Page 46 - IJPS-11-1

International Journal of
Population Studies Age, gender, internet and older adults’ cognition

smoke cigarettes now?”, and the responses were centering, the variance inflation factors were all lower than
dichotomous (0 = No, 1 = Yes). The amount of drinks 4, lower than the commonly adopted cutoff value of 5.0
per week was determined using two questions, “In (Sheather, 2009). All analyses were performed using Stata
the last three months, on average, how many days per 14.2. (StataCorp LP. College Station, Texas).
week have you had any alcohol to drink? (for example,
beer, wine, or any drink containing liquor.)” and “In 3. Results
the last three months, on the days you drink, about 3.1. Sample characteristics
how many drinks did you have?”
As shown in Table 1, the average use of the Internet among
Health-related factors included doctor-diagnosed the sample was nearly 5, almost once a week (M = 4.98,
diseases, such as high blood pressure, diabetes, cancer, SD = 2.59). The deviation from mean Internet use was
depression, ADL, IADL, and self-rated health. Doctor- 1.81e^-16 (SD = 2.59). The mean age was 67.36 years old
diagnosed diseases were measured dichotomously (0 = No, (SD = 10.73) and the deviation from mean age was 2.39e^-
1 = Yes). Depression was measured using the Center for 15 (SD = 10.73). On average, sample’s years of education
Epidemiological Studies Depression (CESD) scale (Radloff, was 13.09, a little above high school graduate. More than
1977). Eight questions about feelings that participants had half of the sample were white (72.38%), women (59.64%),
experience in the past week, such as whether they “felt and married or partnered (62.67%). Most of the sample did
depressed,” “felt activities were efforts,” “slept restless,” “was not smoke now (87.18%). About 60% had been diagnosed
happy,” “felt loneliness,” “enjoyed life,” “felt sad,” and “could with high blood pressure, 26% with diabetes and 16% with
not get going” in much of the time were asked (0 = No, cancer. The average frequencies of doing light and moderate
1 = Yes). Questions on “was happy” and “enjoyed life” were activities were 3.39 (SD = 1.05) and 3.03 (SD = 1.29), in
reversely recoded in this study (0 = Yes, 1 = no). Depression the range of once a week and more than once a week;
was the unweighted sum of the eight items. ADL was and the average frequency of vigorous activities was 2.06
measured by the total number of limitations on bathing, (SD = 1.32), about one to three times per month. The mean
dressing, eating, getting in/out of bed, and walking across a drink per week was about 2.72 (SD = 6.51), depression was
room. ADL ranged from 0 to 5, with higher score indicating 1.36 (SD = 1.91), limitations on activities of daily living was
more limitations. IADL was measured by the total number 0.30 (SD = 0.83), limitations on activities of instrumental
of limited instrumental activities, such as using a phone, daily living was 0.23 (SD = 0.70), and self-rated health was
using money, taking medications, shopping for groceries, 3.13 (SD = 1.02).
and preparing for hot meals. IADL ranged from 0 to 5, with
higher score indicating more limitations. Self-rated health In terms of cognition, differences were found on gender,
status was assessed by a 5-point Likert scale and reversely race, marital status, current smoking status, and being
recoded (1 = Poor, 2 = Fair, 3 = Good, 4 = Very good, and diagnosed with high blood pressure and diabetes. To be
5 = Excellent). more specific, being women, white, married, or partnered
had higher cognition, while being a current smoker, and
2.3. Statistical analysis being diagnosed with high blood pressure and diabetes
had lower cognition.
First, descriptive analysis was conducted to better
understand the characteristics of the sample. Second, Gender differences were identified on marital status,
bivariate analysis – t-test and chi-squared test – were engagement with light, moderate, and vigorous activities,
performed to analyze the associations between cognition having drinks, being diagnosed with high blood pressure
and covariates, and the associations between gender and and diabetes, depression, ADL, IADL, and self-rated health
covariates. Third, multilinear regressions were conducted status. Specifically, compared with older men, older women
to examine the interaction between Internet use and age, as reported more light activities, less moderate and vigorous
well as the three-way interaction of Internet use, age, and activities, less drinks per week, higher levels of depression,
gender on cognition among older adults. Model 1 included more ADL and IADL, and lower self-rated health status,
Internet use, age, and covariates; Model 2 added the and also, they were less likely to be married or partnered,
interaction of Internet use and age to the analysis; Model being diagnosed with high blood pressure and diabetes.
3 added gender; Model 4 added the interaction of Internet
use and gender, and the interaction of age and gender; and 3.2. Multilinear regressions on sample’s cognition
Model 5, the final model, added the three-way interaction Multilinear regression results are summarized in Table 2.
of Internet use, age, and gender. Due to multicollinearity, In Model 1, more Internet use (β = 0.370, p < 0.001)
continuous variables defining the interactions – Internet was associated with a higher level of cognition, while
use and age – were centered at their mean values. After older age (β = −0.091, p < 0.001) was associated with

Volume 11 Issue 1 (2025) 40 https://doi.org/10.36922/ijps.3119