Abstract
The University of New South Wales (UNSW) Canberra is a unique environment where university education and military studies are combined. Entrance into UNSW Canberra is highly competitive whereby students must obtain the appropriate Australian Tertiary Admissions Ranking (ATAR) for university entrance as well as meet additional selection criteria to determine suitability to undertake military training. Since the introduction of the ATAR system in 2009, there has been much debate regarding the use of this single numeric figure, calculated using results from multiple subjects, to predict competency in particular subjects. In order to overcome this limitation of the use of the ATAR score to measure mathematics competency, UNSW Canberra has incorporated a mathematics diagnostic test prior to the commencement of their first academic year to identify ‘at risk’ students who may benefit from additional mathematics support and career guidance.
The current research focuses on reviewing student performance on the Mathematics Benchmarking Quiz (MBQ) and end of semester mathematics and engineering mathematics results from the period 2014 to 2016. Of particular interest is investigating the psychometric properties of the MBQ, examining the relationship between ATAR scores and diagnostic test results, understanding state-based differences in mathematical performance and evaluating the success of a mathematics support program. Key findings revealed that the MBQ is a robust measure of general mathematical ability which holds up to the assumption of uni-dimensionality and displays adequate reliability. Additionally, while the MBQ is a good predictor of mathematical performance, there are other factors that influence results; of note, Queensland students included in the current study appear to be underperforming when compared to their counterparts. The support program put in place for ‘at risk’ students appears to assist in increasing both general mathematical knowledge and end of semester results. Results of the current research add to the wider literature on the use of diagnostic tests, and the potential limitations of the use of ATAR scores in performance prediction.