Cohort Differences in Internet Use Among Older Adults: Evidence from the English Longitudinal Study of Ageing (ELSA)

Paper Title

Cohort Differences in Internet Use Among Older Adults: Evidence from the English Longitudinal Study of Ageing (ELSA)

Publication Info

• Topic area: Age-based digital divide in internet use among older adults.
• Keywords: Ageing, internet use, digital divide, older adults, cognitive ability, education, employment, longitudinal study, HCI.

Background and Problem

• Problem / challenge: Internet use decreases with age, but the specific factors driving this decline are not fully understood. Existing research often assumes functional decline as the primary cause without sufficient empirical evidence.
• Significance: Understanding the factors influencing internet use among older adults is crucial for addressing digital exclusion and ensuring equitable access to technology benefits.
• Motivation and related work: Previous studies have explored physiological, motivational, and structural barriers to internet use among older adults but have not adequately differentiated between age cohorts or examined the relative importance of cognitive, educational, and employment factors.

Solution

• Proposed approach: Analysis of Wave 10 data from the English Longitudinal Study of Ageing (ELSA) to identify patterns and factors associated with internet use across three age cohorts: 50–64 years, 65–79 years, and 80+ years.
• Novelty:
  1. Empirical testing of factors hypothesized to drive age-related differences in internet use.
  2. Identification of cognitive ability, education, and employment as key factors influencing the age gradient in internet use.
  3. Distinction between reasons for non-use among rare and regular internet users.
  4. Critique of the chronological definition of "older adults" in HCI, proposing a shift to 80+ as a more meaningful threshold.
• Procedure and key techniques:
  • Logistic regression analysis to assess the impact of demographic, health, and social factors on internet use.
  • Categorization of internet use into "regular" (daily/monthly) and "rare" (seldom/never) use.
  • Examination of self-reported reasons for not using the internet more, adjusted for sex and education.

Results

• Concrete findings:
  • Regular internet use decreases with age: 97.7% (50–64 years), 91.1% (65–79 years), 65.7% (80+ years).
  • Cognitive ability, education, and employment status account for 16.3% of the age gradient in internet use.
  • Among regular users, older cohorts cite lack of IT skills (43.1% for 80+), lack of trust (24.9%), and no reason to use the internet more (37.7%) as barriers.
  • Physical impairments, social isolation, and depression have minimal impact on the age gradient.
• Advantage over baselines: Challenges the assumption that functional decline is the primary driver of non-use, highlighting cognitive, educational, and employment factors instead.
• Experiments / evaluation: Analysis of data from 6,373 participants in ELSA, representing older adults in England. Logistic regression models tested multiple sociodemographic, health, and motivational factors.
• Limitations and future work:
  • Correlational findings due to single-wave analysis; longitudinal studies needed for causal insights.
  • Exclusion of older adults in institutional settings.
  • Subjectivity in self-reported internet use frequency and reasons for non-use.
  • Future research should explore life histories, cohort effects, and the impact of AI technologies on older adults.

Summary

This study analyzed data from the English Longitudinal Study of Ageing (ELSA) to investigate the age-based digital divide in internet use among older adults. It found that internet use decreases significantly after age 80, with cognitive ability, education, and employment status being key factors influencing this decline. Physical impairments and social isolation were less impactful than previously assumed. Among regular users, older cohorts reported lack of IT skills, trust, and motivation as barriers to greater use. The findings challenge the chronological definition of "older adults" in HCI and suggest a need for nuanced interventions addressing cognitive and educational disparities. Future studies should explore longitudinal patterns, cohort effects, and emerging technologies like AI.