I have not been able to find any non-kinsey institute study that shows this. Kinsey and a number of their studies use prison populations to get there; you can't look at a man that's been raped in prison and say that because he's engaged in homosexual behaviour he is gay or bi.
It feels believable because you see plenty of gays in media and the arts. It isn't that common when you really dive into the data.
Btw, I still operate from the presumption of "born that way" in regards to homosexuality, which means engaging in homosexual behaviour once isn't indicative of homosexuality as much as it is of experimentation. It's only repeated engagement that indicates homosexuality to me.
I'm talking about behavior, not sexual identity. I am not saying that if a man engages in homosexuality that he's a homosexual or bi. I'm taking about how many people engaged in homosexual acts. I mean people engaged in all types of behavior for a variety of reasons. For example, gay for pay porn or drug-addicts that engage in homosexuality for the reason of money.
I know and acknowledged that self=identified homosexuals make up about 2% of the population. It's believable to me because people will do anything to achieve their motives.
ETA: I don't think that sexuality is as fixed people like to believe. There have been studies how sexuality shifts during one's time.