Purpose in life is the feeling that one’s life is goal-oriented and driven (Ryff, Reference Ryff1989) and is considered a cornerstone of eudaimonic well-being (Ryan and Deci, Reference Ryan and Deci2001; Ryff, Reference Ryff2014). The beneficial effects of purpose are now being examined in the context of health. Higher purpose, for example, is associated with better physical function (Kim et al., Reference Kim, Kawachi, Chen and Kubzansky2017), lower risk of cardiovascular events (Kim et al., Reference Kim, Sun, Park, Kubzansky and Peterson2013), and ultimately lower risk of premature mortality (Cohen et al., Reference Cohen, Bavishi and Rozanski2016). These more positive health outcomes may be due, in part, to the healthier behavioral patterns associated with purpose. Individuals who report greater purpose in life, for example, engage in more physical activity (Hooker and Masters, Reference Hooker and Masters2016) and are less likely to smoke (Morimoto et al., Reference Morimoto, Yamasaki, Ando, Koike, Fujikawa, Kanata and Nishida2018).
There is growing evidence that feelings of purpose are also associated with better cognitive outcomes. In particular, purpose has been associated with lower risk of Alzheimer’s disease (Boyle et al., Reference Boyle, Buchman, Barnes and Bennett2010) and dementia (Sutin et al., 2018b) and with greater resilience to Alzheimer’s neuropathology (Boyle et al., Reference Boyle, Buchman, Wilson, Yu, Schneider and Bennett2012). Most research on purpose and cognition has focused on risk of passing a threshold for cognitive impairment. There is some evidence, however, that it is also associated with cognitive function prior to impairment. Higher purpose, for example, is associated with better memory and executive function (Lewis et al., Reference Lewis, Turiano, Payne and Hill2017) and better overall mental status (Kim et al., Reference Kim, Shin, Scicolone and Parmelee2019). It may also be associated with performance on specific cognitive tasks, such as verbal fluency and episodic memory, that are intermediate markers of cognitive health.
Verbal fluency is the ability to generate many words from a single category in a short period of time (Lezak, Reference Lezak2004). It is typically measured as the number of words retrieved correctly from a specific category (e.g. animals) in 60 seconds and is frequently used in research and clinical setting when evaluating cognitive status. Performance on verbal fluency tasks involves the integration of several cognitive functions, including speed, attention, inhibition, and self-monitoring. It is thus a relatively simple task that captures the effective integration of numerous functions (Shao et al., Reference Shao, Janse, Visser and Meyer2014; Troyer et al., Reference Troyer, Moscovitch and Winocur1997). Perhaps because of this integration, performance on verbal fluency tasks has long-term predictive power for consequential cognitive outcomes. In fact, better verbal fluency is associated with lower risk of dementia over time (Sutin et al., 2019b). Episodic memory is likewise a critical cognitive function that is the ability to remember specific events and experiences and is typically measured with immediate and delayed recall of lists of words (Lezak, Reference Lezak2004). Loss of episodic memory is one defining features of ADRD (Alzheimer’s Association, 2020). Similar to fluency, performance on tasks that measure memory earlier in adulthood is associated with risk of Alzheimer’s disease in older adulthood (Josefsson et al., Reference Josefsson, Sundström, Pudas, Nordin Adolfsson, Nyberg and Adolfsson2019). Both verbal fluency and episodic memory are thus clinically relevant intermediate markers of cognitive health on the pathway to dementia.
Purpose in life is typically conceptualized as one component of the broader concept of meaning in life (Steger, Reference Steger and Lopez2009). Specifically, meaning in life is a superordinate construct that includes purpose (feeling that one’s life is goal-directed), coherence (belief that the world is orderly), and significance (sense that one’s life has value) (Martela and Steger, Reference Martela and Steger2016). As such, meaning and purpose, although related, are conceptually distinct constructs (George and Park, Reference George and Park2013). Lay conceptions of the terms, however, may not be distinct, and measures for purpose and meaning may capture similar variance for meaningful outcomes. For example, the association in predicting risk of cognitive impairment is similar across measures of purpose in life (Boyle et al., Reference Boyle, Buchman, Barnes and Bennett2010) and meaning in life (Sutin et al., 2020). This pattern suggests that purpose and meaning may have similar associations with intermediate cognitive outcomes.
