<
 
 
 
 
×
>
hide
You are viewing a Web site, archived on 18:29:24 Nov 08, 2004. It is now a Federal record managed by the National Archives and Records Administration.
External links, forms, and search boxes may not function within this collection.

Non-Custodial Parents' Participation in Their Children's Lives:

Evidence from the Survey of Income and Program Participation

Volume I

Summary of SIPP Analysis

Christine Winquist Nord and Nicholas Zill
Westat, Inc.
Subcontractor to the Lewin Group, Inc.
Prepared for the Office of the Assistant Secretary for Planning and Evaluation, DHHS
August 14, 1996

ACKNOWLEDGEMENTS

This project was funded by the Assistant Secretary for Planning and Evaluation, U.S. Department of Health and Human Services, through contract number HHS-100-93-0012, Delivery Order 11. The study was conducted by Westat, Inc. under a subcontract to The Lewin Group. We wish to thank our Westat colleagues, Margaret Daly and Amy Van Driessche who took on the challenge of preparing all the tables and charts, and Laura Loomis who generously agreed to help annotate key articles. We also gratefully acknowledge our two excellent programmers, Peter Robert Shaw of R.W. Beck and Nina Blecher, a consultant. Both easily navigated the complexities of the Survey of Income and Program Participation. Without them, the analyses would not have gone so smoothly. Thanks also to Burt Barnow and David Stapleton of The Lewin Group. Burt shared his expertise with us, read drafts of the report, and gave us encouragement and comments. David gently, but firmly reminded us of the schedule. It was a pleasure working with Anne Benson, our Project Officer, and Linda Mellgren, both of ASPE. Linda and Anne patiently read and reread several drafts of the report and offered valuable advice and comments. We also thank Barbara Cleveland and Susan Notar of the Office of Child Support Enforcement for sharing their work and providing us with comments. Thanks also to the DHHS Work Group for providing comments on the project: David Arnaudo, Gaile Maller, and Mary Cohen, Office of Child Support Enforcement; Marianne Rufty, David Smith, Ken Maniha, Mark Fucello, Administration for Children and Families; and Paul Legler and Jeanine Smartt, Office of the Assistant Secretary for Planning and Evaluation.

Table of Contents

ACKNOWLEDGEMENTS

EXECUTIVE SUMMARY

INTRODUCTION

RESEARCH METHODS

Data

Creation of Analytic Files

FINDINGS

Joint Custody

Voluntary versus Court-Ordered Agreements

Visitation Provisions

Paternal Involvement

Separated Parents Residing in Different Cities or States

Child Support

Custodial Fathers/Non-Residential Mothers

Custodial Parents Without Child Support Awards

Separated Parents Who Have Never Been Married to Each Other

Children's Well-Being

SUMMARY OF FINDINGS OF REGRESSION ANALYSES

CONCLUSIONS AND POLICY IMPLICATIONS

Encouraging Voluntary Agreements

Requiring Explicit Visitation Provisions

Encouraging and Facilitating Non-residential Parents' Access to Their Children

Promoting Joint Custody Arrangements

Limiting Moves by Divorced Parents

Other Implications of the SIPP Findings

SUMMARY

REFERENCES

APPENDIX A: Child Support Module 1990 and 1991 Panels of SIPP

APPENDIX B: Comparability of SIPP Estimates and Current Population Survey Estimates



EXECUTIVE SUMMARY

INTRODUCTION

Divorce and non-marital childbearing have become commonplace and have dramatically altered children's lives. It can no longer be assumed that most children will spend their entire childhoods living with both parents. To the contrary, approximately half will live in single parent homes at some point before they turn age 18. Unfortunately, a common pattern is for the non-residential parent to become increasingly detached over time, paying minimal or no child support and visiting infrequently if at all. The costs to the children involved and to society at large of this disengagement are far from trivial. Many non-custodial parents do not pay all the child support they owe. Many others have no obligation to pay support. Nonpayment of support forces some families below the poverty level and onto government welfare programs. For others, it means a reduced standard of living and an uncertain future. The costs to children are seen in an increased likelihood of dropping out of school and increased, social, emotional, psychological, and behavioral problems. Not all children are affected and some that are overcome their difficulties in a few years, but others experience long-term setbacks.

The connections between custody arrangements, payment of child support, parental involvement, and child well-being are still not well-understood. Many of the studies on which policy is being made are based on small, unrepresentative samples or on the experiences of divorcing couples in particular states. These studies may not reflect the experience of most custodial parents and their children. If the assumptions about the positive influence of joint custody, for example, or links between payment of child support and visitation are wrong, then the outcomes for families and children may not be to their benefit after all. Although not based on experimental designs, national survey data can be used to cast more light on the issues surrounding visitation, custody, child support, and child well-being and provide policymakers with a more solid base from which to proceed.

The aim of this project was to improve understanding of the relationship between non-custodial parent involvement, children's well-being, child support, and custody arrangements. Two approaches were used. Analyses of data from the Survey of Income and Program Participation (SIPP) were used to provide national estimates of persons living in different custody arrangements, and to examine the connections between custody arrangements, child support payments, parental involvement, and children's well-being in both the divorced and never-married populations. In addition, a review of recent literature was conducted and gaps in the research were noted. The findings of the review were used to guide the SIPP analyses. Also, as part of the project, a limited set of articles was annotated and a bibliography of selected papers on custody, visitation, and child well-being was created. The analyses of SIPP are contained in Volume I of this report. The literature review, selected annotated articles, and the extended bibliography are contained in Volume II. In addition, supplementary tables based on the SIPP were produced. These tables show the demographic characteristics, economic status, and living conditions of custodial parents, and selected measures of children's well-being by the existence of a child support award and whether child support was received, whether the agreement was voluntary or court-ordered, and the type of arrangement. Information on demographic background, the economic status, and the living conditions of the custodial parent and selected measures of child well-being are also shown by the amount of visitation with the non-resident parent. These tables were prepared for all custodial parents, for female custodial parents, for male custodial parents, and for divorced female custodial parents. The sample size for male custodial parents is small in some cells of the tables, so caution should be used in drawing inferences from these tables.

VOLUME I: SUMMARY OF SIPP ANALYSES

The Survey of Income and Program Participation (SIPP) is based on a national probability sample of the U.S. civilian, non-institutionalized population. It is funded and conducted by the U.S. Bureau of the Census. The SIPP is a major source of information about the demographic and economic situation of persons and families in the United States. The SIPP is one of two national surveys containing extensive amounts of child support information. The other survey is the Current Population Survey.

There are three reasons why SIPP was uniquely suited to examine the relationship of child support, child custody, and child well-being: (1) detailed programmatic information; (2) longitudinal nature of the data, and (3) a child support module with questions on award, payment, custody, visitation, and child well-being. There are a variety of questions that can be used to assess the economic well-being of such children and, to a lesser extent, their social well-being. SIPP also contains topical modules which ask about consumer durables owned by each household, the living conditions of households, and the ability of households to meet basic needs. These modules were used to characterize the circumstances in which children are growing up. In addition, the SIPP contains information about the health of persons aged 15 and older.

The analyses relied on a variety of methods. Frequencies and crosstabulations were used to develop profiles of custodial parents with different child support characteristics, such as the number of persons with joint physical and legal custody arrangements, and information about the extent of non-cash support received from non-residential parents. In addition, ordinary least squares regression and logistic regression were used to examine in more detail the predictors of child support and visitation. Multivariate models were also estimated to examine the link between child support awards, custody arrangements, visitation, and payment of child support on children's health and on the receipt of AFDC in the previous year.

Highlights of Descriptive Findings

Multivariate Results and Policy Implications

This study is not a randomized policy experiment or even a non-randomized study of specific policy initiatives. Thus, we must be circumspect about how far we go in drawing policy-related conclusions about the findings. A correlational panel study such as the present one cannot prove that a given policy will work as its advocates content it should. An observed relationship may be due to the operation of other, unmeasured factors. However, the failure to find an expected correlation can provide firmer grounds for believing that a specific policy will not work as anticipated. These results apply to couples who have a written child support agreement. Information on contact and payment of child support was not asked of persons without a written agreement. With these warnings in mind, the SIPP analyses provide support for the following types of activities:

VOLUME II: SYNTHESIS OF LITERATURE

The literature review contains three components: a synthesis of the literature on child development, custody, visitation, and child well-being; an annotated bibliography; and a selected bibliography. The synthesis briefly summarizes recent perspectives on children's development and on the role of the father in families. These two perspectives are important in understanding how and why marital disruption may affect children. The fact that children change over time, developing new skills and capacities and having different needs means that their response to the breakup of their families may differ depending upon their age at the time. Moreover, since the majority of non-custodial parents are fathers, it is important to understand the role that fathers play in children's lives and how that role changes as children grow older. The review then discusses several possible ways in which family disruption may affect children's lives. These are the loss of a parent (usually the father), the adjustment of the custodial parent, parental conflict, economic hardship stemming from the disruption, stressful life changes, including the loss of social supports and other resources. Gaps in existing research are noted. Key findings that helped guide the SIPP analysis are:


NON-CUSTODIAL PARENTS PARTICIPATION IN THEIR CHILDREN'S LIVES: EVIDENCE FROM THE SURVEY OF INCOME AND PROGRAM PARTICIPATION

Christine Winquist Nord and Nicholas Zill

Westat, Inc.

INTRODUCTION

The sustained high rates of divorce and non-marital childbearing in this country have altered children's lives. It can no longer be assumed that children will spend their entire childhoods living with both their parents. Many children undergo the painful experience of witnessing their parents' marriages fall apart. Others are born to parents who never marry. Some of these children never have the experience of living in a stable two-parent home.

Research has shown that a common pattern following the breakup of a family is for one of the parents to become increasingly detached, paying little or no child support and visiting infrequently if at all (Seltzer, 1991; Furstenberg et al., 1983; Peterson and Nord, 1988). Various reforms have been proposed to try to remedy this situation, but there is a need for evidence on the potential impact of different reforms. Maccoby and Mnookin (1992), Braver et al. (1991), and others have conducted a variety of studies that provide intriguing insights into custody and child support issues and the potential effects on families and children. These studies are often confined to specific states, such as California, or sections of the subpopulation, such as persons filing for divorce, and may not be representative of the nation at large. Large national surveys, though not based on experimental designs, can provide data that illuminate the issues and that suggest likely outcomes of different policies. Data from the Current Population Survey (CPS) have been used to provide information on custody arrangements, payment of child support, and contact with the non-residential parent (U.S. Bureau of the Census, 1995, 1991). However, the CPS lacks a longitudinal component and has very limited information on children's well-being. The Survey of Income and Program Participation (SIPP) does have a longitudinal component and contains more information that can be used to describe the conditions of children's lives. For these reasons, we use the SIPP to provide national estimates of persons living in different custody arrangements and to examine the connections between custody arrangements, child support payments, parental involvement, and children's well-being in both the divorced and never-married populations. In particular, we examine the following topics:

This volume details the results from the SIPP analyses. A companion volume (Volume II) contains a synthesis of recent literature on child support, child custody, visitation, and child well-being, as well as annotations of selected articles, and a bibliography of related articles identified during the review, not all of which were included in the synthesis.

In the next section, we provide an overview of the SIPP data and the files that were created for the analyses. The following section describes our findings. We conclude the report with a discussion of the potential policy implications of our research.

RESEARCH METHODS

Data

SIPP is based on a national probability sample of the U.S. civilian, non-institutionalized population. It is funded and conducted by the U.S. Bureau of the Census. An advisory panel with representatives from selected federal agencies oversees the questionnaire design and frequency of interviewing of the Child Support Module. The SIPP is a major source of information about the demographic and economic situation of persons and families in the United States. The primary focus of SIPP is adults ages 15 years or older. Because of SIPP's detailed programmatic information and the longitudinal nature of the data, the child support module supplements and expands the type of information available from the Current Population Survey.(1)

The 1990 panel of SIPP interviewed approximately 20,000 individuals aged 15 and older. However the number of adults who responded to the child support module (those who are custodial parents of children under 21 who had a parent living elsewhere) is much smaller: 3,124. The 1991 panel interviewed approximately 14,000 individuals in the same age range. Of these, 1,619 were eligible for the child support module. The child support module was collected during Wave 3 (October 1990 to January 1991) and Wave 6 (October 1991 to January 1992) of the 1990 panel and during Wave 3 (October 1991 to January 1992) of the 1991 panel.

