Leisure Activities in Very Old Age: Prevalence and Links to Well-Being in (Near-)Centenarians

Abstract

Introduction

This study examined leisure activity engagement among (near-)centenarians and its association with two aspects of well-being: life satisfaction and purpose in life. We analyzed specific activity types and their combinations, as well as three activity parameters: the total number of activities, the variety of different higher-order activity types, and the distribution of activities across these types (i.e., activity diversity).

Methods

We pooled data from two comparable studies conducted in the USA and Germany, yielding a final sample of N = 134 participants aged 96–107 years. Activity engagement was assessed using open-ended questions, with responses coded into 12 activity types further used to compute the activity parameters.

Results

Centenarians' common activities included watching TV, social activities, and experiential activities (e.g., reading). The assessed general activity parameters – number, variety, and diversity of activities – were not related to well-being. Life satisfaction correlated positively with physical and experiential activities and negatively with developmental activities (e.g., visiting a museum). Purpose in life was positively linked to experiential activities. Specific engagement patterns of these activity types, rather than overall activity levels, best predicted well-being outcomes.

Conclusion

Our results contrast prior findings from younger-old samples, suggesting that the relevance of activity types for well-being may shift in very late life. This supports theories emphasizing emotionally meaningful and less demanding activities as more beneficial with increasing age. Findings underline the need for interventions that enable personally relevant forms of engagement to support life satisfaction and purpose in the oldest olds.

Introduction

A substantial body of literature has demonstrated the benefits of an active lifestyle throughout lifespan. Findings show that older adults engaging in a range of activities report better well-being, and mental and physical health [1–3]. These studies suggest that regular activity participation is an important factor for maintaining health and well-being in old age. However, the links between leisure activities and well-being in very late life remain understudied. Given the rising numbers of the very old, including centenarians [4], bridging this gap is crucial for developing strategies to maintain quality of life among the aging population.

The current study investigated leisure activities among near-centenarians (i.e., 95–99 years) and centenarians (i.e., 100+ years), examining their association with two aspects of well-being: life satisfaction and purpose in life. We aimed to identify leisure activities that (near-)centenarians continue to engage in. Additionally, we examined how these activities are linked to life satisfaction and purpose in life using three distinct measures: (1) the total count of specific activities these individuals engage in (e.g., dancing, reading the newspaper), (2) the range of different higher-order activity types they engage in (e.g., physical activities, social activities; = activity variety), and (3) how evenly the specific activities are distributed among these types (= activity diversity). We also sought to identify types and patterns of activity types that are most strongly related to life satisfaction and purpose in life. Identifying these associations can inform tailored interventions and policies that move beyond promoting general activity and instead support specific, evidence-based forms of engagement most closely linked to well-being in the oldest old.

Activities and Aging Well

Although aging may lead to changes in activity engagement, it does not diminish the significance of continued engagement for well-being and health. Thus, activities are at the center of several developmental and aging theories [5, 6]. For instance, the model of Selection, Optimization and Compensation [7] highlights how individuals of advanced age allocate their limited resources by selecting fewer activities, optimizing performance in these areas, and compensating for losses. From this perspective, activity engagement in very old age is expected to become more selective rather than uniformly reduced. Socioemotional Selectivity Theory by Carstensen [8], in turn, emphasizes a motivational shift toward emotionally meaningful goals in late life, suggesting that the subjective relevance and emotional quality of activities may become more important than their number or breadth.

Research suggests that activity engagement is significantly related to older adults’ physical and mental health, and overall well-being [1, 3, 9]. For example, in a cohort of adults in their 70s, leisure activity engagement was positively related to cognitive and physical function and mental health, with standardized effects in the moderate range [3]. Importantly, analyses that differentiate specific activities show that these associations are not uniform: after adjustment for sociodemographic, health, and functional factors, engagement in hobbies/projects and club participation remained linked to lower odds of depression, representing small but meaningful effects, whereas activities such as reading, writing, sewing, or general organizational involvement were not significantly related to depression in adults aged 50 years and older [10]. Similarly, studies examining hedonic and eudaimonic well-being report small positive associations for cognitively and physically engaging activities such as reading, computer use, gardening, and walking, while more passive activities such as television viewing show negligible relationships [11]. It also highlights the connection between meaning or purpose in life and activities in old age, underscoring the significance of this aspect in the context of aging well [2, 12].

Despite its importance, research on activity participation among adults aged 80 years and above is sparse, with few studies linking activities to positive outcomes. Only a handful of studies investigate activities among centenarians, mainly examining physical activities [13, 14] or activities of daily living as indicators of their functional status [15, 16]. Consequently, our understanding of leisure activities of centenarians, with respect to their prevalence and role, remains incomplete. Although there are notable examples of centenarians continuously involved in academics or competitive sports [17], such cases do not represent the experiences of most centenarians. Therefore, investigating the leisure activities of centenarians is crucial to comprehend their significance and relation to well-being in very old age.

