QCA Fuzzy Sets and the F Statistic. 30 minutes long.

Qualitative Comparative Analysis and the F Statistic
BY WENDY OLSEN
PROF AT UNIV OF MANCHESTER
This lecture tells how fuzzy set analysis works, then moves at about 15 minute-mark toward how to run a statistical test of the claim that X is sufficient for Y. Here's an abstract and fuller deatils of online resources.

At the Research Methods Festival, Olsen presented on: "QCA and Fuzzy Set Goodness–of-Fit Tests". This important date 6 July 2016 marked our upload of programmes and supporting documentation to Github for doing QCA with statistical tests afterward. The materials are described in the abstract below.

Title: QCA and Fuzzy Set Goodness–of-Fit Tests. Abstract: The method known as qualitative comparative analysis comprises a set of related case-comparative techniques for studying data that originates as case-wise qualitative, or otherwise detailed, evidence. If it starts off qualitative, the QCA stage involves systematically calibrating the data to fuzzy set or crisp set scores. If the data set is a secondary one with continuous variables, again these must be calibrated. I explain the calibration method, and how consistency is measured. I then augment these QCA techniques with a statistical F test of the result. The F test acts both as a hypothesis test and as a measure of goodness of fit of the data to the pattern predicted under either a ‘sufficiency’ or ‘necessity’ claim. The sufficiency claim is explored using empirical data from India. These data involve 450 female cases, mostly married and with spousal and household evidence available from a related primary research project. The outcome is whether they did remunerated work. The F test involves a comparison of the distance from sufficiency in a fuzzy set context, compared with the distance we expect under a null hypothesis. We found that having land and not having a husband were sufficient together (but not separately) for a woman to work. Users can get the Consistency and F test results (p-values) from either Excel or our online freeware. Wendy Olsen thanks John McLoughlin for his programming help in Python. We also acknowledge funds given by the British Academy: Innovation in Global Labour Research Using Deep Linkage and Mixed Methods project, see also   / mixednetwork   .

The Github address is https://github.com/WendyOlsen/fsgof.

We advertised this programme to JISCMAIL QUAL-COMPARE (now at 185 members).

The Facebook group was converted by Facebook to a study group, see   / mixednetwork  .

There are real examples from India here!

Grateful to the ESRC DFID Poverty Alleviation fund which sponsored the primary field research.

Apologies for any or all errors, and for the failings of the camera system. But it's not too bad.

We are very grateful to British Academy who funded this research.

We are also grateful to the Cathie Marsh Institute for Social Research (CMIST), who held the Symposium on Social Mobility and Labour Markets where this video was recorded.

http://www.cmist.manchester.ac.uk/res...