Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
268 FROM NEURONS TO NEIGHBORHOODS tion is needed in order to produce measurable improvements in childrenâs developmental outcomes. Finally, with welfare reform has come a growing interest in the families of the working poor. This vast natural experiment has also created new opportunities to learn about how various approaches to increasing work and income among families living in poverty affect both child and adult outcomes of paramount interest to the nation. Increasingly, research ad- dressing questions about how resources change over time and their impact on childrenâs development is relying on longitudinal data and experimental designs. DO RESOURCES MATTER FOR CHILDRENâS DEVELOPMENT? Understanding how different family resources affect young childrenâs lives necessitates distinguishing among them; connecting them to such re- sources as money, time, and access to the learning opportunities that they represent; and identifying the different pathways through which these re- sources might influence young childrenâs development. Taking poverty as an example, it is important to know how it manifests itself in young childrenâs lives, how it affects the extent to which their basic needs are met, and through what processes it promotes or undermines their capacity to accomplish the basic developmental tasks outlined in the previous chapters of this report. It would be surprising if the odds of healthy, adaptive development did not differ for children growing up in families with ample, compared with impoverished, resources. Families who occupy different socioeconomic niches because of parental education, income, and occupation have strik- ingly different capacities to purchase safe housing, nutritious meals, high- quality child care, and other opportunities that can foster health, learning, and adaptation (Becker, 1981; Brooks-Gunn et al., 1995). A two-parent family with one highly paid wage earner who makes it possible for the other parent to stay at home with the children is in an entirely different situation from a single parent with a poverty-level wage, for example (Becker and Lewis, 1973; Mason and Kuhlthau, 1992; Timmer et al., 1985). How the trade-offs that families make among employment, cash income, and child care time affect young children is a controversial and poorly understood question. The psychological well-being of mothers and associated patterns of parenting are also much more likely to suffer in families with limited re- sources (Brooks-Gunn and Duncan, 1997). Research has focused increas- ingly on connections among family resources, psychological aspects of fam- ily functioning, and child well-being (McLoyd, 1998). Finally, there is growing interest in how familiesâ access to different social resources, such
FAMILY RESOURCES 269 as relatives, supportive friends and neighbors, and community organiza- tions (churches, family resource centers, safe recreational settings), affects parenting and child development (Coleman, 1985; Edin and Lein, 1997; Jarrett and Burton, 1999; Jencks and Mayer, 1990; Sampson, 1992; Yoshikawa, 1999). We reserve discussion of resources outside the family for Chapter 12. In this section, we describe what is known about the extent to which parental employment, income and poverty, parental schooling, and family structure affect the developing child. We couch the discussion in the con- text of trends that have altered, in many instances dramatically, the socio- economic landscape of young children in the United States. We close our discussion of connections between socioeconomic resources and child de- velopment by addressing the challenges raised by behavioral geneticists (e.g., Rowe and Rodgers, 1997), who argue that genetic factors are at the heart of associations between family resources and child outcomes. In the final section, we review evidence on the various ways in which socioeco- nomic resources affect young childrenâs development. Work and Childrenâs Development Maternal and paternal employment play a powerful role in determining the time and money that families devote to their children. Long-standing concerns about the developmental impacts of fathersâ unemployment and mothersâ employment have now been supplemented by research focusing on the developmental consequences of how parents configure their work, the circumstances of parental work, and the increasing decoupling of work and economic security, illustrated by the growth in working-poor families. Increases in paid maternal employment over the past quarter-century are one of the most dramaticâand best-knownâsocial trends. Between 1975 and 1999, the proportion of children under 6 years of age with mothers in the labor force increased from 38.8 percent to 61.1 percentâa 36 percent increase (Figure 10-1).1 The proportion of young children with mothers working full-time and year-round nearly tripled, from 11 to 30 percent. The increase in maternal employment (including both full- and part-time workers) over this 24 year period was most rapid for infants, rising from 24 to 54 percent, compared with older children. The propor- tion of young children with a mother working part time changed relatively little (ranging between 36 and 40 percent) over that period. A much larger share of young Hispanic (48 percent) than white (29 percent) or black (26 percent) children lived with mothers who did not work for pay in 1997. 1These data are based on all young children who are living with their mothers.
270 FROM NEURONS TO NEIGHBORHOODS 0 10 20 30 40 50 60 70 80 Under 1 year old Under 6 years old 6 to 17 years old, and none younger 1975 1999 Pe rc en ta ge o f m ot he rs b y ag e of c hi ld FIGURE 10-1 Trends in the proportion of mothers in the labor force, by age of child, 1975-1999. NOTES: Since 1975 data for mothers with children under 1 year of age are not available, the data for this column are from the June 1977 Current Population Survey. CPS was redesigned in 1994, so data in this table for 1999 are not strictly comparable with the data shown for 1975 and 1977. SOURCES: Data from 1999 Current Population Survey (annual averages), avail- able online at www.bls.gov/news.release/famee.t05.htm and www.bls.gov/news. re- lease/famee.t06.htm; Hayghe (1997); U.S. Bureau of the Census (1982). Parental employment often, but not always, entails supplementing parental care with substantial amounts of care by others, and herein lies many of todayâs concerns about its effects on very young children, as we discuss in Chapter 11. While employment may increase the cash income of families, work-related expenses may increase as well, leaving them with a differing composition of time and money, but not necessarily greater resources over- all.
FAMILY RESOURCES 271 Employment, of course, can take many different forms. Some parents work full-time, full-year in low-wage jobs that fail to lift their families out of poverty. In other cases, work is intermittent, or multiple low-wage jobs are held by multiple family members, including older children, who all contribute to family income. As we discuss in the chapter on child care, it is not unusual for parents to organize their work hours so that they can keep child care within the family, particularly during infancy. Even among mothers who have received public assistance, a recent analysis found that over a 24-month period, 43 percent either combined work and welfare receipt or cycled between the two, and another 23 percent were not em- ployed but spent substantial time looking for work (Spalter-Roth et al., 1995). What do employment trends mean for young children? Research on maternal employment has been based primarily on middle-income families (Gottfried and Gottfried, 1988; Hoffman, 1989) and has been inconclusive. Most of the evidence indicates that children are either positively affected or unaffected by growing up in a family with an employed mother. Evidence is accumulating,2 however, that suggests that maternal employment in the childâs first year, especially if mothers work long hours, can indeed be a negative factor for infant development (Baydar and Brooks-Gunn, 1991; Belsky and Eggebeen, 1991; Ruhm, 2000; Desai et al., 1989; Vandell and Corasaniti, 1988; Han et al., 2000; Waldfogel et al., 2000). Efforts to determine which infants are most affected have been inconclusive, but the negative findings emerge more often for those in two-parent and middle- income families than for those with fewer family resources. Interestingly, one report (Ruhm, 2000) hinted at similar effects of paternal and maternal employment, suggesting the importance of time investments by fathers as well as by mothers. How strong these negative effects are can, of course, be affected by the quality of the alternative care that the child experiences in the motherâs absence (see next chapter) Research on the children of working-poor parents is just beginning to emerge, despite the fact that this is one of the fastest growing groups of children in the United States. In 1996, about 5 million children lived in a working-poor family, defined as a family with an income below the poverty line and two parents who work the equivalent of a full-time job or a single 2Much of the evidence is based on analyses of the National Longitudinal Survey of Youth (NLSY) The NLSY is a national sample of 12,686 youth who were 14 to 22 years of age when first interviewed in 1979. Beginning in 1986 and every 2 years since, developmental measures were administered to children of civilian mothers in the NLSY (4,971 children of 2,922 mothers in 1986). See Chase-Lansdale et al. (1991) for a fuller description. The survey does not include measures of child care quality.
