Page 58 - IJPS-8-1

International Journal of
Population Studies Cognitive functioning and active life engagement

In our study, the diary measure of time-use may not add amount of earnings of the individual and his/her spouse
up to 24 h, because some categories of time-use including in the previous 12 months, including pensions, part-time
volunteering that rarely happened in rural China and income, and earnings from self-employed activities. It was
personal care activities such as sleeping, bathing, or eating transformed using in form in regression models.
at home were not considered in this study. We also included living arrangement, social support,
The amount of time spent on the activities of each presence of chronic disease(s), and Activities of Daily Life
category was calculated. For each activity, we measured (ADLs) as potential confounders. Living arrangement was
with both participation and intensity (or daily duration coded as a dichotomous variable (living alone = 1 and
of participation). Participation was assessed by a binary others = 0). Social support was measured by three indicators
indicator of whether an individual participated in a given (continuous variables). The first one asked respondent “how
activity. For the subset who participated, the number of many relatives and friends are contacted at least once within
hours spent on that activity was used to assess intensity. 1 month.” The second one asked respondent “how many
We, then, grouped activities into the six broad categories relatives and friends that you feel safe to discuss personal
of the interest for the current study and recalculated these issues are there.” The third one asked respondent “how
measures. many relatives and friends that you can look for help are
there.” Three dichotomous variables were used to indicate
2.2.2. Key predictors: Cognitive functioning the presence of each of the following chronic diseases:
Cognitive functioning was considered as the key predictors diabetes, hypertension, and cardiovascular disease (heart
of active life engagement in this study, which was measured disease or stroke). ADLs were assessed through a series of
both at the baseline level and its decline in follow-up. questions about activities of daily living and instrumental
activities of daily living (ranges from 15 to 45), with a higher
The level of cognitive functioning was measured with score indicating the older person exhibits better capability
the Chinese version of the Mini-Mental State Examination of performing these activities.
(MMSE). The MMSE mainly focuses on four dimensions of
cognitive functioning: orientation, calculation, language or 2.3. Analytical methodology
comprehension, and recall (Folstein, Folstein, and McHugh, This study adopted both univariate and multivariate
1975). In the Chinese version, we, further, modified analyses of our two-wave panel data. Descriptive statistics
some items according to the rural Chinese cultural and were presented to summarize sample characteristics,
socioeconomic conditions, to improve the validity and the cognitive functioning, and the time-use pattern. We also
reliability of the scale implemented in the rural Chinese conducted two sets of analyses to examine the impacts
population. For example, the participants were asked to of cognitive functioning level in 2015 and its decline on
write a sentence, which is impossible for illiterate older the likelihood and the intensity of participating different
adults to complete. We had to drop this item. The modified activities in 2018. According to the two measures of each
Chinese version of MMSE reached a high reliability in this activity, we first run logistic regressions to model the
study (cronbach’s α = 0.85). The combined score of MMSE impacts of cognitive functioning level in 2015 and its
ranges from 0 to 24, with higher scores indicating better decline on probability of participating in these activities or
cognitive functioning. The decline of cognitive functioning prevalence of participation in 2018. Then, we applied OLS
was coded as 1 if the score of cognitive function in 2018 is regressions among participants to model the intensity (or
lower than that in 2015, otherwise was coded as 0. daily duration) of participating in each category of activities,
which given that our outcome was a count of the number
2.2.3. Confounders: Sociodemographic variables and of hours spent. All models were adjusted for potential
physical health confounders including sociodemographic characteristics,
The sociodemographic variables of age, gender, marital living arrangement, social support, presence of chronic
status, education, and household income, which had disease, ADLs in 2015, and respective activities in 2015. In
been identified as important factors of health and time- addition, normal test showed that the distributions of hours
use of older people, were controlled in the data analysis. spent in six activities were skewed. Ln format was used to
We included age as a continuous variable and gender as a transform the dependent variables in OLS regression. All
dichotomous variable (male = 1). Marital status was coded analyses were conducted using Stata 15 software.
as married/living with partner as 1 and 0 as otherwise.
Educational attainment was constructed as a dichotomous 3. Results
variable: “illiterate” or “literate” (primary school or The characteristics of the sample in 2015 and functional
above). Household income was assessed by the total health of sampled older adults are shown in Table 1. The

Volume 8 Issue 1 (2022) 52 https://doi.org/10.36922/ijps.v8i1.1301