We report the results of an experiment we performed to test the matching probabilities for the Ellsberg paradox predicted by the quantum decision model of al-Nowaihi and Dhami (2016). We find that the theoretical predictions of that model are in conformity with our experimental results. This supports the thesis that violations of classical (Kolmogorov) probability theory may not be due to irrational behaviour but, rather, due to inadequacy of classical probability theory for the description of human behaviour.