Although children under age 15 are not interviewed, there are a variety of questions that can be used to assess the economic well-being of such children and, to a lesser extent, their social well-being. SIPP contains topical modules which ask about consumer durables owned by each household, the living conditions of households, and the ability of households to meet basic needs. These modules can be used to characterize the circumstances in which children are growing up. In addition, the SIPP topical module on functional limitations asks whether children aged 6 and older have ever received special education services or if they are limited in their ability to do regular school work because of a physical, learning, or mental health condition. SIPP also obtains information about the health and educational attainment of persons aged 15 and older.

The SIPP has been underutilized to study issues surrounding child support, yet the use of SIPP allows analyses that are not possible with the more commonly used Current Population Survey. Advantages of the SIPP over the CPS include the following:

SIPP is not without its limitations, however. Among these are the following:

Creation of Analytic Files

Several files were created for this project. Using data from the 1990 Panel, we created a file of all parents who responded to the child support modules at either Wave 3 or Wave 6, one year later. This file was used to examine the interrelationship among type of custody arrangement, contact with the non-residential parent, payment of child support, family structure, and economic circumstances. Use of this file, allowed us to examine in more detail whether changes in support tend to precede, follow, or coincide with changes in visitation. It also permitted us to examine the relationship between different types of custody arrangements and patterns of visitation. This file also allowed us to examine a relatively understudied group, namely men who have custody of their children following separation, divorce, or non-marriage. We could thus examine patterns of visitation and payment of child support by female non-residential parents.

We also created a file of all children under 21. Selected parent information, such as marital status and age, was attached to each child's record. This file allowed us to study the effects of various custody arrangements, child support payment patterns, and family structure on selected aspects of children's well-being.

Parallel files were created using the 1991 panel. Although the 1991 panel was substantially smaller than the 1990 panel, the inclusion of the topical modules on living conditions and the ability of households to meet basic needs allowed us to delineate a fuller portrait of the circumstances of children's families following marital breakup.

FINDINGS

In this section, we describe our results. For each topic addressed, we list the set of research questions that guided our research and our preliminary hypotheses, if any. We then describe the results that were actually obtained. Crosstabulations were used to answer many of the questions. The observed associations could change if potentially confounding factors such as marital status, education, and age were statistically controlled. Regression models were also used to answer some of the questions, particularly those that seek to understand the links between three or more factors such as child support, contact, and child custody.

According to estimates from the 1990 Panel of SIPP, there were approximately 12.4 million parents potentially eligible for child support in the United States in the Winter of 1990/91 because they had children under 21 living with them who had a parent living elsewhere.(2)

As can be seen in Figure 1, most custodial parents are mothers (88% or 10.9 million). However, there are an estimated 1.5 million custodial fathers (12% of all custodial parents).

Joint Custody

How many divorced parents in the U.S. have joint custody arrangements regarding their children?

How many have joint physical custody as well as joint legal custody?

What are the demographic characteristics of families that enter into joint custody arrangements? How do they differ from persons with sole custody arrangements?

Is joint custody associated with greater economic well-being for the family in which the child primarily resides?

Is joint custody associated with more frequent contact between the non-residential parent and the children than sole custody with visitation?

Is joint custody associated with payment of more or less child support by the non-residential parent than sole custody with visitation?


H1 Persons who enter into joint custody arrangements will have characteristics that are conducive to greater economic well-being. They will be more educated, have formerly been married, have fewer children, and the residential parent will be more likely to be employed than custodial parents in sole custody arrangements.
H2 Because of their favorable characteristics, they will be less likely to live in poverty and will have better living conditions than families with other custody arrangements.

The SIPP only asks about custody arrangements for persons with a formal written agreement. According to data from SIPP, of the 6.2 million custodial parents with a written child support agreement, 74 percent have a sole custody arrangement of some sort, 21 percent have a joint custody arrangement, and 5 percent have a split custody or other type of arrangement.

Joint legal custody is far more common than joint legal and physical custody. Of the 1.3 million persons with joint custody arrangements, just over 1 million (79.6%) have joint legal only arrangements with one of the parents, usually the mother, having physical custody. The remaining 262 thousand (20.4%) report joint legal and physical arrangements.

Persons who enter into joint custody arrangements, as we expected, differ in several respects from persons with other types of arrangements (See Table 1). They are more likely to be currently divorced than persons with sole custody arrangements (54.5% versus 43.6%) and less likely to be never married (1.21% versus 12.8%). They are less likely to be black (6.9% versus 17.9%) or Hispanic (4.9% versus 7.7%) compared to persons in sole custody arrangements. They also tend to be better educated than persons with sole custody arrangements. Approximately 16 percent are college graduates compared to 10.7% of persons in sole custody arrangements and only 12 percent have less than a high school education compared to 16.8 percent with sole custody arrangements.

A somewhat smaller proportion of persons with joint custody arrangements are under age 25 compared to persons with sole custody arrangements (5.5% versus 9.6%), but the differences in the proportions aged 25-34, 35-44, and 45 and older are not large. Contrary to our expectations, there are not large differences in the number of children in the households of persons with joint custody arrangements and those with sole custody arrangements. There are also not large differences in the work pattern in the previous year between persons with joint custody arrangements and those with sole custody arrangements, though those with joint custody are slightly less likely not to have worked the entire previous year (16.7% versus 19.3%).

As we hypothesized, persons with joint custody arrangements are economically better off than persons with sole custody arrangements (see Table 2). They are less likely to have experienced one or more months of poverty in the previous year (26.5% versus 37.6%), to have received AFDC one or more months in the previous year (7.2% versus 19.6%), or to have moved in the previous year (10.3% versus 16.7%) compared to persons with sole custody arrangements. They also experience less income instability during the course of the year (mean standard deviation of monthly income is 27.9 compared to 32.2), have higher median family incomes ($28,288 versus $24,655), and are more likely to own their own homes (63.1% versus 56.3%).

Table 1 p. 1

Table 1 p. 2

Table 2

Persons in joint custody arrangements not only are economically more advantaged, but their living conditions, in general, appear to be better as well (see Table 3). They are more likely to own a personal computer (25.9% versus 16%) than persons with sole custody arrangements. They are less likely than persons in sole custody arrangements to lack one or more essential consumer durables such as a telephone, stove, or refrigerator (4.7% versus 11.7%), to have not paid their utility bills (18.2% versus 25.0%), to have had their utilities turned off because of lack of payment (1.3% versus 6.2%), and to have had their phone turned off because they didn't pay their bills (3.0% versus 9.8%). They also express more favorable attitudes about their neighborhoods. Fewer of them rate their neighborhood as unsafe (5.3% versus 12.1%) or say that there have been times in the past month that they wanted to go out but didn't because they thought it would be unsafe (6.3% versus 11.6%). They are also less likely to say that they want to move because of the poor quality of their neighborhoods (6.6% versus 10.3%).

The finding that persons with joint custody arrangements are better off economically and appear to have better living conditions than persons with sole custody arrangements more likely is due to the characteristics of people who enter into joint custody arrangements than to the beneficial effects of the arrangement itself. (If states encourage or require more parents to enter into joint custody arrangements, it is possible that fewer economic differences between persons in joint custody and in sole custody arrangements will be observed.)

Another advantage experienced by persons with joint custody arrangements is that the amount of contact with the non-residential parent is higher (see Table 4). The mean number of days of contact in the previous year for persons with joint custody arrangements who had at least some contact was 47.3 days compared to 30.8 days for persons with sole custody arrangements and 32.4 days for persons with sole custody arrangements with visitation provisions. The median days of contact by the non-residential parent is also higher in joint custody arrangements compared to sole custody arrangements (30 days versus 7 days) or sole custody arrangements with visitation provisions (30 days versus 12 days).

Among persons receiving any child support, those with joint custody receive larger amounts of child support than persons in sole custody arrangements (see Table 2). The median amount of child support received by persons in joint custody arrangements is $3000 compared to $1920 of persons in sole custody arrangements and $1980 of persons in sole custody arrangements with visitation provisions. The mean amount received by persons with joint custody arrangements is also higher than that received by persons with a sole custody or with a sole custody arrangement with visitation provisions, $3661 versus $2467 and $2639, respectively.

Table 3 p. 1

Table 3 p. 2

Table 4

Compliance with child support orders, defined as the ratio of child support received to child support that was due in the previous year, is also greater among parents with joint custody arrangements compared to parents with sole custody arrangements (see Table 2). Custodial parents with joint custody arrangements received 76 percent of what they were owed, whereas custodial parents with sole custody arrangements received only 62 percent of what they were owed in the previous year.

Voluntary versus Court-Ordered Agreements

How many divorced parents in the U.S. have voluntary as opposed to court-ordered child support agreements?

What are the demographic characteristics of persons who enter into voluntary as opposed to court-ordered agreements?

Are voluntary agreements associated with greater economic well-being for the family in which the child primarily resides than court-ordered agreements?

Are court-ordered agreements associated with less frequent contact between the non-residential parent and the children than voluntary agreements?

Are court-ordered agreements associated with payment of less child support by the non-residential parent than voluntary agreements?

H1 Voluntary agreements will be associated with greater economic well-being for the family in which the child primarily resides.
H2 Voluntary agreements are probably indicative of a less hostile relationship between former partners than court-ordered agreements. For this reason, voluntary agreements will be associated with more frequent contact and more regular payment of child support than court-ordered agreements.

Most child support agreements are written agreements of some form. Of the 6.7 million custodial parents who have ever had a child support award, 91.5 percent had a written agreement. Only 8.5 percent had a non-written agreement.

The most common type of written agreement is a court-ordered agreement. There are approximately 4.5 million court-ordered written agreements, constituting 73 percent of all written agreements ( see Figure 2).(3) Such agreements do not necessarily mean that the agreement was contested. Another 1.4 million agreements are voluntary written agreements ratified by the court. These agreements represent 23 percent of all written agreements. There are only an estimated 204 thousand agreements (3% of all written agreements) that are something other than a voluntary agreement ratified by the court or a court-ordered agreement, such as a written agreement that was not ratified by the court.

Persons who voluntarily arrange a child support agreement differ in some respects from those who have court-ordered agreements (see Table 5). They are less likely to be never married than those with court-ordered agreements (5.9% versus 11.5%) or to be currently in their first marriage (2.6% versus 4.3%). They are somewhat more likely to be non-Hispanic whites (81.9% versus 73.7%) and to be 45 or older (15.1% versus 8.5%) than custodial parents with court-ordered agreements. They also tend to be better educated: 19.9 percent are college graduates compared to 9.1 percent of custodial parents with court-ordered agreements. They are more likely to have worked full-year, full-time in the previous year than custodial parents with court-ordered agreements (58.2% versus 46.6%).

As we hypothesized, custodial parents with voluntary written agreements fare better economically (see Table 6). Their median family income is higher ($30,531 versus $23,814). They are also less likely than custodial parents with court-ordered agreements to experience one or more months of poverty in the previous year (27.3% versus 38.1%) or to have received AFDC for one or more months (8.8% versus 19.3%). They are more likely to own their own homes (67.9% versus 54.7%). In spite of these differences in economic well-being, there is no real difference between the two groups on our measure of income instability.

Custodial parents with voluntary agreements also appear to have better living conditions than custodial parents with court-ordered awards (see Table 7). They are less likely to lack an essential consumer durable such as a telephone, stove or refrigerator (7.1% versus 10.5%) and are more likely to own a personal computer (25.5% versus 15.2%). They are also less likely to report that in the previous 12 months there was a time when they couldn't pay the full amount of their utility bills (18.9% versus 26.4%), or that service had been turned off because payments were not paid (2.8% versus 5.9%). They are also less likely to report that someone in their household had not seen a doctor or a dentist even though they needed to. Twelve percent of custodial parents with voluntary agreements reported that someone in their household hadn't seen a doctor even though they needed to compared to 16 percent of custodial parents with court-ordered agreements. Similarly 16.3 percent reported that someone in their household hadn't seen a dentist even though they needed to compared to 19 percent of custodial parents with court-ordered agreements. Persons with voluntary agreements are less likely than those with court-ordered agreements to describe their neighborhood as unsafe (6.3% versus 12.8%) or that they would like to move because of the unsatisfactory neighborhood conditions (6% versus 10.7%).

