[Method for calculating the distribution of randomly expected scores in a false-true-do not know-type of test]

Multiple choice tests have been used widely in the evaluation of knowledge. The lowest passing limit is generally chosen arbitrarily. Better and more objective criteria may arise from analyzing the distribution of correct and incorrect answers as expected by chance. In order to calculate the distribution of correct answers and the difference between correct and incorrect answers (core) we propose the use of a method based on a gaussian distribution. The distribution of scores expected by chance is approximated by a gaussian distribution with a mean of zero and a standard deviation SD = square root of n(pA + pE), and the distribution of the total number of correct answers has a mean of npA and SD = square root of npApE, where n is the total number of questions, and pA and pE are the probabilities of having a correct and an incorrect answer, respectively. The formulae are applicable to questions type false/true/do not know and to the more common type of one correct in five options. Once the chance distribution is known, it can be compared with the distribution of scores or correct answers obtained, which can then be used to separate people in two groups: those that answer the test as expected or worse than expected by chance, and those that answer the test better than expected by chance. The first group should not be passed. The passing of individuals in the second group can be decided by additional criteria.