Small, Medium and Large Ns in QCA

When social researchers refer to "small-N," medium-N" and "large-N" it is typically from an (unstated) conventional quantitative orientation: what's required achieve statistical significance. From this perspective, most published QCA studies would be considered small- or, perhaps, medium-N. But QCA is not bound by degrees-of-freedom restrictions and we have different criteria and concerns regarding the number of observations under investigation.

As a case-oriented comparative method, what matters for QCA is the depth and evenness of our knowledge across cases. In small-N QCA studies, the researcher typically possesses a great deal of in-depth knowledge about every case, and QCA is deployed to provide cross-case analysis. In medium-N QCA studies, case knowledge tends to be uneven. The researcher may know a great deal about some cases and less about others. Most classic QCA projects fall into this category. The best of these projects facilitate a two-way dialog between individual case studies and recipe configurations. Many recent QCA projects are increasingly large-N, in which case knowledge is largely lacking and QCA's cross-case comparison is deployed as the primary analytic vehicle. These projects are often embellished with selective case studies, used to illustrate or unpack the solutions identified by the QCA.

It is not the number of observations per se that matter in QCA but, rather, our level of case knowledge and how evenly distributed that knowledge is across our cases. Both of these can vary from project to project. Therefore, I am hesitant to attach specific numbers to defining "small-," "medium-," and "large-N" QCA projects. A well-funded multi-year project that employs a team of country-level experts, for example, will typically be able to support a deeper level of case knowledge across a greater number of cases than an individual researcher working on their own. Publication venue matters as well. Shorter journal articles cannot offer the same opportunities for in-depth case exploration as longer monographs.

With these qualifications in mind, I offer my observation that small-N QCA generally encompass fewer than 10-15 observations, medium-N QCA around 12-50 observations and large-N QCA, more than 30. Again, the key criteria is not the number of observations but, rather, the researcher's distance from their cases. There is indeed a relationship because as the number of observation's increases, one's distance from those cases tends to increase as well. But the researcher lacking theoretical and substantive knowledge of 9 cases is more disconnected than one who, deeply steeped in their 34, has carefully and consciousnessly developed their theoretical understanding.

Of course, this is the case for all of empirical social research and does not just apply to QCA. But from this perspective, perhaps in QCA we should consider refraining from numbers and instead quantify depth of knowledge?