Introduction
Modernisation theory has been classified as one of the core theories within the sociology of ageing (Vauclair and Rudnev, Reference Vauclair, Rudnev, Gu and Dupre2019) which aims to explain why the status of older people decreases over time and differs across societies. In his theory, Cowgill (Reference Cowgill and Gubrium1974) proposed that traditional societies, which are usually associated with agricultural and craft production, assign an important role to older adults and, therefore, an elevated status in their families and communities. In contrast, when societies are modernised by industrial modes of production, there are predictable socio-economic changes that diminish the status of the older people while increasing the status of younger age groups.
When Cowgill and Holmes (Reference Cowgill and Holmes1972) developed their version of modernisation theory, they relied on evidence collected from 14 diverse societies, which revealed that the status of older people was indeed lower in more modernised societies. Even though modernisation theory has become an integral part of the sociology of ageing theories (Aboderin, Reference Aboderin2004; Alley et al., Reference Alley, Putney, Rice and Bengtson2010), only a modest number of empirical studies have put the theory to the test. The existing studies reveal conflicting results which may be due to different operationalisations of older people's social status.
Empirical research in the form of small-scale cross-cultural studies examining perceptions of ageing seemed to provide support to Cowgill and Holmes' main claims (e.g. Bengtson et al., Reference Bengtson, Dowd, Smith and Inkeles1975). However, in a more extensive study including 31 countries, Palmore and Manton (Reference Palmore and Manton1974) found that the association between employment status of older people and the societal level of modernisation followed a curvilinear relationship instead of a linear negative association: objective aspects of older people's status did indeed decline in the early stages of modernisation; however, they seemed to increase again in more advanced stages of modernisation. A more recent study examining the perceived social status of older people with data from 25 European countries also showed that there is a positive association between modernisation and the perceived social status of older adults in countries that have progressed beyond the early stages of modernisation (Vauclair et al., Reference Vauclair, Marques, Lima, Bratt, Swift and Abrams2015). One of the explanations offered for this finding is that the funds and efforts allocated to support older people (e.g. in the form of retirement benefits, policies against age discrimination in employment) increase in more advanced stages of modernisation, which allow older people to maintain their status at higher levels.
In sum, these studies suggest that the link between modernisation and older people's status may be affected by country specifics related to its culture and history. These so-called boundary conditions (Busse et al., Reference Busse, Kach and Wagner2017) need to be considered to arrive at a more accurate and refined version of modernisation theory. Yet the aforementioned studies did not examine these boundary conditions on a large scale, they employed a small number of countries, outdated data or relied on data from the European region (which is highly modernised in general and where differences may often occur due to a cleavage between post-communist and all other countries in the sample). Moreover, the literature is inconsistent in the operationalisation of older people's social status, with some studies examining older people's objective status (e.g. Palmore and Manton, Reference Palmore and Manton1974) and other studies confusing social status with related constructs, such as perceptions and attitudes to ageing (De Tavernier et al., Reference De Tavernier, Naegele and Hess2019), whereas perceived social status itself was only occasionally studied. Hence, the aim of the current study is to follow up on the seminal study of Vauclair et al. (Reference Vauclair, Marques, Lima, Bratt, Swift and Abrams2015) on the perceived social status of older people and explore whether the tenets of modernisation theory apply in a much more diverse cross-sectional sample of countries, with a particular focus on whether the theory holds in different cultural regions.
Background
Social status is a composite characteristic of the hierarchical position of a group in society, or ‘perceived, and in some degree accepted, social [inequality] among individuals that reflects not their personal qualities but rather the “social honour”’ (Chan and Goldthorpe, Reference Chan, Goldthorpe and Chan2010: 11). It is customary to measure social status through subjective evaluations by the general population (see e.g. occupational prestige research; Ganzeboom et al., Reference Ganzeboom, De Graaf and Treiman1992). Developed as a fundamental idea in classical sociology (e.g. Weber, Reference Weber1978), the concept of social status is still useful, in particular to describe the general position of the older adults in a society.
