Replication data for: Ecological dimensions explain the past but do not predict future changes in trust

  • Olga Stavrova (Creator)
  • Anthony Evans (Creator)
  • Mark J. Brandt (Michigan State University) (Creator)



Concerns about declining trust and rising cynicism are recurrent in academic research and the media. Yet, prior studies focused on explaining, rather than predicting, temporal changes in trust. We tested prediction models of trust change across (up to) 98 countries over six measurement waves (from 1981 to 2014). We tested whether different ecological predictors (e.g., pathogen prevalence, population diversity, inequality) explain the past and predict future trust levels across countries. We used societal growth curve models to disentangle between- from within-country effects and evaluated the accuracy of the models’ out-of-sample predictions using the train-test split method: We used data from 1981– 2009 to “train” the models and obtain predictions of trust for the period of 2010–2014. None of our models was more accurate in predicting future trust than a simpler baseline model. Moreover, we did not observe a universal decline in trust. Instead, temporal changes in trust were country-specific, highlighting the locality of cultural change. Most ecological predictors were correlated with between-country differences in trust. Only resource availability and moral opinion polarization were associated with within-country changes in trust: Countries that became less wealthy and more morally polarized over time also became less trustful. These results highlight important differences between explanatory and predictive models and suggest that ecological theories of trust might be of limited use when predicting future cultural shifts.
Date made available16 May 2022

Cite this