This study explores the associations of loneliness with depressive symptoms in a five-year follow-up and describes how the six dimensions of perceived togetherness explain loneliness and depressive symptoms at baseline. The data were collected on 207 residents of Jyväskylä, central Finland, who at baseline in 1990 were aged 80; and 133 residents who at follow-up in 1995 were aged 85. Loneliness was assessed using a questionnaire item with four preset response options, perceived togetherness using the Social Provisions Scale, and depressive symptoms using the CES-D scale. A recursive structural equation model showed that in women but not in men, depressive symptoms predicted more experiences of loneliness. Those who were lonely were more depressed (CES-D score 16 or over) and experienced less togetherness than those who were not. Loneliness was explained by reliable alliance, social integration and attachment; and depressive symptoms were explained by guidance, reassurance of worth, reliable alliance and attachment. A common feature in both loneliness and depressive symptoms was a lower level of perceived emotional togetherness in social interaction.

The intervention did not alleviate depressed mood. Positive changes over time were observed in loneliness, feelings of melancholy, attachment, and guidance but these occurred independently of the intervention. Our secondary analysis suggests that the intervention increased perceived social integration. In sum, the effects of the intervention were moderate only and did not expedite further overcoming depressive mood or loneliness.

The objective of this study was to study the structure and stability of perceived togetherness using cross-sectional and longitudinal confirmatory factor analysis models. Togetherness was measured with the self-administered Social Provision Scale (SPS). The sample consisted of 111 persons interviewed in connection with the Evergreen project's 5-year follow-up at ages 80 and 85. One and two-factor models were specified separately at two different points of measurement. The stability of the factor models'structure and the level of latent factors was studied using longitudinal factor analysis models. The two-factor model fit the data better than the one-factor model. Perceived togetherness was differentiated into factors of emotional and social togetherness. The structure of perceived togetherness remained unchanged during the 5-year follow-up. In terms of means, the latent factors showed a statistically highly significant decline in time (i.e. perceived togetherness decreased with increasing age).

No abstract

No abstract