Yes. At the time I'm commenting, 1 out of 2 questions on the "women's" question is about the gendered nature of the question itself (50%). And 2 out of 11 on the "mens" (18%). The dataset has already been "polluted" as you describe, by the design of the questions.

Thanks. Off to shop for crab meat.

Then comments calling out the unnecessary, ham-handed gendering should be expected data points.