Table 5 p.1

Table 5 p. 2

Table 6

Table 7 p. 1

Table 7 p. 2

Custodial parents with court-ordered agreements are twice as likely as those with voluntary agreements to report that the non-residential parent did not visit the children at all in the previous year, 35.5 percent compared with 17.1 percent ( see Figure 3). Among custodial parents who report any contact with the non-residential parent in the previous year, those with court-ordered agreements report fewer days of contact compared to those with voluntary agreements. The mean number of days of contact in the previous year reported by those with court-ordered agreements was 28 days compared to 52 days for those with voluntary agreements. The median days of contact reported are 7 days and 26 days, respectively.

There are also differences in the amount of child support received and in the receipt of non-cash assistance from the non-resident parent according to whether agreements were court-ordered or voluntary (see Table 6). Custodial parents with voluntary agreements receive more child support and are more likely to receive non-cash assistance than persons with court-ordered agreements. The median amount of child support received by custodial parents with voluntary arrangements is $2,750 compared to $1,900 among persons with court-ordered arrangements. Moreover, nearly one-fifth of them (19.3%) receive non-cash assistance from the non-resident parent compared to 13.2 percent of custodial parents with court-ordered awards.

Compliance with child support orders is also higher among parents with voluntary as opposed to court-ordered agreements (see Table 6). Custodial parents with voluntary agreements received 73 percent of what they were owed, whereas custodial parents with court-ordered agreements received 62 percent of what they were owed.

Visitation Provisions

How many divorced parents in the U.S. do not have visitation provisions in their child support and custody agreements?

Do custodial parents with sole custody, but no visitation provision differ in their demographic characteristics from those with sole custody with visitation provisions?

Is the lack of an explicit visitation provision associated with less economic well-being for the family in which the child primarily resides?

Is there less frequent contact between the non-residential parent and the child in sole custody arrangements with no visitation provisions than in sole custody arrangements with explicit visitation provisions?

Is the lack of an explicit visitation provision associated with payment of less child support by the non-residential parent than sole custody with visitation provisions?

Visitation may or may not be specified in a child support agreement. The lack of explicit visitation arrangements, however, does not mean that visitation does not occur or is discouraged, though it may. Agreements may state that visitation will occur at the "mutual convenience" of the two parents or use terms such as "usual and customary visitation." Such phrases, because they do not lay out a detailed visitation plan, probably are not considered a visitation arrangement by many respondents. Regardless of the type of custody arrangement, the majority of parents report that they do have some type of visitation provision in their agreement. Nearly two-thirds (64.6%) or almost 4 million custodial parents report that they have visitation provisions in their agreements ( see Figure 4). The proportion is very similar for those with joint custody and those with sole custody arrangements, 66.5 percent and 65.6 percent, respectively. It is not clear, however, how custodial parents with joint legal and physical custody arrangements interpreted the question on visitation provisions. Some parents may have considered arrangements made for when the child or children would live with the other parent as a visitation provision, while others may not have considered such arrangements as a visitation provision. Even so, the responses to the question on visitation indicate that approximately 2.2 million persons have no explicit visitation provision in their written child support agreements.

As we saw above, persons with joint custody differ in several respects from those with sole custody. The demographic characteristics of custodial parents with sole custody, but no visitation provisions have characteristics that could make them more vulnerable to economic difficulties. For example, 1.2 percent of custodial parents with joint custody are never married, compared to 8.3 percent of those with sole custody with visitation provisions, and 21.4 percent of those with sole custody without visitation provisions. Similarly, 12.2 percent of those with joint custody reported that they had not graduated from high school, compared to 15.3 percent of custodial parents with sole custody with visitation, and 19.6 percent with sole custody, but no visitation provision.

The same pattern holds true for the economic well-being of the families: those with sole custody, but no visitation provision fare less well than those with joint custody or those with sole custody with visitation (see Table 2). They are more likely than the other two to experience poverty in the previous year: 26.5 percent of persons with joint custody, 32.8 percent of those with sole custody with visitation, and 46.9 percent of those with sole custody without visitation reported experiencing one or more months of poverty in the previous year. They are also more likely to report receiving AFDC in the previous year: 7.2 percent of those with joint custody, 15.1 percent of those with sole custody with visitation, and 28.4 percent of those with sole custody without visitation reported receiving AFDC for one or more months in the previous year.

Custodial parents with sole custody arrangements which do not specify visitation are less likely than those with sole custody with visitation to say that the non-residential parent had some contact with the children in the previous year, 54.3 percent versus 73.3 percent (see Table 4). However, custodial parents with joint custody arrangements are the most likely to report that some visits occurred: 81.6 percent. Among those reporting that some contact took place, parents with sole custody with no visitation report less contact than those with sole custody with visitation or with joint custody arrangements. The mean days of contact reported by parents with sole custody arrangements with no visitation provisions was 27.8 days compared to 32.4 days for those with sole custody with visitation provision, and 47.3 days for those with joint custody arrangements. The median days of contact tell a similar story: 1 day of contact for those with sole custody and no visitation, 12 days for those with sole custody with visitation provisions, and 30 days for those with joint custody.

Like contact, among those who report receiving any child support, the mean and median amounts received by custodial parents with sole custody with no visitation provisions is less than reported by custodial parents with joint arrangements or with sole custody arrangements with visitation provisions (see Table 2). The mean amount of child support received in the previous year by parents with sole custody with no visitation is $2101, compared to $3,661 for those with joint custody, and $2639 for those with sole custody with visitation. The median amounts received are $1512 for parents with sole custody with no visitation provision, $1980 for parents with sole custody with visitation provisions, and $3000 for persons with joint custody arrangements.

Compliance with child support orders is also lower among parents with sole custody arrangements with no visitation provisions compared to parents with sole custody with visitation and to parents with joint custody arrangements. Custodial parents in sole custody arrangements with no visitation provision received 57 percent of what was owed to them compared to 65 percent received by parents with sole custody arrangements with visitation, and 76 percent received by parents with joint custody arrangements.

Paternal Involvement

How often do non-residential fathers visit their children in the course of a year?

Does the amount of visitation by the non-residential father differ by the custodial mother's demographic characteristic?

Does the amount of visitation by the non-residential father differ by the type of custody arrangement or whether it was arrived at voluntarily or court-ordered?

What factors are related to high levels of visits? Conversely, what factors are related to low levels of visits?

Is payment of more child support by non-residential parents in one year predictive of more frequent contact with their children in the following year?

How is years elapsed since an agreement is reached related to paternal involvement?

Does remarriage of the mother affect how often the father visits?

H1 Consistent with previous studies, visits by the father will be higher if the mother was ever married, if she is more educated, and if the children are younger.
H2 Fathers with joint physical custody will have more contact with their children than fathers with other types of arrangements. Fathers will have the least contact with their children if the mother has sole custody.
H3 Involvement will be higher among fathers whose custody arrangement was arrived at voluntarily compared to fathers whose arrangement was court-ordered.

In the 1990 and 1991 panels of SIPP, questions about contact with the non-residential parent were only asked of persons with formal written agreements. The 1992 panel of SIPP, not fully available during the time of this study, obtains information on contact from parents without written agreements as well.

According to data from the 1990 panel of SIPP, a substantial proportion of fathers (31.7%) have not spent time with their children in the previous 12 months, according to the custodial mothers (see Table 8). However, approximately one-quarter (23.8%) of non-resident fathers see their children at least once a week. The mean number of days of contact reported by custodial mothers is 33 days in the previous year, with a median of 12 days.

Frequency of contact varies by the background characteristics of the custodial mother. Mothers who had not married the non-resident father are less likely than other mothers to report that the non-resident father spent time with their children in the previous year (see Table 9). In fact, 40.6 percent of never married mothers and 46.7 percent of mothers who are currently in their first marriage report that the non-resident father did not visit their children at all in the previous year, compared to 35 percent of remarried mothers and 26.7 percent of currently divorced mothers.

Contact with the non-resident father is more likely to occur if the custodial mother is more educated: 82 percent of those who are college graduate report that the father had some contact in the previous year compared to 57.5 percent of those with less than a high school education.

Frequency of contact with the father also varies by the characteristics of the custody and child support agreements ( see Figure 5). Contact is more frequent, at least in sole custody arrangements, when visitation provisions are specified in the agreement than when they are not (see Table 8). Contact is also more frequent when child support agreements are reached voluntarily as opposed to being court-ordered.

Table 8

Table 9 p. 1

Table 9 p. 2

Regression models were estimated to determine what factors are related to high or low levels of visits, controlling for the background characteristics of the custodial mother, characteristics of the child support agreements, previous contact, and payment of child support. As with the bivariate relations described above, these models were restricted to custodial mothers with written child support agreements. For these models, the dependent variable was the natural log of days visited + 1. Days of contact, like other variables that count things, is skewed to the right, which means that most of the cases are clustered towards the lower end of end of the range with some outliers to the right. Transforming the data with a natural log serves to spread out the data at the lower end and pull in the data on the right. Essentially, the transformation is "straightening out" the variable so it is more suitable for use in a regression analysis. The number of days of visits equivalent to the log days are shown below:

Natural Log Transformation

Ln (Days + 1)

Days of Visitation

0

1

2

3

4

5

5.9

None

1.7

6.4

19.1

53.6

147.4

365

Five models were estimated. The first regressed the natural log of days on a set of background characteristics including years elapsed since the agreement was made; where the non-resident father lives; the custodial mother's education, race/ethnicity, age, and marital status; the number of children covered by the award, and the age of the youngest child in the custodial mother's household. For the second model, these same background characteristics were entered plus information on the custody agreement and how it was reached. The third added information on the non-residential father's compliance with the child support order in the year preceding wave 3. The fourth model added information on previous contact, but not on compliance with the child support order. And the fifth included all the previous variables from the other models. The results of these models are shown in Table 10.

Table 10 p. 1

Table 10 p. 2

Below the relationships that were found in the regression models are summarized.

What Factors Predict to Parental Contact?

Court-ordered as opposed to voluntary support agreements (negatively related, e.g., parents with court-ordered agreements as opposed to voluntary agreements are less likely to have contact with their children)

Compliance with child support order in previous year (positively related)

Sole custody without visitation provisions (negatively related)

Other parent lives in a different state, or in a different city or county (negatively related)

Years elapsed since support agreement (negatively related)

Parent education level (curvilinear relationship)

Youngest child is a preschooler (ages 3-5) (positively related)

Parents never married (negatively related)

Factors That Might Be Expected to Predict, But Do Not

Joint legal custody (i.e., over sole custody with visitation provisions)

Joint legal and physical custody (i.e., over sole custody with visitation provisions)

Mother has remarried


Many of the relationships observed at the bivariate level continued to be significant in the multivariate models, though the addition of previous contact removed some of the observed associations. For models that contained the demographic, custody, and child support variables, having a court-ordered agreement, lacking a visitation provision, living in a different state or county, and being a never married custodial parent were all associated with lower contact even after controlling for the other variables. Variables that continued to be predictive of lower contact even after controlling for previous contact and child support were having a court-ordered agreement, sole custody without visitation, and being never married. There appears to be a curvilinear relationship between education of the custodial parent and contact. Custodial parents with high school degrees reported higher contact compared to custodial parents with some college experience. There was no significant difference between those with less than a high school education, those with some college experience, and college graduates in the amount of contact once the other variables were entered into the model. As compliance with the child support order in the year prior to wave 3 increased, so did the amount of contact in the year following Wave 3 even after controlling for all the other variables.

Although joint custody was expected to be associated with greater contact, neither joint legal only nor joint legal and physical custody arrangements were associated with paternal contact once the other variables were entered into the model. As with the question on whether an agreement contained visitation provisions, the question on contact with the non-resident parent may be difficult to interpret for parents with joint legal and physical custody arrangements. For example, some parents in such arrangements may exclude from their calculation of contact times when the child or children are actually living with the other parent, while other parents may include such times in their estimate of frequency of contact.

Separated Parents Residing in Different Cities or States

For how many disrupted families in the U.S. does the non-residential parent live in a different city or county, or in a different state than the parent with whom the children primarily reside?