Modernisation involves a number of societal processes that have been in force throughout the 20th century and continue to affect social statuses of various groups across most countries in the world today. Cowgill (Reference Cowgill1986: 191) identified several processes that happened in the course of modernisation which he considered to be the most important factors determining the social status of older adults. These included the level of economic development, urbanisation, education, and development of health technology. Note that this understanding of modernisation deviates from the mainstream (e.g. by Inglehart, Reference Inglehart1997) and represents a composite of specific modernisation-related phenomena. Modernisation is used in this narrow sense hereafter.
In line with Cowgill, the central role in defining the status of social groups is their position in the labour market. A developed economy yields new requirements in the labour market in which mobility and flexibility are favoured, thereby giving younger people an advantage. Mass education affects mostly younger people, rendering older adults less competitive. Moreover, compared to an illiterate society where older adults are the main source of knowledge, this role is lost in more educated societies. The development of health technology caused the demographic transition towards lower mortality which was strikingly universal across countries and manifested itself in a series of large social changes (Casterline, Reference Casterline, Demeny and McNicoll2003). Most importantly, countries that in addition experienced decline in fertility showed a growing share of older adults which in turn increased competition for jobs. The competition can be partially resolved via social policies such as legal retirement ages, yet this further diminished older adults' status by excluding them from the labour market. Finally, urbanisation loosens family ties and, for example, excludes older adults from the support of their extended families. A review of all these modernisation processes strongly suggests that modernisation adversely affects the status of older adults.
The initial version of the theory stated that the relation between modernisation and the status of older adults is strictly negative, but later accounts suggested that it can be curvilinear: the status is low in pre-agrarian societies, comes to its peak in agrarian societies, then falls at its lowest during industrial ages, and recovers somewhat in the post-industrial context (Palmore and Manton, Reference Palmore and Manton1974; Williamson et al., Reference Williamson, Evans and Powell1982; Cowgill, Reference Cowgill1986). Indeed, in a sample of post-industrial European countries, Vauclair et al. (Reference Vauclair, Marques, Lima, Bratt, Swift and Abrams2015) demonstrated that the association between modernisation and social status of older adults is positive. Yet, this study did not account for the cultural factors that could also affect the perceived social status of older people.
The literature offers only a few insights into cultural aspects of ageism. In their meta-analysis, North and Fiske (Reference North and Fiske2015) concluded that attitudes towards older people were more negative in ‘the East’ compared to countries of ‘the West’. Confirming this result, Peterson and Ralston (Reference Peterson and Ralston2017) found with data from the World Values Survey (WVS) that respondents in East Asian and Middle Eastern countries viewed older people as being a higher burden to society and receiving an unfair share from their government compared to respondents from ‘Western’ countries. Using survey data from students in the United Kingdom and Taiwan, Vauclair et al. (Reference Vauclair, Hanke, Huang and Abrams2017) provided a more nuanced picture: normative perceptions, such as the social status of older people, as well as perceived competence and admiration were more positive in Taiwan. However, personal attitudes tapping into prejudice towards older people were also stronger in Taiwan.
This suggests that attitudes towards older people, stereotypes and their social status might intermingle in different ways across cultures. Consequently, when studying the social status of older adults, culture can play an important role. At the same time, a study in Chinese cities demonstrated that modernisation was negatively associated with expressions of filial piety and material support of older parents, suggesting that modernisation can overrule cultural differences (Cheung and Kwan, Reference Cheung and Kwan2009).
Although studies of attitudes and stereotypes are interesting in the context of ageing studies, they tell us very little about the social status of older adults across the world (see also De Tavernier et al., Reference De Tavernier, Naegele and Hess2019). A literature search with the keywords ‘social status & (aging|ageing) & (aged|older|old) & countries & meta-perceptions’ in Google Scholar showed 105 results, however, when we filtered them by the presence of empirical cross-national comparisons and a relevant measure of social status (meta-perceptions), there were only seven left. With a few exceptions, the topic has been basically abandoned since the 1970s and 1980s. Hence, modernisation theory was put to a test rarely and, even less so, were its potential cultural boundary conditions.
