The Role of Family Socioeconomic Resources in the Black-White Test 
Score Gap among Young Children 

 

 

 

Katherine Magnuson 

University of Wisconsin-Madison 

and 

Greg J. Duncan 

Northwestern University 

 

Forthcoming in Developmental Reviews


The Role of Family Socioeconomic Resources in the Black-White Test Score 
Gap among Young Children 

 

 

Abstract 

This paper reviews evidence on the family origins of racial differences in young 
children’s test scores and considers how much of the gap is due to differences in the economic 
and demographic conditions in which black and white children grow up. Our review of the 
literature finds that the estimated size of the gaps varies considerably across studies. However, a 
surprisingly consistent result is that a collection of measures related to family socioeconomic 
resources appears to account for a little less than half a standard deviation of the black-white test 
score gap, regardless of the assessments used or the populations studied. We discuss the policy 
implications of these findings and suggest avenues for future research. 


The Role of Family Socioeconomic Resources in the Black-White Test 
Score Gap among Young Children 

 

 

Introduction 

In the fall of 1998, the U. S. Department of Education launched the kindergarten 
component of the Early Childhood Longitudinal Study (ECLS-K) – the most comprehensive 
assessment of school readiness among the nation’s kindergarteners. Test score results, drawn 
from a representative sample of over 13,000 children enrolled in both public and private 
kindergartens, show alarming gaps in school readiness between white and Asian students on the 
one hand and black and Hispanic students on the other. 

For math achievement – skills such as recognizing numbers and geometric shapes, 
counting, recognizing patterns – both black and Hispanic students scored about two-thirds of a 
standard deviation below whites, while Asian students outscored whites (Figure 1).1 

Two-thirds of a standard deviation is a large difference. On a 64-item mathematics test, 
white children averaged 25 correct answers whereas Hispanic and black children averaged only 
19 and 20 correct answers, respectively. These scores were derived using Item Response Theory 
(IRT), a method in which all children are first given a set of items designed to assess their skill 
levels, and then a second set of items tailored to their skills. Thus, none of the children was given 
all 64 items (National Center for Education Statistics, 2001). 

[Insert Figure 1 here] 

For reading achievement – skills such as recognizing letters and associating letters with 
sounds at the beginning of words – the racial and ethnic gaps were only slightly smaller. Black 
and Hispanic students scored just under half a standard deviation lower, while Asians again 
performed better than whites, in this case by one-third of a standard deviation. However, the 
reading assessment was given only to children who passed an English language screening test. 
Thus, the gaps shown in Figure 1 may understate differences in reading skills between groups 
because students who were not proficient in the English language were removed from the 
comparison. 

What might be causing these differences? One obvious possibility is that historical racial 
and ethnic inequalities in the US have created disparate socioeconomic circumstances for the 
families in which white, black, Hispanic and Asian children are reared, and these differences 
explain the racial and ethnic achievement gaps. The ECLS-K also asked children’s parents about 
their own schooling, occupations and household incomes. Combining these elements into a 
single socioeconomic status (SES) index and comparing across groups yields a pattern quite 
similar to that found for achievement. The average SES level of black kindergartners was nearly 
two-thirds of a standard deviation below that of whites (Figure 1). Hispanic children had even 
lower socioeconomic standing relative to whites. Also consistent with the test score patterns, 
Asian children enjoyed SES levels that were slightly higher than those of white children. 

1 The authors calculated these and other ELCS-K figures. 


With racial and ethnic gaps in family socioeconomic circumstances matching so closely 
the gaps in test scores, it is tempting to conclude that equalizing the social and economic 
circumstances of young white, black, Hispanic and Asian children would eliminate the 
achievement gaps. But there are many reasons why socioeconomic circumstances may not be as 
consequential for children’s achievement as one might think. Maybe what really matters is a 
collection of psychological dispositions of parents – for example depression – and thus the 
association between socioeconomic status and achievement for children may merely reflect the 
fact that both are higher in the case of better-adjusted parents. Or maybe the association between 
socioeconomic status and achievement stems from the poor health and developmental problems 
of the children themselves, which can lower a child’s academic achievement and reduce a 
family’s resources by limiting parents’ employment. 

Some behavioral geneticists also conclude that socioeconomic conditions are of little 
importance, but put forth a different logic. For them, genetic endowments of ability are a key 
determinant of parents’ education and income, family environments, and children’s test scores. 
Thus they point out that children reared in more affluent families score higher on achievement 
tests because genetic endowments are passed on from one generation to the next (Rowe & 
Rodgers, 1997). If parental mental health, child health, or genetic endowments are what really 
matter for children’s achievement, then increasing parents’ income or education without also 
improving these more fundamental causes would not boost achievement. 

This paper reviews evidence on the family origins of the ethnic and racial differences in 
test scores: how much of the gap is due to differences in the socioeconomic conditions in which 
children grow up? Because there are so few studies that consider these questions for Hispanic 
and Asian children, we focus our attention on the black-white achievement gap. We begin with a 
review of estimates of the size of the achievement gaps and how they change across childhood. 
Next, we review theories as to why family socioeconomic resources might matter for children’s 
early achievement, as well as studies that attempt to account for the black-white achievement 
gaps by examining differences in family SES. We conclude with a discussion of policy 
recommendations and avenues for future research. 

 

Racial Test Score Gaps from Birth to School Entry 

Although we focus on average test score gaps between racial and ethnic groups, we begin 
by pointing out some of the limitations of this approach. A first concern is with the categories 
used by the studies we review – white and black (as well as Hispanic and Asian). Presenting the 
average scores of students by broad, socially-constructed racial and ethnic categorizations 
obscures considerable diversity within these groups. In the United States, black group 
membership is determined largely by phenotypic traits, but nevertheless entails possible cultural 
differences, as well as minority status and its accompanying experiences of discrimination. 
“Black,” for example, refers to a diverse array of African and Caribbean immigrant groups as 
well as African Americans (Garcia Coll et al., 1996; Kao & Thompson, 2003). 

Second, our use of averages obscures considerable variation in skills within each of the 
racial and ethnic groups. Although the average gap in math achievement between white and 
black kindergarteners in the ECLS-K was two-thirds of a standard deviation, nearly one-quarter 
of black children outscored the typical (median) white student in reading achievement. 


Third, how to measure academic achievement, especially in the early elementary grades, 
is a controversial question. Although research typically focuses on standardized assessments, it is 
unclear whether other types of measures would produce different estimates of the achievement 
gaps. Turning again to the ECLS-K for illustration, teachers were asked to rate the mathematics 
skills of children in the study. Black children were rated about half a standard deviation lower 
than white children, a difference similar to what is found with the ECLS-K achievement test. 
There was even a gap -- more than a quarter of a standard deviation in this case -- between 
teacher-rated black and white children’s math skills among black teachers who rated both black 
and white students in their classrooms. 

Fourth, academic achievement is only one aspect of school readiness. When asked to 
name the most important determinants of readiness to learn, kindergarten teachers do not rank 
early achievement indicators such as knowing numbers, shapes, colors and letters very highly. 
Instead they talk about the importance of good physical health and social skills – including being 
able to communicate needs, wants, and thoughts verbally, being enthusiastic and curious in 
approaching new activities, taking turns and knowing how to sit still and pay attention (U.S. 
Department of Education,1993). 

The ECLS-K study also provides information on racial differences in several teacher 
reported measures of problem behavior (e.g., arguing, fighting, disruptiveness) and self-control 
(e.g., controlling temper, respecting the property rights of others). On average, blacks have 
higher levels of problem behavior and lower levels of self control than whites, although the 
magnitudes of the behavior gaps are smaller (.35 of a standard deviation or less) than for 
academic skills. Since few other studies are able to provide estimates of racial gaps in behavioral 
dimensions of school readiness, we focus on academic and cognitive test scores in the remainder 
of this paper. However, understanding the gaps in other dimensions of school readiness is an 
important goal for future research, and likely to be quite complicated given the difficulty of 
accurately assessing children’s behavior and social skills in large studies. 

 

How Large Are Racial Achievement Gaps? 

Virtually all studies of the achievement scores of young white and black children find 
large disparities, although the magnitude of the gap varies with the particular test given. Using 
recent data from the nationally-representative Panel Study of Income Dynamics (PSID), Yeung 
and Pfeiffer (2005) estimate gaps in math and reading skills for preschool age children that are 
remarkably similar to those found the ECLS-K (-.78 and -.43 of a standard deviation for math 
and reading respectively). 

Data from other studies, however, show somewhat larger gaps. Using national data from 
the Children of the National Longitudinal Survey (CNLSY), Phillips and colleagues (1998, Table 
4-2) estimate that 5- and 6-year-old black children scored more than a full standard deviation 
below white children on a picture vocabulary test score. Using the same children and test, Farkas 
and Beron (forthcoming) express racial differences in terms of the average age at which black 
and white children correctly identify 40 words in the test. For whites, this happened two months 
after the children’s 4th birthdays. For blacks, it was a full year later. By this accounting, blacks 
were already a year behind whites in their language development just prior to school entry. 


2 http://nces.ed.gov/nationsreportcard/reading/results2003/natsubgroups.asp, accessed on March 
25, 2004. The 2003 assessments are not comparable to scores from the NAEP long-term trend 
assessments, which are discussed later in the article, because their standard deviations have not 
been corrected to account for measurement error. The NAEP is the only one of the studies we 
review that adjusts its standard errors for measurement error. NAEP gaps will therefore be larger, 
ceteris paribus, than gaps estimated with other data. 

Brooks-Gunn and colleagues (2003) estimated black-white achievement gaps in a sample 
of low birth weight 5-year-olds enrolled in the Infant Health and Development Program (IHDP). 
Although their sample is not nationally representative, the study is noteworthy because it 
contains several measures of achievement as well as IQ. Using the same picture-vocabulary test 
as in the CNLSY, they find a very large racial gap, amounting to 1.3 standard deviations. The 
racial gap in the verbal component of their IQ test was 0.97 standard deviations, still large but 
considerably smaller than for the picture-vocabulary test. Because two different language tests 
administered to the same group of children at the same time produced such different estimates of 
racial skills gaps, we caution against putting too much weight on results from any single test. The 
potential for tests to be biased has long been recognized, with methods of administration, content 
and test conditions all being potential sources of discrepancies in both the size and explanation of 
the gap (Burchinal & Cryer, 2003; Helms, 1992; Jencks, 1998; Rock & Stenner, 2005; Suzuki & 
Valencia, 1997). 

The National Assessment of Educational Progress (NAEP) has measured children’s 
mathematics and reading skills for over thirty years, but because their earliest point of 
assessment is fourth grade, NAEP data reflect the influences of both preschool and early-
elementary-school environments. Large achievement gaps are evident in the 2003 assessments, 
with both black and Hispanic fourth graders scoring about 0.92 of a standard deviation below 
whites in reading and 1.08 standard deviations lower in math.2 

At a minimum, racial gaps amount to about half a standard deviation and some studies 
estimate the racial gap to be more than a full standard deviation. Why the difference in the 
relative size of the achievement gaps? To date, there is no clear explanation. The same studies 
using different measures of skills and achievement, as well as different studies using the same 
measures, arrive at strikingly different estimates of racial test score gaps. Rock and Stenner’s 
(2005) review of the ECLS-K, CNLSY, and IHDP data dismiss several hypotheses that might 
explain differing estimates of early racial test score gaps—including sample selection, test 
content, and racial bias. 

 

Do Racial Gaps Change Over the Course of Childhood? 

It is difficult to measure cognitive ability and achievement at very early ages and to track 
them consistently across childhood. Infants and toddlers are obviously unable to perform the 
reading and verbalizing tasks associated with conventional tests, and their varying abilities to 
focus on test tasks produce considerable measurement error even in the assessments designed for 
their developmental stage. Moreover, the explosive growth in children’s developmental 
capabilities in early and middle childhood makes it difficult to test the “same” dimension of 
ability or achievement at different ages (McCall & Carriger, 1993). Nevertheless, accumulated 


data suggest that the achievement gaps seen between black and white kindergartners are apparent 
by age 3, and may grow larger during preschool and the early school years. 

Brooks-Gunn et al. (2003) find a small amount of growth between ages 3 and 5 in the 
black-white picture vocabulary gap. Phillips, Crouse and Ralph (1998) find that kindergarten 
achievement gaps between black and white children grow by about 0.2 of a standard deviation 
from school entry to high school. Recent analyses of the ECLS-K (Fryer & Levitt, 2005) and 
Panel Study of Income Dynamics (Yeung & Pfeiffer, 2005) also show modest growth in racial 
achievement gaps during the early school years (Figure 2). Both studies find similar growth in 
reading and math gaps-- about 0.2 of a standard deviation over the first few years of school. 
Thus, gaps appear to grow after children enter school, but much and perhaps most of the gap 
appears to be present by school entry. Unfortunately, research to explain why these gaps may 
grow is sparse and largely inconclusive (Fryer & Levitt, 2005; Murnane et al., 2006; Phillips, 
Crouse & Ralph, 1998; Todd & Wolpin, 2005). 

 

[INSERT FIGURE 2 HERE] 

 

Have Racial Gaps Narrowed Over Time? 

A different issue is whether the racial or ethnic gap in school-entry test scores has 
narrowed over the past few decades. The ECLS-K and PSID were conducted in 1998, up to a 
decade after the CNLSY, IHDP, and NICHD assessments. Fryer and Levitt (2004) suggest that 
the smaller gap found in the ECLS-K compared with estimates from prior studies may reflect the 
fact that the young children’s achievement gap has narrowed over time. It is difficult to assess 
this argument, because tests have not been administered to national samples of kindergartners 
more than once. The NAEP has tested reading and math skills consistently since the 1960s, but 
never before fourth grade. 

