There are gender differences in the socioeconomic, health, social, and behavioral predictors of psychological well-being among older Taiwanese adults with some having a stronger association with psychological well-being in women compared to men while others show a stronger association in men.

The dataset used in this research was obtained from A Taiwanese Survey of Mental Health conducted in 2016. The sample size was 2,286 and was representative of the national population. The sample was obtained using the multistage proportional probability to size (Figure 1). The sampling frame consisted of a list containing all districts on the main island which encompasses 98.9% of the entire national population. Each district was categorized into one of seven levels of urbanization (with level 1 being the most urbanized) through a cluster analysis that considered population density, the population over 65 years old, the number of people who had achieved junior college education or higher, the number of Western medical institutions, and the number of farmers and herdsmen. In this study, level 7 was recoded as the highest level of urbanization and level 1 the lowest (Figure 2).

Districts selected from each level with the probability of selection proportional to the size of its older population to minimize data collection bias.

Level 1 consisted of Guishan District (Taoyuan City) and Puzi City (Chiayi County).

Level 2 consisted of Baihe District (Tainan City) and Dacheng Township (Changhua County).

Level 3 consisted of Xinyi Township (Nantou County) and Wanrong Township (Hualien County).

Level 4 consisted of Yuanli Town (Miaoli County) and Jiadong Township (Pingtung County).

Level 5 consisted of Minxiong Township (Chiayi County) and Jiaoxi Township (Yilan County).

Level 6 consisted of Fengshan District (Kaohsiung City), Nantou City (Nantou County), Xincheng Township (Hualien County), Bade District (Taoyuan City), Rende District (Tainan City) and Yilan City (Yilan County).

Level 7, the most urbanized, consisted of Banqiao District (New Taipei City), Wanhua District (Taipei City), Central District, (Taichung City), and Xinzhuang District (New Taipei City).

Census blocks in each district were selected using proportional probability according to the population of residents. Households from these census blocks that had at least one individual aged ≥55 years were selected using the government household registry. The household registration addresses were sorted using simple random sampling for each census block.

Interviewers visited the household to verify that there was an eligible participant. Inclusion criteria were Taiwanese citizenship, age ≥ 55 years, and the ability to provide informed consent. The interviewer read from the questionnaire and filled out the responses at the participants’ homes using paper and pencil. This paper only included participants who were ≥ 65 years of age in our study sample thus the study sample featured 1,533 participants.

PWB was measured using an 18-item version of the Carol Ryff Psychological Well-Being Scale (Li, 2014). The sum of 18 scored items (1 ~ 5 strongly disagree, disagree, neutral, agree, and strongly agree) consisted of the PWB score, a combination of the six dimensions of psychological well-being in Ryff’s model. The median score (M = 69) was used as a “cutoff” mark (Sapranaviciute-Zabazlajeva et al., 2018). Scores of ≤69 were categorized as “low PWB” and scores of >69 were categorized as “high PWB.”

Predictor variables were classified as socioeconomic, health, social, and behavioral factors. Gender was categorized as “male” or “female.” Age, education, marital status, religious beliefs, employment, financial stress, living environment, and financial dependence were classified as socioeconomic. Religion and employment were dichotomized. Education was described as no education, primary, secondary, and tertiary. Marital status was “married” and “not married (single, divorced, separated, or widowed). Financial stress was assessed by the question, “Do you have difficulty in meeting your household expenses?” Responses of “very difficult” and “with some difficulty” were categorized as high, and “no difficulty at all” and “little difficulty” were categorized as low financial stress. The question, “Overall, how satisfied are you with your living environment?” was used to determine satisfaction with their living environment which was dichotomized into “satisfied” and “dissatisfied.” Place of residence was derived from the seven levels of urbanization during the sampling process. The first level of urbanization consisted of the most urbanized areas in Taiwan and was described as “metropolitan,” while the other six levels were described as “other.” Metropolitan locales included Taipei City, New Taipei City, and Taichung City.

Financial dependence was assessed by the question, “What is your main source of income at present?” Participants who received child support or revenue from the state as their main source of income were considered “dependent,” and those who did not were “not dependent.”

Health factors included self-reported health, chronic diseases, sleep quality, and body pain. The questions “In general, how is your health?” and “How was your sleep in the past week” respectively assessed self-reported health and sleep quality. The original five response categories (very good, good, fair, bad, and very bad) were amalgamated into three (poor, fair, and good). The presence of a chronic disease was determined by asking a participant if a healthcare provider had diagnosed them with any of the following: diabetes, hypertension, hyperlipidemia, duodenal ulcer, or mental illness (depression, anxiety, or dementia). The question “Did you have body pain during the past 4 weeks? If so, how much?” was used to assess the body pain variable. The original six-point scale was amalgamated into the three categories of “none,” “mild to moderate,” and “severe.”

