Construction and Evaluation of an Item Bank for an Introductory Statistics Class: A Pilot Study

Abstract

The purpose of the present study is to construct and evaluate the applicability of an item bank for an Introductory Statistics class. The participants of the study were 54 college students enrolled in Statistics in Psychology and Education in the spring and fall semesters of 2010. To establish the test bank, the authors first adopted and re-wrote the items from the midterm and final tests from the previous school year. Students practice the test in advance of the class on the instructional platform, Moodle. A total of 15 units were prepared with 45 items. The results revealed that (1) the point-biserial correlation of 34 items (75.6 %) reached .25, meaning the test items constructed for this study had enough discriminatory power; (2) 80 % of the Rasch item difficulty values ranging from −1 to 1, indicating an appropriate difficulty of the test bank, which not only met the students level, but was also appropriate to help students with preparation for the unit. Students’ achievement performance on the course correlated positively with the number of times they took the preparation tests, suggesting that the implementation of the lesson preparation activity enhanced the learning effectiveness of statistics.