National Health and Nutrition Examination Survey

The Mobile Examination Center (MEC) Interview section on sexual behavior (variable name prefix SXQ) provides information on lifetime and current sexual behavior for both men and women. Major items in this section include age at first sexual intercourse, number of sexual partners, circumcision status (men), and history of sexually transmitted diseases. The sexual orientation question was slightly modified in this cycle.

Survey participants aged 14-69 years were eligible. Several questions in this section are age-specific, and many are gender-specific. Mentally impaired individuals or participants who were unable to understand English, Spanish, Korean, Vietnamese, or Chinese (traditional/Mandarin, simplified/Mandarin, and traditional/Cantonese) were not asked these questions. Only data from participants aged 18-69 years are included in this release. Special use data files for participants 14-17 years are available through the NCHS Research Data Center.

The sexual behavior questions were self-administered in a private room at the MEC, during the MEC Interview, using the Audio Computer-Assisted Self Interview (ACASI) system. The ACASI was conducted in one of the following languages: English, Spanish, Korean, Vietnamese, or Chinese (traditional/Mandarin, simplified/Mandarin, and traditional/Cantonese). The ACASI enables respondents to hear questions through earphones as well as read questions on the computer screen. Respondents move at their own speed and touch the screen to indicate their response. No proxy respondents or translators were used in situations when the respondents could not self-report. The questionnaire is located on the NHANES website at: https://wwwn.cdc.gov/nchs/data/nhanes/2015-2016/questionnaires/sxq_acasi_i.pdf.

For details on the QA/QC process for this component, please refer to the MEC Interview Manual on the NHANES website.

The sexual behavior questionnaire in 2015-2016 is similar to cycles 2013-2014, 2011-2012 and 2009-2010. However, analysts should be cautious when combining data with cycles prior to the NHANES 2009-2010 SXQ data, since some questions have been removed or added, and some questions have had slight wording changes. Please see the Analytic Notes below as well as Appendix 1 (for male) and Appendix 2 (for female) for detailed information on variables included in the SXQ section since 2003-2004.

The variable SXD021 (Ever had vaginal, anal, or oral sex) was derived from whether or not the respondent reported having any of the specific types of sex. For example, if a male answered “yes” to having had any of the following types of sex: SXQ800 (Ever had vaginal sex with a woman), SXQ803 (Ever performed oral sex on a woman), SXQ806 (Ever had anal sex with a woman), or SXQ809 (Ever had any sex with a man; anal, oral); he would be coded as a “1” or “yes” in SXD021. In 2015-16, 9 sample persons were assigned a missing value for SXD021 (SEQN=85190, 86400, 87777, 87947, 88537, 88657, 89444, 92055, 93654).These persons responded “don’t know” or “ refused” to at least one of the questions that were part of the algorithm for deriving SXD021, and had “no” responses for the other questions in the algorithm. Analysts should be aware that any response to other questions in this data set have been retained for these persons.

Responses to SXD031 (age at first sexual intercourse), SXD633 (age first performed oral sex on a woman), and SXD621 (age when first had oral sex) that were 9 years or less were grouped and coded as “12 years or younger.” In previous SXQ data files, young ages were categorized as “9 years or younger.”

Responses to questions about number of sexual partners (SXD171, SXQ824, SXQ827, SXQ636, SXQ410, SXQ836, SXQ624, SXD101, and SXQ724) that were 100 or greater were categorized as “100 or more.” In previous SXQ data set releases, high numbers of partners were categorized as “1000 or more” or “2000 or more.”

Information was collected using two-part (number and unit) questions for both SXD630 (last performed oral sex on a new male partner) and SXD642 (last performed oral sex on a new female partner). This is to allow participants to report how long ago they had performed oral sex either in days, weeks, months, or years. The released variables were edited to standardize the reported length of time to number of days, using the following conversion factors: 7 days per week, 30.4 days per month, and 365.25 days per year.

In general, if a person wasn't asked a question, the response to the question is shown as “missing.” It is up to the analysts to recode the response to zero as they see fit. For example, if a respondent reported that they had zero lifetime vaginal sex partners, they wouldn't be asked the number of vaginal sex partners in the past 12 months. Since the respondent does not receive the question about partners in the past 12 months, the data will be coded as missing for this variable. Analysts may want to recode this to “0.”

For questions that ascertain total number of partners for all types of sex (i.e., SXD171, SXD510, SXD101 and SXD450), if the respondent had reported never having any type of sex, the responses were coded as “0.”

Analysts should keep in mind that respondents aged 60-69 years get asked only some questions. It is important to refer to the codebook to see the target age for each question.

Although the ACASI has many advantages over interviewer-administered questionnaires for sensitive information, inconsistent responses can occur. Since the questions are self-administered, unlikely answers or confusing questions generally cannot be clarified by the interviewer. Although, the questionnaire was programmed to alert respondents of potential data entry errors or inconsistencies, not every possible consistency check was added. To minimize potential respondent frustration and confusion, the interview was not interrupted each time an inconsistent answer occurred in terms of the number of sexual partners for different periods of recall for different types of sexual behavior.

Beginning in 2015-2016, modifications on question wording and response categories were made to the sexual orientation question, specific to males and females, after extensive cognitive and field testing was conducted by NCHS (Dahlhamer et al.). The current question, is now the same as the sexual orientation question in the National Health Interview Survey (NHIS). The question reads, “Which of the following best represents how you think of yourself?” For males (SXQ296), the response options are “Gay,” “Straight, that is not gay,” “Bisexual,” “Something else,” “I don’t know the answer.” For females (SXQ295), the response options are “Lesbian or gay,” “Straight, that is not lesbian or gay,” “Bisexual,” “Something else,” “I don’t know the answer.” According to Dahlhamer et al., the NHANES question from 2013-2014 and before (SXQ292 – males ; SXQ294 – females), “Do you think of yourself as…“ used explanatory text with the main response options (e.g., “homosexual or lesbian, that is sexually attracted only to women),” resulting in questions that intermixed sexual identity with sexual attraction. In 2015-2016, the “unsure” response category shown on the screen was replaced with the response option, “I don’t know the answer.” For all years, if participants do not answer the question, and instead, press “next” to move forward, they get a screen asking them to indicate – “I really meant to answer”, “I don’t know the answer” and “I’d rather not answer.”

Analysis by subgroup characteristics should be conducted carefully. The sample size for the non-heterosexual subgroups is limited and may not meet the minimum sample size requirements for statistical reliability as noted in the NHANES analytic guidelines. Combining multiple years of data, which is often recommended to estimate rarer conditions within a subgroup, analysts will need to consider sample sizes, as well as the changes to the question and response categories made over the time period of analysis. Additionally, the high rate of not selecting a specific answer to the questions should be considered (Miller et al). That is, selecting: “don’t know,” “refused,” “not sure,” or “something else” as a response. Individuals select these answers for a variety of reasons that are well summarized in Table 2 of the NCHS report “Design, Development and Testing of the NHIS Sexual Identity Question.” For example, respondents from the LGBT community select “something else” when they use a different sexual identity label than those presented in the response options. However, those who aren’t part of the LGBT community select this answer because they are not familiar with what the response options mean. For other questions and answers on this topic, see: https://www.cdc.gov/nchs/nhis/sexual_orientation/faqs.htm.

Exam sample weights should be used for analyses. Please refer to the NHANES Analytic Guidelines and the on-line NHANES Tutorial for further details on the use of sample weights and other analytic issues.