ABSTRACT:
Rapid economic growth, lifestyle transitions, and population ageing have intensified the burden of non-communicable diseases (NCDs) in India, with unhealthy diets, tobacco use, alcohol consumption and physical inactivity emerging as major risk factors. In this context, internet use has become an increasingly important determinant of lifestyle behaviour, with the potential to both promote healthier choices through improved access to information and encourage unhealthy practices by displacing physical activity or reinforcing risky behaviours. Using nationally representative data from three rounds of the National Family Health Survey (2005–06, 2015–16, and 2019–21), this study examines socio-economic and geographical variations in lifestyle behaviours among Indian adults and estimates the causal relationship between internet use and dietary and unhealthy consumption outcomes. To address endogeneity arising from unobserved individual traits influencing internet adoption, district-level cell tower density from OpenCellID is employed as an instrumental variable. The analysis applies an Endogenous Treatment Poisson Regression model to estimate the impact of internet use on dietary diversity and a Recursive Bivariate Probit model to assess its association with smoking and alcohol consumption. The findings reveal pronounced socio-economic gradients, with higher education and wealth significantly increasing internet use while reducing the likelihood of smoking and alcohol consumption, alongside substantial gender, rural–urban, and regional disparities. Internet use is positively associated with dietary diversity but also increases the probability of smoking and alcohol consumption. Overall, the results highlight the dualistic influence of internet use on health behaviours, underscoring the need for nuanced public health policies that harness digital technologies for health promotion while mitigating their potential to reinforce unhealthy lifestyles.