English Language test as predictor of success in Medicine Programmes: A Cohort study at University College Dublin 2008- 2010.

AIM Non-EU students applying to study medicine at UCD, whose first language is not English, are required to present an English language competency score alongside second level exam results. IELTS (International English Language Testing System) is the most commonly presented language competency test and consists of an overall score and four individual scores in reading, writing, speaking and listening. This study sought to consider the relationship between IELTS score and academic medical programme performance. METHOD Data was collated for all students who entered the medical programme from 2008 to 2010 and undertook Year One, Two and Three examinations during this period (n=591). Non-EU and EU student performances were compared and correlations between non-EU entry criteria (second level exams, IELTS scores) and outcome measures (module grade point, individual assessment score and semester, year and degree GPA) were analysed. Module assessments were categorised as follows: written exam -short answer questions (SAQ) and/ or essay questions; written exam-multiple choice questions (MCQ) and/or extended matching items (EMQ) and other assessments including practical and spotter exams. As most of the non-EU cohort had generally only completed a half of their programme with UCD before leaving to complete the final half in Malaysia, only the first 2.5 years of the programme were considered. RESULTS Lowest IELTS band scores were in the writing and speaking components. Although performance between IELTS students and native English speakers was comparable in the first year of their programme, IELTS students were on average 0.2 of a GPA weaker in the subsequent two years. This GPA difference, statistically significant in Years Two and Three, was equivalent to a modest performance difference of a fractional grade e.g. B- to a B. IELTS overall and writing band scores correlated to year GPA’s with r values ranging from 0.252 to 0.428. Similar correlations were evident for module grade point scores. Furthermore, a positive statistically significant relationship was found between IELTS writing scores and performance in over 80% of the programme’s writing assessments (r values ranging from 0.151 to 0.341). Reading scores, which were correlate with the Year Three GPA (r=0.389). were found to have a weak positive relationship with the Year Three MCQ/EMQ assessments (r values from 0.151 to 0.179). Analysis of Year GPA means, depending on whether students scored 6.0 or less versus 6.5 or more in the written band of IELTS, once again revealed a statistically significant fractional grade difference. CONCLUSION Although the study has shown that performance in University examinations is correlate with IELTS score, the mean difference in student performance is modest. In order to ascertain whether the current IELTS admission criteria are appropriate, the study will need to be extended to students who take all five years of their programme in UCD and other Irish Medical Schools.