Cognitive decline may be an inevitable companion of aging, but a sweeping new study from China suggests that the protective power of leisure activities does not fade in the twilight years—and that its benefits are deeply shaped by socioeconomic advantages accumulated across an entire lifetime. The research, published in the journal Ageing International, analyzed sixteen years of longitudinal data from older Chinese adults and arrived at a conclusion that challenges one of the most influential ideas in social gerontology: the age-as-leveler hypothesis, which predicts that health inequalities between social groups converge in old age. Instead, the study finds that the cognitive advantages of an active leisure life persist, and in some respects deepen, well into the mid-eighties and beyond, and that who benefits most from these activities depends on a lifetime of socioeconomic circumstance.
The investigation was carried out by Rongjun Sun of Cleveland State University and Zhenmei Zhang of Michigan State University, who mined the Chinese Longitudinal Healthy Longevity Survey, one of the largest and longest-running studies of its kind in the world. The survey randomly selected half of the counties and cities in 22 of China’s 31 provinces, and the researchers drew on six waves of interviews conducted between 2002 and 2018. A baseline cohort of 15,798 adults aged between 65 and 105 was first interviewed in 2002 and then re-interviewed in 2005, 2008, 2011, 2014, and 2018. After accounting for missing data through a complete-case approach, the final analysis rested on 14,854 person-period observations drawn from participants who contributed at least two consecutive waves of data.
At the heart of the study is a deceptively simple question: does the link between staying mentally, socially, and physically active and maintaining cognitive function grow weaker or stronger as people age past 65? The scientific literature offers conflicting answers. Cognitive reserve theory suggests that older adults, whose brains are increasingly burdened by pathological damage, should draw greater compensatory benefit from stimulating activities, and some empirical work supports this—studies have found, for example, that physical activity is particularly beneficial for global cognition and episodic memory among the very old, and that social activity slows cognitive decline more markedly in the oldest-old. Yet other research, including a large analysis of roughly 65,000 participants aged 18 to 90, found either no age differences in the association between cognitively stimulating leisure activities and cognitive performance or benefits that were significant only for younger adults. Sun and Zhang set out to resolve this uncertainty with data robust enough to track associations across the full range of old age, from the young-old to those beyond 85.
The researchers measured cognitive function using the Chinese version of the Mini-Mental State Examination, adapted to fit Chinese cultural settings. Of the 30 original items, 25 were retained in the final impairment index, covering five domains of cognition: orientation, registration, attention and calculation, recall, and language. Higher scores on this reversed index indicate more severe impairment. Seven leisure activities formed the core predictors: physical exercise, outdoor activities, gardening, social activities, playing cards or mahjong, reading newspapers and books, and watching television or listening to the radio. Each was rated on a frequency scale ranging from never to almost every day, with physical exercise captured as regular participation. To capture socioeconomic status across the life course, the study combined childhood indicators—parents’ literacy and whether the respondent was born in an urban or rural area—with adult measures including the respondent’s own literacy, pre-retirement occupational position, and current place of residence. Health controls included a six-item physical performance test, an index of disability in activities of daily living such as bathing, dressing, and eating, and self-rated health. Demographic controls covered age, gender, marital status, living arrangements, and time since baseline.
Methodologically, the study employed a lagged independent variable design, using predictors from one survey wave to forecast cognitive status in the next, thereby establishing a temporal ordering between activity and outcome. Because the cognitive impairment scores were right-skewed and over-dispersed, the authors adopted a Generalized Linear Mixed Model with a negative binomial link function—a statistical framework suited to count-like outcomes with excess variability. Uniquely, the leisure activity variables were treated as random effects, allowing their associations to vary across individuals rather than forcing a single fixed estimate. To guard against bias from sample attrition, the models included dummy indicators for participants lost to follow-up through death or nonresponse, a technique borrowed from prior longitudinal research on functional status in older populations.
The results on the first question were strikingly consistent. Virtually all leisure activities remained significantly associated with lower cognitive impairment across every age group, even into the oldest ages. Expressed as ratios of means from the negative binomial models, values below 1.00 signaled less impairment: for those aged 85 and above, regular exercisers showed cognitive impairment scores 14 percent lower than non-exercisers, holding all covariates constant. Reading, playing cards or mahjong, gardening, and social participation told similar stories. Crucially, when the full suite of life course socioeconomic variables was added to the models, these associations barely budged. Of the 30 coefficients that were statistically significant without socioeconomic controls, only six lost significance, indicating that socioeconomic status does not explain away the leisure-cognition connection—these are not merely privileged people who both stay active and stay sharp.
The second, and arguably more consequential, finding concerns interaction. When the researchers modeled how leisure activity and life course socioeconomic status jointly shape cognition, they uncovered 26 statistically significant interaction effects across the age groups, and nearly all pointed in the same direction: among people who engaged in a given activity, those of higher socioeconomic status experienced less cognitive impairment than their lower-status peers. Notably, these interactions were more numerous at older ages—15 of the 26 occurred after age 80, with the 85-and-over group showing the highest proportion of significant interactions of any age band. Seven of the significant interactions involved childhood socioeconomic variables, such as parents’ education and place of birth, underscoring how early-life circumstances echo into the cognitive landscape of very late life.
