I've always wondered about how they calculated that.

If they just taught people how to use a condom and then sent them off to their merry romps, and then checked how many of them got pregnant after a few months, then the statistic is highly suspicious - after all, it could just be from misuse (put it on the wrong way, then remembered how to use a condom and put it on the right way) or just omitting to use them (and then lying about it).

In this case, that would mean that the 2% statistic just indicates that 2% of the participants were not capable or willing to use a condom.

If that's not the way they did it - then how did they do it? Because beyond having every participant followed around by a man with a clipboard, I don't see any ways to reduce the chance of misuse or lying...