Students' difficulties in practicing computer-supported data analysis: some hypothetical generalizations from results of two exploratory studies

Abstract

In this paper, I will report and summarize some preliminary results of two ongoing studies. The aim is to identify problem areas and difficulties of students in elementary data analysis based on preliminary results from the two ongoing studies. The general idea of the two projects is similar. Students took a course in data analysis where they learned to use a software tool, used the tool during the course, and worked on a data analysis project with this tool at the end of the course. The course covered elementary data analysis tools, such as variables and variable types, box plots, frequency tables and graphs, two-way frequency tables, summary measures (median, mean, quartiles, interquartile range, range), scatterplots, and line plots. The grouping of data and the comparison of distributions in the subgroups defined by a grouping variable was an important idea related to studying the dependence of two variables. The methods for analyzing dependencies differed according to the type of variables: for example, scatterplots were used in the case of two numerical variables, and two-way frequency tables and related visualizations were used in the case of two categorical variables. I have been interested in students' knowledge and competence in using the software tool for working on a data analysis task. For this purpose, students were provided with data and given related tasks. The two studies differed in their basic design. In the " Barriers project, " students were directly interviewed with regard to the data with which they were familiar from the course and which they had used as basis for a class project. This design allowed the researchers to focus on preconceived problem areas. In the "CoSta project," students were allotted approximately one hour for working in pairs on the data and the task before interviewers entered and discussed the results of their inquiry with them. This design provided more room for exploration of the data by the student pairs. However, the subsequent discussion was very dependent on the students' results. In both studies, the interviewers adopted a tutorial or teacher role to an extent that was not intended in the interviews' original design. The Barriers project is a collaborative project between C. Konold (University of Massachusetts, Amherst) and H. Steinbring (University of Dortmund, Germany). The students involved were 12th graders at an American high school who had completed a statistics course that used the software DataScope (Konold & Miller, …