The present study takes a comprehensive approach to test the association between purpose in life/meaning in life and verbal fluency and episodic memory. Specifically, we examine the association between purpose/meaning and verbal fluency in a combined sample of over 125,000 participants from 24 countries and the association between purpose/meaning and episodic memory in a combined sample of over 140,000 participants from 32 countries. We test the preregistered hypothesis that higher reported feelings of purpose/meaning in life are associated with better performance on a verbal fluency task. Although not part of the original preregistration, we also test whether purpose/meaning in life is associated with better performance on an episodic memory task. With up to 32 countries, we examine whether the economic environment of the country (e.g. gross domestic product per capita; GDP) is associated with the strength of the relation between purpose/meaning and cognition because purpose may be a resource for cognitive function when economic resources are less plentiful. We also examine whether the associations differ across measures that assess purpose in life versus meaning in life. Finally, we address whether the associations differ by age, gender, and education to determine if the benefits of purpose/meaning are limited to specific demographic groups or whether the association generalizes across populations.
Participants and procedure
A total of up to 37 samples from 32 countries from 8 established cohorts were included in this research. The cohorts were the Health and Retirement Study (HRS; Sonnega et al., Reference Sonnega, Faul, Ofstedal, Langa, Phillips and Weir2014), the Midlife Development in the United States (MIDUS) study (Brim et al., Reference Brim, Ryff and Kessler2004), the Wisconsin Longitudinal Study (WLS; Herd et al., Reference Herd, Carr and Roan2014), the English Longitudinal Study of Ageing (ELSA; Steptoe et al., Reference Steptoe, Breeze, Banks and Nazroo2013), the National Child Development Study (NCDS; Power and Elliott, Reference Power and Elliott2006), The Irish LongituDinal study on Ageing (TILDA; Kearney et al., Reference Kearney2011), the Brazilian Longitudinal Study of Aging (ELSI; Lima-Costa et al., Reference Lima-Costa2018), and the Survey of Health, Ageing and Retirement in Europe (SHARE; Börsch-Supan et al., Reference Börsch-Supan, Brandt, Hunkler, Kneip, Korbmacher and Malter2013). Each cohort had one sample, except for WLS, which included graduate (WLSG) and sibling (WLSS) samples, and the SHARE, which included samples from up to 28 countries in Europe and Israel (29 countries total). We identified these datasets through the Interuniversity Consortium for Political and Social Research (ICPSR), the Gateway to Global Aging, and the UK Data Service. All data can be obtained through the websites listed below for each of the studies. The preregistration for this study can be found at https://osf.io/rywhu.
