Subjective Age and Its Correlates Among Middle-Aged and Older Adults

Abstract

The present study evaluates discrepancies in subjective age as reported by middle-age persons (aged 44–64 years) in comparison to older adults (aged 65 years and older), using a multidimensional definition of the concept. A convenience sample of 126 middle-aged and 126 older adults completed subjective age measures (felt age, desired age, and perceived old age), attitudes toward older adults, knowledge about aging, and sociodemographic questionnaires. Overall, participants reported feeling younger than they actually were and wanting to be younger than their chronological age. Perceived mean for old age was about 69 years. Discrepancies in felt age and desired age were significantly larger for the older group compared to the middle-aged group. Regarding perceived old age, compared to the younger group, older adults reported that old age begins at an older age. Findings suggest that middle-aged and older adults’ perceptions regarding themselves and regarding old age in general are independent and need, therefore, separate research and practical attention.

Keywords

subjective age older adults middle aged

Introduction

Research has shown that people define both themselves and others in terms of chronological age-group categories such as “older adults” and “middle-aged adults” (Sanderson & Scherbov, 2008; Toothman & Barrett, 2011). The period of old age, in the developed world, is often considered to begin at the age of 65 years (Sanderson & Scherbov, 2008), although variations in the perceptions of the beginning of old age and its determinants have been noted (Gilleard, 2009). The period of middle age is customary assumed to begin between the ages of 35 and 45 years and to end at the age of 65 years (Barclay, Stoltz, & Chung, 2011; Kertzner, 2001). However, this age period is less defined relative to old age mainly because middle-aged individuals may admit to not being young anymore but still differentiate themselves from the elderly (Packer & Chasteen, 2006). Even though age categorization affects the perceptions that people form about themselves and about others (Garstka, Schmitt, Branscombe, & Hummert, 2004), subjective age (SA) has also proven to convey important meanings.

SA is defined as one’s self-evaluation of how old one perceives oneself to be (Barak & Stern, 1986; Montepare, 2009). SA is considered as a multidimensional construct (Kleinspehn-Ammerlahn, Kotter-Grühn, & Smith, 2008) assessing different aspects such as how old individuals feel (felt age), how old they would like to be (desired age), and in what age old age begins (perceived old age) (Kaufman & Elder, 2002; Uotinen, Suutama, & Ruoppila, 2003). While in the first two dimensions, the person himself or herself is the target of the evaluation, the third dimension—perceived old age—refers to a general evaluation of the beginning of this period in life. Overall, two theoretical perspectives have been developed in order to explain the predictors of SA (Stephan, Sutin, & Terracciano, 2015). The first examines the topic from a social psychological perspective and therefore states that SA is associated with social and environmental cues such as social roles and social loneliness (Stephan et al., 2015). The second, examined SA from a biomedical perspective, stating that how young or old an individual feels is associated with information about one’s physical health and functioning (Kleinspehn-Ammerlahn et al., 2008). Indeed, negative changes in health and functioning are often cited as important factors in a shift from a younger to older SA (Bowling, See-Tai, Ebrahim, Gabriel, & Solanki, 2005; Sherman, 1994).

Moreover, overall, studies assessing SA, a unidimensional construct is more commonly used, and among those examining the three dimensions, perceived age has attracted lesser attention in comparison to felt age and desired age. This is surprising since it was argued that SA may reflect different underlying dimensions and that researchers should differentiate between these dimensions and their unique implications (Barak, 2009). Different studies have found that middle-aged and older adults have the conception that older SA means lower life satisfaction and psychological well-being, higher negative affect, and pessimistic perspectives on cognitive functioning (Mock & Eibach, 2011; Schafer & Shippee, 2009). Yet, a youthful SA was found to be an indicator of successful aging and higher life satisfaction (e.g., Demakakos, Gjonca, & Nazroo, 2007; Uotinen et al., 2003). Indeed, the tendency to feel younger or older than one’s actual age is considered a crucial construct in old age, with implications for a variety of physical and psychological outcomes (Stephan, Caudroit, & Chalabaev, 2011) and longevity (Kotter-Grühn et al., 2009).

Overall, examining SA is important since it has been found to predict psychological and health-related outcomes as well as adaptation to life changes among both middle-aged and older adults (Spuling, Miche, Wurm, & Wahl, 2013; Stephan, Caudroit, et al., 2011).

However, findings regarding SA are not conclusive, with many studies reporting that most older and middle-aged adults feel younger, want to be younger, and think that they look younger than they are (Kleinspehn-Ammerlahn et al., 2008; Rubin & Berntsen, 2006), and others demonstrating that middle-aged and older adults feel older than their age (Galambos, Albrecht, & Jansson, 2009; Schafer & Shippee, 2009). These discrepancies might stem from the fact that studies of SA have mainly concentrated on exclusive older adults’ samples (aged 60 or 65 years and older). This is surprising especially since SA has also been shown to be important to persons in their 40s and 50s (Ward, 2010; Westerhof, Barrett, & Steverink, 2003). For example, Westerhof et al. (2003) compared middle-aged and older adults from the United States and Germany, aged 40 to 74 years, and found that both middle-aged and older adults reported feeling younger than their age while the bias toward youthful identities was stronger at older ages. Yet, this study (Westerhof et al., 2003) as well as other studies examining SA (e.g., Bodner, Ayalon, Avidor, & Palgi, 2016) were limited by the use of a unidimensional approach including one general question to measure SA. This might fall short of capturing relevant complexities in the experience of SA, because such a unidimensional construct may be more reflective of health and physical aging than of other domains (such as societal and intellectual age) that are relevant for the aging process (Kornadt, Hess, Voss, & Rothermund, 2016).

