Monty Hall's three doors problem is a well-known "anomaly" in economics. It appears to be an example of a systematic violation of the assumption of subjects' rationality. Many papers have studied the Monty Hall anomaly under different perspectives (i.e. computerizing, learning, grouping, talking, and even teaching how to play it), but the anomaly still survives. The most common explanation for this is that subjects are unable to carry out correctly the Bayesian updating necessary. Our approach is different from all previous attempts to explain and correct the Monty Hall anomaly, inasmuch we developed a more radical "debiasing test". Even if the game remains identical in terms of probabilities and subjects task, we eliminate the necessity for any Bayesian updating, as our new framework does not require subjects to make any probability calculations. Consequently, having eliminated the cause, also the effect should have disappeared. In order to make our results robust, we also run an intermediate treatment. Even though we observe an evident monotonic increase in the switching percentage across the three treatments, as we expected (from 41.5% in the treatment that replicates the standard design, to 45.5% for the intermediate treatment, up to 58% in the new framework), this percentage remains still too low, even though no Bayesian updating is involved. These results drive us to conclude that this anomaly, even if attenuated by design conditions, is not a weak effect, but rather a systematic behavioural regularity. It could be attributed to some psychological underpinnings, such as the status quo bias. We comment on these, and alternative explanations, in our conclusions.
