Abstract Background and aims Higher ventricular mass has been reported in non-white US-Americans with low educational status and in socially isolated people. To assess the impact of education on cardiac mass and function in the general population and to identify mediators. Methods and results Data from a German population-based sample were used (CARLA cohort, n = 1779 at baseline, n = 1436 at the four-year follow-up). Ventricular mass indexed on height (LVMI) and ejection fraction, using Teichholz's formula (EFTZ), were measured. Education was assessed using the ISCED classification. Mediator analyses were performed using the R-macro ‘mediation’ to compute the average direct effect and the average causal mediated effect after confounder adjustment. Sensitivity analyses for unobserved confounders were performed. Considered mediators were BMI, waist-to-hip ratio, HbA1c, and systolic and diastolic blood pressures. We found differences in LVMI and EFTZ, both at baseline and follow-up, between educational levels in women (lowest vs highest educational level: 15.6 g, 95% CI: −25.7, −5.6), but not in men. Similarly, women (lowest vs highest educational level at baseline: 3.3%, 95% CI: 0.8–5.7), but not men, of higher educational levels had a higher EFTZ of comparable magnitude at baseline and follow-up. Of the considered mediators, BMI explained 55.9% at baseline and 54.1% at follow-up of the educational effect, while other potential mediators had no significant effect. Relations remained constant between baseline and follow-up. Conclusions Women with low educational levels tend to have a higher ventricular mass and lower EF, which can be explained by a higher BMI in this group.