Interactive Voice Response Surveys (Ivrs) as a Method for Increasing Representativeness of Rural Respondents in Mobile Phone Surveys: Findings from Malawi.

Malebogo Tlhajoane, London School of Hygiene and Tropical Medicine
Funny Muthema, Institute of Public Opinion and Research
Michael Chasukwa, Institute of Public Opinion and Research
Kelly McCain, London School of Hygiene and Tropical Medicine
Shammi Luhar
Julio E. Romero-Prieto, London School of Hygiene and Tropical Medicine (LSHTM)
Jacob Saikolo, Institute of Public Opinion and Research
Cremildo Manhica, National Institute of Health (INS)
Sarah Walters, .
Georges Reniers, London School of Hygiene and Tropical Medicine (LSHTM)
Boniface Dulani, Institute of Public Opinion and Research

Background: In mobile phone surveys (MPS) reliant on random digit dialling, it may be challenging to compose a representative sample, particularly in settings where mobile phone ownership is low and unequally distributed. Methods: We evaluated the use of fully automated interactive voice response (IVR) surveys among 25,524 unique mobile phone numbers as a screening tool for identifying rural respondents, particularly women, in a mortality MPS in Malawi. Results: Where information on sex was provided, 27.6% of respondents were found to be female. Overall, 33.6% of those who engaged with the survey lived in a rural area, 49.9% of whom went on to complete the mortality MPS at a cost of US$8.91 per IVR number identified, and US$17.9 per mortality survey completed. Conclusions: IVR surveys can be used to improve the representation of rural and female respondents in MPSs. Modifications to the IVR process could be explored further to increase engagement.