The Present Study

In sum, while SA was found to be an important concept for adult persons’ well-being, knowledge in this area is inconclusive, especially in middle-aged persons. Thus, the first aim of the present study was to expand knowledge in this area by assessing discrepancies in SA as reported by middle-age persons (aged 44–64 years) in comparison to older adults (aged 65 years and above), using a multidimensional definition of the concept. Based on Westerhof et al. (2003), we hypothesized that there will be differences between middle-aged and older adults in all dimensions of SA, with older adults reporting wider discrepancies between subjective and chronological age, compared to middle-aged persons.

Moreover, while the literature has examined extensively the consequences of SA (e.g., Choi & DiNitto, 2014; Westerhof, Miche, Brothers, & Wurm, 2014), knowledge about its antecedents is less developed (e.g., Stephan, Demulier, & Terracciano, 2012; Stephan et al. 2015). The identification of correlates of SA is relevant for the exploration of mechanisms that might explain the source of discrepancy between chronological age and SA (Kotter-Grühn, Kornadt, & Stephan, 2015). Thus, the second aim of the present study was to examine the correlates associated with discrepancies in SA in both middle-aged and older adults. Two main factors were examined: attitudes toward older adults and knowledge about aging.

Overall, attitudes toward older adults have been consistently found to be negatively associated with SA (Mock & Eibach, 2011). For example, it was found that when attitudes toward older adults are less favorable, older SA predicts lower life satisfaction and increased negative affect. In contrast, when attitudes toward older adults are more favorable, older SA is no longer associated with these measures of psychological well-being (Mock & Eibach, 2011). In addition, it was argued that negative attitudes toward older adults may lead to perceptions of younger SA, since people who perceive old age negatively would try to dissociate themselves from being old.

The second potential factor examined was knowledge about aging. It has been clearly stated that in order to enhance positive attitudes toward the aged, more knowledge about aging is needed, especially in order to clarify misconceptions that involve images of age (Suh et al., 2012). Indeed, it was found that greater accurate knowledge of aging is associated with more positive attitudes and with less negative attitudes toward older adults (Stuart-Hamilton & Mahoney, 2003; Wurtele & Maruyama, 2013) namely there is a direct association between knowledge about aging and attitudes toward older adults. However, in contrast to the diverse studies examining the associations between attitudes toward older adults and SA (e.g., Mock & Eibach, 2011), to date, to the best of our knowledge, no study examined the association between knowledge about aging and SA. Yet, it should be noted that according to previous studies, limited knowledge about aging negatively contributes to subjective perspectives such as psychological well-being and life satisfaction (Jeon & Shin, 2009; Suh et al., 2012), while higher levels of knowledge about aging were found to be associated with lower worry scores (Neikrug, 1998). Furthermore, it was argued that attitudes toward age and aging may mediate the associations between proximal variables and SA (Montepare, 2009). Therefore, based on these studies (Jeon & Shin, 2009; Montepare, 2009; Suh et al., 2012), we hypothesized that knowledge about aging would be negatively associated with SA, while attitudes toward older adults would mediate the association between knowledge about aging and SA.

Finally, to reach a greater understanding of the factors associated with discrepancies in SA, we examined relationships with the following key background variables: gender, number of children, subjective income, and subjective health status. These key background variables were chosen since they were consistently found in the literature to be associated with SA (e.g., Hubley & Russell, 2009; Montepare, 2009; Montepare & Lachman, 1989). Based on previous studies (Hubley & Russell, 2009; Montepare & Lachman, 1989), we hypothesized that being female, having a higher number of children, lower levels of subjective income and poorer subjective health will be related to larger discrepancies between chronological and SA.

Method

Participants

A convenience sample of 252 lay people participated in the study, half of them were middle age (aged 44–64 years) and half were older adults (aged 65 years and older).

Procedure

Interviews were conducted by trained gerontology students and lasted between 20 and 25 min. Participants received an explanation about the purpose of the study and signed a consent form. No incentive was provided. Face-to-face interviews were conducted, using a structured questionnaire. All participants were recruited opportunistically from public places such as places of work and through personal contacts of the trained interviewers, while ensuring even amounts of middle-aged and older adults. The protocol study was approved by the ethics committee of the Faculty of Social Welfare and Health Sciences of the University of Haifa.

Measures

SA: We examined three of SA ranging from specific aspects of personal aging to a more general view of what it means to be old (Kaufman & Elder, 2002). Based on a previous study (Kleinspehn-Ammerlahn et al., 2008), the first dimension dealt with felt age (“how old do you feel?”), and the second dimension dealt with desired age (“how old would you want to be?”). Finally, the third dimension—perceived old age—was assessed with a more general question about old age (“at what age do you think old age begins?”) (Kaufman & Elder, 2002; Uotinen et al., 2003). Intercorrelations among dimensions were verified before proceeding. For each dimension, we calculated the discrepancy score between SA and chronological age as the dependent variable (i.e., discrepancy = chronological age − SA). We chose to use a discrepancy score between chronological age and the SA since discrepancy score might have different meanings depending on the age of a person (Kotter-Grühn & Hess, 2012).

Subtracting actual age from felt age, desired age and perceived age yielded values that were positive when participants felt older than their chronological age and values that were negative when participants felt younger than their chronological age.

