How to deal with the situation when false alarm rate = 0? about psycho.r HOT 2 CLOSED

Hi @Youguang, It is an expected feature of dprime. One way to deal is to use non-parametric indices (e.g., aprime). Another way (which I personally am not fond of, although I know people are using it) is to adjust the numbers to not have FAs of 0 and HRs of 1 by tweaking the data.

E.g.: “The identification judgments were used to estimate perceptual distances by calculating the z transform of the mean identification percentage for each token and then taking the absolute value of the difference for each pair of tokens. The z transform reaches infinity when percentages equal 0 or 100, so tokens with 0% identifications were assigned values of 1% and tokens with 100% identifications were assigned values of 99% (Macmillan and Creelman, 1991).” here

No other alternatives come to my mind, maybe people have come up with other ideas on the internet? Let us know if you find anything interesting ☺️

from psycho.r.

Thanks for your reply @DominiqueMakowski.

Here are some references.
And dprime.R provided one method, the third approach, to adjust dprime.

from psycho.r.