There are important points to remember when the resulting probabilities are interpreted. First of all, this method does not care what is the “correct” answer. In other words, in a difficult test where most questions only have a small proportion of students getting right, abnormal results can occur. Two very good students may appear to cheat because they share a large number of common (and correct) choices, and the probabilities of answering each one correctly is low (due to the difficulty of the test).
The reason is because in the case of two good students, the chances of picking a particular common choice is no longer “by chance”. In other words, even if most of the rest of the class select other choices of a question, the choice of a good student does not follow the probability distribution.
This does not mean that correct answers should be discarded in the consideration of probabilities. It simply means that you need to take that into account. One possible way to handle this is to separate students into groups based on score first. Then, within a group, observe the distribution of probabilities.