That is a subjective conclusion to draw on the data. Wouldn't associating an aggressive face with an object just as likely drive a more aggressive response when the subjects saw it again?