Re: [QCA] conceptual question SUIN

[Ingo Rohlfing]

If you want to test a specific disjunction for being SUIN, I would create it by hand and feed it into the QCAfit() function. If there are multiple disjunctions or the SUIN analysis is exploratory, I suggest using superSubset() from the QCA package. You can negate the outcome with ~ and you can determine how many disjunctions you want to see by choosing the minimum level of consistency, coverage and ron.

Kind regards

Ingo

Am 24.03.2023 um 16:20 schrieb Alexander Strelkov:

Dear group members, good afternoon

    Conceptual question on my side. When we analyze necessity, we check for conditions necessary for the presence and absence of the outcome (via QCAfit). However, when assessing SUIN this is done only for the presence of the outcome. Is it so and why?

    Does it mean that there is no option to introduce something like "neg.out=TRUE" into a code analyzing SUIN? So that to check the absence of the outcome we would need to construct another column in the dataframe?

    Hope my question makes sense and thank you for your time,

Alexander

--
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/77dc1e0c-3799-4847-a2ab-e15ce534ae1fn%40googlegroups.com.

-- 
Professor für Methoden der Empirischen Sozialforschung (Methods for Empirical Social Research)
phone: +49 851 5092720
fax: +49 851 5092722
Philosophische Fakultät
Innstr. 41
Universität Passau
D-94032 Passau
personal website: https://ingorohlfing.wordpress.com

[Adrian Dușa]

I concur to Ingo’s suggestion.

One analyses SUIN conditions mainly for necessity, in the same way that INUS conditions are analysed for sufficiency.

To add my 2 cents, I would recommend taking a look at the function pof() from package QCA.

Using this function, it is possible to actually test sufficiency and necessity by actually writing expressions, rather than creating the conditions prior to calculating the parameters of fit. It saves one step, and it is (in my opinion) more clear.

Something like:

pof(LIT + IND + STB <- SURV, data = LF)

               inclN   RoN   covN  

---------------------------------- 

1         LIT  0.991  0.509  0.643 

2         IND  0.669  0.786  0.684 

3         STB  0.920  0.680  0.707 

4  expression  1.000  0.367  0.587 

---------------------------------- 

For the negation of the outcome you can use a tilde, as per Ingo’s suggestion:

pof(LIT + IND + STB <- ~SURV, data = LF)

               inclN   RoN   covN  

---------------------------------- 

1         LIT  0.573  0.387  0.414 

2         IND  0.354  0.661  0.403 

3         STB  0.467  0.509  0.399 

4  expression  0.711  0.309  0.464 

---------------------------------- 

I hope this helps,

Adrian