Does Segregation Lead to Lower Birth Weight?: An Instrumental Variable Approach

Area of Science:

• Public Health
• Sociology
• Epidemiology

Background:

• Racial residential segregation is associated with disparities in birth outcomes.
• Previous studies may have obscured the causal impact of segregation on birth weight due to confounding and mediation.
• Understanding the direct effect of segregation on birth weight is crucial for public health interventions.

Purpose of the Study:

• To investigate the causal relationship between racial residential segregation and birth weight in non-Hispanic black and white infants.
• To differentiate the effects of segregation from other covariates using advanced statistical methods.
• To provide a more accurate estimation of segregation's impact on birth outcomes.

Main Methods:

• Utilized singleton birth data for non-Hispanic black and white mothers in the US (2000).
• Linked birth records to metropolitan statistical area (MSA)-level segregation indices (dissimilarity index).
• Employed instrumental variable (IV) analysis, specifically two-stage least squares (2SLS) with a railroad division index, to estimate causal effects, comparing with ordinary least squares (OLS).

Main Results:

• OLS models estimated a 1.2 g decrease in black birth weight per 1% increase in segregation.
• IV (2SLS) models suggested a larger decrease of 2.8 g in black birth weight, though with wider confidence intervals.
• Estimates for white infants showed minimal and inconsistent effects of segregation across OLS and IV models.

Conclusions:

• OLS estimates may underestimate the adverse effect of segregation on black birth weight.
• Instrumental variable models provide evidence consistent with a causal impact of segregation on black birth outcomes.
• Further research is needed to address the imprecision and potential weak instrument bias in IV estimates.