Title: Social change and birth cohort decrease in social support for older adults in China: A cross‐temporal meta‐analysis, 1994–2018
Abstract: Health & Social Care in the CommunityVolume 28, Issue 5 p. 1438-1447 REVIEW ARTICLEFree Access Social change and birth cohort decrease in social support for older adults in China: A cross-temporal meta-analysis, 1994–2018 Zhang Zhao MEd (Master), Zhang Zhao MEd (Master) orcid.org/0000-0001-5320-2210 The Lab of Mental Health and Social Adaption, Southwest University, Chongqing, ChinaSearch for more papers by this authorShuang Jing MEd (Master), Shuang Jing MEd (Master) The Lab of Mental Health and Social Adaption, Southwest University, Chongqing, ChinaSearch for more papers by this authorZhimin Yan PsyD, PhD, Zhimin Yan PsyD, PhD The Lab of Mental Health and Social Adaption, Southwest University, Chongqing, ChinaSearch for more papers by this authorLin Yu PsyD, PhD, Corresponding Author Lin Yu PsyD, PhD [email protected] The Lab of Mental Health and Social Adaption, Southwest University, Chongqing, China Correspondence Lin Yu, The Lab of Mental Health and Social Adaption, Southwest University, No. 2 Tiansheng Road, Beibei District, Chongqing 400715, China. Email: [email protected] for more papers by this author Zhang Zhao MEd (Master), Zhang Zhao MEd (Master) orcid.org/0000-0001-5320-2210 The Lab of Mental Health and Social Adaption, Southwest University, Chongqing, ChinaSearch for more papers by this authorShuang Jing MEd (Master), Shuang Jing MEd (Master) The Lab of Mental Health and Social Adaption, Southwest University, Chongqing, ChinaSearch for more papers by this authorZhimin Yan PsyD, PhD, Zhimin Yan PsyD, PhD The Lab of Mental Health and Social Adaption, Southwest University, Chongqing, ChinaSearch for more papers by this authorLin Yu PsyD, PhD, Corresponding Author Lin Yu PsyD, PhD [email protected] The Lab of Mental Health and Social Adaption, Southwest University, Chongqing, China Correspondence Lin Yu, The Lab of Mental Health and Social Adaption, Southwest University, No. 2 Tiansheng Road, Beibei District, Chongqing 400715, China. Email: [email protected] for more papers by this author First published: 06 May 2020 https://doi.org/10.1111/hsc.13004Citations: 4AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinkedInRedditWechat Abstract Social support not only plays an important role in the physical and mental health of the elderly people but also constitutes an essential resource for healthy ageing. With the rapid economic and social development during the last 40 years in China, the acceleration of urbanisation, and the disintegration of traditional extended families, the social support that Chinese older adults receive may be declining, leading to deterioration in quality of life for the rapidly ageing population. Cross-temporal meta-analysis was employed to investigate changes in older Chinese adults’ social support from 1994 to 2018. One hundred and thirty-six studies (N = 82,722; age ≥ 60) that used the social support rating scale (SSRS) were analysed. Additionally, social support scores were correlated with social indicators to explore the relationship between social support and the environment of social development. Results show that social support scores decreased by 5.09 and 0.73 standard deviations over the past 24 years. Correlation with social indicators suggests that a decrease in social connectedness and an increase in economic imbalance may be responsible for the reduction in social support. What is known about this topic Social support is a key factor in healthy ageing for the elderly people. Declining social support can make ageing difficult. China's aged population is increasing. What this paper adds A cross-temporal meta-analysis method was used to analyse the effects of social support changes in the elderly people. The temporal trend and cohort effect on social support for the elderly people in China were described. The influence of the social environment on social support for the elderly people was analysed through social statistical indicators. 1 INTRODUCTION Social support is a complex and multidimensional concept that can be roughly divided into two categories: (a) objective, which is visible or actual support (material in nature, consisting of social networks and/or family (Tomaka, Thompson, & Palacios, 2006); (b) subjective, consisting of experienced or emotional support (from motion, psychological experience; Tian, Liu, Huang, & Huebner, 2013). Social support, then, represents the material and spiritual help an individual receives from family members, friends, neighbours and other social networks (Taylor, 2011). Social support is probably the most important source of well-being for older people, as it not only provides the resources necessary to help individuals