The Effect of Social Ties on Digitalization in Older Adults: Integrating Propensity Score Matching and Finite Mixture Models on Two Combined Data Sets

Abstract The age digital divide is the gap between older adults and younger individuals in accessing or using different digital technologies. Results from the literature show that social ties with family and friends may have a great influence on the digital inclusion of older adults. To understand the causal link between intergenerational ties and the reduction of age digital divide, we propose to combine data from round 10 of the European Social Survey (ESS) and wave 8 of the Survey of Health, Ageing and Retirement in Europe (SHARE); the former contains information on the digitalization level of older adults, while the latter provides details on the individuals’ network of family contacts and friendships. We propose to reduce imbalance of confounders between treatment groups through propensity score matching techniques. A finite mixture model is then applied to the matched data set. Here, social connectivity indicators measuring the social tie with different members of respondents’ social network are added as predictors of latent class membership probabilities to inform about the causal effect of interest under strong ignorability. Using these approaches, we are able to cluster individuals according to their digitalization level, as well as to measure the effect of social ties on the probability of belonging to the most or least digitalized cluster.