Age-Disaggregated Subnational Patterns of Internet and Mobile Phone Adoption

Michael Y.C. Chong , University of Oxford
Ridhi Kashyap, Oxford University

Digital technologies are transforming population and development processes, yet access remains uneven across geography, gender, and age -- particularly in low- and middle-income countries (LMICs). While younger individuals are typically more connected, most existing evidence on these age patterns comes from high-income countries, leaving uncertain whether similar dynamics hold in LMICs. Whether digital gender gaps persist among younger cohorts in LMICs is also unclear. Reliable subnational, age-disaggregated data on who is connected remain scarce, limiting understanding of demographic digital divides and progress toward the Sustainable Development Goals. This paper introduces a two-stage approach to estimate subnational age- and gender-specific patterns of internet and mobile adoption across LMICs. We first smooth sparse survey data from the Demographic and Health Surveys (DHS) and Multiple Indicators Cluster Surveys (MICS) across ages to generate age curves of digital adoption for subnational areas. We then apply a machine learning model integrating social media, geospatial, and development covariates to predict adoption where survey data are unavailable. The results will provide new global, age-structured subnational estimates, advancing digital demography by revealing demographic dimensions of digital inequality.

See paper

 Presented in Session 26. Flash Session Emerging Data Sources in Demography: Digital Traces, AI and Mobile Phone Data