Article Text
Abstract
Objective There is increasing interest in measuring the influence of socioeconomic position (SEP) across the life-course. However, there are few cohorts that encompass the whole life-course, so many studies use data from cross-sectional studies, hence the need for ensuring the accuracy of recall of the SEP measures used. The aim of this study was to investigate the reliability and validity of two measures of SEP: education and longest job.
Design Prospective cohort study.
Setting Large population survey of older adults from the general population in North Staffordshire, UK.
Main Outcome Measures Health Questionnaire (HQ) including a four-part question on education and a five-part question about the longest job.
Participant In two pilot studies, random samples of patients aged 50+ registered with different general practices were mailed a HQ (nested test–retest repeatability studies were included; n=500; n=160). In the main study, all patients aged 50+ registered with three different general practices were sent a HQ at baseline. 6 years later, 3410 responders consenting to further contact were mailed a follow-up HQ; a subsample of these participants took part in independent qualitative interviews, which included a reconstruction of critical landmarks in their life-course. Ethical approval was granted for all stages of the study.
Analysis Validity was assessed by: percentage completion; internal percentage agreement within each set of questions; percentage agreement of qualitative and quantitative data for age left school and longest job questions; comparing age left school data with historical change in legal school leaving age. Reliability of both questions was calculated from the pilot repeatability studies using ICC and κ, with 95% CI.
Results The adjusted response to all stages of the study was good (62–84%). In the main surveys, completion of the age left school and longest job questions was excellent (94–98%). Internal agreement within each set of questions was good (education: 80–97%; longest job: 95–100%). Comparison of survey and interview data showed good agreement for age left school (85%) and longest job (95%). The change in age left school data concurred exactly with the change in legal school leaving age implemented in 1947. Reliability from pilot test–retests was excellent for recalled age left school (ICC=0.95; 95% CI 0.93 to 0.96) and substantial for longest job (κ=0.69; 95% CI 0.55 to 0.83).
Conclusion The age left school and longest job questions in this study are valid and reliable measures.