The HRS (https://hrs.isr.umich.edu/) is a longitudinal study of aging of Americans aged 50 years and older and their spouses (regardless of age). Cognition was measured on the entire sample in 2010 and 2012, and purpose was measured for half the sample in 2010 and the other half in 2012. Data on cognition and purpose were combined across the 2010 and 2012 assessments. The MIDUS study (http://midus.wisc.edu/) started as a study of midlife health and has continued to assess participants as they age. Purpose and cognition were first measured at MIDUS II in 2004–2006. A previous publication (Lewis et al., Reference Lewis, Turiano, Payne and Hill2017) used MIDUS data to report the association between purpose in life and both memory and executive function that was a composite of serval tasks that included verbal fluency and also the stop and go switch task, number series, digit span backward, and backward counting. The relation between purpose and verbal fluency alone in MIDUS has not yet been reported. The WLS (https://www.ssc.wisc.edu/wlsresearch/) includes two samples: a random sample of individuals who graduated from a Wisconsin high school in 1957 (WLSG) and a selected sibling of the graduates (WLSS). Purpose and cognition were measured in both samples at the 2011 assessment. The ELSA (https://www.elsa-project.ac./uk) is a study of aging of individuals over the age of 50 years in England and their spouses (regardless of age). Meaning and cognition were measured at the baseline assessment in 2002. The NCDS (https://cls.ucl.ac.uk/cls-studies/1958-national-child-development-study/) is a study of individuals all born in the same week in 1958 in England, Scotland, and Wales. Meaning and cognition were measured at the 2008 assessment when study participants were 50 years old. The TILDA (https://tilda.tcd.ie/) is a longitudinal study of aging of individuals over the age of 50 years in Ireland and their spouses. Meaning and cognition were measured at the baseline assessment in 2011 and used in the meta-analysis. In addition, at Wave 4, both meaning and purpose, as well as cognition, were assessed (used in supplemental analysis). ELSI (http://elsi.cpqrr.fiocruz.br/en/) is a longitudinal study of aging of individuals over the age of 50 years and their spouses in Brazil. Meaning and cognition were measured at the baseline assessment in 2015–2016. SHARE (http://www.share-project.org/) is a cross-national study of health and aging in 28 European countries and Israel (29 countries total). We included 21 countries from SHARE in the analysis of fluency (the other 8 countries did not have data on fluency) and all 29 countries for the analysis of memory. We selected Wave 6 to maximize participants with the necessary data to be included in the analysis. Countries that had data on meaning and fluency at Wave 6 were Austria, Germany, Sweden, Spain, Italy, France, Denmark, Greece, Switzerland, Belgium, Israel, Czech Republic, Poland, Luxembourg, Portugal, Slovenia, Estonia, and Croatia. Three countries had meaning and fluency available at other waves and were included in the analysis: Ireland (Wave 2), Hungary (Wave 4), and the Netherlands (Wave 5). The remaining SHARE countries – Lithuania, Bulgaria, Cyprus, Finland, Latvia, Malta, Romania, and Slovakia – were first assessed at Wave 7 and only had the measure of memory. These countries were included in the analysis of meaning and episodic memory. Across all samples, there was no exclusion of participants based on cognitive status.
Purpose/Meaning in life. Purpose in Life was measured with a version of Ryff’s Purpose in Life scale (Ryff, Reference Ryff1989). A 7-item version was administered in HRS and MIDUS and a 6-item version was administered in WLS. Items were rated on a scale from 1 (strongly disagree) to 6 (strongly agree) in HRS and WLS and from 1 (strongly disagree) to 7 (strongly agree) in MIDUS. Meaning in Life was measured with a single-item (“How often do you feel that your life has meaning?”) taken from the Pleasure scale of the control-autonomy-pleasure-self-realization scale (CASP-19) of quality of life in older adulthood (Hyde et al., Reference Hyde, Wiggins, Higgs and Blane2003) in ELSA, NCDS, TILDA, ELSI, and SHARE. This item was measured on a four-point scale and reverse-scored when necessary from 1 (never) to 4 (often) in ELSA, NCDS, TILDA, and SHARE and from 1 (never) to 3 (always) in ELSI. In Romania, the item that measured meaning was translated as the frequency of no meaning in one’s life. To be consistent with the other samples, we scored the item in the direction of higher meaning. In addition, TILDA had both purpose and meaning (and fluency and memory) at Wave 4.
Verbal fluency. Participants completed a standard measure of verbal fluency (Lezak, Reference Lezak2004). Specifically, participants were asked to name as many animals as possible in 60 seconds. The only exception was the WLS samples. Participants in these samples completed a category fluency task, but the category was either animals or food (randomly selected for each participant), and only a random subsample of the WLS participants received this version of the task. In addition, all participants in both WLS samples were administered a letter fluency task in which they had to generate as many words starting with either the letter “f” or “l” as possible in 60 seconds. We included the association with category fluency in the meta-analysis to be consistent with the other samples. We test and report the association with letter fluency to determine if the associations were similar with a slightly different measure of fluency. In both WLS samples, an additional covariate that accounted for category (animal versus food) or letter (“f” versus “l”) was included in the analyses.