Methodological Challenges in Activity Research

Approaches to measuring activity engagement vary considerably across studies. Some examine overall activity levels [1, 3], while others group activities into domains [10, 11, 18, 19]. More recent research uses composite indicators like activity variety and diversity to capture engagement across domains [18–22]. Activity variety refers to the number of different activity domains a person engages in, while activity diversity captures how evenly an individuals’ specific activities are distributed across domains, illustrating the breadth of their participation in diverse activities.

Empirical findings indicate that higher activity variety is associated with better cognitive functioning with small-to-modest effects [19], whereas links with affective outcomes are generally negligible [21]. Activity diversity shows small-to-moderate positive associations with momentary cognitive performance [22] and psychological well-being [20] but remains weakly or inconsistently related to depression and affective states [20].

However, the domains of activity engagement remain debated, with studies identifying up to thirteen activity domains [23]. Factor analyses of the Victoria Longitudinal Study Activity Lifestyle Questionnaire have identified up to 11 activity domains, such as crafts/manual activities, games, religious activities, physical activities/exercise, and social-private/social activities [24, 25]. These allow for nuanced analyses of associations with specific outcomes. Yet, the way individuals combine these activity types may also influence well-being, an aspect often underexplored in research (see [26, 27] for exceptions). Thus, to fully grasp how activities influence positive life outcomes, it is crucial to study not only general activity involvement and composite indicators like activity variety or diversity but also specific activity patterns significantly related to well-being.

The Present Study

Given these conceptual considerations, our study aimed to investigate the multifaceted relationship between leisure activities and well-being (i.e., life satisfaction and purpose in life) in advanced old age. We employed various activity measurements, collectively referred to as activity parameters: a general activity measure (i.e., the number of activities an individual engaged in) and composite indicators (i.e., the variety and diversity of activities across predefined activity types). While related, activity variety and activity diversity capture distinct aspects of engagement. As prior studies typically used only one or the other, we included both to explore whether they offer unique insights regarding their links to well-being. Additionally, we examined the associations of specific activity types as well as their combinations (i.e., profiles) with life satisfaction and purpose in life. Given the limited prior research on activity engagement among centenarians, this study adopted an exploratory approach, addressing the following research questions: (1) what is the association between activity parameters – the number of activities, activity variety, and activity diversity – and well-being among (near-)centenarians? (2) What types of activities are associated with well-being among (near-)centenarians? (3) Which profile of activity engagement best predicts well-being in (near-)centenarians? To address these questions, we employ a combined dataset from two studies conducted in the USA and Germany, with participants aged 95 years and older. Examining overall activity levels, specific activity types, and engagement patterns as distinct but complementary aspects allows for a comprehensive understanding of how activity engagement relates to well-being in very old age. By identifying the forms and patterns of engagement most strongly associated with well-being, the study also provides a basis for developing more targeted recommendations that move beyond encouraging activity in general toward supporting specific, evidence-based forms of engagement in the oldest old.

Methods

Participants and Procedure

This study combined samples of two studies on extreme old age conducted in New York City, USA, and Heidelberg, Germany. The US data came from the Fordham Centenarian Study, involving N₁ = 119 participants aged 95 years and above [28]. Participants were largely recruited through the New York voter’s registry between 2010 and 2013. The German data came from the Second Heidelberg Centenarian Study, involving N₂ = 112 centenarians and/or their proxies [29]. Participants were recruited via resident registries within a 60 km radius of Heidelberg, Germany. The data were collected between 2011 and 2013. Ethical approval was granted by the Institutional Review Boards of Fordham University and Heidelberg University, respectively, and all participants provided informed consent before being interviewed at their residences.

Both studies used Mini-Mental State Exam (MMSE) scores to guide the structured interview procedure and determine whether the activity section was administered. In the Heidelberg sample, activity questions were consistently administered to participants with MMSE short form [30] scores ≥11 (n₁ = 67). In the Fordham sample, interviewers were instructed to attempt the activity section already for participants with MMSE scores between 11 and 14; however, administration within this range was inconsistent, meaning that some participants in this range completed the activity section, whereas others did not. Consequently, only Fordham participants with MMSE scores ≥15 could be assumed to have been administered the activity section in all cases.

