File(s) not publicly available

Sectionality or why section determines grades : an exploration of engineering core course section grades using a hierarchical linear model and the multiple-institution database for investigating engineering longitudinal development

conference contribution
posted on 06.12.2017, 00:00 by G Ricco, N Salzman, R Long, Matthew Ohland
Grades, how they are earned, and the institutional impetuses that drive them, are an issue of central importance in the engineering discipline. (1-4) How grades are earned, how different institutions address grades and grade inequities, how instructional practices and policies affect grades, and other grading notions have been studied widely in engineering education. (5-8) The effect of faculty on student grades, while studied, (9) has not been probed as extensively within engineering education using a hierarchical linear model (HLM).One of the great, open questions in engineering education is whether or not the section makes a difference in a student’s grade. In other words, the effect of sectionality on grades to a large extent is unknown. Sectionality combines instructor effects, effects related to time-of-day of instruction, effects related to any tendency for students to coordinate their enrollment, and other effects. Experience and anecdotal evidence suggest that sectionality affects grades, but large-scale empirical studies of this phenomenon do not exist. Due to the inherent structured nature between course sections and students, standard linear regression models do not offer a robust solution to probing longitudinal systems containing multilevel variables. Hierarchical Linear Models (HLMs) provide a robust solution to studying nested or hierarchical systems when compared with standard regression techniques. We constructed a simple HLM to probe inter-section and intra-section variability in grades within the Multiple Institution Database for Investigating Engineering Longitudinal Development (MIDFIELD) by the calculation of intraclass correlation coefficient (ICCs). (10, 11) We then examined grades from three sets of courses endemic to the first year engineering experience: the first chemistry course; the first calculus course; and the first physics course. Our preliminary results indicate that the choice of a HLM to analyze our longitudinal database is correct due to strong variability in grades explained by the high intraclass correlation coefficient (ICC) for most of our MIDFIELD institutions across all three course types analyzed.

History

Start Page

5861

End Page

5874

Number of Pages

14

Start Date

01/01/2012

ISBN-13

9781622761913

Location

San Antonio, Texas, USA

Publisher

American Society for Engineering Education

Place of Publication

Washington DC

Peer Reviewed

Yes

Open Access

No

External Author Affiliations

Central Queensland University. QCQU; Conference; Purdue University;

Era Eligible

Yes

Name of Conference

American Society for Engineering Education. Conference

Exports