272 FROM NEURONS TO NEIGHBORHOODS parent who works at least 20 hours per week (Wertheimer, 1999). Over half of children in poor, married-couple families and 30 percent of poor children in single-mother families have parents who work these substantial hours. Both Hispanic and black families are more likely than white families to be poor, despite the presence of working adults. The few studies that have focused on maternal employment among low-income children suggest that they are not hurt by and may benefit from maternal employment, particularly with regard to cognitive outcomes (Alessandri, 1992; Hoffman et al., 1999; Moore and Driscoll, 1997; Vandell and Ramanan, 1992). In fact, there appears to be a more consistent advan- tage of maternal employment for children in working class than in middle class families (Desai et al., 1989; Gold and Andres, 1978; Hoffman, 1979; Zaslow, 1987), perhaps as a result of its positive effects on the motherâs sense of well-being, the fatherâs involvement in child-care activities, and the quality of parenting (Hoffman et al., 1999). The limited evidence that is available suggests that infants and toddlers fare better in working-poor families than in poor families in which the parents do not work or work minimally (NICHD Early Child Care Research Network, 1999b). Yet those in working families that are not living in poverty do substantially better than either group of children living in poverty. The childrenâs better adjustment is seen on measures of cognitive, language, and social develop- ment. This pattern of child outcomes is, however, largely attributable to differences across these three groups of families in demographic character- istics (e.g., mothersâ education, family size), mothersâ depression and social support, and parenting quality and attitudes. New evidence from experi- mental studies of welfare reform (discussed below) are beginning to expand understanding of how parentsâ transition to workâoften substantial hours of workâaffects young children living in poverty. The corresponding literature on how fathersâ loss of work and unem- ployment affect children has emphasized the influence of these circum- stances on harmful family dynamics. Unemployment increases financial strain, which in turn may compromise parent-child relationships by creat- ing tension and hostility as well as reducing warmth and supportiveness in the home. These adverse home environments have been found to have negative consequences for childrenâs development in the short and long term (Conger and Elder Jr., 1994; McLoyd, 1989; Tomblin et al., 1997). Investigators have also explored to what extent the circumstances and features of work, such as the flexibility of oneâs work hours, the extent of control over the day-to-day nature of work and the absence of repetitious and boring tasks or the presence of challenging tasks, account for the effects of maternal employment on children (Alessandri, 1992; Greenberger and OâNeil, 1991; Howes et al., 1995a; Jencks et al., 1988; Menaghan and
FAMILY RESOURCES 273 Parcel, 1995; Parcel and Menaghan, 1994; Parke and Buriel, 1998). Re- search has linked these features of work to parental cognitive skills, such as intellectual flexibility, and other personal characteristics, such as self-direc- tion (Kohn and Schooler, 1973), and more recently to childrenâs cognitive achievement and social behavior (Parcel and Menaghan, 1994). In one longitudinal analysis, for example, single mothersâ entry into low-complex- ity, low-wage jobs was associated with declines in the quality of the home environment (Menaghan and Parcel, 1995). This evidence is cause for concern when juxtaposed with projections that the second highest rate of job growth over the next decade will occur in the service economy. These jobs often entail very low wages, few benefits, little autonomy, and non- standard hours (e.g., shift work). They are also disproportionately filled by less-well-educated women who now constitute a sizeable group of mothers who are entering the labor force as a result of welfare reform. In 1994, close to half (41 percent) of children under age 5 whose mothers were employed had mothers whose principal job involved a ânondayâ work shift (defined as the majority of work hours being outside the 8 a.m. to 4 p.m. time period). Young children living in poverty are much more likely to have a mother who works a nonday shift (59 percent of children) compared with young children living above the poverty line (39 percent of children) (Presser and Cox, 1997). We know very little about the developmental implications of shift work. One recent study has reported an association between shift work and marital instability, with the odds of separation or divorce three to six times higher among mothers and fathers who are engaged in shift work, compared with otherwise similar parents not engaged in shift work (Presser, 2000). These findings did not appear to be attributable to spouses in more troubled marriages electing to move into shift work. This is, however, an isolated study that needs to be replicated, particularly in light of the fact that many parents are motivated to engage in shift work as a way of keeping child care within the family. In sum, the familiar trends in parental employment can bode well or ill for young children depending on features of the work, the income it gener- ates, the nature and structure of the job, its timing and total hoursâand, as we see in Chapter 11, on the environments and relationships that children experience when they are not in the care of their parents. Of concern is the fact that the growth in parental employment appears to be in precisely those circumstances that have been found to pose risks to early develop- ment. It is thus critically important to recognize that the characteristics and experiences of working families have changed substantially over the past 25 years. It is especially troubling that young children whose parents are making considerable work efforts are more likely today than in the recent past to be living in poverty.
274 FROM NEURONS TO NEIGHBORHOODS TABLE 10-1 Risk of Adverse Child Outcomes and Environmental Conditions Associated with Poverty Status Risk for Poor Relative to Child Outcomes Nonpoor Children Lead poisoning 3.5a Birth to unmarried teenager 3.1b Short-stay hospital episode 2.0c Grade repetition and high school dropout 2.0c Low birthweight 1.7d Mortality 1.7d Learning disability 1.4c Parent report of emotional or behavior problem that lasted 3 months or more 1.3e SES Mediators Child abuse and neglect 6.8f Depression 2.3g Experiencing violent crimes 2.2h Substance abuse 1.9i aData from NHANES III, 1988-1991 (Brody et al., 1994). Poor children living in families with incomes less than 130 percent of the poverty threshold are classified as poor. All other children are classified as nonpoor. bData from the Panel Study of Income Dynamics (PSID) Based on 1,705 children ages 0-6 in 1968; outcomes measured at ages 21 to 27 (Haveman and Wolfe, 1994, p. 108, Table 4.10c). cData from the 1988 National Health Interview Survey Child Health Supplement (NHIS- CHS), a nationwide household survey. The household member who knew the most about the sample childâs health, usually the childâs mother, reported childrenâs health status. Figures calculated from Coiro et al. (1994) and Dawson (1991). dData from the National Maternal and Infant Health survey collected in 1989 and 1990, with 1988 as the reference period. Percentages were calculated from the number of deaths and number of low-birthweight births per 1,000 live births as reported in Federman et al. (1996, p. 10). eData from the National Health Interview Survey Child Health Supplement (NHIS-CHS). The question was meant to identify children with common psychological disorders, such as attention deficit disorder or depression, as well as severe problems, such as autism. fData from Childrenâs Defense Fund (1994, p. 87, Tables 5-6). Poor families are those with annual incomes below $15,000. gData from a New Haven Epidemiological Catchment Area in 1980 (Bruce et al., 1991). Poverty status was determined by comparing respondentâs 1980 income to the 1980 poverty threshold. Odds ratio of having depressive episode in six months after first interview. De- pressive episode was diagnosed by the DIS. The odds ratio was corrected for age, sex, race, and previous history of depression.
