2020 Data from PMA Ethiopia
Kathryn Grace, Matt Gunther, Jiao Yu - University of Minnesota
Women report all pregnancy outcomes within 3 years
Must have resided near sample cluster (•) for 2 years (centroid displaced up to 10 km)
Beyond 2 years, women report how long they have lived in their current region
Addressing challenges to spatial and temporal aggregation
We use a daily maximum temperature record, CHIRTS Tmax (0.05 arc-degree resolution, 1980-2020)
It’s especially important to weight regional means by population density!
We create indicators at three heat thresholds for each day, 1980-2020.
We identify a heatwave whenever a threshold is passed for at least 3 consecutive days
One for each of 6 heat measures. Lifetime exposure to 30° C Days looks like this:
Analysis of the interaction between lifetime heat exposure (tertiles) and pregnancy heat exposure
Binary logistic regression. Controls include: mother’s age, partnership status, educational background, household wealth, religion, pregnancy intentions at conception, and clustered fixed effects for sample strata (region + urban).