Unless I'm feeling the disease (aka, wasting too much time here; must apply now to be delving into meta) I select from the options:
- Don't upvote.
- Let someone who wants/needs the two point credit (or just is more deeply diseased) edit the thing. If I see a pending edit I do try to look at it.
- Too Painful; Didn't Read.
If, as purported in the SE model, good answers get upvoted and bad answers don't (I try to hold downvotes to WRONG answers, for the most part,) this leaves the door open for someone else to write a coherent answer, or for someone to edit the answer after which it can be upvoted. While I may know it's wrong, and how to put it right, it's not my job, and choosing not to do a thing that's on a basis of unrewarded volunteer activity is not something you should feel guilt over, IMHO.
While I have spent some (diseased) time plowing though the review queues in the past, I rarely bother of late; it doesn't bother me, and it reduces the time-suck this place can be. If that reviewer badge means a lot to you, be my guest - if you're newly empowered and enjoying the novelty, be my guest.
I am also hesitant to dive too heavily into editing stuff where I may be laying what I think the person meant onto what they scribbled, to the point that an edit may change the meaning. I've certainly encountered a few cases where a review of the edit history showed that this had already happened with someone else in the editing seat, either changing the meaning, or making a lucid, comprehensible original question/answer incomprehensible. Teasing out the meaning of something poorly written with no direct knowledge of what was actually meant can be tricky, especially if the writing is really poor.
I will now resist the urge to remove all punctuation and formatting from this post, and add some deliberate misspellings and correctly spelled wrong words to the ones that got past me anyway.