Specificity and sensitivity of SSH data, by study group.

Concordant SSH data are those in which a respondent reported a sibship composition that matches the sibship composition in the HDSS dataset (e.g., if an adult female death was recorded by the HDSS among the sibship, the respondent also reported such a death during the SSH data collection). “Total” is the total number of respondents used for each calculation.aSensitivity is defined as the proportion of concordant SSHs among respondents from sibships with at least one adult female/male death recorded by the HDSS.bSpecificity refers to the proportion of concordant SSHs among respondents from sibships without any adult female/male deaths recorded by the HDSS.cAdjusted odds ratios (aORs) measure changes in the odds of a concordant answer associated with the use of the SSC; they are obtained from adjusted analyses using logistic regressions with stratifying variables as covariates.dThe values of sensitivity and specificity are predicted probabilities obtained from these logistic regressions; p-values are also obtained from these logistic regressions. Standard errors are adjusted for the clustering of respondents within sibships.eΔ is the equivalence margin such that the absolute value of the difference in sensitivity/specificity between the DHS and SSC groups is |DHS−SSC|[56]. We arrived at the Δ recorded above through an iterative process: it is the smallest value of Δ for which both one-sided tests of the null hypothesis of no equivalence are rejected at the pp