Got into a conversation about this yesterday: how much help does a vice-presidential candidate provide in winning that candidate's home state? I vaguely recall that poli science says not much. I went and noodled around wiki and can now draw my own dramatic conclusion: not much.
Wiki has all presidential results by state and year (e.g., here's Texas 1988) so it's simple to compare results before and after a state resident ran for vice president. In the last 30 years, not much happened, although 1992 and 1996 are hard to use because of a strong third party showing. I'd say everyone brought in much less than a 5% bump, with only Bentsen and (sadly) Palin coming in at or slightly above that level.
This small of a bump suggests that veep candidates shouldn't be chosen based on the help they provide in their home state.
OTOH, there's Florida - that's a very big swing state, and a 2% bump could be useful. I've thought a joint ticket of Jeb Bush and Marco Rubio could make winning Florida very difficult for Democrats. I believe Jeb isn't particularly popular in Florida and Rubio is only moderately popular, but people do tend to root for the home team.