ArticlePDF Available

Low Birth Weight, Social Factors, and Developmental Outcomes Among Children in the United States


Abstract and Figures

We used six waves of the National Longitudinal Survey of Youth-Child Data (1986-1996) to assess the relative impact of adverse birth outcomes vis-à-vis social risk factors on children's developmental outcomes. Using the Peabody Individual Achievement Tests of Mathematics and Reading Recognition as our outcome variables, we also evaluated the dynamic nature of biological and social risk factors from ages 6 to 14. We found the following: (1) birth weight is significantly related to developmental outcomes, net of important social and economic controls; (2) the effect associated with adverse birth outcomes is significantly more pronounced at very low birth weights (< 1,500 grams) than at moderately low birth weights (1,500-2,499 grams); (3) whereas the relative effect of very low-birth-weight status is large, the effect of moderately low weight status, when compared with race/ethnicity and mother's education, is small; and (4) the observed differentials between moderately low-birth-weight and normal-birth-weight children are substantially smaller among older children in comparison with younger children.
Content may be subject to copyright.
Low Birth Weight, Social Factors, and Developmental Outcomes 353
Demography, Volume 39-Number 2, May 2002: 353–368 353
We used six waves of the National Longitudinal Survey of Youth–Child Data (1986–1996) to
assess the relative impact of adverse birth outcomes vis-à-vis social risk factors on children’s devel-
opmental outcomes. Using the Peabody Individual Achievement Tests of Mathematics and Reading
Recognition as our outcome variables, we also evaluated the dynamic nature of biological and so-
cial risk factors from ages 6 to 14. We found the following: (1) birth weight is significantly related to
developmental outcomes, net of important social and economic controls; (2) the effect associated
with adverse birth outcomes is significantly more pronounced at very low birth weights (< 1,500
grams) than at moderately low birth weights (1,500–2,499 grams); (3) whereas the relative effect of
very low-birth-weight status is large, the effect of moderately low weight status, when compared
with race/ethnicity and mothers education, is small; and (4) the observed differentials between mod-
erately low-birth-weight and normal-birth-weight children are substantially smaller among older
children in comparison with younger children.
ocial demographers have had a long-standing interest in analyzing the determinants
of adverse birth outcomes, including low birth weight and prematurity (e.g., Cramer 1995;
Frisbie, Forbes, and Pullum 1996; Kallan 1993; Singh and Yu 1996). For the most part,
this interest stems from the close association between adverse birth outcomes and the risk
of infant mortality. That is, the infant mortality rate for low-birth-weight infants is over
20 times that of their normal-weight counterparts (MacDorman and Atkinson 1999). Fur-
thermore, birth outcomes are often viewed as the key intervening variables that link so-
cial factors, such as race and maternal education, to the risk of infant mortality (Eberstein,
Nam, and Hummer 1990). Indeed, controlling for adverse birth outcomes accounts for
nearly all the gap in infant mortality between blacks and whites in the United States (Hum-
mer et al. 1999).
In contrast, the contribution of adverse birth outcomes to child health and develop-
mental outcomes—particularly in combination with social risk factors—is much less well
established in social demography than in the literature on adverse birth outcomes and
infant mortality. This is particularly the case at the national level, where the data require-
ments for such longitudinal linkages between events that occur at birth and outcomes
many years later are especially stringent. Indeed, unlike studies that have relied on large-
scale databases to link birth outcomes and social risk factors to the risk of infant mortal-
ity, there are relatively few data sets that contain the information that is necessary to link
*Jason D. Boardman, Daniel A. Powers, and Robert A. Hummer, Population Research Center and Depart-
ment of Sociology, University of Texas at Austin. Yolanda C. Padilla, School of Social Work and Population
Research Center, University of Texas at Austin. Direct correspondence to Jason D. Boardman, Population Re-
search Center, 1800 Main Building, University of Texas at Austin, Austin, TX 78712; E-mail: This research was funded by NICHD Grant R01-35949. The authors thank W. Parker
Frisbie, Brian Karl Finch, Shelley Blozis, Marilyn Espitia, and the three anonymous reviewers for their helpful
comments on earlier versions of this article.
354 Demography, Volume 39-Number 2, May 2002
adverse birth outcomes with the long-range health and development trajectories of chil-
dren and adolescents. In our study, we used the National Longitudinal Study of Youth–
Child Data (hereafter NLSY–CD), one of the only national-level longitudinal databases
of its kind, to answer questions about the relationship of low birth weight and social risk
factors to the long-term development of children.
Our purpose is to answer three questions regarding the long-term developmental out-
comes of low-birth-weight children. First, are there net negative effects of low birth
weight on cognitive outcomes—as measured by scores on standardized reading and math
achievement tests—among a national sample of U.S. children? Second, if so, does the
effect of birth weight vary across the ages of children? Finally, how do social factors
affect the cognitive development scores of children, net of birth weight, and how do the
effects of these factors vary across the ages of children?
There is a growing body of research on the relationship between adverse birth outcomes
and subsequent child development outcomes. Findings from these studies suggest that
the negative association between low birth weight and cognitive development begins in
early childhood (Hack et al. 1995) and may extend well into and beyond adolescence
(Behrman, Rosenzweig, and Taubman 1994). For example, using data from the 1981
National Health Interview Study–Child Health Supplement, McCormick, Gortmaker, and
Sobol (1990) found a significantly higher incidence of school difficulty (measured ac-
cording to whether the child had repeated a grade or was attending special classes) and
hyperactivity among children aged 4 to 18 who were born with a very low weight. A
greater incidence of developmental problems related to school achievement, in the area
of attention deficit, was also found among six-year-old children who were born with low
weight compared with normal-weight children (Breslau 1996). Likewise, in a recently
well-publicized study, Conley and Bennet (2000) used data from the Panel Study of In-
come Dynamics and found that the probability of a low-birth-weight infant (< 2,500
grams) completing high school in a timely fashion (i.e., by age 19) is 74% lower than the
probability of his or her normal-birth-weight siblings.
The physiological processes underlying the relationship between adverse birth out-
comes and children’s subsequent development is described in the important works of
Barker and colleagues (Barker 1995; Barker et al. 1993), who documented a relationship
between low birth weight and the risk of coronary heart disease among adults. These
researchers articulated the “fetal programming hypothesis” (i.e., the Barker hypothesis),
which states that low birth weight is a risk factor for poor developmental outcomes be-
cause the same processes that cause low birth weight also cause poor subsequent devel-
opment. For example, blood pressure, cholesterol levels, and hormonal levels are believed
to be “programmed” at an early stage of fetal development. Programming of the fetus
occurs when known risks occur at critical (sensitive) periods in early life, which may
have long-lasting impacts on metabolism and physiology. In particular, the nutritional
supply to the placenta may not match placental demand at particularly important develop-
mental moments, which may lead to reduced physical growth and long-term physiologi-
cal problems. The bulk of the research on the fetal programming hypothesis has evaluated
physical health outcomes (Godfrey and Barker 2001); however, there may be reasons to
expect that the processes that affect physiological development may have deleterious ef-
fects on other important developmental outcomes as well.
Social risk factors have been found to mediate significantly the effects of birth out-
comes on the long-term cognitive development of low-birth-weight children (Hack et al.
1995). Studies have consistently shown that socioeconomic disadvantage has a detrimen-
tal effect on cognitive functioning, as well as on a range of outcomes related to school
achievement, including absenteeism, the repetition of grades, and the risk of high school
Low Birth Weight, Social Factors, and Developmental Outcomes 355
dropout (Brooks-Gunn and Duncan 1997; Crooks 1995; McLoyd 1998). A study of white
and black children explored whether both the timing and duration of poverty is relevant
for cognitive development (Smith, Brooks-Gunn, and Klebanov 1997). On the basis of
two separate samples, the authors investigated three cognitive development measures—
IQ, verbal ability, and school achievement—for children aged 2 to 8. They found that the
longer a child was in poverty, the larger the negative consequences for cognitive develop-
ment, although the effects of the timing of poverty were not significant for this young age
group. In addition, controlling for socioeconomic factors, they found that low birth weight
had no significant effect on cognitive outcomes.
Two studies on the relationship between social context and children’s developmental
outcomes found strong ties between family-level socioeconomic characteristics and the
risk of low scores on several achievement tests. First, using the same data set that we
used in our analyses, Guo (1998) found that although the between-child variation is sig-
nificantly larger than the within-child variation, socioeconomic resources are important
predictors of both between- and within-child variation in achievement tests (like the scores
on the Peabody Individual Achievement Tests, PIATs, used in the present analyses). More-
over, Guo’s study found that the disadvantage associated with low socioeconomic status
increases over children’s lives. Similarly, Guo and Harris (2000), building on this previ-
ous research, found that these effects of poverty on children’s development are entirely
mediated by household characteristics (e.g., parenting style, cognitive stimulation in the
home, and physical environment of the home) and appear to be more pronounced in ado-
lescence than in early childhood. Moreover, they indicated that differences in birth weight
account for a portion of the difference in developmental outcomes among poor and
nonpoor children. However, they did not investigate possible differential birth-weight ef-
fects for children of different ages.
A key limitation of current research is that few studies have investigated the effects
of birth weight for different age groups of children. One study, which was restricted to
the period of early childhood (Klebanov et al. 1998), showed that the effects of very low
birth weight on cognitive functioning decreased significantly between the ages 1 and 3.