We therefore used different MMSE cutoffs for the two samples (Heidelberg: ≥11; Fordham: ≥15), not because different levels of cognitive functioning were targeted but because these cutoffs represented the ranges in which activity questions were consistently administered in each study. This allowed us to apply the same effective inclusion criterion across both samples, including only participants for whom activity data were certainly collected. As a result, zero values could be interpreted as true non-endorsement of activities rather than reflecting cases in which activity questions may not have been administered. Both studies employed the activity questions with participants who had a minimum MMSE short form [30] score of ≥11 in the Heidelberg sample (n₁ = 67), and ≥15 in the Fordham sample (n₂ = 84).

A-priori power analyses suggested that a minimum sample size of 127 would be required to detect medium-sized effects (f² = 0.15) with a power of 0.80 at α = 0.05 for up to 12 different activity types as predictors in a regression analysis. Consequently, we combined the two samples to ensure sufficient statistical power. To enable pooling the data as a single analytic dataset and to limit pronounced differences between the samples, we applied optimal pair matching using the R package matchIt. Participants were matched on age, gender, cognition, and subjective health. This procedure aimed to adjust for differences in participant selection and to minimize potential confounding in the pooled sample. Seventeen participants from the larger sample could not be retained because no suitable matches were available. Descriptive statistics of the unmatched samples can be obtained from the online supplementary Table S2 (for all online suppl. material, see https://doi.org/10.1159/000553222).

After matching, our final sample included N = 134 participants aged 96–107 years (M = 99.95, SD = 1.59) with primarily female (82%) participants, as is common in samples of (near-)centenarians. Descriptive statistics for the final and originating samples are displayed in Table 1.

Table 1.

Descriptive statistics of the collapsed and balanced sample divided by sample membership

​	Fordham (n = 67)	Heidelberg (n = 67)	Total sample (N = 134)
	Age, years	99.81 (2.18)	100.09 (0.54)	99.95 (1.59)
Centenarians, %	61.19	94.03	77.61
Gender, %
Female	80.60	83.58	82.09
Education, %
Basic education	19.40	56.72	38.06
Higher education	65.68	34.33	50.05
Postgraduate	14.93	8.96	11.95
Marital status, %
Married/partnered	10.45	5.97	8.21
Widowed	73.13	83.58	78.36
Unmarried	16.42	10.44	3.43
Living situation, %
At home alone	47.76	40.30	44.03
At home with others	19.40	28.36	23.88
Institution/nursing home	32.84	29.85	31.34
Other	0	1.49	0.75
Cognitive functioning	18.12 (2.08)	17.04 (3.05)	17.58 (2.66)
(Subjective) health	2.73 (1.10)	2.51 (0.75)	2.62 (0.94)
Activity parameters
Number of total activities	6.70 (4.15)	5.46 (3.01)	6.08 (3.66)
Activity variety	4.64 (2.16)	3.70 (1.90)	4.17 (2.08)
Activity diversity	0.64 (0.19)a	0.53 (0.19)	0.58 (0.20)f
Well-being
Life satisfaction	3.05 (1.11)b	3.62 (0.85)d	3.33 (1.03)g
Purpose in life	1.31 (0.57)c	1.27 (0.65)e	1.29 (0.61)h

Values are M (SD) or %. ^an = 61, ^bn = 63, ^cn = 62, ^dn = 63, ^en = 57, ^fn = 122, ^gn = 126, ^hn = 119.

Measures

Activities

In both studies, the activity measure was adapted from the Victoria Longitudinal Study’s activity assessment [24, 31]. Unlike the original format with brief activity descriptions, participants were asked to list up to five specific activities within predefined activity types (e.g., Are there any [social private] activities that you like to do, and if so, which are these?) and to rate their frequency on a 6-point scale (0 = no longer do it, 1 = once a month, 2 = about 2–3 times per month, 3 = about once a week, 4 = 2–3 times per week, 5 = daily). To address passive activities often overlooked during pilot testing, the Heidelberg questionnaire included additional yes/no items on radio and TV use, while the Fordham version added similar questions on TV watching, technology, and internet use. These were followed by the same frequency rating scale.

All answers were converted into a dichotomized variable to indicate engagement in a specific activity (frequency ≥1 = 1) or not (0). This approach addressed the nonlinear scaling and significant skewness of the frequency data, thereby simplifying the data analysis process. A full breakdown of the activity measurement is available in Table 2 and in online supplementary Table S1.

Table 2.

