sufficiency analysis and truth tables

[Alexander Strelkov]

Dear community members, good afternoon

    Another query on my side.

I am creating a truthtable for the presence of the outcome (myTTy1) using the attached calibrated data.

I run a conservative solution and save the results

solc3<-minimize(input=TTy1,details = TRUE)

stargazerSol(results = solc3,outcome = "WIN",type = "text",out = "solc3.txt")

Question 1: However, I can't see which cases are applicable to which sufficiency pathway.

Tried adding cases=rownames(calibdata) & cases=calibdata$case but nothing works. I see case numbers in the console, yet in a very convoluted manner and never in the output. What am I doing wrong?

 I am trying to move to a parsimonious and intermediate solution:

solp<-minimize(TTy1,details = TRUE,include = "?",row.dom = TRUE)

Yet I get the following message:

Error in minimize(TTy1, details = TRUE, include = "?", row.dom = TRUE) : 'R_Realloc' could not re-allocate memory (0 bytes)

Question 2: Is it because my truthtable has no '?" in the OUT rows? So I can't actually run a parsimonious or an intermediate solution?

Question 3: If my suspicion about running parsimonious/intermediate solution is correct, then I am "stuck" with reporting 11 pathways, each representing 3-5 cases. Is there any other way to "compress" it?

Your advice is highly appreciated!

Best wishes, 

Alexander

[Breno A. H. Marisguia]

Hey!

So. First, about the parsimonious and intermediate error: installing QCA with `install.packages("QCA", repos = "https://dusadrian.r-universe.dev")` solved the issue for me.

Now, for the questions. Your "rownames" intuition is correct, but try it this way: 

```

library(tibble)

calibdata <- calibdata |>
  column_to_rownames(var = "Case")

```

This will probably solve your cases problems. Then run QCA as normal.

I need to warn you: you have way, way, way too many conditions. This is not good practice. With 12 conditions, that's 2^12 possible configurations (4096). This will generate a substantial amount of logical reminders. Also, as you'll notice, your solutions are all over the place.

Schneider & Wagemann "11.1.2 The choice of the conditions and the outcome"; Mello "Condition Selection"; and Oana, Schneider & Thomann "2.4 Combining Sets" will surely help you.

Breno

--
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 visit https://groups.google.com/d/msgid/qcawithr/ca167df1-adde-4d55-b625-8c230f7f0f13n%40googlegroups.com.

[Kalihputro Fachriansyah]

Hello All. Sorry for jumping in the conversation. 

A greater number of conditions yields larger logical reminders. And when minimising truth table rows for intermediate (or other) solutions, there might be more than one model solutions (model ambiguity). My questions are: how to show all of the generated models, for example to report in word docx? Next, which model is best for further analysis and interpretation? Is it the highest consistency ones?

Thank you,

Best,

Kalihputro

On 29 May 2026, at 15:26, Breno A. H. Marisguia <brenoa...@gmail.com> wrote:



To view this discussion visit https://groups.google.com/d/msgid/qcawithr/CAL62YhDuDqX2HjXUEPpch22PQ3Ye9wRWzJCQwm8WP%2BVJwAB%3DeA%40mail.gmail.com.

[Ingo Rohlfing]

Dear all,

the naming of rownames should be doable with base R. When reading the data into R.

calibdata <- read.csv("calibdata.csv", row.names = "Case")

By assigning the case column as rownames.

rownames(calibdata) <- calibdata$Case

Best

Ingo

[Ingo Rohlfing]

Dear Kalihputro,

opinions on these issues certainly differ. My take is:

- Show all models in an appendix. Each model is a logically correct summary of the truth table, so all should be shown. 

- I think it is fine to focus on one model or a small subset of models when you make transparent how these models are selected. It could be because they have the highest parameters of fit, or because they are theoretically interesting, or substantively interesting, or something else.

Regards

Ingo

[Breno Marisguia]

100% what Professor Rohlfing said. But I would add that it’s always a good idea to go back to your theory and assess whether adding a ton of conditions is really the way to go. QCA does not have to reproduce every possible step or capture every possible observable implication; the minutiae can be saved for a later, more in-depth post-QCA analytical move. Also, check the texts I recommended to Alexander. They're really helpful!

Breno

[Alexander Strelkov]

Dear Breno, Kalihputro, Ingo, 

    Many thanks for your advice, will try implementing.

I do agree that 12 conditions is a lot but my challenge comes from struggling to find a comprehensive script for mvQCA (hence my query in the previous conversation I posted) - doing that would reduce 12 conditions to 7. Taking a two-step approach would be an option but in my case there is no theoretical ground to identify remote or proximate conditions. Will try it with crisp and then see what I could do with mv.

Your advice is very much appreciated!

Best wishes, 

Alexander