Racial and Ethnic Differences in Birthweight: The Role of Income and Financial Assistance

Citation: Cramer (1995) Racial and Ethnic Differences in Birthweight: The Role of Income and Financial Assistance.
Internet Archive Scholar (search for fulltext): Racial and Ethnic Differences in Birthweight: The Role of Income and Financial Assistance
Tagged: uw-madison (RSS), wisconsin (RSS), sociology (RSS), demography (RSS), prelim (RSS), qual (RSS), WisconsinDemographyPrelimAugust2009 (RSS)

Summary

Introduction

"The purpose of this paper is to explore possible explanations of racial and ethnic differences in birthweight, with special attention to income and poverty. Birthweight is examined because it is linked so closely to infant mortality but also predicts other adverse outcomes such as childhood morbidity and developmental problems . . ." (p. 232).

Methods

Methods: Data are from the Youth Panel of the National Longitudinal Surveys of Labor Force Experience (NLSY), which began in 1979. The NLSY contains a large national sample of men and women aged 14-21, with oversampling of blacks, Hispanics and low-income youths. In 1986, information on health and nutrition was collected for each child of the mothers in the sample. Birthweight was also recorded retrospectively at this time, and was measured in grams. Ethnic identification was obtained by self-report, and the three largest groups available for comparison are non-Hispanic whites, blacks, and Chicanos (i.e., Mexican, Central American or South American descent). Several demographic and socioeconomic variables are used-including maternal age, marital status, birth order, birth interval, mother's education, family income and family public assistance. Additionally, psychosocial characteristics-namely self esteem and an armed forces aptitude test-were included as "admittedly weak and speculative proxy variables" for ability to provide adequate prenatal care (p. 326). Finally, measures from the NLSY that are thought to be proximate determinants (smoking and drug use, the timing of initial prenatal medical care, stress, weight-to-height ratio and height) for birthweight were included. Using OLS regression, Cramer specifies "these variables in a very simple causal model of birthweight in which the socioeconomic, demographic, and prenatal parenting variables are treated as exogenous, and the proximate determinants are specified as intervening variables" (p. 239).

Results

Among white Anglo mothers, income effects on birthweight are strong without controls, but when self-esteem and AQFT (the armed forces test) are introduced, income variables are only "moderately and sometimes significantly associated with birthweight, such that birthweight is greater at higher levels of income" (p. 239). In this model, self-esteem is not significant, but AQFT has a strong, positive association with birthweight. Although some of the proximate determinants (e.g., smoking, weight, and stress) have significant effects on birthweight, others do not (e.g., drug use), perhaps due to poor measurement. Importantly, these proximate determinants have "little impact on the estimated effects of the exogenous social variables; thus the model does not explain why the exogenous variables are related to birthweight" (p. 239). To test whether these variables explain ethnic differences in birthweight, equality constraints between regression equations were imposed via the Chow test. The tests indicated that the models are the same for whites, blacks and Chicanos, so the groups are comparable. As expected, Chicanos and especially blacks have lower mean birthweights than whites. In the "social model" (i.e., only exogenous variables; no proximate determinants), the gaps are significantly reduced between ethnic groups (particularly between blacks and whites), but not eliminated. Curiously, the "medical model" (all variables) indicates that the proximate determinants actually widen the ethnic differences, perhaps because blacks and Chicanos are less prone than whites to heavy smoking and excessive thinness.

Conclusion

Supportive of the hypothesis that income accounts for some ethnic differences, yet important gaps remain. Improved measures of income and incorporation of measures of discrimination and racism may help explain the ethnic differences in birthweight more fully.