lable at ScienceDirect

Social Science & Medicine 69 (2009) 76–84

Contents lists avai

Social Science & Medicine

journal homepage: www.elsevier .com/locate/socscimed

Effects of child health on parents’ social capitalq

Jennifer Schultz a, Hope Corman b, Kelly Noonan b, Nancy E. Reichman c,*

a University of Minnesota Duluth, USAb Rider University and NBER, USAc Robert Wood Johnson Medical School, USA

a r t i c l e i n f o

Article history:Available online 13 May 2009

Keywords:Child healthSocial capitalUSAParentsFamilyNew birthSocioeconomic disadvantage

q This research was supported by Grants #R01-HDfrom the National Institute of Child Health and Hgrateful for helpful input from Henry Saffer and Saraassistance from Prisca Figaro and Jessica Fuller.

* Corresponding author. Tel.: þ1 732 235 7977.E-mail address: nancy.reichman@umdnj.edu (N.E.

0277-9536/$ – see front matter � 2009 Elsevier Ltd.doi:10.1016/j.socscimed.2009.04.001

a b s t r a c t

This paper adds to the literature on social capital and health by testing whether an exogenous shock inthe health of a family member (a new baby) affects the family’s investment in social capital. It alsocontributes to a small but growing literature on the effects of children’s health on family resources andprovides information about associations between health and social capital in a socioeconomicallydisadvantaged population. We use data from the Fragile Families and Child Wellbeing study, a longitu-dinal survey of about 5000 births to mostly unwed parents in 20 U.S. cities during the years 1998–2000.Both parents were interviewed at the time of the birth and then again one and three years later. Theinfants’ medical records from the birth hospitalization were reviewed, and poor infant health wascharacterized to reflect serious and random health problems that were present at birth. Social interac-tions, reported at three years, include the parents’ participation in church groups, service clubs, politicalorganizations, community groups, and organizations working with children; regular religious atten-dance; and visiting relatives with the child. Education, employment, wages, and sociodemographiccharacteristics are included in the analyses. The results suggest that infant health shocks do not affect theparents’ social interactions.

� 2009 Elsevier Ltd. All rights reserved.

Introduction

There is considerable evidence of a positive association betweenhealth and social capital. Researchers have presented cogentarguments for why social capital would promote health at both theindividual and the community levels, and the literature hasproduced numerous empirical findings consistent with thosehypotheses. At the same time, it is acknowledged that causalitycould run in the other directiondthat health could affect socialinteractions, and therefore, one’s investment in social capital.Because health and social capital are both forms of human capitaland are affected by many of the same factors, and because it isdifficult to identify a ‘‘shock’’ that may have a direct impact on onebut not the other, it has been difficult to ascertain both causality

-45630 and #R01-HD-35301uman Development. We are

Markowitz and for valuable

Reichman).

All rights reserved.

and directionality. The purpose of this paper is to add to the liter-ature on social capital and health by testing whether an exogenousshock in the health of a family member (a new baby) affects thefamily’s investment in social capital. As such, it will provideevidence, in a specific context, of whether there are effects of healthon an individual’s social capital. Additionally, this paper contributesto a small but growing literature on the effects of children’s healthon family resources, which can have implications for the healthtrajectories of all family members.

We use the baseline and third waves of the Fragile Families andChild Wellbeing (FFCWB) studyda longitudinal survey of about5000 births to mostly unwed parents in 20 U.S. cities during theyears 1998–2000. Both the mothers and fathers were interviewedat the time of the birth and then again one and three years later.Poor infant health is measured in alternative ways using data frommultiple sources to reflect both serious and random health condi-tions that were present at birth. Social interactions include theparents’ participation in church groups, service clubs, politicalorganizations, community groups, and organizations working withchildren; regular religious attendance; and visiting relatives withthe child, as reported in the mothers’ and fathers’ three-yearfollow-up interviews. Data on education, employment, wages, andsociodemographic characteristics are included in the analyses.

J. Schultz et al. / Social Science & Medicine 69 (2009) 76–84 77

Because the FFCWB study oversampled non-marital births, thesample consists of adults who are relatively young and economi-cally disadvantaged. Few studies in the social capital and healthliterature have focused on young adults or disadvantaged pop-ulations, which are important groups to study for several reasons.First, young adults have the potential to benefit more (i.e., fora longer period of time) than their older counterparts frominvestments in social capital. Second, economically disadvantagedindividuals have fewer financial resources that can buffer theeffects of adverse life shocks, and social capital may providea means for overcoming such difficulties. Much previous researchhas found that social capital is highly associated with socioeco-nomic status (SES) (e.g., Saffer, 2008), but few studies of socialcapital and health have focused on economically disadvantagedpopulations. Third, economically disadvantaged men have highrates of crime and poor health, both of which can confer substantialprivate and social costs. Thus, social capitaldif in socially desirableforms (i.e., not stemming from gang-related activity or negativepeer influences)dcan be an important asset for poor men, whorepresent an understudied group in terms of both social capital andhealth.

Background

Social capital has been characterized as a combination of socialorganizations, social networks, and civic participation that canimprove the efficiency of society by facilitating coordinated actions(Putnam, 1993) and has been identified as a factor explainingvariation in a wide range of public health measures. An observedrelationship between social capital and health dates back over 100years when Durkheim, in his classic study of suicide, noted thatsocial integration can enhance well-being. Bourdieu (1980, 1985)highlighted the benefits to individuals from participating in groups,the importance of creating such interactions, and the concept thatbuilding social networks requires investment in group activities(see Portes, 1998 for an extensive review of other contemporarycontributions to the social capital literature). Recent interest insocioeconomic determinants of health has resulted in numerousstudies linking social capital with various aspects of health andwell-being (see Lomas, 1998 for a review).