Attitudes toward older adults were assessed using the Hebrew form of the refined version of the Aging Semantic Differential (refined-ASD; Polizzi, 2003; Porat, 2010). Participants were asked to indicate their attitudes toward older adults using 23 positive and negative semantic differentials (e.g., friendly–unfriendly, pleasant–unpleasant, nice–mean, cooperative–uncooperative). After reversing 11 of the items, a composite index of the average of all items was created, with a higher score indicating high levels of negative attitudes. The original questionnaire’s internal consistency was excellent—α = .97 (Polizzi, 2003). In the current study, the internal consistency was good (Cronbach’s α = .71).

Knowledge about aging was assessed using the Facts on Aging Quiz I which contains 25 true/false statements about aging. The statements cover the basic physical, mental, and social facts and the most frequent misconceptions about aging (Palmore, 1998). In the current study, the item “Over 15% of the United States population are now aged 65 or over” was removed since it is not relevant to the Israeli population. An overall index of knowledge was calculated by summing the correct answers. Total scores ranged from 0 to 24, with a higher score indicating a greater knowledge about aging. The questionnaire’s reliability has been found to range from α = .50 to .80 (Palmore, 1998). In the current study, the internal consistency was good (α = .71).

Sociodemographic characteristics: These included gender (female/male), chronological age, age-group (middle-aged/older adults), place of birth (Israel, Asia/Africa, Europe/America, and Other), marital status (married/not married), number of children, work (yes/no), subjective health status (How would you define your health status? Bad/not good and not bad/good), and subjective perception of income (good/not good and not bad/bad).

Statistical Analyses

The data were cleaned, coded, and analyzed using SPSS version 20.0 (Arbuckle, 2005). Statistical analyses included descriptive statistics (means, standard deviations, and percentages) to describe the sample and the main study’s variables. To assess differences between middle-aged and older adults, t and χ² tests were performed according to the type of variable. Hierarchical regression was applied to assess the contribution of the study variables to the explained variance of SA. Finally, mediation models were estimated with the PROCESS macro for SPSS (Hayes, 2013). We set a stricter p value to signify significance by using Bonferroni correction. Accordingly, only a p value equal to, or smaller than, .0125 was considered significant.

Results

Sociodemographic Characteristics of the Participants

The sociodemographic characteristics of the participants are shown in Table 1. Statistically significant differences were evident in age, place of birth, marital status, and work. A greater percentage of the middle-aged participants were born in Israel and were married and working compared to participants aged 65 years and older.

Table 1.

Participants’ characteristics (N = 252).

Sociodemographic characteristics	Middle aged (N = 126)	Older adults (N = 126)	t test/chi-square values and significance
Mean age (SD)	54.00 (5.7)	72.49 (6.5)	t₍₂₄₆₎ = −24.0**
Gender, N (%)
Male	44 (35.2)	58 (46.8)	NS
Female	81 (64.8)	66 (53.2)
Place of birth, N (%)
Israel	89 (71.8)	62 (49.2)	χ²₍₃₎ = 15.8**
Europe/America	25 (20.2)	23 (18.3)
Asia/Africa	7 (5.6)	36 (28.6)
Other	3 (2.4)	5 (4.0)
Marital status, N (%)
Married	96 (78.0)	87 (69.6)	χ²₍₄₎ = 12.2**
Not married	27 (22.0)	38 (30.4)
Mean number of children (SD)	3.16 (1.4)	3.56 (2.2)	NS
Working, N (%)
Yes	106 (85.5)	41 (32.3)	χ²₍₂₎ = 75.9**
No	17 (13.7)	86 (67.7)
Health status, N (%)
Good	92 (73.6)	76 (59.8)	NS
Not good and not bad	30 (24.0)	45 (35.4)
Bad	3 (2.4)	6 (4.7)
Subjective income, N (%)
Good	58 (46.4)	67 (52.8)	NS
Not good and not bad	52 (41.6)	52 (40.9)
Bad	15 (12.0)	8 (6.3)

**p < .01.NS= Not significant.

Subjective Age

As shown in Table 2, overall participants reported feeling younger than they actually were and wanting to be younger than their chronological age. Perceived mean of old age was about 69 years. Although similar patterns were observed for both age-groups, discrepancies were significantly larger for the older adults group compared to the middle-aged group. Namely, in comparison to middle aged persons who felt 9.8 years younger than they were, older adults felt 13.5 years younger than their chronological age. In addition, in comparison to middle aged that desired to be 17.2 years younger than their chronological age, older adults desired to be 24.4 years younger than their chronological age. Regarding perceived age, compared to the middle-aged group who said that old age begins at the age of 67.8 years, older adults reported that old age begins at the age of 75.1 years. In addition, discrepancy in felt age was positively correlated with discrepancy in desired age (r = .22, p < .001). However, the correlations between discrepancy in felt age and discrepancy in perceived old age (r = −.14, p > .05) and discrepancy in desired age and discrepancy in perceived old age (r = −.13, p > .05) were not statistically significant.

Table 2.

Differences in Discrepancies in Subjective Age Among Middle-Aged and Older Adults.

Item	Total sample (N = 252)	Middle aged (N = 126)	Older adults (N = 126)	t test values and significance levela
Chronological age, M (SD)	63.3 (11.1)	54.0 (5.7)	72.5 (6.5)	t₍₂₄₆₎ = −24.0**
Discrepancy in felt age, M (SD)	−11.9 (12.3)	−9.8 (10.2)	−13.5 (13.9)	t₍₂₂₀₎ = −2.4**
Discrepancy in desired age, M (SD)	−20.8 (16.8)	−17.2 (13.4)	−24.4 (19.0)	t₍₂₂₂₎ = −3.4**
Discrepancy in perceived old age, M (SD)	5.9 (14.0)	13.8 (10.9)	2.6 (11.8)	t₍₂₀₁₎ = −10.3**

^aOnly differences between middle-aged and older adults are shown.

