Rasch Analysis of Student Responses to the Colorado Learning Attitudes about Science Survey
MetadataShow full metadata
The Colorado Learning Attitudes about Science Survey (CLASS) was developed by University of Colorado, Boulder. It investigates students’ beliefs about physics and learning physics1, and has become one of the most popular attitude surveys used in physics. Beginning in 2010, Dr. David Donnelly, Eleanor Close and Hunter Close started using CLASS to investigate students’ change in attitude during introductory physics courses at Texas State University. They used the standard analysis developed by the creators of CLASS. In their study, students’ attitudes did not display significant change during any semesters. However there was significant change during the winter break. This effect is called Winter Break Effect1. The credibility of a survey conclusion is closely related to the quality of the survey and its analysis method. By examining the CLASS and its analysis method, these researchers quickly noticed some problems: the standard analysis only partially counts a student’s responses, and the difficulty of each statement is not weighted when determining this student’s score. This study reanalyzed the data from the aforementioned study by using the Rasch model, which makes full use of the responses from all five survey categories. The results obtained support the existence of the Winter Break Effect, and also found that an instructor significantly influenced students’ attitudes toward physics during her class. In addition, this study compared the Rasch Analysis and the standard CLASS analysis. The Rasch Analysis exhibited advantages over the standard analysis. Moreover, this study evaluated the design of CLASS and provided suggestions to improve its quality.