Much research has examined the relationship between socialcapital and health at both the community and individual levels.Kawachi, Kim, Coutts, and Subramanian (2004) reviewed 31empirical studies investigating the relationship between socialcapital and health that were conceptualized at the community-level or utilized a multi-level framework to incorporate both indi-vidual- and community-level factors. They found, with fewexceptions, a consistent association between social capital andhealth outcomes.

Particularly relevant to the current analysis are studies that haveexamined social capital and health or developmental outcomeswithin the family. Runyan, Hunter, and Amaya-Jackson (1998) usedlongitudinal data to study the impact of social capital on childmaltreatment. They constructed a total social capital score thatincluded the following measures: two-parent family, maternalsocial support, number of children in family, neighborhoodsupport, and church attendance. Child well-being was measuredusing the Batelle Developmental Inventory Screening Test andChild Behavior Check List. The authors concluded that social capitalplays a crucial role in determining developmental and behavioralwell-being in families with low levels of financial and educationalresources. McNeal (1999) found positive effects of parentalinvolvement in schools on their children’s educational outcomesusing the National Educational Longitudinal Study. Kana’iaupuni,Donato, Thompson-Colon, and Stainback (2005) used data from the

Health and Migration Survey in Mexico to study the relationshipbetween child well-being (health status as reported by mother) andsocial networks (network size, kinship roles, interaction, andprovision of financial and emotional support). They found thatnetworks with more extended kin and co-resident ties offer greatersupport resources to mothers with young children, particularlyamong the poorest families. They also found that social support andinteractions with extended kin help sustain children’s health.

Relatively few studies have evaluated the effects of exogenousmeasures of child health (e.g., health measures that are indepen-dent of social capital) on social capital investment by parents. Kazakand Wilcox (1984) found that mothers and fathers of children withspina bifida have small social networks compared to parents ofhealthy children. They also found that families with disabled chil-dren have higher-density networks, as measured by the extent towhich members of the social network know and interact with oneanother. A study of parents with children diagnosed with phenyl-ketonuria had a similar finding (Kazak, Reber, & Carter, 1988),although the differences in network size between families withdisabled children and those without were not large in either study.Overall, the findings from these studies suggest that families withchildren in poor health or with disabilities are not socially isolated.The advantage of these two studies is that they focus on disabilitiesthat were diagnosed at birth and are not believed to be related toprenatal or other parenting behaviors. The disadvantage is thatthey focus on only one type of disability and therefore may not begeneralizable to a broader group of children with serious healthconditions.

Some studies have investigated the effect of poor child healthmore generally on social interactions of the parents, but thisresearch has not focused on health problems that are known to beexogenous. For example, using data from the Wisconsin Longitu-dinal Study, Seltzer, Greenberg, Floyd, Pettee, and Hong (2001)found that parents with disabled children had lower rates of socialparticipation (as measured by participation in social organizationsand frequency of visits with friends and relatives) than parentswithout a disabled child. In a more recent review of studies(generally qualitative analyses of small samples) describing howfamilies cope when a child has kidney failure, Aldridge (2008)found themes of social isolation and distress. The disadvantages ofthis literature are that the severity of the child’s health problemsmay be due, in part, to the intensity with which parents haveinvested in the child’s health and that the factors affecting parentalinvestments in the child’s health may be the same as those affectingthe parents’ investments in their own social capital. Thus, thecausal direction is not clear and it is possible that unobservedfactors underlie the observed associations between child healthand parental social capital.

It is important to consider the point in the life course whenexamining the effect of poor child health on parental social inter-actions. The initial reaction to the diagnosis of a disability in a childcan be devastating, and parents’ adaptations to the reality of raisinga child with a disability are a life-long process (Hauser-Cram, War-field, Shonkoff, & Krauss, 2001), making it difficult to predict howparents may change their behavior as their children grow. Costigan,Floyd, Harter, and McClintock (1997) found that family patterns,routines, and expectations are disrupted by the birth of a child witha disability, but that over time most families regain equilibrium intheir family relationships and well-being. Gallimore, Coots, Weisner,Garnier, and Guthrie (1996) found that although parents makeaccommodations for their disabled children at all ages, the intensityof those accommodations remains stable between the ages 3 and 7and then decreases between ages 7 and 11.

In summary, while there appears to be significant evidencesuggesting a strong relationship between stocks of social capital

J. Schultz et al. / Social Science & Medicine 69 (2009) 76–8478

and health outcomes, relatively little is known about how adverseand exogenous health shocks affect social capital investment.Although a few studies have examined the effects of poor childhealth on parental social capital, they have limitations eitherbecause they investigated very specific child health conditions andmay not be generalizable, or because they did not investigate childhealth problems that are unrelated to parental investments. Theexisting literature points to a need for population-based studiesthat consider a range of child health conditions, investigate theeffects of conditions that are unlikely to be affected by parentalactions, compare individuals at the same life stage, and focus ondisadvantaged populations. The current study fills these gaps byusing population-based individual level data to investigate theeffects of exogenous (broad–based) infant health shocks on thesocial interactions of their largely disadvantaged parentsdmothersand fathersdthree years later.

Theoretical framework

Social capital is an investment and consumption good thatincreases utility and improves the welfare of a household (e.g.Bolin, Lindgren, Lindstrom, & Nysted, 2003; Glaeser, Laibson, &Sacerdote, 2002). In this study, we are interested in investment inindividual social capital, as measured by social interactions overa 12-month period, rather than the stock of social capital at thecommunity level. We thus define social capital as an individualattribute arising from social networks and relationships rather thanas a collective response in the context of a country, region or state. Afamily invests in social capital by forming and maintaining socialinteractions wherein the returns from this investment yield utilityand improve resource allocations (e.g. through information sharingand efficiencies in decision making).

