The study investigated the comparison of the difficulty and discrimination incides of three multiple choice tests using the confidence scoring procedure (CSP). The study was also set to determine whether or not the difficulty and discrimination indices would be improved, if the tests were scored by the confidence scoring procedure. Two null hypotheses were generated and tested at 0.05 significance level. The population consisted of all SS2 Secondary School Students in Gbonyin local government of Ekiti State in Nigeria. A sample 450 students was selected using the purposive and proportional sampling techniques. Questionnaires containing 50-items achievement test in mathematics (in three formats) were constructed and administered on the students. The three multiple-choice tests were mixed and the testees did not know that they were answering different forms of the same test. The test scripts were scored using the confidence scoring method. Data were analyzed using the student t-tests and ANOVA. The result revealed that the contribution of blind guessing to testees was not directly related to the discrimination and difficulty indices of the three multiple-choice tests used. The confidence scoring procedure improved significantly the difficulty index of multiple-choice tests but did not significantly improve the discrimination index of three-index test used. The result showed that confidence scoring procedure reward partial knowledge of testees on the multiple-choice tests. It was recommended that confidence-scoring procedure should be encouraged for scoring multiple-choice tests, it discourages guessing. (Contains 7 tables.)