I know that above 6 independent variables the methodology becomes too complex to operate with, but I am exploring with the fsQCA software and it seems that with 12 independent variables the system simply stops working. Is that normal?
As far as I know 12 independent is maximum in fsQCA, but I don't know about Tosmana or Kerq. Anyway, already 6 factors is much as you seem to have noticed.
Hi Ruben! The number of your conditions (it's better this nomenclature rather 'variables' in QCA) depends on the number of your cases. For instance, as Berg-Schlosser & de Meur (2008) mention in "an internediate-N analysis (10 to 40 cases) would be to select from 4 to 6-7 conditions" (p. 28). You can check as well the possibility to use R for making a fs/QCA analysis (see: Compasss.org). Tosmana only accept 5 conditions and with more than that is not possible to get the Venn diagram! Good luck with your research!!
Thank you very much for your deep answer Priscilla and your wishes! Good luck too!
Dear Ruben, your question has been asked quite a while ago already, but since other researchers might be interested in the same question and thus look it up, let me give you a belated answer nonetheless.
The number of exogenous factors you can include in your analysis does not depend on your number of cases, but the higher the diversity index (the ratio of the number of observed configurations to that of all logically possible configurations), and the more cases per configuration relative to your target population, the more credible your results.
While the fs/QCA programme indeed works with up to 12 exogenous factors, Tosmana can sometimes handle up to 20 such factors (I've done analyses with 23!), about the same number the QCA package for R can manage. But notice that only the QCA package (version 1.1-4) is currently able to identify the entire model space of a solution (i.e. all models that fit your data equally well), whereas this is not the case for the fs/QCA software and Tosmana!
A bit late in the topic, but as algorithms evolve the answers change too.
With fs/QCA it is about 12-13 conditions (depending of course on the number of observed cases), but for QCA in general it is actually possible to get solutions for up to 30 causal conditions, using the latest versions (above 3.0) of the QCA package in R:
https://cran.r-project.org/web/packages/QCA/index.html
The latest and cutting edge algorithm is called CCubes and it has recently been accepted for publication in the R Journal:
https://journal.r-project.org/archive/2018/RJ-2018-062/index.html
I hope this helps,
Adrian
- Badges
- Science method
- Similar topics
- Methodology
- Research Methodology
- Qualitative Research