The study was based on a sample of 374 low-birth-weight, preterm children from the
Infant Health and Development Program. It contained rich measures of socioeconomic
status, including family poverty, family risk factors (e.g., family structure, mothers edu-
cation and cognitive scores, being a teenage parent at the time of the birth, depression,
and social support), and neighborhood poverty. Controlling for socioeconomic character-
istics, Klebanov et al. observed a significant negative effect of birth weight when the
children were age 1 that was no longer evident at ages 2 and 3. The results, however,
should be interpreted cautiously, given that the study was restricted to low-birth-weight
infants and contained a limited sample.
Although this research provided useful evidence that helps to clarify these complex
relationships, important issues remain unresolved. Specifically, there is convincing evi-
dence that biological (birth weight) and social (socioeconomic) risks are related to devel-
opmental outcomes. Whether the relative explanatory power of these characteristics var-
ies with a child’s age, however, is not known. A long-term study of British children who
were born in 1970 followed up with the respondents at ages 5, 10, 16, and 26 (Strauss
2000). The sample of 14,189 full-term infants included 1,064 infants who were born small
for their gestational age (SGA) (their birth weight was less than the fifth percentile for
age at term). The study revealed that the children who were born SGA had academic
difficulties into adolescence. According to Strauss, “children who were born SGA dem-
onstrated significant deficits in a wide range of standardized testing from the ages of 5 to
16 years” (p. 627). Substantively, however, the differences were fairly small. Measures of
academic achievement included teachers’ ratings of general knowledge, standardized tests
of academic achievement, and enrollment in special education. The analysis did not focus
356 Demography, Volume 39-Number 2, May 2002
on different levels of cognitive functioning across the three age periods, but rather on
comparisons between those who were born SGA and those who were of normal weight
for their gestational age. Therefore, no information was provided on how the effects
changed across the age span. At age 26, the respondents were compared on educational
attainment across birth outcomes. Strauss found that the effects of birth weight on educa-
tional attainment disappeared by young adulthood (age 26), although some negative ef-
fects were evident in terms of occupational attainment.
Corman and Chaikind (1993) examined the effects of low birth weight on the school
performance of children in two age groups, 6 to 10 and 11 to 15. The results of their
study were mixed. Using the 1988 Child Health Supplement of the National Health In-
terview Survey, the authors found that low birth weight was not significantly associated
with positive parental reports of school performance in the younger age group. Never-
theless, low birth weight substantially and significantly increased the probability of
grade repetition in the older age group. Analyses were conducted for the entire sample of
children and for a subsample of children who were not attending special education.
However, the study was limited to a cross-sectional analysis and did not follow the chil-
dren over time.
Perhaps the most similar work to our present study was by Lee and Barratt (1993),
who used the first two waves of the NLSY–CD and found a reduction in the effects of
birth weight on achievement scores in as little as two years when they compared children
aged 5 to 6 with those aged 7 to 8. Similarly, they found tentative evidence that the ef-
fects associated with social risk factors, such as low socioeconomic status and risky home
environments, on children’s developmental outcomes may increase over time. It is not
clear, however, that this relationship is consistent across both mathematics and reading
achievement assessments. Nor is it clear that this difference is statistically significant,
given the parameterization of their models. Our work builds on this complete body of
research by using six waves of the NLSY–CD, which allowed us to evaluate the dynamic
relationships among birth weight, social risk factors, and child development across a much
longer period than has been previously examined.
Data for this study came from the NLSY–CD. The NLSY–CD was developed from the
original National Longitudinal Survey of Youth (NLSY), a nationally representative
sample of 12,686 adolescents aged 14 to 22 as of the 1979 survey date, and contains
detailed longitudinal information on health and development outcomes from birth
throughout the teenage years for the children of mothers from the original youth cohort
(Center for Human Resource Research 1993, 1998). Data were collected every two years
from 1986 to 1996 (i.e., six data collection points) regarding the children of the female
NLSY respondents. The NLSY–CD contains a number of commonly used measures to
assess a range of cognitive, social, and psychological aspects of the children’s develop-
ment. Data on the mothers from the NLSY are linked with corresponding NLSY–CD
records on a year-by-year basis.
The NLSY–CD is particularly useful for our analyses because blacks and Hispanics
are oversampled, making it possible to conduct meaningful analyses for a more diverse
sample of U.S. children. The relatively small size of the Hispanic subsample, however,
does not permit detailed analyses of Hispanic subgroups. Given the diversity of the
Hispanic population with respect to our key variables of interest, we restricted our analy-
ses to Mexican American, non-Hispanic black, and non-Hispanic white children. Never-
theless, this is a substantial improvement over most studies in this area that have not
examined Hispanic children at all. Although the data are based on an oversample of
Low Birth Weight, Social Factors, and Developmental Outcomes 357
blacks and Hispanics, we did not weight the data because we included children from all
six waves of the NLSY–CD (for a discussion, see Center for Human Resource Research
We pooled the data for children aged 6 to 14 from the 1986, 1988, 1990, 1992, 1994,
and 1996 NLSY–CD. We limited our analyses to children aged 14 and younger because
starting in 1994, children aged 15 and older were assessed with different developmental
instruments from those used to assess younger children (Center for Human Resource Re-
search 1998:1). This age restriction allowed us to include children for a maximum of five
points in time. In total, our analyses included 1,890 non-Hispanic black, 2,411 non-His-
panic white, and 844 Mexican American children, for a total N of 5,145. Because children
are included anywhere from one to five times in the pooled data set, however, our total
number of observations was 12,295. Last, given that the NLSY–CD data were collected
from NLSY mothers, a number of children in the data set had the same mothers. Specifi-
cally, the 5,145 children in our sample were born to 2,747 women.
Our dependent variables included two widely used achievement tests: (1) The Peabody
Individual Achievement Test of Mathematics (PIAT-M) and (2) the PIAT Reading Rec-
ognition (PIAT-RR) Test. The PIAT-M assessment measures children’s “attainment in
mathematics as taught in mainstream education” (Center for Human Resource Research
1998:60). A multiple-choice test, the PIAT-M starts with basic mathematics skills, such
as number recognition, and ultimately progresses to advanced topics like geometry and
trigonometry. Age-normalized percentile scores are computed for all children (see Dunn
and Markwardt 1970:81–91, 95 for the norm procedures). The PIAT-RR measures two
key reading skills: word recognition and pronunciation ability. After reading a word si-
lently, children are asked to repeat the word aloud; they are assessed on matching letters,
naming names, and reading single words aloud with 84 items. We used PIAT-RR age-
normalized percentile scores. All observations with missing values for the PIAT assess-
ment scores (n = 910; 6.8%) were dropped from our analyses. Ancillary analyses (results
not presented) indicated that these cases were not significantly different from the rest of
the sample with respect to birth weight, poverty status, or race/ethnicity.
Our primary independent variable, birth weight, was measured with mothers’ self-
reports. Although maternal reports of birth weight are clearly less accurate than birth
certificates, they have been used in several related analyses (e.g., Conley and Bennet
2000; Cramer 1995) and generally are considered to be valid. Similarly, we recognized
the heterogeneity within the group of low-birth-weight children (e.g., Frisbie et al. 1996;
Solis, Pullum, and Frisbie 2000). Accordingly, we operationalized birth weight with
three categories: (1) very low birth weight (VLBW: < 1,500 grams), (2) moderately low
birth weight (MLBW: 1,500–2,499 grams), and (3) normal birth weight (NBW: > 2,499
We also included a number of important sociodemographic characteristics that may
affect the relationship between low birth weight and developmental outcomes. We lim-
ited our analyses to three racial/ethnic groups: Mexican Americans, non-Hispanic blacks,
and non-Hispanic whites. In all models, non-Hispanic white is the reference category.
We also included seven other sociodemographic variables in our multivariate models.
The first four were child’s age, measured with a continuous variable in years; sex, a
dummy variable coded 1 = female and 0 = male; marital status, a dummy variable coded
1 = married if the mother was married at the time of the interview and 0 = otherwise;
and age of mother at birth, a continuous variable tapping maternal age measured in
years. The remaining sociodemographic variables were poverty, a variable for all the
respondents based on family income, household size, and the year of the survey, coded 1
= below poverty and 0 = otherwise; mothers education—less than a high school educa-
358 Demography, Volume 39-Number 2, May 2002
tion, high school graduate, some college, and college graduate or higher (reference
category); and (7) HOME, operationalizing the quality of children’s household context
with the Home Observation for Measurement of the Environment–Short Form (HOME–
SF). Widely used, the characteristics of the home environment covered by this scale
(percentile score) measure the extent to which children’s home context provides
cognitive stimulation and emotional support (Center for Human Resource Research
The descriptive statistics of all variables used in the analyses by the year of the
survey are presented in Table 1. These statistics are illustrative because they clearly
show the changing demographic profile of the NLSY–CD sample over the six waves of
data from 1986 to 1996. With respect to our primary independent variable, birth weight,
children in the earliest wave were significantly more likely to be born at weights less
than 2,500 grams compared with those from the most recent survey. This relationship is
understandable because the mothers of the children from the earlier waves of data were
more likely to have characteristics associated with the increased likelihood of adverse
birth outcomes. That is, they were more likely to be black, young, less educated, poor,
and unmarried than those from each successive wave of the survey (Center for Human
Resource Research 1993). Last, it is important to note the increasing mean age of chil-
dren in our sample for each successive wave of data. Specifically, the mean age of chil-
dren from the earliest wave of data is nearly two years less than that of children from the
last wave of data, and these mean ages increase monotonically across the waves. Again,
this pattern is understandable when we consider that the first waves of data contained
few children aged 12 or older. Indeed, in the 1986 sample, only 38 children were 12 to
14 years old.
