Who do phone surveys miss, and how to reduce exclusion: recommendations from phone surveys in nine Indian states

Computer-assisted telephone interviews (CATI) through mobile phones are a low-cost, rapid and safe way to collect data. However, decisions for how such mobile phone surveys are designed and implemented, and their data analysed, can have implications for the sample reached, and in turn affect the generalisability of sample estimates. In this practice paper, we propose a framework for extending the use of CATI-mobile phone surveys in India, which can be applied broadly to future surveys conducted using this method. Across the stages of design, implementation and analysis, we outline challenges in ensuring that the data collected through such surveys are representative and provide recommendations for reducing non-coverage and non-response errors, thereby enabling practitioners in India to use CATI-mobile phone surveys to estimate population statistics with lower bias. We support our analysis by drawing on primary data that we collected in five mobile phone surveys across nine Indian states in 2020. Our recommendations can help practitioners in India improve the representativeness of data collected through mobile phone surveys and generate more accurate estimates.