I'm going to have to strongly disagree with Conor Friedersdorf's argument that civic-minded citizens should volunteer for interest groups, rather than candidate campaigns.
Friedersdorf's argument revolves around politicians, who he doesn't much like (so he considers working for them "elevat[ing] corrupt narcissists to positions of power." He argues that their campaign images are phony, and that at any rate they are corrupted by power once they win.
But that's an argument only for treating politicians as people, not as potential messiahs. And he's right as far as that goes. However, that doesn't mean that candidate campaigns are the wrong way to go.
First, while it's always possible to find examples to the contrary, for the most part politicians really do try to carry out their campaign promises.
But more importantly, the best way to think about working for a candidate isn't about the candidate, but about the party. And whether you like parties or not, they are central to how US politics works. In particular, working within the party to try to push it in the direction you want -- which is really what working for a candidate in any primary election does -- is certainly a very good way to effect change.
Suppose, for example, that you strongly support an issue. Friedersdorf's option, which is certainly a reasonable one, is to find the group that advocates for the policy you support and add your voice (time, effort, energy, money) to theirs. That's useful! But even more useful would be to help convince 40,000 or so Iowans to support a presidential candidate who advocates for the policy, especially if it's a new position for the party, or perhaps one that the party has been wobbly on in the past. You don't even need 40K in Iowa to impress; last time around, 20K would have been enough to run a respectable third place on the Republican side, and in the right circumstances that might be enough to strongly influence the party. And if you influence the party, you eventually influence policy. Because sooner or later the party will be in power, and it's most likely to act on those policies on which it is most strongly committed.
It doesn't have to be a presidential campaign, either. One of the benefits of the Madisonian system is that there are thousands of politicians who matter, hundreds of them at the national level. Directly and indirectly. Even backbench Members of the House make a difference over time, because they can "prove" to their colleagues that some once-taboo policy position has become safe over time, or even that it is now becoming a party requirement. So if you're a Tea Partier, working in 2010 for Ron Johnson or Mike Lee made lots of sense even if you didn't have any particular regard for them personally. The same would be true for liberal Democrats -- if you're looking for something to do to advance the cause of adding a public option to ACA or repealing DOMA or getting a carbon tax passed, my guess is the best thing you can do right now is to find a House primary in a winnable Democratic district in which the candidates for nomination disagree on your issue, and get involved in favor of the candidate on your side. And then next fall, find a marginal House district or Senate race and work on that campaign.
Now, to be fair, I'm not aware of any kind of empirical study comparing the efficacy of spending $1 or an hour of volunteer time, and I'm not sure such a study would be even remotely possible. (It's possible to study the effect of money, and presumably other investments, on election results -- but broadening that to the question of policy effects presumably makes it prohibitively complex). And I definitely don't want to run down participation in interest groups. And of course many interest groups themselves are party-aligned to varying extents. But my strong guess is that if one had to choose, party work, including involvement in nomination battles and also general election fights, is in general more effective. Either way, once again the big point here is that involvement in candidate campaigns should properly be seen as party work. It's not about, or at least in most cases it shouldn't be about, believing in a particular politician.