Combining empirical and computational studies in social cognition: Illustration with implicit gender stereotypes

International audienceThe general aim is to provide an illustration of how empirical and computational research can be combined to understand the formation of a much-studied construct in social cognition: (implicit) stereotypes. Specifically, we focus on implicit gender-math associations - stronger male-math than female-math associations - and on the central role of the self in the construction of (counter)-stereotypic associations. Indeed, although women are still underrepresented in Science, Technology, Engineering, and Mathematics (STEM), some manage to remain in counterstereotypic STEM tracks. We present a line of research investigating implicit gender-STEM stereotypes among counterstereotypic STEM women, STEM men, non-STEM women and non-STEM men. Study 1 examined whether implicit gender-math stereotypes, as measured by the Implicit Association Test (IAT), would be weaker for STEM women than for the other groups. Study 2 investigated implicit gender-math stereotypes with an IAT-adapted MouseTracking technique, hence allowing recording participants’ computer mouse coordinates during the whole decision making process (instead of solely the final decision as in typical reaction time based IATs). In Study 3, a computational study, data were simulated by a connectionist model parameterized to reflect intergroup differences and to integrate, for the first time, strengths from recent developments in embodied models of decision-making with those of social categorization. Together, findings consistently demonstrate weaker implicit gender stereotypes among STEM women and emphasize the key role of the self in the construction of counterstereotypical associations, hence providing support for a core assumption of IAT-related theorization