Current study
In the current study, we rely on a large cross-sectional dataset that includes responses of participants from all the continents. Cowgill theorised that modernisation has a causal effect on the status of older adults and that contemporary differences in levels of modernisation between different societies can actually provide a glimpse into societal changes over time. However, our data are cross-sectional and we are only able to test correlational associations between country modernisation and the perceived social status of older people. Nevertheless, this study makes an important contribution to the literature by putting Cowgill's theory to the test with a large-scale cross-national dataset.
Due to the scarcity of empirical studies on this topic, we build our hypotheses directly on Cowgill's (Reference Cowgill and Gubrium1974, Reference Cowgill1986) theory and the recent evidence at hand (Vauclair et al., Reference Vauclair, Marques, Lima, Bratt, Swift and Abrams2015).
Following Cowgill's theory of modernisation, we expected that:
• Hypothesis 1a (H1a): Country-level modernisation is generally negatively related to the social status of older adults.
• Hypothesis 1b (H1b): There is a curvilinear association where the link with the social status of older people becomes less negative or even positive among the most modernised countries.
The second general hypothesis derives from previous evidence that culture also plays a role in shaping perceptions of older people. Yet, the dearth of cross-cultural studies on ageism as well as the heterogeneity in the conceptualisation and operationalisation of ageist perceptions render the literature rather inconclusive. Hence, we abstained from making specific predictions and just proposed a general cultural moderation hypothesis:
• Hypothesis 2 (H2): Deviations from the general association between modernisation and the social status of older adults can be attributed to distinct cultural zones.
Data and methods
Data
The analyses are based on the sixth wave of the WVS with data collected between 2010 and 2014 in 60 countries (Inglehart et al., Reference Inglehart, Haerpfer, Moreno, Welzel, Kizilova, Diez-Medrano, Lagos, Norris, Ponarin and Puranen2014). WVS is a multinational survey that measures attitudes and values across the world with more than 100 questions and uses large representative national samples. So far, the sixth wave was the only one that included questions on the perceived social status of age groups. The analyses excluded Hong Kong and Taiwan for which country-level data were not available. Sample sizes, average age, percentage female and percentage missing by country are listed in Section A in the online supplementary material. After excluding 13.7 per cent of cases due to missing values, the sample included 78,026 individual responses from 58 countries.
Individual-level measures
The key outcome variable is perceived social status of people in their seventies (hereafter referred to as PSS70) measured with the following question: ‘How [do] you think most people in this country view the position in society of people … over 70?’ The response scale ranged from 1 (extremely low position in society) to 10 (extremely high position in society). As mentioned above, this method is typically applied for measuring social status (Ganzeboom et al., Reference Ganzeboom, De Graaf and Treiman1992).
Since the social status indicates the relative position in a society, it is important to ensure that evaluations of PSS70 are indeed relative. Hence, we included the perceived social status of people in their forties (measured in the same way as PSS70 but referring to people in their forties) as a covariate in the regression analyses. In a similar vein, the descriptive analysis included an adjusted PSS70 computed as residuals of the PSS70 regressed on the evaluation of status of people in their forties. All the regression models controlled for respondents' age, gender and education (nine categories), as well as income (within-country decile groups). All the individual-level variables were standardised at the whole sample.
Country-level measures
Following Cowgill's (Cowgill, Reference Cowgill1986) definition of modernisation, the key independent variable was the level of modernisation operationalised by its relevant consequences, namely a country-level index composed of national economic development, urban population, levels of education and health technology development.
