The BestMyTest TOEFL learner performance data is a free dataset about how our learners do in TOEFL practice. It counts their answers, reviews, and recordings, and it shows where they gain and lose points. Our TOEFL guide pages quote its figures. You can quote them too: the data is under a CC BY 4.0 license, so you only need to name BestMyTest as the source.
The data covers TOEFL practice on BestMyTest from January 2016 to October 2026, on the previous test format and the current one. It holds these figures:
Each guide page gives the exact count behind each figure it quotes.
Every figure is a count or an average of real practice on BestMyTest. We recompute the data every three months. A question-type percentage is the number of correct answers divided by the questions answered. A criterion figure is the average score our tutors or our AI gave, on the scale named next to it. Band figures convert each 0 to 30 section score to the 1 to 6 scale with the table our score report uses. Reading and listening figures combine the previous format, up to January 2026, and the current adaptive format. Select in Passage figures start in July 2026, after we fixed how it was scored. Recordings with no clear speech are left out. The data holds no names, answers, or audio, only counts and scores.
The figures and the anonymous rows behind them are published as files for researchers: one summary file and nine CSV files, such as tutor writing scores and practice-test section scores. Each row is one answer, review, recording, or test section, with its date and scores.
Name the source like this: BestMyTest TOEFL learner performance data, www.bestmytest.com, updated October 2026. The license is CC BY 4.0, so you may share and adapt the figures when you give this credit.