Keywords: Methodological challenges and improvements, including in the areas of sampling, measurement, survey design and survey response or non-response

Data collection in complex environments presents multiple challenges, the central of which is data quality and trust in data collection process. Over the last 3 years, we have developed, tested, and refined an audio check method that unobtrusively recorded surveyors reading aloud questions to participants. In this presentation, we share evidence that this method can be used to detect data fabrication more efficiently and effectively than traditional methods of data quality control. This research is based on comparing failing and passing interviews based on audio checks in 10 African countries (Guinea, Liberia, Sierra Leone, Mali, Senegal, Nigeria, Burkina Faso, Niger, Mauritania, and Cameroon), surveyed in 2017. In the paper, we compare multiple survey attributes between cases with failing and passing audio, including the average length of interviews, the rates of non-response, as well as sample representativeness, and investigate the possible bias of undetected data fabrication. The paper also addresses the cost and time implication of audio checks.