List of activity categories prompted in each sample and their corresponding final activity types with examples of included activity codes

Sample 1: Fordham	Sample 2: Heidelberg	Final activity types (based on Jopp and Hertzog [24])
For each of these activity categories, participants were invited to name up to five activities they engage in and to specify the frequency of these engagements
Social private	Social private	Social private (7; 62.69%); e.g., going out with friends, talking to friends on the phone
Social public	Political	Social public (5; 22.39%); e.g., giving a public talk, attending club meetings, volunteering
	Church	Activities mentioned in these categories were mainly assigned to the religious or social public activity type
	Voluntary
Religious		Religious (2; 22.39%); e.g., attending church service/synagogue, attending in prayer/meditation
Mental	Mental	Developmental (10; 23.13%); e.g., going to the library, visiting concerts/theatres/museums, taking a course at university
Crafts	Crafts	Crafts (4; 13.43%); e.g., doing household repairs, doing woodwork/carpentry or other manual crafts
Relaxing	Relaxing	Experiential (10; 75.37%); e.g., reading for leisure, gardening indoor or outdoor, sewing/knitting/needlework
Games	Games	Games (6; 40.30%); e.g., playing word games, jigsaw puzzling
Physical	Physical	Physical (10; 56.72%); e.g., aerobics, flexibility, walking
Technology usea	Technology use	Technology use (6; 11.94%); e.g., preparing own income tax, playing an instrument, engaging in photography
Internet usea	Internet use	This activity type did not include any entries and was thus excluded from analyses
TV watchinga	TV watchinga	TV watching (4; 80.60%); e.g., watching a documentary, watching the news
	Listening to radioa	This was added as an activity to the experiential activity type
		Travel (3; 1.49%); e.g., traveling out of town
		Household (9; 6.72%); e.g., preparing a meal, driving a car, doing food shopping

Activities of the final activity types that included less than 10% engagement were excluded from further analyses. Numbers in parentheses indicate number of activity codes in each activity type and the percentage of engagement of the final sample in each category.

^aNon-open questions.

Life Satisfaction

Life satisfaction was assessed using the five-item Satisfaction with Life Scale [32]. To accommodate the very old sample, statements were reformulated into direct questions (e.g., Are you satisfied with your life?) and rated on a 5-point scale (0 = not at all to 4 = very much). The internal consistency of the scale was α = 0.74. Mean scores for the analyses were calculated by averaging all available item responses. Higher values indicated greater life satisfaction.

Purpose in Life

Purpose in life was measured using three items from the Valuation of Life Scale [33], capturing a sense of goal-directedness, intentionality, and meaning. Items were reformulated into direct questions (e.g., Does life have meaning for you?). A 5-point response scale (1 = not at all to 5 = very much) was used in the Fordham sample and a 3-point scale (1 = no to 3 = yes) in the Heidelberg sample. To harmonize scores, responses in the Fordham sample were recoded to match the 3-point scale (1 → 1; 2, 3 → 2; 4, 5 → 3). The internal consistency of the scale was α = 0.60. Mean scores were computed by averaging all available responses, with higher values indicating greater purpose in life.

Control Variables

Cognitive functioning and subjective health were considered as relevant background variables for matching the samples. Cognitive functioning was measured by a shortened version of the MMSE, combining the items orientation, recall and recognition, and attention (maximum score of 21 points) [30]. Subjective health was measured by the item “In general, how would you rate your overall health?” rated on a 5-point Likert-scale (1 = poor to 5 = excellent).

Data Analysis

All analyses were conducted using R version 4.3.0 and the packages psych, car, and profileR.

Activity Data Preprocessing

Before data analysis, we coded the individual activities mentioned by participants into 77 activities, primarily derived from the revised Victoria Longitudinal Study activity questionnaire [24], and added activities such as attending a chorus, listening to music/radio, and household tasks. A complete list of activities and their sources is available in the supplementary material. Participants’ information was recoded to 0 (absence) or 1 (presence) of engagement in a specific activity, independent of how often these were mentioned throughout the questions. We then classified these activities into the 11 activity types defined by Jopp and Hertzog [24] and added a household category to accommodate for additional codes. This classification was chosen because it is based on a substantially larger item set and allows a more fine-grained differentiation of activity types than more recent factor solutions based on fewer items (e.g., [25]). Given the limited sample size, we relied on this established classification due to its broad item coverage and clear differentiation of activity types, enabling more nuanced analyses of the links between activity engagement and well-being. Table 2 shows our final 12 activity types and examples for their corresponding codes.

Given differences in breadth across activity types (e.g., watching TV with 4 possible activities vs. developmental activities with 10 possible activities), we dichotomized engagement per activity type (i.e., 0 = absence, 1 = presence) to address the unequal scoring opportunities and enable clearer comparisons across types. Household and travel activities were excluded from the analyses due to low engagement (<10% of the sample). The final set included 10 dichotomized activity types: crafts, TV watching, physical activities, games, technology use, developmental activities (referring to cognitively stimulating and culturally enriching pursuits, such as visiting libraries, museums, concerts, or theatres, or attending educational courses), experiential activities (involving more personal, sensory, or hands-on engagement, including reading for leisure, gardening, or sewing and needlework), social-private activities, social-public activities, and religious activities.

