07.08.2017, 16:56
Hello Satoshi Sasamura,
thank you for your question! Unfortunately I couldn't help you with the question about Mixed Methods, but with this one I feel more confident.
Let us say you have many respondents, and some of them (it seems on first glance) talk very often about a certain topic (let's say icecream) in a certain way (very positive). Now you want to know if there is a correlation between talking about icecream positively and the place where the interview was conducted (maybe being in Chile makes you see icecream more positive then.. being in Iceland). So how to proceed?
First, you retrieve all the overlapping codes (Icecream/very positive) by selecting "Analysis > Complex Retrieval" and then "Overlapping". (You can also click on a square in the Code-Relations-Browser, that will do the same). Then you code the retrieval with a new code ("Icecream seen very positively). Then you right-click on this new code, and select "transform into document variable". That will give you the number of occurences of "Icecream seen positively" for every single document/respondent. Now you can use stats to check for correlations between this number and other variables like country/temperature etc.
I hope this helps! If not, or if I missed the point, please don't hesitate to write to us again.
Kind regards,
Andreas