Statistical Analysis
We evaluated the relationship among low birth weight, social risk factors, and child de-
velopment with a multilevel statistical method that allows for dependence among obser-
vations within children and children within families and provides parameter estimates
that enabled us to describe variations in our outcomes measures that were due to this
clustering. The model used for these data explicitly takes account of the unique features
of the NLSY sample of children in families followed over time and can be expressed in
the following framework:
= β
+ Σ
+ v
+ u
+ e
, (1)
where y
is the PIAT assessment score for the jth individual in the kth family on the ith
measurement occasion (or survey year), x
corresponds to the value of the hth covariate
(h = 1, . . . , H) for that individual, and β
and β
represent intercept and slope param-
eters to be estimated. The disturbance terms (v
, u
, and e
) denote random effects,
which are independently normally distributed with means equal to 0 and variances σ
, and σ
, respectively. The v
(k = 1, . . .,2,747) term represents unobserved family-
level factors affecting y that are shared by all n
children in the kth family (whereas u
represents unobserved traits for the jth child (j = 1, . . ., n
) in the kth family) and that are
assumed to be constant over the m
measurement occasions. The e
term denotes unob-
served heterogeneity in PIAT assessment scores specific to the jth child measured on the
ith occasion (i = 1, . . ., m
). Therefore, Eq. (1) describes a multilevel model in which
measurement occasions (at level 1) are nested within children (at level 2) who are, in
turn, nested within families (at level 3) (see, for example, Goldstein 1995). Controlling
for these sources of variability, observations are independent. Maximum-likelihood esti-
mation of the model in Eq. (1) yields unbiased parameter estimates and standard errors
that are adjusted for the hierarchical nature of the data. We used SAS 8.1 PROC MIXED
Low Birth Weight, Social Factors, and Developmental Outcomes 359
Table 1. Descriptive Statistics for All Variables by Year of Survey: 1986–1996 NLSY–CD
Year of Survey
Variables 1986 1988 1990 1992 1994 1996 Total
Continuous Variables
PIAT-M 46.29 44.97 46.33 46.82 48.66 52.14 47.76
(24.89) (25.26) (25.66) (26.06) (26.31) (26.94) (26.09)
PIAT-RR 55.99 52.57 53.19 54.28 54.09 56.77 54.41
(25.18) (26.58) (27.18) (27.48) (28.27) (27.81) (27.35)
Child’s Age 7.87 8.55 9.03 9.46 9.78 9.81 9.21
(1.78) (2.16) (2.32) (2.39) (2.43) (2.49) (2.38)
HOME Score 46.16 46.08 47.72 48.65 47.56 47.71 47.47
(29.22) (29.14) (29.13) (29.48) (29.06) (28.35) (29.06)
Maternal Age 18.15 19.29 20.39 21.56 22.97 24.64 21.56
(1.98) (2.33) (2.66) (2.95) (3.10) (3.21) (3.49)
Categorical Variables
Birth Weight
VLBW (< 1,500 grams) 1.4 1.1 0.7 0.9 0.9 1.1 1.0
MLBW (1,500–2,499 grams) 9.9 8.4 7.3 6.2 6.2 5.1 6.9
NBW ( 2,500 grams) 88.7 90.5 92.0 92.9 92.9 93.8 92.1
Non-Hispanic black 42.9 37.9 41.4 37.2 34.6 33.6 37.3
Mexican American 12.9 14.6 18.3 18.8 18.5 16.9 17.1
Non-Hispanic white 44.2 47.5 40.3 44.0 46.9 49.5 45.6
Sex of Child
Female 49.9 49.7 50.1 50.8 49.4 48.9 49.8
Male 50.1 50.3 49.9 49.2 50.6 51.1 50.2
Mothers Education
Less than high school 46.7 41.1 31.7 27.4 23.1 19.2 29.6
High school graduate 41.1 41.6 44.4 44.6 42.8 39.7 42.5
Some college 10.9 14.4 19.9 20.8 24.0 26.5 20.5
College graduate 1.3 2.9 4.0 7.2 10.1 14.6 7.4
Poverty Status
Poor 41.3 38.8 31.3 31.0 25.7 21.8 30.4
Not poor 58.7 61.2 68.7 69.0 74.3 78.2 69.6
Marital Status
Married 52.9 54.2 56.9 58.7 63.7 65.9 59.5
Unmarried 47.1 45.8 43.1 41.3 36.3 34.1 40.5
N 1,070 1,992 2,039 2,453 2,406 2,335 12,295
Cell entries for continuous variables represent means, with standard deviations shown in parentheses.
Cell entries for categorical predictors represent percentages.
360 Demography, Volume 39-Number 2, May 2002
(see Littell et al. 1996 for a detailed overview of this application) to estimate these mod-
The results of the multilevel regression models are presented in Tables 2 and 3. These
tables present three sets of model estimates for our null model (without any covariates),
full model (with all covariates), and an interactive model in which we evaluated the ex-
tent to which each covariate varies by the child’s age. The estimates in the null models
provide baseline estimates for the intrachild and intrafamily correlation coefficients. Spe-
cifically, 23.4% and 34.4% of the variability in the PIAT-M and PIAT-RR are due to un-
measured child characteristics, and 36% of this variability (for each outcome) is due to
unmeasured family characteristics. These estimates are important because they illustrate
the dependence among observations nested within children and children nested within
families that otherwise would be lost in more traditional single-level approaches.
The second set of parameter estimates addresses one of the key research questions in
our study—namely, are there any independent negative effects associated with low-birth-
weight status on children’s developmental outcomes? According to these values, VLBW
is associated with a 9.5- and 11.4-point decrease in children’s PIAT-M and PIAT-RR
scores, respectively. And although significantly smaller in magnitude (b = –2.9, p < .001),
the estimated net effect of MLBW on each PIAT assessment is, as well, negative and
statistically significant. This finding is important because these models control for a range
of meaningful social and economic characteristics. In other words, although VLBW and
MLBW children are more likely to be from relatively disadvantaged backgrounds (Cramer
1995; Kallan 1993), both VLBW and MLBW statuses are operating independent of these
socioeconomic characteristics to have adverse effects on developmental outcomes.
These model estimates also illustrate the significance of several social and economic
characteristics on children’s development. Here, it is important to note the distributional
characteristics of our two dependent variables. Specifically, the standard deviations for
PIAT-M (SD = 27.4) and PIAT-RR (SD = 26.1) scores help contextualize the magnitude
associated with the effect of VLBW and MLBW vis-à-vis the social risk factors in the
models. First, non-Hispanic black children scored over 13 points lower (nearly half a
standard deviation) on the PIAT-M and 9 points lower on the PIAT-RR than did non-
Hispanic white children with similar observed social and economic characteristics. And
though the race/ethnicity differential is not as large as the black-white differential, the
Mexican American children, net of our controls, scored 10 points and 6 points lower on
the PIAT-M and PIAT-RR, respectively, than did the non-Hispanic white children. Sec-
ond, it is clear from these results that maternal socioeconomic status, particularly educa-
tion, is a significant predictor of children’s developmental outcomes. Specifically, chil-
dren whose mothers did not complete high school scored 16 and 17 points lower (than did
1. The random-effects model assumes that unobserved heterogeneity is a random variable (differing by
families) that follows a normal distribution and is uncorrelated with the regressors, whereas a fixed-effects
model treats unobserved heterogeneity as a fixed but unknown variable. In either case, the nature of the random
effect is the same. However, there are subtle differences between these models in terms of inference. In the
fixed-effects model, inference is conditional on the effects that are in the sample. Random-effects models are
concerned with marginal inferences with respect to the population of all effects (Hsiao 1986). Our interpretation
considers the NLSY families to be a random sample from a larger population of families. Inferences based on
the population of all families, rather than on the families actually sampled in the NLSY, are more appropriately
handled using a random-effects model. A similar case can be made with regard to the child-level random effect.
