Digital Divide and Social Well-Being in Rapidly Digitizing Societies: A Heckman Correction Analysis of Older Adults

This study investigates the dynamics of internet usage among adults aged 50–74 in Turkey and analyzes the subsequent impact of such usage on individual well-being—specifically social participation—within a rapidly digitizing global landscape. The research utilizes a comprehensive, nationally representative microdata set (Weighted sample size of N=20,207,653) derived from the Turkish Statistical Institute's (TÜİK) 2025 “Household Information Technology Usage Survey.” The primary motivation of this study is to examine the deepening phenomenon of the “gray digital divide” through a multidimensional lens that extends beyond mere physical access to encompass skills, competence, and trust. A prevalent methodological limitation in existing literature is the restriction of analyses to active internet users, which ignores “sample selection bias.” To overcome this constraint and provide statistically unbiased estimates, this study employs the Heckman Two-Step Correction model. The first stage models the decision to use the internet (selection), while the second stage examines the effect of this usage on social participation (outcome). “Internet Access Availability in the Household” is utilized as the exclusion restriction to ensure model identification. The analysis yields striking findings across two axes. First-stage results indicate that education level, household income, age, and settlement type are potent determinants of internet usage. Notably, older adults with low education levels are approximately 90% less likely to engage with the internet compared to their highly educated peers. Furthermore, residing in rural areas and being female are identified as significant demographic disadvantages regarding digital access. The most critical and novel contributions emerge in the second stage, where selection bias is statistically corrected. The high statistical significance (p<0.001) of the Inverse Mills Ratio (Lambda) confirms the necessity of the applied correction method. The adjusted findings reveal that the lack of digital skills exacerbates the risk of social exclusion to a magnitude far greater than standard models predict. Specifically, older adults lacking technical skills (e.g., file transfer, application installation) face an approximately 14-fold higher risk of exclusion from social networks compared to those possessing these skills. Similarly, a lack of content verification competence increases the exclusion risk by 10.5 times. Additionally, the absence of cybersecurity awareness is found to deter engagement with digital platforms, resulting in a loss of social capital. In conclusion, this research demonstrates that the primary barrier to digital integration for older adults in Turkey is no longer solely “physical access” (first-level divide) but rather “skills, competence, and trust” (second-level divide). In light of these findings, policymakers are advised to shift focus from mere device distribution to holistic, applied training programs that enhance cybersecurity awareness, technical skills, and information literacy. Moreover, the establishment of local digital support units for rural elderly populations and the design of age-friendly digital interfaces are vital for enhancing the well-being of this demographic.