Are there results on the null hypothesis of the "surgeon operating on my son" question?

Question

There is a common question used to demonstrate the power of implicit assumptions about gender roles in our society.

Including it here for reference: A man and his son had a terrible car accident and were rushed to the hospital. The man died on the way, but the son was still barely alive. When they arrived, a surgeon was called in to operate. Upon seeing the young boy, the surgeon said, “I can’t operate – this is my son.”

Most people have difficulty thinking of the possibility that the surgeon is the son's Mother.

This riddle strikes me as having elements of the priming experiments whereby subjects primed with certain words perform better/worse in word-stem completion tests, intermixed with the social expectations of our culture. Here the "priming" would be using the male "father", "boy", and "son" throughout the question but asking for a female "mother" as the answer.

I was trying to find results on some sort of "control group" experiment to try and isolate the societal expectations element. Something like subjects being asked the same riddle but with a mother and daughter in a crash and the father a surgeon, or where it's the father and daughter in a car crash.

The only alternative I've seen is about "mother and daughter in crash, father is nurse" which is again combining the two effects.

N.b. I'm expect the priming effect is probably small, I just wanted to know by how much.

Answer 1

The implicit-association test is a larger example of the same phenomenon. Words are paired that are congruent or incongruent with societal expectations; incongruent pairs take most people longer to match.

A caution on this test it is often used to measure bias in individuals and for some as a surrogate for bigotry. However, it does not account for intentional methods to avoid biased action, it only measures the "automatic", involuntary biases imposed by exposure to culture. It should be instructive to test-takers how the results might shape their behaviors in every day life, but shouldn't be taken as a measure of "how sexist" or "how racist" someone is. It is also not apparent that individual results are informative; population-level results trend with population-level biases, but individual performance on the test is not a valid way to find "who is more biased" or anything like that.