Big-5 Personality and Loneliness in Older Adults: A Growth Curve Model for Two Parallel Processes

Background and Purpose: Personality traits have been empirically supported as significant predictors of loneliness, yet longitudinal analysis of the relationship is scarce. While cross-sectional evidence on personality and loneliness could give an impression that personality traits are stable in later life, personality change occurs through the life course. According to the identity theory of personality change, as older adults disengage from their social roles in their middle life stage, their personalities could change with the identity shifts. The perceived level of loneliness also fluctuates in later life. While existing research has investigated the change in personality and age-related trajectories of loneliness separately, longitudinal investigations are needed in order to deepen the understanding of the dynamics between personality and loneliness among older adults. This current study aims to describe the within-person change in personality and loneliness over time and to explore the associations between the change in personality and change in perceived loneliness.

Methods: This study used 2006, 2010, and 2014 waves of data from the Health and Retirement Study (HRS). Individuals with moderate to severe cognitive impairment and cases responded by proxy of the older adults were not asked to rate their perceived loneliness, hence, were excluded from the analysis. The working sample included 8,855 participants. Personality characteristics were measured by the Big-5 personality traits using the Revised Midlife Developmental Inventory Personality Scale. Loneliness was measured with the 3-item UCLA loneliness scale. Longitudinal Growth Curve Models (LGCM) for two parallel processes were run with Mplus 8.

Results: The mean age of the study sample at baseline was 68.29 (SD=9.81). The model fit indices suggested that linear LGCM models fit openness, conscientiousness, extraversion, and agreeableness. The signs of random slopes of these four personality traits all were negative, indicating a decreasing trend over time (openness: B=-0.052, SE=0.006, p<.001; conscientiousness: B=-0.031, SE=0.005, p<.001; extraversion: B=-0.048, SE=0.006, p<.001; agreeableness: B=-0,024, SE=0.005, p<.001). However, the change in neuroticism did not follow a linear trajectory. An LGCM with free time scores were fitted in order to describe the latent change trajectory in neuroticism. The slope suggested that neuroticism increased over time (B=1.271, SE=0.028, p<.001). The mean score of the random slopes of loneliness indicated a decrease in loneliness longitudinally (B=-0.008, p<.05). A higher baseline loneliness score predicts a less decreasing extraversion slope (B=0.031, SE=0.011, p<.004). Baseline neuroticism was negatively associated with the slope of loneliness (B=-0.030, SE=0.012, p<.05). All the personality change slopes were positively associated with the slope of loneliness, suggesting the change in the two parallel processes are positively related.

Conclusions and Implications: Compared to cross-sectional studies, findings of the current LGCM analysis pointed out that baseline personality traits could influence the trajectory of change in loneliness and vice versa. The growth slopes of personality traits and loneliness were positively associated, suggesting the changes in the two parallel processes are related longitudinally. Future research may consider to further explore the linkage between loneliness and personalities over time. Interventions designed to address loneliness in older adults could take personality traits into considerations.