Putting the “T” in LBGT: A transgender question pilot test in the California Health Interview Survey David Grant, Matthew Jans, Royce Park, and Ninez A. Ponce UCLA Center for Health Policy Research Gary Gates, Jody Herman, and Bianca Wilson UCLA Williams Institute 1 Gender identity (GI) measurement Question: Is it possible to reliably measure gender identity in a population-based telephone health survey? Vehicle: California Health Interview Survey (CHIS) Test: 4 versions of gender identity questions Two “2-step” versions Two “1-step” versions Results: Promising 2 Gender identity measurement CHIS has a history of illuminating health and health-related information for underserved and “invisible” groups Asian ethnic groups: Chinese, Filipino, Korean, South Asian, Vietnamese, Cambodian, Japanese Latino ethnic groups: Mexican, Salvadoran, Guatemalan, European, South American Linguistically-defined groups: Chinese, Korean, Vietnamese, Tagalog Non-citizens Lesbian, Gay, Bisexual (LGB) How to expand to include the “T” -- Transgender? 3 Gender identity measurement, cont. Transgender certainly an “invisible” group (often by choice)… 0.3% of US adults (Gates 2011) 0.5% of adults in Massachusetts (Conron 2012) Few probability surveys measure GI Little evidence on transgender disparities Mass. BRFSS (Conron 2012) Higher unemployment and higher poverty Higher educational attainment Higher smoking rates 4 What to ask? In consultation and collaboration with experts from the Williams Institute, 4 GI questions were selected for testing Two “1-step” versions Define “transgender” Ask “are you transgender?” (yes or no) 5 GI pilot test questions: One-step One-step Version 1 Some people describe themselves as transgender when they experience a different gender identity from their sex at birth. For example, a person born into a male body, but who feels female or lives as a woman. Do you consider yourself to be transgender? One-step Version 2 Sex is what a person is born. Gender is how a person feels. When a person’s sex and gender do not match, they might think of themselves as transgender. Are you transgender? 6 What to ask? Two “2-step” versions Sex assigned at birth (male or female) Current gender identity (male, female, or transgender) 7 GI pilot test questions: Two-step Two-step Version 1 1. What sex were you assigned at birth, on your original birth certificate? 2. Do you currently describe yourself as male, female, or transgender? Two-step Version 2 1. What sex were you assigned at birth, on your original birth certificate? 2. Do you currently describe yourself as male, female, transgender, are you not sure yet, or do you not know what this question means? 8 The test and results Adult respondents age 18 to 70 interviewed in English or Spanish Random assignment to one of the four GI versions Midpoint of interview following Qs on sexual orientation N = 2,828 Combined frequencies Gender Identity Not transgender Transgender Not ascertained Total 9 Number 2,783 9 36 2,828 Percent 98.4% 0.3% 1.3% 100.0% GI test results: frequencies 1-step version 1 Not transgender Transgender Not ascertained Total Number 678 2 2 682 Percent 99.4% 0.3% 0.3% 100.0% 1-step version 2 Not transgender Transgender Not ascertained Total Number 718 2 17 737 Percent 97.4% 0.3% 2.3% 100.0% 2-step version 1 Not transgender Transgender Not ascertained Total Number 683 2 5 690 Percent 99.0% 0.3% 0.7% 100.0% 2-step version 2 Not transgender Transgender Not ascertained Total Number 704 3 12 719 Percent 97.9% 0.4% 1.7% 100.0% 10 GI test results Monitoring demonstrated few problems in administering the GI questions, or for respondents to answer them (Jans) Frequency review suggests low levels of nonresponse for two versions (1-step version 1 and 2-step version 1) Review of break-off data (hang-ups) showed several break-offs during this questionnaire section but NONE during or immediately after the GI items 11 Which version performed best? 2-step version 1 was the consensus best performer Short, quick, and clear questions that are easy to administer and answer Allows Respondent to self identify with their current gender status as male, female or transgender (not always possible in the 1-step versions) Allows measurement of T through those that report: 1. “Transgender” or 2. Current gender identity that is different than their sex assigned at birth (“off-the-diagonal” responses) 12 Two-step tweaks Add response categories (male/female) to stem of first question Original: What sex were you assigned at birth, on your original birth certificate? Revised: On your original birth certificate, was your sex assigned as male or female? Given low-frequency and potential for coding errors, add a short confirmation question for “off-the-diagonal” responses: Just to confirm, you were assigned {M/F} at birth and now describe yourself as {M/F}. Is that correct? 13 Tweaks continued… Re-think and conceptualize “gender” in surveys How to handle sex/gender-based skips, such as questions on prostate cancer testing? Gender identity or sex assigned at birth? Standard transition to section on men’s health: These next questions are about men’s health. For confirmed transgender male-to-female respondents, use alternative transition language: These next questions may be relevant to you because you were assigned male at birth. If not, let me know and I will skip them. 14 Moving forward… Add revised 2-step GI questions and confirmation item to full adult CHIS sample in 2015-2016 in all interview languages (English, Spanish, Chinese, Korean, Tagalog, and Vietnamese) Continue to monitor performance through recorded interviews (in all languages) and review of data Monitor need for confirmation of “T” assignment Estimate identifying ~60-100 transgender adults per year 15 Conclusion Pleasantly surprised how easy the GI questions were to administer Expect that YES, GI can be reliably measured on population health surveys such as CHIS Very limited data on Transgender population, especially from probability-based surveys Continue to build knowledge base of both survey capability and Transgender health inequities 16 Thank you dgrant@ucla.edu 17