Consumer loan requests are randomly assigned to gender-balanced real borrowers, who subsequently submit them to a representative sample of loan officers from Chilean banks, whose gender preferences are elicited prior to the experiment. The approval rate among women is 18% lower relative to men, primarily attributed to taste-based discrimination from officers with pro-male preferences. Then, a de-biasing intervention is tested, wherein officers receive information on official statistics indicating that women exhibit higher repayment rates than men. Biased officers update their beliefs, but discrimination remains unchanged, meaning inaccurate statistical discrimination is unlikely to contribute to explaining gender discrimination in consumer lending.