Internalization of negative societal views on old age into self-perceptions of aging : exploring factors associated with self-directed ageism

This is a self-archived version of an original article. This version may differ from the original in pagination and typographic details. Author(s): Title: Year: Version: Copyright: Rights: Rights url: Please cite the original version: CC BY 4.0 https://creativecommons.org/licenses/by/4.0/ Internalization of negative societal views on old age into self-perceptions of aging : exploring factors associated with self-directed ageism © 2023 Ishikawa. Published version Ishikawa, Motoko Ishikawa, M. (2023). Internalization of negative societal views on old age into self-perceptions of aging : exploring factors associated with self-directed ageism. Frontiers in Sociology, 8, Article 1291325. https://doi.org/10.3389/fsoc.2023.1291325 2023 Frontiers in Sociology 01 frontiersin.org Internalization of negative societal views on old age into self-perceptions of aging: exploring factors associated with self-directed ageism MotokoIshikawa 1,2* 1 Faculty of Social Sciences, University of Helsinki, Helsinki, Finland, 2 Department of Social Sciences and Philosophy, University of Jyväskylä, Jyväskylä, Finland Introduction: A growing number of research has provided evidence for the negative impact of ageism on older people’s health and well-being. Among the three different manifestations of ageism, namely institutional, interpersonal and self-directed ageism, significant ageism-health associations have been proved to bestrongest for selfdirected ageism. This supports stereotype embodiment theory, which maintains that lifetime exposure to negative age stereotypes leads to the internalization of ageism as a form of negative attitudes towards own aging and it adversely affects health and well-being in old age. However, little is known about how people internalize negative age stereotypes held in the society into self-perceptions of aging. Methods: This study aimed to explore how socially shared beliefs about old age are internalized into self-perceptions of aging focusing on uncovering factors related to self-directed ageism. Data were derived from the survey that had examined citizen’s attitudes towards old age and aging in Finland. Multinominal logistic regression models were performed to examine the association of sociodemographic and contextual factors with different combinations of societal age stereotypes and two indicators of self-perceptions of aging: subjective views on old age and personal feelings of own old age. Results: The analyses showed that being female, attaining tertiary education, evaluating poor quality of life and awareness of institutional old age discrimination were related to holding negative views on aging towards both society and oneself. Discussion: The findings from univariate and multivariate models suggest that it is not age per se, but structural and cultural circumstances shaped with growing older that turns socially shared negative age stereotypes into negative self-perceptions of aging. Even though the study addressed situations in one country, the findings have an important implication for other rapidly aging societies regarding how social and cultural contexts are closely linked to the formation of self-directed ageism. KEYWORDS ageism, self-directed ageism, negative age stereotypes, internalization, self-perceptions of aging 1 Introduction Ageism is a multifaceted social phenomenon that is defined as the stereotypes, prejudice and discrimination directed towards others or oneself based on age (World Health Organization, 2021). The three dimensions of ageism signifies that stereotypes affect how wethink, prejudice affects how wefeel, and discrimination affects how weact towards people on the basis of their OPEN ACCESS EDITED BY Todd L. Matthews, Sacred Heart University, UnitedStates REVIEWED BY Anna Wanka, Goethe University Frankfurt, Germany Giorgio Piccitto, Bocconi University, Italy *CORRESPONDENCE Motoko Ishikawa [email protected] RECEIVED 09 September 2023 ACCEPTED 06 November 2023 PUBLISHED 17 November 2023 CITATION Ishikawa M (2023) Internalization of negative societal views on old age into self-perceptions of aging: exploring factors associated with self-directed ageism. Front. Sociol. 8:1291325. doi: 10.3389/fsoc.2023.1291325 COPYRIGHT © 2023 Ishikawa. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. TYPE Original Research PUBLISHED 17 November 2023 DOI 10.3389/fsoc.2023.1291325 Ishikawa 10.3389/fsoc.2023.1291325 Frontiers in Sociology 02 frontiersin.org age (Officer et al., 2020). Ageism can operate either explicitly (consciously) or implicitly (unconsciously), and it manifests itself at three different levels as institutional, interpersonal and self-directed ageism, being intertwined and mutually reinforcing (Marques etal., 2020; World Health Organization, 2021). Ageism is regarded to beprevalent, deeply ingrained and more socially accepted than other forms of bias, and older people are most vulnerable to ageism though it can affect any age group (Ayalon etal., 2019; Officer etal., 2020; World Health Organization, 2021). Over the past two decades, a growing body of research has provided evidence for the negative impact of ageism on older people’s health and well-being. A first global systematic review on the impacts of ageism on health, which included