**p < .01.

Attitudes Toward Older Adults

Participants’ attitudes toward older adults were moderately negative, with a mean score of 4.34 (SD = 0.43; on a 1–7 scale, higher score meaning higher negative attitudes toward older adults). No statistically significant differences were found between middle-aged and older adults.

Knowledge About Aging

Overall, the level of knowledge about aging was relatively poor (M = 12.36 out of a maximum score of 24, SD = 3.82). No statistically significant differences in knowledge about aging were found.

Correlations Among the Study Variables

Discrepancy in felt age was positively correlated with subjective health status (r = .22, p < .001), namely, the more participants felt younger, the more they defined their subjective health status as good. Discrepancy in desired age was positively associated with chronological age (r = .24, p < .001) and subjective health status (r = .14, p < .05), namely, the more participants desired to be younger, the more their chronological age was higher and the more they defined their subjective health status as good. Finally, discrepancy in perceived old age was positively associated with chronological age (r = .67, p < .001), subjective health status (r = .24, p < .001), and knowledge about aging (r = .28, p < .001), namely, the more participants perceived that old age begins at a higher age, the more their chronological age was higher, the more they defined their subjective health status as good and the more acknowledgement participants had about aging. It should be noted that when the correlations were computed separately for middle-aged and older adults, they revealed almost the same pattern of associations as the general sample.

Hierarchical Regression Analyses for Identifying SA Correlates

Hierarchical regression was conducted to assess the correlates of the three dimensions of SA (i.e., discrepancy in felt age, discrepancy in desired age, discrepancy in perceived old age). In the first step, we included sociodemographic variables (gender, number of children, subjective health status, and subjective income), in the second step, we included age-group, in the third step, we included knowledge about aging, and in the fourth step, we included attitudes toward older adults. The results are shown in Table 3.

Table 3.

Hierarchical Regressions Analyses for Identifying Correlates of Discrepancy in Subjective Age.

	Discrepancy in felt age	Discrepancy in desired age	Discrepancy in perceived old age
	β	β	β
Step 1
Gender	.13	.19*	−.02
Number of children	−.09	.06	.01
Subjective health status	−.19*	.09	.28**
Subjective income	.04	−.12	.12
Step 2
Gender	.12	.18*	−.05
Number of children	−.06	.03	−.01
Subjective health status	−.22*	.04	.13
Subjective income	.01	−.17*	.18*
Age-group	.17*	.21**	.56**
Step 3
Gender	.12	.18*	−.07
Number of children	−.11	.03	−.01
Subjective health status	−.18*	.03	.09
Subjective income	−.02	−.16*	.04
Age-group	.16*	.21**	.55**
Knowledge about aging	−.09	.05	.23**
Step 4
Gender	.12	.18*	−.07
Number of children	−.11	.03	.01
Subjective health status	−.18*	.04	.10
Subjective income	−.02	−.15*	.05
Age-group	.17*	.21**	.55**
Knowledge about aging	−.10	.07	.26**
Attitudes toward older adults	−.03	.09	.12
Overall R²

*p < .05, **p < .01.

Gender—1. Female, 2. Male; Subjective health status—1. Excellent, 2. Good, 3. Fair, 4. Bad, 5. Very bad; Subjective income—1. Very poor, 2. Poor, 3. Average, 4. Good, 5. Very good; Age-group—1. Middle aged, 2. Older adults.

For discrepancy in felt age, in the first equation, the demographic variables explained 5% of the variance, with subjective health being the main predictor, F(1, 174) = 6.38, p < .05; β = −.19. The second equation added an additional 3% to the explained variance in the dependent variable, all due to subjective health status, F(5, 170) = 3.17, p < .05; β = −.22, and age-group, F(5, 170) = 3.17, p < .05; β = .17. Finally, the third equation which included knowledge about aging and the fourth equation which included attitudes toward older adults did not contribute to the explained variance in the dependent variable.

For discrepancy in desired age, demographic variables explained 7% of the variance, with gender being the main predictor, F(4, 179) = 3.52, p < .05; β = .19. The second equation added an additional 5% to the explained variance in the dependent variable, all due to gender F(5, 178) = 4.51, p < .05; β = .18, subjective health status, F(5, 178) = 4.51, p < .01; β = .41, and age-group F(5, 178) = 4.51, p < .01; β = .21. Finally, the third equation which included knowledge about aging and the fourth equation which included attitudes toward older adults did not contribute to the explained variance in the dependent variable.

For discrepancy in perceived old age, demographic variables explained 7% of the variance, with subjective health status being the main predictor, F(1, 150) = 15.71, p < .01; β = .28. The second equation added an additional 28% to the explained variance, all due to subjective health status, F(2, 149) = 42.83, p < .01; β = .18, and age-group, F(2, 149) = 42.83, p < .01; β = .53). The third equation added an additional 5% to the explained variance in the dependent variable, all due to age-group, F(3, 148) = 34.70, p < .01; β = .55, and knowledge about aging, F(3, 148) = 34.70, p < .01; β = .23. Finally, in the fourth equation, age-group, F(7, 144) = 15.79, p < .01; β = .55, and knowledge about aging, F(7, 144) = 15.79, p < .01; β = .26, added 2% to the explained variance in the dependent variable.