Episodic memory. Participants completed a standard episodic memory task. Specifically, a list of words was read to the participant and the participant was instructed to recall the list immediately and again after a brief delay. All samples used a list of 10 words, except for MIDUS, which used a 15-item list. In addition, in TILDA, participants were asked to recall the word list immediately twice and again after a short delay; we used the first immediate recall score, as well as the delayed recall score. Episodic memory was the number of words recalled correctly across both immediate and delayed recall tasks.
Individual-level covariates. Covariates included self-reported age in years and gender (0 = male, 1 = female) in each sample. Education was reported in years in HRS, WLSG, and WLSS; on a scale from 1 (no school) to 12 (advanced or professional degree) in MIDUS; from 0 (no qualification) to 7 (degree) in ELSA; from 0 (no qualification) to 6 (higher degree) in NCDS; from 1 (some primary, not complete) to 7 (postgraduate/higher degree) in TILDA; from 1 (never studied) to 18 (doctoral degree/PhD) in ELSI. The 1997 International Standard Classification of Education was used to categorize and harmonize education statistics across European countries (UNESCO, 2003) in SHARE. Specifically, level of education ranged from 0 (pre-primary level of education) to 6 (second stage of tertiary education). Race was categorized in each sample based on how the data were collected and reported in each sample (see Table 1). Race was dummy-coded, with white as the reference group. SHARE does not collect information on race or ethnicity and thus was not controlled for in samples from this cohort. Year of assessment (0 = 2010, 1 = 2012) in HRS and category (0 = foods, 1 = animals) in WLS were included as sample-specific covariates for analyses on these samples.
Total N for the meta-analysis = 125,746 for verbal fluency and 141,825 for episodic memory. CI = confidence interval. GDP = gross domestic product per capita. PPP = purchasing power parity. For verbal fluency, age compared samples with a mean age below 65 years (k = 10) to a mean age above 60 years (k = 19), measure compared samples with the purpose in life measure (k = 4) to samples with the meaning measure (k = 25), SHARE compared samples in the SHARE study (k = 21) to samples that were not SHARE (k = 8), and HRS-based compared samples from the HRS suite of studies (k = 25) to samples with other methodologies (k = 4). For episodic memory, age compared samples with a mean age below 65 years (k = 12) to a mean age above 60 years (k = 25). Measure compared samples with the purpose in life measure (k = 4) to samples with the meaning measure (k = 32). SHARE compared samples in the SHARE study (k = 29) to samples that were not SHARE (k = 8). HRS-based compared samples from the HRS suite of studies (k = 32) to samples with other methodologies (k = 4).
Sample-level moderators. We examined sample-level moderators to identify sources of potential heterogeneity in the meta-analysis. We tested characteristics of the sample, specifically mean age of the sample, measure (purpose vs. meaning), samples from SHARE versus samples from other cohorts, and HRS-based versus not HRS-based (the sampling and methodology for ELSA, TILDA, ELSI, and SHARE are based on HRS, whereas the MIDUS, WLS, and NCDS are not). We also tested whether the economic environment of the country was associated with the strength of the association. The economic environment was defined as gross domestic product per capita (GDP) and purchasing power parity (PPP). These indices were obtained from the World Bank (https://www.worldbank.org/) for the year of the assessment in each sample.
All variables were standardized within sample before analysis. In each sample, linear regression was used to predict fluency and memory from purpose/meaning controlling for the covariates. Follow-up analyses tested whether the associations were moderated by age, gender, or education in each sample by entering the interaction between purpose/meaning and the demographic factor, as well as the main effects and other covariates. A random-effects meta-analysis was done on the partial correlation and sample size to summarize the association between purpose/meaning and fluency and memory across all samples. A similar approach was used to summarize the interaction effects for age, gender, and education. SPSS 25 was used for the analysis of the individual samples. The metafor package in R (Viechtbauer, Reference Viechtbauer2010) was used for the meta-analysis. We followed up the meta-analysis with meta-regressions to identify potential sources of heterogeneity. Specifically, we ran meta-regressions to determine whether the associations varied by GDP, PPP, mean age of the sample, purpose measure versus meaning measure, SHARE versus not SHARE, and HRS versus non-HRS.
