
==== Front
Gerontologist
Gerontologist
geront
The Gerontologist
0016-9013
1758-5341
Oxford University Press US

39051130
10.1093/geront/gnae088
gnae088
Measurement Article
AcademicSubjects/SOC02600
Exploring Older Adults’ Subjective Views on Aging Positively: Development and Validation of the Positive Aging Scale
https://orcid.org/0000-0002-1269-6856
Park Miriam Sang-Ah PhD School of Social Sciences, Nottingham Trent University, Nottingham, UK

Badham Stephen PhD School of Social Sciences, Nottingham Trent University, Nottingham, UK

Vizcaino-Vickers Samuel MSc School of Social Sciences, Nottingham Trent University, Nottingham, UK

Fino Emanuele PhD School of Psychology, Queen’s University Belfast, Belfast, Northern Ireland, UK

Gaugler Joseph E PhD, FGSA Decision Editor
Address correspondence to: Miriam Sang-Ah Park, PhD. E-mail: miriam.park@ntu.ac.uk
9 2024
25 7 2024
25 7 2024
64 9 gnae08829 8 2023
21 5 2024
05 9 2024
© The Author(s) 2024. Published by Oxford University Press on behalf of The Gerontological Society of America.
2024
https://creativecommons.org/licenses/by-nc/4.0/ This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial License (https://creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact reprints@oup.com for reprints and translation rights for reprints. All other permissions can be obtained through our RightsLink service via the Permissions link on the article page on our site—for further information please contact journals.permissions@oup.com.

Abstract

Background and Objectives

Historically, aging research has focused primarily on health deterioration and negative aspects associated with aging. This has limited the scope of our understanding of the experience of aging and the relationships between aging and well-being from an integrative biopsychosocial perspective. In the same vein, there is a lack of reliable and valid assessments of aging that capture the positive aspects that characterize and improve the subjective experience of this period of life, particularly one that focuses on psychosocial well-being, including meaningful experiences and activities, group memberships, and general abilities. This study presents the development and validation of the Positive Aging Scale (PAS), a novel self-report assessment.

Research Design and Methods

This was an online cross-sectional study conducted on 501 UK residents aged ≥60 years. A number of self-reported items and measures of positive aging, general health, well-being, and cognitive functioning were administered to the sample. We used exploratory and confirmatory factor analyses and assessed the dimensionality, reliability, and concurrent criterion-related validity of the PAS.

Results

The results suggested that a unidimensional solution represents the data well, with the positive aging factor adequately loading on 8 items, and the solution showing factorial invariance between young-old and old participants (i.e., ≥75 years). Total PAS scores positively correlate with general health, well-being, and cognitive functioning.

Discussion and Implications

The PAS demonstrated strong psychometric properties and the findings highlight correlations between the PAS and key outcomes of positive aging, including general health. Implications for research and interventions are discussed.

Aging
Assessment
Older adults
Positive aging
Well-being
Independent Social Research Foundation 10.13039/100009838
==== Body
pmcBackground and Objectives

There is an increasing recognition within aging and geriatric medicine of the importance of combining objective and subjective measures of positive aging to gauge a comprehensive understanding of the overall quality of life of individuals (Gordon & Hubbard, 2022). Conversely, traditional aging research had focused primarily on the biomedical aspects of aging (Otto et al., 2023). Literature on successful aging (Rowe & Khan, 1997) and healthy aging (World Health Organization, 2015) highlighted the need for moving beyond mere physiological outlooks, focusing on alternative definitions and conceptualizations that emphasize the role of positive aspects of aging, such as those tapping into facets of psychosocial well-being, including engaging with meaningful experiences and activities, group memberships, and general abilities.

Positive psychologists have long noted the importance of those facets and their impact on several life outcomes, including individuals’ physical and mental health, as well as the quality of their attachments, group membership, and relationships (Diener et al., 2002). Recent literature has shown that individuals’ perception of their own aging significantly affects their confidence and general health as older adults (Tully-Wilson et al., 2021). In the same vein, other studies that focused on self-perceptions of aging indicated that individuals who tend to hold a more positive attitude toward their own aging present with increased longevity (Levy et al., 2002), better health behaviors such as healthier diets and exercise routines (Hooker et al., 2019), lower levels of depressive symptoms (Han & Richardson, 2015), and lower odds of developing dementia (Siebert et al., 2018). Notably, most of these studies were conducted by measuring individuals’ perception through the Attitude Toward Own Aging (ATOA) scale (Lawton, 1975), a five-item self-report measure. Although showing satisfactory psychometric properties, the ATOA measures a general attitude toward one’s own aging (e.g., “I am as happy now as when I was younger”), but its brevity and scope do not allow for a comprehensive assessment of the wider construct of positive aging from a biopsychosocial perspective, highlighting a significant gap in the current assessment literature.

Psychosocial well-being in old age has been found to be related with individuals’ commitment to and engagement with meaningful experiences and activities, group memberships and sense of belonging and agency derived from them, and their overall preserved general abilities. In this regard, a recent study conducted on a U.S. nationally representative sample (N = 11,382 aged ≥65 years) has found lower rates of engagement in meaningful activities in participants with either disability, dementia, or depression, compared with counterparts without these conditions (Oh et al., 2021). Research has also shown that membership of and sense of belonging to one’s family and community life predict positive aging and contribute to reduce cognitive decline (Engelhardt et al., 2010). Consistently, Fernandez-Ballesteros (2011) suggested the need to recognize the role of such positive facets in the definition of positive aging, as well as their relationships with health and well-being, beyond negative and deteriorative aspects associated with this period of life.

