The median age doesn't contradict the above model at all. What it tells you is that of the developers employed in the game industry at any given time, half of them are under 31. But it doesn't say a thing about the churn below that age.

You can have a bunch of people who jump in at 23, get super exploited, and burn out in two years balance against one single person over the median age, so long as that stream of younger people is serial. The median is unchanged, but a lot more people leave the industry feeling exploited and abused by it than remain.

That wouldn't be possible without a corresponding behavior in those above the median age as well. If what you said was true and there's a disproportionate group of people below the median age leaving the industry without an equal number of people above the median age leaving, then you'd get a constantly rising median age over time and at this point that age would have be considerably older than 31 years.

I am not disputing that the game industry has a lot of churn. I'm disputing that the game industry has substantially more churn among younger people than older people and that the churn is significantly different from the rest of the software development industry.

It's not so much that young people are uniquely exploited and leave after a couple of years, it's that the number of software developers doubles every 5 years so that there is a huge imbalance of young people in the industry in general.

Stack Overflow is not exactly validated market research. Comparing two completely different sources of statistical data with different collection methodologies is not evidence.

It's not proof, but it's absolutely evidence and it's enough evidence to place the burden that the video game industry consists mostly of young people back on the person making the claim.

If there is such evidence, by all means produce it.