We also did two supplemental analyses in specific samples. First, we tested the association between purpose and letter fluency in the WLS samples to see if the association was similar when a different measure of fluency was used. Second, we tested the association between both purpose and meaning and verbal fluency and memory using Wave 4 of TILDA, which was the wave that included both of these measures (the measures at Wave 1 of TILDA were used the meta-analyses).
Descriptive statistics for all samples are shown in Supplemental Table S1a and S1b for verbal fluency and in Supplemental Table S2a and S2b for episodic memory. Figure 1 shows the association between purpose and fluency in each of the samples and the overall meta-analysis. The association was consistent across samples: higher purpose/meaning in life was associated with better verbal fluency. The positive association was seen in all samples, even though one smaller sample did not reach statistical significance. The one exception was for Israel, where there was a small nonsignificant negative association with fluency. Overall, though, the estimated meta-analytic association across the 29 samples was nearly .10 (p < .001). Figure 2 shows the association between purpose and memory in each sample and the overall meta-analysis. Again, the association was consistent across samples: higher purpose/meaning in life was associated with better episodic memory. The positive association was seen in all 37 samples, even though it did not reach statistical significance in 2 samples (Israel, Malta). Overall, though, the estimated meta-analytic association across the 37 samples was nearly .12 (p < .001).
There was significant heterogeneity across the samples for both fluency (Q = 247.549, p < .001, I2 = 90.41) and memory (Q = 350.978, p < .001, I2 = 90.96). There was, however, no significant difference in the strength of the association for either function between studies that assessed meaning compared to purpose (Table 1). There was only one factor in the meta-regression that was significant for both fluency and memory (Table 1): these association were stronger in samples from SHARE than from the other samples. There was no difference across HRS-based studies or mean age of the sample. There was modest evidence that the associations were moderated by the economic environment of the sample. Specifically, for episodic memory, the associations were stronger in the relatively lower-income countries than in the relatively higher-income countries. There was a similar trend for verbal fluency that was not significant. This difference between fluency and memory is likely because of the eight additional counties in the analysis of memory increased the range of GDP and PPP. Finally, across all samples, there was no evidence that age, sex, or education moderated the association between purpose/meaning and verbal fluency (Supplemental Table S3). For memory, there was modest evidence that the association was slightly stronger at older ages and, consistent with the country-level moderation by GDP and PPP, at lower levels of education (Supplemental Table S4).
Finally, the supplemental analyses supported the main analyses. That is, similar to category fluency, purpose was associated with letter fluency in both WLSG (β = .077, 95% confidence interval [CI] = .049, .106, p < .001; N = 4,488) and the WLSS (β = .047, 95% CI = .009, .085, p = .014; N = 2,507). Both meaning (β = .039, 95% CI = .013, .065, p = .003; N = 4,795) and purpose (β =. 114, 95% CI = .088, .142, p < .001; N = 4,795) were associated with fluency in Wave 4 of TILDA and moderately correlated with each other (r = .42, p < .001). Meaning (β = .044, 95% CI = .020, .068, p < .001; N = 4,861) and purpose (β = .125, 95% CI = .100, .149, p < .001; N = 4,861) were likewise both associated with episodic memory in this wave of TILDA. Of note, in TILDA, the total CASP-Pleasure scale from which the meaning item was taken from was unrelated to fluency (β = .003, 95% CI = −.039, .051, p = .791) and modestly related to memory (β = .029, 95% CI = .005, .054, p = .017; N = 4,861). Finally, as an additional supplemental analysis, we excluded participants who reported ADRD from the analysis in samples that had this information (HRS, ELSI, and SHARE). The pattern of association was similar when these participants were excluded for both fluency (Supplemental Table S5) and memory (Supplemental Table S6).
