A comment on "Raising Health Awareness in Rural Communities: A Randomized Experiment in Bangladesh and India" by Siddique et al. (2024)

Siddique et al. (2024a) report massive effects of a mobile phone-based health awareness campaign in a randomized field experiment conducted in rural Bangladesh and India during the COVID-19 pandemic. Both awareness and compliance with preventive COVID-19 measures were higher when the information was received by voice call rather than text, and even higher for those receiving both. Reproducing the analyses we identify many severe issues, including that the study did not in fact randomize treatment assignment. We further find implausible response patterns in the data, undisclosed sampling criteria that negate the study motivation, and an (unreported) retreatment where some of the respondents were also included in a separate study that provided additional COVID-19 information immediately before the last data collection.