The issue is that the sample population is small and the effects of bias is often subtle so the suggestion that the image is "inappropriate" by overwhelm what my already be a "weak signal."
I don't think the goal was to determine of one image is considered more or less inappropriate it was to determine bias. So why not - "Our university is developing promotional material to attract more women. Would you use this image?" Or even better, show both images and then say "Which image would you choose for promotional material (1) image a, (2) image b, (3) neither images."
Bias is something rarely acknowledged and (thankfully) discouraged. As a result it is difficult to study. But there is a large body of rigorous research that consistently shows that even the most enlightened of us have implicit biases that effect our decision and actions on a daily basis. A very quick google came up with this literature review (http://kirwaninstitute.osu.edu/wp-content/uploads/2014/03/2014-implicit-bias.pdf). I haven't read and I'm not familiar with the Kirwan Institue but it looks like a good survey to my quick skim.