Language Testing Journal published an article (in Open Access) this week by Khaled Barkaoui of York University. It explores the relationship between ELP test scores (IELTS and TOEFL, specifically) and academic achievement. I’ll preface my summary by noting that when I shared this on LinkedIn yesterday there was some blowback in the comments and shares. Some people took issue with the conclusions. As for me… I’m not an academic.
Anyhow…
Barkaoui studied data for 6481 undergrads at York who were admitted to York with a TOEFL or IELTS score between 2009 and 2019. That includes 5215 students who submitted an IELTS score and 1266 who submitted a TOEFL score (which is an interesting bit of data on the market shares enjoyed by each test).
A few things were observed about how test scores hint at semester 1 GPA. To quote the author:
- “[Multilevel modeling] results indicated that IELTS overall scores exhibited a significant and positive association with Semester 1 GPA, whereas the TOEFL total score did not.”
- “Correlation analyses also indicated that IELTS scores exhibited stronger associations with Semester 1 GPA compared to the TOEFL iBT, irrespective of study major. The correlation between TOEFL scores and Semester 1 GPA was negative and weak. On average, a one-band increase in IELTS total score was associated with a half-point rise (0.521) in Semester 1 GPA, while a 10-point increase in TOEFL total score was, on average, associated with a hundredth point (0.01) increase in Semester 1 GPA.”
The author also explores differences in the trajectories of GPA experienced by submitters of each test. Both experienced the typical “U-shaped” GPA trajectory, though the TOEFL submitters experienced a slightly higher initial drop in GPA. Higher IELTS scores were associated with greater academic resilience over time, while TOEFL scores “did not exhibit such a significant interaction with time.”
In concluding, the author notes:
“Taken together, these results suggest that IELTS is not only more effective at predicting academic achievement at the start of university studies (i.e., Semester 1), but can also provide insight into the trajectory of academic achievement over time. This predictive ability can be used to offer students targeted support throughout their academic journey. In contrast, TOEFL iBT scores do not appear to offer similar predictive information in the context of the university where the current study was conducted. This suggests that IELTS scores can be a more useful tool for planning English language support for international students.”
But importantly, he reminds us that:
“Finally, the study highlights the limited predictive power of ELP test scores. Collectively, the two tests only explained a modest percentage of the variance in Semester 1 GPA. IELTS scores exhibited the highest explanatory power, while the TOEFL explained a very small portion of GPA variance. This finding does not mean that ELP plays a negligible or no role in university academic achievement; rather, it underscores that factors beyond language proficiency, such as student motivation, social and academic support, and teaching quality, play a more substantial role in determining academic achievement (cf. Hill et al., 1999). Consequently, universities and researchers should take these factors into consideration when making admission decisions and examining the relationships between ELP and academic achievement.”