Is having the non-residential parent living in another city or county or another state associated with less frequent contact between the non-residential parent and the children?

Is having the non-residential parent living in another city or county or another state associated with payment of less child support by the non-residential parent?

H1 Non-resident parents who live in the same state will be more likely to visit their children and pay child support than parents that live in a different state.


Of the 12.4 million custodial parents, 4.7 million (37.7%) state that the non-residential parent lives in the same city or county as they do ( see Figure 6). An additional, 2.5 million (19.9%) report that the non-residential parent lives in the same state, though not in the same city or county and 2.4 million (19.4%) say that the non-resident parent lives in a different state. Approximately 40,000 report that the non-resident parent is no longer alive (.3%) and 152 thousand (1.2%) say the non-residential parent lives some other place. Disturbingly, an estimated 2.7 million (21.4%) custodial parents have missing information on this question. Reasons for having missing information include not knowing the whereabouts of the non-resident parent or refusing to answer the question. Given that nearly 2 million custodial parents who say that the non-resident parent is alive report that "there is no way of contacting the other parent," it is likely many do not know the whereabouts of the non-resident parent.

As others have found, as distance increases, contact decreases. The proportion of non-residential parents who had any visits with their children in the previous year was 75.2 percent for those living in the same city or county, 69.9 percent for those living in the same state, and 59.8 percent for those living in a different state ( see Figure 7). The mean days of contact also decreases as distance increases. The biggest change occurs as the non-resident parent moves out of the same city or county as the child. Though parents living in the same state see their children more than those living in a different state, the difference in mean days of contact is not large. The mean days of contact by the non-residential parent was 45 days for parents living in the same city or county, 27 days for parents living in a different county or city, but the same state, and 24 days for parents living in a different state.

As we saw in the multiple regression analyses shown in Table 10, the importance of the whereabouts of the non-resident father on contact with the children remains even after controlling for race and ethnicity, education, age, and marital status of the custodial parent and years elapsed since the agreement was formed. Not surprisingly, the influence of how far away the non-resident parent lives on contact weakens once the frequency of previous contact is added into the model.

Payment of child support is more likely to occur if the non-resident parent lives nearby. Among custodial parents with written awards, 83.7 percent report receiving at least some child support if the non-resident parent lives in the same city or county ( see Figure 8). The proportion drops to 77.1 percent who report receiving some child support if the non-resident parent lives in the same state, but in a different city or county. The proportion drops even further to 69.3 percent who report receiving some child support if the non-resident parent lives in a different state. As we will see below, this relationship between how far away the non-resident parent lives and payment of child support disappears once demographic, custody, and contact information are controlled for in multiple regression models. This result should not be surprising. As non-resident parents live further away from their children, the difficulties of making travel arrangements to visit, the costs associated with traveling, and the time needed to make visits all place constraints on the ability of non-resident parents to see their children. However, none of these constraints exist for mailing a child support payment. Thus, one would expect that distance would have a weaker influence on payment than on visitation. There might be some effect due to greater difficulty of enforcing child support compliance across state boundaries, however. Apparently this effect is not substantial, at least in the bivariate relationships.

Child Support

How much money do non-residential parents in the U.S. pay for the support of their children?

What are the factors that lead to greater or lesser payment amounts?

What are the factors that are associated with compliance with child support orders?

Is payment of more child support by non-residential parents associated with greater well-being for the family in which the child primarily resides?

H1 As others have found, many non-residential parents will pay nothing towards the support of their children and many others will pay only part of what they owe.
H2 Custodial parents who are more educated and who are older will be more likely to receive any child support and to receive the full amount that is due compared to other custodial parents.
H3 Families who receive child support will experience greater economic well-being than those who do not receive child support.

The SIPP only asked for amounts of child support received for custodial parents who reported that they had a written child support award and that support was due in the past year. According to these parents, half (50.6%) of the non-resident parents paid the full amount of child support that they owed in the previous year ( see Figure 9). Just over a quarter (27.1%) paid part of what they owed. And, just under a quarter (22.3%) paid nothing at all even though they owed child support. The amounts that were paid, however, were not large. The median amount of child support paid by non-resident parents who paid the full amount owed was $2600, while the median amount for those who made partial payments was $1000. As we shall see later, nearly half of all custodial parents never establish child support awards and the vast majority of those who do not establish awards receive no assistance from the non-resident parent.

Contrary to our expectation, age of the custodial parent had very little association with the receipt of child support (see Table 11). However, as we expected, custodial parents with college educations were more likely to receive child support (84.6%) than were custodial parents with less than a high school education (73.2%) or than those with just a high school education (77.6%). They were also more likely to receive child support than custodial parents who had some college experience (77.8%). Parents who were college graduates were also the most likely to receive the full amount that they were owed in the previous year. Sixty-one percent of college graduates received the full amount owed compared with 46.3% of those with less than a high school education, 49.2 percent of those who were high school graduates, and 50.9% of those with some college experience.

Compliance with child support orders varies by characteristics of the custodial parent. Custodial parents with written child support agreements who were owed child support in the preceding year received about 65 percent of what they were owed ( see Figure 10a). Custodial parents who were in their first marriage received about 70 percent of what they were owed, while never married parents with written agreements received about 60 percent of what they were owed. College graduates were the most likely to experience high compliance: they received approximately 78 percent of the child support monies that they were owed in the previous year.

Table 11 p. 1

Table 11 p. 2

Compliance with child support orders also varies somewhat by the characteristics of custody and child support agreements and where the non-resident parent lives ( see Figure 10b). Compliance with child support orders among persons with joint custody arrangements is higher when the arrangement is joint legal custody as opposed to joint legal and physical (78% of what is owed is received compared to 71%). However, compliance is higher in either kind of joint custody arrangement compared to sole custody arrangements with visitation provisions (65% of what is owed is received) or to sole custody arrangements with no visitation provisions (57% of what is owed is received). Compliance is also higher if the agreement was reached voluntarily as opposed to being court-ordered (73% of what is owed is received compared to 62%).

As we expected, custodial parents who received the full amount due were financially more secure than those who received only partial payments and than those who received no child support (See Table 12). Just under one-third (31.3%) of custodial parents who received the full amount of child support due experienced one or more months of poverty the previous year compared to 41.7% of those who received only part of what was due, and 40.8 percent of those who received none of what was due. Custodial parents receiving the full amount of child support due were also less likely to have received AFDC in the previous year. Approximately 13 percent of custodial parents receiving the full amount of child support owed received AFDC the previous year compared to 19 percent of those receiving partial payments and 27 percent of those receiving none of what was owed them.

Regression models were used to examine which factors are predictive of child support and to examine the interrelationship between visitation and payment of child support. Three different dependent variables were used. The first was whether the non-residential father paid any child support in the year prior to Wave 6 of the survey. The second was the total number of dollars paid in the year prior to Wave 6 by fathers who paid any support during that year. The third was the ratio of the actual dollars paid to the amount that the father was supposed to pay. The third dependent variable can be viewed as measuring the extent to which non-resident fathers complied with the terms of the child support agreements. As with contact, five models were estimated for each of these dependent variables. The first model contained a set of background variables including years elapsed since the agreement was made; where the non-resident father lives; the custodial mother's education, race/ethnicity, age and marital status; the number of children covered by the award; and the age of the youngest child in the custodial mother's household. The second model contained these background characteristics plus information on the custody arrangement and how the child support agreement was reached. The third model added information on previous contact. The fourth model added information on previous child support paid (or extent of compliance with the child support award). And the fifth model contained all the predictor variables contained in the previous models. Table 13 shows the results of these models for any child support paid (as opposed to none). Table 14A shows the results for the dollar amount of child support paid by the father in the previous year and Table 14B shows the models predicting to the ratio of amount paid to amount due.

Table 12

In each case, the strongest predictor of child support was previous child support, i.e., the payment behavior of the father in the year prior to Wave 3 of the SIPP. This was so whether the dependent variable was any payment, amount paid, or degree of compliance with child support order. The addition of the prior support term tended to overwhelm the other predictors, presumably because those same factors worked to shape that prior support behavior as well. Hence, prior support captured and summarized the effects of several of the independent variables. When prior support was omitted from the equation, here are the factors that predicted to whether or not the father paid any support:

Factors That Predict to Payment of Any Child Support (Other than prior support)

Years elapsed since support agreement established (negatively related)

Frequency of visitation in previous year (positively related)

Mother's education level (positively related)

Mother currently married (positively related)

Court-ordered as opposed to voluntary support agreement (negatively related)

Factors that Approach Significance

× Joint physical and legal custody arrangement (positively related)

× Non-residential parent lives in different state (negatively related)

Table 13 p. 1

Table 13 p. 2

Table 14a p.1

Table 14a p. 2

Table 14b p. 1

Table 14b p. 2

Factors That Might Be Expected to Predict, But Do Not

Sole custody without visitation

Joint legal custody

Parents never married

Mother's age


Although frequency of visitation was positively associated with paying some child support, the association was considerably weakened when prior support was included in the model. Furthermore, the addition of visitation frequency to a model that already contained prior support did not enhance the overall predictive power of the model.

Here are the factors that were and were not associated with the amount of child support paid by those fathers who paid any (again, leaving out prior support, which tended to be the strongest predictor):

Factors That Predict To Amount of Child Support Paid By Those Who Paid Any

(Other than prior amount paid)

Number of children covered by agreement (positively related)

Mother's education level (positively related)

Years elapsed since support agreement (negatively related)

Mother is young (under 35 years of age versus age 35 or older) (negatively related, e.g., mothers under 35 receive less child support than mothers 35 years of age and older)

Joint legal custody arrangement (positively related)

African-American family (negatively related)

Court-ordered as opposed to voluntary support agreement (negatively related)

Factors that Approach Significance

Youngest child is under 3

Factors That Might Be Expected To Predict, But Do Not

Parents never married

Sole custody without visitation

Joint legal and physical custody

Non-residential parent lives in different state

Frequency of visitation in previous year

Note that several factors were significant as predictors of amount paid that were not significant with some payment versus none. These were the number of children covered in the agreement, and the mother's age and race. These seem relevant to the size of the award and the father's financial ability to pay (assuming that younger mothers tend to have been married to younger fathers, and that racial minority mothers usually had mates who were also black). On the other hand, the fact that the mother was currently (re)married, which was related to the fact of child support payment, was not related to the amount of that payment.

Here are the factors that were or were not associated with the ratio of the amount paid to the amount of child support owed under the agreement:

Factors That Predict Ratio of Child Support Paid to Amount Owed

(Other than prior compliance ratio)

Years elapsed since support agreement (negatively related)

Frequency of visitation in previous year (positively related)

Mother's education level (positively related)

Mother is young (under 35) (negatively related)

Mother currently married (positively related)

Court-ordered as opposed to voluntary support agreement (negatively related)

Joint custody arrangement (positively related)

Child is under 3 or 18 and over (negatively related)

Factors that Approach Significance

African-American family (negatively related)

Factors That Might Be Expected To Predict, But Do Not

Sole custody without visitation

Parents never married

Many of the factors that predicted to either the fact or the amount of child support payment also predicted to the degree of compliance with the child support award. A notable exception was the number of children covered by the award, which predicted to the amount of child support paid, but not to the fact of payment or to the degree of compliance. Another important exception is the association between paternal contact and the payment of child support. As contact with the father increased in the previous year, the likelihood that any child support would be paid and the degree of compliance with child support awards also increased. The amount of contact with the father in the previous year, however, had no association with the dollar amount paid by fathers who were already paying child support. These results highlight the importance of treating payment of any child support, compliance, and dollar amount paid as distinct factors.

Although we expected that being never married would be negatively associated with compliance and with dollar amount paid, it was not significantly different from being separated or divorced in any of the models. Because these models were restricted to women with written agreements, it may be that the never married with written agreements are different from never married women without written agreements and are more similar to divorced or separated women. We explore this possibility later in the chapter.

From the four sets of regression equations discussed thus far, some preliminary conclusions about the interrelationship of visitation and child support can be drawn. First, past history of contact and past history of payment are the strongest predictors of current contact and current payment of child support. Fathers who visit and who have paid child support in the previous year, tend to do so in the current year.