FAMILY RESOURCES 275 Poverty and Childrenâs Development One of the most consistent associations in developmental science is between economic hardship and compromised child development. The influence of family income, and specifically of poverty, has been of special interest in light of the numerous policies that address poverty in the United States and the intractability ofâindeed, the increase inâthe child poverty rate.3 In 1997, some 5.2 million young children (22 percent of all young children) in the United States were poor, and 42 percent lived at or below 185 percent of the poverty line.4 The strength and consistency of associa- tions between poverty and critical aspects of child development are striking (Brooks-Gunn and Duncan, 1997) (Table 10-1). Developmental research on children in poverty has grown exponen- tially in recent years (see Brooks-Gunn and Duncan, 1997; Chase-Lansdale and Brooks-Gunn, 1995; Huston, 1991; Huston et al., 1994; McLoyd, 1998). This research has yielded suggestive evidence that increasing the incomes of low-income parents with young children will improve the odds of successful early development. What remains to be understood is the nature of the impact and optimal strategies for increasing the incomes of poor parents that best promote their childrenâs development. Even though most children living in poverty grow up to be productive adults, some do not and, without intervention, individual differences among children at school entry that are linked to poverty often persist over time (Stipek, in press). When this evidence is combined with the basic facts about early 3This âofficialâ poverty count is based on a Census Bureau comparison of total family income with a poverty threshold that varies by family size. Expressed in 1997 dollars, the respective poverty thresholds for families with three and four persons were roughly $13,000 and $16,500. Young children living in families with total cash incomes below these thresh- olds were counted as poor. 4Children in families with incomes between 100 and 185 percent of the federal poverty line are designated near poor because they are served by a number of government assistance programs that use 185 percent of the poverty line as the upper limit to determine eligibility. TABLE 10-1 Continued hData from the National Crime Victimization Interview Survey. Results are for house- holds or persons living in households. Data were collected between January 1992 and June 1993 with 1992 as the reference period. Percentages are calculated from the crimes per 1,000 people per year. Reported in Federman et al. (1996, p. 9). iData from the National Comorbidity Survey 1990-1992 (Kessler et al., 1994). Parental sample was restricted to respondents between age 15 and 54. Substance abuse included both alcohol and drug abuse or dependence in the past 12 months, as diagnosed by the Composite International Diagnostic Interview. Poor respondents were those with incomes of less than $20,000 compared with those with $70,000 or more.
276 FROM NEURONS TO NEIGHBORHOODS childhood poverty, there is considerable cause for concern: (1) young chil- dren are now the poorest age group in U.S. society (see Figure 10-2); (2) poverty is considerably more prevalent among children now than 25 years ago,5 despite an unprecedented period of macroeconomic prosperity and substantially higher rates of parental employment; (3) poverty has increased much more for minority than nonminority children; and (4) children living in poverty are falling farther behind their more affluent peers. The United States now has both more poor and more affluent children than it did 25 years ago, creating a widening disparity between the haves and have nots among young children. Our country also has more poor and more affluent children than most other Western countries (Table 10-2). Assessing the developmental consequences of poverty and of differing family incomes more generally is not as straightforward as one might think. Contrary to popular belief, and in contrast to most other measures of socioeconomic status, family income is often quite volatile across a familyâs life cycle and, in particular, a childâs childhood (Duncan et al., 1998a). For example, only half of children and families who are poor in a given year are 0-5 6-17 65+ 18-64 30 25 20 15 10 5 0 Po ve rty ra te (% ) 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 Year FIGURE 10-2 Poverty rates by age, 1975-1997. SOURCE: Bennett et al. (1999). Reprinted with permission. 5As discouraging as these poverty figures are, they would have been worse had it not been for certain changes in government taxes and transfer programs. In 1997, taking into account the effects of all taxes, tax credits (including the earned income tax credit), means-tested income transfers, and non-means-tested income transfers reduces the poverty rate for families with young children by about 9 percentage points, to leave 14 percent living in poverty.
FAMILY RESOURCES 277 persistently poor (Duncan et al., 1994). On average, family incomes in- crease as children age, but average patterns conceal a great deal of year-to- year volatility, making it important to consider how economic resources at different points during the childhood years affect development. The mal- leability of young childrenâs development and the overwhelming impor- tance of the family (rather than school or peer) context suggest that eco- nomic conditions in early childhood may be far more important for shaping childrenâs ability, behavior, and achievement than conditions later in child- hood. Efforts to understand the developmental effects of poverty have relied on both experimental and nonexperimental studies. Experimental designs involving manipulation of family incomes are extremely rare. In four in- come maintenance experiments in the 1960s and 1970s, experimental treat- ment families received a guaranteed minimum income. Impacts on pre- school children, however, were not assessed. School performance and attendance were affected positively in some sites for school-age children, but not for high school adolescents. In two sites reporting high school completion and advanced education, these were higher for the experimental TABLE 10-2 Poverty and Affluence Among Young (Under 6 Years of Age) Children in 16 Countries Nation Percent Poor Percent Affluent Year United States 26.0 6.0 1997 United Kingdom 24.2 6.6 1995 Italy 19.2 4.6 1995 Canada 17.4 2.8 1994 Germany 12.4 2.8 1994 Israel 11.7 6.2 1992 Spain 11.6 8.3 1990 Netherlands 8.6 1.3 1994 France 8.0 4.7 1994 Finland 7.7 1.7 1995 Belgium 6.4 1.7 1992 Austria 5.9 0.7 1987 Denmark 5.6 1.3 1992 Norway 5.3 1.3 1995 Sweden 3.7 1.0 1995 Luxembourg 3.0 3.6 1994 NOTE: âPoorâ is defined as family-size-adjusted income less than 50 percent of country median income. âAffluentâ is defined as family-size-adjusted income greater than 200 per- cent of country median income. Equivalence scale is the square root of family size. SOURCE: Calculations by Lee Rainwater based on data from the Luxembourg Income Study.
278 FROM NEURONS TO NEIGHBORHOODS group (Institute for Research on Poverty, 1976; Kershaw and Fair, 1976; Salkind and Haskins, 1982; U.S. Department of Health and Human Ser- vices, 1983). Experimental Studies A new generation of experimental welfare reform programs have un- dergone recent evaluations. Three incorporate designs that include income- based incentives to work, such as wage supplements and income disregards. Milwaukeeâs New Hope Program involved experimentation with a package of work supports that included wage supplements, child care subsidies, health insurance, and supportive case workers (Bos et al., 1999) for families living in two poor inner-city neighborhoods. Large positive impacts (.25 to .50 standard deviation) were found on teacher-reported behavior and achievement of school-age boys but not girls. Changes outside the family, such as expanded use of after-school and community youth programs, rather than changes inside the family, such as parenting quality, maternal mental health, and family routines (which were unaffected), appeared re- sponsible for the child impacts. Given the multifaceted nature of the New Hope treatment, it was impossible to determine how much of these impacts were caused by the increased family incomes, other components of the intervention, or New Hopeâs particular bundling of these resource compo- nents. Unfortunately, preschool outcomes were not assessed. The Minnesota Family Independence Program (Gennetian and Miller, 2000; Miller et al., 1997) and the Canadian Self-Sufficiency Program (Lin et al., 1998; Morris and Michalopoulos, 2000) are among those that are assessing impacts on young childrenâs well-being. In the Minnesota pro- gram, children were ages 2 to 9 when their families were enrolled, with two-thirds age 6 and younger. The Canadian study involved a broader age range, including a group of children who were less than 3 years old when their parents were enrolled. They were assessed when they were 3 to 5 years of age. Like New Hope, both programs adopted a welfare reform strategy that emphasized rapid entry into work combined with provisions to ensure that family income improved as well.6 Preliminary findings from the two studies are remarkably similar. The youngest children were largely unaffected by their mothersâ participation in the program. This was true of cognitive outcomes, social behavior, emotional well-being, and child health. While both programs increased the time young children spent in child care 6The Minnesota program enabled families to keep both their wages and a generous amount of their former welfare benefits through an income disregard mechanism. The Canadian program supplemented earnings.