From the preprocessed data, we determined three parameters of activity engagement: the total count of specific activities (i.e., how many specific activities an individual engages in), the count of activity types (= activity variety; i.e., how may activity types an individual engages in), and the evenness and breadth of specific activities across the activity types (= activity diversity; how an individual distributes his/her total count of specific activities across the activity types; see Fig. 1 for an example of low and high activity diversity in the final sample). Activity diversity was calculated using Shannon’s entropy, where m represents the 10 activity types and p_ij is the proportion of total engagements in each specific activity type j to the total number of engagements across all activity types, running from j = 1 to m. Activity diversity scores ranged from 0 (no diversity) to 1 (complete diversity):

$${Activitydiversity}_i=-\left(\frac{1}{\ln \left(m\right)}\right){\displaystyle \sum }_{j=1}^m{p}_{ij}{\ln p}_{ij.}$$

Fig. 1.

Examples of a participant with low and high activity diversity.

Analytic Plan

We first conducted zero-order correlations to examine basic associations between activity engagement and well-being. To identify patterns of activity types related to life satisfaction and purpose in life, we then applied Criterion Profile Analysis (CPA) [34]. CPA is a regression-based method that distinguishes between two components: a level effect, which captures the overall amount of activity engagement, and a pattern effect, which reflects the distribution of engagement across our 10 different activity types in relation to an outcome variable. By comparing how much variance each component explains, both individually and together, CPA determines the relative influence of overall engagement versus specific engagement patterns. Unlike traditional regression approaches that focus on the additive effects of individual activity domains, CPA directly evaluates whether the relative distribution of activities as a single profile is associated with well-being, without decomposing this pattern into separate predictors. Thus, it allows to examine whether specific engagement profiles, beyond the general activity level, are associated with well-being. CPA was particularly appropriate given our aim and sample size, as it enables the modeling of activity profiles in relation to outcomes without requiring large samples.

Results

(Near-)centenarians reported engaging in up to 20 different activities (M = 6.08, SD = 3.66). The most frequently mentioned specific activities were watching comedies or unspecified television (n = 108), reading for leisure (n = 61), and talking to friends on the phone (n = 60). Regarding higher-order activity types, the most common were TV watching (n = 108; 80.60%), experiential activities (n = 101; 75.37%), and social private activities (n = 84; 62.69%). The mean score of self-reported life satisfaction was M = 3.33, SD = 1.03, and of purpose in life was M = 1.29, SD = 0.61.

Correlational Analyses

Zero-order correlations linking the activity parameters (number of activities, activity variety, and activity diversity) with life satisfaction and purpose in life revealed no significant associations. This suggests that overall engagement, variety, or diversity of activities were not related to well-being in the matched sample. However, the activity parameters were highly interrelated (all r ≥ 0.71, ps = 0.000).

The zero-order correlations between the specific activity types, and life satisfaction and purpose in life are summarized in Table 3. Both life satisfaction and purpose in life were positively associated with experiential activities (r = 0.30, p = 0.001 and r = 0.21, p = 0.025, respectively). Life satisfaction was further negatively associated with developmental activities (r = −0.23, p = 0.011) and positively with physical activities (r = 0.19, p = 0.036).

Table 3.

Correlations among the control variables, well-being aspects, and activity types in the total sample of (near-)centenarians