Second, the models suggest that, at least in the short run, payment of child support has a greater influence on contact than contact has on payment of child support. In the models predicting to visitation, the degree of compliance with the child support order in the previous year remains a significant predictor of visitation even after controlling for past history of visitation. On the other hand, in the models predicting to child support, once a father's previous history of payment of child support is entered into the model, the frequency of contact in the previous year has no influence on compliance with child support orders or the dollar amount of child support paid, and only a marginal influence on whether any child support is paid at all. The fact that contact appears to have a weaker influence on payment of child support than child support has on contact suggests that efforts to increase contact will not necessarily increase compliance with child support orders or the amount of child support paid, though it may have a small influence on whether any child support is paid at all. However, efforts to ensure compliance with child support orders could lead to more contact. More studies need to be conducted to determine whether this pattern of associations is long-term. We also need studies on how the specific provisions of visitation agreements might affect payment and compliance. The SIPP data source provides us with the information on award amounts due and paid, but there is no information on the specific custody and visitation arrangements.

Custodial Fathers/Non-Residential Mothers

How do the demographic characteristics of custodial fathers differ from those of custodial mothers?

What are the economic and living conditions of custodial father households?

How often do non-residential mothers visit their children?

How regularly do non-residential mothers pay child support?

H1 Custodial fathers will tend to be older and better educated than custodial mothers.
H2 Custodial fathers will be less likely than custodial mothers to experience economic difficulties.
H3 Non-residential mothers will be more likely to visit their children than non-residential fathers.
H4 A smaller proportion of non-residential mothers will be obliged to pay child support and the amounts they pay will be smaller than non-residential fathers.

Custodial fathers represent a small, but growing proportion of custodial parents. As noted earlier, there are an estimated 1.5 million custodial fathers representing 12 percent of all custodial parents. Custodial fathers are distinctively different from custodial mothers (see Table 15). A greater proportion of them had once been married to the non-resident parent compared to custodial mothers (86.4% versus 67.2%). They are also much more likely to have graduated from college than custodial mothers (16.4% versus 9.6%). They are older on average than custodial mothers: 49.1 percent are aged 35-44 and 23.1 percent are 45 or older compared to 36.6 percent of custodial mothers who are aged 35-44 and 9.2 percent who are 45 or older. They are more likely to be non-Hispanic whites and less likely to be non-Hispanic blacks than custodial mothers. Nearly 80 percent are non-Hispanic whites compared to 61 percent of custodial mothers. Only 10 percent are non-Hispanic blacks compared to 25 percent of custodial mothers. The youngest child in their households are also older, on average, than the young children in custodial mother households. In 51.1 percent of custodial father households, the youngest child is age 12 or older compared to only 28.4 percent in custodial mother households.

Table 15

Not surprisingly, custodial father households are more advantaged economically than custodial mother households (see Table 16). They are less likely to have experienced poverty in the previous year than custodial mother households (26.5% versus 45.2%). They are also less likely to have received AFDC in the previous year than custodial mothers (5.2% versus 25.7%). They are more likely to own their own homes (66.0% versus 49.4%) and to have higher median family incomes ($30,023 versus $21,440) than custodial mothers.

As shown in Table 17, a smaller proportion of non-residential mothers were supposed to pay child support in the previous year compared to non-residential fathers (71.2% versus 93.4%). Given the better financial position of custodial fathers, agreements may be less likely to require the payment of child support payments from mothers. The median amounts paid by non-resident mothers is less than non-resident fathers ($1800 versus $2100). Non-residential mothers and fathers are about equally likely, however, to make their payments regularly all the time (50.1% of mothers and 48.8% of fathers do so). A similar proportion also are never regular in paying their child support (25.8% of mothers and 24.6% of non-resident fathers). Non-resident mothers appear somewhat more likely to make their payments within 30 days of the due date all or most of the time (68.7% for mothers versus 61.5% for fathers).

According to the reports of custodial parents, non-resident mothers are more likely to visit their children and to see them more often than non-resident fathers ( see Figure 11). Only 16 percent of non-resident mothers had not visited their children in the previous year compared to 32 percent of non-resident fathers. Moreover, 35 percent of non-resident mothers were able to see their children once a week or more often compared to only 24 percent of non-resident fathers. The mean days of contact by non-resident mothers in the previous year was 52 days compared to 33 days by non-resident fathers. The median days of contact by non-resident mothers was 19 days compared to 12 days by non-resident fathers.

Table 16

Table 17

Custodial Parents Without Child Support Awards

How many custodial parents do not have child support awards?

How do custodial parents without awards differ from parents who have awards?

To what extent do custodial parents without awards receive either cash or in-kind services from their former partners?

How do custodial parents without awards who receive assistance from their former partners differ from those who receive no assistance?

H1 Parents without awards will be more likely to be never married, to have lower educations, to be black or Hispanic, and to have not worked in the previous year.
H2 Parents without awards will be more likely to have been in poverty at least part of the time in the previous year, to have received AFDC at some point in the previous year, and to have lower incomes than custodial parents who have had awards. They will also be more likely to live in unsafe neighborhoods, and to experience difficulties meeting basic needs.
H3 Although some parents without awards will report receiving support from their former partners in the form of cash or services, the majority will report receiving no support.
H4 Among custodial parents with no child support agreement, non-resident parents will be more likely to make some child support payments or provide non-cash items or services if the parent lives in the same state.

Just over half of custodial parents (54.4% or 6.7 million) have had a child support agreement at some point in time. A large minority, 45.6 percent or 5.7 million, have never had a child support agreement of any type.

Parents who have never established child support awards differ in many respects from parents who have ever had a child support award. Parents without awards are much more likely not to have been married to their children's other parent. Nearly 48 percent of custodial parents without awards report that they have never been married or that they are currently in their first marriage, which suggests that they have children from a non-marital union. Among custodial parents who have ever had an award, only 16 percent report that they have never been married or that they are currently in their first marriage.

Parents without awards tend to have less education, to be younger, and to have not been employed at all in the previous year compared to parents who have ever had an award ( see Figure 12). They are also more likely to be black or Hispanic than custodial parents with awards.

Given the differences in demographic characteristics, it should be expected that there will be differences in the economic situation of custodial parents with and without awards. As can be seen in Figure 13, this is the case. Custodial parents without awards are much more likely to have experienced one or more months of poverty in the previous year than custodial parents with awards, 51.8 percent versus 35.5 percent. They are also more likely to have received AFDC for one or more months in the previous year, 30.8 percent compared to 16.9 percent. Their family incomes are also lower. The median family income for custodial parents without a child support award was $18,465 compared to $25,583 for custodial parents with awards.

Custodial parents without awards also appear to experience less support and more change in their lives. Nearly 20 percent of custodial parents without awards reported that they had moved into their current home in the past year compared with 15 percent of custodial parents with awards. Although most custodial parents report that they can call on family or friends living nearby to help them, those without awards are less likely to give this response. Sixty-two percent of custodial parents without awards said that they could obtain help from family when in need compared to 70 percent of custodial parents with awards. Similarly, 56 percent of custodial parents without awards said that they could receive assistance from friends compared to 63 percent of custodial parents with awards.

Those without awards are less likely to report owning their own homes, 44.2 percent versus 57.4 percent. They are also more likely to report that they lack at least one essential consumer durable such as a refrigerator, stove, or telephone: 16.4 percent versus 10 percent.

The unweighted sample size for custodial parents with a non-written agreement was less than 200, but a comparison (not shown) of them and custodial parents with written agreements indicates that the two groups have similar demographic and economic characteristics.

Some custodial parents without awards, report receiving child support or non-cash items or services from the non-resident parent. Of the 6.2 million custodial parents without awards, 849 thousand or 13.6 percent report that the non-resident parent provided child support or non-cash assistance in the previous year. The receipt of such assistance, like those with written agreements, is associated with whether the non-resident parent lives nearby or not: 13.1 percent of custodial parents without an award state that the non-resident parent provided some child support or non-cash items or services if the non-resident parent lived in the same city or county, compared to 10.1 percent if the parent lived in the same state, and 3.8 percent if the parent lived in a different state.

Custodial parents without awards who receive some assistance are less likely to be never married than those who receive no assistance (26.6% versus 34.8%). They are also more likely to have some college experience or to have graduated from college than those who receive no assistance (40.4% versus 32.0%). Their youngest child is also more likely to be under six than those who do not receive support (52.0% versus 43.0%). The youngest child is not necessarily the child of the non-resident parent, though it may be.

Not surprisingly, custodial parents without awards who receive some assistance from the non-resident parent are less likely than those who receive no assistance to experience poverty in the previous year (44.2% versus 51.2%) or to have received AFDC in the previous year (20.9% versus 30.6%). They are also somewhat more likely to own their own homes (49.2% versus 44.6%). They are less likely to have moved into their homes in the last year (14.7% versus 19.2%).

Separated Parents Who Have Never Been Married to Each Other

What proportion of custodial parents who were not married to their children's other parent report having a child support award?

How do custodial parents who have not married their former partner, but establish awards differ from those who do not establish awards?

H1 The majority of custodial parents who have not been married to their former partners will not establish child support awards.
H2 Such parents who do establish awards will tend to be more educated, to be currently married, and to have worked in the previous year.

Most custodial parents were at one time married to their child(ren)'s other parent, though 8.2 percent report that they are in their first marriage, suggesting that the children are from an earlier unmarried union, and 22.2 percent report that they have never been married. Thus approximately 3.8 million custodial parents have children from non-marital unions with a parent living elsewhere.

Ninety-five percent of custodial parents who had not been married to their former partner are women. In contrast, approximately 85 percent of custodial parents who had once been married to their former partner are women. In the remainder of this section, we restrict our attention to custodial mothers who had not married their former partner.

Of the 3.6 million custodial mothers who had never married their former partner, 30.3 percent have ever established a child support award either written or unwritten. One-quarter of these 3.6 million women had established written agreements.

Because of the relatively small sample size of women who never married the non-resident parent who ever established awards (n=272), a detailed profile of their demographic characteristics is not possible. However, a few generalizations about differences between those who have ever had an award and those who have never had an award are possible. Non-Hispanic whites and women with at least a high school education are more likely to establish awards than are blacks, Hispanics, or women with less than a high school education, though the differences are often not large. Thirty-six percent of non-Hispanic white women who had not married the non-resident father were able to establish awards compared to 25.8 percent of black women who had not married the non-resident father. Approximately one-third of women with at least a high school education were able to establish awards compared to 26.3 percent of women with less than a high school education. Currently married mothers were slightly more likely to have established an award than never married mothers (32.9% versus 29.4%), which may reflect more about unmeasured characteristics of the women than about the salutary influences of marriage. A mother's age had almost no association with establishing an award, though women aged 25-34 appeared a bit more likely to have ever had an award compared to women who were younger or who were older. Interestingly, women who had not worked at all in the previous year, but who had looked for work (n=75) were more likely to have ever established awards than women who had worked or than women who hadn't worked and who hadn't looked for work. Approximately 45% of women who had looked for work had ever established awards compared to about 30 percent of those who worked and 27 percent of those who didn't work and didn't look for work. Because of the small sample size, not too much emphasis should be put on this result.

Children's Well-Being

Do custody, child support payments, and paternal involvement affect children's living conditions, economic circumstances, educational attainment, and overall health.

What is the relationship between payment of child support and the economic dependency of children's families as measured by receipt of AFDC?

H1 Children will fare the best if they are under a joint physical or legal custody arrangement, an arrangement that was arrived at voluntarily, and if the father maintains regular contact with the child. Children will fare the worst if the custody arrangements were court-ordered and the father maintains little or no contact.
H2 Consistent with other studies, children will fare better economically if their fathers pay child support and maintain contact with them.

Earlier we saw that a family's living conditions and economic well-being varied by the type of custody arrangement and by whether the child support agreement was court-ordered or voluntary. Figures 14 and 15 summarize some of the observed associations for two measures of living conditions: (1) whether the family did not pay the full amount of their utility bills for one or more months in the previous year and (2) whether the family owns a personal computer. Inability to regularly pay monthly utility bills is an indicator of the economic uncertainty faced by households. Economic concerns contribute to worry and stress experienced by custodial parents, which, in turn, can be detrimental to children's well-being (Elder et al., 1995). Ownership of a personal computer is to a large extent a reflection of a family's economic situation, but it is also an indicator of the educational environment of children's households.

