I study population health in data-scarce settings. One strand of my work develops childhood sex ratios as a proxy for early-life mortality. Because boys die at higher rates in infancy, the sex composition of surviving children can be used as an indicator for stillbirths and infant mortality. I apply this indicator to the historical US, where we lack vital statistics but have excellent census data.
Another strand of my work focuses on maternal-infant health in contemporary sub-Saharan Africa. Using Demographic and Health Survey (DHS) data, I explore how food and water insecurity contribute to child health outcomes.
We use childhood sex ratios, which skew female when infant mortality is high, to estimate 19th-century US infant mortality. For US whites we find roughly 70–80 deaths per 1,000 circa 1850.
I examine how the 2008 food price crisis affected child hemoglobin in Senegal, using a natural-experiment comparison of in-utero and breastfeeding children. I find large negative effects in urban areas, especially Dakar, and none in rural areas.
The repatriation campaign of the 1930s removed a large share of the Mexican-origin population of the United States, including many US-born children. We use childhood sex ratios to estimate the effect of this campaign on early-life mortality among Mexican immigrants in Texas. We find that the campaign caused a major shift toward females among the children of the affected population, implying a large increase in early-life mortality.
Black Health in Slavery and Freedom: New Evidence from Childhood Sex Ratios. With James R. Irwin. We estimate infant mortality for the US Black populations, 1850–1880, using childhood sex ratios.
Comparing the JMP Drinking Water Ladder and the HWISE Scale in the 2022–23 Mozambique DHS. With Indira Bose, Joshua Miller and Sera Young. We compare an infrastructure-based and an experience-based measure of household water access in a single survey. Presented as a poster at the 2026 UNC Water and Health Conference
The Industrialization Puzzle in the US: New Evidence from Childhood Sex Ratios. I use childhood sex ratios to show that US population health deteriorated across the period 1850–1910, not just before the Civil War, recasting the “Antebellum Puzzle” as an “Industrialization Puzzle.”
Sex Ratio at Birth and the HIV Epidemic in South Africa. With Jenny Trinitapoli. We find that HIV prevalence is associated with female-skewed sex ratios at birth in South Africa, and could account for roughly half the gap from a biologically normal ratio. This relationship weakened with the expansion of antiretroviral drugs.