To examine direct and indirect effects of knowledge about aging on SA (discrepancy in felt age, discrepancy in desired age, and discrepancy in perceived old age), we used Hayes’ (2013) macro. According to this method, if zero is not included in the 95% confidence interval, the indirect effect is significant (Cheung & Lau, 2008; Hayes, 2009). This method to testing mediation has been reported as being superior to other methods given that it does not impose the assumption of normality and provides more accurate confidence intervals (Cheung & Lau, 2008; Hayes, 2009). Chronological age, number of children, subjective income, and subjective health status were considered as covariates in the mediation analysis. As shown in Table 4 and in Figure 1, the results suggested a significant indirect effect of knowledge about aging on discrepancy in perceived old age via attitudes toward older adults (LLCI = −.27 and UPCI = −.01), but not for discrepancy in felt age and discrepancy in desired age.

Table 4.

Direct and Indirect Effects of Knowledge About Aging on SA (Discrepancy in Felt Age, Discrepancy in Desired Age and Discrepancy in Perceived Old Age) Via Attitudes Toward Elderly (N = 252).

	Coefficient	SE	Boot LLCI	Boot ULCI
Indirect—Discrepancy in felt age	.02	.05	−.08	.14
Indirect—Discrepancy in desired age	−.07	.06	−.23	.02
Indirect—Discrepancy in perceived old age	−.11	.07	−.27	−.01

SA = subjective age.

Figure 1.

A mediator model depicting direct and indirect effects of knowledge about aging and attitudes toward older adults on discrepancy in perceived old age. **p < .01. Values represent unstandardized regression coefficients (B). The direct effect is presented after extracting mediation effects; The direct effects before extracting mediation effects are presented in brackets. Chronological age, number of children, subjective income, and subjective health status are controlled for.

Discussion

Summary of Findings

The present study constitutes one of the few empirical efforts to examine SA and its correlates as a multidimensional concept among middle-age and older adults. As hypothesized, we found that differences between middle-aged and older adults in all dimensions of SA, with older adults reporting wider discrepancies between subjective and chronological age, compared to middle-aged persons was supported. However, the hypothesis that SA will be associated with knowledge about aging was supported only for perceived old age. In addition, the hypothesis that attitudes toward older adults would mediate the association between knowledge about aging and discrepancy in SA was supported only for perceived old age. Finally, the hypothesis that being female, having a higher number of children, lower levels of subjective income, and poorer subjective health will be related to larger discrepancies between chronological and SA was supported only for subjective health.

Multidimensionality and Level of SA

Recent studies in the area of SA define the concept as encompassing personal dimensions (felt age and desired age) as well as general dimensions (perceived old age) (Kaufman & Elder, 2002). Findings of our study support this conceptualization as we found significant associations only between the personal dimensions of SA, though not with the general dimension. These findings suggest that perceived old age is not just a general dimension but also an independent and separate dimension of SA, which might be tapping into a more global sense of old age rather than a more personal measure (Kaufman & Elder, 2002). However, it might be that the lack of statistical association with perceived old age suggests that this dimension does not fall under the umbrella of SA. In addition, it should be noted that the lack of association between the personal dimensions (“How old do you feel?”) and the general dimension of discrepancy in SA (“At what age do you think old age begins?”) might stem from the general wording of the question of this dimension, rather than from it being an independent dimension. Therefore, it is recommended that future studies will be conducted using a more personal question such as “at what age do you think YOU will consider yourself old?.”

As for the level of SA, similar to previous studies (e.g., Kleinspehn-Ammerlahn et al., 2008; Rubin & Berntsen, 2006), both middle-aged and older adults in the current study, felt and wanted to be younger than their chronological ages. Furthermore, they tended to want to be even younger than they felt. Several explanations might clarify why both middle-aged and older adults in the current study felt and wanted to be younger than their chronological ages. First, the discrepancy between chronological age and SA in the personal dimensions might reflect individuals’ dissociation from the old age in response to the exposure to negative age misconceptions and stigma (Barrett & Montepare, 2015). Second, it might be that younger SA serves as a self-protective strategy allowing individuals to disengage themselves from old age (Weiss & Lang, 2012). Third, holding a youthful SA might have positive consequences on participants since it was found to be associated with indicators of successful aging (Demakakos et al., 2007) and was found to be positively associated with both general and specific self-efficacy (Stephan, Caudroit, et al., 2011). Finally, according to theoretical perspectives, it might be that participants felt young as a result of social and environmental cues (Stephan, Sutin, et al., 2013) or good physical health and functioning (Kleinspehn-Ammerlahn et al., 2008). Indeed, we found that subjective health was a predictor of the different dimensions of SA. Thus, in line with a past study (Stephan, Sutin, et al., 2015), the present study suggests that SA integrates both social and biological perspectives about aging and thus adds to prior theoretical accounts that consider social psychological and health-related factors as sources for age ratings.

In the general dimension, old age was perceived to begin at around the age of 69 years, 4 years older than the considered beginning of old age in the developed world (age of 65 years) (Sanderson & Scherbov, 2008). It is interesting to see that as life expectancy has increased so has the age at which people are considered to be “older adults” (Kaufman & Elder, 2002).