We follow the theoretical model from Bolin et al. in which thefamily is a producer of health and social capital and simultaneouslyinvests in both. In this framework, the returns from investing insocial capital (forming and maintaining connections to otherpeople) are both direct (in that interactions yield utility) andindirect (in that social capital stretches the household’s resources).The factors that determine the family’s ability to transform timeand market goods into commodities enter the demand functionsfor market goods, health and social capital. In its simple form, Bolinet al. present a household utility function where utility at one pointin time, t, is derived from consumption of a vector of commodities(Zt), the health of each member of the family [each parent (P1, P2)and each child (Cj)], and social capital (St):

U�

HP1

t ;HP2

t ;HC j

t ;Zt ; St

�(1)

The household allocates its resources over time to maximizelifetime utility subject to a subjective rate of discounting and initialstocks of health and social capital for each family member. Thestocks of health and social capital will change according to the grosslevels of investment in health and social capital and relevant ratesof depreciation.

The model predicts that the level of social capital will decreasewith age if the rate of depreciation of social capital increases withage. Education may increase the marginal product of time andtherefore may have a positive effect on investment in social capital.Highly educated individuals are also more likely to be informed andhave the skills necessary to engage in social activities. Investment insocial capital may be lower for those who are married or cohabiting ifa partner is considered a substitute for social capital. Presence ofchildren in or outside of the household may increase or decreasesocial capital, depending on whether children are consideredsubstitutes or complements to social capital. For instance, having

children may increase parental involvement in school-relatedactivities. However, having children with other partners maydecrease social interaction by increasing time constraints. Employ-ment effects may go in either direction, as they depend on whetherworking time is a complement to or substitute for social capital.

The key effect of interest in this paper is that of poor child health(defined later) on the social interactions of the parents. We esti-mate the demand for a parent’s social interactions as a function ofthe characteristics from the basic Bolin et al. model plus othervariables that are related to the arguments in Equation (1) and mayaffect the costs or benefits of engagement in social interactions. Theexpected sign of poor child health is ambiguous, as having a child inpoor health could increase both the potential benefits of socialcapital and the costs of the investment. For example, parents mayfeel a greater need to be a part of a formal religious community,which would increase the expected benefits. However, the timedemands on parents of children in poor health might increase theopportunity cost of their time spent in formal social interactions,which would increase the costs of social interactions.

Data

We use data from a recent national birth cohort survey that havebeen linked to medical records of mother respondents and theirnewborns. The Fragile Families and Child Wellbeing (FFCWB)survey follows a cohort of mostly unwed parents and theirnewborn children in 20 large U.S. cities (in 15 states). The study wasdesigned to provide information about the conditions and capa-bilities of new (mostly unwed) parents, the determinants andtrajectories of their relationships, and the consequences for parentsand children of welfare reform and other policies.

The FFCWB study randomly sampled births in 75 urban hospi-tals between 1998 and 2000. By design, approximately 3⁄4 of theinterviewed mothers were unmarried. Face-to-face interviewswere conducted with 4898 mothers while they were still in thehospital after giving birth. The infants’ fathers were also inter-viewed, shortly thereafter in the hospital or at another location (seeReichman, Teitler, Garfinkel, & McLanahan, 2001 for a description ofthe research design). Baseline response rates were 86% amongeligible mothers and 78% among eligible fathers (fathers wereeligible if the infant’s mother completed an interview). Additionaldata have been collected from the hospital medical records (fromthe birth) for a sub-sample of 3684 births.

Follow-up interviews with both parents were conducted over thetelephone approximately three years after the birth of the focalchild. Eighty six percent of mothers who completed baseline inter-views were re-interviewed when their child was between 30 and 50months old. Of the fathers who completed baseline interviews, 77%of fathers who completed baseline interviews completed the three-year follow-up interview. Follow-up interviews also took placeapproximately one year after the birth, but our analyses rely almostexclusively on baseline and three-year follow-up data.

The FFCWB data are well suited for analyzing the effects of childhealth on social capital. They were collected as part of a longitudinalbirth cohort study, and include: (1) detailed data on the child’shealth at birth from hospital medical records; (2) questions aboutparticipation in community organizations and visiting relativesasked to each parent at three years; (3) detailed measures of bothparents’ human capital; and (4) detailed information on the parents’relationship status, living arrangements, and other children.

Descriptive analysis

We use two different analysis samplesda mother sample anda father sample. Below we describe the measures we use in our

Table 1Sample characteristics.

Mothers Fathers

Parents’ Social Interactions (measured at 3 years)Participation in any organization .45 .43Number of organizations .76 (1.04) .76 (1.10)Regular religious attendance .58 .33Visiting relatives with child .89 .85

Child Health (at birth)Severe health condition .02 .02Severe health condition or VLBW .03 .03Moderate or severe health condition .20 .21

Parent CharacteristicsAge, years 25.00 (6.00) 28.09 (7.25)Non-Hispanic Black .49 .47Hispanic .27 .27Other race/ethnicity a .24 .26Immigrant .15 .17<High school graduatea .34 .33High school graduate .31 .35Some college but not graduate .25 .23College graduate .10 .12Lived with both parents at age 15 .42 .44Medicaid birth .65 .60Attended religious services regularly .38 .30Employed .82 .82Hourly wage, dollars 8.15 (9.13) 11.02 (14.50)

Relationship CharacteristicsMarried .24 .31Cohabiting .37 .42Father visited hospital .81 .92Mother had children with another man .33 .32Father had children with another woman .33 .28Parents had other children together .37 .40Father did not complete baseline interview .18 .08

N 2981 2145

Notes: Standard deviations in parentheses.VLBW¼ very low birth weight (<1500 g). All parent and relationship characteristicswere measured at baseline, except when father had children with another woman,which was measured at 1 year.

a Reference category in regression models.

