QCA intermediate solution common condition.

[Humberto Brea Solis]

Hi Adrian

Sorry for posting the message in your previous discussion post. I read afterwards that you have now this new forum. 

I have a simple question and so far I have not been able to find the answer: 

I have computed the intermediate and parsimonious solutions: Let's say that I found the following (from my mind example)

Intermediate solutions: 

~A*B*C*~D
~A*~D*E
~A*B*E

And the parsimonious solution is 
B + C + E

I found it strange that ~A is not part of the parsimonious solution while it is in every one of the intermediate solutions. How is this possible? 

I am using the R package QCA. 

[Ian Greener]

Hi,

I am not Adrian (and he's a lot smarter than me), but the answer is probably in the definitions of the two kinds of solutions. The intermediate solution includes only the counterfactuals you specify as being relevant as candidates in its calculation (the directional expectations). The parsimonious solution includes all counterfactuals as potentially relevant, and so generates the simplest possible solution. It is likely that the exclusion of ~A is as a result of this difference - the parsimonious solution is working through a slightly different minimisation problem.

Best wishes,

Ian

[Adrian Dușa]

Dear Humberto and Ian,

It is a seemingly simple question, but the answer proves to be not exactly trivial and it has escaped me (too) for quite some time before I finally decided to delve in the very details and implement the conjunctural directional expectations (CDEs).

Ian's explanation is shared by the vast majority of QCA users (and teachers!), and while close is not precisely what actually happens.

In fact, the remainders selected for the intermediate solution are selected <not> (just) according to the directional expectations, but:

- according to the various intersections between the conservative and parsimonious solutions,

- that additionally are consistent with the directional expectations

While presenting the concept of CDEs, I had to first make it crystal clear how the remainders are selected. Attached you can find my presentation from the 2019 QCA Expert Workshop from Zurich, which hopefully takes the reader through a step by step explanation of the process. In the presentation, the remainders are then selected according to the "final selection terms" which might or might not be a perfect representation of the directional expectations, function of the structure of the conservative and parsimonious solutions models.

It has originally been presented by Ragin and Sonnett, but not at this detailed level.

This answer might not perfectly explain your example, but it offers a hint as to how your example might be solved. In fact, yours cannot even be solved without the structure of the conservative solution, anyways.

Hope this helps,

Adrian

--
You received this message because you are subscribed to the Google Groups "QCA with R" group.
To unsubscribe from this group and stop receiving emails from it, send an email to qcawithr+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/qcawithr/76f6b4b8-ca2b-475d-8e45-9f261089ba44n%40googlegroups.com.

--

Adrian Dusa
University of Bucharest
Romanian Social Data Archive
Soseaua Panduri nr. 90-92
050663 Bucharest sector 5
Romania

https://adriandusa.eu

[Humberto Brea Solis]

Thank you Adrian and Ian for your help. Thank you Adrian for this detailed explanation and for the presentation as well.!