Calibration of fuzzy set scores
The most important step after data-gathering in fuzzy-set QCA is calibration. Calibration is another word for assigning specific membership scores to cases in a certain condition, on a scale from 0 meaning
'fully out of the set', to 1 meaning 'full membership in the set'. As noted above, the data used for this QCA analysis is survey data. We applied a theory-/case-based calibration method, thus the fuzzy set scores are calibrated on the basis of knowledge derived from theory, specific case knowledge or social knowledge (Schneider and Wagemann 2012). In Table 3.1 we provide an overview of all the calibrated conditions.
Overall, we applied a strict calibration in order to avoid overestimated results. We also performed a series of nine additional analyses as robustness tests by redoing the analyses with even stricter calibrations of the conditions/outcomes related to individual and organizational upward accountability mechanisms and perceived accountability towards society. In addition, we performed robustness tests with analyses excluding the single private law agency in our sample. We refer to these robustness tests when they provide results substantially different from the original analyses reported in this chapter (Schneider and Wagemann 2010). Found patterns are only considered as giving full support for the hypotheses to the extent that these patterns are corroborated by the robustness tests.