The solution to that is picking one and sticking with it from data-collection to results-presentation.
If we limit the definition to non-incarcerated adults who are exclusively having sex with persons of the same gender, it would probably be about 1-2%.
See, there's an example of not sticking to a definition. From your essay, 1-2% is the number of people who admit to being "homosexual" on surveys. As a non-scientist, while I believe one can reasonably assume those who identified "homosexual" on the survey are exclusively homosexually-active, one cannot assume that everyone who is exclusively homosexually-active would or did respond such. (Please feel free to correct me if you have sources indicating that only 1-2% are exclusively homosexually-active.)
If we expand the definition to include anyone who has ever had sexual contact with, or felt a sexual interest in, another person of the same gender (including current and former prison inmates), it might go as high as 10%.
It's most certainly much higher, especially with a definition so broad. "The Janus Report", "The Kinsey Report", and "The Hite Report" all put it over 20% just for having a single "homosexual encounter", and that's just off the top of my head. (Aside: "homosexual encounter" is another term that would need to be defined to be meaningful -- would it require orgasm or just physical contact? Would kissing a cheek count, or would the situation have to be overtly sexual? What about two boys pleasuring themselves simultaneously to the same copy of Playboy?)
Whatever the definition, consistancy counts from beginning to end.
The circumstances and conditions of the surveys I have cited guaranteed the respondents' absolute privacy, and they seem reliable enough to me.
For example, a 1991 national survey of sexually active adults done by the National Opinion Research Center shows that 98.4% of adults were exclusively heterosexual. (Smith TW. "Adult Sexual Behavior in 1989." Family Planning Perspectives 1991, 23: 104.) A 1993 survey by the Battelle Human Affairs Research Centers found that only 1.1% of all Americans are exclusively homosexual. (Rensberger B. "How Many Men in US Are Gay?" Washington Post, April 17, 1993, p. A-1.)
A survey conducted by the Alan Guttmacher Institute in 1993 found that 1% of men consider themselves exclusively homosexual. (Barringer F. "Sex Survey of American Men Finds 1% Are Gay." New York Times, April 15, 1993, p. A-1.) And a 1994 National Health and Social Life Survey at the University of Chicago found that 2.8% of men and 1.4% of women identified themselves as homosexual or bisexual. (Vobejda B. "Survey Finds Most Adults Sexually Staid." Washington Post, October 7, 1994, p. A-1.)
A March 1994 article in the American Journal of Psychiatry cites several surveys of American men showing the prevalence of homosexuality. For example, the National Survey of Men found that only 1.1% of men had been exclusively homosexual during the preceding ten years. Another study found that 2.4% of men are currently homosexual. (Cited by Seidman SN & Rieder RO. "A Review of Sexual Behavior in the United States." American Journal of Psychiatry 1994, 151: 339.)