In addition, to address the assumption of normality of residuals in random-effects models, we obtained empiri-
cal Bayes estimates of the family-level and child-level random effects from our complete model and assessed
normality assumptions heuristically using standard normal probability plots. We found no evidence of
nonnormality of level-2 residuals for either level of analysis (these results are available from the authors on
Low Birth Weight, Social Factors, and Developmental Outcomes 361
Table 2. Multilevel Regression Coefficients: Low Birth Weight, Social Risk Factors, and PIAT-M
Assessment Scores: 1986–1996 NLSY–CD
Interactive Model
Fixed Effects Null Model Full Model Main × Age
Intercept 49.02*** 62.94*** 79.02*** ––
(0.39) (2.78) (6.72)
Child’s Age (Years) –0.42*** –2.30*** ––
(0.07) (0.71)
Birth-Weight Status [ 2,500 grams]
VLBW (< 1,500 grams) –9.47*** –4.32 –0.55
(2.63) (7.17) (0.74)
MLBW (1,500–2,499 grams) –2.97*** –8.13** 0.57*
(1.11) (2.82) (0.28)
Race/Ethnicity [Non-Hispanic White]
Non-Hispanic black –13.39*** –7.29*** –0.67***
(0.79) (1.78) (0.17)
Mexican American –9.92*** –8.41*** –0.17
(0.99) (2.09) (0.21)
Female –0.25 4.49** –0.53***
(0.54) (1.39) (0.14)
Maternal Age 0.18* –0.83*** 0.12***
(0.09) (0.23) (0.02)
HOME Score 0.08*** 0.13*** –0.01**
(0.00) (0.03) (0.00)
Mothers Education [College Graduate]
Less than high school –15.95*** –15.95*** –0.03
(1.32) (3.29) (0.34)
High school graduate –10.93*** –13.19*** 0.23
(1.20) (3.00) (0.32)
Some college –6.56*** –6.38* –0.04
(1.22) (3.17) (0.33)
Poor –1.65** 0.69 –0.24
(0.55) (1.88) (0.19)
Unmarried 0.66 –3.33* 0.29
(0.59) (1.78) (0.18)
Residual Variance
: Child level 158.49 156.15 158.15
(8.08) (7.96) (7.99)
: Family level 248.33 138.27 137.99
(11.91) (8.94) (8.95)
: Observation level 268.24 268.87 265.74
(4.45) (4.46) (4.41)
Chi-Square –– 770.80*** 64.40***
*p < .05; **p < .01; ***p < .001
362 Demography, Volume 39-Number 2, May 2002
Table 3. Multilevel Regression Coefficients: Low Birth Weight, Social Risk Factors, and
PIAT-RR Assessment Scores: 1986–1996 NLSY–CD
Interactive Model
Fixed Effects Null Model Full Model Main × Age
Intercept 55.55*** 75.03*** 95.09*** ––
(0.42) (2.91) (6.32)
Child’s Age (Years) –0.73*** –3.05*** ––
(0.07) (0.65)
Birth-Weight Status [ 2,500 grams]
VLBW (<1,500 grams) –11.39*** –5.63 –0.64
(2.89) (6.81) (0.69)
MLBW (1,500–2,499 grams) –2.85* –7.69** 0.55*
(1.24) (2.64) (0.25)
Race/Ethnicity [Non-Hispanic White]
Non-Hispanic black –9.14*** 7.40*** –1.84***
(0.88) (1.68) (0.16)
Mexican American –5.92*** –4.44* –0.19
(1.10) (1.99) (0.18)
Female 5.59*** 5.16*** 0.03
(0.60) (0.54) (0.13)
Maternal Age –0.19* –1.47*** 0.16***
(0.09) (0.21) (0.02)
HOME Score 0.07*** 0.06* 0.00
(0.01) (0.03) (0.00)
Mothers Education [College Graduate]
Less than high school –16.99*** –11.48*** –0.63*
(1.41) (3.12) (0.31)
High school graduate –9.98*** –7.49*** –0.28
(1.28) (2.85) (0.29)
Some college –5.25*** –6.25* 0.07
(1.27) (3.00) (0.31)
Poor –0.45 –0.34 –0.01
(0.52) (1.72) (0.17)
Unmarried –0.07 2.58 –0.26*
(0.58) (1.64) (0.17)
Residual Variance
: Child level 250.96 242.69 243.80
(9.87) (9.66) (9.59)
: Family level 265.21 180.09 180.59
(13.45) (11.30) (11.28)
: Observation level 211.90 210.72 202.38
(3.53) (3.52) (3.38)
Chi-Square –– 681.50*** 300.30***
*p < .05; **p < .01; ***p < .001
Low Birth Weight, Social Factors, and Developmental Outcomes 363
those whose mothers completed college or higher) on the PIAT-M and PIAT-RR, respec-
tively. And although the positive effects associated with increased educational levels are
monotonic, children whose mothers completed only high school at the time of the assess-
ment scored 11 (PIAT-M) and 10 (PIAT-RR) points lower than did those whose mothers
graduated from college. The effect of maternal education is quite strong and most likely
explains the bulk of the relatively modest effects associated with poverty on PIAT-M
scores (b = –1.65, p < .01) and the insignificant effect of poverty on PIAT-RR scores. We
also found significant effects associated with the levels of cognitive stimulation (HOME
score: b = 0.08, p < .001; b = 0.07, p < .001) in children’s homes to be positively and
significantly related to the children’s performance on both tests. Last, similar to recent
research in this area (Levine, Pollack, and Comfort 2001), we found that maternal age is
positively and significantly related to PIAT-M scores (b = .18, p < .05). We also found a
negative and significant independent effect of maternal age on PIAT-RR scores (b =
–0.19, p < .05). In other words, although children of young mothers are relatively disad-
vantaged when compared with children of older mothers, once these characteristics are
controlled for, children born to younger women actually perform slightly better on the
PIAT-RR test.
Overall, the statistical controls in these models explained little of the between-child
variation but a large share of the between-family variation. Specifically, our full-model
estimates explained roughly 44% (PIAT-M) and 32% (PIAT-RR) of the family-level varia-
tion but only 1% of the between-child variation in PIAT-M scores and 3% of the between-
child variation in PIAT-RR scores. And because there is little variation among siblings
with respect to a number of our predictors (i.e., most of the siblings share similar racial/
ethnic and socioeconomic characteristics) except birth-weight status, it is possible that
birth-weight differentials among siblings account for this relatively small amount of the
variation explained in between-child PIAT-RR scores. In other words, the social context
of children’s households appears to be significantly more influential on children’s devel-
opment than is their birth outcomes.
Our second research question concerns the ways in which the relationships specified
in the full models (Tables 2 and 3) vary significantly with a child’s age to affect develop-
mental outcomes. In other words, we know from the previous models that low-birth-
weight status is negatively and significantly related to developmental outcomes above
and beyond the social and economic context of children’s households and that there are
powerful associations between several of the social risk factors (e.g., race/ethnicity, ma-
ternal education, and HOME score) and our two outcomes. To assess the dynamic rela-
tionship between low birth weight, social risk factors, and developmental outcomes with
age, we estimated an interactive model with additional parameter estimates for the inter-
action between each covariate and child’s age.
We first focus on the interaction of birth weight by child’s age. Tables 2 and 3 show
that the deleterious effect of MLBW status on PIAT scores significantly decreases in mag-
nitude by over half a point with each additional year of age (PIAT-M: b = 0.57, p < .05;
PIAT-RR: b = 0.55, p < .05). The negative effect associated with VLBW status, however,
does not vary by age. It is also important to note that non-Hispanic black children fall
increasingly behind their non-Hispanic white counterparts on both PIAT assessments from
ages 6 to 14. This relationship is most pronounced among reading scores, where the black-
white differential increases by roughly 2 points (b = –1.84, p < .001) with each additional
year of age. Among the PIAT-M scores, the rate at which black children fall behind white
children is not as severe (b = –0.67, p < .001), but follows a pattern similar to that of the
PIAT-RR scores. To illustrate these findings, we plotted the predicted black-white and
low-birth-weight–normal-birth-weight differentials in the PIAT-RR scores by children’s
ages (see Figure 1). According to these estimates, MLBW children score over 4 points
lower than NBW children with similar social and economic characteristics at age 6; but
364 Demography, Volume 39-Number 2, May 2002
by age 14, there is no difference between MLBW and NBW children on the PIAT-RR.
The disparity between black and white children’s scores, however, operates in the oppo-
site direction: at age 6, net of social and economic controls, non-Hispanic black children
score, on average, less than 4 points lower than non-Hispanic white children. By age 14,
however, this differential is greater than 18 points. The same general pattern holds for
PIAT-M scores (this figure is not shown but is available from the authors on request).
We also observed significant interaction terms for child’s age with maternal age, gen-
der (math only), HOME score (math only), education (reading only), and marital status
(reading only). In particular, although girls score slightly higher on the PIAT-M assess-
ment at age 6, on average, they appear to fall a 0.5 percentile point behind boys every
year. Equally important is the increasing difference between the reading scores of chil-
dren whose mothers have less than a high school education and those whose mothers
completed college. At age 6, children with mothers who did not graduate from high school
score an estimated 15 points lower on the PIAT-RR than do those whose mothers gradu-
ated from college; at age 14, this gap is over 20 points (b = –0.63, p < .05). Last, on the
PIAT-RR, children of unmarried mothers fall an estimated 0.3 point behind children whose
mothers are married with every year. At age 6, children whose mothers are unmarried do
not significantly differ from children whose parents are married on the PIAT-RR. By age
14, however, children of unmarried mothers score roughly 2 points lower than do those
whose parents are married. Taken together, this complete set of interaction effects sug-
gests that whereas the effects of birth outcomes either remain constant (VLBW) or de-
crease (MLBW) in significance with a child’s increasing age, the influence of many of
the social risk factors is more pronounced among older children.
Age (Years)
Black-White Differential
PIAT-RR Differential
MLBW-NBW Differential
67891011 12 13 14
Figure 1. Race and Birth-Weight Differentials in PIAT-RR, by Age
Low Birth Weight, Social Factors, and Developmental Outcomes 365
Demographers have long been interested in studying adverse birth outcomes, largely be-
cause of these outcomes’ strong influence on the risk of infant mortality and other severe
medical problems during early childhood. Few large-scale studies, however, have investi-
gated the effects of adverse birth outcomes on longer-term risks during childhood and
adolescence, mainly because of the stringent data requirements necessary to conduct such
analyses. Using the NLSY–CD, we found that there are modest-sized negative effects of
low birth weight on childhood math and reading scores, although the effects on both are
weaker among older children than among younger children. In contrast, we found that the
effects of race/ethnicity, maternal education, gender, home environment, unmarried sta-
tus, and young maternal age on the developmental outcomes exhibit either constant ef-
fects across children’s ages or are even more pronounced for children at age 14 than at
age 6. These results strongly suggest that although the health of children at birth is surely
important for their long-term development, children’s social experiences are clearly
prominent predictors of long-term well-being.