Figure 3 plots the relative scores of black and white fourth graders taking the NAEP 
reading test. As before, we express these differences in standard deviation units, so the -1.04 
difference in 1973 indicates that black children scored one full standard deviation below whites 
in that year. Encouraging in Figure 3 is that the racial gap appears to have declined through the 
1970s and 1980s to 0.71 of a standard deviation in 1988. Similar patterns are found for the 
NAEP reading test gap, which starts at -0.97 in 1973 and falls to -0.74 in 1986. Less encouraging 
is the lack of progress during the 1990s, when the gap may have widened. By 1999 the 
respective gaps in reading and math were -0.90 and -0.84 of a standard deviation. The most 
recent data available for 2004 show a decrease in the gaps of about 25% since 1999. 

 

[INSERT FIGURE 3 HERE] 

Most of the narrowing of the gaps for fourth grade students over the 1970s and 1980s is 
due to larger increases over time in black students’ academic skills compared with increases in 
white students’ skills (Grissmer, Flanagan, & Williamson, 1998). However, the pattern of gains 
differs somewhat for the math and reading tests. Math test scores for black fourth graders rose 
steadily between the early 1970s and late 1980s, with the gains over the entire period amounting 
to an impressive 0.6 of a standard deviation. In the case of reading scores, black gains occurred 


early (nearly 0.5 of a standard deviation during the 1970s) and changed little in the 1980s. Then, 
black students’ reading scores actually fell slightly in the 1990s, before increasing again in 2004. 
It is unclear if the ground gained by the time of the 2004 assessment reflects a transitory 
fluctuation or a positive trend. 

Hedges and Nowell (1998) also take up the question of converging gaps, by comparing 
estimates of black-white test score gaps among high school seniors in six large, national studies 
that span the late 1960s to early 1990s. They found that estimates of test score gaps declined 
from over one standard deviation in the late 1960s and early 1970s to about 0.8 of a standard 
deviation in 1992, a decline of similar magnitude to what is found in the NAEP data. They 
conclude that the reduction in the gap is largely due to improved scores among black students in 
the lower end of the achievement distribution, with the greatest gains occurring in the 1960s and 
1970s. All in all, evidence provides little hope that achievement gaps will simply continue to 
disappear with the passage of time. Encouraging gains for blacks have occurred in the past few 
years, but gains have been the exception rather than the rule since the 1980s. 

 

Family Socioeconomic Status 

Consider the differences between a family with a single parent struggling to make ends 
meet by working at two minimum-wage jobs and a family with one highly-paid wage earner and 
a second parent who stays at home to care for their children. The potential range of material and 
psychological hardships faced by the first family and avoided by the second is vast (Bradley & 
Corwyn, 2002; Evans, 2004; Haveman & Wolfe, 1994; McLoyd, 1998). For example, the first 
family may experience a lower-quality home environment that exposes children to pollutants and 
toxins such as lead and provides fewer learning opportunities in the home, and lower-quality 
child care. In addition, greater stress for the first mother may increase her irritability, and reduce 
her warmth and responsiveness to her children. Strained parent-child relations may, in turn, 
reduce children’s wellbeing. Because there are systematic and often large differences in the 
amount of resources – particularly financial resources – across racial and ethnic groups, it is 
natural to suspect that family resource differences may account for much, if not all, of the racial 
and ethnic differences in early school readiness. 

 

Socioeconomic Status 

Some social scientists lump many indicators of financial and social resources under the 
umbrella of “socioeconomic status” (SES). SES refers to one’s access to economic and social 
resources and the social positioning, privileges, and prestige that derive from these resources 
(Hauser & Warren, 1997; Mueller & Parcel, 1981). Some scholars emphasize the economic, 
educational, and occupational determinants of SES, while others take a more expansive view, 
arguing that cultural differences in habits, tastes, linguistic patterns, preferences, and worldviews 
are the key determinants of SES (Entwisle, Alexander, & Olson, 2005; Lareau, 2003). While 
these approaches differ in what they offer as central explanatory constructs, both perspectives 
ultimately suggest that proximal environments, particularly family and childrearing 
environments, are the critical link between a family’s SES and a child’s wellbeing. We follow in 
this tradition, and in doing so emphasize gradients in social and economic resources, rather than 
cultural underpinnings of categorical social classes. 


Researchers’ differing conceptual views of SES lead to discrepant approaches to 
measurement. Because it may be difficult to measure a family’s access to economic and social 
resources or their position in a social hierarchy directly, social scientists often use one indicator, 
typically occupation or maternal education, or combine multiple indicators (e.g., parental 
education and income) into scales that indicate families’ relative position in a social hierarchy 
(Hauser & Warren, 1997; Mueller & Parcel, 1981). For example, most occupational prestige 
scales of SES are derived from a weighted sum of the average earnings and education level of 
particular occupations. The Hollingshead index, used frequently in developmental studies, differs 
from this tradition in that it is a weighted sum of an individual head of household’s occupational 
prestige and education (Duncan & Magnuson, 2003). 

The differences in SES levels shown in Figure 1 exemplify the single-indicator 
combination approach. Scholars using an index to measure SES often view social stratification as 
an organizing force in individuals’ lives, and presume that one’s social standing (or social class) 
is more important than the particular combination and gradients of economic and social 
resources that determine it. 

An alternative approach to measuring SES is based on the premise that distinct types of 
socioeconomic resources contribute to social inequality and stratification along differing 
economic and social dimensions (Bollen, Glanville, & Stecklov, 2001). For example, although 
parents’ educational attainments, incomes and occupations are related, each may affect children 
in different ways (Duncan & Magnuson, 2003). Rather than using a summary SES measure, 
proponents of this approach consider each component separately. Given the moderate to high 
correlations among SES components, it is difficult to sort out their separate effects, especially in 
school or community samples that lack the full range of population variability. Even with 
sufficient variability, an attempt to estimate the contribution of distinct SES components may 
understate the importance of either the constellation or accumulation of household resources. 
Nevertheless, scholars have concluded that components of SES have differential effects on 
parenting and children’s development, and should not be combined into a single scale (Bornstein 
et al., 2003; Duncan & Magnuson, 2003). 

When viewed from a dynamic systems perspective, regardless of measurement strategy, 
much of research on the effects of SES on children is reductionistic, in that it attempts to 
translate complex ecological processes into associations between a few simple measures (Lerner, 
2003). Developmental systems theory holds that individuals are embedded in multi-level 
ecological contexts (Bronfenbrenner & Morris, 1998). By focusing attention on the integrative, 
interdependent and synthetic nature of individuals and environments, a systems approach 
emphasizes transactional developmental processes, in which reciprocal effects occur across all 
ecological levels (Lerner, 2003). 

Despite the seeming complexity of a dynamic systems framework, it provides several key 
principles that may be useful for studying SES effects on children. First, it suggests that SES is 
one set of conditions that affects, and will be affected by, other contexts in children’s lives. Put 
another way, the effects of SES (or any single SES component such as household income) may 
depend on a host of other key contexts and conditions, as well as child characteristics. This 
insight has spawned a host of risk and resilience research seeking to understand how the 
accumulation of negative contexts associated with poverty influences children’s wellbeing as 
well as to identify which individual and contextual factors promote positive outcomes for 


children raised in economically disadvantaged families (Evans, 2004; Garmezy, 1991; Luthar & 
Zigler, 1991; Kim-Cohen et al., 2004; Masten, 2001; Werner & Smith, 1982). 

Second, developmental systems display both multifinality and equifinality. A set of 
conditions may be linked to several end states and divergent conditions may lead to the same end 
state. This argues that SES may be linked to a broad set of child outcomes and the processes by 
which SES influences children may not be uniform. Finally, a dynamic systems approach 
suggests that the timing and duration of SES resources may matter as much as the presence or 
absence at any one time point (Bronfenbrenner & Morris, 1998). For example, in the case of 
poverty, effects may differ depending on the age at which it is experienced, the timing of the 
outcome being measured, and duration of economic hardship (Duncan & Brooks-Gunn, 1997; 
Elder, 1974). 

Throughout this article we take a multidimensional approach to SES by concentrating on 
four key components of parental socioeconomic status that are particularly relevant for children’s 
wellbeing – income, education, family structure and neighborhood conditions. In the sections 
that follow, we summarize research on the links between SES and children’s achievement, as 
well as the gaps in our understanding. Unfortunately, few studies have fully employed a dynamic 
systems approach to understanding SES effects, although studies often incorporate some aspects 
of this perspective. Therefore, at times our review paints a relatively simplistic portrait of the 
links between SES and children’s achievement, although we attempt to highlight findings that 
are consonant with dynamic systems theory. 

Despite theory suggesting that parents’ occupations may be linked to their parenting 
behaviors (Kohn, 1963; Luster et al., 1989; Menaghan & Parcel, 1991), we do not review the 
parent occupation literature. To date, there is little convincing evidence that parents’ occupation 
per se influences children’s achievement (see Bradley & Corwyn, 2002 for a review), and there 
are no clear interventions that directly target it. 

 

Household Income 

Family economic resources can be measured in a number of ways but most often include 
household income and sometimes wealth. Household income is the sum of income received by 
all members of the household over some time period, typically a calendar year or month. 
Dividing household income by household size or, better yet, a poverty threshold based on 
household size, produces a more refined measure of its per capita command over resources 
(Citro & Michael, 1995). Contrary to popular belief, family income is often quite volatile, and 
this income instability makes a single year’s measure of household income a somewhat 
erroneous measure of “permanent” income. In contrast, averaging income over multiple years 
provides a more accurate portrait of economic resources. Income and wealth are positively 
correlated, but distinct. Wealth is the point-in-time “stock” of a household’s financial assets, 
including both “liquid” forms such as money in a savings account as well as “illiquid” forms 
such the net equity tied up in an owned home. 

At over $30,000, the gap between black and white children’s household income is 
substantial. Rates of poverty are equally disparate; in 2004, the black–white child poverty gap 
was nearly 20 percentage points. Even more striking are racial differences in wealth. In the mid 
1990s, the gap in net worth between the average black family and white family was well over 


$150,000 (Gittleman & Wolf, 2004). If household income, poverty or wealth is consequential 
for children’s development, then such large differences in economic resources may go a long 
way to explaining racial achievement gaps. 

Household income has most often been hypothesized to influence children’s proximal 
environments through at least two distinct, but related pathways—material deprivation and 
parental distress. It is easy to imagine how children are advantaged by higher incomes. Financial 
resources influence the quality of the goods and services parents can purchase for their families 
such as prenatal health care and nutrition; enrich learning environments both in the home and 
outside the home; enable families to live in a safe and stimulating neighborhoods; and, for older 
children, provide access to higher-quality schools and a college education (Bradley & Corwyn, 
2002; Duncan & Brooks-Gunn, 1997; Evans, 2004; Guo & Harris, 2000; Magnuson & Duncan, 
2002; McLoyd, 1998). Thus, children from higher income families may benefit from a range of 
positive developmental contexts which lower income children do not experience. At the other 
end of the spectrum, children’s exposure to multiple negative contexts may lead to elevated 
stress, which overwhelms their adaptive resources and thus lowers their level of wellbeing 
(Evans, 2004). 

The experience of material hardship is distressing for many parents. Being unable to meet 
household and other basic expenses may cause some poor parents to feel frustrated, helpless, and 
depressed (Conger et al., 2002; McLoyd, 1998; Mistry, Biesanz & Taylor, 2004). The worry 
and stress associated with the constant struggle to provide for one’s family leaves many low-
income parents with few psychological or other resources with which to cope with adverse 
experiences (Gallo et al., 2005). Given these strains, it is not surprising that studies find that 
many low-income mothers have poor mental health (Brown, Brody, & Stoneman, 2000; Gallo et 
al., 2005). Psychological distress, in turn, may compromise the quality of parent-child 
interactions, such that low-income parents may be less warm and supportive, as well as harsher 
and more punitive than more affluent parents (McLoyd, 1998). This lower quality parenting may, 
in turn, hinder the development of children’s socio-emotional wellbeing and academic 
achievement (Conger et al., 2002; Mistry, Biensanz, & Taylor, 2004; McLoyd, 1998). 

Despite theoretical predictions and vast correlational evidence, whether family income 
and poverty are causal determinants of children’s achievement and behavior remains a 
controversial issue. As mentioned earlier, some scholars argue that both low family incomes and 
low achievement are the by-products of genetic, psychological, and social differences between 
poor and near-poor families, which are the “true” causes of poor achievement (Mayer, 1997). 

In the absence of experimental studies, researchers have sought to capitalize on 
longitudinal data to better understand the effects of family income on children’s achievement and 
education. Duncan and Brooks-Gunn (1997) coordinated analyses of 12 groups of researchers 
working with 10 different longitudinal developmental data sets. Analyses were designed to 
isolate the effects of permanent income from other SES components, particularly family 
structure and maternal education. This effort resulted in several important findings. First, the 
effects of household income appear to be non-linear, with larger effects of additional income for 
families in the lower end of the income distribution. In addition, family income in early 
childhood appears to be more important for shaping ability and achievement than income during 
middle childhood or adolescence. Finally, persistent poverty and deep poverty are particularly 
detrimental to children’s wellbeing. More recent studies have confirmed these findings, using 
sibling comparison and change models to isolate the effects of income from other parent and 


child characteristics (Ackerman, Brown, & Izard, 2004; Dahl & Lochner, 2005; Dearing, 
McCartney & Taylor, 2001; Duncan et al., 1998). 

How large are the effects of income? Smith et al. (1997) estimated that persistently poor 
young children had achievement scores 0.3 of a standard deviation lower than nonpoor children-- 
a difference large enough to account for substantial share of racial gap in achievement. 
However, trying to rule out a longer list of alternative explanations for poverty-achievement 
correlations, subsequent studies using the same data have arrived at somewhat smaller estimates 
of poverty effects (Phillips et al., 1998). 

A series of experimental welfare reform evaluation studies undertaken during the 1990s 
provided a unique opportunity to observe how increases in family income affect poor children’s 
development (Morris et al., forthcoming). Although all of the experimental programs increased 
parental employment, only some of the programs also supplemented parents’ earnings thereby 
boosting family income. Comparing the effects of programs that only boosted parental 
employment with those that boosted both employment and income provides estimates of the 
extent to which increases in family income benefits children. Preschool and elementary-school 
children’s academic achievement was improved by reforms that increased income, but not by 
programs that only increased parental employment (Morris et al., forthcoming). When these 
same data were used to isolate the impact of income on the achievement of young children, 
estimates suggest a $1,000 increase in income over several consecutive years was associated 
with a .06 standard deviation increase in achievement. 