J. Schultz et al. / Social Science & Medicine 69 (2009) 76–84 79

analyses, present summary statistics, and point out salient character-istics. Unless indicated otherwise, poor child health and all covariatesare measured at baseline. In general, we use each parent’s reports aboutthemselves. However, in cases where fathers’ data are missing, we usemothers’ reports about the father if those are available.

The mother sample is limited to cases for which medical recorddata are available and the mother completed the three-year survey.Of the 3684 mothers with medical record data available, 3192completed the three-year survey. Of the 3192 mothers, 52 wereexcluded from the analyses because of missing information on thesocial interaction variables and 159 were excluded because ofmissing data on one or more covariates. Similarly, the father sampleis limited to cases for which medical record data (for the motherand child) are available, a three-year interview was completed bythe father, and the father had some relationship with the mother atthe time of the birth. Of the 3684 fathers with medical record data(for the mother and child) available, 2490 completed the three-yearsurvey. Of the 2490 fathers, 78 were excluded because they rarelyor never talked with the mother at the time of the birth, 230 wereexcluded from the analyses because of missing information on thesocial interaction variables, and 37 were excluded because ofmissing data on one or more covariates.

Measures of social interaction

We consider two measures of formal social interaction in ourprimary analysesdparticipation in any group or organization andregular religious attendance. In the three-year follow-up surveys,both parents were asked about their own participation in a varietyof different types of groups and organizations during the past 12months. Specifically, they were asked if they had participated in: (1)a group affiliated with their church in the past year, (2) a serviceclub, such as the Police Athletic League, (3) a political, civic, orhuman rights organization, (4) a community organization, such asa neighborhood watch, or (5) an organization working with chil-dren or youth. The last category could potentially include health ordisability related parent support groups, depending on the parents’individual interpretation of the question. We constructeda measure of participation in any of the five types of groups ororganizations in the past 12 months (i.e., a positive response to atleast one of the five questions listed above). Parents were also askedin the three-year interviews how often they attend religiousservices, and we used their responses to construct a measure ofregular religious attendance (at least a few times per month). Insupplementary analyses, we consider two alternative measures ofsocial interactiondthe number of organizations or groups in whichthe individual participated (ranging from 0 to 5) and a measure ofinformal social interactiondwhether the parent visited relativeswith the child at least once per week.

Almost half (45%) of mothers in the sample participated in atleast one type of group or organization, with an average of .76organizations; the figures for the fathers were virtually identical tothose for mothers (Table 1). Almost one third (31% of both mothersand fathers) said that they participated in a church group, 7% ofmothers and 8% of fathers said that they participated in a serviceclub, 3% of mothers and 5% of fathers said that they participated ina political group, 12% of mothers and 14% of fathers said that theyparticipated in a community organization, and almost a quarter(23%) of mothers and 20% of fathers said that they participated in anorganization working with children or youth in the past 12 months(figures not shown in tables). Over half (58%) of the mothersreported that they attend religious services regularly, while onlyone third of fathers reported regular attendance. Most mothers(89%) and fathers (85%) reported that they visit relatives with thechild at least once per week.

Measures of poor child health

In constructing measures of poor child health, we were lookingfor a measure which met two criteria: (1) It characterizes a healthshock that was present at birth and unlikely a function of parentalbehaviors. (2) It includes conditions that are strongly associatedwith long-term morbidity. In terms of the former, our goal was tocapture conditions that are for the most part random (e.g., DownSyndrome, congenital heart malformations), given that the preg-nancy resulted in a live birth.

Our first measure of poor child healthdsevere child health con-ditiondcoded from the medical records and one-year maternalreports of child disability, is whether the infant had an abnormalcondition at birth that met the two criteria above. The coding of theabnormal conditions was conducted by an outside pediatricconsultant who was directed to classify a case as having poor childhealth if the child had a condition that is severe, chronic, unlikelycaused by parents’ prenatal behavior, and in the case of one-yearmaternal reports, likely present at birth. This measure mostlyclosely matches our two criteria for an exogenous health shock. Adisadvantage of this measure (for the analyses) is that it is rare:Only 2% of the children in our samples had a severe child healthcondition as we have defined it (Table 1).

The second measure of poor child health, severe child healthcondition or VLBW is measured as severe child health condition and/or was very low birth weight (<1500 g). Very low birth weight isassociated with a number of serious and long-term child health

J. Schultz et al. / Social Science & Medicine 69 (2009) 76–8480

conditions (Reichman, 2005). Reports of birth weight came fromthe medical records for over 99% of the sample. For the remainingcases, birth weight was ascertained from maternal baseline reports.Three percent of the samples had a severe child health condition orVLBW (Table 1). The advantage of this measure is that we gain a fewmore analysis cases with poor child health. The disadvantage is thatthe VLBW component may not be truly exogenous.

