Mothers’ Time with Children and the Social Reproduction of Cognitive Skills

Despite the theoretical importance mother’s time with children plays in the literature on children’s cognitive development, few studies have empirically examined this relationship. This study uses children’s time diaries from the Child Development Supplement of the Panel Study of Income Dynamics to investigate the effect of maternal involvement during pre-school years on children’s cognitive outcomes assessed at ages five to twelve (N = 1,008). I find a positive and persistent effect of the total quantity of time mothers’ spend with children and the time they spend playing together on children’s language development, but only among children who spend time with verbally skilled mothers. These results are present even after controlling for a range of individual and family characteristics that have not been previously examined—such as children’s health and developmental problems, quality of learning environment and lagged test scores. The results suggest that women differ in their ability to successfully translate childcare time into positive cognitive outcomes. Taken together, this study is the first to identify time use as a mechanism for the social reproduction of cognitive skills—time with children is a pathway through which skilled mothers impart their skills and knowledge onto children. Theories regarding child development contend that the time mothers devote to children is crucial to children’s intellectual development. According to these theories, the time mothers spend with children, along with other financial and material resources, can be seen as investments into the production of child “quality” (Becker, 1991). Through repeated interactions children learn to trust their caregivers, build healthy social relations, and are exposed to intellectual stimulation (Coleman, 1988). These experiences help create social capital or the mechanism that facilitates the intergenerational transmission of knowledge, skills and human capital. Theories in developmental psychology also contend that long periods of separation can disrupt this process leaving mothers less sensitive and responsive to their children’s needs and leaving children less exposed to the stimulation necessary for their cognitive development (Vaughn, Grove and Egeland, 1980; Belsky, 2001). Despite the central role parents’ time with children plays in the theoretical construct of child development and social reproduction, few studies have directly tested these theories. Even less is known about the specific aspects of maternal involvement that matter for children’s achievement outcomes. For example, is it the total quantity of maternal time or how the time is used that has the most significant influence on children’s cognitive development? Are all mothers as effective at translating the time they spend with children into positive achievement outcomes? These questions are examined using time diary data and assessments of child cognitive skills from the Panel Study of Income Dynamics and its Child Development Supplement (PSIDCDS), the only nationally representative and longitudinal survey of individuals and households with children’s time diary data, to examine the relationship between maternal involvement during children’s pre-school years (e.g. age 0 to 5 years old) and their cognitive outcomes assessed when they are 5 to 12 years old. Using extensive controls to account for differences among mothers and children, the results demonstrate that the total quantity of maternal care during pre-school years and the total time children spend playing with their mothers positively relates to children’s language development, but only when children spend time with verbally skilled mothers. Furthermore, these relationships persist even after controlling for children’s lagged test scores. More modest associations were found between the intensity of maternal engagement and test scores. These findings contrast two previous studies that find non-significant or negative associations between the mothers’ time with infants and children’s cognitive test scores (Booth et al, 2002; Huston and Aronson, 2005). There are two main reasons for why I find positive and significant effects. First, previous studies do not address the possibility that certain mothers may be more effective at translating their time with children into positive cognitive outcomes by examining interactions between mother-child time and maternal characteristics such as mother’s education and cognitive skills. Second, previous studies do not adequately control for child specific heterogeneity that biased estimates of maternal time. Additionally, the paper extends the extensive literature on the effects of maternal employment on children’s cognitive development. Several studies have suggested that children from socio-economically advantaged families are the most harmed by early maternal employment (Desai, Chase-Lansdale and Michael, 1989; Vandell and Ramanan. 