Accuracy and errors in judges' attempts to differentiate facial expres
sions that displayed genuine pain, no pain or were dissimulated (i.e.,
masked and exaggerated) were examined. Judges were informed that misr
epresentations in the facial expressions were present and were asked t
o rate their confidence in classifying these expressions. Detailed, ob
jective coding of the patients' facial reactions (e.g., brow lowering,
mouth opening) were related to judges' decisions. Judges' classificat
ion decisions were better than chance, but there were many errors. Ext
reme expressions (i.e., no pain, and exaggerated expressions) were ide
ntified more accurately than genuine and masked expressions. Judges le
vel of confidence was consistent with their level of accuracy. Judges
consistently used rules of thumb based on specific facial cues when ma
king judgments. Certain cues were effectively discriminative. Systemat
ic training in the use of specific cues or the use of articulated deci
sion rules may be helpful in improving judges' accuracy.