​	​	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16
1	Age	​	​	​	​	​	​	​	​	​	​	​	​	​	​	​	​
2	Gender	0.14	​	​	​	​	​	​	​	​	​	​	​	​	​	​	​
3	Sample membership	0.09	0.04	​	​	​	​	​	​	​	​	​	​	​	​	​	​
4	Cognition	−0.08	−0.03	−0.20*	​	​	​	​	​	​	​	​	​	​	​	​	​
5	Subjective health	0.34***	0.10	−0.12	0.04	​	​	​	​	​	​	​	​	​	​	​	​
6	Life satisfactiona	0.29**	0.06	0.28**	0.00	0.31***	​	​	​	​	​	​	​	​	​	​	​
7	Purpose in lifeb	0.06	0.15	−0.04	−0.03	0.16	0.45***	​	​	​	​	​	​	​	​	​	​
8	Crafts	−0.01	−0.16	0.00	−0.07	−0.07	−0.06	−0.13	​	​	​	​	​	​	​	​	​
9	TV watching	0.13	0.02	0.11	0.04	0.08	0.03	−0.12	0.08	​	​	​	​	​	​	​	​
10	Physical	0.01	−0.09	0.03	0.18*	0.13	0.19*	−0.03	0.03	0.30**	​	​	​	​	​	​	​
11	Games	0.08	−0.01	0.03	0.12	−0.06	−0.07	0.07	0.17	0.29**	0.20*	​	​	​	​	​	​
12	Technology use	0.07	0.11	−0.37***	0.11	0.20*	−0.15	0.01	−0.15	−0.05	−0.05	−0.07	​	​	​	​	​
13	Developmental	−0.12	0.03	−0.16	0.24**	−0.06	−0.23*	−0.08	0.10	0.05	0.19*	0.16	0.07	​	​	​	​
14	Experiential	0.15	0.05	0.09	0.07	0.12	0.30**	0.21*	0.07	0.42**	0.30***	0.29**	0.05	0.15	​	​	​
15	Social private	0.12	0.00	−0.28**	0.23**	0.18*	−0.10	−0.04	0.08	0.25**	0.23**	0.29**	0.14	0.24**	0.28**	​	​
16	Social public	−0.03	−0.03	−0.07	0.32***	0.14	0.09	0.12	0.10	0.13	0.11	0.07	0.19*	0.26**	0.18*	0.19*	​
17	Religious	−0.06	0.06	−0.54***	0.12	0.08	−0.05	0.18	−0.11	0.08	0.00	0.03	0.24**	0.17*	0.10	0.23**	0.10

N = 134, gender (1 = male, 2 = female), sample membership (1 = Fordham, 2 = Heidelberg).

^a n = 126.

^b n = 119.

*p < 0.05; **p < 0.01; ***p < 0.001.

Criterion Profile Analysis

To inform about possible patterns of activity types associated with life satisfaction and purpose in life, we conducted two separate CPAs, entering the 10 specific activity types (dichotomized) as predictors of either life satisfaction or purpose in life as an outcome. Predicting life satisfaction, the analysis revealed that the activity types accounted for 27.45% of the variance in this outcome. The pattern component alone explained 26.35% of the variance (F[9, 115] = 4.571, p = 0.000), indicating that the specific combination of activities people engage in was linked to their life satisfaction. By contrast, the level component did not significantly contribute to life satisfaction (F[1, 115] = 1.560, p = 0.999), suggesting that the overall activity engagement was not systematically related to life satisfaction. Significant regression weights were found for the activity types of developmental activities (β = −0.69, p = 0.001), experiential activities (β = 0.95, p = 0.000), and physical activities (β = 0.40, p = 0.027). The pattern associated with high life satisfaction may thus be mainly attributed to the engagement in experiential and physical activities while being less engaged in developmental activities.

Running the CPA to predict purpose in life, the results suggested that the activity types explained 16.99% of the variance. The pattern component explained 16.86% of the variance (F[9, 108] = 2.434, p = 0.015), while the level component did not contribute significantly to the model (F[1, 108] = 0.265, p = 0.608). This indicates that the pattern of activities was related to purpose in life, whereas the overall activity engagement was not systematically associated with purpose in life. Significant regression weights occurred for the activity types of TV watching (β = −0.36, p = 0.026), experiential activities (β = 0.35, p = 0.014), and religious activities (β = 0.28, p = 0.034). Thus, more engagement in experiential and religious activities while watching less TV seemed to best predict higher purpose in life in centenarians. See Figure 2 for a depiction of the profiles most strongly associated to life satisfaction and purpose in life. Detailed results of the regression analyses are provided in online supplementary Tables S3 and S4. Sensitivity analyses including sample membership, education, and living situation as covariates indicated that the pattern of results remained largely unchanged (see online suppl. material).

Fig. 2.

Activity type profiles associated with life satisfaction (n = 129) and purpose in life (n = 119). Only regression weights that reached statistical significance are labeled with their respective standardized beta coefficients. *p < 0.05. **p < 0.01. ***p < 0.001.

Discussion

This study examined the relationship between leisure activities and well-being in advanced old age. We found that the activity parameters – the number, variety, and diversity of activities – were not associated with life satisfaction or purpose in life. In contrast, engagement in experiential activities was positively related to both outcomes. Additionally, developmental activities showed a negative association, and physical activities a positive association with life satisfaction. The CPA revealed that specific engagement patterns, such as more experiential and physical, and less developmental activity, or more experiential and religious and less TV watching, seemed more predictive of well-being than overall activity levels. The most frequently reported activities in the present sample – watching television, reading for leisure, and talking to friends on the phone – are also commonly reported leisure activities in older adult samples. Television viewing has consistently been identified as one of the most prevalent forms of leisure engagement in later life [11, 35] and recent evidence likewise identifies reading as a frequently engaged leisure activity among older adults [11].