Other studies reach varying conclusions regarding the importance of household income. 
Blau (1999) and Mayer (1997) find null to weak impacts but do not focus on the early childhood 
period. Dahl & Lochner (2005) use income increases associated with an expansions in the earned 
income Tax Credit and find significant impacts on both reading and math achievement. A 
balanced reading of the evidence appears to support the idea that poverty probably matters for 
children’s achievement, although not as much as some of the early studies suggested. In 
addition, young children may be more susceptible to negative effects of poverty than older 
children, particularly deep and persistent poverty. 

Whether household wealth has an independent influence on children’s achievement is 
more uncertain. Recent forays in understanding the effects of family wealth on children have 
yielded mixed conclusions about whether family wealth predicts children’s achievement, holding 
constant permanent income (Orr, 2003; Phillips et al.,1998; Yeung & Pfeiffer, 2005). 

 

Parental Human Capital 

Human capital constitutes a second form of SES-based family resources, and includes the 
collection of parental skills acquired in both formal and informal ways that are valuable in the 
labor market and at home (Becker, 1981). Formal schooling is the most familiar and most 
studied form of human capital, and a large body of literature has demonstrated that obtaining 
higher levels of education increases individuals’ future employment and earnings (Card, 1999). 
Most studies consider parents’ education to be fixed during a child’s life. Yet it has become 
increasingly common for adults to accrue education in a discontinuous fashion, and to extend 
their schooling well into adulthood (Astone et al., 2000; Jacobs & Stoner-Eby, 1998). Attending 


school during adulthood is particularly common for economically disadvantaged mothers (Rich 
& Kim, 1999). 

Parents’ completed schooling varies widely by race. We found that among the ECLS-K 
kindergarteners, 18% of black mothers had failed to complete high school as compared with only 
7% of white mothers. At the other end of the scale, 28% of white, but only 9% of black mothers 
had completed a four-year college. There may also be disparities in the quality of schooling that 
black parents received compared with white parents, as studies routinely find that blacks attend 
lower quality schools than whites (Phillips & Chen, 2003). However, few studies have been able 
to consider whether the quality of parents’ schooling influences their children’s achievement (see 
Phillips et al., 1998, for an exception). With research suggesting that an additional year of 
parental education has equally strong associations with white and black children’s achievement 
(Davis-Kean, 2005; Phillips et al., 1998), it may be that differences in the average school quality 
white and black parents experience is not very consequential for their children’s development. 

Higher levels of parental education (and higher quality of education) may benefit children 
indirectly by leading to a higher family income. Less obvious are the ways in which parents’ 
skills may directly improve child well-being, for example, by improving parenting behaviors and 
parents’ abilities to accomplish their parenting goals (Bronfenbrenner, 1986; Hoff, 2003; Laosa, 
1980; Michael, 1972). Research has found that parents with higher levels of education promote 
their children’s achievement by holding higher expectations for their children, providing more 
stimulating learning materials and activities, engaging in higher quality instruction, using more 
varied and complex language and speech patterns, as well as becoming involved in and 
supportive of their children’s learning (Davis-Kean, 2005; Hoff, 2003; Laosa, 1983; Neitzel & 
Stright, 2004; Raviv, Kessenich, & Morrison, 2004; Richman, Miller, & Levine, 1992). With 
such a range of possible mechanisms linking parents’ schooling to their children’s development, 
it is interesting to note that most studies have considered these pathways in isolation. 
Consequently, the extent to which these pathways uniquely explain transmission of educational 
advantage is uncertain. In addition, the way in which these processes interact to promote and 
reinforce children’s achievement is not well understood. 

Studies of children’s development routinely find that children with more highly educated 
parents have higher levels of cognitive development and academic achievement than children of 
parents with lower levels of education (Haveman & Wolfe, 1995). Remarkably, the association 
between children’s cognitive development and parental education is evident as early as 3 months 
of age (Roe & Bronstein, 1988). Yet, most comparisons do not clearly isolate parental education 
as the cause of high achievement (Davis-Kean & Magnuson, 2004; Mercy & Steelman, 1982). 
Few studies are able to disentangle parents’ educational attainment from other sources of 
advantage, such as cognitive endowments, that may lead to higher levels of achievement among 
both parents and children. The few U.S. studies that have tried to isolate the effects of parental 
education per se find positive and policy-relevant effects of maternal and paternal education on 
children’s outcomes, with an additional year of schooling associated with an increase in 
children’s test scores of about 0.15 of a standard deviation (Neiss & Rowe, 2000; Rosenzweig & 
Wolpin, 1994). Results from studies of European data, however, have been less conclusive 
about whether parental education has positive effects on children’s development (Black, 
Devereux, & Salvanes, 2004; Plug, 2003). 

Perhaps because parental education is often employed as a control variable, few studies 
have sought to move beyond estimating average linear effects, and to more fully understand how 


effects and mechanisms of influence may differ according to family and child characteristics, 
particularly children’s developmental stage. There is some evidence that the effects of parental 
education may be non-linear, such that increasing the education of mothers with a high school 
degree or less boosts children’s achievement more than increasing the schooling of college-
educated mothers (Haveman & Wolfe, 1995; Magnuson, 2004). For example, a recent study 
suggests welfare recipients randomly assigned to participate in mandated education or training 
improved their young children’s academic school readiness by as much as a quarter of a standard 
deviation (Magnuson, 2003). 

 

Family Structure 

Single-parent families. Increasing rates of cohabitation, divorce, and non-marital 
childbearing have profoundly changed children’s experiences of family life. Currently, about 
one-third of all children are born outside of marriage, and over half will reside in a single-parent 
family at some point in their childhood. Relatively high rates of multiple partner fertility also 
contribute to the complex family structures of many low-income families (Carlson & 
Furstenberg, 2005; Meyers, Cancian and Cook, 2005). As is the case for education and income, 
family structure differences across racial groups are also large. Our calculations of the ECLS-K 
data indicate that 15% of white kindergartners live in single-parent households, whereas 50% of 
black children do. Racial differences are also apparent in the pathways into single-parent 
families, with black children being more likely to be born outside of marriage, and white 
children more likely to experience divorce. 

Economic and psychosocial resources in single-parent families differ markedly from 
those in two-parent families (McLanahan & Sandefur, 1994). Young children in single-mother 
families face poverty at five times the rate of preschoolers in intact families (50% vs. 10%), and 
the declines in children’s household incomes after a divorce are dramatic and lasting (US Census 
Bureau, 2004). Financial and time constraints faced by single parents may limit their ability to 
supervise and discipline their children, as well as provide emotionally supportive and cognitively 
stimulating home environments (Amato, 1993; Clarke-Stewart, 2000; Coley, 1998; Miller & 
Davis, 1997). Parental conflict and strain found in divorcing families may also be harmful to 
children’s development, especially if it compromises parents’ mental health and reduces the 
quality of their parenting (O’Connor et al., 2000). Children in single-parent families may 
experience several transitions in family life, including subsequent cohabitations, remarriages, 
separations, and divorces. This instability and complexity in their family relationships may pose 
additional risks to child wellbeing (Sigle-Rushton & McLanahan, 2003). Finally, due to the 
prevalence of low levels of father involvement among single-parent families, children in these 
households often have fewer relationships with male role models, which might be detrimental for 
their social development (Black, Dubowitz, & Starr, 1999; Coley, 1998). 

On average, children raised by single parents have lower levels of social and academic 
well-being than children from intact marriages (McLanahan & Sandefur, 1994). Studies find 
that most children have a difficult time during and shortly after the divorce process, and that 
problems are larger for children’s behavior than for school achievement (Ackerman et al., 2002; 
Amato, 1993; Amato & Keith, 1991; Hetherington & Stanley-Hogan, 1999; McLanahan, 1997; 
O’Connor et al., 2000). Children in elementary school are the most profoundly affected by 
divorce, and although difficulties may re-emerge later in life, studies suggest that the vast 


majority of children from divorced families do not exhibit severe or enduring problem behaviors 
(Chase-Lansdale, Cherlin, & Kiernan, 1995). 

Divorce is not the only path to a non-intact family, and it is important to distinguish 
families affected by divorce from never-married, cohabiting, and blended (remarried) families. 
Most research has focused only on the effects of divorce on children, or has lumped together all 
single-parent families (McLanahan, 1997). The few studies that have addressed this question 
find little difference between children of divorced and never-married parents; both groups are at 
risk for poorer achievement and behavior as compared with children from intact families 
(Carlson & Corcoran, 2001; Cooksey, 1997; DeLeire & Kalil, 2002; McLanahan, 1997). 

Although most children from non-intact families fare worse than children growing up in 
intact families, studies suggest that pre-existing differences in parent and family characteristics 
(e.g., educational attainment and income), rather than family structure per se, account for much, 
but not all, of the gap (Auginbaugh et al., 2005; Brown, 2004; Carlson & Corcoran, 2001; 
Clarke-Stewart et al., 2000). For example, findings from a recent adoption study suggest that the 
academic problems in divorced families may be explained by the transmission of genetic 
influences, but these factors do not account for elevated levels of problem behavior (O’Connor et 
al., 2000). 

Teen parenthood. Rates of teenage childbearing have been steadily declining over the 
recent past, dropping 22% between 1991 and 2000, from 62.1 to 48.1 births per 1,000 15-19 year 
olds (Ventura, Mathews & Hamilton, 2002). Despite this drop, the US continues to have high 
rates of teenage parenthood compared with other European countries. In the US, teen birth rates 
differ substantially by race. Our analysis of the ECLS-K study revealed that about 1 in 5 black 
children were born to mothers less than 20 years old, nearly twice the rate for white children. 

Typically, children of teen mothers face a constellation of socioeconomic-related 
hardships throughout their lives including single-parenthood, poverty, and lower maternal 
educational attainment (Coley & Chase-Lansdale, 1998). As outlined earlier, each of these 
circumstances may influence children’s outcomes through a variety of processes. However, a 
mothers’ young age may also pose a unique risk to children’s healthy development by limiting 
mothers understanding and expectations of motherhood (Sommer et al., 1993). More 
specifically, compared with older mothers, young mothers may have imprecise or erroneous 
knowledge about children’s development and parenting as well as negative perceptions about 
both their parenting role and their children’s behavior. Research suggests that this lack of 
cognitive preparedness and negative attitudes may translate into lower quality parent-child 
interactions, particularly lower levels of responsiveness, which in turn may result in poor 
developmental outcomes (Fergusson & Woodward, 1999; Miller et al., 1996). 

Although studies find that children of adolescent parents on average have poorer 
outcomes than other children, they also suggest great variability in outcomes (Coley & Chase-
Lansdale, 1998). In seeking to understand why some children of adolescent mothers do better 
than others, researchers have explored whether differences in maternal characteristics and 
parenting quality explain these differences. As is the case for single parents, it appears as if 
some of the poor outcomes are more accurately attributed to a host of other maternal 
characteristics, such as antisocial behavior and lower levels of education, rather than mothers’ 
young ages per se (Jaffee et al., 2001). Consequently, the extent to which children would 


benefit from their mothers’ postponing their childbearing by a few years is uncertain, although 
likely modest (Jaffee et al., 2001; Levine, Pollack, & Comfort, 2001; Turley, 2003). 

 

Neighborhoods 

Neighborhoods vary widely along a number of dimensions that may shape children’s 
development, although it is likely that adolescents are more susceptible to neighborhood 
conditions than young children, who spend little unsupervised time in neighborhood settings 
(Brooks-Gunn, Duncan, & Aber, 1997). The risks posed by low quality neighborhoods are most 
striking in high poverty urban communities plagued by violence, gangs, drug activity and 
environmental pollutants. In these communities, parents may not allow children to walk to 
school alone or play outside for fear that they will be in harm’s way (Jarrett, 1997). How might 
such a neighborhood influence children? Pathways identified in the research literature include: 
stress, which may be elevated as a result of community violence; lack of social organization, 
including few positive role models and a lack of consensus around common values, which might 
in result in little control over youth’s problem behavior; few institutional resources such as 
enriching schools and police protection; negative peer influences, which might spread problem 
behavior; and exposure to environmental toxins such as lead paint and dust (Ellen & Turner, 
1997; Evans, 2004; Jencks & Mayer, 1990; Sampson, Morenoff & Ganon-Rowley, 2002). 

Broad-based studies of neighborhood effects on preschool children indicate that despite 
the many possible ways in which neighborhoods may affect children, the associations between 
neighborhood characteristics and children’s outcomes are quite modest. A review of studies 
suggests that only between 5-10% of the variation in children’s achievement and behavior 
outcomes are explained by neighborhood characteristics, after controlling for family 
characteristics (Leventhal & Brooks-Gunn, 2000). 

A recent experiment provides a compelling test of the extent to which neighborhood 
residence matters for children’s outcomes. The Moving to Opportunity (MTO) experiment 
offered housing-project residents in five of the nation’s largest cities a chance to move to low-
poverty neighborhoods. Analyses of children’s achievement test scores and problem behavior 
collected four to seven years after families moved revealed no statistically significant differences 
between program and control group children, even among those who were preschoolers when the 
program began (Orr et al., 2003; Sanbonmatsu et al., forthcoming). So despite improvements in 
neighborhood conditions, particularly neighborhood safety, there were virtually no favorable 
treatment effects on children’s test scores, school success or engagement and problem behaviors. 
Why not? 

It appears that one of the possible factors behind the program’s lack of effects on children 
is that the large improvements in neighborhood poverty enjoyed by experimental-group families 
were accompanied by no more than modest improvements in the quality of the schools attended 
by their children (Sanbonmatsu et al., forthcoming). Evidence from a qualitative study of MTO 
families suggests that parents took advantage of school choice programs and enrolled their 
children in schools near their original neighborhoods (Popkin, Harris, & Cunningham, 2001). 
Another possible explanation is that while MTO-related neighborhood advantages appeared to 
improve the mental health of mothers, they did not translate into other kinds of household 
resources or advantages that might have improved children’s well-being (Kling et al., 2004; Orr 
et al, 2003). After moving, adults in the MTO program were still similar to their control-group 


counterparts in terms of their employment status, welfare dependence, family income, parenting 
practices, and connections to their children’s schools and to the parents of their children’s 
friends. 

