Open tonofshell opened 5 years ago
It turns out that if we get a large amount of data from each participant, we get quite reliable results with these numbers of subjects. The number of subjects required really depends on the specific question that is being asked, and the variability across people in the answer to that question. In this case, almost all people showed qualitatively similar responses, and so very large numbers of subjects were not required for the questions we were asking.
With 16 usable participants for experiment 1 and 23 for experiment 2, there was a very small sample size in your study. What unique challenges did this pose in your analysis and how did you overcome them?