If subjects do not favor any of the 4 alternatives a priori, the proportion of correct responses on your 4AFC task is a measure of sensitivity unaffected by response bias (see Green & Swets, pp. 45–49, 1966).
You can find a complete description of calculations in Stanislaw, H and Todorov, N “Calculation of signal detection theory measures” Behavior Research Methods, Instruments, & Computers 1999, 31 (1), 137-149
Hi, you might want to investigate the Az measure proposed by Dorfman and Alf.
Dorfman, D. D., & Alf, E. (1969). Maximum likelihood estimation of parameters of signal detection theory and determination of confidence intervals — rating method data. Journal of Mathematical Psychology, 6, 487-496.
Depends on whether you want to convert an accuracy measure into a d' or analyze it at the level of individual trial. For the former, just consult http://link.springer.com/article/10.3758%2FBF03208311?LI=true. For the latter, you might try http://www.columbia.edu/~ld208/jmp12.pdf - I've never applied the latter technique.
Note that if you want to estimate both d' and bias, there is a method presented in DeCarlo doi: 10.1016/j.jmp.2012.02.004
- Badges
- Science topic