Date created: | Last Updated:

Category: Project

Description: In comparative research, a researcher makes an ecological inference when inferences about a lower-level unit are based on a higher-level analysis using aggregated data. In quantitative research, it is known that ecological inference carries the risk of an ecological fallacy. Thus far, ecological inference and the risk of an ecological fallacy have been ignored in methodological and empirical research using Qualitative Comparative Analysis (QCA). In this paper, I show that QCA researchers should be more attentive to ecological inference when using aggregated micro data. I present a simple example and run Monte Carlo simulations to demonstrate that an ecological fallacy can occur in QCA. The set relation found with aggregated micro data might be the same as the inference based on a micro-level analysis. However, it is also possible that one infers the absence of a set relation on the aggregated level when it is present on the micro level, and vice versa. The lesson is that empirical QCA researchers should work with micro data when it is available and take into account that they might commit an ecological fallacy when forced to use aggregated data due to the unavailability of micro data.