Abstract
BACKGROUND: Social roles common to adulthood (e.g. employment, parenthood) and their timing and combinations have been shown to relate to health. However, there is a lack of contemporary data to study complex patterns. We applied a pragmatic algorithmic approach to data from the ALSPAC (Avon Longitudinal Study of Parents and Children) G1 cohort born in 1991-1993 to build valid annual indicators at age 16-31 of being in (or out of) six key roles. METHODS: In February 2023 we searched the data catalogue and identified 449 relevant variables (318 root questions, 131 auxiliary) indicating if an individual had been or was currently in the following roles: further/higher education; employment; living away from the parental home; cohabiting with a partner/spouse; parenthood; being a carer. Variables were incorporated into four algorithms to determine whether each individual was in/out of a role per age year. We addressed missing indicator values with multiple imputation methods. We assessed how well indicators captured annual role status by using them to derive descriptive statistics and comparing with those from national and survey data from the same period. RESULTS: Descriptive statistics on transitioning to or leaving a particular role by age 30 were comparable to national official data. For example, by age 30, at least 27% of men and 50% of women indicated having left the parental home (at median age 23-24); of these individuals, 19-30% subsequently indicated living with their parents again by age 30 (median interval 2-3 years, interquartile range 1-4 years). However, employment and parenthood appeared to be under-captured, relative to the other four roles. CONCLUSIONS: These indicators can now be used by other researchers, for example to study a particular role or construct (e.g. Not in Education, Employment, or Training) over time, in different social and health contexts. They are best used to derive composite variables for use in life course research, rather than for precise year-on-year sequence analysis and parenthood should be used with caution.