Luke Muschialli, Victoria L Parsons, Soazig Clifton, Elise Paul, Rebecca Light, Katharine Sadler, Claudia Gascoyne, Anika Knuppel, Emily Trickett, Nigel Field, Christopher Bonell, Wendy Macdowall, Kirstin R Mitchell, Catherine H Mercer, Pam Sonnenberg, Andrew Copas
Evidence of changing bias between Natsal-3 and Natsal-4 was observed in key SRH estimates, which may be due to changes in participation or reporting. Resulting overestimations or underestimations of the magnitude of change over time in cross-survey comparisons have implications for interpreting apparent population-level trends to inform SRH policy and practice.
OBJECTIVES: The fourth National Survey of Sexual Attitudes and Lifestyles (Natsal-4; conducted 2022-2024) identified substantial changes in reported sexual and reproductive health (SRH) in Britain compared with Natsal-3 (2010-2012). Lower response rates in Natsal-4 necessitated mixed-mode adaptations, raising concerns that estimates of change over time could be biased by changes in who took part (participation bias) and how they answered questions (reporting and instrument bias).
METHODS: We identified two birth cohorts of participants eligible to be sampled in Natsal-3 and Natsal-4, enabling comparison of people of the same generation across the two surveys: those born 1964-1980 (ie, Gen-X) and 1981-1994 (ie, Millennials). If bias remained stable between surveys, we would expect consistent estimates from members of the same birth cohort when sampled in each survey for events occurring before Natsal-3 data collection (ie, before 2010) and for outcomes where prevalence can logically only increase or only decrease over time.
RESULTS: In Natsal-4 compared with Natsal-3, a higher proportion of Gen-X and Millennial cohorts reported having had same-sex sex before 2010; a lower proportion of men reported opposite-sex sex before 16 years old; and a lower proportion of women reported a first pregnancy outcome and live birth before 2010 (all of which should be stable across surveys). The largest change in bias was observed among women and men in the Gen-X cohort through their increased reporting of never having had a sexual experience and never having had sex, which logically should only decrease over time.
CONCLUSIONS: Evidence of changing bias between Natsal-3 and Natsal-4 was observed in key SRH estimates, which may be due to changes in participation or reporting. Resulting overestimations or underestimations of the magnitude of change over time in cross-survey comparisons have implications for interpreting apparent population-level trends to inform SRH policy and practice.