewhere). Thus, a very substantial percentage have complex parenting roles. Stratifying the sample by union status shows that nearly threequarters of men in unions have complex fathering responsibilities compared to just 8.5% of single fathers (results not shown). Clearly, fathers who are remarried or cohabiting have more opportunities for complex family relationships. The components of the complexity measure are described in Table 1. Based on the prior fertility of nonresident fathers about 8% have obligations to more than one set of nonresident children (6% of single fathers and 9% of nonresident fathers who are married or cohabiting). A similar percentage (8%) are currently residing with biological children who are not the children of the current spouse or partner (if relevant) (3% of single fathers and 11% of coupled fathers). In addition, 14% of all nonresident fathers and one-quarter of coupled nonresident fathers have partners with resident biological children, and a considerable share of nonresident fathers have partners who are also absent parents -- 9.5% of all nonresident fathers and 14% of nonresident fathers in unions. Finally, 42% of coupled nonresident fathers have shared biological children with a spouse or cohabiting partners (28% for all nonresident fathers). To summarize for men who are married or cohabiting, we show the distribution of nonresident fathers across degrees of parenting complexity. About one-quarter (28%) of nonresident fathers have no complexity, i.e., they are responsible for one set of nonresident biological children. Almost one-half of nonresident fathers have one form of parenting complexity, i.e., they are responsible for one set of nonresident children and one other type of child. Approximately one-quarter of nonresident fathers possess two or more forms of parenting complexity, i.e., they have two other types of children in addition to the focal nonresident children.
8
Multivariate Results Table 2 presents multivariate analyses of the effects of parenting complexity, constructed as a dichotomous variable, on our two outcome variables.2 The first two columns show the effect of complexity on visitation and the second two on whether child support was paid to the focal child. Focusing on the first column, nonresident fathers with simple parenting responsibilities visit their children more often than those with complex configurations. The second column indicates that these effects persist with the inclusion of the control variables. As for child support, the coefficient for complexity is not statistically significant but operates in the expected direction. However, the last column shows that with the inclusion of the control variables, the coefficient becomes statistically significant. Nonresident fathers with simple parenting obligations have 85% higher odds of paying child support than nonresident fathers with complex parenting obligations. The variable that seems to explain the suppression of the effect of parenting complexity is union status. This is not surprising because by definition partnered nonresident fathers can experience greater levels of parenting complexity. To examine the effects of the elements of complexity, Table 3 includes our detailed parenting measures. These analyses are restricted to nonresident fathers who are coupled because they face the most potential parenting complexity. The results for the visitation model suggest that nonresident fathers prior fertility is not significantly related to visitation. The spouse or partners fertility appears to be related to visitation. Men who live with stepchildren visit their nonresident children just as often as men who do not have resident stepchildren. Men who live with women who have nonresident children (nonresident stepchildren variable) visit their own nonresident children less often than other nonresident fathers, an effect that remains statistically significant at the .10 level with the inclusion of the control variables. In addition, fathers who have biological children with their current partner are significantly less likely to visit their nonresident children than fathers who do not have new children. With the inclusion of controls, the effect is diminished, but remains statistically significant at the .05 level. Another notable finding in this model is the positive effect of spouse or partners income on visitation with nonresident children, suggesting that more then the partner or spouses fertility influences nonresident fathers contact with his children. The third column presents the zero-order effects of parenting complexity on payment of child support. Fathers who have more than one set of nonresident children have significantly lower odds of paying child support than those who have only one set of nonresident children. We rely on this dichotomous measure for the total sample analysis because single fathers can only have two forms of complexity sets of nonresident children and resident children from a prior union while fathers who are cohabiting or married can have many more types of complexity.