cope with challenges but also exerts a tremendous positive influence on their physical and mental well-being (Harandi, Taghinasab, & Nayeri, 2017). Research has shown that high levels of social support protect one against negative physical health outcomes, such as cardiac disease and hypertension, leading to a decreased mortality rate in older adults (Liu, Hernandez, Trout, Kleiman, & Bozzay, 2017). In terms of its effects on mental health, increased social support predicts better self-reported general and emotional subjective well-being (Heinze, Kruger, Reischl, Cupal, & Zimmerman, 2015; Tajvar, Fletcher, & Grundy, 2016). Besides, social support plays a unique role in reducing loneliness and depression in the elderly people, and even affects the brain through improved cognitive ability (Heinze et al., 2015). With rapid industrialisation, change is the most important fact in social life (de la Sablonnière, 2017). Older adults are the ones who witness the changes in living environment, but their adaptability is often neglected and more social support is needed (Dai et al., 2016). Traditional filial piety culture in Asia, typical of China, where older adults are treated with special care and indulgence; and society, especially the family, has a responsibility to provide them with financial and emotional support (Chow, 2004). However, a series of social and family transformations—such as disintegration of the extended family, decrease in the community network size—all of which have gradually shaken the foundation of intergenerational filial piety, which may in turn lead to changes in social support for the elderly people (Aartsen & Jylhä, 2011). In the context of the same era, research has shown that, as society changes, the rate of loneliness and depression detected among Chinese elderly people is increasing, whereas subjective well-being is declining. These results provide an indirect signal of changes in social support for older adults (Chen, Hicks, & While, 2014; Yan et al., 2014; Yu et al., 2016). Although efforts have been devoted to investigating social support for older adults in China, little attention has been paid to investigating either trends in such social support or the relationship between social changes and social support. On the one hand, longitudinal research is time-consuming and labour-intensive; on the other hand, the cross-sectional study has a small sample size and span, narrow distribution, a brief period of study and problems with accuracy. Thus, such studies cannot draw general conclusions about changes in social support (Asante & Castillo, 2018; Bélanger et al., 2016). Various restrictions put stringent demands on research methods for topics like social support. However, with a population in China of about 250 million over age 60 (National Bureau of Statistics of China, 2019), it is necessary to explore direct evidence for changing trends in social support for older adults so as to provide references for decision-making in welfare policy and regarding social pensions for older adults. A cross-temporal meta-analysis (CTMA) is a special meta-analysis method that examines trends in variables over time using a chronological sequence of studies that are independent of each other, but that use the same tools (Twenge, 1997a, 1997b). CTMA provides a possibility for intertemporal comprehensive research, which can simplify the laws behind the data, fill in for limitations from missing longitudinal studies and make causal conclusions more convincing. Compared to conventional meta-analysis, CTMA is more specific, as it compares multiple scores within the same questionnaire, whereas conventional meta-analysis compares scores from multiple questionnaires with the same structure (Twenge, 2011). In addition, CTMA focuses on changes to the sample mean score rather than calculating effect size (Cohen's d), which is a better method for recording a ‘cohort effect’ (Twenge, 2011). 2 SOCIAL SUPPORT AND THE LARGER SOCIAL ENVIRONMENT The social cognitive theory holds that behaviour is influenced by a variety of factors, including individuals and the environment (Bandura, 1977). Similarly, ecosystem theory contains discussions about the different layers of the social environment that affect individual development (Bronfenbrenner, 1994). Although numerous studies focus on individual factors, a significant problem remains: failure to explore adequately the impact of social factors, such as economic development and policy reform, and the inability to study how they affect individual health (Braveman & Gottlieb, 2014). That social and environmental factors—such as employment, disease and income—have an impact on health behaviours has been supported by evidence from various studies (John-Henderson, Stellar, Mendoza, & Francis, 2015). Social demographic variables with large and convenient data sets are informative for discussing change over time. Even more important, associating general conclusions drawn by CTMA with demography helps to organise research objects from a macro perspective and systematically sort them out so as to arrive at some general laws concerning the social environment's effects on the research objects (Twenge, 2000). This study adopts Twenge's (2000) practice of dividing demographic variables into three constructs: social connectedness, economic conditions and overall threat. 