Disentangling General and Specific Institutional Trust: A Bifactor Analysis Across Generations

Institutional trust is frequently measured as a unidimensional construct or via single-item indicators with limited reliability and validity. This paper introduces and validates a new, multidimensional instrument designed to measure trust across four distinct domains: national government, courts, health institutions, and scientific institutions. By applying a bifactor modeling approach, we demonstrate that institutional trust simultaneously comprises a general institutional trust propensity and domain-specific trust evaluations of separate institutions. We validate this instrument using a three-stage design in the Czech Republic involving a pretest (N=261), an exploratory sample of adults under 50 (N=468), and a confirmatory sample of older adults aged above 50 (N=767). This split-sample design allows for rigorous testing of measurement invariance and factor structure across multiple independent samples and age cohorts. Our primary finding is that the bifactor model offers superior fit compared to standard hierarchical or correlated-factor models. This supports the existence of a “General Institutional Trust” factor (the shared variance representing a generalized trust in the system) alongside distinct, specific factors for trust in government, courts, health, and science institutions. Crucially, we show that once this general trust factor is controlled for, the correlations between specific institutions diminish significantly. This implies that prior research, by failing to account for the general trust baseline, may have overestimated the interdependence of trust across different institutions. Finally, we establish concurrent and discriminant validity by linking these specific factors to distinct related constructs. For instance, trust in government is uniquely associated with political alienation and conspiracy mentality, while trust in health institutions is uniquely associated with subjective well-being and perceived health. The presented instrument provides political scientists with a reliable tool to assess trust (defined close to legitimacy) in multiple democratic institutions and to disentangle generalized institutional system support from evaluations of trust in separate institutions.