Objectives: We studied the associations between perceived togetherness, depressive symptoms, and loneliness over a six-month period among 222 people aged 75–79 who reported loneliness or depressive mood at baseline.

Method: The present cross-lagged models utilized baseline and six-month follow-up data of a randomized controlled trial that examined the effects of a social intervention on loneliness and depression (ISRCTN78426775). Dimensions of perceived togetherness, i.e. attachment, social integration, guidance, alliance, nurturance, and reassurance of worth, were measured with the Social Provisions Scale, depressive symptoms with a short form of the Geriatric Depression Scale, and loneliness with a single item.

Results: After controlling for baseline loneliness and depressive symptoms, baseline higher attachment in all participants and baseline higher opportunity for nurturance in the social intervention group predicted lower depressive mood at follow-up. No cross-lagged associations between the dimensions of perceived togetherness at baseline and loneliness at follow-up were observed. In addition, depressive symptoms and loneliness at baseline tended to negatively predict the dimensions of perceived togetherness at follow-up.

Discussion: Depressive symptoms and loneliness appear to be precursor for perceived togetherness, rather than dimensions of perceived togetherness to be antecedents of loneliness and depressiveness among older people.