Improving Questions on Sexual Partnerships: Lessons Learned from Cognitive Interviews for Britain’s Third National Survey of Sexual Attitudes and Lifestyles (“Natsal-3”)

Catherine R. H. Aicken, Michelle Gray, Soazig Clifton, Clare Tanton, Nigel Field, Pam Sonnenberg, Anne M. Johnson, Catherine H. Mercer

Patterns of sexual partnership formation and dissolution are key drivers of sexually transmitted infection transmission. Sexual behavior survey participants may be unable or unwilling to report accurate details about their sexual partners, limiting the potential to capture information on sexual mixing and timing of partnerships. We examined how questions were interpreted, including recall strategies and judgments made in selecting responses, to inform development of a module on recent sexual partnerships in Britain's third National Survey of Sexual Attitudes and Lifestyles ("Natsal-3"). Face-to-face cognitive interviews were conducted with 14 men and 18 women aged 18-74 years, during development work for Natsal-3. People with multiple recent partners were purposively sampled and questions were presented as a computer-assisted self-interview. Participants were generally agreeable to answering questions about their sexual partners and practices. Interpretation of questions designed to measure concurrent (overlapping) partnerships was broadly consistent with the epidemiological concept of concurrency. Partners' ages, genders, ethnicity, and participants' perceptions of whether partner(s) had had concurrent partnerships were reported without offense. Recall problems and lack of knowledge were reported by some participants (of all ages), especially about former, casual, and/or new partnerships, and some reported guessing partners' ages and dates of sex. Generally, participants were able to answer questions about their sexual partners accurately, even when repeated for multiple partners. Cognitive interviews provided insight into the participants' understanding of, ability to answer, and willingness to answer questions. This enabled us to improve questions used in previous surveys, refine new questions, and ensure the questionnaire order was logical for participants.