The first aim of the present study was to assess the effect of stereotype threat on recall of a visual event involving cars by femalewitnesses. It was expected that stereotyped women (by the interviewer) would provide a poorer description of the cars involvedthan non-stereotyped women. The second aim was to test if a modified cognitive interview (MCI) could reduce this negativeeffect. We expected that MCI reduce the harmful effect of stereotype threat on women’s recall in comparison with a controlinterview (CI). Sixty-nine women watched a videotape of a criminal event involving cars. Two days later, they had to recall theto-be-remembered event and were given either a neutral or a threatening (= stereotyping) instruction. They were then interviewedusing a CI or an MCI. Results showed that threatened women were less accurate and made more errors and confabulations thannon-threatened women. They also produced more incorrect details about the cars involved. Moreover, the MCI elicited morecorrect details but without impairing accuracy for both threatened and non-threatened women. Our study supports the existingevidence that women’s testimony may be affected by the social context and highlights the need to use the MCI during judicialinterviews.