http://www-psych.stanford.edu/~lera/psych115s/notes/signal/

The method I used is z-score "hits minus false alarms"

When I run an ANOVA on simple proportional accuracy scores I find no interaction between condition A (two levels) and condition B (two levels), but when I run an ANOVA on the same data but using d prime indexes instead, I do find an interaction between condition A and B.

Is this because the d prime method has greater

**power**to detect a small difference between the interacting conditions that simple proportional accuracy does not? Would a increasing sample size (ie more power) therefore be likely to find the same result for simple accuracy eventually?