Suitable normative information on the Iowa Gambling Task (IGT) is not currently available, though it is clear that there is great individual variability in performance on this assessment tool. Given that the task is presumed to measure the emotion-based learning systems that are thought to form the biological basis of 'intuition,' there is some reason to think that education (especially tertiary education) might explicitly de-emphasise the role of emotion-based learning in decision-making. This suggests the paradoxical finding that better-educated participants should show poorer performance on the IGT. We recruited 30 participants (all female, all aged 18-25) to participate in a 'real money' version of the IGT. There was no significant difference in performance in blocks 1-3 of the task (trials 1-60). However, there was a substantial effect of education on the final two blocks (trials 61-100), such that the less-well-educated participants produced twice as much of an improvement over baseline as did their university-educated colleagues. A range of possible explanations for this remarkable finding are discussed. The most likely appears to be that tertiary education specifically discourages the use of emotion-based learning systems in decision-making. These findings bear on the extent to which education has a role to play in our reliance on cognition and emotion in decision-making, including the likely role of education in the generation and maintenance of false beliefs.