both structuraland individual-level studies from 45 countries, found that ageism led to significantly worse health outcomes: 95.5% of the 422 studies and 74.0% of the 1,159 associations between ageism and health showed evidence of the adverse effects of ageism (Chang etal., 2020). Significant ageism-health associations were observed across 11 health domains: exclusion from health research, devalued lives of older persons, lack-of-work opportunities, denied access to healthcare and treatments, reduced longevity, poor quality-of-life and well-being, risky health behaviors, poor social relationships, physical illness, mental illness and cognitive impairment. These domains represent a broad range of health outcomes, which conforms with World Health Organization’s (2020) definition of health as a state of complete physical, mental and social well-being and not merely the absence of disease or infirmity. The systematic review also found that the association between ageism and health outcomes was strongest for self-directed ageism, that is, when ageism was turned against oneself and operationalized as selfperceptions of aging measure. For instance, negative self-perceptions of aging significantly reduced longevity (Zhao etal., 2017), increased risk of harmful tobacco and alcohol use (Villiers-Tuthill et al., 2016), and impeded recovery from severe disability (Levy etal., 2012). Moreover, a recent study revealed that the connection between COVID-19 health worries and anxiety symptoms was stronger among older adults with high levels of self-directed ageism (Bergman etal., 2020). Other systematic review and meta-analysis of longitudinal studies corroborate the findings of adverse health effects of ageism among older adults. A systematic review, which examined the longitudinal consequences of self-perceptions of aging in participants 50 years or older by synthesizing 21 studies across five western countries, revealed that having more positive self-perceptions of aging was consistently associated with better self-rated health and less obesity, greater longevity, better performance of the activity of daily living, less depression and better cognitive functioning (Tully-Wilson etal., 2021). Self-perceptions of aging (SPA) refer to a person’s general evaluation of own aging process and reflect internalized stereotypes of aging. SPA and its reliable scale used in numerous longitudinal studies―the Attitude Towards Own Aging (ATOA) scale (Lawton, 1975; Liang and Bollen, 1983; Tully-Wilson etal., 2021) ―are regarded as a measure to appraise self-directed ageism. Thus, the results of this systematic review suggest that maintaining positive self-perceptions of aging provides a buffer against the negative impact of ageism on various health outcomes of older people. Likewise, a meta-analysis of 19 longitudinal studies found a small significant overall effect of subjective aging― operationalized either as subjective age or self-perceptions of aging― on health, health behaviors and survival over time (Westerhof etal., 2014). The analyses, however, revealed heterogeneity, with stronger effects on the more proximal outcome of health compared to survival, over a shorter time period, in younger participants in their second half of life, and in welfare regimes with less state provisions. Taken together, previous research including the recent systematic reviews as well as meta-analysis highlights that ageism is a global health issue, which constitutes an important, and hitherto neglected, social determinant of health (Officer etal., 2020; World Health Organization, 2021). 1.1 Theory on ageism-health associations The ample evidence of significant ageism-health associations provides support for stereotype embodiment theory (SET). Drawing on research showing that positive and negative age stereotypes held by older individuals can have beneficial and detrimental effects, respectively, on a variety of cognitive and physical outcomes, SET proposes that age stereotypes are embodied when their assimilation from the surrounding culture leads to self-definitions that, in turn, influence functioning and health (Levy, 2009). There are four components constituting the theory. First, age stereotypes (i.e., beliefs about older people in general) typically begin to develop in childhood and are reinforced and internalized across the life span. Second, these stereotypes can operate unconsciously. Third, as one grows older, age stereotypes become increasingly relevant for the aging person when identifying oneself with old person. Fourth and the finally, these age stereotypes gained salience from self-relevance shape self-stereotypes of aging (Levy, 2003; Wurm etal., 2017) and they exert influence on an array of health outcomes through multiple pathways: psychologically, behaviorally, and physiologically. Accordingly, SET offers a lifespan perspective on the internalization and impact of both positive and negative age stereotypes on individual aging process (Wurm etal., 2017; Tully-Wilson etal., 2021), which occurs in a top-down way from society to the individual as well as over time from childhood to old age (Levy, 2009). Significant associations between ageism and health as well as longitudinal health consequences of self-perceptions of aging identified in aforementioned studies specifically validate the third and fourth components of SET, namely the adverse