Imagine looking at the 5 deadliest plane crashes in 1970 and comparing them to the 5 deadliest plane crashes in 2010. You find that more people died in these crashes in 2010 than 1970. You wouldn't conclude that air travel is now less safe than in the 70s, would you?

There was a lot more air travel in 2010 compared to 1970. The 5 deadliest flights makes up a much smaller portion of all flights in 2010 than it did in 1970. It's a misleading comparison. You should instead look at the number of deaths per passenger or per passenger mile.

You have to think of the same point when looking at the "top 350 companies in the USA in 1965 vs 2017".

Are the companies the same size in 2017? Do CEOs have the same amount of responsibility in 2017? Is the complexity of managing the company the same in 2017 as it was in 1965? What percentage of all companies do these top 350 make up?

These are all factors you might want to take into account when looking at a specific number of top companies.

I'd be curious to know what the numbers would look like if they looked at the top 1% or top 0.1% of companies in the US instead. Or perhaps even the average and median companies. My guess is that those stats wouldn't be politically as useful. I imagine that you'd still see a growing disparity, but a smaller one.