I'm studying the endowment effect through the exchange paradigm. Participants received an object A or B. Then the experimenter asks the participants if they accept to trade their object with the alternative object.
I would like to compare the proportion of participants who refuse to trade A with B, with the proportion of participants who accept to trade B with A. In other words, in a contingency table, I want to compare the proportion of failure of Group 1 with the proportion of success of Group 2. Note that I have a small sample (n < 30). I don't know tests that allow a direct cross-tabulation comparison like this. I have tried many alternatives (e.g. Fisher's exact test) but they do not allow me to answer this question exactly...
Would you have any ideas or suggestions?