As hypothesized, families with joint custody agreements and families with contact with the non-resident parent are less likely to experience economic uncertainty and are more likely to own a computer than are families with sole custody arrangements or those with no contact with the non-resident parent. Award status has no association with a family's economic uncertainty ( Figure 14), but families who have ever had a written award are more likely than those who have not to own a personal computer ( Figure 15). Such a pattern makes sense in that households with awards tend to have higher incomes and will, therefore, be better able to pay for luxuries such as a personal computer. Higher incomes, however, are not necessarily linked with how well families are able to budget their incomes. Similarly, receipt of child support is linked to the inability to pay utility bills regularly, but has no association with ownership of a personal computer. It is precisely the households which experience the greatest uncertainty in their monthly budgets (those who receive only partial payments), who are most likely not to have paid the full amount of their utility bills in one or more months in the previous year. No matter what a family's income level, fluctuations in monthly income make it more difficult to regularly meet basic living expenses. Whether one receives the full amount, partial amount, or no payment, however, is different from the size of one's income and therefore is not linked to ownership of a personal computer.

Children's living conditions are not the only aspects of children's lives that are affected by characteristics of child support agreements, custody arrangements, and visitation. As Figures 16 and 17 show, many of these factors are also related to children's likelihood of attending college (among 19-20 year olds) and health (among children aged 15 and older)(4). Approximately 56 percent of 19-20 year olds whose parents had a written custody agreement have had college experience compared to 27 percent of those whose parents have never had a child support award. Children whose parents had a written award are also somewhat more likely to be described as being in excellent health with no disabilities than children whose parents had no award (48.5% versus 41.7%).

Although the cases become small, having a joint custody arrangement or having a voluntary child support agreement appears to substantially increase the chances that a 19-20 year old will have some college experience. Approximately 75 percent of 19-20 year olds whose parents had a joint custody arrangement and more than 70 percent whose parents had a voluntary child support agreements had some college experience compared to 55.8 percent of all youth covered by a written agreement. Attendance at college also increases if the full amount of child support was paid compared to only part or none. Contact with the non-resident father, however, has no association with college attendance.

Joint custody arrangements and voluntary agreements are also associated with reports of better health among children: 59 percent of children in joint custody arrangements are reported to be in excellent health with no limitation compared to 45 percent of those in sole custody arrangements with visitation and 49 percent of those in sole custody arrangements without visitation. Similarly 54 percent of youth whose parents had a voluntary written agreement were reported to be in excellent health with no limitation compared to 46 percent of youth whose parents had a court-ordered agreement. Contact with the non-resident father is associated with better reports of health. We suspect that reports of health contain a psychological or emotional component, as well as being an indicator of physical health. One reason for this belief is the bivariate associations between health and contact and health and custody arrangements. As noted above, youth whose parents had a sole custody arrangement with visitation provisions were less likely to be rated in good health than those in sole arrangements with no visitation, but contact itself is associated with better ratings of health. Some of the non-resident parents will not see their children even though they have visitation provisions. It is possible that either the custodial parent or the children themselves or both would experience greater disappointment which could color their responses to assessments of the child's health when visitation is expected but does not occur than when it is not expected.

Logistic regression models were used to examine the influence of demographic factors, custody and visitation arrangements, visitation, and payment of child support on the economic situation of the children's families and on the children's health. As with the adult models, a series of regressions were run. The first model contained a set of background variables including years elapsed since the agreement was made, where the non-resident father lives, and the custodial mother's education, race/ethnicity, age and marital status, the number of children covered by the award, and the child's age. The second model added information on custody arrangements and how the child support arrangement was reached. The third, and last model, added information on visitation and payment of child support.

Table 18 shows the models predicting to whether the children's families received AFDC in the previous year. The observed relationships are summarized below.

Factors That Predict To Welfare Dependence

Custodial parent currently (re)married (negatively related)

Education level of custodial parent (negatively related)

Age of mother (negatively related)

African-American family (positively related)

Court-ordered as opposed to voluntary support agreement (positively related)

Amount of child support paid in previous year (negatively related)

Number of children covered by support agreement (positively related)

Years elapsed since support agreement (positively related)

Joint legal custody arrangement (negatively related)

Parents never married (positively related)

Age of youngest child (negatively related)

Nonresidential parent lives in different city or county of same state (positively related)

Hispanic family (positively related)

Table 18 p. 1

Table 18 p. 2

Factors That Might Be Expected To Predict That Do Not

Frequency of visitation by nonresident parent in previous year

Nonresident parent lives in different state

Sole custody without visitation provision


Even after controlling for demographic characteristics of the custodial parent, how far away the non-resident parent lives, years elapsed since the separation, and information on custody, child support, and visitation, having a court-ordered as opposed to a voluntary agreement increases the likelihood of being on AFDC. Not surprisingly, receiving higher amounts of child support reduces the likelihood of receiving AFDC. Contact, however, has no association with receipt of AFDC, contrary to our expectations. This is in keeping, however, with our earlier finding that contact has no effect on the dollar amount of child support paid. Contact appears to be linked less to income and income-related factors than it is to other aspects of children's lives, as we shall see below.

Table 19 contains the models predicting whether youth were in excellent health with no limitation. The observed relationships are summarized below.

Factors That Predict To Youth In Excellent Health with No Limitations

Frequency of visitation by nonresident parent in previous year (positively related)

Custodial parent has less than a high school education (negatively related)

Nonresident parent lives in different state (positively related, but only significant when frequency of contact included in model)

African-American family (negatively related)

Years elapsed since support agreement (positively related)

Table 19 p. 1

Table 19 p. 2

Factors that Approach Significance

Custodial parent currently (re)married (negatively related)

Other race family (negatively related)

Factors That Might Be Expected To Predict That Do Not

Amount of child support paid in previous year

Joint legal custody arrangement

Sole custody without visitation provision

Court-ordered as opposed to voluntary support agreement

With respect to children's assessed health, how an agreement was reached, type of custody arrangement, and amount of child support paid have no effect on health. However, contact with the non-resident father is associated with a greater likelihood of being in excellent health with no limiting condition, even after controlling for all the other variables. As noted above, we suspect that the health item is capturing at least in part psychological or emotional well-being, as well as physical health. It appears that father-child contact does promote greater child well-being even though it may not contribute to economic well-being.

SUMMARY OF FINDINGS OF REGRESSION ANALYSES

In this section, we summarize the results of a series of multivariate analyses that examined the relationships between custody and support arrangements, on the one hand, and the subsequent behavior of non-residential parents, on the other. The custody and support variables included whether the support agreement was voluntary as opposed to court ordered, whether it had an explicit visitation provision, and whether there was a joint custody arrangement (either joint legal custody only, or joint physical custody as well). Other predictor variables examined in the regressions were the frequency of contact by the nonresidential parent in the past year, support paid by the parent in the past year, the number of years that had passed since the support agreement was issued, and whether the nonresident parent lived in another state or city at the time of the survey.

The measures of behavior of the nonresidential parent that served as dependent variables in the regressions were the frequency of contact with the children in the current year, whether the parent provided any child support in the current year, the dollar amount of support provided by parents who provided any support during the year, and the ratio of the amount paid to the amount owed under the support order (compliance ratio). Two measures of family well-being were also used as dependent variables. One was freedom from welfare dependence (whether or not the family had received any welfare payments (AFDC) between Wave III and Wave VI of the SIPP). The other was an indicator of the health status of older teens and young adult children in disrupted families (whether they were in "excellent" health with no health condition that limited their daily activities).

The results of the analyses are presented in summary form in Table 20 below. They are described in more detail in the following pages of the report.

Table 20. Relationships Between Custody and Support Arrangements and Behavior of Non-residential Parents and Family Well-being Indicators

(Multiple Regression Analyses)

Predictor Variables

Behavior of non-residential parent

Well-being indicators

Frequency of contact

Probability of support

Amount of support

Compliance ratio

Freedom from welfare dependence

Health status of youth

Voluntary support agreement

Y (+)

Y (+)

Y (+)

Y (+)

Y (+)

N

Visitation provision

Y (+)

N

N

N

N

N

Joint legal custody

N

N

Y (+)

N

Y (+)

N

Joint legal & physical

Past contact

N

Y (+)

M (+)

Y (+)

N

N

Y (+)

Y (+)

N

N

N

Y (+)

Past support

Y (+)

Y (+)

Y (+)

Y (+)

Y (+)

N

Elapsed time

Y (–)

Y (–)

Y (–)

Y (–)

Y (–)

Y (+)

Residence in other state

Residence in other city

Y (–)

Y (-)

M (–)

N

N

N

M (–)

N

N

Y (-)

Y (+)

N

Key:Y = significant relationship between predictor and dependent variable (p < .05)

M = marginally significant relationship (.05 < p < .10)

N = no significant relationship

(+) positive relationship between predictor and dependent variables

(–) negative relationship between predictor and dependent variables

Voluntary versus court-ordered agreements. The distinction between voluntary child support agreements and support orders that were imposed by a court proved to be an important one as far as predicting to the future behavior of nonresidential fathers was concerned. Fathers who had voluntarily agreed to make regular financial contributions toward the care of their children maintained more frequent contact with those children than fathers who had child support obligations imposed upon them. Fathers in the former group were more likely than those in the latter to pay some support, to pay larger amounts of support, and to comply more fully with the terms of the support order. Mothers who reached voluntary support agreements with their former husbands were less likely to be welfare dependent than mothers who had to obtain court-ordered child support. Voluntary agreements were not associated with better (or worse) health for young adults from disrupted families, however.

Visitation provision in child support agreement. Nonresidential fathers had significantly more contact with their children when the child support agreement had an explicit provision specifying the frequency and schedule of visitation than when such a provision was not present in the agreement. On the other hand, the presence of a visitation provision was not associated with a greater likelihood of paying child support, nor with the payment of greater amounts of support, nor with better compliance with child support orders. Having a visitation provision was associated with a reduced likelihood of welfare dependence, but it was only significant at the .10 level. The presence of a visitation provision did not predict to better health for youth from disrupted families.

Joint custody arrangements. The distinction between joint custody and sole custody arrangements also proved to be important for predicting to the future behavior of nonresidential fathers. Fathers with joint legal and physical custody were more likely to pay some support and to comply more faithfully with the terms of the support agreement. Fathers with joint legal only arrangements paid larger amounts of child support than fathers in sole custody arrangements where the mother had custody. Mothers who had entered into joint legal custody arrangements were less likely to be welfare dependent than those with sole custody.

On the other hand, fathers with joint custody arrangements, regardless of the type, did not have significantly more contact with their children than when the mother had sole custody but the father had explicit visitation rights. (Both fathers with joint custody and those without joint custody but with visitation provisions saw their children more frequently than when the mother had sole custody and there was no visitation provision.) As noted earlier, however, it is not clear how respondents interpreted the question about contact when the parents had joint physical custody. There was no significant association between the presence of a joint custody arrangement and the health status of youth from disrupted families, once other related factors were taken into account.

Past contact. The frequency with which nonresidential fathers saw their children in the previous year was the strongest predictor of how much contact they would have in the current year. Frequency of contact in the recent past was also a significant, though much weaker predictor of whether the father would pay any child support in the current year, and how well he would comply with the child support schedule. Frequency of past contact did not predict to how much support the father would pay, among those who paid anything.

In addition, once past child support was entered into the regression analysis, past contact added little or nothing to the prediction of current child support. Nor was more frequent contact associated with less welfare dependence, once related factors were taken into account. However, more frequent contact by the nonresidential father was associated with better health status among older teenagers and young adults from disrupted families.

Past child support. The payment of child support in the previous year was by far the strongest predictor of payment of child support in the current year. This was so whether the predicted variable was payment of any support, amount paid among those who paid any, or compliance with child support orders. Past child support was also a significant predictor of the amount of contact the nonresidential father would have with his children during the current year. Even when past contact was entered in the regression as well, past child support significantly strengthened the prediction of current contact. Past support was a stronger predictor of current contact than past contact was of current child support.

The payment of child support in the recent past was associated with a lower likelihood that the custodial mother would be dependent on welfare. This was so even though the average amounts of child support paid by nonresidential fathers were quite modest, averaging just $2,739 per year among custodial mothers with a written agreement who had child support due that year. Payment of child support in the recent past, however, did not seem to make a difference as far as the health status of older children from disrupted families was concerned.

