How do these people know it makes a difference? Do they measure their productivity? Is there any scientific research on this?

Similarly, a top flight programmer who has mastered emacs or vim or Sublime Text won't be operating at a conscious level but will be controlling the environment in front of them at a far higher level, especially if the conceptual model of the text buffer is the same no matter what they're doing with it.

An extreme comparison would be to sit two people of the same ability down in front of the Linux kernel source tree, one armed with emacs, cscope, syntax highlighting, org-mode and cc-mode and the other with nano and see who is faster at some non-trivial task.

Sure you don't. You can feel the difference. The problem occurs when people treat their editors as an extension of themselves and can't tell the difference between "not liking you" and "not liking your editor", eventually getting defensive. That's a "cognitive level" I don't want to get myself into.

I've interviewed people who were exposed to Nano at school, and have never tried out any other editor. They are missing out on productivity improvements available in many other editors. More middling of the scale I've also interviewed someone who used Eclipse as their first editor and has proceeded to use it ever since despite not working on Java projects any more. That isn't bad per se, but I rank someone higher if they at least try out others every now and then.

The fact that people feel strongly about something does not mean it is important. Lots of people have very strong feelings about the New Jersey Devils or the Miami Heat. That doesn't mean that sports or teams matter.

I'm not saying your conclusion is wrong, but your comment does not constitute a valid argument in favor of that conclusion.

Which would probably be important if you were interviewing them to be local sportscasters in those markets, no?

I meant important to them individually, not the whole world