The present study endorses a definition of positive aging that is rooted in the holistic perspective of older adults’ overall well-being. This perspective recognizes the role of meaningful everyday experiences, group memberships, personal resources, and extrinsic factors in determining positive aging (Annele et al., 2019; Cosco et al., 2014; Lewis 2014). This is in line with the conceptualization proposed by Fernández-Ballesteros (2011), highlighting the necessity to look upon the positive aspects of aging that emphasize the role of older adults’ sense of belonging and agency to gauge an integrative perspective over their experience of this period of life, besides other equally important physiological markers.

Based on such definition, the present study aimed to improve the assessment of positive aging by developing and validating the Positive Aging Scale (PAS), a novel self-report assessment. Specifically, the main objective of the present study was to develop and validate a measure tapping into facets of psychosocial well-being, focusing on meaningful everyday experiences, group memberships, and general ability. First, we conducted a pilot study to gather older adults’ own thoughts on what positive aging means and developed an initial pool of items. In this pilot study, we conducted semistructured interviews with older adults asking them directly and openly about positive aging, and questions included “what does positive aging or aging well meant to you?” and “can you tell us about someone you know that you think is aging positively and tell me why you think so?” The initial pool of items then comprised our initial scale of positive aging and the questionnaire including the scale was administered to a sample of older adults from the UK general population. Second, we investigated the dimensionality of the scale and tested for factorial invariance of the solution between young-old and old (≥75 years old) individuals to ensure adequate scale functioning in both groups. Finally, we tested the reliability and concurrent criterion-related validity of the PAS, hypothesizing significant associations between positive aging and general health, well-being, and cognitive functioning.

Research Design and Methods

Study Design and Procedure

This was a cross-sectional study in a sample of UK-resident older adults. We used convenience sampling and recruited participants via emails, social media outlets, and community websites. Participation was entirely voluntary, with no incentives being provided. Individuals meeting the following inclusion criteria were recruited for the study: (a) they had to be at least 60 years old or older; (b) they had to be self-reportedly fluent in English, so to be able to understand and complete the survey; and (c) they had to be a resident in the United Kingdom. Moreover, to take part in the study, interested candidates were asked to read, understand, and sign informed consent by means of a dedicated electronic form. The procedure lasted approximately 15 min and consisted of completing a set of self-report measures through Qualtrics (qualtrics.com). The study was reviewed and received favorable ethics approval by Schools of Business, Law and Social Sciences Research Ethics Committee at Nottingham Trent University.

Participants

Five hundred forty-eight UK residents aged ≥60 were recruited. However, preliminary data screening led us to retain and use a final set of 501 valid observations (more details in the Results section). Table 1 reports the distribution of participants by age, gender, ethnicity, marital status, employment, education, and household condition. As mentioned in the previous paragraphs, we considered two main groups in relation to their age, namely young-old and old, with cut-point set at 75 years of age (see Cohen-Mansfield, 2013).

Table 1. Participants’ Sociodemographic Characteristics (N = 501)