effect of internalized negative age stereotypes or self-directed ageism on health and well-being in old age. In contrast, much fewer studies have focused on the ways in which age stereotypes, as socially shared beliefs about older people and old age, are incorporated into self-stereotypes. A comprehensive systematic review found a total of 14 determinants of ageism against older people, of which 13 associated robustly with other-directed forms of ageism, and only one had an effect on self-directed ageism (Ayalon and TeschRömer, 2017; Marques etal., 2020). Self-directed ageism was mostly determined by older adults’ mental and physical health status. This finding implies bi-directional nature of the association between selfdirected ageism and health. Meanwhile, another study found lower healthy life expectancy and fewer proportion of older people within a country as explanatory factors for an individual or a country being ageist (Officer etal., 2020). Nevertheless, it is again relevant to ageism on interpersonal level but not to self-directed ageism. To sum it up, little is known about processes between age stereotypes and self-stereotypes, especially in terms of how negative age stereotypes generally held in the society develop self-directed ageism via self-relevance. To date, no robust determinants of selfdirected ageism have been found at the interpersonal/intergroup and institutional/cultural level of analysis (Marques et al., 2020). The scarce knowledge about determinants of self-directed ageism presents Ishikawa 10.3389/fsoc.2023.1291325 Frontiers in Sociology 03 frontiersin.org a major research gap, while the theory maintains that individuals assimilate age stereotypes from the surrounding culture (Levy, 2009), and specifically research has increasingly demonstrated detrimental effect of self-directed ageism on health and well-being. A sociological approach that sheds light on socio-cultural influences on individuals in shaping self-directed ageism is now called for. 1.2 The current study To address the identified knowledge gap, the current study aims to examine under what conditions, including sociodemographic characteristics and socio-cultural contexts, people are susceptible for internalizing socially perceived beliefs about the process of aging and about older people. In other words, it investigates how socially shared beliefs about old age are linked to individuals’ self-perceptions of aging. The research questions are as follows: (1) Does increasing age cause socially shared negative age stereotypes to shape negative self-perceptions of aging among older adults? (2) If older age is not the major predictor of self-directed ageism, what other factors are associated with internalization of negative societal age stereotypes into self-perceptions of aging? This study utilizes dataset from the survey in which participants with a broad age range were asked about their perceptions of age stereotypes shared in the society and self-perceptions of aging. Inquiring views on aging on both societal and individual levels has an advantage in undertaking the research questions of the present study. Besides, two questions introduced in the survey to assess self-perceptions of aging capture multifaceted nature of ageism: one is about respondents’ negative attitudes towards aging and the other refers to fear of aging. According to SET, lifetime exposure to negative age stereotypes leads to the internalization of ageism as a form of negative attitudes towards own aging. Meanwhile, terror management theory maintains that threats of death, physical deterioration, and insignificance evoked by the encounter with older adults manifest in ageism as a fear of aging directed towards others as well as selves (Martens etal., 2005; Lev etal., 2018). Thus, the two questions are both relevant to the manifestation of ageism, but it presumably emerges through different pathways. This suggests that these questions should beanalyzed separately, since determinants of self-directed ageism may differ between the two. Answering to the research questions by analyzing this survey data fills a gap that has been identified in both the theory (SET) and empirical studies. Given the accumulated knowledge on adverse health effects of ageism, the most notable is that results of this study about factors associated with self-directed ageism will highlight potential vulnerable population groups and life situations to which socialand health policy as well as interventions should pay more careful attention. 2 Materials and methods 2.1 Data The survey examining citizens’ attitudes towards old age, aging and older population in Finland was the data source for this study. VTKL – The Finnish Association for the Welfare of Older Adults commissioned the survey to Aula Research Oy. The targets of the survey were Finnish citizens aged 16 years and older and sampling was conducted to correspond to the distribution of target population according to residential region, age, and gender. The survey was conducted between January and March 2022 through online questionnaire for all age groups and by telephone interviews for those who were 85+ years old (Vaarama, 2022). The response rate of the online survey was 19%, whereas 46% of people contacted by telephone responded to the questions. The survey received 2,056 respondents. However, weremoved 30 cases without information on