VARIETIES OF LONELINESS – TOWARDS A NEW TRUST PREDICTOR Exploring the relation between loneliness and political trust

Abstract

In this thesis, the focal relationship between loneliness and political trust is tested. Literature from the field of social psychology has found loneliness to impair several variations of trust. Simultaneously, loneliness theoretically fits as a predictor in the same relationship when analyzing the field of political behavior. A negative relation between the two variables is thus hypothesized. In order to use loneliness as an independent variable, it is categorized into five sub-varieties in accordance to previous literature and established loneliness tests. These sub-varieties are then used in corresponding models in order to discern potential differences in effects among them. The relation is tested through two sub-studies. In study 1, cross sectional data from the European social study was used (n = 100 000-350 000 depending on model). All tested varieties indicated significant results that were aligned with the hypothesis, however with rather small effects (standardized coefficients: .005-.042). Used method was OLS regression. In study 2, panel data from the Swiss household study was used (n = 3789 to 3950 respondents/groups). Two out of four varieties were significant at p<0.05, and they both supported the hypothesis through three models. The effect was of similar size as of study 1, but with limited R-squared. Used method was random effects regression. This thesis concludes that loneliness appears to have a small but significant effect on political trust. In terms of the varieties the results were inconclusive, with no variety having a considerably larger effect than other by aggregate.