Judd, N.; Kievit, R.

Telomere length is increasingly used as a proxy for an aging exposome\", an index of non-genetic environmental (and behavioral) exposures that an individual encounters, leading to differences in rates of aging. One of the largest associations with telomere length is schooling, leading to the suggestion of long-lasting protective effects of (additional) education. However, for ethical and practical reasons, education cannot generally be randomized, precluding an understanding of the causal effect of education on health outcomes. Here, we leverage a well-established policy change in the UK to test whether an additional year of education caused differences in telomere length decades later. Employing a fuzzy regression discontinuity design, we find no effect of education on telomere length in the UK BioBank across multiple analytic frameworks (continuity, local randomization & Bayesian approaches) and a series of robustness tests, such as increasing the analysis window around the cutoff, and changing the prior, similarly find no effect of an additional year of education on telomere length. This robust finding contrasts the prevailing \"aging exposome account\" of telomere length and offers a cautionary note in attributing causal power to environmental factors based on observational differences.