Thus, results from the interim evaluation of the MTO residential mobility program 
indicate that residential mobility programs are not cure-alls for the achievement problems of 
children who live in public housing in a high-poverty neighborhood. It appears that interventions 
focused exclusively on neighborhoods, rather than on the larger constellation of factors directly 
related to the child, family and school contexts are unable to solve the myriad problems of 
children growing up in high-poverty urban neighborhoods. Nevertheless, it remains possible that 
small effects are meaningful, and as dynamic theory suggests, that neighborhood poverty may 
interact with other dimensions of SES resources in important ways. 

 

Nurture and Nature 

There is little doubt that children are born with differing social and cognitive capacities 
and that at least some of these differences are genetic in origin. At the same time, virtually all 
scholars agree that environmental conditions, including family environments, also shape 
children’s social and cognitive development (Plomin, 1994). Although, the relative importance 
of genetically and environmentally determined differences in explaining SES and racial gaps in 
achievement remains controversial (Huston, McLoyd, & Garcia Coll, 1997; Nisbett, 1998; Rowe 
& Rodgers, 1997), efforts to parse out genetic or environmental determinants of achievement or 
other behaviors are increasingly being called into question as accumulating evidence indicates 
that gene expression is the result of a complex interplay between genetic endowments and 
environmental experiences (Collins et al., 2000; Lerner, 2003). 

Although a full accounting of the controversies and consensus in the field of behavioral 
genetics regarding the achievement gap is beyond the scope of this article (Helms, 1992; 
Herrnstein & Murray, 1994; Huston, McLoyd, Garcia Coll, 1997; Jensen, 1969; Nisbett, 1998; 
Rowe & Rodgers, 1997), several points about the role of genes in group achievement differences 
are important. The first relates to the likely genetic bias in most studies of the contribution of 
SES to the test-score gaps. Because it is likely that genetic processes are linked with both 
children’s academic abilities and parents’ economic resources, researchers face the challenge of 
gene-environment correlations (Scarr, 1992). Several dimensions of parents’ genetic 
endowments are likely to influence the levels of socio-economic resources they provide to their 
children as well as to be, at least in part, inherited by their children (Walker, Petrill, & Plomin, 
2005). Therefore, analyses that estimate the contribution of family and neighborhood SES to 
test-score gaps may overstate the importance of SES unless research designs somehow 
incorporate genetic influences in their modeling (Rowe, 1994; Walker, Petrill, & Plomin, 2005). 

Suggesting that genetic factors may bias estimates of the importance of parental SES, 
however, does not imply that gaps in achievement between low and high SES children are 
largely determined by the inheritance of genetic traits. Environmental experiences are important 
in regulating the expression of genetic traits. A recent adoption study illustrated the importance 
of environment in shaping cognitive development by comparing the pre- vs. post-adoption 
outcomes of low-IQ children adopted around age 5 into low-, middle- and high-SES families 
(Dumye, Dumaret & Tomkiewicz, 1999). IQ growth was strikingly different by the SES of the 
adopting families (defined by father’s occupation), with the gains associated with adoption into 


high- and middle-SES families much larger than the IQ gains for children adopted into low-SES 
families. Thus, evidence suggests that SES, and not solely genetic endowments, was responsible 
for the differential gains. Perhaps more pertinent to racial achievement gaps, the Minnesota 
Transracial Adoption Study (Scarr, Weinberg, & Waldman, 1993) demonstrated that low-SES 
black children adopted into high-SES white families had IQs of 97 at age seven – close to the 
overall population mean. A different IQ test was administered at age 17, and as a result of 
different norming criteria the IQ scores of the black adopted children were lower, but the 
decrease in their scores (from 97 to 88) was similar to the drops of non-adopted white children 
(from 117 to 109) and other family members. Moreover, the correlations between the IQ scores 
of the adopted children and their adoptive parents did not decline over this time. 

Exactly how genetics should be incorporated into research on SES effects on 
achievement is not obvious. Research indicates that the heritability of achievement varies 
substantially between children reared in low- and high-SES families, and is strongest in 
advantaged families (Guo & Stearns, 2002; Rowe, Jacobson & Van den Oord, 1999; Turkehimer 
et al., 2003). For example, Turkheimer et al. (2003) found that the family (shared) environments 
of low-SES children accounted for about 60% of the variation in aptitude, whereas genetic 
factors accounted for almost none of the aptitude differences. For high-SES children, these 
fractions are almost exactly reversed. Thus, individual differences in higher-SES environments 
appear to be driven more by genetic processes, whereas individual differences in low SES 
environments appear to be more closely linked to environmental influences. This apparent 
differential heritability of cognitive ability in high- and low-SES environments complicates any 
simple-minded attempt to adjust for genetic bias to SES effects in estimating group differences. 

 

Can Family Socioeconomic Resources Explain Racial Achievement Gaps? 

Both theory and evidence suggest that the family socioeconomic environments in which 
children are reared may account for at least some of school-entry achievement gaps. This section 
summarizes results from a collection of what we refer to as accounting studies, which estimate 
the extent to which school-entry racial and ethnic achievement gaps are the result of SES 
differences across groups. These studies do so primarily by observing the extent to which the gap 
between young white and black children’s test score performance is reduced when measures of 
family SES are included in regression analyses with few other controls (for a discussion of 
black- white high school achievement gaps see Hedges & Newell, 1998) . 

As with our section on SES, we begin with a warning about the difficulty of establishing 
causal relationships: accounting studies assume that SES differences cause achievement 
differences. To the extent that this is not the case, estimates of how much of the achievement 
gaps are accounted for by SES or other unmeasured differences will overstate the impact of 
policies directed at eliminating the SES differences. On the other hand, poor measurement of 
SES resources or an incorrect specification of the functional form of these associations may lead 
research to understate the role of SES in explaining gaps. 

 

Explaining the Gaps 

Figure 4 shows representative results from two recent studies of black-white differences 
in test scores around the point of school entry. Math and reading results are presented from our 


own calculations using ECLS-K data and from Yeung and Pfeiffer (2005) using PSID data. The 
first bars show the simple, unadjusted mean racial and ethnic differences. As we note in the 
introduction, black children score two-thirds of a standard deviation lower than whites in math 
and close to half a standard deviation lower in reading. The second set of bars indicates the size 
of racial gaps, holding constant family socioeconomic resources. 

How much of these racial gaps are due to differences in socioeconomic resources? 
Following Fryer and Levitt (2004), we adjust for an SES composite, an indicator of whether the 
family is poor, age at entry into kindergarten, age of mother at the time of the birth, a basic 
measure of family structure (two parents no siblings, two parents and siblings, single parent no 
siblings, single parent with siblings) and whether the mother received benefits from the Women, 
Infants and Children (WIC) Nutrition program. This handful of SES-related measures accounts 
for nearly all (about .46 of a standard deviation) of the black-white difference in math and 
reading test scores. Fryer and Levitt’s analysis excludes the poverty and family structure 
indicators and includes the number of books in the house and the child’s birthweight. Given our 
resource-based definition, these two measures may not be appropriate as SES indicators since 
they may reflect parents’ preferences, values and health more than their social and economic 
resources. Nevertheless, their results are virtually identical to ours. 

 [Insert Figure 4] 

A set of analyses conducted by Yeung and Pfeiffer (2005) with data from the Child 
Supplement to the PSID closely parallel Fryer and Levitt’s (2004) methods. Yeung and Pfeiffer 
estimated the magnitude of black-white gaps on two subtests of the Woodcock Johnson 
Achievement Tests, one measuring children’s early reading skills and the other measuring early 
mathematical skills. They also considered the extent to which family SES explains these gaps. 
Of particular interest are the analyses conducted with preschool children (ages 3-5) which 
yielded racial gaps of a similar magnitude to those found by Fryer and Levitt (2005) (Figure 4). 
Moreover, Yeung and Pfeiffer found that measures of the log of family income, parents’ 
educational attainment, occupational prestige and net wealth explained nearly the entire reading 
gap and about 0.36 of a standard deviation of the math gap. 

The conclusion that SES differences explain nearly all of the racial and ethnic test score 
gaps is unusual in previous studies. For example, Figure 5 summarizes results from a study of 
gaps in the picture/vocabulary scores (PPVT-R) of black and white 5- and 6-year-olds from the 
CNLSY (Phillips et al., 1998). Not only is the unadjusted gap much larger in the CNLSY than 
ECLS-K and PSID data – more than one standard deviation – but measures of family income and 
wealth accounts for only about 0.2 standard deviations of the gap (Phillips et al., 1998, Table 4-
3, column 8). One possible source of difference in the NLSY data is the way in which the 
household income variable is calculated. It involves adding several reports of different types of 
income together, which as Cole and Currie (1994) demonstrate introduces more opportunity for 
“measurement” errors as decisions are made about how to handle each source of income. Two 
NLSY conventions are noteworthy: total household income does not include the earnings of 
mothers’ non-married cohabiting partners and income is top coded at $100,000. In addition, for 
youth residing with their parents, youths’ income was coded as the total household income when 
information on parental income was missing and the questions about parental income changed 
after 1986. Cole and Currie (1994) “cleaned” the total household income variable for the 1979 
NLSY female sample, and found that they corrected the net household income variable for about 
10% of the sample each year from 1980-1990. Certainly measures of household income in other 


studies also include some measurement error, although studies like the PSID devote considerably 
more resources to cleaning their income data, and do not employ the same coding conventions. 

It is also worth noting that alternative specifications of the CNLSY employed by Phillips 
and colleagues (1998) show that including just a few additional indicators of family SES, such as 
occupational prestige or parental education, does little to further explain the gap. However, a 
much larger set of family background measures that includes indicators of the mothers’ school 
experiences and cooperation with the interviewer, explained a greater proportion of the gap (.47 
of a standard deviation, Phillips et al., 1998, Table 4-5, Column 4). 

Analyses conducted with data on five-year olds in the IHDP study are also presented in 
Figure 5 (Brooks-Gunn et al., 2003). The IHDP study administered both the PPVT-R (the same 
measure used in the CNLSY) and two sub-scales of the Wechsler Preschool and Primary Scale 
of Intelligence (WPPSI)—verbal and performance (non-verbal thinking and problem solving). 
As with the CNLSY, the verbal test score gaps amount to about a full standard deviation. 
Nevertheless, about 0.45 of a standard deviation in the gaps for the verbal measure appears to be 
the result of the SES differences between white and black children, as measured by the families’ 
3-year average income-to-need ratio. The performance gap is slightly less than a standard 
deviation, but the same measure of family income appears to explain about 0.4 of a standard 
deviation of the gap. 

Following Fryer and Levitt’s (2004, 2005) methods, Murnane and colleagues (2006) 
conducted a set of analyses with data from the NICHD Study of Early Child Care. Although this 
study is not nationally representative, like the IHDP it follows a relatively large cohort of 
children over time, and like the PSID it assessed children’s skills with subscales of the 
Woodcock Johnson. These data yield raw estimates of black-white gap of about one standard 
deviation for both reading and math. Using a nearly identical set of measures as used by Fryer & 
Levit (2004), Murnane and colleagues (2006) found that about 0.5 of a standard deviation of the 
racial kindergarten achievement gaps was explained by SES. 

[Insert Figure 5 here] 

Finally, both Fryer and Levitt (2005) and Yeung and Pfeiffer (2005) took advantage of 
the longitudinal nature of their data to consider whether the same set of SES measures explain 
the somewhat larger achievement gaps found among older children. Despite modest growth in 
the gaps over the early school years, Fryer and Levitt (2005) found that their small set of control 
variables explain about the same amount of the gap (0.5 of a standard deviation) in the spring of 
third grade as they did in the fall of kindergarten. Similarly, Yeung and Pfeiffer (2005) found 
that their SES measures explain roughly the same amount of the preschoolers’ reading gap (ages 
3-5) as the school-age children’s gap (ages 8-10). However, SES appears to explain a slightly 
larger amount of the math gap when children were older (.47 vs. .37). Because the proportion of 
the racial achievement gap explained by SES resources was consistent over time, it is unlikely 
that the increase in gaps is due to family socio-economic differences. 

Although the findings from studies provide divergent sizes of the estimated gaps, an 
interesting consistency is that a handful of SES indicators explain roughly the same absolute 
amount of the gap among young children. With one exception (CNLSY, Phillips et al., 1998), a 
rather small collection of SES-related measures (including parents’ income, educational 
attainment and occupational prestige, plus age) appears to account for a difference of between .4 
to .5 of a standard deviation in white-black reading and math test score gap, regardless of the 


assessments used or the populations studied. In cases where the raw gaps are about half a 
standard deviation, this implies that family background accounts for nearly the entire gap. When 
the raw gaps are considerably larger, this implies that family background can account for only a 
portion of the gap. 

 

Is it really SES? 

Simply documenting that SES accounts for about .4 to .5 of standard deviation of the 
black-white test score gap does not prove that differences in SES have caused differences in 
children’s test scores. For example, although Fryer and Levitt (2004) are able to account for 
virtually all of the racial and ethnic gaps in kindergarten achievement using measures of family 
background, they lack any measure of genetic endowments and are thus unable to discount the 
possibility that what appear to be family socioeconomic effects are really caused by other family 
characteristics. 

Phillips and colleagues (1998) approached the issue of genetic bias in two ways. First, 
they took advantage of the fact that the CNLSY assessed mothers’ academic aptitude using the 
comprehensive Armed Forces Qualifying Test. Although not a direct measure of genetic 
endowment, and not available for fathers, the scores can provide a rough gauge of the potential 
magnitude of biases in studies that lack such measures. Racial differences in mothers’ test scores 
accounted for about one-third of the test score gap, but a 0.75 standard deviation difference still 
remains. Family environment, which includes more than just indicators of SES, was still able to 
account for about half (0.38 of a standard deviation) of this remaining gap. 

