Most of this is a copy from my meta post here at Open Source:
I recently made a review that you can find here (Note that the "Emacs" spelling changes were made after I had done the review). Also note that the author subsequently approved of the edit.
The edit made the sole change: open source to free software. If you're wondering why, I rejected the edit, here's my stance:
I rejected the edit on the grounds that as the reviewer of posts, I should strive to make sure that edits fix the issues with the post: concerning formatting, quoting, attributions, grammar, punctuation... making the post look pretty. I'm not there to verify information - I believe that is the responsibility of the author. If information is incorrect, that's fine. If you're wondering how we deal with that, we have a largely sophisticated system: voting. If something is wrong, down vote. In essence, if you have something that is right, leave a comment with a source to back up your claim, and let the author make that edit. Otherwise, you could be deviating from the author's intent in that you may be changing the meaning of something the author has to > say.
Thinking that this applied to all sites, I've brought it here in the ultimate meta of metas. On Open Source, it seems that if you're not going to go out and find a link yourself to verify the terms that are being changed, or the new information that is being added, then you shouldn't be reviewing at all.
I also want to make a note that I raised that (and this) meta post out of concern for other reviewers, many of which may not be necessarily completely knowledgable, and to help set a guideline for whether changing small words/terms should be allowed, or rejected.
I just want to ask: If I were to extend this thinking to the other communities in which I participate, should I do this work as a reviewer? And is this the expected behaviour of reviewers in general?