The present research found that in over 140,000 participants from up to 32 countries, higher purpose/meaning in life was associated with better performance on tasks that measure specific cognitive functions. The consistency across samples was striking: 28 out of 29 samples had a positive association and 27 of those associations were statistically significant for verbal fluency. Likewise, there was a positive association in all 37 samples with episodic memory, and only 2 were not statistically significant. These associations were similar across countries from three continents (Europe, and South and North America), multiple languages, tasks that measure two cognitive functions, two different measures of fluency (letter and category), and measures of both purpose and meaning. Further, the meta-analysis of the interactions indicated that the association did not vary by sociodemographic characteristics for fluency and only modestly for memory. That is, purpose/meaning was beneficial across different ages (slightly stronger at older ages for memory), both genders, and levels of educational attainment (slightly stronger at lower educational levels for memory). The present study thus indicates a robust association between feeling that one’s life has purpose and meaning and better verbal fluency and episodic memory.
There may be a number of mechanisms that contribute to the association between purpose/meaning and cognition. First, purpose is associated with health-related behaviors and outcomes that support healthier cognitive aging. Individuals higher in purpose, for example, engage in more physical activity (Hooker and Masters, Reference Hooker and Masters2016), are less likely to smoke (Morimoto et al., Reference Morimoto, Yamasaki, Ando, Koike, Fujikawa, Kanata and Nishida2018), maintain better physical function (Kim et al., Reference Kim, Kawachi, Chen and Kubzansky2017), and have a lower burden of disease (Czekierda et al., Reference Czekierda, Banik, Park and Luszczynska2017). This healthier lifestyle may help promote better performance on cognitive tasks, such as verbal fluency and memory. Second, purpose is associated with personality traits, such as higher emotional stability, extraversion, and conscientiousness (Scheier et al., Reference Scheier2006), which are also implicated in better verbal fluency and memory (Sutin et al., Reference Sutin, Stephan, Damian, Luchetti, Strickhouser and Terracciano2019a). Third, individuals higher in purpose tend to have larger social networks (Scheier et al., Reference Scheier2006), greater social support (Musich et al., Reference Musich, Wang, Kraemer, Hawkins and Wicker2018), and fewer feelings of loneliness (Mwilambwe-Tshilobo et al., Reference Mwilambwe-Tshilobo2019). Such social integration helps support cognitive health (Kelly et al., Reference Kelly, Duff, Kelly, McHugh Power, Brennan, Lawlor and Loughrey2017), including lower long-term risk of cognitive impairment (Sutin et al., Reference Sutin, Stephan, Luchetti and Terracciano2018a). Social integration may also include more social interactions that help promote better cognitive function.
There is also increasing evidence that feelings of purpose and meaning are protective against the development of Alzheimer’s disease and other severe cognitive impairments (Boyle et al., Reference Boyle, Buchman, Barnes and Bennett2010; Sutin et al., 2020, 2018b). The processes and behaviors associated with a purposeful/meaningful life may help support healthier cognition across the lifespan. Both fluency (Sutin et al., Reference Sutin, Stephan and Terracciano2019b) and memory (Josefsson et al., Reference Josefsson, Sundström, Pudas, Nordin Adolfsson, Nyberg and Adolfsson2019) have been identified as intermediate markers of cognitive health that are predictive of incident cognitive impairment. Similar to cognitive (Stern et al., Reference Stern, Barnes, Grady, Jones and Raz2019) and social (Ihle et al., Reference Ihle2019) reserve, purpose/meaning may serve as a reserve of well-being that supports healthier cognitive aging. The present research is a step toward building a model of purpose/meaning and risk of cognitive impairment. Interestingly, there are strong lay beliefs that purpose promotes healthier cognitive aging and is protective against cognitive impairment (Vaportzis and Gow, Reference Vaportzis and Gow2018).
