Probabilistic testing involves having the examinee assign probabilities to each of the options of a multiple-choice item. These probabilities reflect the student's perception of the correctness of each option. What is presented in the paper is a rationale for probability testing, the current theoretical and empirical findings, and some suggested directions for further research. The rationale given for considering probabilistic testing includes the following points. First, testing involves making decisions under uncertainty as do many situations faced every day and as such should be solved by using a subjective probability decision theoretic paradigm. Second, using multiple-choice testing situations may be a good way of teaching the subjective probability decision theoretic paradigm. Third, probability testing procedures should lead to more reliable and possibly more valid tests. Fourth, probability testing in conjunction with specific utility functions yields a way of incorporating and handling "risk" and "guessing" behavior in testing situations. An annotated bibliography is also included to introduce potential researchers to the general area of confidence testing. (Author)