Previous research has suggested that changing the percentage of the course grade associated with exam grades in STEM courses can change the gender gap in the course. It has also been shown that assessments with the highest stakes have the lowest (relative) scores for female students. Previous research by the authors has shown that the implementation of retake exams can eliminate the gender gap in introductory physics courses. This paper explores several different hypotheses for why retake exams are associated with a zeroed gender gap. Two independent measurements comparing exams with different stakes are used in support of the argument that the entire gender gap on introductory physics exams may be due to the stakes associated with those exams. In other words, these data support the idea that a gender grade gap on exams is not measuring a gender difference in the physics knowledge or physics ability of these students. Implications suggest that instructors should choose lower stakes assessment options if they are interested in exam measurements that are not influenced by differences in students' performance related to exam stakes.

13 pages, 3 figures