Second, Phillips et al. (1998) constructed a “cognitive genotype index” by combining the 
mother’s test score, class rank in high school, and the interviewer’s assessment of the mother’s 
understanding of the interview. They also constructed an SES index based on a host of 
education, income, occupation and other background factors. The simple correlation between 
children’s test scores and their SES index was 0.62. After accounting for the potential confound 
with the genotype index, this correlation was lowered to 0.46 – a 26% reduction. A decrease of 
this magnitude is comparable to what is found in more specialized studies of such mother-child 
correlations between biological and adoptive families (Loehlin, Horn & Willerman, 1989; Scarr 
& Weinberg, 1976). Although far from trivial, this reduction indicates that genetic factors are 
unlikely to account for the bulk of the association between SES and children’s achievement. 
Nevertheless, we remind readers that this approach to understanding the role of genetics is likely 
an oversimplification, given the evidence reviewed above regarding the lower levels of 
heritability and increased importance of environmental influences among low-SES families. 

 

Can Trends In Family Socioeconomic Resources Account for the Convergence of Racial 
Test Scores? 

Another way of investigating the role of SES is to assess whether racial trends in SES 
account for trends in the racial test score gaps. If income and schooling have causal impacts on 
achievement, then we would expect to see that a narrowing of the income and schooling gaps 
between blacks and white would produce corresponding reductions in child achievement gaps. 
So if we consider, for example, the narrowing of the gap between 1970 and 1985 and its 
fluctuations thereafter, then data indicating that blacks and whites became more alike in terms of 


incomes, parental schooling and family structures up to but not after 1985 would be consistent 
with the idea that family SES is a telling determinant of the gap. 

Historically, black (and Hispanic) families faced racism, discrimination, and segregation, 
which limited their educational, economic, and occupational attainment. By the 1960s, the civil 
rights movement had set in motion societal changes that opened opportunity structures and 
improved the SES standing of ethnic and racial minority populations. Since the 1970s, black 
families have experienced larger gains than white families. For example, poverty rate for black 
children has dropped over 9 percentage points since 1980, from 42% to 33%, whereas white 
children’s poverty rates have shown little overall decline. Likewise, since 1979 the share of 
mothers with a high school diploma has increased more among blacks than whites (37% vs. 
13%). 

However, not all trends in indicators of SES over the past 20 years favor black families. 
Since 1979, increases in the percentage of mothers with college degrees were larger for white 
children compared with black children (17% vs. 11%). Though the rates of single-parent families 
differ dramatically across white and black families, the trends in these rates over time have been 
similar. Changes since the 1990s suggest that some trends have favored blacks and others whites. 
Consequently, it is unclear how much of the convergence in young children’s school readiness 
scores should be attributed to changes in SES. 

Analyses that have attempted to account for the sources of convergence have nearly 
always focused on older children, primarily because time-series data with reliable measures of 
achievement family background are not available for young children. Nevertheless, two studies 
suggest that between 15% and 25% of the convergence of scores before the 1990s may be due to 
improvements in children’s parental backgrounds, such as increases in mothers’ educational 
attainment (Cook & Evans, 1998; Grissmer, Flanagan, & Williamson, 1998). 

 

Summary and Policy Discussion 

On average, black students begin school with far worse academic skills than whites, and 
the gap seems to grow during the school years. The very different family incomes, parental 
schooling, and family structures of black and white students are surely the cause of at least some 
of the initial academic gap. Most accounting studies find that SES differences explain a little less 
than half of a standard deviation in the school-entry achievement gaps. Time-series studies find 
that trends in SES can account for about one quarter of the racial trends in high school 
achievement. Because the accounting studies are unable to adjust for a full set of genetic and 
other confounding causes of achievement, and may also lack exact measurement or proper 
functional form, their estimates of the role of family SES are likely to be imprecise. 

To the extent that differences in the socioeconomic backgrounds of young white and 
black children play a causal role in creating achievement gaps, what implications does this have 
for policy? The answer to this question is far from clear, because wise policy decisions require 
an understanding of both causal mechanisms and cost-effective interventions that produce 
desired changes. 

To illustrate, suppose that we believed that boosting maternal schooling by one year 
causes children’s kindergarten achievement scores to increase by one quarter of a standard 
deviation. With the achievement gap between whites and blacks at one half to three quarters of a 


standard deviation, a policy that could increase maternal schooling for all black mothers by an 
average of one or two years, without also changing the schooling of white mothers, would go a 
long way toward eliminating the achievement gaps. 

A first issue is whether it is useful to even frame policy questions in terms of relative 
achievement gaps rather than absolute achievement levels. Interventions can be designed to 
improve black (and/or white) children’s relative skills and absolute levels of academic skills at 
differing points in the skill distribution. However, it is not immediately obvious which is more 
detrimental to blacks and to society in general – lower levels of achievement among black 
children or lower achievement of black as compared with white children. Although both may be 
detrimental, the policies and programs that reduce black-white achievement gaps may differ 
from those that improve absolute levels of achievement among all children (Ceci & Papierno, 
2005). For example, if boosting low levels of achievement is of greater concern, then a program 
that generated higher achievement among all low-achieving children, regardless of race, would 
be successful. In contrast, if reducing racial gaps is a higher priority, then policies and programs 
must improve black children’s skills but hold constant, or reduce, white children’s skills. This 
would likely be achieved only by policies that targeted successful interventions to black children, 
which may be politically infeasible. 

While we recognize that achievement gaps between blacks and whites per se may be 
detrimental, we suggest that low levels of skill achievement are likely to be more harmful for 
blacks, regardless of their standing relative to whites. Low absolute levels of academic skills not 
only limit an individual’s life chances, but also may take a toll on our nation’s prospects for 
economic growth, which depend on a highly skilled workforce (Heckman & Krueger, 2003). 
Consequently, we believe that boosting the absolute level of low-achieving students’ skills 
should be the top policy priority, with perhaps less emphasis placed on addressing achievement 
gaps. 

 

Maternal Schooling 

With respect to maternal schooling, it is critical to identify policies to improve mothers’ 
educational attainment. With large gaps in parental education levels between racial and ethnic 
groups, interventions that increase rates of high school completion may have a large payoff for 
future generations. The historical record of the past 40 years clearly shows large improvements 
in the completed schooling of black parents. But there is little agreement on why schooling has 
increased and, most importantly, how policy might promote future increases in schooling. 

Unfortunately, few of the various academic programs developed to increase high school 
graduation among at-risk adolescents have proved effective. A recent review of 16 random-
assignment evaluations of dropout-prevention programs found only one of them to show clear 
benefits (Dynarski, 2001). On the other hand, rigorous evaluations of a small number of 
intensive teen mentoring programs have produced more promising results (Dubois et al., 2002; 
Roth & Brooks-Gunn, 1998). But the successes of even these intensive programs are not 
guaranteed, particularly when they are implemented on a large scale. 

Studies of low-income populations routinely report that even without any programmatic 
intervention, close to 50 percent of disadvantaged mothers return to school after having children 
(Furstenberg, Brooks-Gunn & Morgan, 1987; Love et al., 2002; McGroder et al., 2000). Yet 


even with high rates of continued schooling, educational attainment among economically 
disadvantaged parents remains much lower than among advantaged families. Thus, another 
intervention approach is to promote educational activities among parents. For example, programs 
targeting teen mothers may provide supports and incentives to stay in school after the birth of a 
child, or welfare programs may make cash benefits contingent on mothers’ participation in 
education and training. Have such programs worked? Evaluations suggest that to date, these 
types of interventions have not been successful in boosting mothers’ educational activity above 
the relatively high level of participation of control group mothers (McGroder et al., 2000; Quint 
et al., 1997). 

The high enrollment of disadvantaged mothers suggests a more promising approach to 
promoting parental education: offsetting the costs of education, particularly higher education, 
and increasing access to adult educational opportunities. Indeed, public spending on higher 
education, including more generous financial aid and the expansion of community colleges, has 
been consistently linked to increases in college attainment and enrollment (Dynarski, 2002; 
Sefter & Turner, 2002; Turner & Bound, 2003; but see also Heckman and Krueger, 2004). Still 
another approach to promoting schooling is to increase the age at which children may leave 
school or begin to work. Changes in these policies over the past century have had a modest 
positive effect on youths' educational attainment (Acemoglu & Angrist, 1999; Oreopoulus, Page 
& Stevens, 2003). 

 

Family Income 

There is a large gap in family income between white children and black children. What 
policies might help to close these gaps? One strategy, embodied in several of the welfare reform 
programs described above, is to promote parents’ attachment to the labor market and reduce low-
income families’ use of welfare benefits. But even the most generous welfare reform programs 
were able to boost average family incomes by only one or two thousand dollars per year. Other 
types of employment interventions, such low-cost job search programs, have produced relatively 
small absolute earnings gains for women—a few hundred dollars over the course of a year or two 
(Heckman, LaLonde & Smith, 1999). More intensive, training-based employment programs have 
produced proportionately bigger earnings gains for women -- a few thousand dollars over several 
years -- but none produced the kinds of long-term income increases that would begin to narrow 
the income or poverty gap between white and ethnic and racial minority families. Results of 
employment interventions for disadvantaged adult men are even less encouraging. Across all 
types of programs, only about one-third of the interventions produced increases in either 
employment or earnings, and no particular type of program emerged as a cure-all (Heckman et 
al., 1999). 

An alternative approach to increasing family incomes and reducing poverty is to 
supplement the incomes of low-income working families through such policies as the Earned 
Income Tax Credit (EITC). A refundable federal tax credit for low-income working families 
with children, the EITC was expanded during the 1990s, and is now the largest cash transfer 
program for low-income families. In 2003, the maximum EITC benefit for a family with two 
children was about $4,200, and nearly 19 million families received the credit (US Committee on 
Ways and Means, 2004). In 1997 the program lifted about 2.2 million children out of poverty 
(Council for Economic Advisors, 1998). Some evidence suggests that the income increases 


occasioned by recent expansions in EITC benefits may have boosted children’s achievement 
(Dahl & Lochner, 2005). 

 Moreover, an encouraging result from the research literature is the apparently selective 
nature of income impacts, with poverty mattering more in the preschool period than later in 
childhood. This suggests the wisdom of child-focused redistributive efforts using, say, tax 
credits, European-style child allowances or future increases in the EITC in which benefits are 
restricted to families with preschool children. The political feasibility of such programs is 
enhanced by the fact that by covering only a fraction of children, they are considerably cheaper 
than programs that cover all children. 

 

Family Structure 

We know very little about the potential of interventions to promote stable two-parent 
families and marriage among disadvantaged populations (Dion et al., 2002). At any rate, the 
current round of federal marriage promotion programs will have to be remarkably successful in 
order to make much of a dent in the huge family structure differential between blacks and whites. 
Furthermore, it appears that for children’s achievement to benefit substantially from marriage, 
there must also be an increase in family resources such as income. Whether higher rates of 
marriage will necessarily improve other aspects of socioeconomic circumstances is unclear 
(Sigle-Rushton & McLanahan, 2003). First, many unmarried couples are already cohabiting, so 
marriage may not improve families’ economic circumstances. Second, it is unclear whether 
marriage will result in longer lasting and more prosperous relationships. We hope that 
evaluations of marriage programs will yield information on the feasibility of increasing marriage 
rates, as well as whether doing so will improve children’s well-being. 

One way to reduce the rate of teen childbearing is to reduce adolescents’ sexual activity 
and increase contraceptive use. A review of experimental and quasi-experimental evaluations of 
teen pregnancy prevention programs suggests that, more often than not, programs designed to 
prevent youths’ sexual behavior fail to reduce the frequency or delay the onset of sexual 
behavior (Kirby, 2001). Of twenty-eight programs focused on abstinence, sexual education, and 
HIV prevention, only 10 delayed the age of sexual initiation. Furthermore, only 19 programs 
measured the frequency of youths’ sexual activity, and of these programs 13 had no meaningful 
effects on the outcomes of interest. In addition, the programs were rather unsuccessful at 
increasing teenagers’ use of contraception. Only four of the eleven sexuality program evaluations 
that measured teenagers’ use of contraception found positive program impacts. However, a 
handful of more intensive interventions that provided mentoring and constructive after-school 
activities have been more successful (Coley & Chase-Lansdale, 1998; Kane & Sawhill, 2003). 
We do not know, however, if the success of these programs can be replicated on a larger scale. 
As is the case for dropout prevention programs, it seems that successful pregnancy prevention 
programs are possible, but they require an intensive effort, and additional research is necessary to 
identify characteristics associated with program success (Coley & Chase-Lansdale, 1998). 

 

Directions for Future Research 

 We close with several recommendations for new directions in research. First, our review 
of racial and ethnic test-score gaps focused primarily on black-white gaps, because only very few 


studies have considered Hispanic-white gaps. The proportion of children of Hispanic origin is 
growing steadily (Hernandez, 2005), and as evident in the ECLS-K data presented in the 
introduction, they too enter school with fewer academic skills than white children. As with other 
racial and ethnic groups, Hispanic populations are quite heterogeneous, and reflect different 
experiences in the United States. Moreover, the average level of Hispanic families’ socio-
economic resources is quite low, raising the possibility that SES differences may explain much 
of the Hispanic-white achievement gaps. 

Because many Hispanic children have parents who are immigrants, an examination of the 
role of SES in these gaps is particularly complex (Fuligni & Yoshikawa, 2003). The assessment 
of Spanish-speaking children’s reading and language achievement may be compromised by their 
limited English language proficiency. In addition, some SES resources may not be comparable in 
immigrant and native-born families. For example, the content and structure of education in other 
countries may not readily correspond to the US educational system (Fuligni & Yoshikawa, 
2003). Perhaps the most important problem is that data on Hispanic children’s achievement are 
relatively scarce. Few of the older large-scale longitudinal studies include a representative 
sample of Hispanic children, because these studies began before the most recent wave of 
Hispanic immigration occurred. The ELCS-K is one exception, and Fryer & Levitt (2004) find 
that a handful of family background measures, including SES, account for most of the Hispanic-
white reading and math gaps at school entry. Because of the need to better understand disparities 
in academic achievement, it is crucial that researchers grapple with these methodological 
challenges and further explore the ways in which family SES influences Hispanic-white 
achievement gaps. 