	<75 years old	≥75 years old	Overall	
Demographics	Female (N = 250)	Male (N = 86)	Female (N = 116)	Male (N = 48)	Nonbinary (N = 1)	Female (N = 366)	Male (N = 134)	Nonbinary (N = 1)	
Age	
 Mean (SD)	68.2 (4.09)	68.2 (4.15)	78.9 (3.91)	79.5 (4.45)	78.0 (NA)	71.6 (6.41)	72.2 (6.89)	78.0 (NA)	
Ethnicity	
 African	1 (0.4%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	1 (0.3%)	0 (0%)	0 (0%)	
 Any other Asian background	1 (0.4%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	1 (0.3%)	0 (0%)	0 (0%)	
 Any other ethnic background	2 (0.8%)	1 (1.2%)	0 (0%)	0 (0%)	0 (0%)	2 (0.5%)	1 (0.7%)	0 (0%)	
 Any other mixed/multiple ethnic background	1 (0.4%)	2 (2.3%)	0 (0%)	1 (2.1%)	0 (0%)	1 (0.3%)	3 (2.2%)	0 (0%)	
 Bangladeshi	0 (0%)	0 (0%)	1 (0.9%)	0 (0%)	0 (0%)	1 (0.3%)	0 (0%)	0 (0%)	
 Caribbean	1 (0.4%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	1 (0.3%)	0 (0%)	0 (0%)	
 White and Asian	2 (0.8%)	1 (1.2%)	0 (0%)	0 (0%)	0 (0%)	2 (0.5%)	1 (0.7%)	0 (0%)	
 White British	222 (88.8%)	74 (86.0%)	111 (95.7%)	47 (97.9%)	1 (100%)	333 (91.0%)	121 (90.3%)	1 (100%)	
 White Other	20 (8.0%)	8 (9.3%)	4 (3.4%)	0 (0%)	0 (0%)	24 (6.6%)	8 (6.0%)	0 (0%)	
Marital status	
 Divorced	43 (17.2%)	6 (7.0%)	19 (16.4%)	2 (4.2%)	0 (0%)	62 (16.9%)	8 (6.0%)	0 (0%)	
 Married or in a domestic partnership	140 (56.0%)	70 (81.4%)	45 (38.8%)	34 (70.8%)	1 (100%)	185 (50.5%)	104 (77.6%)	1 (100%)	
 Separated	5 (2.0%)	2 (2.3%)	3 (2.6%)	1 (2.1%)	0 (0%)	8 (2.2%)	3 (2.2%)	0 (0%)	
 Single (never married)	22 (8.8%)	6 (7.0%)	4 (3.4%)	2 (4.2%)	0 (0%)	26 (7.1%)	8 (6.0%)	0 (0%)	
 Widowed	40 (16.0%)	2 (2.3%)	45 (38.8%)	9 (18.8%)	0 (0%)	85 (23.2%)	11 (8.2%)	0 (0%)	
Employment	
 Employed full-time (≥40 hr/week)	7 (2.8%)	1 (1.2%)	0 (0%)	0 (0%)	0 (0%)	7 (1.9%)	1 (0.7%)	0 (0%)	
 Employed part-time (<40 hr/week)	13 (5.2%)	1 (1.2%)	0 (0%)	1 (2.1%)	0 (0%)	13 (3.6%)	2 (1.5%)	0 (0%)	
 Unemployed and currently looking for work	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	
 Unemployed and not currently looking for work	3 (1.2%)	1 (1.2%)	0 (0%)	0 (0%)	0 (0%)	3 (0.8%)	1 (0.7%)	0 (0%)	
 Student	1 (0.4%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	1 (0.3%)	0 (0%)	0 (0%)	
 Retired	217 (86.8%)	77 (89.5%)	113 (97.4%)	47 (97.9%)	1 (100%)	330 (90.2%)	124 (92.5%)	1 (100%)	
 Homemaker	1 (0.4%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	1 (0.3%)	0 (0%)	0 (0%)	
 Self-employed	6 (2.4%)	4 (4.7%)	3 (2.6%)	0 (0%)	0 (0%)	9 (2.5%)	4 (3%)	0 (0%)	
 Unable to work	2 (0.8%)	2 (2.3%)	0 (0%)	0 (0%)	0 (0%)	2 (0.5%)	2 (1.5%)	0 (0%)	
Education	
 Associate degree	13 (5.2%)	2 (2.3%)	12 (10.3%)	5 (10.4%)	0 (0%)	25 (6.8%)	7 (5.2%)	0 (0%)	
 Bachelor’s degree	73 (29.2%)	29 (33.7%)	21 (18.1%)	19 (39.6%)	0 (0%)	94 (25.7%)	48 (35.8%)	0 (0%)	
 Doctorate	5 (2.0%)	5 (5.8%)	4 (3.4%)	4 (8.3%)	0 (0%)	9 (2.5%)	9 (6.7%)	0 (0%)	
 High school degree or equivalent	24 (9.6%)	5 (5.8%)	17 (14.7%)	4 (8.3%)	0 (0%)	41 (11.2%)	9 (6.7%)	0 (0%)	
 Less than a high school diploma	21 (8.4%)	6 (7.0%)	17 (14.7%)	1 (2.1%)	0 (0%)	38 (10.4%)	7 (5.2%)	0 (0%)	
 Master’s degree	48 (19.2%)	15 (17.4%)	14 (12.1%)	8 (16.7%)	0 (0%)	62 (16.9%)	23 (17.2%)	0 (0%)	
 Professional degree	18 (7.2%)	8 (9.3%)	2 (1.7%)	2 (4.2%)	0 (0%)	20 (5.5%)	10 (7.5%)	0 (0%)	
 Some college, no degree	48 (19.2%)	16 (18.6%)	29 (25.0%)	5 (10.4%)	1 (100%)	77 (21.0%)	21 (15.7%)	1 (100%)	
Household condition	
 Alone	90 (36.0%)	14 (16.3%)	67 (57.8%)	14 (29.2%)	0 (0%)	157 (42.9%)	28 (20.9%)	0 (0%)	
 Couple living together	142 (56.8%)	61 (70.9%)	44 (37.9%)	32 (66.7%)	1 (100%)	186 (50.8%)	93 (69.4%)	1 (100%)	
 Living with 1–3 others	11 (4.4%)	9 (10.5%)	4 (3.4%)	2 (4.2%)	0 (0%)	15 (4.1%)	11 (8.2%)	0 (0%)	
 Living with ≥4 others	7 (2.8%)	2 (2.3%)	1 (0.9%)	0 (0%)	0 (0%)	8 (2.2%)	2 (1.5%)	0 (0%)	
Note: SD = standard deviation.

Measures

The PAS is the novel self-report assessment of positive aging. The version of the scale used for testing included 18 items. We developed the PAS by way of a multiple-step process. This started with interviewing a sample of older adults from the local community, using a protocol for identifying and discussing major antecedents, manifestations, and implications of positive aging. Semistructured interviews were conducted between January and July 2021 by the first author of the paper, who had previously undergone dedicated training. The protocol was developed by the researchers coauthoring the present paper and underwent peer review with colleagues from the first authors’ department.

The main investigator and first author of the paper wrote and iteratively reviewed the PAS items, drawing upon the findings from the semistructured interviews, which highlighted the importance of the following content categories: psychosocial well-being, meaningful activities, group memberships, and abilities. A total of 23 items were initially developed, reflecting older adults’ narratives and personal definitions of positive aging. This initial version of the scale was submitted for peer review within the wider local research team, composed of experts in psychology of aging and mental health. Upon discussion, to minimize repetition, some of these items were either removed or merged with other items.

Finally, a pool of 18 items was deemed the most appropriate for use. Participants were asked to rate how much they agreed or disagreed with items such as: “Have meaningful social interactions with people regardless of their geographical distance from me” and “Engage in things and activities that are worthwhile or valuable in my opinion.” The scale originally used a 6-point response option scale, ranging from 1 (“disagree strongly”) to 6 (“agree strongly”).

The Cognitive Failures Questionnaire (CFQ; Broadbent et al., 1982), a 25-item measure of experiences of cognitive failures in everyday life, such as slips and errors of perception, memory, and motor functioning. Items include: “Do you find you forget appointments?” and “Do you find you forget people’s names?” Previous studies showed satisfactory test–retest reliability (rtt = 0.80), from 6 to 65 weeks (Broadbent et al., 1982). We averaged across reverse-coded individual items’ ratings to obtain a general cognitive functioning score.

The Mental Health Continuum-Short Form (MHC-SF; Lamers et al., 2011) is a 14-item, self-report measure of hedonic, eudaimonic social, and eudaimonic psychological well-being. Individuals are asked to rate how often, in the past month, they felt as described by each of the 14 items, using a scale from 0 (“never”) to 5 (“everyday”). Items included “Felt that you had something important to contribute to society” and “Felt good at managing the responsibilities of your daily life.” Previous research showed values of Cronbach’s alpha above 0.74 across the three factors, in multiple countries (Lamers et al., 2011).

