I was half-joking in my previous post when I said that marathon training might delay teen sex more than abstinence programs. Of course, that would not hard to do because the definitive study of abstinence programs finds they don't delay teen sex at all.
Marathons fit in with the question of how people who don't intend to run marathons admire those who do, and I wasn't thinking about it as a potential intervention. The reason marathon training might delay teen sex is that the studies by the National Campaign to Reduce Teen Pregnancy show that comprehensive lifestyle programs are the most effective programs of all: these are the programs which involve the adolescents in some kind of activity, improve support systems, and take a lot of time.
The half-joking reasons why marathon training might delay teen sex are that they will be too tired and scheduled to find a partner and find time alone with the partner. But in the ideal case, first, it could create a close-knit and supportive group that meets regularly for long periods of time, so the group will reinforce its norms more strongly than a more loose group. Second, setting a high goal and working towards it may raise participants' self-efficacy, the confidence that they can accomplish their goals in general. Self-efficacy may be particularly important in the case of religious teens that have sex even though they think they shouldn't. Third, working towards a big goal may make that goal more salient in their minds, and may make sex slightly less salient. The endorphins wouldn't hurt. Plus, as one commenter said, maybe it would delay puberty, though I'm not sure if marathon training is safe so young.
This is assuming the maximal best case where all of the participants are super into the idea. Obviously in the real world, most would drop out, and only the most risk-seeking kids would stay in the program so it would be hard to find the right control group to test whether it works.
Casual team sports take up some time, but might not have enough intensity to raise participants' self-efficacy.