This article makes four important contributions. First, the findings presented here
highlight the otherwise masked heterogeneity within the group that is conventionally de-
fined as “low birth weight.” Although some have cautioned against the continued use of
the 2,500-gram threshold for low birth weight (Kline, Stein, and Susser 1989) and others
have effectively demonstrated a nonlinear relationship between birth weight and health
risks (Boardman, Finch, and Hummer 2001; Solis et al. 2000), researchers continue to
use this binary categorization of birth weight (e.g., Conley and Bennet 2000) as a conve-
nient predictor of a wide range of health and developmental outcomes. We found a strong
relationship between birth weight and developmental outcomes such that the VLBW chil-
dren (< 1,500 grams) in our sample not only scored significantly lower than the NBW
children but also scored roughly 6.5 (math) and 8.5 (reading) percentile points lower than
the MLBW children (1,500–2,499 grams). And not only did we document an increased
risk of poorer developmental outcomes among VLBW than among MLBW children, we
also presented evidence that the developmental trajectories of these two subpopulations
of children are measurably different. In other words, VLBW children face greater chal-
lenges than do MLBW children in terms of positive developmental outcomes, but they
also may face different challenges. This finding has important public health implications,
given the relative number of VLBW children compared with MLBW children. Specifi-
cally, in 1999, whereas 7.6% of all children born in the United States were less than 2,500
grams at the time of delivery, children born weighing less than 1,500 grams made up only
1.2% of the total number of children who were born (Ventura et al. 2001). Of the 3.96
million births that occurred in 1999, this difference represents an at-risk population of
over 250,000 for MLBW children, compared with under 50,000 for VLBW children.
Second, the differences in the developmental trajectories of VLBW and MLBW chil-
dren that we documented are also important because they build on the work of Conley and
Bennett (2000). These authors used longitudinal data from the Panel Study of Income
Dynamics to show that children who were born with low birth weights are substantially
disadvantaged in graduating from high school in a timely fashion. Together with their
earlier analysis showing a robust influence of parental birth weight on filial birth weight,
these findings suggest a cycle of biological disadvantage such that low birth weight is
strongly influenced by parental birth weight and then goes on to have an adverse influence
on socioeconomic outcomes. Because of data limitations and a subsequently small number
of individuals who were born under 1,500 grams, Conley and Bennett limited their
operationalization of low birth weight to a threshold of 2,500 grams. It is possible that the
effects that they reported were driven primarily by children with VLBW (i.e., less than
1,500 grams). Indeed, according to our results, the at-risk population that has been conven-
366 Demography, Volume 39-Number 2, May 2002
tionally defined as low birth weight, with respect to academic achievement among adoles-
cents, may not necessarily include children who weighed 1,500 to 2,500 grams at birth.
Third, our findings are also important because they suggest that the relative impact of
MLBW vis-à-vis the characteristics of children’s social contexts is small in magnitude. In
particular, the independent net effect of maternal education appears to far outweigh the
effect of MLBW as a predictor of children’s test scores. Moreover, the deleterious effect
of this important characteristic of children’s social context on children’s academic test
scores was more pronounced among older children. These results, which are much more
in line with the current thinking of the broader literature on low birth weight (Hack et al.
1995), suggest that children’s home environments and the socioeconomic and demo-
graphic backgrounds of their parents have a much more powerful influence on children’s
cognitive development than does the weight at which the children were born.
Finally, it is particularly important to point out that our analyses demonstrate large
racial/ethnic disparities in developmental outcomes, such that black and Mexican Ameri-
can children scored below non-Hispanic white children on both the math and reading
tests. These racial/ethnic differentials persisted despite a wide range of controls for
children’s social and economic characteristics. Furthermore, for non-Hispanic black chil-
dren, the differential with non-Hispanic white children was wider among older children
than among younger children, whereas the Mexican American-white gap was constant
across ages. Here, it is important to consider school-based disparities in access to educa-
tional resources among non-Hispanic white, non-Hispanic black, and Mexican American
children as a potentially important mediator between race/ethnicity and performance on
standardized tests. For example, according to Ferguson (2001:381) non-Hispanic black
children (and to a lesser extent Hispanic children) are more likely than non-Hispanic
white children to respond affirmatively to the following statements about their schools:
(1) “too many teachers are doing a bad job”; (2) “not enough emphasis on the basics
such as reading, writing, and math”; (3) “too many kids get passed to the next grade
when they should be held back”; and (4) “classes are too crowded.” Likewise, 13% of
non-Hispanic black children and 16% of Hispanic children, compared with only 6% of
non-Hispanic white children, report being afraid of “being attacked or harmed at school”
(Mayer, Mullens, and Moore 2000). Moreover, Roscigno (1998) found evidence that
school characteristics (i.e., racial composition; socioeconomic characteristics; teachers’
expectations; and, to a lesser extent, student-teacher ratios) are strongly related to stu-
dents’ scores on math and reading achievement tests and that black-white differences in
these characteristics accounted for roughly 14% of the observed racial gap in test scores.
And although differences in class size along racial/ethnic lines have been reduced since
the publication of Coleman et al.’s (1966) influential report, stark differences persist in
the average level of school resources and quality of teachers for black, white, and His-
panic children. For example, first-time teachers (Henke, Chen, and Geis 2000), and un-
qualified teachers (Mayer et al. 2000) are more likely to work in high-minority and high-
poverty schools, respectively: whereas 38% of all teachers work in high-poverty schools,
64% of all unqualified teachers work in high-poverty schools (Mayer et al. 2000). Given
the overrepresentation of non-Hispanic blacks and Hispanics in high-poverty areas
(Jargowsky 1997), the quality of teachers may be an important characteristic that medi-
ates the observed racial/ethnic differentials in both PIAT assessments. It also stands to
reason that these differentials would help account for the larger magnitude of the ob-
served black-white differentials among older children. In other words, not only do non-
Hispanic black children face a disadvantage compared with non-Hispanic white children
in their access to good-quality education, but this disadvantage, with respect to reading
and math scores, also appears to be cumulative. Clearly, aggressive steps need to be
taken to help end the racial/ethnic disparities in parental, school, and neighborhood re-
sources on which children’s well-being depends.
Low Birth Weight, Social Factors, and Developmental Outcomes 367
Barker, D.J. 1995. “Fetal Origins of Coronary Heart Disease.” British Journal of Medicine
Barker, D.J., C. Osmond, S.J. Simmonds, and G.A. Wield. 1993. “The Relation of Small Head
Circumference and Thinness at Birth to Death from Cardiovascular Disease in Adult Life.” Brit-
ish Journal of Medicine 306:422–26.
Behrman, J.R., M.R. Rosenzweig, and P. Taubman. 1994. “Endowments and the Allocation of
Schooling in the Family and in the Marriage Market: The Twins Experiment.” Journal of Politi-
cal Economy 102:1131–74.
Boardman, J.D., B.K. Finch, and R.A. Hummer. 2001. “Race, Birth Weight, and Maternal Reports
of Asthma for a Nationally Representative Sample of Young Children in the United States.”
Population Research and Policy Review 3:1–20.
Breslau, N. 1996. “Psychiatric Sequela of Low Birth Weight at 6 Years of Age.” Journal of Abnor-
mal Child Psychology 24:385–401.
Brooks-Gunn, J. and G.J. Duncan. 1997. “The Effects of Poverty on Children.” The Future of
Children 7:55–71.
Center for Human Resource Research. 1993. NLSY Child Handbook-Revised Edition. Columbus:
Ohio State University, CHRR User Services.
———. 1998. 1996 Child and Young Adult Data: Users Guide. Columbus: Ohio State University,
CHRR User Services.
Coleman, J.S., E.Q. Campbell, C.J. Hobson, J. McPartland, A.M. Mood, F.D. Wienfield, and R.L.
York. 1966. Equality of Educational Opportunity. Washington, DC: U.S. Government Printing
Conley, D. and N.G. Bennett. 2000. “Is Biology Destiny? Birth Weight and Life Chances.” Ameri-
can Sociological Review 65:458–67.
Corman, H. and S. Chaikind. 1993. The Effect of Low Birthweight on the Health, Behavior, and
School Performance of School-Aged Children. Working Paper Series, No. 4409. Cambridge,
MA: National Bureau of Economic Research.
Cramer, J.C. 1995. “Racial and Ethnic Differences in Birthweight: The Role of Income and Finan-
cial Assistance.” Demography 32:231–47.
Crooks, D.L. 1995. “American Children At Risk: Poverty and Its Consequences for Children’s
Health, Growth, and School Achievement.” Yearbook of Physical Anthropology 38:57–86.
Dunn, L.M. and F.C. Markwardt, Jr. 1970. Peabody Individual Achievement Test Manual. Circle
Pines, MN: American Guidance Service.
Eberstein, I.W., C.B. Nam, and R.A. Hummer. 1990. “Infant Mortality by Cause: Main and Inter-
action Effects.” Demography 27:413–30.
Ferguson, R.F. 2001. “Test-Score Trends Along Racial Lines, 1971–1996: Popular Culture and
Community Academic Standards.” Pp. 348–90 in America Becoming, edited by N.J. Smelser,
W.J. Wilson, and F. Mitchell. Washington, DC: National Academy Press.
Frisbie, W.P., D. Forbes, and S.G. Pullum. 1996. “Compromised Birth Outcomes and Infant Mor-
tality Among Racial and Ethnic Groups.” Demography 33:469–81.
Godfrey, K.M. and D.J. Barker. 2001. “Fetal Programming and Adult Health.” Public Health Nu-
trition 4:611–24.
Goldstein, H. 1995. Multilevel Statistical Models. 2nd ed. New York: Halstad Press.
Guo, G. 1998. “The Timing of the Influences of Cumulative Poverty on Children’s Cognitive Abil-
ity and Achievement.” Social Forces 77:257–88.
Guo, G. and K.M. Harris. 2000. “The Mechanisms Mediating the Effects of Poverty on Children’s
Intellectual Development.” Demography 37:431–47.
Hack, M., N.K. Klein, and H.G. Taylor. 1995. “Long-Term Developmental Outcomes of Low Birth
Weight Infants.” The Future of Children 5:176–96.