Comparison Between Age-Groups in SA

As previously stated, studies comparing SA among different age-groups arrived to inconclusive findings with some showing that middle-aged and older adults do not differ in their felt age, suggesting that after a shift in SA by late middle age, chronological age might not play a major role (e.g., Kornadt et al., 2016) and others showing that older adults tend to experience larger discrepancies among felt age, desired age, and chronological age than middle-aged persons (e.g., Montepare & Lachman, 1989). Results of our study corroborate the trends of these latter studies, suggesting that there are differences in SA between middle-aged and older adults. However, this lack of uniformity in the literature may stem from the age range that was examined and also with regard to whether studies focused on cross-sectional comparisons or on longitudinal changes in SA (Kornadt et al., 2016).

Correlates of Discrepancies in SA

As stated above, an important aim of the current study was to examine the correlates of SA in each one of its dimensions. Although we hypothesized that SA will be associated with knowledge about aging, this association was supported only for perceived old age. This finding might stem from the fact that the instrument used to assess knowledge—the Facts on Aging Quiz—assesses knowledge about aging processes and misconceptions rather than individual experiences and tendencies, explaining therefore the association with the more general dimension of SA and the lack of direct association with the more personal-individual dimensions of SA. This explanation is reinforced by the findings showing that, attitudes toward older adults mediate the association between knowledge about aging and discrepancy in SA, only in the perceived old age, that is, the general dimension of the concept. Thus, having more knowledge about aging might reduce the level of discrepancy in perceived old age, but indirectly, through the mediating effect of attitudes toward older adults: participants who were more knowledgeable about aging tended to report less negative attitudes toward older adults, and this reduction in negative attitudes toward older adults directly reduced the discrepancy in perceived old age. These findings support a previous study which found that negative attitudes toward older adults may lead to perceptions of younger SA, since people who perceive old age negatively would try to dissociate themselves from being old (Gwinner & Stephens, 2001).

Among the different sociodemographic variables, we found that subjective health status and a female gender were related to the different dimensions of SA. Indeed, subjective health status was consistently found to be related to SA (Kotter-Grühn et al., 2015). In addition, it has lately been shown that the pace of biological aging differs among individuals. Whereas some individuals age biologically according to their chronological age, others show accelerated increase or decrease in their aging process (Belsky et al., 2015). Therefore, it is not surprising that individuals may perceive themselves as younger or as older than their chronological age. Thus, our findings regarding subjective health status are in line with the biomedical perspective, stating that how young or old an individual feels is associated with his or her physical health and functioning (Kleinspehn-Ammerlahn et al., 2008). However, it should be noted that the evidence regarding gender-related differences in SA has been mixed. Whereas some researchers have found aging women to experience younger SAs, others have found no differences in men’s and women’s SAs (Arbeau, Galambos, & Jansson, 2007; Barak, Mathur, & Lee, 2001). Therefore, it is recommended that future research would evaluate the longitudinal effects of gender on SA.

Limitations of the Study

Several limitations should be noted. First, the use of a single item to measure each of the dimensions of SA may not reflect the whole dimension. Future studies, using multi-item scales to measure SA, are needed. Second, the quantitative survey method used did not allow respondents to expand on their views and experiences regarding their SA. Third, we used a multidimensional construct to examine SA while using a latent construct would have shown the different contribution of each dimension. Fourth, we did not assess the objective health status of the participants. Future research may account this information. Fifth, questionnaires were administered as face-to-face interviews which might affect social desirability. Finally, the use of a convenience sample does not allow for generalizability, nor does it provide an accurate representation of all middle-aged and older adults in Israel. Yet, respondents’ demographic background was similar to the general population at these age-groups in Israel at the time of the study (Central Bureau of Statistics, 2016).

Conclusions

Despite these limitations, the findings of the present study have both theoretical and practical implications. Theoretically, they strengthen the need to understand SA as a multidimensional construct. Our results demonstrated that adult persons’ perceptions regarding themselves and regarding old age in general are independent and need therefore, separate research and practical attention. Such a conceptualization might advance the understanding of the antecedents and consequences of SA, as we showed that it is composed of separate and unique dimensions. Going beyond previous findings demonstrating that domains of physical functioning account for unique SA variance (e.g., Kornadt et al., 2016), we also demonstrated other predictors of SA that may help to explain these dimensions’ variance. In addition, the age-group differences found in our study provide evidence for the different processes each group has in distancing themselves from old age and stress the need to relate to SA differently in each age-group. Practically, findings of the present study have practical implications at a societal and individual level. At the societal level, till today, chronological age is the only category used by societies to assign positions and responsibilities within given social structures. Our findings suggest that SA also plays a crucial role which should be taken into account while trying to make successful allocations of positions and responsibilities. At the individual level, findings of our study corroborate an increasing body of knowledge stressing the importance of subjective perceptions of age as a core element in adult persons’ well-being. Therefore, it might be worthwhile to incorporative different domains of aging or other age-related self-concepts, like health-related selves and past-future selves in intervention programs aimed at increasing well-being in a variety of adult persons’ groups.

Finally, the current study paves the way for future research aiming at better understanding of the sources of SA among middle-aged and older adults. In addition, further studies should be conducted to reassess the associations between knowledge about aging, attitudes toward the elderly and SA, using a longitudinal design.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Author Biographies

Shiri Shinan-Altman is a social worker and a researcher at Bar Ilan University, School of Social Work. Her research concern illness representations, social work in health care setting and formal and informal caregiving.

Perla Werner is a full professor in the Department of Community Mental Health at the University of Haifa, Israel. She has served as Vice Rector of the University and as Dean of the Faculty of Social Welfare and Health Sciences. Prof. Werner has published more than 250 peer-reviewed articles and received numerous competitive grants. Her areas of expertise are the psychosocial aspects of dementia.