FAMILY RESOURCES 279 and, in particular, time in center-based arrangements, they had no effect on parenting behavior, the high rates of depressive symptoms that character- ized the mothers in both programs, or the home environment (as measured by the HOME scale). To the extent that positive child outcomes emerge from these experiments, they are restricted largely to school-age children. A fourth set of studies were based on work and education strategies that did not include economic incentives, such as those included in the three experimental programs described above (Hamilton, 2000; McGroder et al., 2000). These were conducted in 11 sites of the National Evaluation of Welfare-to-Work Strategies (NEWWS). There were no consistent impacts on young childrenâs development. Nonexperimental Studies Nonexperimental research using longitudinal data has shifted from studying poverty as an unchanging status (poor versus not poor) to under- standing how particular characteristics of poverty affect development for different age groups. This has focused attention on the depth, duration, and timing of poverty in childhood (Brooks-Gunn and Duncan, 1997; McLeod and Shanahan, 1993). What have we learned? Of particular im- portance is emerging evidence suggesting that family income may exert its most powerful influence on children during the earliest years of life (Duncan et al., 1998b). Using data from a nationally representative sample of chil- dren and families, Duncan and his colleagues related childrenâs completed schooling to family income averaged over three age spans: 0-5, 6-10, and 11-15 years. Family income during childrenâs preschool years, which are most distant from their decisions about leaving school, appeared far more important than income during middle childhood. Income during adoles- cence mattered, but primarily for entry into college. Moreover, early child- hood income effects were particularly strong in the lower income ranges. Controlling for income later in childhood as well as for demographic char- acteristics of households, a $10,000 increment to income averaged over the first five years of life for children in low-income families was associated with a 2.8-fold increase in the odds of finishing high school. This analysis suggests that for children in families experiencing economic hardship, in- come in the preschool years matters more for childrenâs education attain- ment than does income later in childhood. We have also learned that a householdâs long-term economic status has a much greater association with achievement and behavior problems than do single-year income measures (Blau, 1999). There appear to be larger impacts of income increments on low-income than higher-income families (Duncan et al., 1998b; Mayer, 1997; Smith et al., 1997), although this is not found consistently (Blau, 1999). Finally, although weâve seen that
280 FROM NEURONS TO NEIGHBORHOODS poverty combined with parental work appears to be more beneficial for children than nonworking poverty, correlational evidence suggests that young childrenâs outcomes may be affected positively by a transition from welfare to work only if that transition lifts the familyâs income above the poverty line (Moore and Driscoll, 1997; NICHD Early Child Care Re- search Network, 1999b). What we donât yet know, but hope to untangle with forthcoming evidence from welfare experiments that include child assessments, is whether these work- and poverty-related developmental patterns are due to work or income per se. They may instead be due to preexisting differences among parents who fail in their efforts to secure employment, work in low-wage jobs, or secure higher-wage jobs that are also, not inconsequentially, more likely to provide health insurance, family leave, and other benefits. In sum, young children are more likely than any other age group in this society to live in poverty, and poverty during the early years is more power- fully predictive of later achievement than is poverty at any subsequent stage of development. Children living in poverty are more likely today than in the recent past to have working parents, many of whom work consistently and for substantial hours. While the weight of the evidence indicates that parental work is usually a neutral or positive influence in the lives of young children, particularly for those living in poverty, its benefits appear to be attenuated or lost in the presence of low wages that sustain rather than ameliorate poverty, low job complexity, and perhaps employment that occurs during a childâs first year of life. The new generation of welfare reform studies provides some of the only experimental evidence available about the effects of providing increased income to working-poor families with young children, particularly those who previously had a history of reliance on public assistance. They suggest that, in the absence of positive effects on young childrenâs home environ- ments, parental mental health, and parenting, increases in family income and reductions in poverty are not in and of themselves sufficient to benefit young children. Yet there is no evidence that the children are harmed, and the evidence of positive outcomes for school-age children raises hopes for improved outcomes for the young children as they reach school age. At the same time, this evidence raises the question of âwhat would it take?â to improve the well-being of younger children in the context of efforts to improve the work effort and earnings of their mothers. The research on parentsâ mental health reviewed in Chapter 9 and the early intervention literature reviewed in Chapter 13 suggests that a promising answer would involve making the most of the opportunity that welfare reform presents to link families to both mental health and early intervention services.
FAMILY RESOURCES 281 Parental Schooling Large, positive associations between parental schooling levels and childrenâs achievement and behavior are among the most substantial and replicated results from developmental studies. It would thus be reasonable to expect that the recent increases in the educational attainment of parents of young children would bode well for them. Between 1974 and 1997, the proportion of children whose mothers had not graduated from high school dropped nearly in half, from 30 to 17 percent, while the proportion whose mothers had graduated from college nearly doubled, from 13 to 24 percent (Figure 10-3). Trends in fathersâ schooling were similar, although not quite as dramatic. Changes in parental schooling levels among young black chil- dren were even more favorable than among white children, although in 1998 it was still the case that much larger fractions of black (21 percent) than white (8 percent) children had mothers who had not completed high school. In stark contrast, however, the low schooling levels of Hispanic 0 5 10 15 20 25 30 35 Living in poor families Living in affluent families Mothers have a college education Mothers work full-time, full-year Not living with two parents 1974 1997 Pe rc en t o f c hi ld re m FIGURE 10-3 Trends in the socioeconomic resources of young children, 1974- 1997. SOURCE: Untabulated data from the Current Population Survey.
282 FROM NEURONS TO NEIGHBORHOODS immigrant parents led to distressingly low maternal schooling levels for Hispanic children as a whole; in 1998 nearly half (45 percent) of young Hispanic children had mothers who lacked a high school diploma. Parental education levels are strongly associated with the home literacy environment, parental teaching styles, and investments in a variety of re- sources that promote learning (e.g., high-quality child care, educational materials, visits to libraries and museums) (Bradley et al., 1989; Laosa, 1983; Michael, 1972). These dimensions of what economists refer to as human capital are indisputably linked to early learning and educational attainment once children begin formal schooling. Nevertheless, surprisingly little is known about the developmental im- plications of these trends. In policy terms, if a program could somehow increase the number of young mothers completing high school, how many spillover benefits would be expected in their childrenâs development? Evi- dence from developing countries indicates that educating mothers through at least the primary grades benefits public health, reduces fertility, and improves their childrenâs literacy and verbal skills (Dexter et al., 1998; Hobcraft et al., 1984; Richman et al., 1992). For the purposes of this U.S.- based review, the issues are more subtle, with policy changes typically involving the motherâs completion of a general education development (high school equivalency) certificate, a final year or two of formal high school, or a year or two of junior college. What evidence is there that these kinds of changes benefit children? Two studies have taken advantage of the fact that young mothers may acquire more formal schooling between the births of first and subsequent children to estimate whether achievement and behavioral differences be- tween earlier- and later-born siblings are related to increases in motherâs formal schooling. The results are mixed, however, with one study (Kaestner and Corman, 1995) reporting no effect of increased maternal education on young childrenâs achievement scores, and the other (Rosenzweig and Wolpin, 1994) reporting that an additional year of maternal schooling had a modestly positive effect, and more specifically that a mothersâ enrollment in school during a childâs first three years had a significant and large posi- tive effect on childrenâs receptive vocabulary. Thus, the jury is still out regarding the role of increased parental educa- tion in promoting the well-being of children. From a policy point of view, it may well be that the increments in skills associated with the completion of high school or an associate degree are too small to make much of a difference for children. This does not, however, answer the question of whether larger changes in parental education or changes that involve ac- quiring basic literacy would benefit young children.
