In this experiment, it was examined whether it is possible to detect what was presented subliminally when feedback (correct or incorrect) was given. Sixteen healthy right-handed undergraduate and graduate students were participated as subjects. Three (facial emotions: angry, happy, neutral) x three (feedback: available, no, random) factors were located within a subject. After a mask stimulus (neutral face) was presented, subjects were required to answer what was presented as a prime face by pressing one of three buttons. After subject's response, feedback was given as "correct" or "wrong" in feedback-available condition and "00000" in no feedback condition. In random feedback condition, the probability of presenting "correct" and "incorrect" was fixed as 3:7. ANOVA was performed on result of RT, hit rate, and d'. As for RT, the main effect of feedback was significant. As for hit rate, the interaction between facial expressions and feedback conditions was significant. Multiple comparison corrected by LSD on angry face revealed marginal significance between feedback-available and no-feedback, and significant difference between random-feedback and no-feedback. On neutral face, significant difference was found between feedback-available and no-feedback, and feedback-available and random feedback. In angry face condition, hit rate was lower than other conditions, and in neutral face condition, feedback-available condition showed lower hit rate than other conditions. About d', significant difference was not found among conditions. From above, it could be said that in no-feedback condition primed stimuli tend to be judged as neutral, but as long as feedback was given, subject's judgments shifted to answer that the prime was angry face.
View full abstract