Reliability and Validity of Colorado Learning Attitudes about Science Survey (CLASS)

Abstract

This study aimed to propose revisions in the psychometric properties of Colorado Learning Attitudes for Science Survey (CLASS) based on the evidence found from this study. CLASS is a 42-item instrument that measures students’ attitudes towards physics and learning physics. In determining the validity of CLASS, 227 senior high school students from different schools in Metro Manila and  in Cavite, Philippines participated in the study. Statements in CLASS underwent exploratory factor analysis through principal factor analysis to determine its factor structure. Results showed that from the original 42 items in CLASS, 13 items did not reach acceptable correlation coefficient and factor loadings, thus they were deleted. From the remaining 29 items, four categories emerged: Personal Interest and Real-World Connection, Sense Making/Effort and Problem Solving, Conceptual Connections, and Applied Conceptual Understanding, compared to the eight categories from the original version CLASS. Specifically, the four factors were composed of the following number of items: 8 for Personal Interest and Real-World Connection, 7 for Sense Making/Effort and Problem Solving, 8 for Conceptual Connections, and 6 for Applied Conceptual Understanding. Reliability also reached acceptable overall Cronbach’s alpha value of 0.745. This study proposes this revised 29-item instrument of CLASS as a valid instrument which measures students’ attitude in the areas of personal interest, real world connection, personal effort and approaches in a physics course, and to problem solving.