FAMILY RESOURCES 283 Family Structure Family structure is often included among the dimensions that scientists study when trying to understand how the availability or lack of resources in families affects child development. Not surprisingly, the configuration of resources in single-parent families is often quite different from that in two- parent families (McLanahan and Sandefur, 1994). Single parents are most often mothers, and single-mother families face much higher rates of poverty than two-parent families. Among working adults, unmarried women main- taining families have the highest risk of living in poverty (Klein and Rones, 1989; Thompson and McDowell, 1994). Many children in single-parent households have fewer relationships with male role models or nonmaternal adults that might be important for their social development (Levine-Coley, 1998). Time constraints faced by single parents may affect their ability to supervise their children and participate in their activities (Amato, 1993; Levine-Coley, 1998). These factors may, in turn, be associated with dimin- ished emotional supports and lower levels of cognitive stimulation in the home environment (Amato, 1993; Levine-Coley, 1998; Miller and Davis, 1997). The circumstances and adaptations of parents vary greatly among single-parent families, as do the amount and types of resources they make available to their children. Accordingly, there is growing interest in how single parenting comes about and what alternative forms of support exist. On average, children raised by single parents have lower levels of social and academic well-being than do children from intact marriages (Cherlin, 1999; McLanahan and Sandefur, 1994)âa finding that has fueled widespread concern about the large and persistent decline in the proportion of young children living with two parents. Between 1974 and 1997, the proportion not living with two parents rose from 18 to 31 percent (Figure 10-3). Two- parent family structures have declined much more rapidly among black (a 16 percentage point decline) and Hispanic (18 percentage points) than white (10 percentage points) families. As of 1998, only 35 percent of young black children lived with two parents, compared with 63 percent of young Hispanic and 79 percent of young white children. Most of this decline can be accounted for by the increase in the proportion of young children living with never-married mothers rather than divorced or separated mothers. Indeed, in 1998, more than three-quarters of young children living in mother-only families had mothers who had never been married. What do we know about how these trends may be affecting the devel- opment of young children? Studies focused on divorce find that most chil- dren have a difficult time during and shortly after the divorce process (Hetherington and Stanley-Hagan, 1999), and that the problems are larger for their behavior than for school achievement (McLanahan, 1997). Nev- ertheless, although difficulties may reemerge later in life, recent reviews
284 FROM NEURONS TO NEIGHBORHOODS suggest that the vast majority of children from divorced families do not exhibit severe or enduring problem behaviors (Amato and Keith, 1991). The very few studies that have investigated the effects of divorce on pre- school children have found that divorce typically has small negative effects on preschoolersâ social adjustment, but no effects in other domains. The largest effects of divorce on children are found among children in primary school (Amato and Keith, 1991). It is also of interest that studies using better designs (e.g., better controls for differences between divorcing and stable couples, more representative samples) tend to find smaller effects (Amato and Keith, 1991). Divorce is but one of several routes into single parenthood, and it is important to distinguish it from other routes, particularly childbearing by unmarried women. Unfortunately, most research has focused only on the effects of divorce on children or has pooled together all single-parent fami- lies (McLanahan, 1997). The few studies that have addressed this question have found few differences between children of divorced and never-married parents; both groups are at risk for poorer achievement and behavior com- pared with children from intact families (Cooksey, 1997; McLanahan, 1997). As already noted, this risk is largely accounted for by differences in the socioeconomic resources available to single parent families. A forth- coming study on âfragileâ families will focus directly on children born to never-married parents (see Box 10-1). But many questions remain unanswered. Are these differences caused by the family structure, or do they reflect preexisting differences between children in intact and single-parent families? Does the effect of single- parent family structure depend on the age of the child? Unfortunately, research to date has not supplied clear answers to these questions. In sum, the central challenge facing those who study children of single parents is one of disentangling the effects of family structure from the effects of the diminished resources that typically characterize single-parent families. Those who have tackled this challenge largely agree that, while growing up in a single-parent family increases the odds that children will do less well in school and exhibit behavior problems, these outcomes derive largely from the socioeconomic realities of single parenthood (e.g., lower income, less parental time from both mothers and fathers), rather than from any direct effects of living only with one parent. Nevertheless, this topic remains an area of active controversy among scientists and politicians alike. Role of Genetic Factors As a part of a review of evidence linking differences in family resources to differences in child outcomes, it is critical to address the challenges to this literature that have been raised recently by behavioral geneticists. If
FAMILY RESOURCES 285 BOX 10-1 The Fragile Families and Child Well-Being Study The Fragile Families and Child Well-Being Study is designed to study unmarried parentsâtheir relationship and their resourcesâto learn more about them generally and to determine how outside factors and public policies affect them and the health and well-being of their children. Re- searchers are particularly interested in examining how Temporary Assis- tance to Needy Families (TANF) work requirements and time limits, as well as stricter paternity establishment and child support enforcement (as set down in the Personal Responsibility and Work Opportunity Reconcil- iation Act) will affect unmarried parents and their children. Specifically, the Fragile Families and Child Well-Being Study will be examining four major issues: 1. The circumstances and resources of new, unmarried parents, es- pecially fathers (i.e., How many parents have steady jobs? How many fathers want to be involved in raising their children?) 2. The nature of the relationship between unmarried parents. (i.e., How many of these couples are involved in stable relationships? Do they expect to marry? Are they experiencing high levels of conflict or domestic violence?) 3. Factors that affect new, unmarried parentsâ relationship (i.e., What pushes new unmarried parents together or pulls them apart? How are their living arrangements and parentsâ behavior affected by public poli- cies?). 4. The long-term consequences for parents, children, and society of new welfare regulations, especially the implementation of stronger pater- nity establishment rules, stricter enforcement of child support payment requirements, and changes in the delivery and financing of health care and child care. The study began in 1997, and data will continue to be collected through 2004. More than 3,600 unmarried couples are participating in the study, which is being conducted in 21 cities across the United States. In addition, information is also being collected from more than 1,100 mar- ried couples to be used as a comparison group. Researchers sought out participants in the hospital and interviewed the mother within 24 hours of her childâs birth. The fathers were then interviewed as soon as possible after that. Follow-up interviews will be conducted with both the mother and the father when their child is 12, 30, and 48 months old. In addition, the childrenâs health and development will be assessed at home when they are 48 months old. Funding for the study is provided by the National Institute of Child Health and Human Development and a consortium of national and local foundations. Principal investigators are Sara McLanahan and Irving Garfinkel (see McLanahan and Garfinkel, 2000).
286 FROM NEURONS TO NEIGHBORHOODS family resource differences derive primarily from parentsâ genetic endow- ments (e.g., cognitive, mental health) as behavioral geneticists argue (see, for example, Rowe, 1994), then any putative effects of resources on childrenâs development could in fact derive from genetic endowments rather than from any of the features of families that we have just discussed. This in turn would render policy efforts aimed at these features either moot or exceedingly difficult. There is little doubt that genetic influences need to be added to the long list of potentially important factors that deserve attention in studies that assess the effects of family resources on childrenâs development. The issue is how to ask the pertinent questions and how to explore them in research. Two kinds of evidence suggest that, even net of genetic endowments, family resources have important impacts on child development. These studies have used measures of family socioeconomic status (SES) (i.e., parental education and occupational status, income, family structure, and other measures of the family environment) to capture family resources. The first study compares the importance of socioeconomic factors on childrenâs achievement before and after statistical adjustments for parental genetic endowments. Phillips et al. (1998) used data from a nationally representative sample of mothers (the National Longitudinal Survey of Youth) to do exactly this. Specifically, they adjusted for what they called the âmotherâs cognitive genotype indexâ using her score on the battery of Armed Forces Qualifying Tests (numerical operations, arithmetic and math knowledge, paragraph comprehension), her class rank in high school, and the interviewerâs assessment of the motherâs understanding of the interview when assessing the association between socioeconomic status and childrenâs achievement. They found that genetic factors accounted for only about one-quarter of the SES-achievement association. Although far from trivial, this finding suggests that maternal cognitive endowments do not account for most of the socioeconomic contributions to childrenâs achievement. The second approach compares the association of socioeconomic status to child outcomes between children raised by biological and adoptive par- ents. If this association is due primarily to genetic factors, then the correla- tion between child outcomes and the SES of adopted (and thus genetically unrelated) parents should be much lower than the correlation between child outcomes and the SES of their biological parents. Loehlin, Horn, and Willerman (1989) found that the correlation between SES and child IQ for their sample of adopted children was only 18 percent less than the correla- tion for biological children. Scarr and Weinberg (1976) found similar patterns in their sample of black adopted and biological children. The magnitude of these reductions are in line with those found in the direct approach of Phillips et al. (1998), and also suggests that SES impacts on childhood IQ cannot be attributed primarily to genetic factors.