gender or birth year and 29 cases who did not respond to the question on societal views on old age, 10 cases who did not answer the question on subjective views on old age, and 3 cases without response to the question on personal feelings of own old age. Additionally, weremoved one case whose birth year was 1900 and 2 cases whose birth year was 2007 or later. Thus, wehad 1,981 respondents in our final study cohort. Given that recent rise in population aging seems to beassociated with negative attitudes towards aging and older people in a global context (North and Fiske, 2015), Finland is an interesting country for investigating how age stereotypes shared in the society and people’s attitudes towards own aging are related. In Finland, as a country belonging to the Nordic welfare regime, public authorities have been responsible for universal provisions of income security and care services for older people. However, the country has been aging most rapidly in Europe (Eurostat, 2023a), which poses a serious challenge to upkeep the public provisions. Hence, the recent demographic change and its possible negative consequences in old age social policies may influence both societal and self-perceptions of aging. 2.2 Outcome variables To examine the research questions on how socially shared beliefs about old age are linked to individuals’ self-perceptions of aging, this study created new outcome variables by merging values of two questions. The followings explain the original variables on societal age stereotypes and self-perceptions of aging and how these were integrated into the new variables. 2.2.1 Age stereotypes shared in the society Age stereotypes were assessed in the survey as people’s perceptions of societal views on old age by asking participants as follows: ‘How positively or negatively do yousee old age is regarded in Finnish society in general?’ As the present study focuses on investigating negative perceptions, response alternatives with a six-level Likert scale were dichotomized as negatively (‘very negatively’ and ‘somewhat negatively’) and positively or neutrally (‘very positively’, ‘somewhat positively’, ‘neither positively nor negatively’ and ‘cannot say’). 2.2.2 Self-perceptions of aging Self-perceptions of aging were measured in the survey using two questions: one asked participants’ attitudes towards aging in a relatively general manner and the other explored more explicitly personal feelings of own old age. The first question, ‘How positively or negatively do yousee old age?’ was intended for assessing subjective views on old age and this was inquired before the previously mentioned question on societal views on old age. The same six response alternatives as societal age stereotypes (from ‘very positively’ to ‘very negatively’) were recoded similarly in a dichotomous manner. The second question focused more on participants’ personal feelings regarding fear of aging: ‘When youthink of your own old age, what Ishikawa 10.3389/fsoc.2023.1291325 Frontiers in Sociology 04 frontiersin.org does it feel like?’ To examine negative perceptions, five-level Likert scale for response was dichotomized as feel fear (‘very fearful’ and ‘somewhat fearful’) and feel safe or neutral (‘very safe’, ‘somewhat safe’ and ‘neither safe nor fearful’). 2.2.3 Creating two outcome variables: combinations of societal age stereotypes and self-perceptions of aging Next, the two outcome variables were created for the present study by merging previously described dichotomous variables. The variable on societal age stereotypes was merged with each variable on selfperceptions of aging: one with subjective views on aging (attitudes towards aging) and another with personal feelings of own old age (fear of own old age). The combined new variables had four categories: (1) those who take a positive/neutral stance towards both societal views on old age and self-perceptions of aging, (2) those who have negative societal views on old age, while self-perceptions of aging are positive/neutral, (3) those who have positive/neutral societal views on old age, while selfperceptions of aging are negative and (4) those who take a negative stance towards both societal views on old age and self-perceptions of aging. Respondents located in the last category refer to those who internalize negative age stereotypes in the society into self-perceptions. 2.3 Explanatory variables Explanatory variables included the following demographic and socioeconomic characteristics: gender, age in ten-year groups, living arrangements (‘living with a partner’ or ‘not living with a partner’), highest attainment in education (‘basic education’, ‘secondary education’ or ‘tertiary education’) and income level (‘enough money for daily needs’ or ‘not enough money or reluctance to answer’). The cumulative advantage –disadvantage perspective has theoretically and empirically explained the interindividual divergence in a given characteristic (e.g., money, health, or status) with the passage of time, which reproduces a greater heterogeneity and inequality in late life (Dannefer, 2003). Thus, these explanatory variables were expected to berelated to self-directed ageism shaped over the course of one’s life. Besides, gender is especially important in studies on ageism, as the so-called ‘double standard of aging’, referring to discrimination against older women both because of their gender and their age, has generated different meanings of aging for women and men (Sontag, 1978; Arber and Ginn, 1995). Self-rated quality of life (‘very good’, ‘good’, ‘neither good nor bad’, ‘bad’ or ‘very bad’) was also included as an explanatory variable. Given the socio-cultural influence on shaping age stereotypes (Levy, 2009), participants’ opinions about whether following institutions are discriminating against older people were incorporated in the analysis: politicians, media, health care services (e.g., doctor’s appointment), social services (e.g., home care, service housing, meal service), the Social Insurance Institution of Finland, private companies (e.g., shops, pharmacies, restaurants), transport and employers. 