The present research aims at testing the effects of the mixed debate in a classroom situation on reducing the stereotype threat according to which girls perform less well than boys in mathematics. Our work is based on studies conducted on stereotype threat (Steele and Aronson in J Person Soc Psychol 69:797-811. 10.1037//0022-3514.69.5.797, 1995), as well as those aiming to reduce the impact of stereotype threat on stereotyped populations (Huguet and Régner in J Educ Psychol 99(3):545-560. 10.1037/0022-0663.99.3.545, 2007; Marx and Roman in Personal Soc Psychol Bull 28:1185-1197. 10.1177/01461672022812004, 2002; Spencer et al. in J Exp Soc Psychol 35:4-28. 10.1006/jesp.1998.1373, 1999). We will more particularly study Smith and Postmes' (Br J Soc Psychol 50(1):74-98. 10.1348/014466610X500340, 2010) work on small group discussions used to reduce the influence of the threat. Results indicate that having a debate as such reduces the impact of the stereotype threat in girls. Possible interpretations of the results and future research perspectives will be put forward.