This study investigates the effect of correctional instructions when detecting lies about relational aggression. Based on models from the field of social psychology, we predict that correctional instruction will lead to a less pronounced lie bias and to more accurate lie detection. Seventy-five teachers received videotapes of students' true denial messages (e.g., students were innocent) and untrue denial messages (students were involved in relational aggression). A random half of the teachers were either given no further instructions or were told not to use stereotypical non-verbal cues when forming judgements (correctional instruction condition). As predicted, teachers in the correctional instruction condition were slightly better at detecting true than invented stories. (Contains 1 table and 3 notes.)