Our main measure is a direct, but broad, measure of poor childhealthdwhether the child had an abnormal condition that meetsthe criteria for severe child health condition or has a more moder-ately severe condition that is considered random (not a function ofparental behavior). This measure includes conditions that may ormay not have poor long-term prognoses (examples are hydro-cephaly and cleft palate). We call this measure moderate or severechild health condition. Again, the coding was conducted by anoutside pediatric consultant who systematically reviewed themedical record data on child conditions, as well as data from theone-year interviews on physical disabilities of the child. About onefifth of the children in the samples were coded as having a moderateor severe child health condition as we have defined it (Table 1). Theadvantages of this measure are that all of the conditions areconsidered random (exogenous) and that there are more cases ofpoor child health to analyze. The disadvantage is that most of theconditions do not fall in the ‘‘severe’’ category.

Covariates

The choice of covariates was guided by the theoretical modelpresented earlier. These include a basic set of sociodemographiccharacteristicsdage, race/ethnicity, immigration status, education,whether the parent lived with both of her/his parents at age 15 (apotential indicator of intergenerational poverty), and whether thebirth was covered by Medicaid (a proxy for poverty). Race/ethnicityand immigrant status are included to capture differences in culturalnorms and experiences across groups and because social trust hasbeen found to vary by race (e.g., Alesina & La Ferrara, 2002; Glaeser,Laibson, Scheinkman, & Soutter, 2000; Toussaint, Kiecolt, & Morris,2000). We also control for regular attendance at religious servicesat baseline in the models of regular religious attendance at threeyears, allowing us to capture the dynamics in that outcome (cor-responding baseline controls were not available for the othermeasures of social participation).

The father sample is slightly older than the mother sample (28.1versus 25.0 years) (Table 1). About half of both mothers and fathersare non-Hispanic Black and a quarter of each group is Hispanic.About 1 in 7 of each group was foreign-born. Education levels werelow: About a third (31%) of the mothers were high school graduatesbut did not attend college and only 10% were college graduates. Thefigures are similar for fathersd35% were high school graduates butdid not attend college and 12% were college graduates. About twothirds of the births (65% of the mother sample; 60% of the fathersample) were covered by Medicaid, indicating that a largeproportion of the sample is poor or near poor. About a third (38% ofmothers, 30% of fathers) regularly attended religious services at thetime of the birth.

We include measures of the individual’s employment and wage.For mothers, the measure of employment was whether theyworked within the two-year period preceding the child’s birth(82%). For the fathers, the measure of employment was whetherthey were working at the time of the child’s birth (also 82%). Fatherswere asked how much money they earned in their current or mostrecent job held for two weeks or more. For fathers who had neverbeen employed for two or more consecutive weeks, their reportedreservation wage (how much they would need to be paid per hourto accept a job offer) was used, and for those with missing

information on wages (12%), we set hourly earnings to zero andincluded a flag variable for missing data on wage (not shown intables). Mothers were asked how much money they earned the lasttime that they worked two or more consecutive weeks, and thoseresponses were used to construct a measure of hourly wage. Formothers who had never been employed for two or more consec-utive weeks, the hourly wage was set to zero (reservation wage wasnot asked of the mothers).

We also include a number of variables capturing the parents’relationship and family structure. We consider whether the parentswere married, cohabiting, or neither married nor cohabiting at thetime of the birth. Over three quarters (76%) of the mother samplewas unmarried and almost half (49%) of mothers who wereunmarried lived with the child’s father. Sixty nine percent of thefather sample was unmarried and more than half (61%) of thosewho were unmarried lived with the mother. We also includewhether the father visited the hospital during the birth hospitali-zation (81% for the mother sample; 92% for the father sample), tocapture relationship quality rather than status; whether themother and father had any previous children together; whether themother had children with another partner; whether the father hadother children with another partner (this variable was measured atone year due to data availability); and whether the father did notcomplete a baseline interview.

Finally, we include whether the focal child is male, whether thebirth was a multiple (2% in both the mother and father samples; notshown in tables), and indicators for the mother’s state of residenceat the time of the baseline interview.

If our measures of poor child health are random, we would expectthem to be unrelated to maternal behavioral characteristics, such aseducational attainment and marital status. To assess this assump-tion, we compared the characteristics of mothers with children inpoor health (using the moderate or severe condition variable) tothose of children not in poor health. We found significant differencesonly for the two variables describing the biological characteristics ofthe infantdchild gender and multiple birth. Thus, these bivariateresults are consistent with the assertion that poor child health, as wehave defined it, is unrelated to maternal behaviors.

Multivariate analysis

We use probit models to predict both participation in anyorganization and regular religious attendance. Table 2 presentsthe probit results based on moderate or severe child healthcondition as the measure of poor child health. Each cell containsthe probit coefficient on top, the marginal effect in brackets, andthe standard error of the probit coefficient, which is corrected forcity clustering of observations using the Huber–White method, inparentheses.

Poor child health is positively associated with parents’ socialinteractions (assessed at three years), but never statisticallysignificant at conventional levels (this was also the case whencontrolling for no covariates, not shown). Race, nativity, collegeeducation, and marital status are associated with both participationin any social organization and religious attendance in the expecteddirections, for both mothers and fathers. Hispanic ethnicity ispositively associated with religious attendance, as is baseline reli-gious attendance. Medicaid births are negatively associated withboth social interaction outcomes for mothers, but positively asso-ciated with participation in any organization for fathers. Some ofthe strongest effects are for religious attendance. For example, themother being non-Hispanic Black increases her likelihood ofregular religious attendance at three years by 20 percentage points,and attending religious services regularly at baseline increases thelikelihood of subsequent regular attendance by 41 percentage

Table 2Multivariate probit estimate of effects of poor child health and covariates on parents’ social interactions at three years.