Elapsed time since support agreement. In comparison to the situation that prevailed in the late 1970s and the early 1980s (Furstenberg, Nord, Peterson, & Zill, 1983), there have been significant increases in the average amount of contact that nonresidential fathers maintain with their children following separation, divorce, or childbearing outside of marriage. There have also been gains in the proportion of nonresidential fathers who pay at least some child support during the course of a year, as well as some growth in the average amount of support paid (U.S. Bureau of the Census, 1995; Zill & Nord, 1994). However, it is still the case that both parental contact and child support become less frequent as time goes by.

Thus, even when other factors were controlled, the number of years that had elapsed since the support agreement was reached proved to be a significant negative predictor of how frequently nonresidential fathers would see their children in the current year. Elapsed time was also a predictor of whether the fathers would pay any support during the current year, of how much support they would pay if they paid any, and of how well they would comply with their support obligations. Elapsed time was associated with a greater likelihood of welfare dependence, after related factors such as the age of the mother and the age of the child were taken into account.

Somewhat surprisingly, the number of years that had elapsed since the support agreement was found to be positively associated with the health status of older teen and young adult children in disrupted families, after controlling for custody, child support, and contact. Although it cannot be determined with the SIPP data, the positive association may indicate a diminishing effect of marital conflict on child well-being; as the conflict between parents grows more distant in time, and the family reaches a new equilibrium, the mental health of the child gets better. Such a supposition is supported by research which indicates that children go through a crisis period following a divorce, but that most children rebound psychologically and emotionally within a couple years of the divorce (Morrison and Cherlin, 1995; Chase-Lansdale and Hetherington, 1990).

Residence in another state or city. Just as the passage of time led to less contact between the nonresidential father and the child, so also greater distance between their respective residences was associated with less frequent visitation. Nonresidential fathers who lived in another state saw their children less often than those who lived in the same state, even when other related factors were controlled. The same was true of fathers who lived in a different city or county of the same state, though to a less pronounced degree.

Fathers who lived in another state or city were somewhat less likely to pay child support in the current year than those who lived in the same city and state. The statistical significance of the difference was only marginal, however, at the .10 level. Among fathers who paid any support, those living in another jurisdiction did not pay less support, on average, than those living in the same city and state. Father's residence in another jurisdiction was associated with less compliance with child support obligations, although again the effect was of marginal significance. Father's residence in another state or city was not linked to a greater frequency of welfare dependence for the custodial mother and her children.

Again there was an unexpected finding with respect to the direct indicator of child well-being. Here it was that father's residence in another state was associated with better health in the older teen and young adult children of divorce. However, this relationship did not emerge until the frequency of contact with the father was included in the regression analysis. It will be recalled that teenage children of divorce were reported to be in better health when the nonresidential parent visited often. Thus, the two relationships seem at least somewhat contradictory.

One possible interpretation for this positive relationship between the father living in a another state and the youths health is that fathers who are very involved with their children through the payment of child support and frequent contact, despite living far away, may be showing their greater commitment to their children which contributes to the children's better health and psychological well-being. Another interpretation is that living in another state reduces the opportunity for conflict between the separated parents and the reduced conflict contributes to the children's greater well-being. It is impossible with the SIPP data to determine whether either or both of these interpretations is correct.

CONCLUSIONS AND POLICY IMPLICATIONS

To conclude our investigation of non-custodial parents' participation in their children's lives, we examine the implications of the SIPP findings for various policies aimed at enhancing the functioning of postdivorce families with children and promoting the healthy development of young people who must grow up in these families. Specifically, we discuss the extent to which the SIPP findings support or fail to support the following kinds of policy options:

Obviously, the present study is not a randomized policy experiment nor even a nonrandomized study of specific policy initiatives. Thus, we must be circumspect about how far we go in drawing policy-related conclusions from the findings. As shown above, parents who enter into voluntary agreements differ in several respects from those who have court-ordered agreements. Although our models tried to control for these differences, it is possible that parents with voluntary agreements differ from those with court-ordered agreements in ways that we have not measured. Such unmeasured characteristics could predispose parents with voluntary agreements to pay more support and to maintain more contact with the children. If parents with a different set of characteristics were cajoled into entering into a voluntary agreement, they might not behave in the same manner.

A correlational panel study such as the present one cannot prove that a given policy will work as its advocates contend it should. An observed relationship may be due to the operation of other, unmeasured factors. However, the failure to find an expected correlation can provide firmer grounds for believing that a specific policy will not work as anticipated. With these cautions in mind, let us examine what the data seem to indicate about the viability of the policy approaches outlined above.

Encouraging Voluntary Agreements

To begin with, the results of our analyses seem broadly supportive of attempts to provide mediation and encourage divorcing couples or never married parents to enter into support agreements voluntarily. The multivariate analyses indicate that court-mandated support orders are often associated with more negative outcomes for families, even after controlling for the demographic and socioeconomic characteristics of the parents and other possible mediating factors. Custodial mothers with voluntary as opposed to court-ordered agreements reported significantly more days of contact with the non-resident father, larger amounts of child support received, and better compliance by the father with child support orders. Such mothers were also less likely to have received AFDC in the previous year.

Although having a voluntary agreement had no detectable direct effect on children's health, it did have an indirect effect. Voluntary agreements were associated with more frequent contact by the nonresidential father, and that in turn was associated with better health among older teens and young adult children. Thus, the predominance of the evidence was clearly consistent with the notion that voluntary agreements are something worth striving for. Encouraging parents to consider ways to structure and manage their child care and child support obligations might well give them a sense of ownership and self-direction. This might, in turn, induce them to comply more faithfully with the requirements of the agreement. Of course, couples exhibiting substantial conflict would require mediation or other services before discussing the details of a support agreement.

As mentioned above, the correlational evidence from the present study does not demonstrate that it was the voluntary agreements per se that made the difference in the subsequent behavior of the nonresidential parent. It is possible that there is a third factor or factors that leads parents to enter into voluntary agreements and to maintain contact and pay child support. But at least we can say that there is nothing in the present findings that seriously challenges the argument in favor of voluntary agreements, where such agreements are feasible.

Requiring Explicit Visitation Provisions

Fathers' rights groups and other divorce reformers want to promote the continued involvement of nonresidential parents in their children's lives by requiring that custody and support agreements have explicit provisions specifying the visitation rights and responsibilities of each parent. They have argued that such provisions would help to ensure that noncustodial parents have access to their children and would lead to more ample and frequent payment of child support by these parents.

The findings of the present study support the notion that explicit visitation provisions in child support agreements are linked to more frequent contact between the nonresidential parent and child. Indeed, when other factors were controlled, families with sole custody arrangements that contained visitation provisions evinced as much contact by the nonresidential parent as families with joint legal custody arrangements. As before, however, we cannot say from the present results that it was the visitation provision as such that led to the more frequent contact because we do not know what was in the visitation provision. Moreover, visitation provisions per se were not associated with significantly more frequent or abundant child support payments, nor were they linked with greater freedom from welfare dependence for the custodial parent.

There was no direct linkage found between visitation provisions and better mental or physical health for youth from disrupted families. There was an indirect link, though, inasmuch as visitation provisions were associated with more frequent contact by the nonresidential parent, and more frequent contact was significantly related to better health for older teens and young adults from disrupted families.

In sum, the present study findings are supportive of the call for explicit visitation provisions. The SIPP results suggest that an important first step in maintaining contact between children and their non-resident parents, is to encourage explicit visitation provisions in agreements. However, the existence of visitation provisions will not guarantee that visitation will occur. The SIPP findings also contradict one of the arguments made for such provisions, namely, that they will lead to the payment of child support. If the aim is to encourage nonresident parents to pay what is due or to pay more child support, one must do more than ensure that visitation provisions are specified.

Encouraging and Facilitating Non-residential Parents' Access to Their Children

Existing research indicates that there are a variety of reasons why fathers tend to lose contact with their children over time. These reasons include the psychological and emotional pain experienced by fathers due to their diminished role and the artificial nature of visitation, and lack of access to their children due to actions of their former spouse (Price et al., 1994; Umberson and Williams, 1993).

Although the SIPP data reveal nothing about obstructive actions taken by the custodial parent or the psychological adjustment of the non-residential parent, they do suggest that continued contact with the non-residential parent has a beneficial influence on older teens and young adults. Moreover, contact has a positive association with both the payment of child support and with compliance with child support orders. Although this relationship diminishes with the addition of past child support behavior to the models, the positive sign remains suggesting that contact is not hindering the payment of child support and is actually exerting pressure towards the payment of child support. The SIPP results provide support for efforts that encourage contact between children and their non-residential parents.

Promoting Joint Custody Arrangements

Another way of encouraging both parents to remain involved in their children's lives following a divorce or birth outside of marriage is through joint custody arrangements. In these arrangements, both parents participate in decisions about how their children are cared for and raised, or at least have a legal right to do so. In the case of joint physical custody, the child may actually live with and be cared for by one parent during part of the week, and with the other parent during the rest of the week.

As the SIPP data demonstrate, joint custody is still an arrangement that is implemented by a minority of divorced families. And most of these opt for joint legal custody only, not joint physical custody. Joint custody is considerably more popular now than it was ten or twenty years ago (Furstenberg & Nord, 1985; Kelly, 1994; Seltzer & Meyer, 1996).(5) Does joint custody appear to have the beneficial effects on parental behavior and family well-being that its advocates have claimed? And, does it matter whether the joint custody arrangement is joint legal only or joint legal and physical?

The results of our analyses were broadly supportive of the arguments for joint custody, though the influence of joint custody differed somewhat depending upon whether it was joint legal only or joint legal and physical. The multivariate analyses indicate that nonresidential parents with joint legal and physical custody were more likely to have paid some child support in the current year (significant at the .10 level) and to have complied more fully with their support obligations than non-resident fathers in other arrangements. There was no difference in the probability of paying support or in the degree of compliance with the child support orders between fathers with joint legal only custody arrangements and other fathers. However, fathers with joint legal only arrangements paid larger amounts of support than other fathers. This was the case even after controlling for the demographic and socioeconomic characteristics of the parents and other possible mediating factors. Mothers in joint legal custody arrangements were significantly less likely to have received AFDC in the previous year.

Nonresidential fathers with either type of joint legal custody did not have more frequent visitation with their children than nonresidential fathers whose former wives had sole custody but the father had explicit visitation privileges. However, these three groups had significantly more visitation than nonresidential fathers whose former wives had sole custody and there was no visitation provision in the support agreement. Thus, joint custody does seem to be associated with above-average visitation, though not as much as sometimes thought. As discussed earlier,however, it is not clear precisely what visitation meant for joint custody families, especially in the case of joint physical custody arrangements.

Joint custody was not directly linked with better health for older children from divorced families. Also, as with voluntary agreements, the findings do not prove that the higher rates of child support found in joint custody families were directly attributable to the joint custody arrangements. Nonetheless, we can again say that there is nothing in the findings that seriously challenges arguments in favor of joint custody, where such arrangements are feasible.

Limiting Moves by Divorced Parents

Another policy that has been advocated by those who seek to promote more postdivorce involvement by nonresidential parents in their children's lives is one that would limit the rights of divorced parents to move to another city, county, or state. Such moves are seen as making it more difficult for both parents to maintain contact with the children. Critics on the other side have argued that such a policy would represent an unconstitutional restraint on parents' freedom, and could be harmful to a family's economic well-being as well, by circumscribing the employment and remarriage possibilities of the custodial parent.

The SIPP analyses show that the number of days on which non-resident fathers see their children does indeed decrease as the father's distance from the child increases. This is also consistent with what previous research has shown (Furstenberg and Nord, 1985). The decline appears to be nonlinear in that the greatest decrease occurs when the father lives in a different city or county, even if he remains in the same state.

The father's residence in another state was found to be associated with a lower probability of his paying child support in the current year (though the relationship was only significant at the .10 level), and with less compliance with support obligations. There was no relationship between residence in another state and the amount of support paid by those fathers who paid any.

Research has been inconclusive about the importance of continued contact with the non-resident father for children's well-being. A consistent finding has been that most children want to maintain contact with their fathers. The first several years after a marriage breaks up are often stressful for children, partly because of the loss of contact with one parent, usually the father. The SIPP analyses indicate that older children of divorce fare better, at least with respect to health, the more contact they have with their nonresident fathers. As noted earlier, we suspect that the health measure has a substantial psychological or emotional component.

