An Application of Predictive Modelling of Undergraduate Student Intake - Research and Statistical Modelling

Sep 5, 2026

Student load management in Australian universities is changing rapidly, making reliable load prediction crucial for planning. This paper examines how improving prediction methods can reduce institutional risk, focusing on the modelling of multiple components of commencing undergraduate EFTSL. Using real historical data, it investigates and evaluates statistical approaches for predicting commencing undergraduate student load based on application information, addressing accuracy, appropriateness of methods, and the uncertainty challenges posed by evolving funding and demand dynamics.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages21

Reading Time53 min

AccessFree to download

Abstract

Introduction

Background

An Application of Predictive Modelling of Undergraduate Student Intake Author: Ewa Seidel Senior Information Analyst (Statistics) Planning Services Unit Flinders University [Email:](Email: ewa.seidel@flinders.edu.au)[ ](Email: ewa.seidel@flinders.edu.au)[ewa.seidel@flinders.edu.au](Email: ewa.seidel@flinders.edu.au) Abstract Processes around student load management within Australian Universities are rapidly evolving. The efficacy of student load predictions has significant consequences on planning. Accordingly, institutions are increasingly interested in improving load prediction methods and understanding their accuracy. Flinders is seeking to develop better methods of predicting load that reduce risk. This paper discusses alternative statistical modelling methods and their appropriateness for use in predicting multiple components of student load. Through the application of real historical data, this paper specifically investigates and evaluates a method of predicting commencing undergraduate student load, based on applications. Introduction In response to recommendations made in the Bradley Review of Australian Higher Education ( Professor Denise Bradley, Noonan, Nugent and Scales 2008) , the Australian Government is aiming to increase the rate of higher education attainment. By 2025, the Government hopes to reach the target of 40 per cent of all 25 to 34 year olds holding a bachelor level qualification or higher (Australian Government 2009) . In order to achieve this goal, Commonwealth Grant Scheme (CGS) funding will be based on demand, will have no cap. Naturally, universities will be more competitive in the new environment. While Flinders University may choose to increase the resulting offers for some courses, this may not necessarily translate to an increase in load for all courses, due to the increased competitiveness. Accompanying such shifts in student load is a likely increase in the associated uncertainty, necessitating development of accurate and more precise student load projections to base robust planning mechanisms on. This paper will focus on CGS load due to it corresponding to the majority of total student load and its changing funding structure. Commencing and continuing equivalent full time student load (EFTSL) are generally predicted separately. Commencing load, the focus of this paper, corresponds to students who enrol for the first time in a particular higher education course at a particular education institution, while continuing students include those who are not commencing ( DEEWR 2011) . Student load is dynamic and complex in nature, the details of which are frequently sought and are of great importance. Load forecasts are first generated when prospective students apply for higher education. Early forecasts are used to estimate student demand and, for example, to determine how budgets will be distributed across academic areas. This information is particularly important since it is also used to understand and plan for teaching and room resources required to meet demand. Due to their great importance, early forecasts will be the primary focus in this paper. Background Flinders University has an emphasis on research and provides higher education to over 17,000 domestic students mostly from the southern regional Adelaide population, and almost 4 ,000 International students from around 100 countries. The University has a data warehouse which stores demographic, enrolment, admission and academic progression information about all current and previously enrolled students. All students who wish to enrol in higher education based on their Australian Tertiary Entrance Ranking (ATAR) scores achieved in year 12 apply through the South Australian Tertiary Admissions Centre (SATAC) . Each applicant can include up to 6 courses as ranked preferences , and is asked to provide detailed personal information with their application. SATAC also stores ATAR scores and details of offers when they be

FAQs

Why is predictive modelling of student load important for planning at universities?

Because load predictions directly influence budgeting and resource planning, and the paper highlights that uncertainty and risk increase in the evolving funding environment.

What type of student load does the paper focus on?

It concentrates on commencing load, defined as students enrolling for the first time in a course at an institution, modeled separately from continuing load.

What data source is used to predict commencing undergraduate student load?

The study applies real historical data and specifically investigates predicting commencing undergraduate load based on applications.

Sep 5, 2026

Uploaded by Alex Sinclair