There are two points worth making, if only because any scientist such as this one should be well aware of them.
1) The NHSLS was not a longitudinal study. Therefore, it does not account for generational differences in homosexual identity reporting (or formation).
2) 8% and 2% is within the statistical margin of error of the relevant subgroups..
It's also worth noting that the percentages reporting identity, activity, and ongoing activity were higher for men than for women in NHSLS. It's best (more scientific) to give the respective figures for each sex rather than the combined average, IMHO.
Assuming this is a well-done study, how can you make a claim on margin error without even knowing the sample size?