At the same time, and somewhat paradoxically, the regression analyses indicated that older children with divorced parents in a different state were healthier, on average, than those whose nonresident parents were in the same state and city. This relationship only emerged when a term representing the frequency of contact with the parent was also included in the analysis. We inferred that this association might represent greater motivation on the part of the non-resident fathers who remain involved which is beneficial for their children. Another possibility is that residence in another state might be linked to less opportunity for continued conflict between separated parents.

In sum, our results indicate that physical distance does play an important role in impeding continued parental involvement in children's lives. However, the findings also indicate that it is possible for disrupted families and children to make a positive adjustment when parents live in different cities or states. A policy of strictly limiting parental moves does not seem supported by existing research.

Other Implications of the SIPP Findings

Relationships between contact and support. Many researchers and policymakers are interested in determining the nature of the relationship between contact with the non-resident parent and payment of child support. Two questions are of particular interest. First, will policies that increase non-resident parents' payment of any child support or their compliance with child support orders result in greater contact between them and their children? And, second, will policies that increase contact between non-resident parents and their children result in higher compliance with child support orders? Existing research, generally based on studies that collect information on contact and payment of child support at the same point in time, have yielded inconsistent results (Seltzer, 1995; Seltzer et al., 1989; Veum, 1993). The SIPP findings, which use information collected one year apart, suggest that past payment of child support is more closely linked with future contact between parent and child than past contact is with future payment of support. The SIPP models indicate that contact in the past year has a small influence on whether any child support is paid in the current year and on compliance with the child support order, and no effect on the dollar amount paid by men who are currently paying something. However, once past child support payment behavior is entered into the model, contact has no influence on payment of child support in the current year. Compliance with child support orders in the past year is, however, a significant predictor of the amount of contact that the nonresident parent has with his children in the current year, even when past contact and other factors are taken into account.

These results have at least two policy implications. First, policies that simply aim to increase contact between nonresidential parents and their children cannot be relied on to increase compliance with child support orders. More contact may have a positive effects on child well-being, but may not increase payment of support. On the other hand, more vigorous enforcement of child support obligations may indeed result in more nonresidential parents seeking increased contact with their children. However, we cannot assume that parents who do not currently pay will pay and spend more time with their children because of enforcement efforts. Nor can we assume that if they do pay, that their relationship with their children will be the same as the relationship between children and fathers who currently are paying, especially if enforcement efforts are viewed as harsh or unfair. To the extent that nonresident fathers wind up spending more time in positive interactions with their children and playing a more active role in their lives, this could be a good thing. However, increased contact could bring the possibility of renewed conflict between parents. Further research aimed at identifying when positive or negative consequences should be expected seems warranted.

Families without written agreements. Because the visitation and contact questions were only asked of families with agreements in the SIPP 1990 and 1991 panels, we were unable to examine issues of contact and support for parents who did not have written child support awards. However, from existing research we do know that having an award is an important first step in obtaining any child support at all (Peterson and Nord, 1990).

Nearly half (45.6%) of the custodial parents in the SIPP sample reported that they had never had an agreement of any type. Almost half of these were parents who had never married their former partners. Such parents do not need to go through legal channels to separate, yet nearly a quarter of them had reached written agreements and another 5 percent had non-written agreements. Research that examines why it is that some never-married parents are able to establish written child support awards should be encouraged. Such research might suggest ways to increase the proportion of never married couples with children who establish awards and visitation agreements.

SUMMARY

In brief, the SIPP analyses provide support for the following types of activities:

The SIPP results also provide some support for efforts to encourage joint custody arrangements, though often there was little difference between either type of joint custody arrangement and sole custody arrangements in which visitation provisions had been specified. Differences, when they were observed appeared to be in the payment of child support and not in the amount of contact with the children or in the well-being of children.

The SIPP results do not support efforts to restrict the move of parents to other cities or states. Although contact did decrease as fathers moved farther away from their children, residence in another city had not adverse effect on the likelihood that fathers would pay any child support or their compliance with child support orders. It also had no effect on the dollar amount paid by fathers who were paying some child support. Residence in another city also had no negative effect on the health of teenagers and young adults. Residence in another state, once previous child support behavior was entered into the models, also had no adverse effect on payment of child support or on children.

REFERENCES

Braver, Sanford H., Sharlene A. Wolchik, Irwin N. Sandler, Bruce S. Fogas, and Daria Avetina. 1991. "Frequency of Visitation by Divorced Fathers: Differences in Reports by Fathers and Mothers." Amer. J. Orthopsychiatr. 61(3): 448-454.

Chase-Lansdale, P. Lindsay and E. Mavis Hetherington. 1990. The impact of divorce on life-span development: Short and long term effects. In P.B. Baltes, D.L. Featherman, R.M. Lerner, eds. Life-Span Development and Behavior Vol. 10. Hillsdale, NJ: Lawrence Erlbaum Assoc. Pages 105-150.

Elder, Glen H., Jr., Jacquelynne S. Eccles, Monika Ardelt, and Sarah Lord. 1995. "Inner-City Parents Under Economic Pressure: Perspectives on the Strategies of Parenting." Journal of Marriage and the Family 57(3): 771-784.

Furstenberg, F.F., Jr. & C.W. Nord. (1985). Parenting apart: Patterns of childrearing after marital disruption. Journal of Marriage and the Family, 47(4), 893-904.

Furstenberg, Frank F., Christine Winquist Nord, James L. Peterson, and Nicholas Zill. 1983. The life course of children of divorce. American Sociological Review 52(5): 695-701.

Kelly, J.B. (1994). The determination of child custody. The Future of Children, 4(1), 121-142.

Maccoby, Eleanor E. and Robert H. Mnookin. 1992. Dividing the Child: Social and Legal Dilemmas of Custody. Cambridge, MA: Harvard University Press.

Morrison, Donna Ruane and Andrew J. Cherlin. 1992. The divorce process and young children's well-being: A prospective analysis. Journal of Marriage and the Family 57(3): 800-812.

Notar, Susan A. 1995. An examination of state custody and visitation laws. Paper prepared for the U.S. Commission on Child and Family Welfare, June 1995.

Peterson, James L. and Christine Winquist Nord. 1990. The regular receipt of child support: A multistep process. Journal of Marriage and the Family 52(2): 539-551.

Price, David A., Robert G. Williams, Jessica S. Pearson, and Nancy A. Thoennes. 1994. Child Access Demonstration Projects: Final Wave I Report. Reported submitted to the Office of Child Support Enforcement, U.S. Department of Health and Human Services, Washington, DC, September 30, 1994.

Seltzer, J.A., & Meyer, D.R. 1996. Child support and children's well-being. Focus, 17(3), 31-36.

Seltzer, Judith A. 1995. Will child support enforcement increase father child contact and parental conflict after separation? Paper presented at the conference, The Effects of Child Support Enforcement on Nonresident Fathers, Princeton University, September 13-15, 1995.

Seltzer, Judith A. 1991. Relationships between fathers and children who live apart: The father's role after separation. Journal of Marriage and the Family 53(1): 79-101.

Seltzer, Judith A., Nora Cate Schaeffer, and Hong-Wen Charng. 1989. Family ties after divorce: The relationship between visiting and paying child support. Journal of Marriage and the Family 51(4): 1013-1032.

Umberson, Debra and Christine L. Williams. 1993. Divorced fathers: Parental role strain and psychological distress. Journal of Family Issues 14(3): 378-400.

U.S. Bureau of the Census. 1995. Child support for custodial mothers and fathers: 1991. Current Population Reports (Series P-60, No. 187). Washington, DC: U.S. Department of Commerce, Economic and Statistics Administration.

U.S. Bureau of the Census. 1991. Child support and alimony: 1989. Current Population Reports (Series P-60, No. 173). Washington, DC: U.S. Department of Commerce, Economic and Statistics Administration.

Veum, Jonathan R. 1993. The relationship between child support and visitation: Evidence from longitudinal data. Social Science Research 22(3): 349-360.

Zill, N., & Nord, C.W. (1994). Running in place: How American families are faring in a changing economy and an individualistic society. Washington, DC: Child Trends.

APPENDIX A:

Child Support Module 1990 and 1991 Panels of SIPP

APPENDIX B:

Comparability of SIPP Estimates and Current Population Survey Estimates

Table A-1 compares estimates obtain from the 1990 Panel of SIPP with estimates obtained from the April 1992 CPS.

Table A.1 Comparison of Child Support Data from the Survey of Income and Program Participation (SIPP) and the Current Population Survey (CPS)

(As of October, November, December 1990, or January 1991 in the SIPP. As of April 1992 in the CPS.

Characteristics

SIPP

Winter 1990/91

CPS

Spring 1992

Custodial Parents

Total.............................millions

Female....................................

Male......................................

Custodial Parents with Awards

Total.............................millions...

Female...............................................

Male.................................................

Mean Child Support Received by Persons Receiving Any in Previous Year

Total..............................dollars...

Female...............................................

Male.................................................

12.4

10.9

1.5

6.7

6.3

0.4

2,725

2,739

2,435

11.5

9.9

1.6

6.2

5.5

0.6

2,961

3,011

2,292

Source: CPS Numbers: U.S. Bureau of the Census. 1995. Current Population Reports. Series P90-187. Child Support for Custodial Mothers and Fathers: 1991. U.S. Government Printing Office: Washington, DC, Table A-3. SIPP Numbers: Analysis of SIPP data by authors.

The child support questions asked in these two surveys, although similar, are not identical. The questions used to obtain the above estimates from the two surveys appear below.

Screener Questions:

SIPP: Is...the parent of the children under 21 years of age who live in this household? [If yes, ask next question]

Does...have any children in this household under 21 years of age who have a parent living elsewhere? [Do not include stepparents or parents who would be living at home except for military or other job related absences.]

CPS: Does...have any children under 21 years of age in this household you have a parent living elsewhere?

Type of Award Questions:

SIPP: These next few questions concern child support. Have child support payments ever been agreed to or awarded for (any of) ...'s children living here? [If yes, ask]

In some cases, child support agreements are made and then later modified or revised. The following questions relate to the most recent initial agreement and any subsequent modifications of that agreement.

Was this agreement a voluntary written agreement ratified by the court, a court-ordered agreement, some other type of written agreement, or a non-written (verbal) agreement?

CPS: Were child support payments for ...agreed to or awarded?

No child support order/agreement

Voluntary written agreement

Court award

Other

Child Support Received Questions:

SIPP: [These questions were only asked of persons with a written child support agreement. They refer to the most recent such agreement.]

Were any payments due in the last 12 months? [If yes, ask]

What is the total amount that ...was supposed to have received in child support payments during the past 12 months (from the most recent agreement)?

What is the total amount that ... actually received in child support payments under that agreement, during the past 12 months?

CPS: During calendar year 1991 were you or your child(ren) covered by this award SUPPOSED to receive any child support? [If yes, ask]

In total, how much in child support payments were you SUPPOSED to receive in 1991 from child support agreement for (child(ren) covered by this award?

How much in child support payments did you ACTUALLY receive in 1991 for (child(ren) covered by this award?

As can be seen from the above questions, SIPP asks about child support received in the previous year only of persons with written awards. The CPS, on the other hand, does not ask the respondent to distinguish written from non-written awards. The question on child support received is asked of all persons with any type of award who said that support was supposed to be due in the previous year, regardless of whether the award was written.

(1)See Appendix A for a copy of the child support module asked in both the 1990 and 1991 panels of SIPP.

(2)See Appendix B for a comparison of SIPP and CPS estimates on selected child support variables.

(3)This estimate and the estimate of persons with voluntary agreements depends upon the respondents' understanding of the terms "court-ordered agreement" and "voluntary written agreement ratified by the court." In practice, the distinction between voluntary and court-ordered agreements can be blurred. For example, a voluntary agreement could be reached after a mandatory mediation session.

(4)Although we would have preferred to examine the health of all children, data on health were not available for children younger than 15.

(5)SIPP is the first national survey to distinguish joint legal only custody from joint legal and physical custody. The increase in joint custody over time thus refers to both types of custody arrangements combined.