There's a good line in a Richard Powers interview somewhere-or-other in which he talks about how he threw everything he knew into his first novel, because he didn't know if he'd have a chance to write a second. I like that. A bit too much, perhaps.
I'm really susceptible to going off on wild tangents of reading up on a subject that interests me, on the sometimes-dubious grounds that I need to know more about it for the novel. Just as an example, some subjects that I've researched in writing the current book:
* mining in Colorado
* late 19th-century labor activism and violence
* weather patterns
* what happens if you blow up a dam (hello, Homeland Security!)
* Salton City and its decline
* states' rights
* utopian societies in the United States
* U.S. troops being sent into Russia toward the end of World War I
* other conspiracy theories, including one that there are secret tunnels under Denver International Airport
* the effects of cocaine and peyote (a lifetime without drug use has left me sadly deficient in some areas)
* the layout of San Diego
...and a bunch of other things that I'm just not thinking of right now.
The point being, when the writing isn't going well at all I look for ANY distractions, and researching something ostensibly for the book is one where I can keep the illusion up that I'm WORKING. And that's probably counterproductive. I gotta quit that.