More Americans than ever identify as bisexual, according to a report released by the CDC yesterday. But that's not just changing social norms; it may be due to the way the US government is phrasing the question. Unfortunately, the CDC's questions about sexuality aren't always consistent — or even all that sex-savvy.

The CDC asked more than 9,000 people in the US between the age 18 and 44 about their sexual experiences, who they're attracted to, and to self-identify their sexual orientation in the 2011-2013 National Survey of Family Growth. This survey matters because the portion on sexuality is used to identify US populations that might be at higher risk for STIs or HIV. In the survey, 1.3 percent of women and 1.9 percent of men identified either gay or lesbian. (It’s not immediately clear how people who identify outside the gender binary were classed — or if they were included at all. The NSFG only includes a male and a female questionnaire.)

In addition, 2 percent of men identified as bisexual, compared with 1.2 percent of men in the 2006-2010 NSFG survey. For women, the proportion of women who identify as bisexual rose to 5.5 percent of women, from 3.9 percent in the previous survey. Those differences may have to do with changing attitudes toward bisexuality and gay identities — but the findings are quite different from those in another CDC survey, the 2013 National Health Interview Survey. In that study, only 0.9 percent of women and 0.4 percent of men identified as bisexual. It's a weird difference, given that the time period for the 2013 NHIS and 2011-2013 NSFG overlapped.

Self-identification doesn't tell the whole story, in terms of sexual behaviors or attraction. For instance, the survey found that the ways in which people identify doesn't always line up with their sexual behaviors or who they're attracted to. Case in point: 17.4 percent of women and 6.2 percent of men said they had same-sex sexual contact in their lifetime. That doesn't jibe with another figure though: 19 percent of women and about 8 percent of men said they weren't exclusively attracted to members of another gender. These findings aren't especially new, however. For the most part, the survey's results were very similar to those reported in the 2006-2010 NSFG survey — except with regards to bisexuality.

The CDC's methodology is key

The first thing to note here is that the government doesn't know why the results from both surveys — the NSFG and the NHIS — are so different. "With all these different estimates, there are lots of reasons why [these differences] could be," says Casey Copen, CDC demographer and co-author of the NSFG study. "It could be a trend, it could be the question wording — it could be lots of different things."

And the studies' methodologies could play a role too. The NHIS is conducted with in-person interviews instead of through a computer — as the portion about sexuality in the NSFG is. People who didn't have to disclose their sexuality to a total stranger may have felt more comfortable, some people suggest. "There has been data in the past showing that that makes a difference," Copen says. But there's conflicting information about that even within the CDC; when the researchers behind the NHIS looked into this they didn't find the same trend, Copen says. There was no difference between face-to-face interviews and interviews conducted through a computer.

But there's another way study methodology could impact the findings, says Gary Gates, a demographer who focuses on the LGBT population who recently retired from the Williams institute. In the NHIS, people were asked about their sexual orientation only. In this week's study, people were asked about their sexual behaviors and attractions, as well as their orientation. When people are only asked if they're gay, straight, or bi, they might not stop to really think about it. But asking more questions about what someone truly thinks and feels may actually move some respondents to identify as bisexual within the context of the survey.

"There's some evidence that suggests that people who acknowledge having some same-sex sexual attraction are more likely to identify as LGB," Gates says.

The problems with methodology don't stop there. In this week's survey, women were marked as having same-sex sexual experience if they had performed oral sex on another woman, received oral sex from another woman, or if they had had "any sexual experience of any kind" with another woman. Men, on the other hand, were only marked as having same-sex sexual experience if they had given oral sex to or received oral sex from a man, or if they had given anal sex to or received anal sex from a man. They had no "any sexual experience of any kind" option. That difference may have played a role in the large gap that the CDC found between men and women with regards to same-sex sexual experiences — and subsequently in the number of men who said that they identified as bisexual.

"We're disappointed in the inconsistencies in definition of same-sex experiences for men versus women," Kate Estrop, co-president of the Bisexual Resource Center, told The Verge in an email. A broader definition of same-sex experience for women may have contributed to the numbers for women with same-sex experience being almost three times higher than for men with same-sex experiences, she says. Gates agrees. "The difference in how same-sex sexual behaviors are measured could be one reason that these figures are so much higher among women." Estrop also points out that a more rigid definition of "sexual experience" for men with same-sex partners "lends to the idea that sex and sexuality itself is less fluid for men than for women, which is not always the case in reality."

The survey's flaws are so great that Estrop says she feels the numbers fail to represent the actual population. The survey also fails in one of its main purposes: to understand human sexuality, and make sure at-risk groups get the sexually transmitted infection protection they need. If bisexual men aren't properly counted, Estrop says, they "will continue to have health disparities — including STIs — that are not being identified."

The NSFG was the first big government survey to routinely include questions about same-sex sexual activity. It also separates the questions of sexual attraction and behavior. But there's still a lot that the CDC could do to ensure that people are adequately represented in all the agency's surveys. "If you don't ask certain questions and get answers, then policy makers think you don't count. They ignore you," Gates says. Getting human sexuality right on a federal level is important not just for health — but for the dignity of the people being counted.

Update January 8th, 4:39PM: This story was updated with information about people who identify outside of the gender binary.