Henke, R., X. Chen, and S. Geis. 2000. Progress Through the Teacher Pipeline: 1992–93 College
368 Demography, Volume 39-Number 2, May 2002
Graduates and Elementary/Secondary School Teaching as of 1997. NCES 2000-152. U.S. De-
partment of Education, National Center for Education Statistics. Washington, DC: U.S. Gov-
ernment Printing Office.
Hsiao, C. 1986. Analysis of Panel Data. Cambridge, England: Cambridge University Press.
Hummer, R.A., M. Biegler, P.B. DeTurk, D. Forbes, W.P. Frisbie, Ying Hong, and S.G. Pullum.
1999. “Race/Ethnicity, Nativity, and Infant Mortality in the United States.” Social Forces
Jargowsky, P.A. 1997. Poverty and Place: Ghettos, Barrios, and the American City. New York:
Russell Sage Foundation.
Kallan, J.E. 1993. “Race, Intervening Variables, and Two Components of Low Birth Weight.” De-
mography 30:489–506.
Klebanov, P.K., J. Brooks-Gunn, C. McCarton, and M.C. McCormick. 1998. “The Contribution of
Neighborhood and Family Income to Developmental Test Scores Over the First Three Years of
Life.” Child Development 69:1420–36.
Kline, J., Z. Stein, and M. Susser. 1989. Conception to Birth: Epidemiology of Prenatal Develop-
ment. New York: Oxford University Press.
Lee, H. and M.S. Barratt. 1993. “Cognitive Development of Preterm Low Birth Weight Children at
5 to 8 Years Old.” Developmental and Behavioral Pediatrics 14:242–48.
Levine, J.A., H. Pollack, and M.E. Comfort. 2001. “Academic and Behavioral Outcomes Among
the Children of Young Mothers.” Journal of Marriage and Family 63:355–69.
Littell, R.C., G.A. Milliken, W.W. Stroup, and R.D. Wolfinger. 1996. SAS System for Mixed Mod-
els. Cary, NC: SAS Institute.
MacDorman, M.F. and J.O. Atkinson. 1999. “Infant Mortality Statistics From the 1997 Period
Linked Birth/Infant Death Data Set.” National Vital Statistics Reports 47(23). Hyattsville, MD:
National Center for Health Statistics.
Mayer, D.P., J.E. Mullens, and M.T. Moore. 2000. Monitoring School Quality: An Indicators Re-
port. NCES 2001–30. Washington, DC: National Center for Education Statistics.
McCormick, M.C., S.L. Gortmaker, and A.M. Sobol. 1990. “Very Low Birth Weight Children:
Behavior Problems and School Difficulty in a National Sample.” Journal of Pediatrics 117:687–
McLoyd, V.C. 1998. “Socioeconomic Disadvantage and Child Development.” American Psycholo-
gist 53:185–204.
Roscigno, V.J. 1998. “Race and the Reproduction of Educational Disadvantage.” Social Forces
Singh, G.K. and S.M. Yu. 1996. “Adverse Pregnancy Outcomes: Differences Between U.S.- and
Foreign-Born Women in Major U.S. Racial and Ethnic Groups.” American Journal of Public
Health 86:837–43.
Smith, J.R., J. Brooks-Gunn, and P.K. Klebanov. 1997. “Consequences of Living in Poverty for
Young Children’s Cognitive and Verbal Ability and Early School Achievement.” Pp. 132–89 in
Consequences of Growing Up Poor, edited by G.J. Duncan and J. Brooks-Gunn. New York:
Russell Sage Foundation.
Solis, P., S.G. Pullum, and W.P. Frisbie. 2000. “Demographic Models of Birth Outcomes and In-
fant Mortality: An Alternative Measurement Approach.” Demography 37:489–98.
Strauss, R.S. 2000. “Adult Functional Outcome of Those Born Small for Gestational Age: Twenty-
Six-Year Follow-up of the 1970 British Birth Cohort.” Journal of the American Medical Asso-
ciation 283:625–32.
Ventura, S.J., J.A. Martin, S.C. Curtin, F. Menacker, and B. Hamilton. 2001. “Births: Final Data
for 1999.” National Vital Statistics Reports 49(1).
... invest in their children, which may compensate for the initial health disadvantage. 10,23 This study has two aims. First, we compare cognitive development across early infancy and mid-adolescence in four groups of children: MAR-conceived low birthweight (MAR LBW); MAR-conceived non-low birthweight (MAR NLBW); naturally conceived (NC) low birthweight (NC LBW); and naturally conceived non-low birthweight (NC NLBW). ...
... MAR children, on average, come from socioeconomically advantaged backgrounds, 36 which may compensate for the negative consequences of being born LBW. 23 We used a representative UK longitudinal cohort study to investigate the cognitive development of MAR Figure 1 Predicted scores in cognitive development and 95% confidence intervals of MAR and naturally conceived children by their weight status at birth at different age points. Baseline models include: sex, multiple birth. ...
... 21,48 Second, it may be possible that we observe a convergence as the negative effect of being born LBW for NC conceived children fades away over time, which is consistent with previous studies documenting that the negative effect of LBW become smaller as children grow older. 23,49,50 Our results partially differ from those of a previous study which has analysed the cognitive development of MAR children using the same data until age 11. 3 Our results are consistent with theirs until age 7, after which they differ since we find that MAR NLBW show consistently higher cognitive ability with respect to all other groups before accounting for family characteristics. This difference is likely driven by the fact that we distinguish between NLBW and LBW children, whereas the other study considered MAR children in general. ...
Full-text available
Background Previous research has documented that children conceived through medically assisted reproduction (MAR) are at increased risk of poor birth outcomes, such as low birthweight (LBW), which are risk factors for stunted longer-term cognitive development. However, parents who undergo MAR to conceive have, on average, advantaged socioeconomic backgrounds which could compensate for the negative effects of being born LBW. Previous studies have not analysed whether the negative effects of LBW are attenuated among MAR conceived children. Methods We draw on the UK Millennium Cohort Study (sweeps 1–6) which contains a sub-sample of (N = 396) MAR-conceived children. The dependent variable measures cognitive ability at around ages 3, 5, 7, 11 and 14. We examine the cognitive development of four groups of children: MAR-conceived low birthweight (MAR LBW); MAR-conceived non-low birthweight (MAR NLBW); naturally conceived low birthweight (NC LBW); naturally conceived non-low birthweight (NC NLBW). We estimate the two following linear regression models for each sweep: (i) a baseline model to examine the unadjusted association between cognitive development and low birthweight by mode of conception; and (ii) a model adjusted by socio-demographic family characteristics. Results In baseline models, MAR LBW children [age 3: β = 0.021, 95% confidence interval (CI): -0.198, 0.241; age 5: β = 0.21, 95% CI: 0.009, 0.418; age 7: β = 0.163, 95% CI: -0.148, 0.474; age 11: β = 0.003, 95% CI: -0.318, 0.325; age 14: β = 0.156, 95% CI: -0.205, 0.517], on average perform similarly in cognitive ability relative to NC NLBW at all ages, and display higher cognitive scores than NC LBW children until age 7. When we account for family characteristics, differences are largely attenuated and become close to zero at age 14. Conclusions Despite the higher incidence of LBW among MAR compared with NC children, they do not seem to experience any disadvantage in their cognitive development compared with naturally conceived children. This finding is likely explained by the fact that, on average, MAR children are born to socioeconomically advantaged parents.
... Children from families with higher socioeconomic status (SES) living with learning disability are more likely to get into integrated special services programs tailored to their special needs compared to those from families with lower SES [8,[14][15][16][17]. Additional studies have also associated children from lower SES with learning disabilities to have a higher probability of repetition of grades, and in other instances, dropping out of high school, thereby amounting to reduced economic workforce [9,[18][19][20]. ...
... Few studies have examined sociodemographic factors and their relationship to learning disability in individual states in the United States [8,18,21]. To the best of my knowledge, no study has explored the relationship of socioeconomic and sociodemographic factors and learning disability in preterm children for all the states combined. ...
Full-text available
Background In 2019, 1 in every 10 infants born in the United States was preterm. Prematurity has life-threatening consequences and causes a range of developmental disabilities, of which learning disability is a prevalent complication. Despite the availability of special services for children living with learning disability, gaps still exist in terms of access due to socioeconomic factors. The aim of this study is to evaluate socioeconomic and sociodemographic correlates of learning disability in preterm children. Methods This cross-sectional study used data from the 2016–2018 National Survey of Children’s Health. Weighted multivariable analyses were conducted to ascertain the association of sociodemographic and socioeconomic factors on learning disability among preterm children. The main outcome variable was the presence of learning disability. Results Among 9555 preterm children in our study population, 1167 (12%) had learning disability. Learning disability was significantly associated with health insurance, food situation, and poverty level after adjustment for other variables. Children currently insured had lower odds of having learning disability compared to those without health insurance (OR = 0.79, 95% C.I. = 0.70–0.91). Also, children living in households that cannot afford nutritious meals are more likely to have learning disability compared to those that can afford nutritious meals at home (OR = 1.55, 95% C.I. = 1.22–1.97). Conclusion These findings highlight the need for intervention efforts to target these children living with a learning disability to achieve the 2004 Individuals with Disabilities Education Act of promoting educational equality and empowerment of children living with a learning disability.
... Of particular concern are differences in health at the start of life, often captured by low birth weight (LBW) (<2500 g). LBW is an important and commonly used indicator of infant health; it predicts infant mortality, early neurodevelopmental problems, and later life outcomes, such as academic performance and adult health (Boardman et al., 2002;Conley & Bennett, 2001;Goldenberg et al., 1996;Hummer et al., 1999). Understanding the extent and causes of ethnoracial disparities at life's "starting gate" (Conley et al., 2003) is a critical element of research and policy agendas focused on increasing racial equity in opportunities. ...