Specific Activities Are Related to Life Satisfaction and Purpose in Life

Our finding that activity number, variety, and diversity were unrelated to life satisfaction or purpose in life contrasts with prior research. For instance, Lee et al. [20] found that in older adults, activity diversity was associated with both eudaimonic well-being (e.g., personal growth, purpose in life) and hedonic well-being, as reflected by increases in positive affect and reductions in negative affect over time. Positive affect is commonly considered a core component of hedonic well-being, alongside life satisfaction. In contrast, our results did not show associations between general activity parameters and either life satisfaction or purpose in life. One possible explanation is that in very late life, maintaining well-being may depend less on the breadth of activities and more on selective engagement in a few personally rewarding domains. Another consideration is the high intercorrelation among the three activity parameters in our very old participants, which may limit their distinctiveness and complicate interpretations of their specific relationships with well-being outcomes.

Among activity types, experiential and physical activities were positively correlated with life satisfaction while developmental activities were negatively associated. In addition, profile analysis showed that specific combinations of activities, rather than the overall level of activity, explained individual differences in life satisfaction. Specifically, concurrent engagement in physical and experiential activities while not engaging in developmental activities seemed particularly beneficial for life satisfaction.

These results contrast with findings in a younger-old sample by Ryu and Heo [36]. They reported positive associations between cultural activities (akin to developmental activities) and life satisfaction in adults between 60 and 90 years, while hobbies and indoor activities (akin to experiential activities) were not related to life satisfaction. These findings suggest a varied impact of activity types on life satisfaction in older compared to very old adults, highlighting the complexity of activity engagement effects across the lifespan.

Three mechanisms may explain the strong association between experiential activities and life satisfaction in very old age. First, cognitive decline may make simpler, experiential activities more manageable and rewarding; at the same time, developmental activities may become more bothersome and may even put a spotlight on cognitive restrictions. Second, according to socioemotional selectivity theory [8], shifting the focus from long-term goals and personal development early in adulthood to prioritize emotional well-being potentially may lead to a preference for experiential activities in advanced old age. Third, following the model of selection, optimization, and compensation [7], individuals in advanced old age may strategically invest their limited resources into less demanding activities. In this context, engaging in experiential activities may represent a feasible form of adaptation to maintain life satisfaction.

Contrary to prior research [23, 36], social activities were not linked to life satisfaction in our sample. This may reflect age-related barriers such as reduced mobility and smaller social networks, which can limit social engagement and alter its role in well-being among the oldest old [37, 38].

The reasoning outlined above also helps explain why experiential activities are significantly linked to purpose in life. Echoing the earlier discussion, their relationship can be understood through Carstensen’s socioemotional selectivity theory [8], suggesting that (near-)centenarians focusing on short-term experiential activities may experience increased goal-directedness and meaningfulness. Consistent with this interpretation, recent evidence shows that experiential activities such as reading or gardening are positively related to eudaimonic well-being in older adults, whereas passive activities like television viewing show weak or even negative associations [11]. In line with this pattern, our regression analysis further revealed that religious activities are positively linked with purpose in life, while TV watching had a negative link. This outcome likely reflects the established connection between religiosity and a sense of purpose [39], as opposed to the passive engagement associated with TV watching, which is often linked to lower purpose in life [12].

Overall, activity engagement showed stronger associations with life satisfaction than with purpose in life, consistent with recent findings distinguishing hedonic and eudaimonic well-being among older adults with cancer experience [11]. Life satisfaction reflects a broad evaluative judgment and may therefore be more immediately influenced by everyday experiences, whereas activities may foster purpose in life primarily when they are experienced as meaningful and goal-congruent – an aspect not assessed in the present study.

Strengths and Limitations

Research on centenarians fills a critical gap in aging studies, which often overlook the oldest old. Our study contributed to this field by examining activity engagement in this group using a relatively large sample. However, several limitations must be acknowledged.

First, merging the samples improved statistical power for correlation and regression analyses but may have masked cultural differences. Activity engagement is shaped by contextual factors that differ between Germany and the USA, such as living conditions [40, 41]. To account for such contextual factors, we included educational background and living situation in supplementary analyses. However, participants’ financial resources (e.g., income) could not be incorporated due to nonequivalent measurement across the two studies, which precluded meaningful harmonization. Differences in financial resources may therefore still have influenced activity engagement and well-being. Also, the study was primarily powered to detect medium-sized associations, and smaller effects may not have been detectable.