FAMILY RESOURCES 287 An even more dramatic illustration of the role of parent SES is provided by a recent study of children adopted between 4 and 6 years of age into families that varied widely in socioeconomic status (Duyme et al., 1999). This study directly addresses the question of the extent to which the envi- ronment, defined by the SES (fatherâs occupation) of adoptive families, can alter the cognitive development of children who tested in the very low range (IQs between 60 and 86) prior to adoption. The results are compelling. All children, whether adopted by low-, middle-, or high-SES families, had higher IQs after adoption. But more to the point, the children adopted by higher- SES families had significantly larger gains in IQ than did children adopted into lower-SES families (see Figure 10-4). Because the children and their adoptive parents are genetically unrelated, these SES effects carry no genetic influence. A very different line of reasoning leads many behavioral geneticists to doubt that family socioeconomic status matters for childrenâs development. Key here are the striking similarities in the abilities and personalities of twins and other siblings reared apart from one another. Indeed, these similarities are almost as large as those found for siblings who grow up together and, in the language of the behavioral geneticists, share the same environment (see Chapter 2). By behavioral geneticistsâ accounting, childrenâs âshared environmentsâ account for very little (almost always less than 10 percent, usually less than 5 percent) of the variability of ability and personality found in the population (Bouchard et al., 1990). Some have concluded from this evidence that the developmental consequences of per- Before adoption (age 3-5) After adoption (age 11-18) 100 95 90 85 80 75 70 High SES Middle SES Low SESIQ FIGURE 10-4 Mean differences in IQ by SES of adoptive family. SOURCE: Duyme et al. (1999).
288 FROM NEURONS TO NEIGHBORHOODS sistent family environmental influences, such as socioeconomic status and parenting, are remarkably small. Scarr (1992), for example, argues that family environments in the ânormal developmental rangeâ have little or no effect on childrenâs development. Harris (1995, 1998) relies, in part, on this evidence to argue that âparents donât matter.â Problematic in this reasoning is that socioeconomic status is not a permanent family characteristic shared by siblings. There is abundant evidence that the nature and effects of family socioeconomic influences vary sufficiently across time and among children to suggest that they are more properly conceived as belonging to the ânonsharedâ than the âsharedâ environmental category. Specifically, longitudinal studies based on nation- ally representative data have shown that family income is quite volatile (Duncan, 1988) and that siblings several years apart in age often experience quite different childhood incomes (Duncan and Raudenbush, 1999). Few children who live in single-parent families do so for their entire childhood (Duncan and Rodgers, 1998). This leads to the possibility that the effects on child development of economic conditions and single-parent family struc- ture, for example, may depend on the stage of childhood in which they are experienced. In fact, a study of the completed schooling of siblings in a national sample found that differences in family income specific to stages of childhood accounted for approximately 17 percent of the variation in dif- ferences in completed schooling (Teachman et al., under review). Thus, socioeconomic status contributes importantly to both the shared and non- shared environments of children, and one cannot use evidence on the unimportance of siblingsâ shared environments to argue that socioeconomic status does not matter for childrenâs development. In sum, the review of current thinking about genetic influences pre- sented in Chapter 2 reminds us of the importance of understanding the interplay between genetic and environmental influences over the course of development. This discussion of genetics and family resources counters more extreme portrayals of socioeconomic status as primarily reflecting genetic influences on development. At the same time, by revealing that environmental influences tell only part of the SES story, the evidence also reminds us that genetic influences warrant greater attention in studies that assess the effects of both shared and nonshared family resources on child development. INFLUENCES ON YOUNG CHILDRENâS DEVELOPMENT The processes by which family resources matter for childrenâs well- being have been difficult to elucidate (Belsky et al., 1986; Brooks-Gunn and Duncan, 1997). In this section, we summarize evidence on three pathways by which they may affect childrenâs development. The first focuses on the
FAMILY RESOURCES 289 parentâs own mental health. The second involves parental beliefs about childrearing. The third focuses not on parents directly, but rather on the home environments they create for their children. Although we discuss these pathways separately, it is likely that they interact and accumulate within families in disparate ways. Moreover, the processes we describe presume that parents and parental environments affect children, but not vice versa. However, as we have discussed throughout this report, children actively shape both their personal relationships and their environments more broadly (Bell, 1968, 1974; Sameroff and Fiese, 1990). Thus, some of the apparent associations between parental factors and childrenâs develop- ment undoubtedly reflect the pervasive and reciprocal ways in which chil- dren and parents affect each other. Parent Psychological Distress One way in which familiesâ economic resources may shape childrenâs lives is through their impact on parentsâ mental health. Low-income par- ents are at greater risk for depression and other forms of psychological distress, such as low self-worth and negative beliefs about control (see bottom panel of Table 10-1; Gazmararian et al., 1995; Pearlin and Schooler, 1978; Rosenberg and Pearlin, 1978). Over 40 percent of the poor women in two large samples participating in work and training programs, for example, scored at or above the cutoff for clinically significant depressive symptoms (Quint et al., 1997; U.S. Department of Health and Human Services, 1995). Nationally representative estimates of mental health prob- lems indicate that approximately 10 percent of poor and less-educated people in the United States have current major depressive episodesâtwice the rate of others who are more advantaged (Blazer et al., 1994). The psychological cost of economic hardship is compellingly portrayed by ethnographic work with poor families. Based on hundreds of interviews with low-income welfare recipients and working single mothers living in three cities, Edin and Lein (1997) describe their constant struggles to pro- vide food, housing, and other necessities, as well as to keep their children out of danger. Despite ongoing hardship, most of the mothers in the study adapted to their situations. They budgeted carefully and spent considerable time and energy making money in alternative ways. Despite their efforts, however, arrangements for child care, housing, and medical care were often precarious. Any one of a number of events, such as a family or extended- family illness, could cause major disruptions to their employment and fam- ily lives. The chronic and pervasive stress that Edin and Lein document suggests important potential links among economic hardship, mental health, and parenting. Psychological distress is more prevalent among low-income popula-
290 FROM NEURONS TO NEIGHBORHOODS tions because they experience more negative life events and have fewer resources with which to cope with adverse life experiences (Kessler and Cleary, 1980; McLeod and Kessler, 1990). In addition, Kessler (1982) demonstrated that low levels of education, income, and occupational status each make independent contributions to the variation seen in maternal psychological distress. While all socioeconomic dimensions may play a role, most developmental research has emphasized the effects of economic hardship on parentsâ mental health (McLoyd, 1997). The connection between economic hardship and mental health is im- portant because, as discussed in Chapter 9, poor mental health is related to harsh, inconsistent, and detached parenting. Research in this field has emphasized the associations among economic decline, economic strain, pa- rental psychological well-being, and childrenâs outcomes (e.g., McLoyd, 1997). For example, in the case of depression, mothersâ responses to the needs of their children tend to be less consistent and positive. Conse- quently, research on low-income families has explored whether depressive parenting patterns, or elements of these patterns, account for the relation- ship between economic hardship and childrenâs maladjustment. These as- sociations are often dependent on the age and gender of the children and, as with each aspect of socioeconomic status, they account for only part of the association between poverty and child well-being (McLeod and Shanahan, 1993; Watson et al., 1996). The work of Elder and colleagues (Elder, 1979; Elder et al., 1984, 1985) on children of the Great Depression found strong associations among economic hardship, parental psychological well-being, and childrenâs well- being in intact families. Fathers who experienced job loss and economic deprivation were more distressed psychologically and prone to explosive, rejecting, and punitive parenting. Preschool-age children in these families, especially boys, were more likely to exhibit problem behaviors, while ado- lescent girls were more likely to have lower feelings of self-adequacy and to be less goal-oriented. Adolescent boys fared better than either adolescent girls or younger children. Elder and colleagues (1985) speculated that the gender and age differences reflected different experiences in families during the deprived times. During this time of economic hardship, adolescent boys sought economic opportunities outside the home, which reduced the time they spent with their families, gave them a useful role to play, and may have reduced the amount of negative family interactions they experienced. Younger children and adolescent girls did not have the same access to buffers provided by extrafamilial activities. In more recent applications of Elderâs framework, similar processes have been found to operate in Midwestern farm families experiencing eco- nomic decline (Conger et al., 1992, 1994) and single-parent black families that experience chronic economic strain (McLoyd et al., 1994). These
