According to wikipedia, there's a little research and it's been positive, but it's not the sort of research I find persuasive. I do have mountains of anecdata from myself and several friends whose opinions I trust more than my own. PM me if you want a pdf of the book.

Thirded. The most helpful part for me was internalising the idea that even annoying/angry/etc outbursts are the result of people trying to get their needs met. It may not be a need I agree with, but it gives me better intuition for what reaction may be most effective.

I would like to offer further anecdotal evidence that NVC techniques are useful for understanding your own and other people's feelings and feeling empathy toward them.

When it comes to research about paradigms like that it's hard to evaluate them. If you look at nonviolent communication and set up your experiment well enough I think you will definitely find effects.

The real question isn't whether the framework does something but whether it's useful. That in turn depends on your goals.

Whether a framework helps you to successfully communicate depends a lot on cultural background of the people with whom you are interacting.

If you engage in NVC, some people with a strong sense of competition might see you as week. If you would consistentely engage in NVC in your communcation on LessWrong, you might be seen as a weird outsider.

You would need an awful lot of studies to be certain about the particular tradeoff in using NVC for a particular real world situation.

I don't know of many studies that compare whether Windows is better than Linux or whether VIM is better than Emacs. Communication paradigms are similar they are complex and difficult to compare.

I found NVC is very intuitively compelling, have personal anecdotal evidence that it works (though not independent of ESRogs, we go to the same class).