1992; Harvey, 1999; Waldfogel, Han, and Brooks-Gunn. 2002; Hill et al., 2005), but none identify the mechanisms that bring about these effects. My findings suggest that maternal employment may differentially affect children because women differ in their ability to influence their children’s cognitive development. Children’s language development benefits more from exposure to verbally skilled mothers than from exposure to less skilled mothers. This study is among the first to identify mothers’ time with children as pathway for the social reproduction of cognitive skills. PREVIOUS STUDIES The literature on maternal care and children’s cognitive development can be organized into three major areas. The first concentrates on identifying the effect of early maternal care on children’s cognitive outcomes by using maternal employment as a proxy for quantity of care children receive. The second uses time diary data to describe the relationship between mothers’ labor supply and the time they devote to children. The third and most recent area of research uses time diary data to examine the effect of maternal involvement on child outcomes. In the following section, I review the literature and discuss how my study may offer some important contributions to each of these three areas of research. Maternal Employment and Children’s Cognitive Outcomes The literature on the effects of maternal employment on child cognitive outcomes is extensive (see Waldfogel, Han and Brooks-Gunn (2002) and Ruhm (2004) for a complete literature review). These studies all attempt to identify the influence of maternal time on children’s cognitive development by using employment as a proxy for both the quantity and the quality of care children receive. The majority of recent studies suggest that there are deleterious effects associated with employment during children’s first year but that the influence of maternal employment after the first year is more ambiguous (Baydar and Brooks-Gunn, 1991; Blau and Grossberg, 1992; Harvey, 1991; Brooks-Gunn, Han and Waldfogel, 2002; Ruhm, 2004; James-Burdumy, 2005; Hill et al, 2005). Many studies also demonstrate that the influence of maternal employment differs by the characteristics of mothers and families, suggesting that maternal employment may be more harmful for children from advantaged backgrounds—children from wealthier families, non-Hispanic white children and children from intact families. Desai, Chase-Lansdale and Michael analyze a sample of 503 NLSY children and find that maternal employment only negatively influences children from higher income families but not children from middle or lowincome families (1989). Waldfogel, Han and Brooks-Gunn examine a sample of 1,872 NLSY children who were followed from birth to age 7 or 8 using both OLS and family fixed effects and find a persistent negative effect of first year employment on cognitive tests for non-Hispanic white children but not for African American or Hispanic children (2002). Other studies have also found that early maternal employment is associated with positive cognitive outcomes among low-income families (Vandell and Ramanan, 1992; Harvey, 1999) and female-headed households (Harvey, 1999). Taken together, these studies suggest that there is a negative effect of maternal employment during children’s first year but that the overall effect may differ by the characteristics of families. None, however, can identify the mechanism through which maternal employment affects child outcomes or explain why the effect of employment should vary across children and families. Time Diary Studies of Mother’s Time with Children Several time diary studies show that increases in early maternal employment over the past several decades have also been accompanied by behavioral changes in how mothers spend their time outside of work (Bryant and Zick, 1996; Bianchi, 2000; Sandberg and Hofferth, 2001). This literature highlights the fact that there is not a one-to-one corresponding relationship between the amount of time mothers spend at work and the amount of time they spend with children. First, time use studies suggest differences in the amount of time employed and nonemployed mothers devote to childcare are not as great as one might expect considering the amount of time employed women spend working. Analyses using the National Institute of Child Health and Development (NICHD) Study of Early Child Care to examine differences in maternal involvement among working and stay-at-home mothers with 7 month old infants—a time during which childcare responsibilities are arguably the most demanding—show that employed mothers spent only about 12 hours less per week with infants than their nonworking counterparts despite spending on average nearly 30 hours per week at work (Booth et al, 2002; Huston and Aronson, 2005). Second, time diary studies also show that working mothers have changed how they spend their childcare time, suggesting that employment status poorly captures the type of maternal involvement children receive. An analysis of time diaries collected from 226 married couples with children from the 1981 Study of Time Use finds that much of the difference in childcare time between working and stay-at-home mothers is due to differences in the types of childcare activities that only involve children peripherally (Nock and Kingston, 1988). For example, nonemployed women spend significantly more time involved in activities that require low levels of parent-child engagement such as doing housework while children play or watch television in another room. Other studies find that employed mothers spend a greater proportion of their time engaged in “social interaction” (i.e. talking, holding and other forms of interaction) rather than “instrumental care” (i.e. feeding, bathing, diapering, and providing physical care) (Booth et al, 2002; Huston and Aronson, 2005). Therefore, in addition to the quantity of care, the type of maternal care that children receive may differ by employment status. Time diaries show that women’s entry into the labor force is associated with behavioral changes in time