FAMILY RESOURCES 291 studies confirmed the cluster of economic insecurity and decline, poor pa- rental mental health, punitive and less involved parenting, and poor adoles- cent outcomes. But there are important exceptions. Among the farm families and children, some became more involved in social institutions, such as schools and churches, and the adolescentsâ resilience had much to do with their connections to these social influences (Elder Jr. and Conger, 2000). Two studies have focused on young children. In one, reduced financial resources among black, rural, single-parent families were associated with lower maternal self-esteem, and lower self-esteem was associated with dete- rioration in family routines and the quality of mother-child interactions (Brody and Flor, 1997). These family processes were related to 6- to 9- year-oldsâ self-regulation, which in turn was associated with both academic and behavioral problems. A second study (Harnish et al., 1995) of ethni- cally diverse low-income children entering first grade found that the quality of mother-child interaction partially accounted for the effects of socioeco- nomic status and maternal depressive symptomatology on childrenâs exter- nalizing behavior. Substance abuse constitutes another risk factor associated with de- creased mental health and economic hardship among parents (Table 10-1). Research on children of substance-abusing parents has focused largely on drug exposure during pregnancy and childrenâs subsequent developmental outcomes (Harden, 1998; Mayes, 1995; U.S. Department of Health and Human Services, 1999c). Since the effects of drugs vary dramatically by drug type, substance-abusing parents may display a variety of patterns of impaired parenting. For example, drugs such as alcohol or marijuana may depress parentsâ moods, possibly resulting in withdrawn behavior, whereas cocaine may increase activity and elevate moods, possibly resulting in un- predictable or impulsive behavior. Few empirical studies have evaluated parenting among substance-abus- ing parents. Most of the evidence comes from studies that have docu- mented high occurrences of abuse and neglect among these parents; more harsh, negative, angry, threatening, and punitive interactions; and less re- sponsiveness to their children (Bauman and Dougherty, 1983; Bernstein et al., 1986; Colten, 1980; Leif, 1985). However, because drug abuse often co-occurs with other psychiatric problems and disadvantaged circumstances, it is hard to know whether the parenting practices of substance-abusing parents are uniquely impaired by their drug habits (Mayes, 1995). Finally, both the reporting and incidence of child maltreatment are higher among low-income than high-income families (Table 10-1; Trickett et al., 1991; Waldfogel, 1998). Studies of the etiology of child maltreat- ment suggest that while child abuse and neglect capture different behav- iors, children who are abused are also often neglected, and differences in
292 FROM NEURONS TO NEIGHBORHOODS co-occurrence patterns may have both different causes and different effects (Aber, 1994; National Research Council, 1993). Some have speculated, for example, that persistent poverty is more closely related to neglect, while abuse is precipitated by sudden economic loss (Aber, 1994). Research suggests that economic hardship is only one of several risk factors that may contribute to child abuse and neglect. Others include parental beliefs about childrearing and unrealistic expectations of childrenâs capabilities, social isolation, and psychopathology (Aber, 1994). Nevertheless, the contribu- tion of economic hardship to child maltreatment suggests that some of the negative effects of poverty on children may result from higher rates of abuse and neglect (McLoyd, 1997). In sum, the toll that low socioeconomic status takes on parentsâ mental health appears to have important effects on child well-being. This implies a pressing need to integrate economic and mental health policy at numerous levels, ranging from federal decision making to the implementation and evaluation of both economic interventions, such as welfare reform, and early interventions for children and families in local communities. We also have much more to learn about connections among low socioeconomic status, parental mental health, parenting behaviors, and child well-being. Important questions concern the differential effects of economic hardship on parents of infants, toddlers, older children, and adolescents; the pro- gression of effects on parents and children over time; and identification of factors that assist or undermine coping. In light of the more serious prob- lems that are associated with cumulative risk (Sameroff et al., 1987; Seifer, 1995), we also need to understand how different types and manifestations of psychological distressâwhich often occur together and in conjunction with protective factorsâcombine to affect child development. Parental Beliefs Only modest differences have been found in the typical parenting prac- tices and parent-child interactions of low-income and higher-income par- ents (Miller and Davis, 1997; Radziszewka et al., 1996). Higher-SES parents have been found to rely more than lower-SES parents on shame, guilt, and reasoning as disciplinary strategies and less on commands and imperatives (Kohn, 1969). These modest differences have been ascribed to parentsâ values or beliefs (Hoff-Ginsberg and Tardif, 1995; Kohn, 1959, 1963, 1969, 1976; Kohn and Schooler, 1973). One of the most often cited differences thought to affect parenting practices and childrearing is that lower-class parents value conformity, whereas higher social-class parents value self-direction (Gecas, 1979). In fact, mothers who value conformity have been found to voice more concern that being too responsive to a childâs crying and fussing will spoil the child (Luster et al., 1989). These
FAMILY RESOURCES 293 mothers were also less likely to emphasize the importance of reading and more likely to endorse controlling their childâs activities, with discipline if necessary. Mothers who valued self-direction, in contrast, were more likely to emphasize reading and exploration and were less concerned about disci- plining children or spoiling them by responding to their crying. Unfortunately, much of this work has stopped short of relating parenting values to childrenâs developmental outcomes (e.g., Harwood, 1992; Holden, 1995; Sigel et al., 1992). Consequently, it is unclear how much social class differences in values, beliefs, and parenting practices account for the differential development of children. Further complicating this research are the facts that important parental values (e.g., about aca- demic achievement) do not differ by social class (Warren et al., 1993) and that social class is only one of many potential influences on parentâs belief systems (Sigel et al., 1992). Finally, class differences in values have declined over time, suggesting that values may be a less important source of differ- ences in parenting practices than they once were (e.g., Alwin, 1984; Hoff- Ginsberg and Tardif, 1995; Wright and Wright, 1976). In sum, although researchers have consistently found a modest rela- tionship between socioeconomic status and some parental beliefs, research has failed to establish that these differences in values explain differences in parenting, or are consequential for young children. All told, class-related differences in parental values are unlikely to explain more than a small portion of variation in the development of children. The Home Learning Environment Family socioeconomic resources are closely associated with the home learning environments of poor children. Seminal work by Bradley and Caldwell (1980) identified important aspects of the home environment that are related to childrenâs well-being (see Chapter 9). Their widely used HOME (Home Observation for Measurement of the Environment; Bradley and Caldwell, 1984) scale assesses the type and frequency of interactions and learning experiences parents provide for their children, both inside and outside the home. Stimulation, emotional support, structure, and safety are associated with the well-being of both low-income and high-income chil- dren (Bradley et al., 1994). Although there is considerable overlap between the HOME scores of high- and low-income families, on average, high- income families received higher scores. This may be due in part to the fact that several of the HOME items depend on having more income (e.g., books in the home; Bradley et al., 1994). Poverty and persistent poverty are strongly associated with less optimal home environments (Garrett et al., 1994). Recent work has suggested that the home learning environment (as
