Goldstein: Sharing the mistakes or sharing the things that go wrong is an underrated skill. Most people are not rewarded for doing that, but that's something you can publish on your own website; so it's really great to be able to say we tried this and it didn't work.

DeCarlo: The randomized clinical control trial is the gold standard and recently there've been a couple of them on HIV prevention that some of which the sites have been based at San Francisco at CAPS that came out with kind of not as good as they hoped results.

And in fact one of them, Project Explore, which was across six cities across the U.S., I believe what they said was, "We didn't score a home run." And because of that a lot of academic journals won't publish your findings. They don't publish things that don't work. They don't publish things that might have worked or possibly didn't work.

So it's hard to get that information out there, but one of the things the researchers were talking about in Project Explore in particular was they had this very complex multi-session, multi-group intervention and they tend to think that maybe less is more. So maybe you don't need to have 15 groups, maybe you only need to have five groups.

So something like that is really important for community organizations to know and you're not going to learn that through journal articles, because it's a "failure" and it's really important to get that information out some other way if journal articles aren't going to print it.