Social factors consisted of family relationships and time spent with friends. The strength of family relationships was measured by three questions: “How frequent is your interaction?,” “How is the quality of the relationship?,” and “How well do you get along with family members?” Each response was allocated a score of 1 ~ 5 (Chiang et al., 2013). A higher score for the three questions indicated a better relationship. A median of 12 was used as a cutoff to create the categories of “low” and “high.” Time spent with friends was assessed by the amount of time participants spent socializing with their friends as “little,” “fair,” and “a lot.”

Behavioral factors included caregiving, participation in community activities, and internet use. The question, “Does anyone in your family need your care due to chronic disease, mental illness, old age, dementia, disability, cancer, major injury, etc.?” to determine if they performed a caregiving role. Community participation was dichotomized (yes or no) and included participation in organizations, religious, charity, or other social functions. Time spent with friends examined the amount of time spent interacting with their friends. Daily internet use was dichotomized as yes or no.

Data was analyzed using SPSS for Windows version 22 (SPSS, Chicago, IL, USA). The study sample (N = 1,533) was created after data cleaning and selecting all participants ≥65 years with complete data fields, and then descriptive characteristics were determined. Bivariate and multivariate logistics analysis were performed with PWB, then stratified by gender using a 95% confidence interval (CI) to report the percentage of the odds ratio (OR).

Each study participant provided written informed consent after the purpose and contents of the survey were explained. This study was approved by the Taipei Medical University Joint Institutional Review Board (no.: 201608036).

In contrast to other studies, marriage was not found to be a significant predictor of PWB when controlled for other socio-economic, health, social, and behavioral factors. It can then be suggested that marriage quality or marriage happiness would be a better predictor of PWB among older Taiwanese adults.

Son preference is a cultural phenomenon that was dominant in pre-industrialized Taiwan before the 1960s (Lin, 2009). This preference for male children meant that boys were more educated than girls, facilitating resource substitution. Resource substitution is a theory that states that education improves the PWB of women more than men due to inequity in socioeconomics. This postulates that women depend more on education to achieve well-being than their male counterparts (Ross and Mirowsky, 2006). Thus the effect of a single factor such as education is amplified in the marginalized group and is less significant in the preferred group (Ross and Mirowsky, 2010). This could account for the amplified effect education had on PWB for women in comparison to men. Education can potentially promote autonomy, positive relations with others, and personal growth, the foundations of psychological well-being.

The study’s negative associations of financial dependence and unemployment with PWB are consistent with what is in the literature (Ryff and Singer, 2008; Mousteri et al., 2018; Moxham et al., 2018; Pak, 2020). Distinctive to this study, these relationships were notably observed in men. Taiwan’s traditional Confucian values often depict men as providers and having a higher socioeconomic status than women (Lin, 2009); therefore, financial dependence and unemployment may negatively affect the PWB of men more than women in this cultural context. Older men unlike women are expected to support their families financially. If such expectations become aligned with a man’s sense of purpose and autonomy then when they are no longer able to work and became financially dependent, their well-being could be diminished.

It is believed that religion provides positive intrapersonal factors, e.g., hope, and interpersonal factors, e.g., social interactions with others improve mental health (Joshi et al., 2008). Religion can also promote happiness and a sense of purpose which are key in assessing PWB. This accounts for religion being a strong predictor of PWB in both men and women (Li, 2014).

Unsatisfactory physical living conditions could be a source of contention within a family. This could potentially be internalized more by women who may feel more burdened to maintain harmony in the home due to traditional Confucian values (Chao and Roth, 2000). Additionally, women tend to be more sensitive to environmental security issues than men (Bentes et al., 2017). This highlights potential directions for further research.

Metropolitan regions had a notably more positive influence on the PWB of older Taiwanese adults irrespective of gender highlighting their advantages over their limitations. Happiness in older populations is associated with the availability of good-quality services such as transportation (Krefis et al., 2018). Additionally, the presence of green and blue spaces, both of which are popular in Taiwan’s metropolitan areas have been found to promote good mental well-being (Krefis et al., 2018). Older adults in rural regions often encounter limited access to adequate infrastructure. Additionally, they may experience empty nests, as their younger family members migrate to the city to improve their livelihoods, leaving them without caretakers (Srivastava, 2009). This may result in loneliness and feelings of abandonment, potentially contributing to psychological distress.

Self-rated health (SRH) reflects one’s perception of their health while the presence of chronic disease reflects a medically diagnosed illness. Distinctive to previous studies, this study examines gender differences in the perception of good health and a medically diagnosed chronic disease and its impact on PWB. Consistent with the literature, (Williams et al., 2017) as higher SRH significantly improved PWB outcomes in men and women. However, the relationship between the presence of chronic disease and PWB was only significant for men. It can be suggested that when controlled for SRH, chronic disease is not a significant predictor of PWB for women as they are more aware of their health status and manage chronic diseases more efficiently. Thus, reducing potential effects on PWB. Men tend to be more reluctant to seek and articulate health issues to avoid the perception of being “weak” (Matud, 2017). A medical diagnosis forces men to confront their ill health, potentially contributing to psychological distress. Thus, it can be deduced that the inclusion of medically diagnosed conditions is particularly useful when examining outcomes for men.