Economic development was indicated by a logarithm of Gross Domestic Product per capita measured as purchasing power parity in current international dollars (GDPpc). Level of urbanisation was the percentage of urban population in country. Average schooling level was indicated by the school-life expectancy in tertiary education. It reflects an estimated number of years that country nationals of a specific age are expected to spend in tertiary education. Health technology development was measured by its consequences as the percentage of population aged 65+. There is some controversy in measuring the consequences of health technology development. In contrast to Vauclair et al. (Reference Vauclair, Marques, Lima, Bratt, Swift and Abrams2015), who used life expectancy, the share of older people corresponds more closely to Cowgill's more recent understanding of this pillar:
Modernized societies have older populations, that is, higher proportions of old people … The status of the aged is highest when they constitute a low proportion of the population and tends to decline as their numbers and proportions increase. (Cowgill, Reference Cowgill1986: 189)
This approach is similar to the United Nation's Human Development Index, which also combines data on economic development, educational involvement and demographic modernisation (United Nations Development Programme, 2020).
All country indicators are from 2012 or the closest year available, to correspond roughly to the year of data collection for the WVS. They were obtained from the World Bank database (World Bank, 2021) and standardised before the analyses. A principal component analysis detected a single strong component explaining 93 per cent of total variance based on which a corresponding summative index was calculated. Both Cronbach's alpha and McDonalds' omega were 0.83.
To examine cultural differences, we categorised countries into cultural zones as indicated in Inglehart (Reference Inglehart1997), but due to a smaller set of available countries, some zones were combined, resulting in seven groups: Latin America, New World, South-East Asia, sub-Saharan Africa, Muslim and post-communist countries (excluding Slovenia)Footnote 1 and non-post-communist Europe. Some countries were part of more than one cultural zone (see Section A in the online supplementary material).
Analytical strategy
First, we describe the bivariate relations between country PSS70 and country-level modernisation. Next, we applied regressions in order to examine associations between country-level modernisation and its components with PSS70. Given the nested structure of the data, multilevel regressions were run (Hox et al., Reference Hox, Moerbeek and van de Schoot2017). Such models correct standard errors for non-independence among respondents within each country. More specifically, we applied a random intercept model. Since the main interest was at the country level, we did not include random slopes. We tested the overall effect of the modernisation index as well as each of its components as predictors of PSS70 separately in order to gauge whether different components vary in their associations with PSS70. Finally, a set of additional analyses tested the robustness of the results. The data used in this analysis are available from the WVS website. The R code is available at the Open Science Framework (https://osf.io/g2kh5).
Results
Bivariate relations
Figure 1 illustrates the relations between the PSS70 and country's level of modernisation. The PSS70 is indicated by country means of residuals as obtained from a regression model which predicted social status of people in their seventies with the social status of people in their forties. The scatterplot already points to cultural specificities: first, it shows that the highest PSS70 is in the cultural zone comprised of predominantly Muslim countries, including countries as diverse as Kuwait, Malaysia, and Uzbekistan. In contrast, the lowest PSS70 is found in post-communist countries, with Estonia, Russia, and Ukraine occupying the bottom of the graph. Second, the figure also demonstrates that the overall association between the modernisation index and PSS70 is clearly negative, r(56) = −0.57, p < 0.001. Yet, there are also clear differences in the association across cultural zones. Among Muslim countries the PSS70 does not seem to be associated with their modernisation level, r(18) = −0.03, p = 0.910, whereas the association found in post-communist countries is even more strongly negative than the average, r(10) = −0.91, p < 0.001.
Association of modernisation index and the relative perceived social status of people in their seventies (PSS70) compared to the status of people in their forties.
Multilevel regressions
As indicated by the intraclass correlation coefficient, 14 per cent of the total variance of PSS70 is due to differences between countries. This value is relatively high, pointing to the importance of considering country differences when examining the perceived social status of older people.