2
9
Resident biological children from a prior union do not influence child support payments. The effects of stepchildren depend on their residence. Nonresident fathers who reside with stepchildren have lower odds of paying child support than fathers who do not live with stepchildren. Men who have nonresident stepchildren share similar odds of paying child support as men who do not have nonresident stepchildren. Nonresident fathers who have biological children with their new spouse or partner have lower odds of making child support payments than men who do not have new biological children. The fourth column shows the model that includes the other covariates. The magnitude and statistical significance of the effect of more than one set of biological children actually increases. Resident biological children have no effect on child support payments. Resident stepchildren are associated with lower odds of making child support payments and nonresident stepchildren continue to have no impact on financial payments to nonresident children. In the model with all the covariates, the effect of new biological children diminishes in size and is no longer statistically significant. This effect is explained by the inclusion of the spouse/partners earnings, suggesting that stepmothers earnings offset the negative effect of additional biological children on child support payments and may be partially used to subsidize nonresident father involvement. The control variables operate in the expected direction. Older children who live farther away from their fathers and have fewer siblings visit their fathers more often. Men who are older and have spouses or partners with higher earnings visit more frequently than fathers who are younger with lower earning spouses or partners. The greater the distance between the father and the child decreases the odds of paying child support. Children with more siblings are more likely to receive child support. Men who are older with high educational levels and have greater earnings are more likely to make child support payments. Also financial payments are greater when their spouse or partner earns more income. Men in cohabiting unions are less likely to make payments than men who are married. Nonresident fathers who have greater levels of conflict are more likely to pay child support. Possibly men who pay child support have more opportunities to have conflict with the mothers of their children. Finally, Table 4 shows the effects of level of complexity or accumulation of complexity on father involvement with nonresident children. Again we limit the analyses to only coupled nonresident fathers because they possess more possible levels of complexity than single nonresident fathers. Men who have two or more types of children visit their children as often as men with 0 complexity, (i.e., those who have the one set of nonresident biological children). The last two columns suggest a stronger relationship between parenting complexity and child support payments; nonresident fathers with the most complex parenting obligations have the lowest odds of paying child support. This result remains significant even with the inclusion of the other covariates. Most of the control variables have similar effects in these models as those presented in Table 3. DISCUSSION Social science and policymakers have typically framed the issue of nonresident fatherhood in terms of fathers and one set of absent children. In general, the possibility that nonresident fathers may have quite complex parenting responsibilities has been downplayed or 10
ignored. Our study is the first to use national data to provide a complete portrait of nonresident fathers parenting obligations. We find that roughly half of all nonresident fathers have parenting responsibilities beyond a single set of nonresident children, and that nearly three-quarters of those who are remarried or cohabiting face potential obligations to other children. Moreover, these estimates are likely to be conservative because some men may be reluctant to count other children when answering survey questions. Recent rates of nonmarital childbearing, divorce, cohabitation, and remarriage suggest that complex family configurations have remained quite common. Our findings support a family systems approach to investigating nonresident fatherhood. We find implications of fathers complex parenting ties for visitation and child support paid to nonresident children. Prior research has not directly empricially evaluated whether fathers who have complex parenting roles have lower levels of involvement with their nonresident children. We find that nonresident fathers with simple parenting responsibilities visit their nonresident children more often and have higher odds of paying child support net of controls for characteristics of the father, child, and income of spouse/partner. We also include a measure of the accumulation of nonresident fathers parenting responsibility to determine whether the level of complexity influences ties to nonresident children. The wider the array of nonresident fathers parenting, the less likely he is to make child support payments. Yet this wider circle of children does not appear to deter fathers visitation with his nonresident children. Our results show that it is important to account for specific types of parenting. Recent studies that have accounted for parenting circumstances have excluded some categories of parenting that have significant effects of father involvement (e.g. Cooksey and Craig 1998; Manning and Smock 1999, 2000). A measure that has not been included in prior studies is whether men have more than one set of nonresident children. We find that men who have more than one set of nonresident children have lower