Unmasking the truth: Experimental evidence of facemask compliance in Bangladesh, Kenya, and Nigeria during the COVID-19 pandemic.

High levels of compliance with public health measures are critical to ensure a successful response to the COVID-19 pandemic and other public health emergencies. However, most data on compliance are self-reported and the tendency to overreport due to social desirability could yield biased estimates of actual compliance. A list experiment is a widely used method to estimate social desirability bias in self-reported estimates of sensitive behaviours. We estimate rates of compliance with facemask mandates in Kenya, Nigeria, and Bangladesh using data from phone surveys conducted in March-April 2021. Data on compliance were collected from two different survey modules: a self-reported compliance module (stated) and a list experiment (elicited). We find large gaps between stated and elicited rates of facemask wearing for different groups depending on specific country contexts and high levels of overreporting of facemask compliance in self-reported surveys: there was an almost 40 percentage point gap in Kenya, 30 percentage points in Nigeria, and 20 percentage points in Bangladesh. We also observe differences in rates of self-reported facemask wearing among key groups but not using the elicited responses from the list experiment, which suggest that social desirability bias may vary by demographics. Data collected from self-reported surveys may not be reliable to monitor ongoing compliance with public health measures. Moreover, elicited compliance rates indicate levels of mask wearing are likely much lower than those estimated using self-reported data.