Does any of these studies take into account the amount of work/time needed to end up with more or less the same robustness level of the result?
For illustration, assume hypothetically that it would take ten times as much work to get a given program written in JavaScript to the same robustness level as the “same” program in TypeScript. Would these studies still claim TypeScript wouldn’t help on robustness (reducing bugs) compared to JavaScript?

I have the same impression as Alex Chan, I feel much more productive and “safe” with a modern statically typed language than with a dynamic one. I find types a very valuable tool for helping my thinking.