First, we tested the role of modernisation in explaining country differences in PSS70. The analytical strategy consisted of examining the modernisation index as a predictor, then each of its components, its squared terms and, finally, all components together to inspect their predictive power and determine if earlier results were confounded by each other's effects. Table 1 shows that the modernisation index had a strong and highly significant negative effect on PSS70 (model M1). Moreover, the components of this index (except for urbanisation) showed substantial negative associations with PSS70 when entered into the model separately (M2–M5). The squared term of the modernisation index was small and non-significant; however, the squared terms of GDPpc, percentage of population aged 65+ and level of urbanisation, when entered separately, had significant and positive coefficients which suggests the presence of a curvilinear association in which the negative link with modernisation becomes weaker among the most-modernised countries.
Standardised fixed effects of multilevel regression of perceived social status of people in their seventies on facets of modernisation
Notes: The model includes individual-level controls: perceived social status of people in their forties, age, gender, education level and income. GDPpc: Gross Domestic Product per capita measured as purchasing power parity in current international dollars. AIC: Akaike information criterion. BIC: Bayesian information criterion.
Significance levels: * p < 0.05, ** p < 0.01, *** p < 0.001.
When all four components of the modernisation index were entered together in the same model (M6), most of them became non-significant except for the share of people aged 65+. Moreover, its squared term remained significant as well (M7). Bayesian and Akaike information criteria are the smallest for M3, confirming that the share of people aged 65+ is mostly responsible for the significant effects of modernisation.Footnote 2
The predictive power differed widely across predictors. GDPpc and urbanisation made the smallest contributions (R 2 = 0.24 and 0.25, respectively), whereas years of schooling explained 44 per cent of between-country variance and the share of people aged 65+ explained 60 per cent of variance. The latter two surpassed the predictive power of an overall modernisation index (R 2 = 0.40).
To sum up, the results supported H1a, but mostly due to a single component of the modernisation index, the share of population aged 65+. Its association with PSS70 was strong and negative and only weakly curvilinear, becoming slightly less negative across countries with the highest shares of older adults in the population.
Next, we tested H2 by regressing PSS70 on the seven cultural zones and the modernisation index. The results are presented in Table 2. M9 included only the cultural zones as predictors and showed that the lowest PSS70 was in the post-communist countries, the highest in the Muslim zone and sub-Saharan Africa, and the remaining cultural zones did not differ from each other. M10 added the modernisation index which showed again a significant negative association with PSS70. The differences between cultural zones, except for the Muslim and post-communist ones, became non-significant. It implies that modernisation did not account for all the differences between countries; the cultural differences of the Muslim and post-communist countries have an independent link with PSS70. In M11 and M12, we only used these two cultural zones and merged all the others into a single reference group.
Standardised fixed effects of multilevel regression of perceived social status of people in their seventies on facets of modernisation
Notes: The model includes individual-level controls: perceived social status of people in their forties, age, gender, education level and income. AIC: Akaike information criterion. BIC: Bayesian information criterion.
Significance levels: * p < 0.05, ** p < 0.01, *** p < 0.001.
Figure 1 suggested different associations of PSS70 with modernisation across cultural zones. We formally tested this by introducing an interaction term between the two cultural zones and the modernisation index as well as its components. The respective model (M11) shows that the interaction effect between the modernisation index and post-communist cultural zone was significant and negative, implying that modernisation had a stronger negative association with PSS70 within this group than in most other countries in the sample. The interaction between the Muslim cultural zone and modernisation was positive, meaning that the association of PSS70 with modernisation was weaker in this cultural cluster. Figure 2 illustrates these differences: the modernisation index had a weak negative link in Muslim countries (η 2 = 0.14), whereas there was a strong negative link in post-communist countries (η 2 = 0.51) and a moderate negative association was present in the remaining countries (η 2 = 0.30). Additionally, we tested interaction effects of cultural zones with each of the components of the modernisation index separately, controlling for the share of the population aged 65+ (see Section C in the online supplementary material). None of the interaction effects were highly significant.
Interaction between the two cultural zones and the modernisation index in their effects on social status of people in their seventies (predicted by model M11).
