Hello,
I'm looking for a way to fit a Signal Detection Model to a n*n confusion
matrix of data like the one provided below (This hypothetical data stems
from a n-forced-choice experiment).
I'm looking for a procedure that can separate the perception of the
signal from the criterion that the participants use.
Does anybody know of a simple procedure, and maybe an implementation of
this (a program?).
Many Thanks,
Dieter
Example data:
Signal: One Two Three Four
Response: One 0.189555 0.273401 0.444942 0.197321
Two 0.99889 0.993752 0.721348 0.341097
Three 0.996513 0.614513 0.481794 0.865671
Four 0.918038 0.133882 0.539384 0.337964
|