2.1 Social connectedness Social connectedness refers to individuals’ connections to social networks, and urbanisation, birth rates, death rates and divorce rates are often measures of social connectedness in macrosystem studies (Asante & Castillo, 2018; Bailey, Cao, Kuchler, Stroebel, & Wong, 2018; Cornwell, Laumann, & Schumm, 2008). Using these measures, the evidence shows that Chinese society is experiencing a breakdown in social connectedness. First, urbanisation is referred to as ‘stranger society’, which may weaken kinship and friendship ties (Dijst, 2014). Additionally, decreases in both birth and death rates, along with an increase in the divorce rate, have gradually developed into trends (Robards, Evandrou, Falkingham, & Vlachantoni, 2012; Su, Liang, Yang, & Liu, 2018; Wang, 2019). The decline in fertility has further weakened the source of family support (Shanas, 1979). The divorce rate has shaken the family structure—with particularly negative consequences—because family is the main support network for individuals, especially in China (Melchiorre et al., 2013; Yang, 2013). As for mortality, though a declining rate may seem positive, studies have shown that unless individuals have stronger social relationships, a decline in mortality does not mean a general increase in social support (Holt-Lunstad, Smith, & Lay-ton, 2010). 2.2 Economic conditions Economic conditions, in this study, refer to a nation's financial status. Economic conditions objectively reflect changes in people's daily expenses and living conditions, and because an economic downturn directly threatens people's survival, it usually serves as an objective index of social support (Fiori, Antonucci, & Cortina, 2006). China's economy is undergoing a period of great prosperity; thus, gross domestic product (GDP) growth may translate into more social support. Some evidence indicates that a healthy economy can positively predict social support for older people (Aboderin, 2004). On the contrary, economic deterioration means the possibility of a decrease in social support. It needs to be measured by the overall indicators of national economic development: price level, income gap and employment status (Mittelsteadt, Adamowicz, & Box all, 2001). 2.3 Overall threat Overall threat, as a measure of overall health, may be derived from the number of ill, injured or impaired individuals seeking medical attention, as well as medical expenses. ‘Personal medical expenditure’ and ‘the number of hospital beds’ can intuitively reflect the overall health of residents. As life expectancy increases, the health of older groups becomes more worrisome (Jaul & Barron, 2017). Threats to health arise when older people are restricted in their activities by either themselves, their families or their communities. In turn, the social support network of the elderly people may be diminished by health threats, and lower social support may likewise accelerate health deterioration (Campos, Ullman, Aguilera, & Dunkel Schetter, 2014; Gow, Pattie, Whiteman, Whalley, & Deary, 2007). 3 OVERVIEW This study has attempted a) to reveal trends in the social support of Chinese older adults since the early 1990s and b) to identify the relevant social indicators of change. Given the changes in social statistic indicators, we expected a decrease in the social support of Chinese older adults over the last 24 years. We employed a cross-temporal meta-analysis to study birth cohort differences in social support and employed correlation and regression analyses to identify correlations between social indicators and social support. 