... An important indicator of an infant's health at birth, LBW predicts mortality in the first year of life, a child's academic performance, and adult health (Boardman et al., 2002;Conley & Bennett, 2001;Hummer et al., 1999). Disparities across ethnoracial groups in LBW indicate racial inequality from the first moments of life. ...
Full-text available
In this article, we describe, decompose, and examine correlates of the geography of ethnoracial inequalities in low birth weight (LBW) in the United States. Drawing on the population of singleton births to U.S.-born White, Black, Latinx, and Native American parents in the first decade of the twenty-first century (N = 28.2 million births), we calculate county-level LBW rates and rate ratios. Results demonstrate a stark racial hierarchy in which Black infants experience the most significant disadvantage, but we also document substantial local-level variation organized in what we call a regionalized patchwork of inequality, with high-disparity counties bordering low-disparity counties coupled with regional clustering. Examining the component parts of local disparities – the LBW rates for Whites and groups of color - we find strong evidence that spatial variation in ethnoracial LBW inequalities is driven by greater variation in infants of color's health across counties than by variation in Whites' health. Further, LBW rates for groups of color are only weakly to moderately correlated with Whites' LBW rates, suggesting that the same contexts can produce racially divergent health outcomes. Examining contextual factors that predict LBW disparities, we find that more segregated, socioeconomically unequal, and urban counties have larger LBW disparities. We conclude by positing an approach to health disparities that conceptualizes ethnoracial differences in health as fundamentally relational and spatial phenomena produced by systems of White advantage.
... The patterns of heterogeneity of this effect by parental SES are mixed. Nevertheless, among those studies that found heterogeneity for at least some outcomes (11 out of 18 studies reviewed), evidence for compensation among high-SES families is more commonly found than reinforcement (8 out of 11 studies) (Boardman et al. 2002;Kelly et al. 2001;Lin et al. 2007;Torche and Echevarría 2011;Yi et al. 2015). ...
... The remaining seven studies found fairly consistent heterogeneous effects of BW by parental SES. Among them, five found compensation (or smaller effects of BW) in high-SES families, and reinforcement (or larger effects of BW) in low-SES families(Boardman et al. 2002;Kelly et al. 2001;Lin et al. 2007; Torche and Echevarría 2011;Yi et al. 2015), while only the remaining two found reinforcement in high-SES and compensation in low-SES families (Cabrera-Hernández 2016;Figlio et al. 2014). ...
Full-text available
In post-industrial societies, a college education is the main channel for upper classes to prevent their children falling down the social ladder, while, for working classes, it is the best bet for upward mobility. Despite attaining post-compulsory education was equalised and a driver of social mobility in the last decades, inequalities by socioeconomic status (SES) in college graduation, the main social lift, remained relatively unchanged. We are only starting to understand the complex interplay between biological and environmental factors explaining why educational inequalities gestate before birth and persist over generations. Besides, further research is needed to unravel why advantaged students are more likely to get ahead in education than equally-skilled, but disadvantaged peers. This thesis bridges interdisciplinary literature to study how parental SES affects educational attainment during childhood in Germany, evaluating the implications for social justice. It contributes to the literature by (1) analysing the consequences of prenatal health shocks on skill formation; (2) examining the effect of cognitive and non-cognitive skills on the transition to secondary education; and (3) assessing SES-heterogeneity in these associations. Drawing from compensatory theories, I demonstrate how negative traits for educational attainment—low birth weight and cognitive ability—are less detrimental for high-SES children from the early stages of the status-attainment process due to mechanisms like parental investments and aspirations, and teachers’ bias in assessments. The German educational system enforces early tracking into academic or vocational pathways from age 10, supposedly according to ability. Thus, the case of Germany represents an institutional starting gate to evaluate equal opportunity, where compensating for negative traits might be difficult. To test compensatory theories, I utilise the Twin Life Study and the National Educational Panel Study applying quasi-causal empirical designs. The findings challenge the liberal conception of merit as the sum of ability plus effort in evaluating equal opportunity.
... If given a causal interpretation, health-related selection to education may be explained by disengagement from the school environment and peers as a result of increased absenteeism and stigmatization (Basch 2011;Hale and Viner 2018;Needham 2009); delayed cognitive development (Bhutta et al. 2002;Boardman et al. 2002); reduced educational expectations held by parents, teachers, and subsequently the children and adolescents themselves (McLeod and Fettes 2007;Roeser et al. 1998); or compromised future orientation due to the reduced expected utility of pursuing a long-term time-consuming investment, such as higher education (Becker and Mulligan 1997;Haas et al. 2011). Among factors potentially confounding the relationship, parental education and other aspects of socioeconomic position-along with gender-have been the most common to adjust for in health selection research (Hale et al. 2015;Melkevik et al. 2016). ...
Full-text available
This article reconsiders the role of social origin in health selection by examining whether parental education moderates the association between early health and educational attainment and whether health problems mediate the intergenerational transmission of education. We used longitudinal register data on Finns born in 1986-1991 (n = 352,899). We measured the completion of secondary and tertiary education until age 27 and used data on hospital care and medication reimbursements to assess chronic somatic conditions, frequent infections, and mental disorders at ages 10-16. We employed linear probability models to estimate the associations between different types of health problems and educational outcomes and to examine moderation by parental education, both overall in the population and comparing siblings with and without health problems. Finally, we performed a mediation analysis with g-computation to simulate whether a hypothetical eradication of health problems would weaken the association between parental and offspring education. All types of health problems reduced the likelihood of secondary education, but mental disorders were associated with the largest reductions. Among those with secondary education, there was further evidence of selection to tertiary education. High parental education buffered against the negative impact of mental disorders on completing secondary education but exacerbated it in the case of tertiary education. The simulated eradication of health problems slightly reduced disparities by parental education in secondary education (up to 10%) but increased disparities in tertiary education (up to 2%). Adolescent health problems and parental education are strong but chiefly independent predictors of educational attainment.
Full-text available
Medically assisted reproduction (MAR) plays an increasingly important role in the realization of fertility intentions in advanced societies, yet the evidence regarding MAR-conceived children’s longer-term well-being remains inconclusive. Using register data on all Finnish children born in 1995–2000, we compared a range of social and mental health outcomes among MAR- and naturally conceived adolescents in population-averaged estimates, and within families who have conceived both through MAR and naturally. In baseline models, MAR-conceived adolescents had better school performance and the likelihood of school dropout, not being in education or employment, and early home-leaving were lower than among naturally conceived adolescents. No major differences were found in mental health and high-risk health behaviours. Adjustment for family sociodemographic characteristics attenuated MAR adolescents’ advantage in social outcomes, while increasing the risk of mental disorders. The higher probability of mental disorders persisted when comparing MAR adolescents to their naturally conceived siblings. On average, MAR adolescents had similar or better outcomes than naturally conceived adolescents, largely due to their more advantaged family backgrounds, which underscores the importance of integrating a sociodemographic perspective in studies of MAR and its consequences.
Full-text available
A total of five hundred and ninety-one (591) school children of 6+ and 7+ years of age living in Mysore city were recruited as subjects using purposive sampling technique. Both schools and households visit were carried out to recruit the subjects. A self-developed questionnaire was administered to the parents to collect the data on socio-demographic conditions and birth history of children. The anthropometric measurements of children were recorded by using standard methods and techniques. The anthropometric data were further used to calculate indices such as BMI-forage , Height-forage and Weight-forage according to child growth reference of World Health Organization. The findings of the study indicates that the prevalence of malnutrition was distinctively higher among school children born with Low birthweight (LBW)(Stunting=81.3%, underweight= 82.6% and Undernutrition=54.0 %) than school children born with Normal Birthweight (Stunting= 37.6%, underweight=32.6% and Undernutrition=25.12 %). Within Low birthweight groups significant association of gender, age and education with level of stunting (Ht/Age), Age and education with level of underweight (Wt/age) were noticed. Within normal birthweight group, significant association of gender with level of underweight (wt/age) and undernutrition (BMI/age), age and education with level of stunting (Ht/age), education with level of undernutrition (Wt/age) were noticed. Overall, the present study concluded that birth weight determines the physical growth and nutritional status of children in late childhood years i.e. at school age period. The study suggests the implementation of intervention programs and stimulating programs to reduce the long-term consequences of infants born with low birthweight.
Policy Points Policies that increase county income levels, particularly for middle-income households, may reduce low birth weight rates and shrink disparities between Black and White infants. Given the role of aggregate maternal characteristics in predicting low birth weight rates, policies that increase human capital investments (e.g., funding for higher education, job training) could lead to higher income levels while improving population birth outcomes. The association between county income levels and racial disparities in low birth weight is independent of disparities in maternal risks, and thus a broad set of policies aimed at increasing income levels (e.g., income supplements, labor protections) may be warranted. Context: Low birth weight (LBW; <2,500 grams) and infant mortality rates vary among place and racial group in the United States, with economic resources being a likely fundamental contributor to these disparities. The goals of this study were to examine time-varying county median income as a predictor of LBW rates and Black-White LBW disparities and to test county prevalence and racial disparities in maternal sociodemographic and health risk factors as mediators. Methods: Using national birth records for 1992-2014 from the National Center for Health Statistics, a total of approximately 27.4 million singleton births to non-Hispanic Black and White mothers were included. Data were aggregated in three-year county-period observations for 868 US counties meeting eligibility requirements (n = 3,723 observations). Sociodemographic factors included rates of low maternal education, nonmarital childbearing, teenage pregnancy, and advanced-age pregnancy; and health factors included rates of smoking during pregnancy and inadequate prenatal care. Among other covariates, linear models included county and period fixed effects and unemployment, poverty, and income inequality. Findings: An increase of $10,000 in county median income was associated with 0.34 fewer LBW cases per 100 live births and smaller Black-White LBW disparities of 0.58 per 100 births. Time-varying county rates of maternal sociodemographic and health risks mediated the association between median income and LBW, accounting for 65% and 25% of this estimate, respectively, but racial disparities in risk factors did not mediate the income association with Black-White LBW disparities. Similarly, county median income was associated with very low birth weight rates and related Black-White disparities. Conclusions: Efforts to increase income levels-for example, through investing in human capital, enacting labor union protections, or attracting well-paying employment-have broad potential to influence population reproductive health. Higher income levels may reduce LBW rates and lead to more equitable outcomes between Black and White mothers.
