Journal of Statistics Education, V7N3: delMas Appendix C

A Model of Classroom Research in Action: Developing Simulation Activities to Improve Students' Statistical Reasoning

Robert C. delMas and Joan Garfield
University of Minnesota

Beth L. Chance
University of the Pacific

Journal of Statistics Education v.7, n.3 (1999)

Copyright (c) 1999 by Robert C. delMas, Joan Garfield, and Beth L. Chance, all rights reserved. This text may be freely shared among individuals, but it may not be republished in any medium without express written consent from the authors and advance notification of the editor.

Appendix C: Multivariate Analyses of Variance

Separate multivariate analyses of variance (MANOVA) were conducted for scores based on the three different reasoning criteria. Each analysis consisted of two within and one between factors: item by test (pretest vs. posttest) by group (initial vs. new activity). Analysis of scores under the Correct criterion produced a significant three-way interaction [F(4, 912) = 3.22, p = .012]. Separate t-tests which compared the two groups for each item found no statistically significant differences between the initial and new activity on the pretest. All differences on the posttest were statistically significant at = .05. The MANOVA also produced a statistically significant interaction between group and test [F(1, 228) = 36.72, p < .001]. While the difference in the amount of change from pretest to posttest may depend on a particular item, overall the results suggest that the new activity group outperformed the initial activity group on all five items with respect to choosing the correct answer on each item.

Similar results were found for the Correct or Good criterion. A significant three-way interaction [F(4, 912) = 4.25, p = .002] was found. Separate t-tests produced no statistically significant differences between the initial and new activity on the pretest items. Differences between the two groups on the first four posttest items were statistically significant at = .001. Consistent with the t-test results, the MANOVA produced a statistically significant interaction between group and test [F(1, 228) = 32.84, p < .001]. With the possible exception of the fifth problem, the overall results suggest that the new activity group outperformed the initial activity group on the posttest with respect to choosing either a correct or good response for each item.

Analysis of scores under the third criterion (correct, good, or larger to smaller choice) did not produce a significant three-way interaction. The two-way interaction of group by test was statistically significant [F(1, 228) = 12.02, p = .001]. T-test results indicated that while there was no difference between the two groups on the pretest [t(159.58) = 1.54, p = .125], the difference on the posttest was significant [t(164.78) = 4.88, p < .001]. Under all three criteria, posttest scores of students using the new activity were significantly higher than those of students who used the initial activity.

Return to delMas, Garfield, and Chance Paper