Recent changes to the TOEFL
Educational Testing Service (ETS) has revised the TOEFL a number of times since the exam was first offered in 1964. The creation of the TOEFL internet-based test (iBT), which became available to students in 2005, improved the TOEFL assessment in a number of ways. The goal of the iBT was to evaluate functional English capabilities in native speakers of other languages. A speaking section was added to ensure that international students were in possession of a skill crucial to the successful navigation of university life, and ineffective sentence completion exercises were replaced with assessments of written English that required understanding of context. Internet testing also allowed the TOEFL to be offered on a greater range of dates and in a larger number of locations, and the iBT's combination of electronic and human grading eased the workloads of ETS staff members. The TOEFL iBT can be considered one of the most substantial improvements in standardized testing in recent memory, and international students benefit from an exam that much more effectively demonstrates their suitability for university study in the English-speaking world.
A new version of the TOEFL iBT reading section was introduced on November 1, 2011. The total number of included reading passages was reduced from five to either three or four. TOEFL administrations that include three reading passages allow test-takers 60 minutes to complete the associated questions, while students are given 80 minutes if their TOEFL features four reading passages. Prior to 2011, the iBT reading section was divided into separately timed parts, but the entire section is now continuously timed and students may use their allotted 60 or 80 minutes however they wish. ETS cited user feedback as the reason for these changes, which permit students to use a greater range of test-taking strategies. Many TOEFL educators unaffiliated with ETS agree with this assessment of the latest iBT reading section. Continuous timing allows test-takers to devote more time to reading passages that they find more difficult, and students may also skip back and forth among reading passages, change answers, and delay responses to the most challenging questions. ETS did not report any changes to the reading passages or questions themselves, and there is no external evidence that the reading section revisions concern any issues other than timing and number of passages. Preparation is thus affected only in terms of strategy; the necessary content knowledge and skills remain the same as before.
According to ETS, the mean iBT reading score for all test-takers was 20.1 in 2010, the last full year in which students took the now-discontinued version of the reading section. In terms of 2010 percentile rankings, the midpoint was somewhere between 21 (the 47th percentile) and 22 (the 52nd percentile). The changes to the TOEFL reading section do not seem to have had a significant effect on mean reading scores for the total group, which were 20.1 in 2013 and 20.3 in 2015. TOEFL reading percentile rankings for 2013 and 2015 were also largely consistent with those reported in 2010. Native speakers of some languages, however, may have benefited from the revisions. The average reading score for Korean speakers, for example, increased from 21 in 2010 to 22 in 2015, and primary speakers of Vietnamese and Indonesian improved by an average of two points within the same timeframe (from 18 to 20 and 19 to 21 respectively). Data from future testing years will help clarify the effects of TOEFL iBT revisions on specific groups of students and the overall set of test-takers.
Between the first administration of the TOEFL iBT in September of 2005 and the end of 2015, mean scores increased for three of the exam's four sections. Reading scores rose at the highest rate, from 19.1 in 2005 to 20.3 in 2015. Speaking scores, at 19.4 in 2005 and 20.3 in 2015, jumped nearly a full point. The increase in writing scores (20.1 to 20.6) was more modest, while listening scores declined by more than half of a point (from 20.5 to 19.9). These sectional averages add up to total scores of 79.1 in 2005 and 81.1 in 2015. As of 2015, the sum of mean sectional iBT scores for graduate school applicants was 84.8, compared to an average total score of 79 for aspiring undergraduate students. In the same year, applicants for professional licenses and applicants for employment using the TOEFL iBT as one of the qualifying criteria averaged total scores of 84 and 83 respectively. In the total group of 2015 test-takers, females outperformed males in terms of mean composite scores (82 to 80). 2015 ETS data also shows that among languages with a statistically significant number of test-takers, the highest TOEFL iBT scores were received by native speakers of Dutch (average total score of 100), Danish (98), and German (97). Language backgrounds from which students tended to struggle on the TOEFL included Japanese (average total score of 71), Arabic (70), and Kurdish (61).