The association between purpose/meaning and better cognition was apparent in almost all of the individual samples, but the meta-analysis indicated significant heterogeneity for both cognitive tasks. The meta-regressions revealed only one potential source of the heterogeneity across both fluency and memory: samples from SHARE had slightly stronger associations than samples from other studies. Interestingly, the difference between HRS-based and non-HRS-based studies was not quite significant. It was, however, suggestive that the associations may be somewhat stronger in studies that used HRS methodology. There was also suggestive evidence of small differences by the economic development of the country, with slightly stronger associations in countries that have lower GDP per capita and less purchasing power parity. This moderation was significant for episodic memory and a trend for verbal fluency. This pattern suggests that, similar to other psychological factors (Luchetti, Terracciano, Stephan, Aschwanden, & Sutin, Reference Luchetti, Terracciano, Stephan, Aschwanden and Sutinin press), purpose/meaning may serve as a psychological resource for cognitive function in less economically robust countries. This analysis, however, was limited by the concentration of high-income countries and should be expanded to include more lower- and middle-income countries.
Purpose in life and meaning in life are two related yet distinct constructs. Meaning in life is typically conceptualized as a superordinate construct that is composed of several lower-order constructs, including purpose (Martela and Steger, Reference Martela and Steger2016). Although conceptually distinct, there tends to be overlap in the measurement (e.g. many scales include items on both purpose and meaning; Steger et al., Reference Steger, Frazier, Oishi and Kaler2006), and, when differentiated, the correlates tend to be similar, even if the magnitude may differ somewhat (Costin and Vignoles, Reference Costin and Vignoles2020). It was thus expected that similar associations would emerge across measures of purpose and meaning. And, in fact, there was no difference in the strength of the association between measures of purpose versus meaning in the meta-analysis for both fluency and memory. It should be noted that the assessment of meaning was based on a single item that performed as well as the 6-7-item measure of purpose. Independently from the number of items, the pattern of results suggests that measures of purpose and meaning capture similar variance associated with better performance. It may be that the lay understanding of the construct does not make a distinction between purpose and meaning. Regardless of whether participants made this distinction, however, the association with fluency and memory was the same.
Purpose/meaning in life may be a potential target of intervention for healthier cognitive aging. Purpose/meaning can be increased through intervention (Park et al., Reference Park2019); whether such interventions that increase purpose also benefit cognitive function still need to be tested. In addition to fostering a general sense of well-being, such interventions may support healthier cognitive aging through both proximal and distal pathways. In a proximal pathway, experimental manipulations of purpose may directly lead to better performance on cognitive tasks that has an immediate benefit and that also may consolidate into maintenance of cognitive function over time. In a more distal pathway, interventions for long-term change in purpose/meaning may help promote behaviors (e.g. more physical activity) and relationship quality (e.g. lower loneliness) that helps protect against long-term risk of cognitive impairment. And, in fact, improvements in psychological well-being through intervention in older adulthood reduce risk factors for cognitive decline (Delhom et al., Reference Delhom, Satorres and Meléndez2020; Wuthrich et al., Reference Wuthrich, Rapee, Draper, Brodaty, Low and Naismith2019). More research needs to test these possibilities and evaluate the full potential of interventions that focus on purpose/meaning.
The present study had several strengths, including a total sample size of up to over 140,000 participants from up to 37 samples in 32 countries. There are also some limitations that could be addressed in future research. First, the data are cross-sectional. Second, with cross-sectional data, we could not test fluency or memory as a mediator between purpose/meaning and dementia. Third, although participants were from 32 countries, most of these countries were high income. Future research could examine the longitudinal relations between purpose/meaning and fluency and memory, whether fluency and memory mediate the association between purpose/meaning and risk of cognitive impairment, and whether these associations are apparent in low- and middle-income countries.
Despite these limitations, the present research provides robust evidence that purpose and meaning in life are associated with better verbal fluency and episodic memory. Further, these associations do not seem to be limited to one demographic group or geographical region. Given that purpose/meaning can be increased through intervention (Park et al., Reference Park2019), this robustness suggests a promising novel target to promote healthier cognitive aging across the lifespan.
