Using mobile phone survey paradata for process evaluations and improvements: best practices and lessons learned from nine surveys in seven sites.

The use of mobile phone surveys (MPS) for regionally or nationally representative data allows for quick, efficient and affordable data collection for monitoring trends and generating results to guide action. By digitizing this process, data flows can be expanded to include metadata and paradata that allow survey administrators to evaluate and improve survey processes and parameters. Between 2017 and early 2022, the Centers for Disease Control and Prevention provided technical support to country partners to implement MPS gathering indicators on noncommunicable diseases within adult populations in seven countries. These surveys resulted in 37 591 completed interviews containing no personal identifiable information. When combined, these surveys result in over 25 million rows of paradata representing timestamped interactions between the data collection platform and each survey respondent. Using exploratory data analysis, five key metrics were identified which had implications on MPS process optimization: timing of engagement, question randomization, contacts to complete, errors and mode effect. The use of survey paradata allows for real-time process evaluations and identifies factors that can improve efficiency and effectiveness of MPS methods.