Objective: This study examined the impact of underreporting on tests of the cumulative advantage and disadvantage hypothesis (CAD), which predicts age-related increases in health disparities between individuals with higher and lower education.

Methods: Using the English Longitudinal Study of Ageing (ELSA), we identified underreporting by comparing selfreported hypertension and diabetes with biomedically measured hypertension (systolic blood pressure≥140 mm Hg and/or diastolic blood pressure≥90 mm Hg) and diabetes (fasting glucose level≥7 mmol/l and/or HbA1c≥6.5%). In a sample of 11,859 respondents aged 50 to 85 (54% women, 97% White), we assessed the
associations between underreporting and the main analytic constructs in tests of the CAD (education, age, sex, and cohort).

Results: The results showed that self-reported measures underestimated the prevalence of hypertension and diabetes. Underreporting showed weak to moderate associations with the main constructs in tests of the CAD, being more pronounced in individuals with lower education, in older age, in more recent cohorts, and among men. When correcting for underreporting using biomedical measures, the overall prevalence of hypertension and diabetes increased substantially, but education differences in age trajectories of both conditions remained similar.

Conclusions: Underreporting affected conclusions about the prevalence of hypertension and diabetes, but it did not affect conclusions about the CAD hypothesis for either condition.