Negated outcome results - intermediate and enhanced intermediate solutions

[Charlie]

Hi,

I wonder whether you can help.

I am going through enhanced standard analysis on R for the negated outcome (~REN). I am getting these results for the intermediate solution:

When I get the enhanced intermediate solution, the results are as follows:

As you can see, different EC rows are used for each solution:

So my question is: how can I interpret those results? Does it mean I have missed a step that could have ensured identical uses of ECs across the two solutions? Does it also mean I need to discard one solution in favour of the other? Any suggestions on how best to interpret these results would be welcomed!

Thanks a lot in advance for your help!

All the best,

Charlie

[Adrian Dușa]

Hi Charlie,

I was wondering, why would this an "enhanced" standard analysis?

When you're only using directional expectations, this is standard analysis proper. It could be, depending on how you created the truth table(s) but absent the relevant script commands I cannot say for sure.

What I see is not "different" EC rows, but rather one that is common for both presence and absence of the outcome (for instance remainder number 17), which should suggest something like a simultaneous subset relations type of problem (one that should lead to an enhanced standard analysis, should that remainder be excluded from the minimization).

Otherwise, the use of the function identical() is not exactly useful, in the sense that it will always be FALSE if comparing 17 to 17 and 27:

> identical(17, c(17, 27)) # FALSE

I would rather use intersect(), as in:

> intersect(17, c(17, 27)) # 17

Hope this helps,

Adrian

[Ingo Rohlfing]

Dear Charlie,

the results ware what I would expect when comparing the results of a standard analysis and an enhanced standard analysis. With ESA, you have excluded one remainder from being used as a simplifying assumption. (row 27, it seems) This explains why the ESA solution is a bit more complex than the SA solution. They would only be identical if the input into ESA, whatever this is, would not concern any remainder in your data. 

You have to decide whether you think the better analysis is the standard analysis or enhanced standard analysis. Is it more plausible - theoretically, empirically, methodologically - to exclude remainders from the minimization process or not? Once this has been decided, you should use the resulting QCA solution for theoretical discussion.

Regards

Ingo