Cross-lagged relationships between sense of purpose in life, memory performance, and subjective memory beliefs in adulthood over a 9-year interval

Abstract

Objectives

Previous studies have found a positive association between having a sense of purpose in life and memory functioning in old age. We extend these findings by examining the relationships between sense of purpose, memory performance, and subjective memory beliefs over time in a large sample of adults in mid to later adulthood.

Method

We used data from 3633 participants of the second and third wave of the MIDUS study. Cross-lagged panel analysis investigated the relationships between the variables at the two points, which were approximately 9 years apart, while controlling for gender, age, education, positive and negative affect, and self-rated health.

Results

Sense of purpose in life, memory performance, and subjective memory beliefs were all cross-sectionally related to each other at both times. Longitudinally, sense of purpose was a positive predictor of subjective memory beliefs. Memory performance and subjective memory beliefs positively predicted each other over time. Furthermore, all three variables showed correlated changes over time. Exploratory analyses suggest that the covariates of affect and self-rated health are possible mediators or confounders in respectively the relationship between subjective memory beliefs and later sense of purpose, and sense of purpose and later objective memory performance.

Conclusion

Our findings underscore once more the relevance of sense of purpose in life as a predictor of positive late life functioning, as it is related to both performance-based and subjective cognitive outcomes. More work is needed to understand mechanisms underlying the purpose-memory association in order to develop and implement purpose interventions.

Keywords:

• Psychological well-being
• cognition
• aging
• longitudinal

Disclosure statement

The authors report no conflict of interest.

Notes

1 We also tested models where age, gender, and education were only added as covariates at T2, similar as one would in regression analyses. These models showed worse fit, although the primary conclusions with regard to our main hypotheses were not altered.

Additional information

Funding

This work was supported by the Research Foundation Flanders under grant 1109417 N (to LD).