If an "M" popped on the screen, the respondent had 500 milliseconds (a half second) to press a key on the keyboard before him or her; if a "W" appeared, the person was told to do nothing.
I think that is an invalid method of testing. he better way would have been to choose between two buttons, imo.