Well, I reported a few days ago on the impressive increase to the mean TOEFL score found in the data released by ETS. I expressed some puzzlement at the increase, as it is pretty huge. I’m still not entirely certain why it happened, but after talking it out with some experts, my conclusions are:
- The change is mostly due to the shorter test. I guess the shorter version is “easier.” While the reported mean score did not change in 2019, that was partly because of rounding, and a drop in the mean writing score. If we look carefully there were fractional increases in 2019 which hint at a trend.
- ETS may have adjusted the e-rater which scores essays. That’s a normal thing. I think they are on iteration 19 or something like that. I suspect that caused writing scores to increase. That makes up 25% of the overall increase… but the shorter test should have no effect on it. Perhaps they wanted to address the long-term drop in average writing scores.
- The increase is caused in large part by China (presumably the number one TOEFL market) and Korea (presumably the number two TOEFL market). Increases in the mean score probably reflect advances in preparation techniques in those countries. Coincidentally I spent the month before the score data release reporting on those advances.
Let me know what you think in the comments.