Mothers (N¼ 2981) Fathers (N¼ 2145)

Participation in any organization Regular religious attendance Participation in any organization Regular religious attendance

Coefficient Coefficient Coefficient Coefficient

[ME] [ME] [ME] [ME]

(SE) (SE) (SE) (SE)

Child HealthModerate or severe health condition .07 .04 .14 .08

[.03] [.01] [.05] [.03](.07) (.09) (.07) (.10)

Parent CharacteristicsAge .01 .08* �.01 .02

[.01] [.03] [�00] [.01](.02) (.04) (.02) (.04)

Non-Hispanic Black .23** .52** .17* .23*[.09] [.20] [.07] [.08](.06) (.10) (.08) (.10)

Hispanic �.08 .52** �.05 .15[�.03] [.19] [�.02] [.05](.08) (.07) (.12) (.09)

Immigrant �.44** .31** �.27 .36**[�.17] [.11] [�.10] [.13](.10) (.10) (.16) (.08)

High school graduate .09 �.02 .11 �.15*[.04] [�.01] [.04] [�.05](.07) (.06) (.08) (.07)

Some college, but not graduate .41** �.04 .37** �.19[.16] [�.01] [.15] [�.07](.05) (.07) (.09) (.11)

College graduate .77** .25** .65** �.17[.29] [.09] [.25] [�.06](.08) (.12) (.11) (.15)

Lived with both parents at age 15 .03 .03 .09 .05[.01] [.01] [.04] [.02](.05) (.07) (.08) (.08)

Medicaid birth �.11** �.14* .10* .06[�.04] [�.05] [.04] [.02](.04) (.07) (.04) (.06)

Attended religious services regularly 1.15** 1.17**[.41] [.43](.05) (.06)

Employed �.07 �.10 .03 .02[�.03] [�.04] [.01] [.01](.05) (.09) (.08) (.10)

Hourly Wage �.00 �.01 .01* .00[�.00] [�.00] [.00] [.00](.00) (.01) (.00) (.00)

Relationship CharacteristicsMarried .26** .03 .17 .26

[.10] [.01] [.07] [.09](.10) (.10) (.08) (.13)

Cohabiting .03 .01 �.15 .06[.01] [.00] [�.06] [.02](.07) (.06) (.08) (.07)

Visited hospital .05 .05 .03 �.13[.02] [.02] [.01] [�.05](.06) (.08) (.09) (.11)

Mother had child with another man .02 .07 .04 .07[.01] [.03] [.01] [.02](.06) (.06) (.06) (.06)

Father had children with another woman �.02 �.03 .19* .10[�.01] [�.01] [.08] [.04](.06) (.04) (.08) (.05)

(continued on next page)

J. Schultz et al. / Social Science & Medicine 69 (2009) 76–84 81

Table 2 (continued )

Mothers (N¼ 2981) Fathers (N¼ 2145)

Participation in any organization Regular religious attendance Participation in any organization Regular religious attendance

Coefficient Coefficient Coefficient Coefficient

[ME] [ME] [ME] [ME]

(SE) (SE) (SE) (SE)

Parents had other children together .10* .06 .12 �.08[.04] [.02] [.05] [�.03](.05) (.06) (.07) (.06)

Father did not complete baseline interview �.03 .10 �.09 .57**[�.01] [.04] [�.04] [.21](.06) (.07) (.18) (.18)

Notes: **significant at 1% level; *significant at 5% level. All models include quadratic terms for age; indicators for the child’s gender, multiple birth, and mother’s state ofresidence at baseline; and (for fathers) an indicator for missing wage (estimates not shown). ME: marginal effect; SE: standard error.

J. Schultz et al. / Social Science & Medicine 69 (2009) 76–8482

points for mothers and 43 percentage points for fathers. Controllingfor all of these factors, employment and wages have insignificant orextremely weak associations with both social interaction outcomes.The associations between the family characteristics and socialparticipation outcomes are mixed, but generally insignificant.There are no statistically significant effects of child gender ormultiple birth (not shown).

In Table 3, we present estimates of the effects of poor childhealth from alternative model specifications. These models includealternative measures of poor child health described earlier (severechild health condition or severe child health condition or VLBW) ormoderate or severe health condition, which was used for Table 2.Estimates are shown for the two primary outcomes (participationin any organization and regular religious attendance), as well astwo alternative outcomes (number of organizations and visitedrelatives with the child). The estimates for the number of organi-zations are Ordinary Least Squares regression coefficients and thosefor the other outcomes are probit marginal effects; for all tests ofsignificance, the standard errors were corrected for city clusteringof observations using the Huber–White method. The other esti-mates are probit marginal effects. Only the religious attendancemodels control for baseline religious attendance; otherwise, all ofthe models include all of the controls from Table 2. Covariate esti-mates are not presented, but are very similar to those in Table 2(and for visiting relatives, are very similar to those for formal socialinteractions) and are available upon request. Again, poor childhealth does not appear to affect parents’ social interactions. Formothers and fathers, there are no significant effects and the sign isinconsistent across specifications. Overall, the results indicate thatadverse infant health shocks do not have negative effects onparents’ social interactions when their children are a few years old.

Table 3Effects of poor child health on parents’ social interactions, alternative model specificatio

Participation in anyorganization (ME)

MothersSevere Child Health Condition .05

Severe Child Health Condition or VLBW .02

Moderate or Severe Child Health Condition .03

FathersSevere Child Health Condition .10

Severe Child Health Condition or VLBW .06

Moderate or Severe Child Health Condition .05

Notes: Each of the 24 models includes the same set of covariates as in Table 2 (except thattendance). ME¼marginal effect (from probit model); OLS¼Ordinary Least Squares; VLconventional levels.

