Confirmatory Factor Analysis of the BCSSE Scales

Abstract

The Beginning College Survey of Student Engagement (BCSSE, pronounced “bessie”) measures entering first-year students’ high school academic and co-curricular experiences as well as their expectations for participating in educationally purposeful activities during the first year of college. Since BCSSE’s launch in 2007, nearly 900,000 entering first-year students enrolled at 506 institutions have participated in the survey.

BCSSE was updated in 2013 to align with the updated version of NSSE. The new version maintained BCSSE’s focus on gathering information from entering first-year students regarding their high school experiences and their expectations for engagement during their first year in college. It also included new items to increase alignment with NSSE, improve clarity and applicability of survey language, and refine existing measures. The revision also included new items and nine scales (aka, BCSSE scales). One of the goals during the revision process was to assure that these scales have strong psychometric properties. Much of the development work was completed during the NSSE update process and included extensive student cognitive interviews, literature reviews, expert consultations, pilot testing, statistical analysis of pilot data, and interviews with administrators responsible for use of the BCSSE data on their campuses. BCSSE was updated again in 2019 to include transfer and delayed-entry students (older students with no prior college experience) for the online survey mode. This report only evaluates the psychometric properties of BCSSE scales using data from first-year students. A future psychometric report will include data from transfer and delayed-entry students. However, pilot data collected from transfer and delayed-entry students in 2018 did not indicate any deviation in scale internal consistency compared to first-year students.

Two statistical techniques were used to reevaluate the psychometric properties of the BCSSE Scales (See Cole & Dong, 2014). First, internal consistency reliability was investigated using Cronbach’s alpha. The second technique used a split-sample approach to identify possible factor structures and the second to test the identified structure.