Qualitative Comparative Analysis by Patrick A. Mello

Author:Patrick A. Mello [Mello, Patrick A.]
Language: eng
Format: epub
ISBN: 9781647121440
Google: wjU1zgEACAAJ
Barnesnoble:
Publisher: Georgetown University Press
Published: 2021-12-01T00:00:00+00:00

How does this look in practice? To illustrate the use of the RoN measure, let us take a simple example involving the condition X, the outcome Y, and hypothetical data on just four cases. Table 6.5 shows that, formally speaking, X is an almost perfect superset of Y because the values for X are almost always larger than or equal to the values for Y. On this basis, the condition X might be considered a necessary condition for Y.

The right-hand side of table 6.5 further shows the results for the calculation of the consistency and coverage for necessary conditions, and the relevance of necessity. As expected, we can see that at 0.95, the set-theoretic consistency is very high, satisfying the formal threshold for necessary conditions (equal to or above 0.90). The coverage is lower, but at 0.59 it would not immediately prompt concern. However, we can see that the RoN indicator is closer to 0 than to 1, suggesting a potentially trivial necessary condition.

Why is this? This being a hypothetical example, we have no substantive knowledge of the underlying data. But what we can see is that X shows little variation, with three of four cases at values equal to or close to 1. With data patterns like this, the consistency measure would always satisfy the criterion for a necessary condition. However, the RoN measure suggests that we should be cautious before treating it as a relevant necessary condition for the outcome. Ultimately, dealing with data patterns like this is a matter of interpretation. There can be situations where a condition is almost a constant, but still the condition may have substantive importance and relevance as a necessary condition. However, such an interpretation would need to be justified explicitly. As a rule of thumb, any potential necessary condition that meets the consistency benchmark of 0.9 should be checked for its coverage and relevance. If the latter two metrics fall below 0.5, this suggests that we may be dealing with a trivial necessary condition. In order to make an informed judgment on this, we should always examine the empirical distribution of our cases and their set-theoretic membership scores. A good way to do this are histograms and X–Y plots of the raw and calibrated data (on this, see also chapter 10).

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.