We experimentally study the effect of financial education on investment attitudes in a large sample of high school students in Italy. Students in the treated classes were taught a course in finance and interviewed before and after the study, while controls were only interviewed. Our principal result is that the difference-in-difference estimates of the effect of the course are not statistically significant. However, the course in finance reduced the virtual demand for cash, and increased the level of financial literacy and the propensity to read (and the capacity to understand) economic articles in both treated and control classes compared with pre-treatment baseline levels. A breakdown of the cognitive process, which is statistically significant for the classes treated, suggests that error and ignorance reduction was sizable, and that the progress in financial literacy was stronger in subgroups which exhibited lower ex-ante knowledge levels.