You are simply asserting that "sexist stereotypes contribute to more skewed representation", which is no better than asserting that wet roads cause rain.
First, there is nowadays strong evidence that people generally do not hold explicitly sexist views. So the fact that people know that there are, for example, fewer female software engineers does not mean that they think women are incapable of being software engineers. Knowing that there are fewer female software engineers and therefore a software engineer you encounter is more likely to be male is not sexist, it is simple knowledge of the facts.
So what about "implicit bias"? Well, the poster child for implicit biases, the Implicit Association Test, is actually highly dubious, as results (a) are highly inconsistent [the same person will get wildly different scores on different instances of the test] (b) quite weak and a bit different than reported and (c) do not predict biased behaviour.
What do I mean with "a bit different than reported"? IIRC, it was reported that the IAT showed men being biased against women. Which was a bit of mischaracterisation: for male subjects, there was no difference in "bias" for/against women or men, but women had a positive bias towards other women. So there was a difference, but not quite what was reported. (This in-group bias was also discovered in a different study).
Finally, what about the backwards causality? Let's take Stereotype Threat. Results were always weak and somewhat questionable, and basically crumbled in the replication crisis, the effect disappearing when controlling for publication bias.
You are simply asserting that "sexist stereotypes contribute to more skewed representation", which is no better than asserting that wet roads cause rain.
First, there is nowadays strong evidence that people generally do not hold explicitly sexist views. So the fact that people know that there are, for example, fewer female software engineers does not mean that they think women are incapable of being software engineers. Knowing that there are fewer female software engineers and therefore a software engineer you encounter is more likely to be male is not sexist, it is simple knowledge of the facts.
So what about "implicit bias"? Well, the poster child for implicit biases, the Implicit Association Test, is actually highly dubious, as results (a) are highly inconsistent [the same person will get wildly different scores on different instances of the test] (b) quite weak and a bit different than reported and (c) do not predict biased behaviour.
What do I mean with "a bit different than reported"? IIRC, it was reported that the IAT showed men being biased against women. Which was a bit of mischaracterisation: for male subjects, there was no difference in "bias" for/against women or men, but women had a positive bias towards other women. So there was a difference, but not quite what was reported. (This in-group bias was also discovered in a different study).
Finally, what about the backwards causality? Let's take Stereotype Threat. Results were always weak and somewhat questionable, and basically crumbled in the replication crisis, the effect disappearing when controlling for publication bias.
https://replicationindex.com/2017/04/07/hidden-figures-repli...