Finally, M12 tested if the coefficient of the share of population aged 65+ reported above remained significant after controlling for the two cultural zones. Although the squared term of the share of population aged 65+ vanished, its main effect remained significant. Interestingly, the coefficients of the two cultural zones also decreased substantially, indicating that at least some of the differences between them were due to differences in the share of older adults. The other indicators of modernisation were tested in the same manner (see Section D in the online supplementary material), and the linear associations of all the predictors were negative and significant, but their squared terms were relatively low and became non-significant. At the same time, the coefficient of the two cultural zones remained significant, suggesting that the general trend of the association is linear and adjusted by specificities of the Muslim and post-communist countries. This confirms both H1a and H2.
Robustness tests
To ensure that the associations we found were not the result of some confounding factors, we ran several robustness tests which are detailed in Section E in the online supplementary material. Alternative indicators of economic development (namely gross national income measured as purchasing power parity and with the Atlas method) were significant predictors when entered separately but had a zero effect when entered together with the share of population aged 65+. Likewise, alternative indicators of country's level of education (population enrolment in secondary and tertiary education, government spending on secondary and tertiary education as percentage of GDP) were non-significant after controlling for the share of population aged 65+.
The coefficients of the country share of employed older adults which was found to interact with modernisation measures across countries in the European region (Vauclair et al., Reference Vauclair, Marques, Lima, Bratt, Swift and Abrams2015) did not show significant associations with the status of older people – neither before nor after controlling for the components of the modernisation index and the two cultural zones. Government spending on older adults as percentage of GDP had a negative and significant coefficient when entered by itself, with the GDPpc or with the two cultural zones, but lost its significance when entered together with the modernisation index, the population aged 65+ or the level of education. Country life expectancy (as an alternative to population aged 65+) was not significant when controlling for the modernisation index or its components.
Earlier we mentioned that Slovenia was excluded from the post-communist zone as an influential outlier. Including this country in the post-communist zone only slightly changed the main effects of the post-communist zone, however, its interactions with the modernisation index became substantially smaller.
Finally, we excluded other potential outliers, namely Germany and Japan, identified in the descriptive analysis (figure in Section B in the online supplementary material). This increased the correlation between PSS70 with the modernisation index from −0.57 to −0.59; with the share of people aged 65+ from −0.70 to −0.76. The multilevel regression results without these two countries showed virtually the same results. To summarise, we found that the robustness tests fully confirmed the results described above.
Discussion and conclusion
Cowgill's (Reference Cowgill and Gubrium1974, Reference Cowgill1986) modernisation theory provided powerful explanations for the decline in the social status of older adults around the world. Still, only a few empirical studies directly addressed the main claims of the theory. The current study provided a correlational large-scale test of the theory by examining the relative social status of people over the age of 70 across 58 countries from different regions of the world.
In line with Cowgill's theory and our hypothesis H1a, PSS70 was strongly and negatively related to the level of modernisation. Different controls did not change this result. The curvilinear association was not supported for the overall modernisation index, and only marginally supported for some of its components, and completely vanished when two specific cultural zones (Muslim and post-communist) were considered. Thus, an overall curvilinear relation between modernisation and PSS70, as claimed by our hypothesis H1b and Palmore and Manton (Reference Palmore and Manton1974), was not supported. One reason for the differences in results might be the differences in operationalising older people's status. In particular, Palmore and Manton measured observed status, including participation in workforce and attained level of education, whereas our study focused on the subjectively perceived social status of older adults by individuals from the same country. Even though these two ways to measure social status are complimentary, the subjective approach better reflects overall prestige and social status and is not necessarily limited to education and occupation.
