Purpose: There is support for the role of Internet use in promoting well-being among older people. However, there are also contradictory findings which may be attributed to methodological issues. First, research has focused on frequency of online activity rather than how engagement in different types of online activities may influence well-being. Secondly, previous studies have used either cross-sectional designs, which cannot elucidate causality or intervention designs with uncontrolled extraneous variables. In this longitudinal observational study, we test the indirect impact of online engagement for social, informational, and instrumental purposes on older adults' well-being via reducing loneliness and supporting social engagement.

Design and method: A population sample of 1,165 adults aged 60-77 (M = 68.22, SD = 4.42; 52.4% female) was surveyed over 3 waves. Using longitudinal mediation analysis with demographic controls, the indirect effects of types of Internet use on well-being through loneliness and social engagement were estimated.

Results: Participants engaged online for 3 purposes: social (e.g., connecting with friends/family), instrumental (e.g., banking), and informational (e.g., reading health-related information). Social use indirectly impacted well-being via decreased loneliness and increased social engagement. Informational and instrumental uses indirectly impacted well-being through engagement in a wider range of activities; however, were unrelated to loneliness.

Implications: Findings highlight that Internet use can support older adults' well-being; however, not every form of engagement impacts well-being the same way. These findings will inform the focus of interventions which aim to promote well-being.