Supplementary analyses

We conducted several additional specification checks andauxiliary analyses (estimates are not shown, but are available uponrequest). First, we estimated models predicting each of the fivetypes of social interaction alone and found no associations withpoor child health. Second, we estimated a full set of modelsrestricting the sample to mothers having first births (40% of themother sample) to eliminate potential confounding by healthstatus of the mothers’ other children (information on the health ofnon-focal children was not available) and found the estimatedeffects of poor child health to be insensitive to this samplerestriction. Third, we ran models using low birth weight (<2500 g)as the measure of poor child health. The advantage of this measureis that it is well measured, widely available, and often used asa proxy for poor child health. The disadvantage is that it is unlikelyto be exogenous. Although most associations between low birthweight and parents’ social interactions were insignificant, theywere negative and significant for regular religious attendance (formothers) and for visiting relatives with the child (for fathers).Finally, based on findings that social trust and religious attendanceare positively correlated and that frequent religious attendance ispositively associated with larger and denser social networks (Elli-son & George, 1994; Glaeser, Laibson, Scheinkman, & Soutter, 1999),we estimated models of participation in any organization, numberof organizations, and visiting with relatives that controlled forbaseline religious attendance (the only available measure of socialinteraction at baseline). We found that although baseline religiousattendance was highly significant and positively associated withthe various outcomes, including it did not change the insignificanteffects of poor child health.

ns.

Number of organizations(OLS coefficient)

Regular religiousattendance (ME)

Visiting relativeswith child (ME)

�.02 .09 .04

�.02 .05 .02

.06 .01 .02

.14 �.09 .02

.04 �.08 �.04

.09 .03 �.02

at baseline religious attendance is included only in the models for regular religiousBW¼ very low birth weight (<1500 g). No coefficients are statistically significant at

J. Schultz et al. / Social Science & Medicine 69 (2009) 76–84 83

To place our data (from a largely disadvantaged sample) incontext, we explored the well-documented positive associationbetween adult health and social interaction. For mothers, we useda self-reported measure of overall health status from the firstfollow-up interview (this information was not collected at base-line). We found that mothers who reported excellent or very goodhealth were about 9 percentage points more likely than those whoreported good, fair, or poor health to participate in at least oneorganization. We used the same self-reported measure of overallhealth status for fathers, but from the baseline interview, and foundthat fathers were also 9 percentage points more likely to participatein any organization if they reported their health to be very good orexcellent. Positive and significant associations were found betweenboth parents’ health and the number of organizations, as well. Theestimated effects of parents’ own health on their own socialinteractions were smaller in the multivariate context than whennot controlling for any factors, and the effects of poor child healthwere not affected by including parents’ health.

Finally, to further explore our findings, we ran models for alloutcomes that stratified the sample by neighborhood poverty,which may be related to availability of organizations that facilitatesocial interactions. For both mothers and fathers and for both poorand non-poor neighborhoods defined various ways, we founda pattern of insignificant associations between exogenous shocks inpoor child health and parents’ social interactions.

Conclusion

The contribution of this study is three-fold. First, we contributeto the literature on social capital and health by exploiting anexogenous health shock and estimating effects from health to socialcapital, rather than the other way around. We found that infanthealth shocks do not reduce the social interactions of the child’sparents (as we have measured them). The estimated effects of poorchild health on our measures of social interactions were insignifi-cant and inconsistent (in sign) for mothers, and positive butinsignificant for fathers. Thus, we do not find any compellingevidence that an exogenous shock to health ‘‘causes’’ changes inspecific types of social interactions (which can be viewed asinvestments in social capital) in a specific three-year window.

Second, we have added to the small but growing literature onthe effects of poor child health on family resources, which can haveimplications for the health and well-being of all family members.Recent studies have found that mothers of children born in poorhealth are less likely to live with the child’s father one year later(Reichman, Corman, & Noonan, 2004) and that poor child healthleads to reductions in both mothers’ and fathers’ labor supply(Corman, Noonan, & Reichman, 2005; Noonan, Reichman, & Cor-man, 2005). The findings from this study indicate that poor childhealth does not take yet another toll on family resources byreducing the parents’ social interactions, at least in the first threeyears of the child’s life.

Third, we explored relationships between health and socialinteractions in a cohort of mostly disadvantaged young adults andtheir young children. We found positive associations betweenparents’ health and their social interactions as has been found inthe broader literature. Similar to findings of Schultz, O’Brien, andTadesse (2008) and others, we found that education, children, andreligiosity have positive impacts on adults’ participation in formalactivities and on one type of informal interaction (visiting relatives)even in our relatively disadvantaged population.

Although our results inform the debate on the direction of thecausality between health and social capital, more work needs to bedone. This study was specialized in several ways. We consideredonly one type of health shock, a diagnosis of a serious condition of

a newborn, and linked it to social interactions when the child wasapproximately three years old. It could be that similar shocks toolder children or to the parents themselves would result indifferent effects, or that the effects of the infant health shock haveshorter or longer term effects on parents’ social interactions (thatsaid, the uniform age of the children was a strength of our study).We explored a limited set of social interaction outcomes (e.g., wewere not able to characterize certain types of informal interactions,such as friendships or participation in support groups, with ourdata); it is possible that adverse child health shocks have significanteffects on those types of interactions but not on the types that wemeasured. We were able to capture the dynamics in only one socialinteraction outcome (regular religious attendance), by controllingfor the pre-birth level of that outcome. However, if the child healthshocks are truly exogenous (and the evidence supports this),controlling for the initial level of the outcome may not be necessaryfor obtaining unbiased estimates. Our sampledalthough pop-ulation-based, interesting, and policy relevantdis not representa-tive of the entire U.S. population; rather, it consists of urban, mostlyunmarried parents with a new baby. It could be that the same shockwould affect non-urban or higher-SES parents differently.