A single-item, self-reported question requiring participants to indicate their perceived level of General Health, using a response option scale ranging from 1 (“very poor”) to 5 (“very good”; Wu et al., 2013).

Analytical Plan

Data were initially screened by evaluating (a) missing responses, (b) disengaged response patterns, (c) multivariate outliers, and (d) item response categories. Disengaged responses were detected by computing response variance (SD < 0.30) across all initial PAS items. Multivariate outliers were detected through robust Mahalanobis’ distance (p < .01). There were no missing responses. This method uses correspondence analysis to define Mahalanobis distances, relying on singular vectors and the determinant from the eigenvalues obtained through correspondence analysis. Further, we examined item response categories, monitoring the patterns of participants’ use of the response categories. Listwise deletion led to removal of 35 cases, and 12 cases were identified as multivariate outliers and were therefore removed from the final analyses. After identifying and removing these problematic data points, we divided the remaining sample into two stratified random portions, by gender and age group (i.e., ≥75 years). The first subsample was used to run dimensionality tests (n1 =  250) and the second subsample to run confirmatory multigroup invariance analyses (n2 = 251). Finally, the entire sample was used to test for concurrent criterion-related validity analyses.

Model evaluation and selection followed the following steps. First, we examined the polychoric correlation matrix obtained on the entire sample, particularly, looking at potential multicollinearity and removing items that presented high (>0.60) correlations with several others and whose descriptors could indicate redundancy. Then, we used this revised version of the PAS in dimensionality tests and exploratory factor analysis (EFA), using factor loadings, commonalities, and the interpretability of the pattern matrix as criteria to inform our evaluation. The resulting model from EFA was finally tested via confirmatory factor analysis (CFA). We then used criteria such as residual intercorrelations, factor loadings, and model fit to inform item selection, leading to the final version of the scale.

To test the dimensionality of the PAS, several concurrent methods were used, described as follows. First, parallel analysis with principal axis factoring, a technique comparing the eigenvalues from the empirical correlation matrix to the eigenvalues calculated on randomly generated data sets, with eigenvalues representing the variance explained by each factor was conducted. Factors with empirical eigenvalues greater than those obtained at random are considered for retention. Secondly, Velicer’s simple structure (VSS) method was used, comparing several alternative factors solutions to a specific solution accounting for the highest-loading item, whereas all other loadings are fixed to zero. The test is iterative, repeating over several alternative k-factor solutions. Here, the best candidates for retention are those identified by the solution that maximizes the fit between the simplified pattern matrix and the relevant empirical matrix. Third, Velicer’s minimum average partial (MAP) test. MAP evaluates average squared correlations after partialling out factors from 1 to k − 1 (k number of items), with the correlation matrix that minimizes residual variance being considered the best candidate for retention. Fourth, EFA was conducted to assess for the number of factors considered as best candidates for retention. A series of alternative solutions were compared, considering factor loadings, total variance explained (≥50%), and commonalities (≥0.70) as main interpretation criteria. In particular, we considered satisfactory loadings as those with a ≥0.3 magnitude on the relevant factor showing and not cross-loading on multiple factors (Costello & Osborne, 2005). In this regard, we must emphasize that cutoff values for loadings and cross-loadings are somewhat arbitrary, with alternative recommendations being available in the literature. The minimum cross-loading magnitude of 0.30 explains approximately 10% of overlapping variance with other items in a given factor (Tabachnick & Fidell, 2013). However, in light of the exploratory nature of this analysis, we considered 0.30 as minimally acceptable for EFA, whereas we required strong loadings (i.e., ≥0.5) in confirmatory analyses. We used polychoric correlations and weighted least squares estimation to account for the ordinal nature of the data.

Subsequently, CFA was conducted with mean- and variance-adjusted weighted least squares estimation, robust standard errors, and Wu and Estabrook’s (2016) identification method to assess the fit of the measurement model to the data. The following fit indices and criteria were used: comparative fit index (CFI) ≥ 0.95, root mean square error of approximation (RMSEA) < 0.06, and the standardized root mean square residual (SRMR) < 0.08 (Kenny, 2022). We evaluated factor loadings and residual covariances to establish the model with best fit to the data. The reliability of the scale and of all the measures in the analyses were assessed by means of McDonald’s omega coefficient (1999; see also Green & Yang, 2009).

To investigate the factorial invariance of the PAS between the two previously defined age groups, a multigroup CFA with ordinal data was conducted (Jorgensen et al., 2019; Svetina et al., 2020). This consisted of evaluating a series of progressively constrained models, in the following order: (a) a baseline model fitted to data from both groups (i.e., configural model); (b) a model with factor loadings and response category thresholds constrained to equality between the groups (metric model); a model with factor loadings, thresholds, and intercepts constrained to equality between groups (scalar model). We used ≥0.10 ΔCFI difference and the likelihood ratio test (alpha = 0.01) to compare the fit of the nested models. Lastly, for concurrent criterion-related validity analyses, we used Spearman’s correlation coefficient (alpha = 0.05).

All the analyses were run in R version 4.2.2 (R Core Team, 2022), with the following packages: lavaan (Rosseel, 2012), lavaanPlot (Lishinski, 2021), MBESS (Kelley, 2022), psych (Revelle, 2022), OuRS (Sunderland & Beaton, 2023), and semTools (Jorgensen et al., 2019).

Results

Table 2 presents univariate descriptive statistics (N = 501).