Another reason for the discrepancy in results might be due to the fact that the current study accounted for cultural differences, which was not considered in previous studies. Moreover, there was a relatively high proportion of less-developed countries in our sample and only a few highly developed countries, especially from Europe, which might have rendered it more difficult to detect the curvilinear associations that have been proposed previously. This might also explain why the current study results are different to the findings of Vauclair et al. (Reference Vauclair, Marques, Lima, Bratt, Swift and Abrams2015), which detected a positive association between modernisation and PSS70 in European countries. When considering a subset of the WVS European countries, the association was indeed positive but non-significant, r(10) = 0.17, p = 0.611. Thus, the lack of a curvilinear relation might have occurred simply because there were too few highly developed European countries in the WVS dataset. This emphasises that the negative association between modernisation and PSS70 currently dominates in the world, and it might be overturned only in a few highly modernised countries. Therefore, although Cowgill's theory is a weak explanation of the status of the older adults within Europe, it is well supported globally.
The share of older people
Our analysis demonstrated that the significance of the modernisation index was mostly due to the share of older adults in the population. The share of older adults had the strongest association with PSS70 even compared with the modernisation index itself. On the contrary, urbanisation, economic development and education were not significant when controlling for the share of older adults.
The questions remain why the importance of the share of older people is the main predictor of the perceived status of older people in society. One explanation might be derived from threat-based theories of prejudice (e.g. Stephan and Stephan, Reference Stephan and Stephan2000): countries with higher numbers of older adults also have higher levels of competition between generations. This can create perceptions of threat for in-group resources or processes, for instance in the form of economic burdens to provide for the increasing share of older people, which then translates into a hostile form of ageism (North and Fiske, Reference North and Fiske2015) and, therefore, a devaluation of older people in the form of lower perceived social status. Yet we cannot be sure about the proposed relations between stereotypes, attitudes, and social status of older adults at the country level which is yet to be examined.
Interestingly, the share of older people as a predictor remained significant even after controlling for overall life expectancy, whereas life expectancy itself was not significant. It means that the status of older people is specifically associated with population ageing and is not a matter of people on average living longer.
These results might also be interpreted in terms of the theory of diminishing marginal utility. It is conceivable that any society demands a contribution from older adults. The scarcity of older adults renders them more valuable to the labour market as well as increases their symbolic capital. Hence, they are awarded higher social status. It is also evident that the ‘supply’ of older people differs across countries and time but seems to grow faster than the ‘demand’. Given our results, it is conceivable that the status of older people is determined mostly by the higher ‘supply’ rather than by a low demand. In other words, a higher number of older adults is inconsistent with the relatively stable societal need for them. This highlights a disproportion between the global and rapid processes of ageing and societal structures that effectively accommodate the population of older adults, failing their integration into the workforce and missing opportunities for their potential contribution. As a result, it affects older people's social status and, consequently, most likely their wellbeing (cf. Demakakos et al., Reference Demakakos, Biddulph, Oliveira, Tsakos and Marmot2018).
Although Cowgill (Reference Cowgill1986: 194) hinted at ‘an element of supply and demand’ of older people in his theory, he nevertheless considered other aspects of modernisation to be more important. In contrast, our results suggest that the share of older adults has the strongest association with their social status. In some way, this calls Cowgill's version of modernisation theory into question. It might be that the supply/demand model is self-sufficient even without considering other pillars of modernisation. One might even speculate that the social status of older adults, determined by the supply/demand disproportion, causes the perceptions of threat, burden or spread of stereotypes, whereas these perceptions and stereotypes function to reproduce and reinforce an already existing social order (rather than determine status change). Future research could aim to examine these causal pathways.