Second, nearly all of the research on racial gaps has focused specifically on children’s 
academic achievement. Yet, other dimensions of children’s school readiness may also be of 
consequence for school success. Children’s social skills, problem behavior, school engagement, 
and attention may be correlated with their learning and later school achievement (Hinshaw, 
1992; Ladd, Birch, & Buhs, 1999; McClelland, Morrison, & Holmes, 2000). As noted earlier, 
data from the ECLS-K suggest that gaps in teacher ratings of young children’s classroom 
behavior are smaller than gaps in their achievement. However, estimating differences in social 
skills and behavior may be more complicated than assessing children’s relative academic skills, 
as valid and reliable measurement of children’s behavior may be difficult to accomplish in large-
scale survey studies (Konold, Walthall, & Pianta, 2004). A recent review of gaps in school 
readiness found very few studies of black-white gaps in children’s behavior (Rock & Stenner, 
2005). Consequently, we recommend that future studies explore how SES may influence racial 
disparities in behavioral and social dimensions of school readiness. 

Two studies reviewed here indicate that black-white achievement gaps grow during the 
early school years. Fryer and Levitt’s (2005) analysis rules out some obvious explanations for 
the growth in the gap, but does not identify likely causes. For example, they find that less than a 
third of the growth in the gaps is due to differences in school or class quality (see also Murnane 
et al., 2006; Phillips, Crouse, & Ralph, 1998; Todd & Wolpin, 2005). Many hypotheses about 
what may be causing this growth remain to be examined, and we highlight two particularly 
promising possibilities-- teachers’ differential treatment of and expectations for black and white 
students as well as the emergence of stereotype threat (Steele & Aaronson, 1995; Weinstein, 
Gregory, & Strambler, 2004). Ferguson’s (1998) review concluded that teacher’s lower 
expectations for black students are likely to account for a portion of the gap, and it seems 


possible that this influence may become increasingly important during the early school years (see 
also Farkas, 2003). McKown and Wesintein (2003) found that children’s awareness of cultural 
stereotypes increases during middle childhood, and that this elevated awareness was linked to the 
underperformance of minority children on challenging tests. These findings provide some 
evidence that children may experience stereotype threat –anxiety resulting from a fear of 
confirming cultural stereotypes. Further research into these two areas may provide more insight 
into the growth in racial achievement gaps. 

A related topic that deserves more research is whether testing competence and 
performance differs for white children and black children. As noted earlier, it is well known that 
testing conditions influence an individual’s performance on achievement tests. Because most 
assessments of young children’s skills are conducted by one-on-one interview, this has led to 
some speculation that children’s comfort and experience with interviewers of a different racial 
background may influence their performance (Kim, Baydar, & Greek, 2003). Unfortunately, this 
possibility has not been widely explored, either because large studies have not made information 
on interviewers’ race available or because the interviewer was of the same race as the child. One 
exception is the CNLSY, in which 75% of black children were assessed by white interviewers. 
Using these data, Kim and colleagues (2003) found that matching child and interviewer race had 
some positive effects on black’s children tests scores overall, but a significantly larger influence 
on older children (ages 8-9). Whether this is due to an increased level of comfort with the 
interviewer or some other factor is unclear. Yet their findings suggest that additional research is 
needed to better understand how the gap between performance and competence may differ for 
black and white children, and how it may change over time. 

Although resiliency research has identified several factors that may protect children from 
the negative effects of low-SES resources (Kim-Cohen et al., 2004; Luthar & Zigler, 1991; 
McLoyd, 1998), nearly all studies have assumed that the influence of SES is unidirectional and 
have neglected the active role that children play in their own development. Consequently, 
possibly fruitful avenues for future research include examining the development of children’s 
awareness of their own and others’ socioeconomic resources, the individual and ecological 
factors influencing their understanding of economic hardship, and how individual differences in 
perceptions and understanding of socio-economic resources, in turn, affect development. Given 
the economic inequality in our culture and the salience of social comparison processes among 
children, it is remarkable that so little research has been conducted on this topic (Crosby & 
Mistry, 2003). 

Our understanding of how SES affects children may also be furthered if research is 
informed by dynamic systems theory. For example, more careful attention to the nature of 
associations among components of SES is needed. Studies have indicated that changes in family 
income and maternal education have independent effects on children’s achievement, but the 
cumulative and interactive effects of SES components are not well understood. For example, do 
effects of income differ for children with more or less educated parents? Do the effects of 
mothers’ education differ depending on family structure or fathers’ educational attainment? 
Moreover, little attention is given to the correct functional form for the effects of income, 
education, and occupation. Linear effects are often assumed in developmental research, despite 
increasing evidence that the effects of both parents’ education and income are non-linear, with 
increments in parents’ education and income yielding less benefit for more advantaged children. 
Furthermore, household income, family structure, and even parental education are not static over 


the life course, nor are their effects on children. Thus, studies should continue to explore how the 
effects of SES resources differ by their timing, duration, and persistence in children’s lives. 

Finally, we close by underscoring that policy approaches to redressing school 
achievement gaps should focus more on boosting achievement among all low-skilled children 
than on reducing gaps between different groups. Given the intervention record, this probably 
means assigning only a modest role to the current range of programs designed to increase 
parents’ socioeconomic resources, many of which have not proved to be effective. At the same 
time, we need a better understanding of the kinds of policies that have boosted the 
socioeconomic resources in the past as well as evaluations of new program and policy initiatives 
in order to identify more successful avenues of intervention. There are a wide range of 
interventions for children, including policies and programs that directly target children’s 
aptitudes and mental and physical health, as well as some that intervene to change parenting 
behavior. Going beyond SES interventions, child and parenting interventions may prove to be 
the most cost-effective ways of addressing disparities in achievement. 


References 

 

Acemoglu, D., & Angrist, J. (1999). How large are the social returns to compulsory schooling 
effects? Evidence from compulsory schooling laws. (Working Paper No. 444): Cambridge, MA: 
NBER working paper. 

Ackerman, B.P., Brown, E. D., & Izard, C. (2004). The relations between contextual risk, earned 
income, and the school adjustment of children from economically disadvantaged families. 
Developmental Psychology, 40, 204-216. 

Ackerman, B.P., Brown, E.D., Schoff D’Eramo, K., & Izard, C. E. (2002). Maternal relationship 
instability and the school behavior of children from disadvantaged families. Developmental 
Psychology, 38, 694-704. 

Amato, P. R. (1993). Children’s adjustment to divorce: Theories, hypotheses and empirical support. 
Journal of Marriage and the Family, 55, 23-38. 

Amato, P.R., & Keith, B. (1991). Parental divorce and the well-being of children: A meta-analysis. 
Psychological Bulletin, 110, 26-46. 

Astone, N. M., et al (2000). School reentry in early adulthood: The case of inner-city African 
Americans. Sociology of Education, 73, 133-154. 

Aughinbaugh, A., Pierret, C. R., & Rothstein, D. S. (2005). The impact of family structure 
transitions on youth achievement: Evidence from the children in the NLSY79. Demography, 42, 
447-468. 

Becker, G. (1981). A Treatise on the Family. Cambridge, MA: Harvard University Press. 

Black, S. E., Devereux, P. J., & Salvanes, K.G. (2004). Why the Apple Doesn’t Fall Far from the 
Tree. (Working Paper No. 10066): Cambridge, MA: NBER working paper. 

Black, M., Dubowitz, H, & Starr, R. H. (1999). African American fathers in low income, urban 
families: Development, behavior, and home environment of three-year-old children. Child 
Development, 70, 967-978. 

Blau, D. M. (1999). The effect of income on child development. The Review of Economics and 
Statistics, 8, 261-276. 

Bollen, K. A., & Glanville, J.L., & Stecklov, G. (2001). Socioeconomic status and class in studies of 
fertility and health in developing countries. Annual Review of Sociology, 27, 153-185. 

Bornstein, M.H., Hahn, C.S., Suwalsky, J. T. D., & Haynes, O. M. (2003). Socioeconomic status, 
parenting and child development: The Hollingshead Four Factor Index of Social Status and the 
Socioeconomic Index of Occupations. In M. H. Bornstein, & R.H.Bradley (Eds.), Socioeconomic 
Status, Parenting, and Child Development (pp. 29-82). Mahwah, NJ: Lawrence Erlbaum. 

Bradley, R. H. & Corwyn, R. F. (2002). Socioeconomic status and child development. Annual 
Review of Sociology, 53, 371-99. 

Bronfenbrenner, U. (1986). Ecology of the family as a context for human development: Research 
perspectives. Developmental Psychology, 22, 723-742. 


Bronfenbrenner, U. & Morris, P. A. (1998). The ecology of developmental process. In R. M. Lerner 
(Ed.), Handbook of Child Psychology, Vol. 1 (5th edition). (pp. 993-1028). New York: John 
Wiley & Sons, Inc. 

Brooks-Gunn, J., Duncan, G. J., & Aber, J. L. (1997). Neighborhood Poverty: Context and 
Consequences for Children (Vol. I & II). New York: Russell Sage. 

Brooks-Gunn, J., et al., (2003). The Black-White test score gap in young children: Contributions of 
test and family characteristics. Applied Developmental Science, 7, 239-252. 

Brown, A. C., Brody, G. H., & Stoneman, Z. (2000). Rural black women and depression: A 
contextual analysis. Journal of Marriage and the Family, 62, 187-198. 

Brown, S. L. (2004). Family structure and child well-being: The significance of parental 
cohabitation. Journal of Marriage and Family, 66, 351-367. 

Burchinal, M.R., & Cryer, D. (2003). Diversity, child care quality, and developmental outcomes. 
Early Childhood Research Quarterly, 18, 401-426. 

Card, D. (1999). The causal effect of education on earnings. In O. Ashenfelter, & D.E. Card (Eds.), 
Handbook of Labor Economics: Volume 3A (pp.1801-1863). New York: Elsevier. 

Carlson, M.J., & Furstenberg, F. (2004). Complex families: Documenting the prevalence and 
correlates of multi-partnered fertility in the United States." Fragile Families working paper 
#2003-14-FF. Princeton, New Jersey: Bendheim-Thoman Center for Research on Child 
Wellbeing, Princeton University. 

Carlson, M.J., & Corcoran, M.E. (2001). Family structure and children’s behavioral and cognitive 
outcomes. Journal of Marriage and the Family, 63, 779-792. 

Ceci, S. & Papierno, P. B. (2005). The rhetoric and reality of gap closing: When the “have-nots” 
gain but the “haves” gain even more. American Psychologist, 60, 149-160. 

Citro, C. & Michael, R. (eds.) (1995). Measuring poverty: A new approach. Washington, D.C.: 
National Academy Press. 

Clarke-Stewart, K. A. et al. (2000). Effects of parental separation and divorce on very young 
children. Journal of Family Psychology, 14, 304-326. 

Chase-Lansdale, P. L., Cherlin, A. J., & Kiernan, K.E. (1995).The long-term effects of parental 
divorce on the mental health of young adults: A developmental perspective. Child Development, 
66, 614-634. 

Cole, N. & Currie, J. (1994). Reported income in the NLSY: Consistency checks and methods for 
cleaning the data. Cambridge, MA: NBER working paper #162. 

Coley, R.L. (1998). Children’s socialization experiences and functioning in single mother 
households: The importance of fathers and other men. Child Development, 69, 219-230. 

Coley, R.L., & Chase-Lansdale , P. L. (1998). Adolescent pregnancy and parenthood: Recent 
evidence and future directions. American Psychologist, 53, 152-166. 

Collins, W. A., Maccoby, E. E., Steinberg, L., Hetherington, E. M., Bornstein, M. H. (2000). 
Contemporary research on parenting. The case for nature and nurture. American Psychologist, 
55(2), 218-32. 


Conger, R. D., Ebert Wallace, L., Sun, y., Simmons, R. L., McLoyd, V. C., & Brody, G. H. (2002). 
Economic pressure in African American families: A replication and extension of the family 
stress model. Developmental Psychology, 38, 179-193. 

Cook, M. D., & Evans, W. N. (1998). Families or schools? Explaining the convergence in White and 
Black academic performance. Journal of Labor Economics, 18, 729-754. 

Cooksey, E. (1997). Consequences of young mothers’ martial histories for children’s cognitive 
development. Journal of Marriage and the Family, 59, 245-261. 

Council for Economic Advisors. (1998). Good news for low income families: Expansions in the 
Earned Income Tax Credit and the Minimum Wage. Washington, D.C: Author. 

Crosby, D. A. & Mistry, R. S. (2003). Children’s Perceptions of Wealth and Poverty: Do Child, 
Family, and School Characteristics Relate to Views on Economic Inequality? Paper presented at 
the biennial meeting of the Society for Research in Child Development, Tampa, FL. 

Dahl, G. & Lochner, L. (2005). The effects of family income on child achievement. University of 
Wisconsin, Institute for Research on Poverty Working Paper, 1305-05. 

Davis-Kean, P. (2005).The influence of parent education and family income on child achievement: 
The indirect role of parent expectations and the home environment. Journal of Family 
Psychology, 19, 294-304 

Davis-Kean, P., & Magnuson, K. (2004). The Influence of Parental Education on Child 
Development. Mimeo, University of Michigan. 

Dearing, E., McCartney, K. & Taylor, B. A. (2001). Change in family income-to-needs matters more 
for children with less. Child Development, 72, 1779-1793. 

DeLeire, T., & Kalil, A. (2002). Good things come in 3’s: Multigenerational coresidence and 
adolescent adjustment. Demography, 39, 393-413. 

Dion, M. R., et al. (2002). Helping unwed parents build strong and healthy marriages: A conceptual 
framework for interventions. Princeton NJ: Mathematica Policy Research. 