2.4 Analysis Multinomial logistic regression models were performed to examine the association of participants’ sociodemographic backgrounds, quality of life and opinions on old age discrimination with the two outcome variables (Kleinbaum and Klein, 2002). First, separate logistic models were fitted for each explanatory variable to analyze their unadjusted associations with the outcome variables (Model 1). Since all explanatory variables were categorical with k classes, they were transformed as k-1 dummy variables. For example, for education the equation is defined as: logitx xx secondary tertiary πβ ββ () () =+ + 01 2 (1) where logit(π(x)) is the logarithmic transformation of the expected value of binary outcome variable, β 0 is the intercept of linear predictor and β 1 and β 2 are the regression coefficients for the secondary and tertiary education dummy variables, respectively. The regression coefficients were transformed as relative risks (RR) thus describing the probability of having the value 1 of outcome variable for secondary and tertiary education as compared to basic education used as the reference category. Further, 95% confidence intervals (CI) for RRs were calculated using standard errors. After that a full logistic regression model was fitted including all explanatory variables simultaneously (Model 2) to study how the mutual adjustment for other explanatory variables affects these associations. In the case of p explanatory variables, the equation is: logitx xx xx pp πβββ ββ () () =+ +++… 0112233 (2) Multinomial logistic regression model can beextended to have an outcome variable with k categories estimating k-1 sets of estimates. Since the outcome variables had 4 classes, three sets of regression estimates were estimated. In this case, RRs describe the probability to belong in a certain category of outcome variable as compared to the reference categories of both the outcome and explanatory variables. The models were estimated using Stata/SE 17.0 for Windows statistical software (StataCorp, College Station, TX, UnitedStates). 3 Results 3.1 Sample characteristics and share of having negative societal views on old age and having negative self-perceptions of aging Table1 displays characteristics of the total sample according to categorized groups and share of having negative societal views on old age and having negative self-perceptions of aging in respective groups of the sample. These are background information for the multinominal logistic regression models explained later. The share of those who had negative subjective views on old age and those who felt fear about own old age was smaller than the share of those who negatively perceived societal views on old age in almost all groups. This supports the argument that people endorse certain but not all socio-cultural age stereotypes (Wurm etal., 2017). It is also noteworthy that while older participants, except the oldest old, had more negative perceptions on how old age was regarded in the society, those who had negative subjective views on aging and were fearful about own old age tended to decrease as age of the respondents increased. Ishikawa 10.3389/fsoc.2023.1291325 Frontiers in Sociology 05 frontiersin.org 3.2 Association of different factors with combinations of societal age stereotypes and subjective views on old age Table 2 presents the results of multinominal logistic regression analyses that investigated the association of explanatory variables with combinations of societal age stereotypes and subjective views on old age. Reference category was those who took a positive/neutral stance towards both societal views on old age and subjective views on old age. The results show the likelihood of being placed in other three categories (negative societal views–negative subjective views, negative societal views– positive/neutral subjective views, positive/neutral societal views–negative subjective views) in respect of respondents’ characteristics compared with the reference category. Model 1 refers to univariate analysis in which each variable was included into the model separately, whereas multivariate analysis controlled for all variables is shown in Model 2. The primary focus of this study was to examine who had negative perceptions of societal views on old age and hold negative subjective views on old age as well. Multivariate analysis controlled for all variables revealed that respondents who were female, had the highest educational level, rated quality of life as worse, and saw politicians, social services, private companies and employers as perpetrators of ageism were more likely to think that old age was regarded negatively in Finland and were themselves liable to view aging in a negative light. It can beargued that people with such characteristics are susceptible for internalizing negative societal age stereotypes into their own perceptions about aging. However, as the table presents, other categories of the outcome variable also had many predictors with statistical significance. Hence, the following subsections describe findings to highlight the role of each factor in the associations. 