Cultural differences
Our results also demonstrate the explanatory role of culture in country differences in PSS70. It has been argued before that ageist perceptions are inextricably linked with cultural values (North and Fiske, Reference North and Fiske2015; Vauclair et al., Reference Vauclair, Hanke, Huang and Abrams2017) and our results provide evidence that this is also true for PSS70. Even though most differences across cultural zones were due to differences in modernisation, the predominantly Muslim and post-communist countries demonstrated a substantial deviation from the global trend. In particular, in Muslim countries the status was higher than in most countries, and higher than the modernisation level would predict it to be. The association between modernisation and the status of the older adults was much weaker across Muslim countries. One reason for that might be a higher overall religiosity among Muslims and a literal interpretation of the Quran that puts a moral obligation on children to provide material support and respect for their parents and older relatives (Elsaman and Arafa, Reference Elsaman and Arafa2012), which means that the older adults are reserved a special place in the family and are afforded higher spiritual status with advancing age (Johnson, Reference Johnson, Bengtson, Kirkwood and Johnson2005). This cultural imperative is also known as filial piety but has been mostly associated with East Asian cultures. In contrast, our results indicated that PSS70 in East Asian countries was well aligned with their corresponding level of modernisation, and this concurs with existing evidence from China (Cheung and Kwan, Reference Cheung and Kwan2009). In fact, all major religions recognise filial obligation as an important moral virtue, but more traditional cultures in which religious doctrines play a greater role may be more likely to continue cherishing filial piety despite the effects of societal modernisation. Some indirect evidence supports this reasoning in regard to Islam in particular (e.g. Abdulrahim et al., Reference Abdulrahim, Ajrouch and Antonucci2015; Latif et al., Reference Latif, Ali and Zafar2019). Given the scarcity of literature on the status of older adults across Muslim and non-Muslim countries, we consider it an important new finding that requires further investigation.
On the other hand, post-communist countries have shown a consistent deviation from the modernisation trend in the direction of lower PSS70. It can be explained by the rapid societal changes that have deteriorated the social and human capital of older generations. Cohn (Reference Cohn1982) had already pointed out that the speed of social change might be even more important for the status of older people than the level of development. The fall of the communist regimes harshly affected human and financial capital of the then-middle-aged and older people, who by the time of the survey in 2014 became those in their seventies or older. The change in economic and political systems made many of their professional and social abilities obsolete, a situation which brought them to the lower steps of the social status hierarchy.
Limitations
One of the limitations of this study is the measurement of social status with a single item and the possibility of different meanings across countries. It is possible that respondents from different cultures might have interpreted ‘social status’ differently, e.g. some thinking of their symbolic capital whereas others might have thought of their access to material wealth. Another limitation is related to convenience sampling of countries, which allows conclusions to be made for only the 58 countries in the sample. Moreover, we closely followed Cowgill's understanding of modernisation which is unusual, as it only involved the aspects that might affect the older population. This limits the generalisation of our results across other modernisation studies. In particular, the percentage of the population aged 65 and higher is definitely a characteristic of modernised societies, but it might not be part of the driving force of modernisation. Therefore, it has a questionable association with a more common understanding of modernisation (e.g. by Inglehart, Reference Inglehart1997). Yet, inclusion of this aspect allowed us to find an important association with the status of older people. Finally, although the hypothesised causal relations are theory-driven, our study does not provide evidence for temporal or causal changes from modernisation towards the status of older adults. There is a possibility that both modernisation and the decrease of status of older people occurred due to some common cause.
Despite these limitations, the study provided rare empirical evidence of the validity of Cowgill's modernisation theory as well as its cultural boundary conditions. We demonstrated the strong link of the share of older people with their social status, and emphasised cultural specificities related to Muslim and post-communist cultural zones. Future studies should further investigate the specific features of the Muslim and post-communist countries that are associated with higher or lower social status of older adults. The suggested theoretical explanation involving supply and demand of older adults and their implications for status perceptions should also be developed and validated in future research.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0144686X22001192.
Author contributions
C-MV suggested the idea for the paper, contributed to the theoretical parts and revised the paper. MR conducted the statistical analysis, wrote the first draft and contributed to further revisions.
Financial support
This work was supported by the Fundação para a Ciência e Tecnologia (FCT) (C-MV, grant number IF/00346/2014). The work of MR is an output of a research project implemented as part of the Basic Research Program at the National Research University Higher School of Economics (HSE University).
Conflict of interest
The authors declare no conflict of interest.
Ethical standards
Ethical approval was not required.