Dubois, D. L., Holloway, B.E., Valentine, J. C., & Cooper, H. (2002). Effectiveness of mentoring 
programs for youth: A meta-analytic review. Journal of Community Psychology, 30, 157-197. 

Duncan, G., & Brooks-Gunn, J. (1997). Consequences of Growing Up Poor. New York: Russell 
Sage. 

Duncan, G. et al., (1998). How much does childhood poverty affect the life chances of children? 
American Sociological Review, 63, 406-423. 

Duncan, G., & Magnuson, K.A. (2003). Off with Hollingshead: Socioeconomic resources, parenting 
and child development. In M. H. Bornstein, & R.H.Bradley (Eds.), Socioeconomic Status, 
Parenting, and Child Development (pp. 83-106). Mahwah, NJ: Lawrence Erlbaum. 

Duyme, M., Dumaret, A. C., & Tomkiewicz, S. (1999). How can we boost IQs of “Dull Children”? 
A late adoption study. Proceedings of the National Academy of Sciences, 96, 8790-8794. 

Dynarski, M. (2001). Making do with less: Interpreting the evidence from recent federal evaluations 
of dropout prevention programs. Paper presented at the Dropouts: Implications and Findings 
Conference, Harvard University, Cambridge, MA. 


Dynarksi, S. (2002). The consequences of lowering the cost of college. American Economic Review, 
92, 279-85. 

Elder, G. (1974). Children of the Great Depression: Social change in life experience. Chicago, IL: 
University of Chicago Press. 

Ellen, I. G., & Turner, M. A. (1997). Does neighborhood matter? Assessing recent evidence. 
Housing Policy Debate, 8, 833-866. 

Entwisle, D. R., Alexander, K. L., Olson, L. S. (2005). First grade and educational attainment by age 
22. American Journal of Sociology, 110, 1458-1502. 

Evans, G. (2004). The environment of childhood poverty. American Psychologist, 59, 77-92. 

Farkas, G. (2003). Racial Disparities and Discrimination in Education: What Do We Know, How Do 
We Know It, and What Do We Need to Know? Teachers’ College Record, 105, 1119-1146. 

Farkas, G. & Beron, K. (forthcoming). The detailed age trajectory of oral vocabulary knowledge: 
Differences by class and race. Social Science Research. 

Ferguson, R. F. (1998). Teachers’ perceptions and expectations and the black-white test score gap. 
In C. Jencks, & M. Phillips (Eds.), The Black-White test score gap (pp. 273-317). Washington, 
DC: Brookings. 

Fergusson, D. M., & Woodward, L. J. (1999). Maternal age and psychosocial outcomes in early 
adulthood. Journal of Child Psychology and Psychiatry, 43, 479-489. 

Fryer, R. & Levitt, S. D. (2004).Understanding the Black-White test score gap in the first two years 
of school. Review of Economics and Statistics, 86, 447-464. 

Fryer, R. & Levitt, S. D. (2005). The Black-White test score gap through third grade. National 
Bureau of Economic Research. Working paper, January. 

Fuligni, A. & Yoshikawa, H. (2003). Socioeconomic resources, parenting, poverty and child 
development among immigrant families. . In M. H. Bornstein, & R.H.Bradley (Eds.), 
Socioeconomic Status, Parenting, and Child Development (pp. 107-124). Mahwah, NJ: 
Lawrence Erlbaum. 

Furstenberg, F. F., Brooks-Gunn, J., Morgan, S. P. (1987). Adolescent mothers in later life. New 
York: Cambridge University Press. 

Gallo, L. C., Bogart, L. M., Vranceanu, A., & Matthews, K. A. (2005). Socioeconomic status, 
resources, psychological experiences, and emotional responses: A test of the reserve capacity 
model. Journal of Personality and Social Psychology, 88, 386-399. 

Garcia Coll, C. T., et al. (1996). An integrative model for the study of developmental competencies 
in minority children. Child Development, 67, 1891-1914. 

Garmezy, N. (1991). Resilience in children’s adaptation to negative life events and stressed 
environments. Pediatrics, 20, 459-466. 

Gittleman, M., & Wolff, E. N. (2004). Racial differences in patterns of wealth accumulation. Journal 
of Human Resources, 39, 193-227. 


Grissmer, D., Flanagan, A., & Williamson, S. (1998). Why Did the Black-White score gap narrow in 
the 1970s and 1980s? In C. Jencks, & M. Phillips (Eds.), The Black-White test score gap (pp. 
182-226). Washington, DC: Brookings. 

Guo, G., & Mullan Harris, K. (2000). The mechanisms mediating the effects of poverty on 
children’s intellectual development. Demography, 37, 431-447. 

Guo, G., & Stearns, E. (2002). The social influences on the realization of genetic potential for 
intellectual development. Social Forces, 80, 881-910. 

Hauser, R. M. (1996). Trends in Black-White test score differentials (Working paper No. 96-29: 
Center for Demography and Ecology). Madison, WI: University of Wisconsin-Madison. 

Hauser, R. & Warren, J. R. (1997). Socioeconomic indexes for occupations: A review, update, and 
critique. Sociological Methodology, 27, 177-298. 

Haveman, R., & Wolfe, B. (1994). Succeeding generations: On the effect of investments in children. 
New York: Russell Sage Foundation. 

Haveman, R. & Wolfe, B. (1995).The determinants of children’s attainments: A review of methods 
and findings. Journal of Economic Literature, 23, 1829-1878. 

Heckman, J., LaLonde, R., & Smith, J. (1999). The economics and econometrics of active labor 
market programs. In Ashenfelter, & Card (Eds.), Handbook of Labor Economics (pp. 1865-
2097). NY: Elsevier. 

Heckman, J., & Krueger, A. (2003). Inequality in America – What role for human capital policies? 
Cambridge, MA: MIT Press. 

Hedges, L., & Nowell, A. (1998). Black-white test score convergence since 1965. In C. Jencks, & 
M. Phillips (Eds.), The Black-White test score gap (pp. 149-181). Washington, DC: Brookings. 

Hedges, L., & Nowell, A. (1999). Changes in the black-white gap in achievement test scores. 
Sociology of Education, 72, 111-135. 

Helms, J.E. (1992). Why is there no study of cultural equivalence in standardized cognitive ability 
testing? American Psychologist, 47, 1083-1101. 

Hernstein, R. J., & Murray, C. (1994). The bell curve: Intelligence and class structure in American 
life. New York: Free Press. 

Hinshaw, S. P. (1992). Externalizing behavior problems and academic underachievement in 
childhood and adolescence: Causal relationships and underlying mechanisms. Psychological 
Bulletin, 111, 127-155. 

Hernandez, D. (2005). Demographic change and the life circumstances of immigrant families. 
Future of Children, 14, 17-47. 

Hetherington, M.E., & Stanley-Hogan, M. (1999). The adjustment of children with divorced parents: 
A risk and resiliency perspective. Journal of Child Psychology & Psychiatry, 40, 129-140. 

Hoff, E. (2003). The specificity of environmental influence: Socioeconomic status affects early 
vocabulary development via maternal speech. Child Development, 74, 1368-1378. 

Huston, A., C., McLoyd, V. C., & Garcia Coll, C. (1997). Poverty and behavior: The case for 
multiple methods and levels of analysis. Developmental Review, 17, 376-393. 


Jacobs, J. & Stoner-Eby, S. (1998). Adult enrollment in high education and cumulative educational 
attainment, 1970-1990. Annals of the American Academy of Political and Social Science, 559, 
91-108. 

Jaffee, S., Caspi, A., Moffitt, T.E., Belsky, J., & Silva, P. (2001). Why are children born to teen 
mothers at risk for adverse outcomes in young adulthood? Results from a 20-year longitudinal 
study. Development and Psychopathology, 13, 377-397. 

Jarrett, R. (1997). African American family and parenting strategies in impoverished neighborhoods. 
Qualitative Sociology, 20, 275-288. 

Jencks, C. (1998). Racial bias in testing. In C. Jencks, & M. Phillips (Eds.), The Black-White test 
score gap (pp. 55-102). Washington, DC: Brookings. 

Jencks, C., & Mayer, S. (1990). The social consequences of growing up in a poor neighborhood. In 
Lynn, & McGeary (Eds.), Inner-City Poverty in the United States (pp. 111-186). Washington, 
D.C.: National Academy Press. 

Jensen, A. J. (1969). How much can we boost IQ and scholastic achievement? Harvard Educational 
Review, 39, 1-123. 

Kao, G. & Thompson, J.S. (2003). Racial and Ethnic Stratification in Educational Achievement and 
Attainment. Annual Review of Sociology, 29, 417-42. 

Kane, A., & Sawhill, I.V. (2003). Preventing teen childbearing. In I.V. Sawhill (Ed.), One percent 
for the Kids (pp. 56-75). Washington, DC: Brookings Institution Press. 

Kim, H., Baydar, N., & Greek, A. (2003). Testing conditions influence the race gap in cognition and 
achievement estimated by household survey data. Applied Developmental Psychology, 23, 567-
582.

Kim-Cohen, J., Moffit, T. E., Caspi, A., & Taylor, A. (2004). Genetic and environmental processes 
in young children’s resilience and vulnerability to socioeconomic deprivation. Child 
Development, 75, 651-668. 

Kirby, D. (2001).Emerging answers: Research findings on programs to reduce teen pregnancy. 
Washington, DC: National Campaign to Prevent Teen Pregnancy. 

Kling, J., et al. (2004). Moving to Opportunity and tranquility: Neighborhood effects on adult 
economic self-sufficiency and health from a randomized housing voucher Experiment,” Mimeo, 
Industrial Relations Section Working Paper, Princeton University. 

Kohn, M. L. (1963). Social class and the parent-child relationship: An interpretation. American 
Journal of Sociology, 68, 471-480. 

Konold, T. R., Walthall, J. C. & Pianta, R. (2004). The behavior of child behavior ratings: 
Measurement structure of the child behavior checklist across time, informants, and child gender. 
Behavioral Disorders, 29, 372-383.

Krueger, A., & Whitmore, D. (2001). The effect of attending a small class in the early grades on 
college-test taking and middle school test results: Evidence from project STAR. Economic 
Journal, 11, 1-28. 

Ladd, G. W., Birch, S. H., Buhs, E. S. (1999). Children’s social and scholastic lives in Kindergarten: 
Relates spheres of influence? Child Development, 70, 1373-1400. 


Laosa, L. (1980). Maternal teaching strategies in Chicano and Anglo-American families: The 
influence of culture and education on maternal behavior. Child Development, 5, 759-765. 

Laosa, L. (1983). School, occupation, culture and family. In E. Sigel and L. Laosa (Eds.), Changing 
Families (pp.79-135). NYC, NY: Plenum Press. 

Lareau, A. (2003) Unequal childhoods. Los Angeles, CA: University of California Press. 

Lee, V., & Burkam, D. (2002). Inequality at the starting gate: Social background differences in 
achievement as children begin school. Washington, DC: Economic Policy Institute.

Lerner, R. (2003). What are SES effects of? A developmental systems perspective. In M. H. 
Bornstein, & R.H.Bradley (Eds.), Socioeconomic Status, Parenting, and Child Development 
(pp.231-255). Mahwah, NJ: Lawrence Erlbaum. 

Leventhal, T., & Brooks-Gunn, J. (2000).The neighborhoods they live in: The effects of 
neighborhood residence on child and adolescent outcomes. Psychological Bulletin, 126, 309-337. 

Levine, J.A., Pollack, H., & Comfort, M. (2001). Academic and behavioral outcomes among 
children of young mothers. Journal of Marriage and the Family, 63, 355-369. 

Loehlin, J.C., Horn, J. M., & Willerman, L. (1989). Modeling IQ change: Evidence from the Texas 
Adoption Project. Child Development, 60, 993-1004. 

Love, J. M., et al. (2002). Making a difference in the lives of infants and toddlers and their families: 
The impacts of early Head Start. Volume 1: Final Technical Report. Princeton, NJ: Mathematica 
Policy Research. 

Luster, T., Rhoades, K., & Haas, B. (1989). The relation between parental values and parenting 
behavior: A test of the Kohn Hypothesis. Journal of Marriage and the Family, 51, 139-147. 

Luthar, S. & Zigler, E. (1991). Vulnerability and competence: A review of research on resilience in 
childhood. American Journal of Orthopsychiatry, 61, 6-22. 

Magnuson, K. (2004). The effect of increases in maternal education on children’s academic 
trajectories: Evidence from the NLSY. Mimeo: Columbia University. 

Magnuson, K. (2003).The effect of increases in welfare mothers’ education on their young children’s 
academic and behavioral outcomes: Evidence from the national evaluation of welfare-to-work 
strategies child outcomes study. University of Wisconsin, Institute for Research on Poverty 
Discussion Paper, 1274-03. 

Magnuson, K.A., & Duncan, G. (2002). Parents in poverty. In Bornstein (Ed.), Handbook of 
Parenting (pp. 95-121). Mahwah, NJ: Lawrence Erlbaum. 

Masten, A. (2001). Ordinary magic: Resilience processes in development. American Psychologist, 
56, 227-238. 

Mayer, S. (1997). What money can’t buy: The effect of parental income on children’s outcomes. 
Cambridge, MA: Harvard University Press. 

McCall, R. & Carriger, M. (1993). A meta-analysis of infant habituation and recognition memory 
performance as predictors of later IQ. Child Development, 64, 57-69. 


McClelland, M., Morrison, F. J., Holmes, D. L. (2000). Children at risk for early academic 
problems: The role of learning-related social skills. Early Childhood Research Quarterly, 15, 
307-329. 

McGroder, S. M., et al. (2000). National Evaluation of Welfare-to-work Strategies Impacts on Young 
Children and their Families Two Years After Enrollment: Findings from the Child Outcomes 
Study. Washington, DC: United States Department of Health and Human Services Office of the 
Assistant Secretary for Planning and Evaluation Administration for Children and Families. 