4 METHOD 4.1 Measurements and literature search Based on the CTMA principle, the social support rating scale (SSRS) was selected as the tool for the study, which is the most prevalent questionnaire for measuring social support in China (Xiao, 1994). It contains 10 items in three subscales: subjective social support, objective social support and the utilisation of social support. The total score of social support is the sum of scores from the 10 items. The objective support score is the sum of items 2, 6 and 7; the subjective support score is the sum of items 1 and 3–5 and the utilisation of support is the sum of items 8–10. Items 1–5 and 8–10 were rated on a 4-point Likert scale, ranging from 1 (‘not at all’) to 4 (‘very much’). Items 6 and 7, if ‘No source’ is answered, give a score of 0; if ‘have a source’ is the answer, each source provides 1 point. Overall, the higher the score, the greater the level of individual social support. In order to achieve a relatively comprehensive data set, we collected data from Chinese academic databases as well as international databases. For domestic academic databases, we selected the two most popular: Wan Fang Data and China National Knowledge Infrastructure. Three keyword search terms were used: SSRS, the elderly people and social support. For the international databases, we selected Web of Science, ProQuest and Wiley, using four keywords search terms: SSRS, old people, social support and Chinese. 4.2 Inclusion criteria The inclusion criteria for the meta-analysis are as follows: (a) Participants must be mainland Chinese older adults (excluding those from Hong Kong, Macao or Taiwan). (b) Participants had no reported physical or mental illness and were not less than 60 years old. (c) Participants’ level of social support was determined by the SSRS scale. (d) Studies used must report the sample size, mean score of social support and standard deviations. For those that did not report the mean score of social support or standard deviations but provided the score for subscales of social support, we applied Formulas 1-2 and 1-2 to calculate these. (e) Finally, studies used must not reproduce data already published elsewhere. In addition, if the year when data collection occurred was not provided in the study, it was coded as two years prior to publication (Twenge & Im, 2007). Eventually, according to the data collection results, 136 publications that included 82,722 older adults were obtained for the period of 1994–2018 (Table 1). (1) (2) TABLE 1. Distribution of studies data in cross-temporal meta-analysis research, 1994–2018 Year of data collection Number of studies Total sample size 1994 1 150 1996 1 3,069 1997 2 438 1999 1 117 2000 2 217 2001 2 527 2003 7 1,641 2004 1 182 2005 2 997 2006 3 1,238 2007 5 2,377 2008 5 6,636 2009 10 5,398 2010 6 4,399 2011 6 4,924 2012 14 5,612 2013 18 12,467 2014 14 8,718 2015 19 12,231 2016 9 7,066 2017 6 2,503 2018 2 1,815 Total 136 82,722 Note: represents the average after combination, st represents the standard deviation after combination, ni represents the sample size, xi represents the average deviation and si represents the standard deviation. Based on the general steps of meta-analysis and suggestions from Twenge, a database was established for this study by assigning a unique number to each document and recording the basic data (N, M, SD), publication date and data collection date of each document (Twenge, 2000). Additional information was encoded next, including the type of journal and the data collection area (Table 2). TABLE 2. Cross-temporal meta-analysis research variable coding table Variable Code Number of studies Type of journal 1 = Core publication 53 2 = General publication 50 3 = Collection of academic studies 33 Region 0 = No clear regional information 3 1 = East coast 59 2 = Developing central region 26 3 = Western development region 42 4 = Northeast China 6 4.3 Sources of social indicators According to Twenge (2000), social indicators should fit basic criteria: the indicators must be easily obtained and quantified as meaningful continuous variables, and they must represent general trends. As mentioned earlier, we extracted data for social connectedness, economic conditions and overall threats through the National Bureau of Statistics of China website (National Bureau of Statistics of China, 2019). 4.4 Data analyses Correlation and regression methods were used to analyse the ‘cohort effect’, that is, the possible changes to social support for Chinese older adults and factors leading to these changes. In addition, we used a cross-lagging approach to analyse the influence of socioeconomic factors on social support. In this manner, we matched the social indexes with the score of social support at three points in time: 10 years prior to data collection, 5 years prior to data collection and the year of data collection. The lagging social indexes suggested how older adults’ social support related to their socioeconomic status when they were younger. If higher socioeconomic status can alter mental variables, the correlations should be significant when the scores are matched to social indexes 5–10 years in the past (Twenge, 2000). 