Objective: Some previous studies have reported the improved survival of very-low-birth-weight (VLBW) neonates with no disabilities. However, 16% of these neonates have developmental disorders. Considering the lack of research on the developmental status of five-year-old VLBW children and the importance of early detection and treatment, in this study, we aimed to assess the developmental status of five-year-old VLBW children. Materials & methods: This historical cohort study was conducted on five-year-old children. The participants were divided into VLBW and normal-birth-weight (NBW) groups. Data were gathered using the Ages and Stages Questionnaire (ASQ). This questionnaire consisted of five developmental domains, including communication, gross motor, fine motor, problem-solving, and personal/social skills. Data were reported by measuring descriptive statistics, including mean, standard deviation, number, and percentage, and analyzed by Mann-Whitney U test and independent t-test in SPSS version 22. Results: A total of 106 five-year-old children, including two groups of VLBW and NBW, participated in this study. The results of Mann-Whitney U test showed a significant difference between the groups regarding the scores of communication (P=0.002), gross motor (P<0.001), fine motor (P<0.001), and problem-solving (P<0.001) skills. However, no significant difference was found between the groups regarding the personal/social developmental status (P=0.559). Conclusion: According to the results, a higher risk of developmental delay was observed in VLBW infants as compared to NBW neonates; therefore, it is recommended to perform developmental screening tests for timely detection of high-risk children and early diagnostic and therapeutic interventions.
Early health problems predict lower educational attainment, but it remains unclear whether this is due to health problems weakening school performance or due to other mechanisms operating above and beyond school performance. We employed counterfactual-based mediation analysis on a register-based sample of Finnish adolescents born in 1988–1993 (n = 73,072) to longitudinally assess the direct (unexplained by school performance, as measured by grade point average) and indirect (pure mediation and mediated interaction via school performance) effects of early adolescent somatic and mental health problems on the noncompletion of upper secondary education and track choice (vocational vs. general). Mental disorders were associated with the largest increases in both noncompletion and choosing the vocational track, but somatic conditions also showed small but robust associations. Weakened school performance mediated up to one-third of the differences in noncompletion and around half of the differences in track choice. When the same analyses were conducted within sibships, the total effects of health problems on educational pathways were weaker, but the contribution of school performance remained similar. In counterfactual simulations that assigned everyone an above-median school performance—that is, eradicating below-median school performance—about 20–40 percent of the effects of mental disorders on educational pathways remained. Our results suggest that while impaired school performance is an important component in health-related selection to education, it does not fully explain the shorter and less academically oriented educational careers of adolescents with health problems. These adolescents may benefit from additional educational support regardless of their formal school performance.
Full-text available
Research has uncovered many mechanisms that exacerbate racial inequalities in achievement. Due to specialization within the field, however, little focus has been devoted to the multitiered and often interconnected institutional nature of these processes. Matching data from the restricted-use National Educational Longitudinal Survey and the Common Core of Data, I hierarchically model the influence of family/peer and educational institutional processes simultaneously on the black-white gap in achievement. The modeling strategy used offers a more comprehensive understanding of the reproductive interinstitutional dynamics at work, suggesting strong linkages between family/peer group attributes and access to educational resources. I conclude by suggesting the need to extend this line of inquiry a step further still, developing a theoretically driven contextual and spatial understanding of educational opportunity and achievement.
Full-text available
Two key questions are addressed regarding the intersection of socioeconomic status, biology, and low birth weight over the life course. First, do the income and other socioeconomic conditions of a mother during her pregnancy affect her chances of having a low-birth-weight infant net of her own birth weight, that of the father, and other family-related, unobserved factors? Second, does an individual's birth weight status affect his or her adult life chances net of socioeconomic status? These questions have implications for the way we conceive of the relationship between socioeconomic status and health over the life course, specifically in sorting out causal directionality. We use intergenerational data from the Panel Study of Income Dynamics, for the years 1968 through 1992. Results of sibling comparisons (family-fixed-effects models) demonstrate that maternal income does not appear to have a significant impact on birth weight. However, low birth weight results in lower educational attainment net of other factors. These findings suggest that, when considered across generations, causality may not be as straightforward as implied by cross-sectional or unigenerational longitudinal studies.
This report is the second in a series that follows 1992-93 college graduates' progress through the teacher pipeline using data from the Second Follow-up of the Baccalaureate and Beyond Longitudinal Study. It examines academic characteristics and preparation for teaching of those who took various steps toward teaching and is organized by a conceptual teacher pipeline that represents a teacher's career. The pipeline includes preparatory activities and teaching experiences. The report examines the rate at which graduates with varying demographic and academic characteristics entered the teacher pipeline and describes the steps that pipeline entrants took toward teaching, noting experiences of those who taught. It also discusses the rate at which those who had taught since completing their degree had stopped teaching and all pipeline members' expectations for teaching in the future. Many graduates who taught soon after college did not expect to spend much time teaching or to make it a career. White, non-Hispanics were the most inclined to teach. Asian/Pacific Islanders were the least inclined to teach. Commitment to teaching related to college entrance examination scores and undergraduate grade point averages. Patterns in teaching behavior have continued from participants' first year out of college to their fourth. (Contains 47 references.) (SM)
Context Although studies have documented cognitive impairment in children who were born small for gestational age (SGA), other studies have not demonstrated differences in IQ or other cognitive scores. The need exists for long-term studies of such children to assess functional outcomes not measurable with standardized testing. Objective To determine the long-term functional outcome of SGA infants. Design Prospective cohort study. Setting and Participants A total of 14,189 full-term infants born in the United Kingdom on April 5 through 11, 1970, were studied as part of the 1970 British Birth Cohort; 1064 were SGA (birth weight less than the fifth percentile for age at term). Follow-up at 5, 10, 16, and 26 years was 93%, 80%, 72%, and 53%, respectively. Main Outcome Measures School performance and achievement, assessed at 5, 10, and 16 years; and years of education, occupational status, income, marital status, life satisfaction, disability, and height, assessed at 26 years, comparing persons born SGA with those who were not. Results At 5, 10, and 16 years of age, those born SGA demonstrated small but significant deficits in academic achievement. In addition, teachers were less likely to rate those born SGA in the top 15th percentile of the class at 16 years (13% vs 20%; P<.01) and more likely to recommend special education (4.9% vs 2.3%; P<.01) compared with those born at normal birth weight (NBW). At age 26 years, adults who were SGA did not demonstrate any differences in years of education, employment, hours of work per week, marital status, or satisfaction with life. However, adults who were SGA were less likely to have professional or managerial jobs (8.7% vs 16.4%; P<.01) and reported significantly lower levels of weekly income (mean [SD], 185 [91] vs 206 [102] £; P<.01) than adults who were NBW. Adults who were SGA also reported significant height deficits compared with those who were NBW (mean [SD] z score, −0.55 [0.98] vs 0.08 [1.02]; P<.001). Similar results were also obtained after adjusting for social class, sex, region of birth, and the presence of fetal or neonatal distress. Conclusions In this cohort, adults who were born SGA had significant differences in academic achievement and professional attainment compared with adults who were NBW. However, there were no long-term social or emotional consequences of being SGA: these adults were as likely to be employed, married, and satisfied with life.
Is childhood such a critical period that, by the end of this period, cumulative poverty would have exerted maximum effect on children's cognitive outcomes? Or are cognitive outcomes more a function of the length of exposure to poverty regardless of the life stage in which the child is exposed to poverty? The NLSY, which measures each child's cognitive development repeatedly over time, was analyzed to answer these questions. We distinguish between ability and achievement. Ability is a more stable trait than achievement and tends to be determined by both environmental and genetic factors early in life. Achievement on the other hand is more acquired. This study shows that long-term poverty has substantial influences on both ability and achievement, but the time patterns of these influences are distinctly different. Childhood appears to be a much more crucial period for the development of cognitive ability than early adolescence. In contrast, poverty experienced in adolescence appears to be more influential to adolescent achievement than poverty experienced earlier in life.
This report explores why some schools may be better than others at helping students learn. It responds to the congressionally mandated Special Study Panel on Education Indicators, which asked the National Center for Education Statistics to examine indicators of the health of the nation's educational system. The report reviews the literature on school quality and is intended to help policy makers and researchers understand those characteristics that are most likely related to student learning. It identifies the availability and reliability of national indicators and assesses the current status of schools by examining and critiquing these national indicator data. The report claims that school quality, as it affects student learning, is demonstrated by the training and talent of the teaching force, what goes on in the classrooms, and the overall culture and atmosphere of the school. Within these three areas, the document identifies 13 indicators of school quality that recent research suggests are related to student learning. Findings indicate that students learn more from teachers with high academic skills and who teach subjects related to their undergraduate or graduate studies than they do from teachers with low academic skills and who teach subjects unrelated to their training. (Contains 146 references.) (RJM)