Much, much smaller population. From OP: " So here is a more comprehensive data bank of 982 individuals across 14 publications." It's that, vs. the millions of

Much, much smaller population. From OP: " So here is a more comprehensive data bank of 982 individuals across 14 publications." It's that, vs. the millions of employees in big tech. So first you're dealing with the fact that smaller samples are going to have higher variance. Beyond that, though, since diversity has become important to businesses, tech employment has grown much faster than tech journalism. If you're actively hiring a lot of people, you can shift the mix by bringing in a larger proportion of people that are different than the initial population that you started with. If you're not growing, then that means people have to leave the original population, either by firing or quitting, and then be replaced. The latter takes longer than the former, particularly when you're comparing big tech, which has a huge amount of turnover at a given company, and tech journalism, which has much less.