5 RESULTS 5.1 Correlations between the mean score of social support and the year of data collection To visualise the social support trend from 1994 to 2018, we used the year of data collection as the abscissa, and social support and its three-dimensional scores as the ordinate to create a scatter plot. From the scatter plot, the mean values of the total social support score, objective support and utilisation of support all showed a decreasing trend at the year of data collection (Figure 1). In addition, the results of curve estimation show that the linear model fit the relationship between social support and the year of data collection well (F = 8.08, p < .05, R2 = 0.06). FIGURE 1Open in figure viewerPowerPoint Changes in Chinese older adults’ social support scores, 1994–2018 As with a general meta-analysis, the results of CTMA can also be influenced by such characteristics of studies as the year of publication, journal type and source (region; Twenge, 2011). In order to control for the influence of these factors, we used the social support score and its three subscale scores as dependent variables, and the year of collection data, region and journal type as independent variables for the stepwise regression analysis. The results show that under a controlled sample size and after the inclusion of three independent variables, the cohort effect on the scores for social support, subjective support, objective support and utilisation of support are still significant (Table 3). This shows that the relationship between the level of social support and the year of data collection was not significantly affected by journal type, subject source or other factors. TABLE 3. The correlation between the mean value of social support factors and age Variable Unweighted sample size Weighted by sample size Social support score −0.24** 0.06 −0.29*** 0.08 −0.39*** 0.15 Subjective support score −0.06 0.01 −11*** 0.01 −0.24*** 0.06 Objective support score −0.32*** 0.11 −0.30*** 0.09 −0.37*** 0.14 Utilisation of social support −0.21* 0.04 −0.21*** 0.05 −0.27*** 0.07 Note r is the correlation coefficient of unweighted sample size, β1 is a standardised regression coefficient weighted by sample size, β2 is the standardised regression coefficient weighted by sample size and controlled additional variables (region, journal type) and and are the coefficients of determination. * p < .05, ** p < .01, *** p < .001. 5.2 Magnitude of change To explain the magnitude of change during these 24 years, we referred to previous studies by calculating the effect of amount d or the interpretation rate r2, which is expressed in Formulas 3 and 4, where SD is the mean standard deviation over 24 years (Twenge & Im, 2007). (3) (4) First, social support scores and the mean value of each factor were taken as the dependent variable and the year of data collection as the independent variable. Then, the sample size was weighted to establish the regression equation y = Bχ + C, where B indicates the unstandardised regression coefficient, x indicates the year, C indicates the constant and y indicates the mean score of older Chinese adults’ social support. The mean score of M1994 and M2018 was obtained by substituting 1994 and 2018 into the regression equation. Finally, we calculated the mean change between M1994 and M2018 and divided it by the mean standard deviation of all studies over the 24-year period to get the value d (Table 4). This method prevents an ecological fallacy (Twenge & Im, 2007). TABLE 4. Changes in levels of social support for Chinese older adults Variable SD d r 2 Social support score 39.72 34.63 −5.09 6.95 −0.73 0.12 Subjective support score 20.20 18.90 −1.30 4.22 −0.31 0.02 Objective support score 9.81 7.26 −2.54 2.69 −0.95 0.18 Utilisation of social support 7.38 6.64 −0.74 2.21 −0.34 0.03 Note Mchange = M2018 − M1994, d = (M2018 – M2014)/SD. As displayed in Table 4, the mean score for social support decreased by 5.09 between 1994 and 2018. During this period, the standard deviation decreased by 0.73, reflecting a middle effect (0.5 < d < 0.8; Cohen, 1992). We converted d (−0.73) into a variance explained by year, which resulted in a 12% ratio change. 5.3 Correlations between the social support mean score and social statistics The CTMA reveals that the social support of older adults in China shows a downward trend. The reasons for this trend can be analysed via the relationship between social support and social statistical data (Twenge, 2011). Because of the difference in data units, the analysis yielded correlations between these social statistics and the z-scored scale combinations of social support. We found that social connectedness predicts social support well (Table 5). The level of urbanisation, birth rate and divorce rate shows fairly consistent correlations over time. Of these, the birth rate is positively correlated with social support, whereas the level of urbanisation and divorce rate are negatively correlated with social support (p < .01). In addition, the results show that the correlation between social support and social statistics is strong