Bristol_Jonesey
I very rarely bother with Clip Gain but instead use Track Volume envelopes.

 
Just a few minutes ago I used some tracks' volume envelopes to get rid of a rather abrupt start to a phrase sung by three vocals.  It sounded like some sort of recording error, or maybe Melodyne did it.  I silenced the volume right before the phrase then ramped it back up quickly, just enough to lengthen the attacks of the note.  Worked very well.
 

If I can hear my dogs in between vocal phrases, I cut.  I have never automated gain in this scenario though. I am definitely going to give that a shot.
 
Sincerely,
Guy with no recording booth

Karyn
 
I tend to automate the clip gain rather than output volume to ensure no noise is fed to pre-fader FX...

I've always used volume envelopes but does this make more sense?
 
My general approach, as a vocalist who does all my own vocals/harmonies, at some point in the song when I'm pretty satisfied, I quickly go thru and add double nodes between parts and then just pull the volumes down ctrl/mouse to remove all or most of the background noise like my chickens squawking in the yard, birds, or whatever...
 
it would be nice if I could select all the harmony parts and add multiple track nodes at the same time but I haven't figured out how to do that/doesn't seem to be possible.

It's a really tough topic and what applies to one, doesn't to someone else, and varies between what type of music your creating, and even the end result, how it's mastered.
 
I've found with some songs I've gone overboard on silencing the background, it's changed the whole feel, and not for the better. Other tracks are woeful with background noise left in. But in general, I manually do all the vocal tracks. Complete silence where required such as a lead break, and usually somewhere between 6 and 18db down between vocal lines, splitting and fading in and out if necessary. I always leave traces of everything that's related to 'vocal' noise. I've tried automation, but get much better results going painstakingly slow between each vocal line. Takes ages but it pays off. I tend to use automation on guitars and gates on the bass, but the more work I do, the better I get at the recording stage, getting rid of most unwanted stuff before it hits the disk.
 
Ears before eyes as someone else said. Same with Melodyne. Very easy to group notes, bars, or entire clips, for processing. But you cannot beat manually listening to, and modifying accordingly, each and every note.
 
Depends on how much time you have I guess, and what level you are prepared to settle for.
 
regards

holy crap!! everyone had such good suggestions, it's also, i think, one of those times where everyone is correct. different situations require different methods. what i've learned most from what everyone has said is to "listen". my music is pretty noisy, what i mean is, most times you're not going to hear a car alarm between lines or phrases. 

So in your case there's a good argument for not doing too much with the vocal track(s).
 
Be careful though, the more mush you DO leave intact is going to build up, so if there's mush on your vocals, rhythm guitars, bass, drums you run the risk of accumulating an awful lot of unwanted frequencies.
 
This is where a clear idea of how you want the final product to sound comes into play.