Comparing exploratory factor models of the Brief Electricity and Magnetism Assessment and the Conceptual Survey of Electricity and Magnetism

Abstract

While numerous studies have analyzed the conceptions probed by the Force Concept Inventory (FCI), assessments dedicated to electricity and magnetism lack similar analyses. This paper investigated the conceptions explored by the Brief Electricity and Magnetism Assessment (BEMA) and the Conceptual Survey of Electricity and Magnetism (CSEM) using exploratory factor analysis techniques. Exploratory factor analysis (EFA) was performed on both assessments using 5368 and 4941 postinstruction student responses for the BEMA and CSEM, respectively. A 6-factor EFA generated model was found for the CSEM, and was fit against another sample of 4964 student responses using confirmatory factor analysis to supply evidence for the possible generalizability of the model. The 5-factor EFA generated model for the BEMA could not be fit against another sample when trying to check for generalizability. The EFA generated factor models for the BEMA and CSEM were then compared and found to be similar in conceptual content, with the exception of one or two factors. Thus, from a factor analysis perspective, the BEMA and CSEM were found to be similar in conceptual content as revealed by student responses. With a better understanding of these electricity and magnetism assessments, future research into this domain of physics will then be able to make stronger conclusions based on students' results within these assessments.

Assessing the longitudinal measurement invariance of the Force Concept Inventory and the Conceptual Survey of Electricity and Magnetism

“…The results suggested that the optimal factor models identified in their study were not generalizable. In Eaton et al's study, a six-factor model identified using EFA was found to be more general as it was also validated with another population sample using CFA [54].…”

“…In the original paper, an 11-factor model was established, which is too many for a 32-item CI. Recently, two additional studies have tried to analyze the latent factor structures of the CSEM [54,55]. Zabriskie and Stewart [55] explored the factor structures of the CSEM through a MIRT approach using two datasets from different institutions, which produced two different optimal factor models, a nine-factor model and an eight-factor model, for the two populations, respectively.…”

“…Regarding the factor structures of CSEM, there exists only one study, which identified an EFA driven 6-factor model labeled as Eaton6 shown in Table II [54]. Among the factors, F1 concerns students' understandings of action or reaction pairs in the context of electricity and magnetism.…”

“…For the three items associated with F5 in Eaton6, item 18 is excluded from the new model due to a negative estimate of the factor loading, and items 19 and 20 are reassigned into F5, which targets the concept of force caused by an electric field (as suggested by the creators of the CSEM [2]). In addition, items 16 and 17 are reassigned to F5, since these were loaded poorly on F2 in Eaton6 [54], and were also suggested by the creators of the CSEM as probing student understandings similar to those measured by the other items in F5 of Eaton5M [2]. With these changes, F5 in Eaton5M is conceptually interpreted as relations between electric fields, electric forces, and electric potentials, which is similar to the original F6 in Eaton6 but with a few more items included.…”

“…Item 24 is moved to F4 from the original F1 because it probes student understandings of electromagnetic interaction regarding the forces between two current-carrying wires. Items 16 and 13 are removed from the new F4 since they probe understandings of electrostatics instead of magnetism [54].…”