The purpose of this paper is to evaluate the psychometric properties of The Colorado Learning Attitudes about Science Survey (CLASS). The 50-item instrument was administered to 311 college students from a public institution in the United States enrolled in General Chemistry I Laboratory. Confirmatory factor analysis and Cronbach's [alpha] estimates were used to examine the reliability and validity of the CLASS instrument scores. Detailed information regarding this process is discussed. Initial results revealed a lack of evidence for simultaneous measurement of all of the constructs underlying the nine CLASS subscales. However, further work, including the administration of two small, single-scale instruments taken directly from CLASS, suggests that concise, easy-to-administer measures of individual constructs can be derived from CLASS. (Contains 8 tables.)