The results of this study were consistent with the literature as good family relationships and friendships were positively associated with higher PWB (Tramonti et al., 2019). Taiwanese culture exhibits strong familial values (Chiao et al., 2019); however, this association was only observed in women in this study. As women are traditionally tasked with the responsibility of maintaining harmony and cohesion within the family (Yang, 2016), thus, older Taiwanese women may be more sensitive than men to difficulties and dysfunction within the family. Consequently, positive family relationships may have a more positive influence on their PWB compared to men. One can suggest that women might place greater emphasis on family relationships, which could explain why there was a significant association between the variable “spending a fair amount of time with friends” and not “spending a lot of time with friends.” Friendships play a role in higher PWB among older adults (Thanakwang et al., 2012; Greenfield and Reyes, 2014) as they promote social support and pleasant experiences as aging progresses (Ng et al., 2021). From this study, spending time with friends appeared to be a better predictor of PWB for women than men, while women’s PWB is more heavily influenced by family relationships.

Consistent with the literature, community participation promoted higher PWB in both men and women (Gyasi et al., 2019). Community participation is associated with increased social support, health and well-being (Jones et al., 2013). Community participation may be a suitable approach to PWB interventions aimed at capturing both male and female older Taiwanese adults.

Generally, the literature states that internet use is associated with PWB in both men and women (Quintana et al., 2018; Wang et al., 2019). It is difficult to ascertain why this type of association was significant for women but not for men without information. Finding out the purpose of internet use may give insight into these differences. As the world becomes more digital, it is worthwhile to deeply explore internet use among older populations and its impact on PWB.

The findings of this study provide insight into the predictors of psychological well-being.

Overall, employment, education, religious beliefs, satisfaction with the living environment, living in a metropolitan region, financial dependence, self-rated health, chronic disease, community participation, time spent with friends, family relationship score, and daily internet use were all significant predictors of PWB. Financial dependence and chronic disease both distinctively negatively impact PWB outcomes.

Employment positively predicts PWB in men. Financial dependence and chronic disease both predict negative PWB in men. While education positively predicts PWB for both men and women, it has a greater impact on women than men. Satisfaction with the living environment was found to be a positive predictor for PWB in women. While spending time with friends positively predicts PWB for both men and women, it has less impact on women than men. Additionally, family relationships positively predicted PWB in women but had no significant impact on men.

In conclusion, this study supports the hypothesis. There are gender differences in the socioeconomic, health, social, and behavioral predictors of psychological well-being among older Taiwanese adults.

Several interventions could be considered to improve the psychological well-being of older Taiwanese adults. Educational Programs targeted at older adults could be beneficial as education is a strong predictor of PWB for women and financial dependence is a strong predictor of PWB for men. Programs focused on vocational training or skill development may help empower them, increasing their sense of purpose and self-esteem. Such programs can also encourage older adults to create other sources of income. Separate educational programs can be specifically tailored to suit the needs of older Taiwanese men and women.

Social Support Workshops targeted at older adults could also be beneficial. These workshops could be aimed at strengthening and improving family dynamics as family relationships are a strong predictor of PWB for women. These workshops can involve counseling services surrounding mental and physiological health, especially for men as the diagnosis of a chronic disease is a strong predictor of lower PWB outcomes for men. This can be facilitated by community and religious groups as well as healthcare institutions.

Community engagement initiatives could also be beneficial to PWB in both men and women. Such initiatives could include social activities, cultural activities, and volunteer opportunities. This will promote social connections and a sense of belonging and purpose. Community health-based programs can provide opportunities for screening for chronic diseases and assistance with disease management.

Urban Planning and Infrastructure Development could be beneficial in improving PWB in older adults. Initiatives aimed at creating age-friendly environments such as accessible mental health services for older adults with chronic conditions, accessible public spaces, reliable and frequent public transportation services, and accessible recreation will promote active aging and social inclusion, especially outside the metropolitan centers.

This study is cross-sectional thus causation cannot be determined. Additionally, PWB can be affected by recent positive or negative events, thus the recorded PWB may be different from the participant’s norm.

Variables were self-reported which might be susceptible to recall and social desirability bias. Variables such as financial stress, satisfaction with the living environment, and time spent with friends were subjective. Objective tools of measurement of these variables would improve the robustness of the data.

The dataset was fixed which limited scope. Therefore, new hypotheses and research questions that arose during data analysis could not be explored nor could additional relevant variables be considered. Additionally, individuals without household registration and those institutionalized in retirement homes were excluded thus limiting generalizability.

While the results of this study share several similarities with the results of other studies, the influence of the cultural context of Taiwan must be considered. This can reduce the extent to which it can be applied globally in other populations.

It is highly recommended that this study be replicated through several longitudinal studies focusing on gender-specific analyses of predictors of PWB. Specifically, future research should explore detailed examinations of predictors such as marital happiness rather than marital status, place of residence and the living environment, family dynamics, chronic health conditions, and financial dependency.

It is also recommended that intervention studies be conducted to evaluate the effectiveness of education programs, urban planning and infrastructure development, and the establishment of social support networks.