Table 2. Descriptive Statistics (N = 501)

Item	Frequencies of responses (%) per category	Descriptive statistics	
1	2	3	4	5	6	M	SD	Median	Min	Max	Skewness	Kurtosis	
PAS 1	0.00	0.01	0.02	0.23	0.20	0.54	5.24	0.94	6.00	1.00	6.00	−0.99	0.41	
PAS 2	0.01	0.04	0.07	0.28	0.30	0.31	4.73	1.13	5.00	1.00	6.00	−0.76	0.31	
PAS 3	0.00	0.01	0.02	0.26	0.25	0.46	5.10	0.98	5.00	1.00	6.00	−0.85	0.28	
PAS 4	0.01	0.02	0.06	0.29	0.26	0.36	4.86	1.08	5.00	1.00	6.00	−0.69	0.11	
PAS 5	0.01	0.01	0.03	0.23	0.13	0.59	5.25	1.04	6.00	1.00	6.00	−1.25	1.26	
PAS 6	0.00	0.00	0.01	0.25	0.19	0.55	5.27	0.90	6.00	2.00	6.00	−0.76	−0.65	
PAS 7	0.01	0.02	0.05	0.18	0.28	0.46	5.07	1.09	5.00	1.00	6.00	−1.19	1.14	
PAS 8	0.00	0.03	0.07	0.18	0.24	0.48	5.06	1.12	5.00	1.00	6.00	−1.07	0.44	
PAS 9	0.00	0.01	0.05	0.22	0.28	0.45	5.11	0.95	5.00	2.00	6.00	−0.73	−0.37	
PAS 10	0.01	0.02	0.07	0.23	0.27	0.41	4.95	1.11	5.00	1.00	6.00	−0.92	0.47	
PAS 11	0.00	0.00	0.02	0.25	0.35	0.38	5.08	0.85	5.00	2.00	6.00	−0.46	−0.58	
PAS 12	0.00	0.02	0.05	0.24	0.33	0.36	4.96	0.99	5.00	1.00	6.00	−0.76	0.20	
PAS 13	0.01	0.01	0.05	0.22	0.33	0.38	5.02	0.99	5.00	1.00	6.00	−0.94	0.93	
PAS 14	0.02	0.03	0.03	0.15	0.19	0.58	5.20	1.19	6.00	1.00	6.00	−1.66	2.43	
PAS 15	0.03	0.11	0.26	0.32	0.23	0.06	3.79	1.16	4.00	1.00	6.00	−0.17	−0.42	
PAS 16	0.01	0.03	0.09	0.26	0.32	0.29	4.71	1.12	5.00	1.00	6.00	−0.72	0.16	
PAS 17	0.00	0.02	0.04	0.19	0.26	0.49	5.17	0.99	5.00	2.00	6.00	−1.03	0.43	
PAS 18	0.00	0.00	0.01	0.14	0.18	0.67	5.50	0.78	6.00	1.00	6.00	−1.47	1.93	
Notes: PAS = Positive Aging Scale; SD = standard deviation.

In addition, we found 14 out of 18 items for which the aggregated frequency of response for the adjacent response categories 1–3 was below 10%. For this reason, we decided to collapse those response categories across the entire set of PAS items, reducing the response option scale to 1–4 (recommended anchors for future research using the PAS: 1 = “disagree strongly,” 4 = “agree strongly”). Subsequently, we used a random split to define two subsamples (n1 = 250; n2 = 251), respectively, used in dimensionality/exploratory and confirmatory analyses.

The polychoric correlation matrix (N = 501) showed substantial multicollinearity, that is, several high (≥0.70) intercorrelations. For this reason, we decided to remove problematic items from further analyses, specifically, items 3, 5, 6, 11, 12, and 18.

Reliability estimates for the additional measures used in the study are detailed as follows: CFQ (ω = 0.92; 95% confidence interval [CI] = 0.90–0.93), MHC-SF hedonic well-being (ω = 0.84; 95% CI = 0.81–0.87), MHC-SF eudemonic (social) well-being (ω = 0.77; 95% CI = 0.74–0.80), and eudaimonic (psychological) well-being (ω = 0.83; 95% CI = 0.81–0.86). Detailed results from the analysis of the reliability of the PAS are provided in the following paragraphs.

Exploratory Analyses

Parallel analysis suggested four factors with empirical eigenvalues (Figure 1).

Figure 1. Parallel analysis, scree plot (n1 = 242).

However, this solution was challenged by the result from further VSS and MAP tests. In fact, VSS showed the best performance of the unidimensional model (Max VSS = 0.85). Similarly, MAP pointed out to unidimensional model (Min MAP = 0.03) as the top-performing solution. Based on these results, we ran and compared a series of EFAs by extracting, respectively, from one to four factors (KMO = 0.88; Bartlett’s test: Chi Square66 = 511.77, p < .001). The results are presented in Table 3.

Table 3. Weighted Least Squares Exploratory Factor Analysis (n1 = 250)

Item	One-factor	Two-factor	Three-factor	Four-factor	
F1	H2	F1	F2	H2	F1	F2	F3	H2	F1	F2	F3	F4	H2	
PAS 1	0.61	0.36	0.57		0.39			0.84	0.71			0.85		0.75	
PAS 2	0.48	0.22	0.56		0.30			0.64	0.43			0.66		0.43	
PAS 4	0.60	0.34	0.55		0.37			0.37	0.37			0.34		0.36	
PAS 7	0.73	0.52	0.75		0.61		0.74		0.74		0.82			0.80	
PAS 8	0.58	0.35	0.76		0.47		0.83		0.65		0.75			0.60	
PAS 9	0.67	0.45	0.56		0.47				0.44				0.98	0.97	
PAS 10	0.68	0.45	0.40	0.36	0.44	0.41		0.34	0.47	0.34				0.47	
PAS 13	0.80	0.64	0.39	0.52	0.71	0.62			0.68	0.53				0.67	
PAS 14	0.59	0.36		0.66	0.47	0.71			0.50	0.66				0.51	
PAS 15	0.41	0.18		0.72	0.42	0.67			0.32	0.65				0.34	
PAS 16	0.70	0.50	0.33	0.47	0.53	0.59	0.34		0.59	0.54	0.42			0.62	
PAS 17	0.73	0.54	0.34	0.50	0.57	0.57			0.57	0.46				0.58	
Proportion of variance explained (%)	41.03		28.13	19.06		22.45	16.06	15.36		18.11	15.40	13.94	11.80		
Notes: FN values represent standardized factor loadings (values < 0.30 were suppressed). H2 values represent commonalities.

