A comparison of methods used to teach probability in a university elementary statistics course

This study examined the effects of the use of computer simulations coupled with a traditional teaching approach on students' learning of probability and student confidence in a university elementary statistics course. Specifically, achievement and attitudinal scores of students enrolled in elementary statistics courses taught using traditional probability were compared to the achievement and attitudinal scores of students enrolled in elementary statistics courses taught using traditional probability instruction supplemented with computer simulations.;Subjects were students enrolled in Math 115-Elementary Statistics at The University of Mississippi during the spring semester of 2001. A sample of 342 subjects---158 freshmen, 68 sophomores, 69 juniors, 46 seniors, and 1 graduate student---agreed to participate in this study. Subjects in the treatment group received instruction supplemented with computer simulation models for a time span of four weeks. Subjects in the control group received traditional instruction. Pretests and posttests measuring students' knowledge of probability were administered. The Confidence in Learning Probability Scale was used to measure attitudinal difference between groups.;An analysis of covariance was conducted on the posttest means of knowledge of probability. The analysis of covariance yielded significant results. However, the mean posttest score of the control group exceeded the treatment group. A dependent samples t-test was conducted to assess significance of gains from pretest to posttest. Posttest scores for the treatment group were significantly higher than pretest scores. Likewise, posttest scores for the control group were significantly higher than pretest scores. A factorial analysis of variance was conducted on the posttest means to examine interaction between method of instruction and prior knowledge of probability. Results of the factorial analysis of variance showed no significant interaction between method of instruction and prior knowledge of probability. A Mann-Whitney U test yielded no significant difference in attitudinal scores between the control and treatment groups.