The effects of covariates on the instantaneous hazard of mortality§,*.

§Number of death events = 10,963; number of records = 648,237; sample consists of individuals born under observation and surviving to at least 6 months of age.*p-value ≤0.05,**≤0.01,***≤0.001. n is reduced compared to total sample due to missingness.aBeta is the estimated coefficient of the relationship between a given independent variable (e.g., gender) and the outcome of interest (here, the log hazard of mortality); i.e., a one-unit change in the independent variable is associated with a Beta increase in the log hazard of dying at any time.bSE is standard error of the estimated Beta.cReference category is female.dReference category for head of household’s occupation is agriculture.eReference category is not-uxorilocally married.fReference category is legitimate.gReference category is low prevalence of minor marriage; based on S1 Table; see Supplementary Methods for details.hReference category is birth cohort 1; see S1 Table.iCoefficient = -8.66 x 10–3.The effects of covariates on the instantaneous hazard of mortality§,*.