References

Aldridge, M. D. (2008). How do families adjust to having a child with chronic kidneyfailure? a systematic review. Nephrology Nursing Journal, 35(2), 157–162.

Alesina, A., & La Ferrara, E. (2002). Who trusts others? Journal of Public Economics,85(2), 207–234.

Bolin, K., Lindgren, B., Lindstrom, M., & Nysted, P. (2003). Investments in socialcapitaldimplications of social interactions for the production of health. SocialScience & Medicine, 56, 2379–2390.

Bourdieu, P. (1980). Le capital social: notes provisoires (Social capital: provisionalnotes). Actes de la Recherche en Sciences Sociales, 30, 3–6.

Bourdieu, P. (1985). The forms of social capital. In J. G. Richardson (Ed.), Handbook oftheory and research for the sociology of education (pp. 241–258). New York:Greenwood.

Corman, H., Noonan, K., & Reichman, N. E. (2005). Mothers’ labor supply in fragilefamilies: the role of child health. Eastern Economic Journal, 31(4), 601–616.

Costigan, C. L., Floyd, F. J., Harter, K. S. M., & McClintock, J. S. (1997). Family processand adaptation to children with mental retardation: disruption and resilience infamily problem-solving interactions. Journal of Family Psychology, 11(4),515–529.

Ellison, C. G., & George, L. K. (1994). Religious involvement, social ties, and socialsupport in a southeastern community. Journal for the Scientific Study of Religion,33, 46–61.

Gallimore, R., Coots, J., Weisner, T., Garnier, H., & Guthrie, D. (1996). Familyresponses to children with early developmental delays II: accommodationintensity and activity in early and middle childhood. American Journal on MentalRetardation, 101(3), 215–232.

Glaeser, E. L., Laibson, D., & Sacerdote, B. (2002). An economic approach to socialcapital. The Economic Journal, 112(483), 437–458.

Glaeser, E.L., Laibson, D., Scheinkman, J.A., &. Soutter, C.L. (1999). What is socialcapital? The determinants of trust and trustworthiness. NBER Working PaperNo. 7216.

Glaeser, E. L., Laibson, D., Scheinkman, J. A., & Soutter, C. L. (2000). Measuring trust.Quarterly Journal of Economics, 115, 811–846.

Hauser-Cram, P., Warfield, M., Shonkoff, J., & Krauss, M. (2001). Children withdisabilities: a longitudinal study of child development and parent well-being.Monographs of the Society for Research in Child Development, 66. (3, Serial No.266).

Kana’iaupuni, S. M., Donato, K. M., Thompson-Colon, T., & Stainback, M. (2005).Counting on kin: social networks, social support, and child health status. SocialForces, 83(3), 1137–1164.

Kawachi, I., Kim, D., Coutts, A., & Subramanian, S. (2004). Commentary: reconcilingthe three accounts of social capital. International Journal of Epidemiology, 33,682–690.

Kazak, A. E., Reber, M., & Carter, A. (1988). Structural and qualitative aspects of socialnetworks in families with young chronically ill children. Society of PediatricPsychology, 13(2), 171–182.

Kazak, A. E., & Wilcox, B. (1984). The structure and function of social networks infamilies with handicapped children. American Journal of Community Psychology,12, 645–661.

Lomas, J. (1998). Social capital and health: implications for public health andepidemiology. Social Science & Medicine, 47(9), 1181–1188.

McNeal, R. B. (1999). Parental involvement as social capital: differential effectivenesson science achievement, truancy, and dropping out. Social Forces, 78, 117–144.

J. Schultz et al. / Social Science & Medicine 69 (2009) 76–8484

Noonan, K., Reichman, N. E., & Corman, H. (2005). New fathers’ labor supply: doeschild health matter? Social Science Quarterly, 86(s1), 1399–1417.

Portes, A. (1998). Social capital: its origins and applications in modern sociology.Annual Review of Sociology, 24, 1–24.

Putnam, R. D. (1993). Making democracy work: Civic traditions in modern Italy.Princeton, NJ: Princeton University Press.

Reichman, N. E. (2005). Low birth weight and school readiness. The Future ofChildren, 15(1), 91–116.

Reichman, N. E., Corman, H., & Noonan, K. (2004). Effects of child health on parents’relationship status. Demography, 41(3), 569–584.

Reichman, N. E., Teitler, J. O., Garfinkel, I., & McLanahan, S. (2001). FragileFamilies: sample and design. Children & Youth Services Review, 23(4/5),303–326.

Runyan, D. K., Hunter, W. M., & Amaya-Jackson, L. (1998). Children who prosper inunfavorable environments: the relationship to social capital. Pediatrics, 101, 12–18.

Saffer, H. (2008). The demand for social interaction. The Journal of Socio-Economics,37, 1047–1060.

Schultz, J. F., O’Brien, M., & Tadesse, B. (2008). Social capital and self-rated health:results from the US 2006 Social Capital Survey of one community. Social Science& Medicine, 67(4), 606–617.

Seltzer, M. M., Greenberg, J. S., Floyd, F., Pettee, Y., & Hong, J. (2001). Life courseimpacts of parenting a child with a disability. American Journal on MentalRetardation, 106, 265–286.

Toussaint, J., Kiecolt, J., & Morris, E. Race and social trust. Paper presented at theAnnual Meeting of the American Sociological Association, Montreal, Quebec,Canada, 2000.