Vitality Watch

Dossier pubmed observational 18-sep-2026
Human cohort, case-control, or epidemiological association

Purpose in Life Among Older Adults: A Machine Learning Analysis of Potential Predictors.

Key claim: In a nationally representative U.S. older-adult sample, machine learning ranked self-esteem and positive affect as the strongest predictors of purpose in life, with psychological resources outweighing demographics.

Abstract

A sense of purpose is a central component of mental well-being in older adulthood. Although many predictors of purpose have been identified, most studies have considered a limited set, leaving their relative importance unclear. The present study addressed this gap by examining a large set of psychological, social, and demographic variables in a nationally representative sample of U.S. adults from the Health and Retirement Study. Extreme Gradient Boosting (XGBoost) was used to evaluate 47 predictors. Self-esteem emerged as the strongest predictor, followed by positive affect. The leading predictors were organized into six domains: adaptive self-related beliefs, affective well-being, perceived control, optimism, social engagement, and spirituality/religiosity. Demographic variables were less predictive overall. These findings indicate that individual differences in purpose are associated with a broad set of psychological resources involving self-regulation, interpersonal processes, optimism, and positive affect.

Why this matters

In a nationally representative U.S. older-adult sample, machine learning ranked self-esteem and positive affect as the strongest predictors of purpose in life, with psychological resources outweighing demographics.

Source

Link

Briefing card

Purpose in Life Among Older Adults: A Machine Learning Analysis of Potential Predictors.
Stage: observational · pubmed · 18-sep-2026

In a nationally representative U.S. older-adult sample, machine learning ranked self-esteem and positive affect as the strongest predictors of purpose in life, with psychological resources outweighing demographics.

Cross-references (0)

None recorded — doctrine links and citations appear here as scans and citation sweeps find them.

External: doi:10.1177/07334648261490460

build build 880 · 2bcef25 · 2026-11-18