Second, our sample included only centenarians with relatively intact cognitive functioning, likely leading to an overestimation of activity levels possibly biasing associations between activity engagement and well-being. Third, we used straightforward methods to align with the data, which restricted our ability to depict varied activity profiles among centenarians using more advanced approaches like cluster analysis or directly control for confounding variables. This reduced the nuance with which distinct engagement patterns and their associations with well-being could be examined. Moreover, the cross-sectional design only allowed for the exploration of associations, leaving the direction of effects unclear. It thus remains uncertain whether activities enhance well-being in late life, whether higher well-being leads to more activity, or whether these processes are mutually reinforcing.

Fourth, the well-being measures showed acceptable, yet somewhat limited internal consistency, especially for purpose in life. Mean scores were based on all available responses, with some participants contributing only a subset of items. These factors may have introduced measurement error, potentially attenuating associations and limiting the precision of our conclusions.

Conclusion and Outlook

The results indicate that in very old age, the overall levels of activity engagement are less closely related to well-being than specific patterns of engagement. Profiles characterized by a combination of experiential, physical, and religious activities, alongside lower engagement in developmental or passive activities, seemed most strongly associated with life satisfaction and purpose in life in this cognitively relatively healthy sample of (near-)centenarians. Despite these insights, more research is needed to understand how activities influence well-being in very old age. Centenarians themselves have emphasized the role of meaningful activities in their longevity [42], and prior research suggests that meaning mediates the link between leisure and well-being (e.g., [43]). Identifying which activities centenarians perceive as meaningful and how these shape well-being remains a valuable direction for future research. To further clarify this study’s interpretations, future research should also compare different age groups among older adults or apply longitudinal designs to examine age differences in the relevance of activity types for well-being. As activities are also associated with affect (e.g., [20, 22]), future research should additionally address the emotional components of well-being in advanced old age.

Building on these findings, promoting specific, evidence-based forms of activity engagement may be more beneficial for supporting well-being in the oldest old than encouraging general activity. Fostering participation in emotionally enriching and personally relevant activity types, such as experiential, physical, and religious activities, could help maintain life satisfaction and a sense of purpose in very late life. Tailored interventions and community programs that enable access to such activities may contribute meaningfully to quality of life in this rapidly growing population, reinforcing public health goals for aging well.

Acknowledgments

We are grateful for Prof. Dr. Peter Hilpert and Dr. Alexander Stahlmann for their statistical advice during the data analysis for this paper. We thank all study participants and their families for their support.

Statement of Ethics

This human study was approved by the Ethics Committee of the Faculty of Behavioural and Cultural Studies at Heidelberg University, Approval No. 04.10.2011, and the Institutional Review Board at Fordham University, Approval No. 07.07.2010. All adult participants provided written informed consent to participate in this study.

Conflict of Interest Statement

The authors have no conflicts of interest to declare.

Funding Sources

This work was supported by the Swiss National Science Foundation (186239 to D.S.J. and M.S.); the Velux Stiftung (1637 to C.R.); the University Research Priority Program “Dynamics of Healthy Aging” at the University of Zurich; the Brookdale Leadership in Aging Fellowship (to D.S.J.); the Robert Bosch Foundation (to D.S.J. and K.B.); and the Dietmar Hopp Foundation (to D.S.J. and K.B.). The funders had no role in the design, data collection, data analysis, and reporting of this study.

Author Contributions

Melanie Stahlmann initiated and led the publication project, conceptualized the specific research questions and analyses for the present paper, conducted the statistical analyses, and was responsible for writing the original draft as well as revising and editing the manuscript. Prof. Dr. Daniela S. Jopp led the data collection and project administration of the original studies, contributed to the development of the present analyses, acquired funding, provided supervision throughout the project, and contributed to writing both the original draft and revisions of the paper. Dr. Christina Röcke contributed to the methodological approach and analyses and was involved in writing both the original draft and revisions of the manuscript. Dr. Charikleia Lampraki was responsible for coding the activity data and contributed to the review and editing of the manuscript. Prof. Dr. Kathrin Boerner collaborated with Prof. Dr. Daniela S. Jopp in the data collection and project administration and contributed to the conceptualization of the present paper, as well as to reviewing and editing the manuscript.

Funding Statement

This work was supported by the Swiss National Science Foundation (186239 to D.S.J. and M.S.); the Velux Stiftung (1637 to C.R.); the University Research Priority Program “Dynamics of Healthy Aging” at the University of Zurich; the Brookdale Leadership in Aging Fellowship (to D.S.J.); the Robert Bosch Foundation (to D.S.J. and K.B.); and the Dietmar Hopp Foundation (to D.S.J. and K.B.). The funders had no role in the design, data collection, data analysis, and reporting of this study.

Data Availability Statement

The data that support the findings of this study are not publicly available due to these data being collected with the help of private funding agencies, and no data sharing policies were put into place. The data are available from Prof. Dr. Daniela S. Jopp (daniela.jopp@unil.ch) upon reasonable request.