Extended Data Table 2 Understanding potential bias from unobserved indoor air pollution exposure

From: Robust relationship between air quality and infant mortality in Africa

  1. a, Potential bias at the household level in the estimated effect of PM2.5 exposure on infant health \(\hat{\alpha }\) as a function of the relationship between indoor and outdoor pollution exposure. Cells show the expected relative magnitudes of time-averaged differences in PMin versus PMout for all combinations of δ0 and δ. b, Relationship between ambient PM2.5 and fraction of households using clean cooking fuels at the DHS cluster level. Standard errors are in parentheses, and asterisk denote statistical significance (two-sided): *P < 0.10, **P < 0.05, ***P < 0.01. RMSE, root mean squared error.
Back to article page