A PSYCHOMETRIC INVESTIGATION OF THE CONCEPTUAL STRUCTURE OF THE REVISED NATIONAL SURVEY OF STUDENT ENGAGEMENT (NSSE 2.0)

Abstract

Student engagement plays an important role in student learning and success in elementary through postsecondary education. There has been growing interest in student engagement and its influence on student persistence, retention, and overall success in education over the past several decades. The importance of engagement has led researchers and university administrators to seek out measures of engagement that will provide information for improving undergraduate student success and that provide a means of assessing educational quality. The National Survey of Student Engagement (NSSE) was developed in the hopes of providing valid and reliable information about institutional quality and student engagement, and is administered to hundreds of thousands of freshmen and seniors each year at over 500 institutions. In 2013, the NSSE underwent a major revision; however, the psychometric properties of the revised instrument (i.e., NSSE 2.0) have not been thoroughly investigated and validated for the instrument’s intended uses and score interpretations. This study addressed the dearth of research investigating the psychometric properties and validity evidence of the revised NSSE 2.0, including the validity of the proposed conceptual structure for measuring student engagement. The findings of the current study partially support that the NSSE 2.0 has adequate psychometric quality to make confident decisions utilizing the NSSE engagement indicator scores at the institution level. The Academic Challenge and Learning with Peers themes were deemed adequate psychometrically for their intended use; however, the findings for the Experiences with Faculty theme were inconclusive and the Campus Environment theme was lacking in psychometric quality. At the item level, the majority of the engagement indicator items appeared to be working well based on item parameter estimates and information functions, though there was local dependency among items of the engagement indicators. Both the CFA and IRT results indicated that there was a multidimensional or higher order nature to the NSSE engagement indicators. However, due to mixed results across engagement indicators and themes and issues with goodness of fit of numerous models, the NSSE’s suggested model structure of the construct of student engagement with ten engagement indicators organized into four themes was not fully confirmed.