3.2.1 Gender Being female increased the risk of considering both societal and subjective views on old age as negative, as well as perceiving negative societal attitudes towards aging while holding positive/neutral subjective views on old age. Multivariate analysis controlled for all variables even strengthened the gender effect for these categories. However, gender was not related to the likelihood of viewing old age negatively on a personal level despite having positive/neutral societal views on aging. The findings indicate that negative attitudes among females were directed mainly to societal beliefs about aging rather than to their own views on old age. 3.2.2 Age Respondents aged between 56 and 75 years compared with the youngest age group were more likely to perceive aging negatively on both societal and individual levels. Age-related changes in one’s life course such as retirement, approaching old age, and care needs for close relatives and/or oneself might make respondents think in this manner. However, adjusting for other variables attenuated association effect of age. Older age also increased the risk of having negative societal views on aging while subjectively viewing old age in a positive/ neutral light. This applied to the respondents aged 46 years and older except the oldest old. Multivariate analysis controlled for all variables kept the significant association for age between 66 and 85. Conversely, both univariate and multivariate analyses revealed that older respondents were less likely to negatively view old age on a personal level despite holding positive/neutral societal views on old age. Age effects on all categories of the outcome variable suggest that older people were inclined to shape negative perceptions of aging towards society and much unlikely towards themselves. 3.2.3 Living arrangements Whether the respondents lived with a partner or not was generally not related to different combinations of societal age stereotypes and subjective views on old age. 3.2.4 Education Compared to those who had only basic education, respondents with tertiary education were more likely to perceive old age as negative on both societal and personal levels, as well as to have negative societal views on old age while subjective views were positive or neutral. Adjusting for other variables even increased the association effect of the highest education for these categories. However, educational level was not related to the category of having positive/neutral societal views on old age and simultaneously holding negative subjective views on old age. The results indicate that negative attitudes towards aging among people with tertiary education were directed mainly to the society rather than to themselves. 3.2.5 Income level When income level was analyzed separately from other variables, living under insufficient economic situation increased the risk of belonging to the two categories: negative societal views on old age– negative subjective views on old age, and negative societal views on old age–positive/neutral subjective views on old age. However, adjusting for other variables attenuated association effect of income level. 3.2.6 Quality of life Among different explanatory variables, poor quality of life appeared to bemost robustly associated with respondents’ negative attitudes towards aging on both societal and individual levels. The multivariate analysis revealed that the risk of having such way of thinking was 15 times higher for those who evaluated quality of life as bad and 21 time higher for those with very bad quality of life, compared with people with very good life quality. Poorer self-rated quality of life was significantly related to the other two categories too, but the risk was higher for negatively perceiving subjective views on old age while having positive/neutral societal view on aging. This indicates that those who rated worse quality of life were more likely to shape negative perceptions of aging towards themselves rather than towards the society. 3.2.7 Institutions involved in old age discrimination The univariate analysis showed that seeing each social institution as discriminating against older people increased the risk of negatively perceiving aging on both societal and individual levels, as well as holding negative societal views on aging while subjective views were positive/neutral. Though adjusting for other variables attenuated most of the association effects, regarding politicians, social services, private companies and employers as perpetrators of ageism remained to besignificantly associated with having negative perceptions of both societal and subjective views on old age. Meanwhile, opinions on old age discrimination were not related to the likelihood of viewing old age negatively on a personal level despite having positive/ neutral societal views on aging. These findings suggest that whether each institution was discriminating