McLanahan, S. (1997). Parent absence or poverty: Which matters more? In G. Duncan, & J. Brooks-
Gunn (Eds.), Consequences of growing up poor (pp. 35-48). New York: Russell Sage 
Foundation. 

McLanahan, S., & Sandefur, G. (1994). Growing up with single parent: What hurts, what helps. 
Cambridge, MA: Harvard University Press. 

McLoyd, V. (1998). Socioeconomic disadvantage and child development. American Psychologist, 
53, 185-204. 

McKown, C. & Weinstein, R. (2003). The development and consequences of stereotype 
consciousness in middle childhood. Child Development, 74, 498-515. 

Menaghan, E. G. & Parcel, T. L. (1991). Social sources of change in children’s home environments: 
The effects of parental occupational experiences and family conditions. Journal of Marriage and 
the Family, 57, 69-94. 

Mercy, J. A., & Steelman, L. (1982). Familial influence on intellectual attainment of children. 
American Sociological Review, 42, 532-542. 

Meyer, D. R., Cancian, M, & Cook, S. (2005). Multiple partner fertility: Incidence and implications 
for child support policy. Social Service Review, 79, 577-601. 

Meyer, B., & Rosenbaum, D. T. (2001). Welfare, the Earned Income Tax Credit and the labor supply 
of single mothers. Quarterly Journal of Economics, 116, 1063-1114. 

Michael, R. T. (1972).The effect of education on efficiency in consumption. New York: Columbia 
University Press. 

Miller, C. L., Miceli, P. J., Whitman, T. L., & Borkowski, J. G. (1996). Cognitive readiness to parent 
and intellectual-emotional development in children of adolescent mothers. Developmental 
Psychology, 32, 533-541. 

Miller, J.E., & Davis, D. (1997). Poverty history, marital history, and quality of children’s home 
environments. Journal of Marriage and the Family, 59, 996-1007. 

Mistry, R., Biensanz, J. & Taylor, L. (2004). Family Income and Its Relation to Preschool Children's 
Adjustment for Families in the NICHD Study of Early Child Care. Developmental Psychology, 
40, 727-45. 

Morris, P. A., Duncan, G. and Clark-Kauffman, E. (forthcoming). Child well-being in an era of 
welfare reform: The sensitivity of transition in development to policy change. Developmental 
Psychology. 

Mueller, C., & Parcel, T. L. (1981). Measures of socioeconomic status: Alternative and 
recommendations. Child Development, 52, 13-30. 


Murnane, R. J., Willet, J. B., Bub, K. L., & McCartney, K. (2006). Understanding trends in 
Racial/Ethnic achievement gaps during elementary school. Paper presented at the American 
Economic Association Meetings, Boston, MA. 

National Center for Education Statistics (2001). User’s Manual for the ECLS-K Base Year Public-
Use User’s Manual. Washington, DC: U.S. Department of Education, Office of Educational 
Research and Improvement. 

National Center for Education Statistics. (2003, October). Reading. Retrieved February 14, 2005, 
from : http://nces.ed.gov/nationsreportcard/reading/results2003/natsubgroups.asp 

Neiss, M., & Rowe, D. C. (2000). Parental education and child’s verbal IQ in adoptive and 
biological families in the National Longitudinal Study of Adolescent Health. Behavior Genetics, 
30, 487-495. 

Nisbett, R. E. (1998). Race, genetics and IQ. In C. Jencks & M. Phillips (Eds.), The Black-White test 
score gap (pp. 86-102). Washington, DC: Brookings. 

Neitzel, C. & Stright, A. D. (2004). Parenting behaviours during child problem solving: The role of 
child temperament, mother education and personality, and the problem-solving context. 
International Journal of Behavioral Development, 28, 166-179. 

O’Connor, T. G., Plomin, R., Caspi, A., & DeFries, J.C. (2000). Are associations between parental 
divorce and children’s adjustment genetically mediated? An adoption study. Developmental 
Psychology, 36, 429-437. 

Oreopoulus, P., Page, M., & Stevens, A.H. (2003). Does human capital transfer from parent to 
child? The intergenerational effects of compulsory schooling. (Working Paper No.10164): 
NBER. 

Orr, A. J. (2003). Black-white differences in achievement: The importance of wealth. Sociology of 
Education, 76, 281-304. 

Orr, L., et al. (2003). Moving to Opportunity Interim Impacts Evaluation, Washington, D.C., U.S. 
Department of Housing and Urban Development, Office of Policy Development & Research, , 1-
178. 

Phillips, M. & Chin, T. (2003). School inequality: What do we know? University of California, Los 
Angeles, Mimeo. 

Phillips, M., Crouse, J., & Ralph, J. (1998). Does the Black-White test score gap widen after 
children enter school? In C. Jencks, & M. Phillips (Eds.), The Black-White test score gap (pp. 
229-267). Washington, DC: Brookings. 

Phillips, M., et al. (1998). Family background, parenting practices, and the Black-White test score 
gap. In C. Jencks, & M. Phillips (Eds.), The Black-White test score gap (pp. 103-144). 
Washington, DC: Brookings. 

Plomin, R. (1994). Genetic and experience: The interplay between nature and nurture. Thousand 
Oaks, CA: Sage. 

Plug, E. (2003). Estimating the effects of mothers’ schooling on children’s schooling using a sample 
of adoptees. American Economic Review, 94, 358-368. 


Popkin, S.J., Harris, L. E., & Cunningham, M. K. (2001). Families in transition: A qualitative 
analysis of the MTO experience. Washington, DC: Urban Institute. 

Quint, J. C., Bos, J. M., & Polit, D. F. (1997). New chance: Final report on a comprehensive 
program for young mothers in poverty and their children. New York: Manpower Research 
Demonstration Corporation. 

Raviv, T., Kessenich, M., & Morrison, F. J. (2004). A mediational model of the association between 
socioeconomic status and three-year-old language abilities: The role of parenting factors. Early 
Childhood Research Quarterly, 19, 528-547. 

Rich, L. M., & Kim, S. (1999). Patterns of later life education among teenage mothers. Gender and 
Society, 13, 798-817. 

Richman, A. L., Miller, P. M., & LeVine, R. A. (1992). Cultural and educational responsiveness in 
maternal responsiveness. Developmental Psychology, 4, 614-621. 

Rock, D. A., & Stenner, A. J. (2005). Assessment issues in the testing of children at school entry. 
Future of Children, 15, 15-34. 

Roe, K., & Bronstein, R. (1988). Maternal education and cognitive processing at three months as 
shown by infants’ vocal response to mother vs. stranger. International Journal of Behavioral 
Development, 1, 389-395. 

Rosenzweig, M. R., & Wolpin, K. I. (1994). Are there increasing returns to intergenerational 
production of human capital? Maternal schooling and child intellectual development. Journal of 
Human Resources, 29, 670-693. 

Roth, J., & Brooks-Gunn, J. (1998). Promoting healthy adolescents: Synthesis of youth development 
program evaluations. Journal of Research on Adolescence, 8, 423-459. 

Rowe, D. C. (1994) The Limits of family influence: Genes, experience and behavior. New York: 
Guilford. 

Rowe, D. C. & Rodgers, J. L. (1997). Poverty and behavior: Are environmental measures nature and 
nurture? Developmental Review, 17, 358-375. 

Rowe, D. C., Jacobson, K. C., & Van den Oord, E. (1999). Genetic and environmental influences on 
vocabulary IQ: Parental education level as moderator. Child Development, 70, 1151-1162. 

Sampson, R.J., Morenoff, J.D., & Ganon-Rowley, T. (2002). Assessing neighborhood effects: Social 
processes and new directions in research. Annual Review of Sociology, 94, 774-780. 

Sanbonmatsu, L., Kling, J., Duncan, G. & Brooks-Gunn, J. (forthcoming). Neighborhoods and 
academic achievement: Results from the Moving to Opportunity experiment. Journal of Human 
Resources. 

Scarr, S. (1992). Developmental theories for the 1990s: Development and individual differences. 
Child Development, 63, 1-19. 

Scarr, S. & Weinberg, R. (1976). IQ performance of black children adopted by white families. 
American Psychologist, 31, 726-739. 

Scarr, S., & Weinberg, R., Waldman (1993). IQ correlations in Transracial adoptive families, 
Intelligence, 17, 541-555.. 


Sefter, N., & Turner, S. (2002). Back to school: Federal student aid policy and adult college 
enrollment. Journal of Human Resources, 37, 336-52. 

Sigle-Rushton, W., & McLanahan, S. (2003). For richer or poorer: Marriage as an anti-poverty 
strategy in the United States,” Mimeo, Princeton University, Center for Research on Child 
Wellbeing Working Paper #01-17-FF. 

Smith, J., Brooks-Gunn, J., & Klebanov, P. (1997). The consequences of living in poverty on young 
children's cognitive development. In G. Duncan, & J. Brooks-Gunn (Eds.), Consequences of 
Growing Up Poor (pp. 132-189). New York: Russell Sage. 

Sommer, K., et al (1993). Cognitive readiness and adolescent parenting. Developmental Psychology, 
32, 12-25.

Steele, C. & Aronson, J. (1995). Stereotype threat and the intellectual test performance of African 
Americans. Journal of Personality and Social Psychology, 69, 797-811. 

Suzuki, L.A. & Valencia, R.R. (1997). Race-ethnicity and measured intelligence. American 
Psychologist, 52, 1103-1114. 

Todd, P. and Wolpin, K. I. (2005). The production of cognitive achievement in children: Home, 
school, and racial test score gaps. Unpublished manuscript, University of Pennsylvania. 

Turkheimer, E., Haley, A., Waldron, M. D’Onofrio, and Gottesman, I. (2003). Socioeconomic status 
modifies heritability of IQ in young children. Psychological Science, 14, 623-628. 

Turley, R.N.L. (2003). Are children of young mothers disadvantaged because of their mother’s age 
or family background. Child Development, 74, 465-474. 

Turner, S., & Bound, J. (2003). Closing the gap or widening the divide: The effects of the G.I. Bill 
and World War II on the educational outcomes of Black Americans. Journal of Economic 
History, 63,145-177. 

US Census Bureau (2004). People in Families with Related Children Under 18 by Family Structure, 
Age, and Sex, Iterated by Income-to-Poverty Ratio and Race: 2002. Accessed from 
http://ferret.bls.census.gov/macro/032003/pov/new03_100_01.htm accessed April 26, 2004 

U.S. Department of Education, Office of Educational Research and Improvement. (1993, 
September). Public school kindergarten teachers' views on children's readiness for school: 
Contractor Report to the National Center for Education Statistics. (Fast Response Survey 
System NCES 93-410). Washington, DC: Author. 

US House Committee on Ways and Means. (2004). Green Book. Retrieved April 26, 2004, from 
http://waysandmeans.house.gov/Documents.asp?section=813.

Ventura, S. J., Mathews, T. J., & Hamilton, B.E. (2002). Teenage births in the United States: State 
trends, 1991-2000, an update. National Vital Statistics Reports, 50, 1-2. 

Walker, S. O., Petrill, S. A., & Plomin, R. (2005). A genetically sensitive investigation of the effects 
of the school environment and socio-economic status on academic achievement in seven-year-
olds. Educational Psychologist, 25, 55-73. 

Weinstein, S., Gregory, A., & Strambler, M. J. (2004). Intractable self-fulfilling prophecies: Fifty 
years after Brown v. Board of Education. American Psychologist, 59, 511-520. 


Werner, E.E., & Smith, R. S. (1982). Vulnerable but invincible: A study of resilient children. New 
York: McGraw Hill. 

Yeung, W. J. & Pfeiffer, K. M. (2005). The black-white test score gap: Lessons from the Panel 
Study of Income Dynamics. New York University, Unpublished Manuscript.


 Figure 1: Test Score and Socioeconomic Differences For 
Kindergartners From Different Race/Ethnic Groups-0.638-0.401-0.561-0.722-0.427-0.6250.150.3350.112-1.5-1.4-1.3-1.2-1.1-1-0.9-0.8-0.7-0.6-0.5-0.4-0.3-0.2-0.100.10.20.30.40.5123Standard Deviation Difference from WhitesBlacksHispanicAsianMath ScoresReading ScoresSocioeconomic 
Status


 Figure 2: Black-White Test Score Gaps in Academic Skills 
PSID VerbalECLS-K ReadingECLS-K MathPSID Math-1.5-1.4-1.3-1.2-1.1-1-0.9-0.8-0.7-0.6-0.5-0.4-0.3-0.2-0.100.10.2Standard Deviation Difference from WhitesAge 8-10Age 3-5KindergartenKindergartenThird GradeThird GradeAge 3-5Age 8-10 



Figure 3: Black-White Gap in Fourth Grade NAEP Reading 
Scores-0.74-0.9-1.04-0.92-0.84-0.79-0.71-0.79-0.83-0.8-1.5-1.4-1.3-1.2-1.1-1-0.9-0.8-0.7-0.6-0.5-0.4-0.3-0.2-0.100.10.219651970197519801985199019952000YearStandard Deviation Differerence from 
Whites 
Note: The figure is based on authors’ calculation of NAEP long-term trend data available from NCES. The 
estimates prior to 1999 are comparable to those presented by in Phillips et al. (1998). 


 Figure 4: Accounting for the Black-White Test Score Gap with 
SES, Data from the ECLS-K and PSID-1.5-1.4-1.3-1.2-1.1-1-0.9-0.8-0.7-0.6-0.5-0.4-0.3-0.2-0.100.10.2Standard deviation difference from whitesRaw gapAdj. for SES 
ECLS-K 
MathECLS-K 
ReadingPSIDReadingPSID Math 



Figure 5: Accounting for Black-White Test Scores Gaps with 
SES, Data from the CNLSY and IHDP-1.5-1.4-1.3-1.2-1.1-1-0.9-0.8-0.7-0.6-0.5-0.4-0.3-0.2-0.10Standard deviation difference from whitesRaw gapAdj. for SES 
IHDP-WPPSI 
Verbal (Age 5)
IHDP-WPPSI 
Performance 
(Age 5)
IHDP PPVT-R 
(Age 5)
NLSY PPVT-R 
(Age 5/6)