Each of the multidimensional models represented a poorly interpretable solution, showing fewer than three items being loaded on by their relevant factor and several cross-loadings, challenging their overall meaningfulness and interpretability. Conversely, the unidimensional solution showed all the PAS items being loaded ≥0.41 by a single positive aging factor, although with some commonalities being lower than 0.30. Based on these exploratory findings, we retained the unidimensional solution for further inspection.

CFA and Invariance Testing

The results from CFA showed that the unidimensional model from EFA had unsatisfactory fit to the data (CFI = 0.96; RMSEA [90% CI] = 0.09 [0.07–0.11]; SRMR = 0.07). After a series of iterations based on the previously indicated criteria, we found that the model accounting for eight of the original items was the best to represent the data (in the interest of brevity, we only report the retained items’ numbers: 1, 4, 7, 9, 10, 13, 14, and 17; we refer the reader to the data and code shared online, for further details). The finalized scale can be found as a supplementary file (PAS; Supplementary File S1). The fit of the model was excellent (CFI = 0.99; RMSEA = 0.05 [0.02–0.08]; SRMR = 0.04). All factor loadings on the general positive aging factor were high (i.e., ≥0.56; Figure 2) and the single positive aging factor showed satisfactory reliability (ω = 0.86; 95% CI = 0.83–0.89).

Figure 2. Confirmatory factor analysis (n2 = 241), standardized factor loadings.

Based on these results, we proceeded to the test of invariance of the unidimensional PAS model. We found <0.1 ΔCFI and no statistically significant differences between models nested for their relevant constraints, indicating metric and scalar invariance between young-old and older adults. Table 4 presents detailed findings from invariance tests.

Table 4. Multiple Group CFA, Invariance Testing Between Young-Old (n = 168) and Old (n = 83) Participants

Model	Constrained parameters	CFI	ΔCFI	RMSEA	SRMR	Chi-square	df	ΔChi-square	Δdf	p Value	
Configural	No constraints.	0.99	0	0.06	0.05	29.68	40	NA	NA	NA	
Metric	Loadings, thresholds.	0.99	0	0.05	0.05	42.30	55	13.30	15	.579	
Scalar	Loadings, thresholds, intercepts.	0.99	0	0.05	0.05	51.92	62	5.86	7	.556	
Notes: CFA = confirmatory factor analysis.

Criterion-Related Validity

Spearman’s correlation analyses showed positive correlations between PAS scores and cognitive functioning (respectively: rs = 0.21, p < .001), general health (rs = 0.20, p < .001), hedonic well-being (rs = 0.20, p < .001), eudaimonic (social) well-being (rs = 0.13, p < .01), and eudaimonic (psychological) well-being (rs = 0.25, p < 0.001), respectively.

Discussion

The present study tested and validated the PAS, a novel self-report assessment based on a biopsychosocial conceptualization of positive aging. Using the newly developed scale, we tested positive aging in older adults and examined its relationship to measures of general health, well-being, and cognitive functioning. The results pointed toward a unidimensional solution as the best representation of the data, with the single factor loading on eight items and showing acceptable reliability. The solution was age group-invariant and PAS scores positively correlated with general health, well-being, and cognitive function, demonstrating its relevance for testing outcomes related to positive aging.

The findings pointed toward engaging in meaningful and stimulating activities, keeping close social ties and community memberships, and continuing to enjoy things one values in life as significant indicators of psychosocial well-being and general functioning in older adults. Such aspects of positive aging and their potential impact on well-being in later life are echoed in recent literature. For instance, Heinz et al. (2023) showed that older adults found meaning and purpose in life by actively and mindfully engaging in daily activities and nurturing a goal-oriented mindset. Lu et al. (2022) emphasized the role of the “social capital” in aging well, that is, the importance of social relations and the meaning of group memberships in fostering a sense of belonging that positively affects the objective and subjective experience of healthy aging.

The PAS will be instrumental for further research on healthcare policy, prevention, and intervention in the community. In fact, until recently, most interventions had been characterized by a restricted definition of aging, focusing on older adults’ health issues and mere physiological outcomes rather than their holistic well-being (Uchino & Rook, 2020). The PAS captures a measure of aging well that extends beyond retaining independence (cf., Watkins et al., 2023) to encompass remaining connected to and supported by family and society. Interventions for older adults that are more nuanced are much needed, specifically those targeting factors that may be closely associated with or traceable with positive aging. The PAS therefore offers a more balanced assessment of positive aging that can be used to evaluate the potential impact of preventative interventions, beyond a mere focus on the negative detrimental influences of age on individuals’ well-being. Nevertheless, we must acknowledge that the evidence here presented needs to be corroborated by further examination of the PAS across other samples and contexts.

Limitations and Suggestions for Future Research

This study presents several limitations. First, it did not encompass aspects of biological or physical functioning that are concerned with positive aging, as it solely focuses upon the psychosocial variables (Fernandez-Ballesteros, 2011). Second, further research would benefit from using longitudinal designs to assess the extent to which the PAS can measure variations of positive aging in time, especially by distinguishing between the intra- and interindividual level. Third, in light of the importance of culture and the role of collective values, norms, and beliefs associated with positive aging across different contexts, the UK-based sample utilized in the present study represents a major threat to the cross-cultural generalizability of the PAS. Fourth, differences between PAS scores and objective measures of positive aging (e.g., those associated with individuals’ physical health) require further investigation (Fernandez-Ballesteros, 2011). Fifth, criterion validity but not convergent validity of the PAS was evaluated, nor was the relationship between PAS scores and ATOA scores addressed. This could have served to clarify specificity of the PAS in assessing positive aging, compared to attitudes toward own aging, warranted for future research.