against older people was more Ishikawa 10.3389/fsoc.2023.1291325 Frontiers in Sociology 06 frontiersin.org TABLE1 Sample characteristics for the total sample and share of having negative societal views on old age and having negative self-perceptions of aging in respective groups. Distribution in the sample (%), N= 1,981 Share of having negative societal views on old age (%) Share of having negative subjective views on old age (%) Share of feeling fear about own old age (%) Gender Male 47.7 36.7 14.2 22.1 Female 52.3 50.8 15.6 32.9 Age in ten-year groups 16–25 11.3 30.0 16.1 30.0 26–35 12.4 37.4 17.5 37.8 36–45 13.2 37.0 15.3 36.6 46–55 14.8 42.7 13.7 33.8 56–65 18.9 51.5 18.1 29.1 66–75 17.0 54.2 11.6 18.2 76–85 8.9 54.8 11.3 10.2 86 + 3.5 29.0 14.5 10.1 Living arrangements Living with a partner 46.9 43.7 12.6 24.9 Not living with a partner 53.1 44.4 17.0 30.3 Education Basic education 19.9 37.2 13.7 26.6 Secondary education 54.6 43.3 14.7 28.9 Tertiary education 25.4 51.2 16.5 26.2 Income level Sufficient 48.1 40.4 12.9 18.4 Insufficient/ reluctance to answer 51.9 47.4 16.8 36.4 Quality of life Very good 8.7 28.9 5.8 6.9 Good 46.0 40.4 11.5 21.2 Neither good nor bad 32.2 46.9 15.7 31.4 Bad 11.1 59.4 30.1 53.4 Very bad 2.1 63.4 36.6 68.3 Institutions involved in old age discrimination Politicians No 71.4 34.6 13.6 23.0 Yes 28.6 67.7 18.2 39.8 Media No 86.9 40.2 14.8 26.1 Yes 13.1 69.9 16.2 38.6 Health care services No 76.7 37.4 13.6 24.9 Yes 23.3 66.2 19.5 37.1 Social services No 74.1 36.9 13.2 23.1 Yes 25.9 64.5 19.9 41.1 (Continued) Ishikawa 10.3389/fsoc.2023.1291325 Frontiers in Sociology 07 frontiersin.org relevant with respondents’ perceptions of societal views on old age than their self-perceptions of aging. 3.3 Association of different factors with combinations of societal age stereotypes and personal feelings of own old age Table3 shows the association of respondents’ sociodemographic and contextual factors with combinations of societal age stereotypes and personal feelings of own old age. Reference category was those who took a positive/neutral stance towards societal views on old age and felt safe or neutral about their own old age. The likelihood of being placed in other three categories (negative societal views–feeling fear about own old age, negative societal views–feeling safe/neutral about own old age, positive/neutral societal views–feeling fear about own old age,) in comparison to the reference category was specified in the table. Multivariate analysis controlled for all variables revealed that respondents who were female, had the highest educational level, rated quality of life as worse, and considered politicians, social services and employers discriminating against older people were more likely to think that old age was regarded negatively in the society and were fearful about their own old age. These factors appeared to beassociated with internalization of negative societal age stereotypes into selfperceptions of aging. However, since other categories of the outcome variable also had many significant predictors, findings are described according to the explanatory variables as follows. 3.3.1 Gender Being female increased the risk of being placed in all the three categories of the outcome variables. Multivariate analysis controlled for all variables even strengthened the gender effect for all the categories. The results suggest that compared to males, females were more likely to think that socially shared beliefs about aging were negative and they were also inclined to feel fear about own old age. 3.3.2 Age Respondents aged between 26 and 75 years compared with the youngest age group were more likely to perceive aging negatively on societal level as well as to feel fear about own old age. However, adjusting for other variables attenuated association effect of age. In terms of having negative societal views on old age while feeling safe/ neutral about own old age, older age increased the risk of being placed in this category. In contrast, both univariate and multivariate analyses revealed that older respondents were obviously less likely to feel fear about own old age while having positive/neutral societal views on old age. The results of age effect indicate that older age was relevant to shaping negative perceptions of societal age stereotypes and was much unlikely to berelated to fear of aging. 3.3.3 Living arrangements Whether the respondents lived with a partner or not was generally not related to different combinations of societal age stereotypes and personal feelings of own old age. 3.3.4 Education Attaining tertiary education was significantly associated with being placed in the two categories: negative societal views on old age– feeling fear about own old age, and negative societal views on old age– feeling safe/neutral about own old age. Adjusting for other variables even increased the effect of tertiary education. However, educational level was not related to the category of having positive/neutral societal views on old age and simultaneously feeling fear about own old age. This indicates that negative attitudes among people with tertiary education were directed mainly to societal beliefs about aging rather than to personal feelings of own old age. 3.3.5 Income level Univariate analysis revealed that living under insufficient economic situation increased the risk of belonging to the two