I have 3 subjects. Each subject completed the same experiment with 4 randomly interspersed stimulus types. For each stimulus type the subject completed 120 trials on a 2AFC (2-alternative forced choice) task. From these I can calculate a percent correct.

I would like to know if the overall percent correct is statistically different between the four stimulus types, but I am not sure what statistical test to do or how to combine the three subjects.

I am wondering if I need to do something with d' or t-tests?


You're looking for the Pearson Chi-squared test. It generalizes a two-sample t test to more than two samples.

There are a number of good presentations on the Internet. Here are two with particularly clear and non-overly-technical explanations for applied work:

http://math.hws.edu/javamath/ryan/ChiSquare.html http://courses.washington.edu/dphs568/course/day12.htm


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.