Loneliness, verbal fluency, and social engagement over time: A mediation analysis

Background Lonely individuals are more likely to socially withdraw, an effect attributed to social distrust and preference for solitude (Ernst and Cacioppo, 2000). However, it is alternatively possible that loneliness causes cognitive deficits, which in turn precipitate social withdrawal. Loneliness has been shown to be associated with decreased verbal fluency in the literature (Schnittger et al., 2012; Boss, 2015; Shankar et al., 2013). We wanted to investigate whether loneliness a) was related to changes in social engagement over time, and b) predicted verbal fluency, which in turn precedes social withdrawal, possibly due to communicative difficulties. Methods Data from 4715 participants who were not institutionalised, did not have dementia at baseline, and were involved in waves 3, 4, and 5 of the English Longitudinal Study of Ageing (ELSA), were used. Residualised change scores were calculated to evaluate changes in engagement in social activity between waves 3 and 5. A mediation model (within a Structural Equation Modelling framework) was used to evaluate whether verbal fluency (measured using an animal naming task at wave 4) mediated the association between loneliness at wave 3 (measured using the UCLA Loneliness scale; Russell, 1996) and residualised change scores in engagement in social activities between waves 3 and 5. Results Overall, engagement in social activities decreased from a median of 3 activities to 2 activities between waves 3 and 5. This change was predicted by both loneliness (Beta = -0.056, p<0.001) and verbal fluency (beta = 0.083, p<0.001) such that loneliness predicted a decrease and verbal fluency an increase in social engagement. Verbal fluency was a mediator of the association between loneliness and changes in social engagement between waves (indirect effect = -0.009, p