Applying Rasch Measurement to Validate the Colorado Learning Attitudes About Science Survey

• In Event: Equity in Science Assessment and Measurement

Abstract

We collected data from 728 high-school students to revalidate the Colorado Learning Attitudes about Science Survey (CLASS) using the Partial Credit Rasch Model (PCM). Results suggest a ceiling effect and a number of items are over easy to differentiate students. The Likert Scale is not equal interval, as evidenced by half of the items with an invalid category (i.e., interval is zero). We suggest that the identified issues to be fixed for future use of CLASS. PCM detected three distinct item clusters, with two of them highly correlated to each other. CLASS had significant DIF on 11 items between male and female students, which is not unexpected as research has shown that physics learning attitudes is distinct by gender.

Author

• Lehong Shi, University of Georgia