Perhaps the most provocative pattern was that most of these significant interactions stood alone, without a corresponding significant main effect of the activity itself. Outdoor activity illustrates the phenomenon vividly: its main effect on cognitive impairment was never significant at any age, yet six of its interactions with socioeconomic variables were, including two in the oldest age group. In practical terms, this means that for older adults of lower life course socioeconomic status, outdoor activity showed virtually no association with cognition, whereas for their higher-status counterparts the same activity was linked to meaningfully lower impairment. Under the theory of “education as learned effectiveness,” this makes intuitive sense: education endows individuals not merely with credentials but with durable skills—goal setting, sustained attention, problem-solving styles—that change how people engage with activities and therefore how much cognitive benefit they extract from them. Two people may garden or play mahjong with equal frequency, yet reap very different rewards depending on the cognitive toolkits their life trajectories have built.
The authors interpret these findings as a direct challenge to the age-as-leveler hypothesis, which holds that socioeconomic disparities in health widen from young adulthood through midlife but then narrow after age 60, as biological processes become the dominant driver of health outcomes, as selective survival produces a more homogenous elderly population, and as social policies such as pensions weaken the ties between status and wellbeing. Cumulative disadvantage theory predicts the opposite—that inequalities compound with age—and the present study aligns with that camp. Sun and Zhang’s contribution goes a step further than most prior tests, because they examined not just raw socioeconomic gradients in health but the conditional nature of a purportedly protective behavior. The persistence of interactions into the oldest ages suggests that individual differences in cognition become more complex, not simpler, as people age, consistent with the biocultural model of lifespan development advanced by Paul Baltes and colleagues, which holds that old age depends increasingly on deploying culture-based resources—material, social, economic, and psychological—to compensate for declining biology.
The study is not without limitations, as the authors are careful to acknowledge. The Mini-Mental State Examination, even in its culturally adapted Chinese form, remains sensitive to education, and a general impairment index obscures variation across distinct cognitive domains. Life course socioeconomic status was measured roughly, omitting parental occupation, childhood household environment, and the complexity of pre-retirement work. The leisure measures captured neither the intensity, duration, nor quality of participation, and excluded activities such as traveling, volunteering, and entertainment. And while the lagged design strengthens temporal inference, the authors caution against causal claims, noting that bidirectional dynamics—cognitive decline prompting withdrawal from activities, and vice versa—cannot be fully disentangled from observational data. Even so, the scale of the dataset, the sixteen-year span, and the methodological care make this one of the most compelling demonstrations to date that the cognitive benefits of an engaged life are neither uniform nor automatically accessible.
For a world facing rapidly growing populations of the very old, the implications are sobering and actionable in equal measure. Promoting leisure activity among older adults remains a sound public health strategy—the main effects were real and durable—but the study suggests that activity-based interventions alone may not reach everyone equally. Older adults carrying lifetimes of socioeconomic disadvantage may extract less cognitive protection from the same pursuits, meaning that effective policy must address the underlying resource gaps, not just the behavior. As Sun and Zhang conclude, the interplay between who we were, what we do, and how our minds hold up in old age is a story of compounding complexity—and the first longitudinal portrait of that complexity reveals inequalities that refuse to fade with age.
Subject of Research: The association between leisure activity engagement and cognitive impairment among Chinese older adults aged 65 and older, and how this association varies by life course socioeconomic status and age.
Subject of Research: Medicine
Article Title: Does the Association Between Leisure and Cognition Vary by Life Course Socioeconomic Status and Age?
Article References: Sun, R., & Zhang, Z. (2026). Does the Association Between Leisure and Cognition Vary by Life Course Socioeconomic Status and Age?. Ageing International, 51(2), Article 15. https://doi.org/10.1007/s12126-026-09661-z
Image Credits: AI Generated
DOI: 10.1007/s12126-026-09661-z
Keywords: cognitive impairment, leisure activity, life course socioeconomic status, cognitive aging, Chinese Longitudinal Healthy Longevity Survey, age-as-leveler hypothesis, cognitive reserve, oldest-old, health inequality, longitudinal study, Mini-Mental State Examination, negative binomial model
Cite Scienmag News
APA MLA Chicago
Glenn Wilkins. (September 5, 2026). Leisure-Cognition Link May Depend on Socioeconomic Status and Age. Scienmag. https://scienmag.com/leisure-cognition-link-may-depend-on-socioeconomic-status-and-age/
Copy citation Download RIS
Tags: active leisure and cognitive preservationage-as-leveler hypothesis challengeage-as-leveler hypothesis in social gerontologyaging and cognitive resilienceaging and social disparitiesChinese longitudinal health studiesChinese Longitudinal Healthy Longevity Survey analysiscognitive benefits of active lifestyles across the lifespanhealth inequalities in old ageimpact of socioeconomic advantages on agingimpact of socioeconomic factors on cognitive resilienceinfluence of socioeconomic background on aging outcomesinfluence of socioeconomic background on leisure benefitsLeisure activities and cognitive health in aginglifelong socioeconomic factors and aginglong-term effects of leisure on aginglong-term effects of leisure on cognitionlongitudinal aging studies in Chinasocial determinants of cognitive agingsocial determinants of cognitive declinesocioeconomic status and cognitive declinesocioeconomic status and lifelong advantages