categories: having negative societal views on old age–feeling fear about own old age, and having positive/neutral societal views on old age–feeling fear about own old age. This may indicate that disadvantage in economic situation was subject to fear of own old age rather than negative perceptions of societal views on aging. However, adjusting for other variables attenuated association effect of income level. Distribution in the sample (%), N= 1,981 Share of having negative societal views on old age (%) Share of having negative subjective views on old age (%) Share of feeling fear about own old age (%) The Social Insurance Institution of Finland No 81.3 39.5 13.7 23.7 Yes 18.7 63.9 20.5 45.3 Private companies No 92.6 42.0 14.0 26.7 Yes 7.4 70.1 27.2 40.8 Transport No 90.3 41.4 14.3 26.0 Yes 9.7 69.3 21.4 44.3 Employers No 77.7 37.0 12.7 23.6 Yes 22.3 68.8 22.6 42.3 TABLE1 (Continued) Ishikawa 10.3389/fsoc.2023.1291325 Frontiers in Sociology 08 frontiersin.org TABLE2 Association of individual characteristics and contextual factors with combinations of societal age stereotypes and subjective views on old age. CategoriesaNegative societal views on old age–Negative subjective views on old age Negative societal views on old age–Positive/neutral subjective views on old age Positive/neutral societal views on old age–Negative subjective views on old age Model 1bModel 2cModel 1 Model 2 Model 1 Model 2 Variables RRd (95% CI) RR (95% CI) RR (95% CI) RR (95% CI) RR (95% CI) RR (95% CI) Gender Male ref. ref. ref. ref. ref. ref. Female 1.62 (1.20–2.19) 1.89 (1.36–2.63) 1.83 (1.50–2.24) 2.08 (1.66–2.60) 1.02 (0.65–1.59) 0.97 (0.61–1.55) Age in ten-year groups 16–25 ref. ref. ref. ref. ref. ref. 26–35 1.61 (0.83–3.14) 0.92 (0.45–1.88) 1.28 (0.83–1.99) 0.85 (0.53–1.36) 0.84 (0.42–1.68) 0.79 (0.38–1.67) 36–45 1.46 (0.75–2.84) 0.73 (0.35–1.50) 1.25 (0.82–1.92) 0.78 (0.49–1.25) 0.63 (0.31–1.30) 0.54 (0.25–1.18) 46–55 1.57 (0.82–3.01) 0.76 (0.37–1.53) 1.63 (1.08–2.46) 1.06 (0.67–1.65) 0.48 (0.22–1.03) 0.37 (0.16–0.84) 56–65 2.87 (1.58–5.21) 1.49 (0.78–2.86) 2.09 (1.41–3.10) 1.38 (0.90–2.13) 0.40 (0.18–0.87) 0.30 (0.13–0.70) 66–75 1.98 (1.05–3.74) 1.31 (0.66–2.60) 2.61 (1.76–3.88) 1.97 (1.28–3.04) 0.18 (0.06–0.54) 0.15 (0.05–0.47) 76–85 1.79 (0.85–3.78) 1.32 (0.59–2.92) 2.84 (1.81–4.45) 2.26 (1.39–3.68) 0.36 (0.12–1.09) 0.32 (0.10–1.02) 86 + 1.13 (0.42–3.07) 0.98 (0.34–2.81) 0.83 (0.42–1.64) 0.82 (0.40–1.70) 0.60 (0.20–1.86) 0.45 (0.14–1.47) Living arrangements Living with a partner ref. ref. ref. ref. ref. ref. Not living with a partner 1.25 (0.93–1.69) 1.20 (0.86–1.68) 1.04 (0.85–1.26) 1.09 (0.87–1.36) 2.10 (1.30–3.39) 1.61 (0.97–2.68) Education Basic education ref. ref. ref. ref. ref. ref. Secondary education 1.41 (0.93–2.16) 1.59 (1.00–2.51) 1.22 (0.94–1.59) 1.28 (0.96–1.71) 0.81 (0.48–1.38) 0.96 (0.54–1.68) Tertiary education 2.04 (1.28–3.23) 2.98 (1.77–5.00) 1.65 (1.23–2.21) 1.99 (1.42–2.80) 0.76 (0.39–1.47) 1.02 (0.49–2.10) Income level Sufficient ref. ref. ref. ref. ref. ref. Insufficient/ reluctance to answer 1.65 (1.22–2.23) 0.91 (0.62–1.33) 1.26 (1.04–1.53) 1.02 (0.79–1.31) 1.17 (0.75–1.82) 0.68 (0.40–1.16) Quality of life Very good ref. ref. ref. ref. ref. ref. Good 2.83 (1.20–6.67) 3.16 (1.31–7.65) 1.56 (1.07–2.27) 1.52 (1.01–2.30) 1.86 (0.65–5.37) 2.09 (0.72–6.10) Neither good nor bad 4.61 (1.95–10.88) 4.73 (1.90–11.77) 1.98 (1.35–2.91) 1.69 (1.09–2.64) 2.69 (0.92–7.82) 3.74 (1.23–11.43) Bad 13.58 (5.55–33.26) 14.74 (5.59–38.85) 2.96 (1.85–4.74) 2.79 (1.63–4.79) 6.52 (2.10–20.26) 9.92 (2.97–33.12) Very bad 19.83 (6.15–63.96) 21.14 (5.93–75.38) 3.69 (1.57–8.64) 3.22 (1.25–8.28) 10.82 (2.37–49.33) 16.26 (3.25–81.31) Institutions involved in old age discrimination (ref.: not discriminating) Politicians 3.66 (2.66–5.04) 1.56 (1.02–2.38) 4.05 (3.24–5.08) 2.32 (1.73–3.12) 1.01 (0.56–1.84) 1.12 (0.52–2.41) Media 2.80 (1.84–4.28) 0.81 (0.48–1.37) 3.54 (2.62–4.79) 1.29 (0.89–1.86) 0.64 (0.23–1.80) 0.58 (0.18–1.80) Health care services 3.64 (2.62–5.06) 1.52 (0.96–2.40) 3.10 (2.45–3.94) 1.39 (1.00–1.93) 0.81 (0.41–1.61) 0.86 (0.36–2.05) Social services 3.57 (2.59–4.92) 1.58 (1.01–2.47) 3.00 (2.39–3.78) 1.37 (1.00–1.87) 1.12 (0.62–2.00) 1.36 (0.62–2.98) The Social Insurance Institution of Finland 3.32 (2.35–4.69) 0.99 (0.61–1.60) 2.49 (1.93–3.22) 0.88 (0.62–1.25) 0.86 (0.42–1.77) 0.67 (0.27–1.66) Private companies 5.03 (3.12–8.12) 2.35 (1.31–4.23) 2.77 (1.85–4.16) 1.22 (0.75–1.99) 1.23 (0.43–3.52) 1.11 (0.34–3.62) Transport 3.72 (2.38–5.82) 0.94 (0.53–1.64) 2.99 (2.12–4.24) 1.03 (0.67–1.57) 0.88 (0.31–2.49) 0.77 (0.24–2.49) Employers 4.95 (3.55–6.91) 2.60 (1.73–3.92) 3.57 (2.79–4.58) 1.84 (1.36–2.49) 1.45 (0.79–2.65) 1.69 (0.82–3.51) Values showing the significance level of p < 0.05 are bolded. aThose who have positive/neutral attitudes towards both societal views on old age and subjective views on old age as reference category (n = 1,024, 51.7%). Distribution of other categories: negative societal views on old age–negative subjective views on old age (n = 212, 10.7%), negative societal views on old age–positive/neutral subjective views on old age (n = 661, 33.4%), positive/neutral societal views on old age–negative subjective views on old age (N = 84, 4.2%). bAll variables included into the model separately. cAll variables included into the model simultaneously. dRatio of probability of outcome in respective categories vs reference category.