use that make employment status a poor proxy for maternal involvement. The question, however, still remains regarding whether these behavioral changes in time use— particularly in terms of differences in maternal engagement and in the types of activities performed together—relate to cognitive development. Time Diaries and Children’s Cognitive Outcomes To date only two published studies have used direct measures of mother-child time from large, representative time diary surveys to examine the relationship between the maternal involvement and children’s cognitive development. Both analyze time diaries administered to mothers with 7-month old infants from the NICHD Study of Early Child Care. Booth et al. examine a sub-sample of 326 married or cohabiting mothers and find that the amount of time mothers spend with their infants is not significantly correlated with measures of cognitive skills assessed when children are 15-months old (2002). Huston and Aronson analyze a more representative sample of mothers by including single mothers (N=1,053) and utilize tests of cognitive skills measured at 24 and 36 months (2005). Overall, they also find that maternal involvement does not significantly predict cognitive outcomes; in one case, they even find that involvement relates negatively to toddlers’ expressive language skills. These two studies provide an important first step towards directly identifying the effect of maternal time on children’s cognitive development. However, they also present new avenues for future research. Neither study adequately controls for characteristics of children that may bias estimates of maternal time. For example, children may also differ in their initial endowments such as innate cognitive ability, health and physical development. Mothers may respond to observed difficulties faced by children by spending more time with children with learning or health problems. In this case, the estimated effects of maternal time will be negatively biased (i.e. one will observe a negative relationship between maternal care and outcomes), if those aspects of initial ability are not considered. Failure to account for these characteristics may explain why Huston and Aronson (2005) find a negative relationship between maternal involvement and early language skills. Additionally, neither study controls for mother’s own cognitive skills nor examines interactions between maternal time and maternal characteristics. Previous studies find that the effect of maternal employment is particularly harmful for children from advantaged backgrounds, which suggest that certain mothers are better at translating the time they spend with children into positive cognitive outcomes (Desai, Chase-Lansdale and Michael, 1989; Vandell and Ramanan, 1992; Harvey, 1999). Examining interactions between maternal time, on one hand, and mothers’ own human capital and cognitive skills, on the other, may provide insight into how cognitive skills are differentially reproduced across families. Finally, both studies are limited because both utilize measures of cognitive skills that are assessed at very young ages (e.g. between 15 to 36 months). There is little evidence that suggests assessments of cognitive development that are measured at such early ages can accurately predict later outcomes. On the other hand, cognitive tests assessed as early as age seven have been shown to predict adult outcomes, such as educational attainment and labor market performance (Robertson and Symons, 1996; Currie and Thomas, 1999). 1 Both Huston and Aronson (2005) and Booth et al. (2002) are limited to examining the influence of maternal care for infants at 7-months. This study, on the other hand, expands the age range to examining mothers’ time with children 5 years old and younger in 1997. Integrating and Extending the Literature: Time Diaries and Child Outcomes This study aims to integrate studies using proxies to examine the effect of maternal involvement on child outcomes and time diary studies describing mothers’ time with children in order to use direct measures of maternal involvement to assess the effects of maternal care for children’s cognitive outcomes. This study extends the maternal employment literature in several important ways. First, the study uses direct measures of maternal involvement using children’s time diaries completed when children were between the ages of 0 and 5. Second, I explore alternative measures of maternal involvement and identify the particular aspects of maternal involvement that relate to cognitive test scores. Third, this article extends Booth et al. (2002) and Huston and Aronson (2005) by better controlling for maternal and child characteristics that may bias estimates of maternal involvement by including controls for mother’s cognitive skills, extensive controls for child developmental, health and behavioral problems during infancy, children’s lagged test scores to account for initial differences among children in innate ability and quality of children’s schooling environment. To determine whether cognitive skills are differentially transmitted from mothers to children, I investigate interactions between maternal time with measures of mother’s own human capital and cognitive skills. Furthermore, I extend Booth et al. (2002) and Huston and Aronson (2005) by examining the influence of mother-child time during a wider age range (i.e. when children are between 0 to 5 years old) and utilizing measures of cognitive ability that are assessed at older ages, when children are between 5 and 12 years old.