Future research would benefit from further examining the criterion-related validity of the PAS, for example—but not limited to—by exploring the relationships between PAS scores and objective measures of physical and mental health in the older adults. In particular, whether PAS scores can predict outcomes and markers of healthy aging warrants further study, such as associations with indices of functional impairment and disability and biomarkers like blood pressure and lipids. Also, cross-cultural examination of the property of the PAS is much needed, especially because the definition of positive aging differs widely across contexts (Abolhasani & Bastani, 2019; Bowling, 2006; Cosco et al., 2013). For this reason, research on the theoretical underpinnings of positive aging defined and operationalized through the PAS will expand our current understanding of the role of culture on positive aging.

Conclusion

The present findings highlight the psychometric property of the PAS, a novel self-report assessment of positive aging that draws upon a biopsychosocial perspective over this period of life. The PAS is invariant between young-old and older adults and total scores correlate positively with measures of general health, well-being, and cognitive functioning.

Supplementary Material

gnae088_suppl_Supplementary_Materials

Funding

This research was funded by the Independent Social Research Foundation.

Conflict of Interest

None.

Data Availability

The study reported in the manuscript was not preregistered. The data, analytic methods, and R code materials are available to other researchers for replication purposes and can be accessed at https://osf.io/ws2dy/?view_only=1855f938b65e4d9086ca6950c1276774.

Author Contributions

Miriam Sang-Ah Park (Conceptualization [equal], Data curation [equal], Investigation [equal], Methodology [equal], Project administration [equal], Supervision [equal], Validation [equal], Writing—original draft [equal], Writing—review & editing [equal]), Stephen Badham (Conceptualization [equal], Methodology [equal], Project administration [equal], Supervision [equal], Writing—original draft [equal]), Samuel Vizcano-Vickers (Investigation [equal], Project administration [equal], Writing—review & editing [equal]), and Emanuele Fino (Formal analysis [equal], Validation [equal], Visualization [equal], Writing—original draft [equal])
==== Refs
References

