The Business of Teaching Statistics as an Experimental Science: Or an Experiment in the Science of Teaching Business Statistics

Article excerpt


The results of a two semester experiment teaching business statistics as a computer lab based course, rather than a lecture based (or lecture based with lab component) course, shows that students show significant improvement in data interpretation and analysis at the cost of a slight degradation of probability technique performance. The evidence is from over 200 students (representing 3 lab sections and 1 control group section) on 5 exam questions and the course capstone project. Indirect assessments further suggest that students recognize the value of the lab approach both from the immediate rewards of deeper conceptual understanding to the longer lasting effects from future use of the techniques.


As instructors, we should be concerned with the results of our teaching (meaning student learning) rather than any perceived glory or status involved in the method of our teaching. I mention this as I believe that the headlong rush to incorporate technology into the classroom has not been as carefully considered as it should be. One approach to any change in teaching method is to conduct a scientific experiment, using a control group and careful analysis, on the students in such a course.

This paper describes a year long experiment teaching statistics as a (computer) lab science instead of as a traditional lecture based course. Over the course of my year as a Distinguished Teaching Fellow with UW System, I developed a project analyzing the learning goals and teaching methods of the Statistics course at the University of Wisconsin--Oshkosh (Econ 210). I conducted three sections of the course with a computer lab instead of a lecture only course. The lab occupied roughly a quarter of the class contact hours.

My findings are that students in the lab based course gained a facility with computer oriented statistical analysis and lost the trepidation to approach computer work and numerically oriented problems so often found even in students who successfully completed other statistics programs. Furthermore, the students from the lab based course did not show any degradation of skill with standard statistical techniques and methods. The lab students took away with them a very important job skill. They were able to recognize their new facility with computer oriented statistical analysis with very positive comments on class evaluations as well as with continuing feedback to the course instructor over succeeding semesters.


During 2000 three sections of our normal statistics course was converted from a lecture based course to a lab based course. There were a total of 13 sections of the course offered during that time. In the Spring of 2000, only one section was converted to a lab course and students were not apprized of the change until the course begun. Sufficient other sections were available for students who wished to transfer out if they felt strongly they did not want to participate in the experiment; two students did not complete the course in the Spring of 2000. Fall 2000 saw two lab sections offered, again without notations in registration material though student word of mouth may have biased student selection of these sections; there were no drops that term.

The statistics course offered in the Economics department serves both the Economics Department majors and minors but also all majors from the College of Business. Other departments in the College of Letters and Sciences offer their own statistics courses, such as Sociology, Psychology and Mathematics. Therefore, our student base is generally rather homogenous as to major and career aspiration. All students were undergrads and the non lab section has 35% females while the lab section has 34% females. This closely follows general enrollment in the Economics Department and in the College of Business here.

The course sections under consideration here, 3 lab based sections and 1 non lab (lecture) section were all taught using the same basic text, with the same professor, at roughly the same times of day, with the same lecture notes in an attempt to keep the control group as consistent with the experimental groups as possible. …