Page 39 - IJPS-6-2

Wang and Kubanga

Table 4. Ordinal logistic regression on internet use by age group.
Model 1 Model 2 Model 3 Model 4
Young-old Old-old Young-old Old-old Young-old Old-old Young-old Old-old
Demographic
Women 1.38* 0.94 1.47** 0.92 1.50** 1.14 2.19** 1.05
Living alone 0.73* 0.90 0.98 1.11 0.99 0.98 0.91 0.91
Socioeconomic
Years of education 1.38*** 1.35*** 1.37*** 1.38*** 1.22*** 1.26**
Married or partnered 1.03 1.76 1.02 2.18 1.15 1.71
Retired 0.38*** 0.48 0.46*** 0.61 0.37*** 0.45
Living in poverty 0.32*** 0.51 0.32*** 0.66 0.27*** 0.75
Health-related
Self-rated health 1.15 1.18 1.11 1.21
Number of diseases 1.00 1.21 1.03 1.32
Difficulties in activities 0.89 0.78 0.77 0.76
of daily living
Difficulties in 0.81 0.86 1.15 1.05
instrumental activities
of daily living
Mental health-related
Depression 0.87* 0.90
Discrimination 1.19 1.02
Cognitive functioning 1.15*** 1.14**
R square 0.004 0.000 0.111 0.081 0.118 0.097 0.156 0.124
-Log likelihood 1364.58*** 255.20*** 1165.23*** 233.20*** 1102.75*** 217.71*** 526.80*** 194.56***
Effect sizes stand for odds ratio. *p<0.05; **p<0.01; ***p<0.001.
difficulties. They may be more likely to interpret their limitations in ADL as results of older age, internalize ageist
stereotypes and believe that they are too old, and act out those stereotypes in the form of a self-fulfilling prophecy
(Chrisler, Barney, and Palatino, 2016). Consequently, they might reduce the use of the Internet, which is consistent with
a recent study (Choi, Kim, Chipalo, et al., 2020). Choi et al. (2020) also used the 2016 Wave of the HRS data and found
that negative self-perception of aging was associated with reduced internet use only among older women but not older
men, despite using a different analytical method.
Previous studies have indicated that older adults with depressive symptoms were more likely to have less social
resources, such as living alone, lacking or having lost close relationships, and having low levels of social support, so they
might be more likely to use the Internet and look for social interactions with others online (Bessiere, Kiesler, Kraut, et al.,
2008; Choi and DiNitto, 2013). However, in this study, depression was found to be associated with decreased odds of more
frequent use of the Internet among older African American men and not significant among women. This may be explained
by different influences of depression on older men and women and their different coping strategies. The negative effects of
depression on daily functioning such as decreased primary activity, mobility, and the ability to conduct household chores
are more severe for men than for women (Forlani, Morri, Ferrari, et al., 2014). Older men who are affected by depression
may not have time or be competent to use the Internet. Older women’s internet use may also be affected by depression,
but they may have been more likely to use the Internet to seek social support compared to men (Girgus, Yang, and Ferri,
2017). For example, older women experiencing depression might be more likely to chat with their families and friends
online, post their experiences and feelings in their social media accounts, or look for more coping strategies online.
Age difference on correlates of internet use was also revealed in this study. Among old-old adults, those who aged 75
or older, more years of education and better cognitive functioning were the only two significant correlates of more
frequent use of the Internet. However, among young-old adults, being retired, living in poverty, and depression were also
predictive of internet use frequency besides education and cognition. Different from retirement, poverty, and depression,
education and cognition are relevant to one’s capacity to use the Internet. It seemed that compared to young-old adults, the
capacity to use the Internet may be more essential for internet use among old-old adults. This could be explained by the
complexity of the Internet and the required technology knowledge and skills to use it. Older adults with higher education

International Journal of Population Studies | 2020, Volume 6, Issue 2 33