in both the past and the present. TABLE 5. Correlation between social indicators and social support scores for Chinese older adults weighted by sample size, 1994–2018 (number of studies = 136) 10 years prior 5 years prior 0 years prior Social connectedness Urbanisation level −0.43*** −0.35*** −0.43*** Divorce rate −0.22** −0.30*** −0.32*** Birth rate 0.45*** 0.39*** 0.31*** Death rate 0.29** −0.43*** −0.29** Economic conditions GDP −0.25** −0.33*** −0.36*** CPI −0.34*** −0.48*** −0.41*** Gini coefficient −0.52*** −0.37*** −0.22* Unemployment rate −0.54*** −0.29** −0.31*** Overall threat Personal medical expenditure −0.29** −0.29** −0.32*** The number of hospital beds −0.18* −0.24** −0.35*** Regression with total scales Economic conditions −0.44*** −0.41 *** −0.42*** Social connectedness & Overall threat −0.15 −0.32*** −0.33*** Note Regression with a total scale is the standardised regression coefficient (Beta). * p < .05, ** p < .01, *** p < .001. Economic conditions also predict social support and show fairly consistent results as well. The data from the GDP, CPI, Gini coefficient and unemployment rate all negatively correlate with the year of data collection (p < .05). The same is true of overall threat, with both personal medical expenditure and the number of hospital beds showing significant negative correlations over time (p < .05). These analyses suggest that several aspects of the larger environment may have influenced levels of social support. During the analysis, we found that social connectedness and overall threat correlate at about 0.94 (r = 0.938, p < .001), using the z-score. Hence, the two composites display too much multicollinearity to be entered separately into a regression equation. A new regression equation was devised, with economic and social connectedness/overall threat as the independent variables and social support as the dependent variable. The results show that when social connectedness/overall threat index is controlled for, economic conditions predict social support well. When economic conditions are controlled for, the social connectedness/overall threat index predicts social support well in the first 5 years and the current year. 6 DISCUSSION In this study, changes in the times and the factors that influence social support for the elderly people in China were discussed. Consistent with the hypothesis, social support for older adults in China decreased from 1994 to 2018. In addition to existing factors in the literature, we supplemented with the impact of social and demographic data in the macrosystem, as few studies have dealt with these factors. Our findings suggest that economic conditions can predict the extent of available social support. The study confirms a negative relationship between CPI, Gini coefficient and unemployment rate, and social support for older adults. In opposition to previous studies, we found that GDP negatively predicts social support (Aboderin, 2004). Our results showed that strong and universal relationships between socioeconomic factors and the outcomes of social support for older people can reflect more complex and long-term causal relationships. In the past 24 years, China's GDP has increased by 18.51 times (National Bureau of Statistics of China, 2019). However, we conclude from the results that economic growth has not benefited all segments of the population equally or at the same rate, especially the elderly people. There may be a few reasons for this: first, in most countries, the main forms of care and support for the elderly people are home and informal care (Centre for Policy on Ageing, 2014). Especially in the context of a collectivist culture, the government expects families, communities and NGOs to play an important role in guaranteeing the welfare of the elderly people (Dai et al., 2016; Hu, Li, & Shi, 2019). What makes China's social welfare system and others different is that the dynamic relationship between the central and local governments determines that the welfare expenditure of local governments is not unified by the central government. Rather, it is determined by the total economic volume of the local government, which means that in the absence of external pressure or central fiscal incentives, local governments are unwilling to increase social welfare levels (Zhu, 2016). High-level elderly people welfare policies are quite limited and have strict guidelines on matters such as age boundaries, physical disabilities and low income (Hu et al., 2019). Second, Xie and Zhou (2014) point out that China's GDP and Gini coefficient are showing a ‘double high’ positive correlation, meaning that rapid economic development and imbalances co-exist, which is consistent with our conclusions. The data show that although the growth rate of the Gini coefficient has slowed, in fac