odds of making child support payments but visit their children as often as fathers with only one set of nonresident children. The partner or spouses prior fertility influences both child support and visitation but it varies depending on the residence of the stepchildren. Researchers often account for resident stepchildren but exclude nonresident stepchildren from analyses (Manning and Smock 1999, 2000). Nonresident stepchildren have negative effects on visitation and no effect on child support payments. Perhaps womens nonresident parenthood takes precedence when both spouses/partners are nonresident parents. In contrast, we find that resident stepchildren have negative effects on child support payments and no effect on visitation suggesting they influence economic resources made available but not time spent with children. Cooksey and Craig (1998) also find that resident stepchildren have no impact on visitation, but our findings are not consistent with their results showing nonresident stepchildren have no significant effect on whether nonresident fathers visit monthly. Explanations for this discrepancy may be due to differences in the coding of the dependent variable, their slightly older sample (in the second wave the youngest father is 23), their use of children as the unit of analysis, and their analytic sample that includes men not in unions who by definition do not have stepchildren. In previous work some researchers have grouped all biological resident children together and do not distinguish between those born into the current union and those born to the father from another union (Cooksey and Craig 1998). We find that the fathers biological resident 11
children from prior unions do not influence their fathers involvement with nonresident children. Yet children born into the new union have negative effects on visitation but not on child support. In this paper the effects of new biological children on whether child support was paid are explained by the spouse or partners earnings. The visitation findings are supported by Manning and Smock (2000) but the child support results are counter to those reported by Manning and Smock (1999). This discrepancy may be due to several reasons. The goal of the Manning and Smock (1999) paper was to model change in payments using longitudinal data, they include men not in unions who by definition do not have new biological children, the partner/spouses earnings are not included in their analyses, and they measure level of payment and not whether support was paid. In this paper we cannot determine exactly why nonresident fathers with more complex parenting obligations visit less often or pay less child support. We are able to control for one indicator of nonresident fathers economic resources, it is quite difficult to measure other important dimensions of fathers lives, such as time and emotional resources. Our findings suggest that the reason that nonresident fathers with complex parenting obligations visit or pay child support less often is not due completely to the economic constraints (measured by income) that a wide network of responsibilities may impose on fathers. Yet, we have not been able to account for all measures of economic well-being or constraints and more in-depth analyses with detailed measures are required. We speculated that mens contact and support of nonresident children may be more responsive to their own prior fertility rather than their spouse or partners prior fertility. Our results do not support this notion. It appears that partners prior fertility is a more significant predictor of visitation than nonresident fathers own fertility. With regard to child support, men and womens prior fertility both have significant effects. We also anticipated that resident children may have a greater effect on ties to children than nonresident children. However, we find that the effects of children cannot be simply delineated by residence. For example, nonresident stepchildren have a negative effect on visitation and resident stepchildren have a negative effect on child support. Our findings indicate that the spouses or partners of nonresident fathers act as important sources of influence on nonresident father involvement. Not only does their prior fertility matter, in this study we were able to account for the income of nonresident fathers spouse or partner. The spouse or partners income has positive effects on both child support payment and visitation. It appears that the higher incomes of spouses or partners may make it easier for men to both visit and financial support their children. Further attempts should be made to consider how other features of the nonresident fathers spouse or partner influence ties to nonresident children. These women are all nonresident stepmothers (they could be other types of parents as well) and are often treated as being invisible parents when in fact they appear to matter. This paper is limited to only an examination of two types of involvement: visitation and child support payments. Obviously men can be involved in childrens lives in many ways and are not restricted to our relatively simple measures of involvement. Future studies should explore a fuller range of parenting activities. This work suggests that efforts to engage men in the lives of their nonresident children should pay particular attention to those men who have potentially competing parenting 12
responsibilities. It is this situation of being responsible for children who are his, hers and ours that often results in men spending less time and resources with their nonresident children. Unfortunately, nonresident children appear to lose out when their father has potential obligations to other children via biological or social fatherh...