References

Arbeau

K. J.

Galambos

N. J.

Jansson

S. K.

(2007). Dating, sex, and substance use as correlated of adolescents’ subjective experience of age. Journal of Adolescence, 30, 435–447. doi:10.1016/j.adolescence.2006.04.006

Arbuckle

J. L.

(2005). AMOS 6 user’s guide. Spring House, PA: AMOS Development Corporation.

Barak

(2009). Age identity: A cross-cultural global approach. International Journal of Behavioral Development, 33(1), 2–11.

Barak

Mathur

Lee

(2001). Perceptions of age-identity: A crosscultural inner-age exploration. Psychology & Marketing, 18, 1003–1029.

Barak

Stern

(1986). Subjective age correlates: A research note. Gerontologist, 26, 571–578.

Barclay

S. R.

Stoltz

K. B.

Chung

Y. B.

(2011). Voluntary midlife career change: Integrating the Transtheoretical Model and the life-span, life-space approach. The Career Development Quarterly, 59(5), 386–399.

Barrett

A. E.

Montepare

J. M.

(2015). It’s about time”: Applying life span and life course perspectives to the study of subjective age. Annual Review of Gerontology & Geriatrics, 35(1), 55–77.

Belsky

D. W.

Caspi

Houts

Cohen

H. J.

Corcoran

D. L.

Danese

Sugden

(2015). Quantification of biological aging in young adults. Proceedings of the National Academy of Sciences, 112, E4104–E4110.

Bodner

Ayalon

Avidor

Palgi

(2016). Accelerated increase and relative decrease in subjective age and changes in attitudes toward own aging over a 4-year period: Results from the Health and Retirement Study. European Journal of Ageing, 14(1), 17–27.

10.

Bowling

See-Tai

Ebrahim

Gabriel

Solanki

(2005). Attributes of age-identity. Aging & Society, 25, 479–500.

11.

Central Bureau of Statistics. (2016). Israeli population by age, religion and marital status. Jerusalem, Israel: Government Press.

12.

Cheung

G. W.

Lau

R. S.

(2008). Testing mediation and suppression effects of latent variables: Bootstrapping with structural equation models. Organizational Research Methods, 11(2), 296–325.

13.

Choi

N. G.

DiNitto

D. M.

(2014). Felt age and cognitive-affective depressive symptoms in late life. Aging & Mental Health, 18(7), 833–837. doi:10.1080/13607863.2014.886669

14.

Demakakos

Gjonca

Nazroo

(2007). Age identity, age perceptions, and health: Evidence from the English Longitudinal Study of Ageing. Annals of the New York Academy of Sciences, 1114, 279–287. doi:10.1196/annals.1396.021

15.

Galambos

N. L.

Albrecht

A. K.

Jansson

S. M.

(2009). Dating, sex, and substance use predict increases in adolescents’ subjective age across two years. International Journal of Behavioral Development, 33(1), 32–41.

16.

Garstka

T. A.

Schmitt

M. T.

Branscombe

N. R.

Hummert

M. L.

(2004). How young and older adults differ in their responses to perceived age discrimination. Psychology & Aging, 19(2), 326–335. doi:10.1037/0882-7974.19.2.326

17.

Gilleard

(2009). Old age in the dark ages: The status of old age during the early middle ages. Ageing & Society, 29(7), 1065–1084. ‏doi:10.1017/S0144686X09008630

18.

Hayes

A. F.

(2009). Beyond Baron and Kenny: Statistical mediation analysis in the new millennium. Communication Monographs, 76(4), 408–420.

19.

Hayes

A. F.

(2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. New York, NY: Guilford.

20.

Hubley

A. M.

Russell

L. B.

(2009). Prediction of subjective age, desired age, and age satisfaction in older adults: Do some health dimensions contribute more than others? International Journal of Behavioral Development, 33(1), 12–21. ‏doi:10.1177/0165025408099486

21.

Jeon

S. N.

Shin

H. G.

(2009). The effects of perceived health and ageism experience on successful aging. Journal of the Korean Gerontological Society, 29(4), 1383–1396.

22.

Kaufman

Elder

G. H.

(2002). Revisiting age identity: A research note. Journal of Aging Studies, 16(2), 169–176. ‏doi:10.1016/S0890-4065(02)00042-7

23.

Kertzner

R. M.

(2001). The adult life course and homosexual identity in midlife gay men. Annual Review of Sex Research, 12, 75–92.

24.

Kleinspehn-Ammerlahn

Kotter-Grühn

Smith

(2008). Self-perceptions of aging: Do subjective age and satisfaction with aging change during old age? The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 63(6), 377–385.

25.

Kornadt

A. E.

Hess

T. M.

Voss

Rothermund

(2016). Subjective age across the life span: A differentiated, longitudinal approach. The Journals of Gerontology Series B: Psychological Sciences & Social Sciences, 71(6), 1–11. doi:10.1093/geronb/gbw072

26.

Kotter-Grühn, D., Kleinspehn-Ammerlahn, A., Gerstorf, D., & Smith, J. (2009). Self-perceptions of aging predict mortality and change with approaching death: 16-year longitudinal results from the Berlin Aging Study. Psychology and Aging, 24(3), 654.

27.

Kotter-Grühn

Hess

T. M.

(2012). So you think you look young? Matching older adults’ subjective ages with age estimations provided by younger, middle-aged, and older adults. International Journal of Behavioral Development, 36(6), 468–475. doi:10.1177/0165025412454029

28.

