Disentangling Corruption Across Public Institutions: Perceptions, Experiences, and Citizen Evaluations

Despite extensive efforts to improve corruption measurement, most comparative research still relies on unidimensional indicators that conflate heterogeneous phenomena and overlook how citizens evaluate corruption. This paper contributes to emerging multi-level and mixed-measurement approaches by disentangling different forms of corruption and examining how perceived and experienced corruption operate through distinct evaluative mechanisms to shape institutional trust and citizen satisfaction. Building on expectancy-based theories of public evaluations (Brandsma & Schillemans, 2013; James et al., 2016; Marvel & Girth, 2016; Van Ryzin, 2013), the study interprets perceived corruption as capturing primarily normative evaluations of institutional integrity (how institutions ought to behave), while experienced corruption reflects predictive expectations about institutional behaviour, rooted in direct encounters with public officials (Favero & Kim, 2021; Hjortskov, 2020; Kumlin et al., 2024). From this perspective, perceptions and experiences should not be treated as interchangeable measures of corruption, but as analytically distinct dimensions operating at different levels. The empirical analysis uses original survey data from Canada and Switzerland (n = 5,597), covering twelve types of public institutions and different corruption types, including bribery, abuse of office, misappropriation of public funds, and nepotism. The design moves beyond aggregate indices by combining institution-specific measures of corruption perception, corruption experience, trust, and satisfaction, thus enabling a fine-grained assessment of how corruption is evaluated across sectors. The two-country comparison follows a most-similar systems logic, allowing institutional effects to be examined under different conditions of trust and corruption salience. The findings reveal three key patterns. First, in both countries, perceived corruption systematically exceeds reported experience, confirming that focusing exclusively on experience-based indicators, especially in “low-corruption” is not enough to understand corruption risk and impact. Second, regression models show that trust is the strongest and most consistent predictor of satisfaction, particularly in Canada. However, perceived corruption frequently exerts an independent and statistically significant effect on satisfaction, even after controlling for trust. By contrast, direct experience of corruption has limited explanatory power, affecting satisfaction only in a small number of high-exposure institutions. Third, cross-national differences are theoretically informative: in Switzerland, a high-trust, low-experience context, perceptions of corruption play a more prominent role than trust itself in shaping satisfaction, suggesting heightened sensitivity to perceived norm violations rather than experienced misconduct. These results underscore the importance of treating corruption as a multi-level, multi-causal phenomenon rather than as a unidimensional one. By empirically separating perceptions from experiences and situating them within an expectations-based framework, the paper demonstrates how different corruption measures capture distinct relational dynamics between citizens and institutions. The study, therefore, advances corruption measurement by showing why integrating micro-level experiences with meso- and macro-level perceptions is essential for understanding how corruption affects trust, satisfaction, and democratic legitimacy.