294 FROM NEURONS TO NEIGHBORHOODS distinct from other aspects of the home environment, such as warmth and safety) might be particularly important for understanding childrenâs cogni- tive development. Several studies have found that the more positive home learning environments of high-income versus low-income children account for as much as half of the gap in test scores of preschool children, and as much as one-third of the gap in the achievement scores of school-age chil- dren (Smith et al., 1997). Miller and Davis (1997) found stronger associa- tions between a childâs poverty history and the quality of the home learning environment than between poverty and parent-child interactions. While income is a strong correlate of the home learning environment, so too are education and occupation. Miller and Davis also found that, after controlling for history of poverty, maternal educational attainment was still significantly and positively associated with the cognitive stimula- tion provided to the child at home. For example, mothersâ provision of verbal stimulation differs by education and occupation. As we discussed in more detail in Chapter 6, higher-SES mothers, compared with lower-SES mothers, âtalk more, provide more object labels, sustain conversational topics longer, respond more contingently to their childrenâs speech and elicit more talk from their childrenâ (Hoff-Ginsberg and Tardif, 1995:177; see also Hart and Risley, 1995; Hoff-Ginsberg, 1991). Efforts to understand why maternal education might be a particularly important aspect of socioeconomic status in determining mothersâ verbal interactions with their children have pointed to the fact that a motherâs educational attainment, but not her occupational status, correlates with her teaching style (Laosa, 1983). Specifically, mothers with higher levels of education use more verbal reinforcement, inquiry, modeling strategies, and reading with their preschool children. What remains to be understood is whether these findings are attributable to mothersâ relative schooling per se, or to genetic differences or other characteristics that distinguish mothers who acquire different levels of schooling and might affect such relevant aspects of parenting as the use of complex verbal strategies with their children (see Borduin and Henggeler, 1981). As described earlier, Parcel and Menaghan (1994) argue that jobs that are routinized, have low autonomy, and provide little opportunity for sub- stantively complex work erode parentsâ cognitive skills and, in turn, de- crease the likelihood that they will provide a cognitively stimulating envi- ronment for their children. In a longitudinal study, Parcel and Menaghan (1994) show that the complexity of mothersâ and fathersâ occupations has a positive association with the home learning environment that is indepen- dent of parental education, wage rate, and hours of work. However, they also found that the positive effects of job complexity on the home environ- ment depend on family and work demands and stresses, such as the birth of an additional child and the spouseâs work conditions.
FAMILY RESOURCES 295 In sum, the support for learning that characterizes young childrenâs home environments is strongly associated with both their cognitive devel- opment and their familyâs socioeconomic niche. As with maternal mental health, therefore, improving the literacy and learning environment of the home offers a potentially promising focus for efforts to promote early learning in poor families. The challenge, as illustrated by our review of parenting interventions (see Chapters 9 and 13), is finding effective ap- proaches to accomplishing this goal. SUMMARY AND CONCLUSIONS The past quarter-century has produced many changesâsome favor- able, many notâin familiesâ time, money, education, and other socioeco- nomic resources. Income inequality has increased, producing both more poverty and more affluence among families with young children. The average parental schooling level has increased. More young children are growing up in single-parent homes, and many more mothers with young children now hold full-time jobs than before. Finally, more children are growing up in poverty today than was the case 25 years ago. These trends hold both the promise of improved child well-being and the risk of increased problems. Their effect on an individual child will depend on the mix of positive and negative influences affecting his or her own family. Their effect on this generation of young children will depend on the broader landscape of how many children are affected by which influences, and what steps society takes in response to them. On balance, however, the evidence suggests that while improved maternal education may have modestly positive effects on early development, the effects of shifting family structures and, to an even greater extent, of maternal em- ployment will depend on a number of accompanying conditions. However, the persistent economic hardship that affects so many children is likely to be highly detrimental, especially during the earliest years of life. If confirmed in future research, this evidence that poverty during the early childhood years is especially harmful suggests that tax and transfer policies affecting family economic status should pay much more attention to improving familiesâ incomes while children are young. The emerging evidence from welfare reform experiments suggests, however, that the suc- cess of such efforts (when the criteria for success emphasize the well-being of young children) may hinge on simultaneously linking families and chil- dren to early intervention and mental health services. Nevertheless, be- cause many children growing up in poverty become productive adults, it is most accurate to portray low socioeconomic status as reducing the chances of success rather than leading inevitably to diminished attainments. We found suggestive associations, but little strong evidence, that an
296 FROM NEURONS TO NEIGHBORHOODS intervention aimed at generating modest increases in parental education would produce measurable benefits for childrenâs development. Associa- tions found between parental occupation and childrenâs development sug- gest that characteristics of employment may have a modest impact on childrenâs development The literature on single-parent family structure shows that children living in single-parent families are at greater risk for poor developmental outcomes compared with children reared in two-par- ent families, although we have a limited understanding of the processes involved. The research on maternal employment and childrenâs development is generally reassuring to working parents. Nevertheless, we have learned that maternal employment is too complex a phenomenon for simple com- parisons between young children with and without working mothers to reveal consistent differences. Rather, it is the circumstances of work, such as the income it generates, the proportion of the day the infant is spending in the presence of a security-giving, trusted caregiver, and related effects on family functioning that lie at the heart of how maternal employment affects young children. In particular, there is now evidence that nonstandard working hoursâwhich now make up a major share of jobs for poor work- ing womenâpose risks for children; and that going to work for long hours during the childâs first year poses a risk to child development perhaps especially when trade-offs are involved from time in sensitive and stable parental care at home to time in poorer quality alternative care, as they often are. Some of the most promising efforts to understand how a familyâs re- sources affect young children have focused on the mental health of parents, associated effects on their parenting, and the quality of the home environ- ment, notably the support it provides for learning. Punitive parenting, reduced monitoring, parental psychological distress, and substance abuse, as well as less parental support for childrenâs early learning, are all more prevalent in low-income families. While these factors have often been studied in isolation, they are likely to occur in clusters which, in turn, place children at higher risk of poor outcomes.
S297 Growing Up in Child Care 11 econd only to the immediate family, child care is the context in which early development unfolds, starting in infancy and continuing through school entry for the vast majority of young children in the United States. It is the setting in which most children first learn to interact with other children on a regular basis, establish bonds with adults other than their parents, receive or fail to receive important inputs for early learning and language development, and experience their initial encounter with a school-like environment. Early and extensive en- rollment in child care has become the norm in U.S. society. Indeed, if children were only sporadically or briefly exposed to child care, it would not be the visible policy issue that it is today. In 1994, 10.3 million children under the age of 5 were in child care while their mothers worked, including 1.7 million infants under 1 year of age (U.S. Bureau of the Census, 1997). The vast majority of 5-year-olds are in kindergarten (88.5 percent in 1995) (Hofferth et al., 1998). Younger children have also been enrolling in center-based child care, preschool, and pre-kindergarten programs at increasing rates so that, by 1997, 45 percent of 3- and 4-year-olds and 22 percent of children younger than 3 were in these types of programs (Capizzano et al., 2000; Ehrle et al., 2000). But enrollment in child care begins long before this. In 1999, the National Household Education Survey, which asks all families about nonparental child care arrangements regardless of the employment status of the mother, reported that 61 percent of children under age 4 were in regularly scheduled