My point being that I don't actually believe that people need to be the same gender (or skin color or etc) to be relatable as role models.
I think making role models less demographically binned would be an important and positive step.
.
"Female students have generally related more to my female colleagues"
Do you have any reason to believe that this isn't because every time they're getting started someone carefully creates that situation, as is being created here?
Measurements are frequently measurements of effects, rather than of causes.
.
"Which is perfectly fine"
How can you be sure that refusing demographic cross pollination by intent is not a negative thing?
It has been proven that our society is full of gender and racial biases, and to fight them we need to accept this fact so we can have a sensible plan of attack. Apparently we can't just make those biases disappear by wishing them away. It is easier to dismantle them by first fighting the symptoms. If you just tell people to ignore their biases, you will ask them to ignore their personal day-to-day experience, and that's asking a lot. It's hard for people to believe one thing if what they see around them goes against that belief. To change beliefs we must first change the experience.
All for what? a social experiment to fight biases with "positive discrimination"? The hope removing half of every possible role model is worth getting a female role model over a male role model?
In the end its the parent that decide and it might be worth if there is already plenty of male role models, so long one is aware of the costs.
Over the past 8 years, I've taught programming to people from ages 6 to 50, of all genders and cultural backgrounds. Female students have generally related more to my female colleagues, just like black students generally relate more to my black colleagues. Which is perfectly fine, and why you want a diverse set of educators in the first place.