Does Sexual Orientation and Gender Identity (SOGI) Question-Wording Influence Responses?: Evidence from an Experimental Test in a Non-Traditional Sample

• Activity Number: 182 - Contributed Poster Presentations: ASA LGBT Concerns Committee
• Type: Contributed
• Date/Time: Monday, July 30, 2018 : 10:30 AM to 12:20 PM
• Sponsor: ASA LGBT Concerns Committee
• Author(s): Deirdre Middleton* and Matt Jans and Naomi Freedner and Lee Harding and Ronaldo Iachan and Scott Worthge and James Dayton
• Companies: --None-- and ICF and ICF and ICF and ICF and MFour and ICF
• Keywords: SOGI; Sexual Orientation; Gender Identity; Non-probability survey ; Question Experiment ; Mobile Panel

Abstract

Survey methodologists face a unique challenge when writing sexual orientation and gender identity (SOGI) questions. The questions must be understood by the non-LGBT population (i.e., straight and cisgender people), but also meaningful to and respectful of the LGBT community. While SOGI questions tend to perform well in general population surveys, there is evidence that question-wording can make a difference in identification estimates. This pilot randomized mobile nonprobability panel respondents to one of two GI measures and one of two SO measures. For GI, a one-step measure was compared to a two-step measure, and for SO the comparison was between two versions that are used in large-scale general population surveys. Panel members from Washington state were invited to the questionnaire when they triggered a geofence around a grocery or convenience store in that state. The effect of question-wording on both SO and GI measures is assessed, and results are benchmarked to population-based surveys of Washington. Results are discussed in the context of improving SOGI measurement and measuring the LGBT population with novel methods.