Abolhasani, F., & Bastani, F. (2019). Successful ageing in the dimensions of life satisfaction and perception of ageing in the Iranian elderly adults referring to the health center in the west of Tehran, Iran. Iran Journal of Nursing, 31 (116 ), 61–74. 10.29252/ijn.31.116.61
Annele, U., Satu, K. J., & Timo, E. S. (2019). Definitions of successful ageing: A brief review of a multidimensional concept. Acta Bio Medica: Atenei Parmensis, 90 (2 ), 359. 10.23750/abm.v90i2.8376
Bowling, A. (2006). Lay perceptions of successful ageing: Findings from a national survey of middle aged and older adults in Britain. European Journal of Ageing, 3 (3 ), 123–136. 10.1007/s10433-006-0032-2 28794758
Broadbent, D. E., Cooper, P. F., FitzGerald, P., & Parkes, K. R. (1982). The cognitive failures questionnaire (CFQ) and its correlates. The British Journal of Clinical Psychology, 21 (1 ), 1–16. 10.1111/j.2044-8260.1982.tb01421.x 7126941
Cohen-Mansfield, J., Shmotkin, D., Blumstein, Z., Shorek, A., Eyal, N., Hazan, H., & CALAS Team. (2013). The old, old-old, and the oldest old: Continuation or distinct categories? An examination of the relationship between age and changes in health, function, and well-being. International Journal of Aging & Human Development, 77 (1 ), 37–57. 10.2190/AG.77.1.c 23986979
Cosco, T. D., Prina, A. M., Perales, J., Stephan, B. C., & Brayne, C. (2013). Lay perspectives of successful ageing: A systematic review and meta-ethnography. BMJ Open, 3 (6 ), e002710. 10.1136/bmjopen-2013-002710
Cosco, T. D., Prina, A. M., Perales, J., Stephan, B. C., & Brayne, C. (2014). Operational definitions of successful ageing: A systematic review. International Psychogeriatrics, 26 (3 ), 373–381. 10.1017/S1041610213002287 24308764
Costello, A. B., & Osborne, J. B. (2005). Best practices in exploratory factor analysis: Four recommendations for getting the most from your analysis. Practical Assessment, 10 (7 ), 1–9. 10.7275/jyj1-4868
Diener, E., Lucas, R. E., & Oishi, S. (2002). Subjective well-being: The science of happiness and life satisfaction. Handbook of Positive Psychology, 2 , 63–73. 10.1093/oxfordhb/9780195187243.013.0017
Engelhardt, H., Buber, I., Skirbekk, V., & Prskawetz, A. (2010). Social involvement, behavioural risks and cognitive functioning among older people. Ageing & Society, 30 (5 ), 779–809. 10.1017/S0144686X09990626
Fernandez-Ballesteros, R. (2011). Positive ageing: Objective, subjective, and combined outcomes. E-Journal of Applied Psychology, 7 (1 ). 10.7790/ejap.v7i1.238
Gordon, E. H., & Hubbard, R. E. (2022). Frailty: Understanding the difference between age and ageing. Age and Ageing, 51 (8 ), afac185. 10.1093/ageing/afac185 35973066
Green, S. B., & Yang, Y. (2009). Reliability of summed item scores using structural equation modeling: An alternative to coefficient alpha. Psychometrika, 74 (1 ), 155–167. 10.1007/s11336-008-9099-3
Han, J., & Richardson, V. E. (2015). The relationships among perceived discrimination, self-perceptions of ageing, and depressive symptoms: A longitudinal examination of age discrimination. Ageing & Mental Health, 19 (8 ), 747–755. 10.1080/13607863.2014.962007
Heinz, M., Benton, N., & Gleissner, L. (2023). Older adults documenting purpose and meaning through photovoice and narratives. Gerontologist, 63 , 1289–1299. 10.1093/geront/gnad008 36753514
Hooker, K., Mejía, S. T., Phibbs, S., Tan, E. J., & Stevens, J. (2019). Effects of age discrimination on self-perceptions of ageing and cancer risk behaviors. Gerontologist, 59 (Suppl. 1 ), S28–S37. 10.1093/geront/gny183 31100138
Jorgensen, T. D., Pornprasertmanit, S., Schoemann, A. M., & Rosseel, Y. (2019). semTools: Useful tools for structural equation modeling (0.5-2). https://10.0.127.102/CRAN.package.semTools
Kelley, K. (2022). MBESS: The MBESS R package. R package version 4.9.2. https://CRAN.R-project.org/package=MBESS
Kenny, D. A. (2022). Measuring model fit. http://davidakenny.net/cm/fit.htm
Lamers, S. M. A., Westerhof, G. J., Bohlmeijer, E. T., ten Klooster, P. M., & Keyes, C. L. M. (2011). Evaluating the psychometric properties of the Mental Health Continuum-Short Form (MHC-SF). Journal of Clinical Psychology, 67 (1 ), 99–110. 10.1002/jclp.20741 20973032
Lawton, M. P. (1975). The Philadelphia Geriatric Center Morale Scale: A revision. Journal of Gerontology, 30 (1 ), 85–89. 10.1093/geronj/30.1.85 1109399
Levy, B. R., Slade, M. D., & Kasl, S. V. (2002). Longitudinal benefit of positive self-perceptions of ageing on functional health. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 57 (5 ), 409–417. 10.1093/geronb/57.5.p409
Lewis, J. (2014). The role of the social engagement in the definition of successful ageing among Alaska native elders in Bristol Bay, Alaska. Psychology & Developing Societies, 26 (2 ), 263–290. 10.1177/0971333614549143
Lishinski, A. (2021). lavaanPlot: Path diagrams for ‘Lavaan’ models via ‘DiagrammeR’ (0.6.2). https://CRAN.R-project.org/package=lavaanPlot
Lu, S., Chui, C., Lum, T., & Meeks, S. (2022). Promoting social capital for healthy aging: Towards an integrative framework. The Gerontologist, 63 (4 ), 628–636. 10.1093/geront/gnac062
McDonald, R. P. (1999). Test theory: A unified treatment. Taylor & Francis.
Oh, A., Gan, S., Boscardin, W. J., Allison, T. A., Barnes, D. E., Covinsky, K. E., & Smith, A. K. (2021). Engagement in meaningful activities among older adults with disability, dementia, and depression. JAMA Internal Medicine, 181 (4 ), 560–562. 10.1001/jamainternmed.2020.7492 33492334
Otto, R. B., Fields, N. L., Bennett, M., & Anderson, K. A. (2023). Positive aging and death or dying: A scoping review. Gerontologist, 63 (9 ), 1497–1509. 10.1093/geront/gnad006 36744720
R Core Team. (2022). R: A language and environment for statistical computing. R Foundation for Statistical Computing.
Revelle, W. R. (2022). psych: Procedures for psychological, psychometric, and personality research. Northwestern University. R Package Version 2.4.6. https://CRAN.R-project.org/package=psych
Rosseel, Y. (2012). lavaan: An R package for structural equation modeling. Journal of Statistical Software, 48 (1 ), Article 1. 10.18637/jss.v048.i02
Rowe, J. W., & Kahn, R. L. (1997). Successful ageing. Gerontologist, 37 (4 ), 433–440. 10.1093/geront/37.4.433 9279031
Siebert, J. S., Wahl, H.-W., Degen, C., & Schröder, J. (2018). Attitude toward own ageing as a risk factor for cognitive disorder in old age: 12-Year evidence from the ILSE study. Psychology and Ageing, 33 (3 ), 461–472. 10.1037/pag0000252
Sunderland, K., & Beaton, D. (2023). ours: Outliers and robust structures. R package version 0.1.0.9004. https://github.com/derekbeaton/OuRS
Svetina, D., Rutkowski, L., & Rutkowski, D. (2020). Multiple-group invariance with categorical outcomes using updated guidelines: An illustration using Mplus and the lavaan/semTools packages. Structural Equation Modeling: A Multidisciplinary Journal, 27 (1 ), 111–130. 10.1080/10705511.2019.1602776
Tabachnick, B. G., & Fidell, L. S. (2013). Using multivariate statistics (6th ed.). Pearson Education, Inc.
Tully-Wilson, C., Bojack, R., Millear, P. M., Stallman, H. M., Allen, A., & Mason, J. (2021). Self-perceptions of ageing: A systematic review of longitudinal studies. Psychology and Ageing, 36 (7 ), 773–789. 10.1037/pag0000638
Uchino, B. N., & Rook, K. S. (2020). Emotions, relationships, health and illness into old age. Maturitas, 139 , 42–48. 10.1016/j.maturitas.2020.05.014 32747039
Watkins, S., Raisborough, J., & Connor, R. (2023). Aging as adaptation. Gerontologist, 63 (10 ), 1602–1609. 10.1093/geront/gnad049 37098134
World Health Organization. (2015). World report on ageing and health. Author.
Wu, H., & Estabrook, R. (2016). Identification of confirmatory factor analysis models of different levels of invariance for ordered categorical outcomes. Psychometrika, 81 (4 ), 1014–1045. 10.1007/s11336-016-9506-0 27402166
Wu, S., Wang, R., Zhao, Y., Ma, X., Wu, M., Yan, X., & He, J. (2013). The relationship between self-rated health and objective health status: A population-based study. BMC Public Health, 13 , 1–9. 10.1186/1471-2458-13-320 23280303
