Development of an Early Alert System to Predict Students at Risk of Failing based on their Early Course Activities

Abstract

The emphasis on increasing student retention and graduation rates at institutions of higher education is driving the need for creation and implementation of early alert systems. Such early alert systems could be used in identifying students in academic trouble before failure. Early identification of students who are at a risk of dropping or failing a course will help instructors to adapt their course delivering techniques with student's learning styles and improve overall performance of a class. This paper discusses an early alert system to identify students who are at risk of failure based on their activity at the beginning of semester. The proposed alert system considers various indicators, including the homework assignments and the mid-term exam corresponding to the first quarter, along with in-class participation as input parameters. Data collected in large sections of Mechanics of Materials course over four semesters were employed for development and validation of the early alert system. The data analysis showed that the proposed model is capable of predicting the final scores of the students with an acceptable accuracy (R2=0.69). Feasibility of using the model was also validated using over 100 additional data points, which were randomly selected from the initial dataset. Good correlation was observed between the data and model predictions, with over 94% of the data points falling within the limits of a 90% confidence interval. The proposed model has possible implications in the similar engineering courses provided that the required data are collected during early semester activities. This tool enables the instructor to detect and reach out to the at-risk student and provide proactive assistance, so that they are able to succeed in the course. Proactive assistance may include referrals to appropriate resources, providing tailored activities to improve the weakness of students and one-to-one academic skill building workshops.

Meeting Name

124th ASEE Annual Conference and Exposition (2017: Jun. 25-28, Columbus, OH)

Department(s)

Civil, Architectural and Environmental Engineering

Keywords and Phrases

Academic assessment; Early alert system; Engineering course; Prediction model; Regression analysis

International Standard Serial Number (ISSN)

2153-5965

Document Type

Article - Conference proceedings

Document Version

Citation

File Type

text

Language(s)

English

Rights

© 2017 American Society for Engineering Education (ASEE), All rights reserved.

Publication Date

01 Jun 2017

This document is currently not available here.

Share

 
COinS