Kotter-Grühn

Kornadt

A. E.

Stephan

(2015). Looking beyond chronological age: Current knowledge and future directions in the study of subjective age. Gerontology, 62(1), 86–93. ‏doi:10.1159/000438671

29.

Mock

S. E.

Eibach

R. P.

(2011). Aging attitudes moderate the effect of subjective age on psychological well-being: Evidence from a 10-year longitudinal study. Psychology & Aging, 26(4), 979–986. doi:10.1037/a0023877

30.

Montepare

J. M.

(2009). Subjective age: Toward a guiding lifespan framework. International Journal of Behavioral Development, 33(1), 42–46. doi:10.1177/0165025408095551

31.

Montepare

J. M.

Lachman

M. E.

(1989). You’re only as old as you feel”: Self-perceptions of age, fears of aging, and life satisfaction from adolescence to old age. Psychology & Aging, 4(1), 73–78. doi:10.1037/0882-7974.4.1.73

32.

Neikrug

S. M.

(1998). The value of gerontological knowledge for elders: A study of the relationship between knowledge on aging and worry about the future. Educational Gerontology: An International Quarterly, 24(3), 287–296.

33.

Packer

D. J.

Chasteen

A. L.

(2006). Looking to the future: How possible aged selves influence prejudice toward older adults. Social Cognition, 24(3), 218–247.

34.

Palmore

E. B.

(1998). The facts on aging quiz. New York, NY: Springer.

35.

Polizzi

K. G.

(2003). Assessing attitudes toward the elderly: Polizzi’s refined version of the Aging Semantic Differential. Educational Gerontology, 29(3), 197–216. doi:10.1080/713844306

36.

Porat

(2010). Attitudes, knowledge and exposure to elderly persons among medical students at one faculty of medicine in Israel. Gerontology, 37(4), 11–22.

37.

Rubin

D. C.

Berntsen

(2006). People over forty feel 20% younger than their age: Subjective age across the lifespan. Psychonomic Bulletin & Review, 13(5), 776–780. ‏doi:10.3758/BF03193996

38.

Sanderson

Scherbov

(2008). Rethinking age and aging. Washington, DC: Population Reference Bureau.

39.

Schafer

M. H.

Shippee

T. P.

(2010). Age identity, gender, and perceptions of decline: Does feeling older lead to pessimistic dispositions about cognitive aging? Journal of Gerontology: Social Sciences, 65B, 91–96. doi:10.1093/geronb/gbp046

40.

Sherman

S. R.

(1994). Changes in age identity: Self-perceptions in middle and late life. Journal of Aging Studies, 8, 397–412.

41.

Spuling

S. M.

Miche

Wurm

Wahl

H. W.

(2013). Exploring the causal interplay of subjective age and health dimensions in the second half of life. Zeitschrift für Gesundheitspsychologie, 21(1), 5–15. doi:10.1026/0943-8149/a000084

42.

Stephan, Y., Caudroit, J., & Chalabaev, A. (2011). Subjective health and memory self-efficacy as mediators in the relation between subjective age and life satisfaction among older adults. Aging & Mental Health, 15(4), 428-436.

43.

Stephan

Demulier

Terracciano

(2012). Personality, self-rated health, and subjective age in a life-span sample: The moderating role of chronological age. Psychology & Aging, 27(4), 875–880. doi:10.1037/a0028301

44.

Stephan

Sutin

A. R.

Terracciano

(2015). How old do you feel? The role of age discrimination and biological aging in subjective age. PLoS One, 10(3), e0119293. ‏doi:10.1371/journal.pone.0119293

45.

Stuart-Hamilton

Mahoney

(2003). The effect of aging awareness training on knowledge of, and attitudes towards, older adults. Educational Gerontology, 29(3), 251–260.

46.

Suh

Choi

Lee

Cha

(2012). Association between knowledge and attitude about aging and life satisfaction among older Koreans. Asian Nursing Research, 6(3), 96–101. ‏doi:10.1016/j.anr.2012.07.002

47.

Toothman

E. L.

Barrett

A. E.

(2011). Mapping midlife: An examination of social factors shaping conceptions of the timing of middle age. Advances in Life Course Research, 16(3), 99–111. ‏doi:10.1016/j.alcr.2011.08.003

48.

Uotinen

Suutama

Ruoppila

(2003). Age identification in the framework of successful aging. A study of older Finnish people. The International Journal of Aging & Human Development, 56(3), 173–195. doi:10.2190/6939-6W88-P2XX-GUQW

49.

Ward

R. A.

(2010). How old am I? Perceived age in middle and later life. The International Journal of Aging & Human Development, 71(3), 167–184. ‏doi:10.2190/AG.71.3.a

50.

Weiss

Lang

F. R.

(2012). ‘They’ are old but ‘I’ feel younger: Age-group dissociation as a self-protective strategy in old age. Psychology and Aging, 27(1), 153–163. doi:10.1037/a0024887

51.

Westerhof

G. J.

Barrett

A. E.

Steverink

(2003). Forever young? A comparison of age identities in the United States and Germany. Research on Aging, 25(4), 366–383. doi:10.1177/0164027503025004002

52.

Westerhof

G. J.

Miche

Brothers

A. F.

Wurm

(2014). The influence of subjective aging on health and longevity: A meta-analysis of longitudinal data. Psychology & Aging, 29(4), 793–802. ‏doi:10.1037/a0038016

53.

Wurtele

S. K.

Maruyama

(2013). Changing students’ stereotypes of older adults. Teaching of Psychology, 40, 59–61. doi:10.1177/0098628312465867