Conflict of interest
Description of authors’ roles
A. Sutin designed the study, obtained and analyzed the data from the individual cohorts, and wrote the paper. J. Strickhouser performed the meta-analysis and assisted with writing the paper. M. Luchetti, Y. Stephan, and A. Terracciano provided critical feedback throughout the project and assisted with writing the paper.
Research reported in this publication was supported by the National Institute on Aging of the National Institutes of Health under Award Numbers R01AG053297 and R01AG068093. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Data can be downloaded or access requested from parent studies (urls provided in the Method). The data for this study were not preregistered because the data come from large public health datasets. The analyses for this research were preregistered on OSF (https://osf.io/rywhu).
We gratefully acknowledge the parent studies whose public data made this work possible: Health and Retirement Study (HRS): The Health and Retirement Study is sponsored by the National Institute on Aging (NIA-U01AG009740) and conducted by the University of Michigan. Midlife in the United States (MIDUS): MIDUS is sponsored by the MacArthur Foundation Research Network on Successful Midlife Development (MIDUS I), the National Institute on Aging (P01-AG020166; MIDUS II), and grants from the General Clinical Research Centers Program (M01-RR023942, M01-RR00865) and the National Center for Advancing Translational Sciences (UL1TR000427). Wisconsin Longitudinal Study Graduate and Sibling samples (WLSG and WLSS): since 1991, the WLS has been supported principally by the National Institute on Aging (AG-9775, AG-21079, AG-033285, and AG-041868), with support from the Vilas Estate Trust, the National Science Foundation, the Spencer Foundation, and the Graduate School of the University of Wisconsin-Madison. Since 1992, data have been collected by the University of Wisconsin Survey Center. English Longitudinal Study of Ageing (ELSA): Funding for the English Longitudinal Study of Ageing is provided by the National Institute of Aging [grants 2RO1AG7644-01A1 and 2RO1AG017644] and a consortium of UK government departments coordinated by the Office for National Statistics. The content is solely the responsibility of the authors and does not represent the official views of the parent studies or funders. National Childhood Development Survey (NCDS): we thank The Centre for Longitudinal Studies, Institute of Education for the use of these data and to the UK Data Archive and Economic and Social Data Service for making them available. The TILDA study was supported by the Irish Government, the Atlantic Philanthropies, and Irish Life PLC. The ELSI-Brazil baseline study was supported by the Brazilian Ministry of Health (DECIT/SCTIE – Department of Science and Technology from the Secretariat of Science, Technology and Strategic Inputs (Grant 404965/2012-1); COSAPI/DAPES/SAS – Healthcare Coordination of Older Adults, Department of Strategic and Programmatic Actions from the Secretariat of Health Care) (Grants 20836, 22566, and 23700); and the Brazilian Ministry of Science, Technology, Innovation and Communication. The SHARE data collection has been funded by the European Commission through FP5 (QLK6-CT-2001-00360), FP6 (SHARE-I3: RII-CT-2006-062193, COMPARE: CIT5-CT-2005-028857, SHARELIFE: CIT4-CT-2006-028812), FP7 (SHARE-PREP: GA N°211909, SHARE-LEAP: GA N°227822, SHARE M4: GA N°261982), and Horizon 2020 (SHARE-DEV3: GA N°676536, SERISS: GA N°654221) and by DG Employment, Social Affairs & Inclusion. Additional funding from the German Ministry of Education and Research, the Max Planck Society for the Advancement of Science, the National Institute on Aging (U01_AG09740-13S2, P01_AG005842, P01_AG08291, P30_AG12815, R21_AG025169, Y1-AG-4553-01, IAG_BSR06-11, OGHA_04-064, HHSN271201300071C) and from various national funding sources is gratefully acknowledged (see www.share-project.org). This paper uses data from SHARE Waves 2 (10.6103/SHARE.w2.700), 4 (10.6103/SHARE.w4.700), 5 (10.6103/SHARE.w5.700), 6 (10.6103/SHARE.w6.700), and 7 (10.6103/SHARE.w7.700).
To view supplementary material for this article, please visit https://doi.org/10.1017/S1041610220004214