Much research shows that students take account of their perceived comparative advantage in mathematics relative to verbal skills when choosing college majors and career tracks. There is also evidence for an important role for comparative advantage in explaining the gender gap in college STEM major choice. For these reasons, it is important to understand why student perceptions of comparative advantage may differ from true comparative advantage as determined by actual abilities. One plausible pathway is through teachers. We study gender differences in teacher evaluations of student comparative advantage relative to comparative advantage as measured by test scores. We show that findings are very sensitive to the methods used; commonly used methods are not equivalent and can give different results as they target different estimands. Using two recent UK cohort surveys, we show that these conceptual issues matter in practice when we evaluate whether teachers are likely to over-estimate female comparative advantage in English relative to mathematics. Our preferred estimates provide no evidence that teachers exaggerate the female advantage in English relative to mathematics and generally suggest the opposite. We conclude that differences in teacher judgement by gender do not provide another reason for the gender gap in STEM.