In decisions under risk more numerate people are typically more likely to choose the option with the highest expected value (EV) than less numerate people. Prior research indicates that this link between numeracy and decision quality cannot be explained by differences in the reliance on explicit calculation of EVs between the more and less numerate—raising the question, which other pre-decisional processing mechanisms might give rise to it. The current work uses the attentional Drift Diffusion Model (aDDM, Krajbich et al., 2010) as a unified computational framework to formalize, quantify and disentangle several candidate mechanisms. The model is used to derive hypotheses about three key aspects of pre-decisional information search and processing—namely attention allocation, amount of deliberation, and attentional modulation of information integration—which may differ between more and less numerate people and which may explain why more numerate people tend to make better decisions. The analyses show that while numeracy barely affects how people allocate their attention across the options (captured using eye-tracking), it systematically modulates how much information people require before committing to a choice (captured in the aDDM’s boundary separation parameter) and also how strongly their processing of information is distorted by what they attend to prior to making a choice (captured in the aDDM’s attentional amplification parameter). Together, the latter two aspects mediate the effect of numeracy on decision quality to a large extent. Thereby, the current account offers a new, integrative and formally rigorous angle on the cognitive underpinnings of the link between numeracy and decision quality.