Understanding Micro-Level Causes of Corruption: Integrating Experimental and Configurational Approaches

Why do individuals engage in corruption? This question shapes part of the current debate in the field of corruption research, which has developed from a focus on macro-level, index-based investigations to mixed-method and multi-level approaches to understanding the causes of corruption. Especially in the context of corruption in public administration, individual motives have gained relevance (de Graaf, 2007; Jancsics, 2014; Navot et al., 2016; Roman & Miller, 2014; Thomann et al., 2025; Weißmüller & Zuber, 2023). Yet micro-level explanations face a persistent challenge: corrupt behaviour cannot be understood through isolated motives alone, but emerges from the interaction between psychological dispositions, organizational norms, and situational opportunity structures. Conceptually, this paper builds on an interactionist understanding of corruption as a situationally embedded misuse of public office, in which accountability deficits arise from the conjunction of individual and contextual conditions. The paper introduces a novel mixed-methods design for analysing micro-level causes of corruption that explicitly addresses this causal complexity. It combines a survey experiment with fuzzy-set Qualitative Comparative Analysis (fsQCA) to integrate the strengths of experimental and configurational approaches. While experiments offer strong causal leverage over isolated factors, they often abstract from the complexity of real-world decision environments. Configurational methods, by contrast, foreground conjunctural causation and equifinality but typically rely on observational data, limiting causal inference. By embedding experimental manipulations within a configurational framework, this study bridges these approaches. The survey experiment provides controlled variation in key individual-level conditions, such as normative cues, perceived detection risk, and moral justifications, while fsQCA is used to identify combinations of experimental treatments and contextual characteristics that are sufficient for corrupt behaviour. This design captures core features of corruption processes, such as equifinality, causal asymmetry, and conjunctural causation, that are difficult to model with additive statistical approaches such as OLS regression. Building on established theoretical insights into the micro-foundations of corruption (e.g. Dong & Torgler, 2009; Sundström, 2016; Thomann et al., 2025; Weißmüller et al., 2022; Weißmüller & Zuber, 2023), the study translates these insights into experimentally testable and configurationally analysable conditions. By combining experimental control with configurational analysis, this study offers a methodological contribution that enhances both causal inference and theoretical sensitivity to complexity in the study of individual-level corruption. Beyond its methodological contribution, the paper advances corruption diagnostics at the micro level by offering a more fine-grained way of identifying distinct risk profiles within public organizations. Rather than asking whether single factors increase the likelihood of corruption, the proposed approach uncovers multiple causal pathways through which different constellations of motivations, constraints, and opportunities lead to corrupt behaviour. In doing so, the study contributes to the development of mixed-source and multi-level strategies for measuring and understanding corruption and provides a basis for more targeted integrity management and accountability interventions in public administration.