In short
- The CEFR describes ability, a Lexile measure describes a text by sentence length and word frequency, and the GSE is a 10 to 90 scale that Pearson has aligned to the CEFR.
- Research on readability formulas shows that sentence length and word frequency predict difficulty only roughly, and for a language learner the unknown words decide comprehension.
- Pearson publishes its own GSE-to-CEFR table; the MetaMetrics documents cited here publish no Lexile-to-CEFR table, so none is offered.
- When a pupil arrives with a Lexile or GSE figure, record it, place the pupil with your own reading check, and never report the figures as if they were the same thing.
A pupil transfers in with a Lexile measure on her report. A coursebook promises GSE 43 to 50. Your school reports in CEFR levels. Teachers ask for a conversion table, and the honest answer is that no exact one exists. This article explains what each of the three scales measures, how each is produced, and what research says about readability formulas and second-language readers. It ends with a way to use a Lexile or GSE figure alongside a CEFR level without pretending they are the same.
The short answer: three scales, three different things
The CEFR, the Lexile Framework and the Global Scale of English measure different things, so there is no exact conversion between them. The CEFR (Council of Europe, 2020) describes what a learner can do, in six levels, with separate scales for each skill. A Lexile measure is a number for the difficulty of a text, and for a reader's ability, built from sentence length and word frequency (Lennon & Burdick, 2004). The GSE is a numeric scale from 10 to 90 that Pearson has aligned to the CEFR bands (de Jong, Mayor & Hayes, 2016). One is a description, one is a readability measure, one is a finer ruler laid over the first.
That is why a Lexile measure cannot tell you a pupil's CEFR level, and a CEFR level cannot tell you which Lexile band a pupil reads at. The GSE is the exception, because its owner built it to map onto the CEFR. Even there, the mapping is Pearson's own, and it describes ability, not texts. CEFR Levels A1 to C2 Explained covers what the levels themselves mean.
What does each scale measure, and how?
| Scale | What it measures | How it is produced | Good for | Limits for an English learner |
|---|---|---|---|---|
| CEFR | Language ability: what a learner can do, by skill, in six levels | Can-do descriptors scaled by the Council of Europe; a level is assigned by teacher judgement, a placement check or an exam aligned to the framework | Shared reporting across schools, coursebooks and exams; setting goals | Broad bands; no number for a text; two B1 judgements can differ |
| Lexile | Text difficulty, and a reader's ability, on one numeric scale ending in L | Text measure from mean sentence length and mean word frequency in a large corpus; reader measure from a reading test | Matching a reader to texts of similar difficulty; tracking growth in reading alone | Counts only sentence length and word frequency; its documentation speaks of readers and texts, not of language learners; says nothing about the grammar a learner has met, or about listening and speaking |
| GSE | Language ability as a number from 10 to 90, by skill | Learning objectives rated by teachers and calibrated with Rasch analysis; aligned to the CEFR by Pearson | Finer tracking inside a CEFR level; progress below A1 | A proprietary scale; the alignment is Pearson's own; still a description of ability, not a measure of a text |
The Lexile scale is a developmental scale for reading. Lennon and Burdick (2004), writing for MetaMetrics, say the Lexile measure is based on "two factors: Word frequency and sentence length, which are more formally called semantic difficulty and syntactic complexity". Word frequency here is not how often a word appears in the passage, but how often it appears in a corpus of nearly 600 million words. Both the reader and the text are placed on the same scale, and MetaMetrics describes a well-matched text as one the reader comprehends at about 75 percent.
The GSE is a different kind of object. De Jong, Mayor and Hayes (2016) describe it as "a standardised, granular English proficiency scale from 10–90" that is "psychometrically aligned" to the CEFR. Its learning objectives were rated by teachers and calibrated on a single scale with Rasch measurement. It exists because the CEFR levels are wide: the same paper notes that A2, B1 and B2 are each about twice as wide as A1 and C1.
Why is there no exact conversion?
Because the scales are answers to different questions. A CEFR descriptor asks: what can this learner do with a text? A Lexile measure asks: how long are the sentences, and how common are the words? A GSE score asks: how far along a calibrated set of can-do objectives is this learner? Three questions, three answers, and no formula links them.
Even within its own terms, a readability number carries error. Stenner, Burdick, Sanford and Burdick (2006), of MetaMetrics, estimate the error from the Lexile theory itself at 64L for a standard passage of about 125 words. For whole books the error shrinks to single digits, because the whole text is processed. They add that uncertainty in expected comprehension comes mostly from the reader measure, not from the text measure. In other words, a book's number is fairly stable; a pupil's number is not.
Then there is the question of what a formula leaves out. Crossley, Greenfield and McNamara (2008) compared traditional readability formulas with an index built from word frequency, syntactic similarity between sentences and content-word overlap, on texts used with second-language readers. The index predicted reading difficulty for those readers better than the traditional formulas. They conclude that variables tied to how readers process a text predict difficulty better than the surface counts of traditional formulas.
For a language learner, the deciding factor is the unknown words. Hu and Nation (2000) found that learners of English needed to know about 98 percent of the words in a story to read it without help. No readability formula knows which words your pupils have met. A text can have short sentences and common words and still be unreadable to a class that has never seen half of them. Natova (2021) built qualitative and quantitative tools for judging text complexity against the CEFR, and tested them with students at B1 and B2. Most of those students understood at least 90 percent of the vocabulary of the text at their level, which supports the tools at those two levels only.
What alignment does exist?
Pearson publishes a GSE-to-CEFR table in its technical paper, and labels the alignment as its own. De Jong, Mayor and Hayes (2016) give the ranges below. They define being at a CEFR level as having "at least 50% probability of being able to perform all language activities at that given level". A learner about 80 percent of the way through a level is likely entering the next.
| CEFR level | GSE range |
|---|---|
| Below A1 | 10 to 21 |
| A1 | 22 to 29 |
| A2 and A2+ | 30 to 35 and 36 to 42 |
| B1 and B1+ | 43 to 50 and 51 to 58 |
| B2 and B2+ | 59 to 66 and 67 to 75 |
| C1 | 76 to 84 |
| C2 | 85 to 90 |
For Lexile measures, the MetaMetrics documents cited in this article give no CEFR alignment, so this article offers none. Tables that pair Lexile bands with CEFR levels circulate online. Unless the owner of the scale publishes one, treat such a table as somebody's estimate, not a fact.
How to use a Lexile or GSE figure alongside a CEFR level
The rule is to use each figure for what it measures, and to let your own reading check decide placement. A Lexile measure can help you choose between two texts; it cannot place a pupil in a CEFR band. A GSE score can show movement inside a level; it comes with Pearson's alignment, which you may use as long as you say whose it is.
- Record the incoming figure as it came, with its scale and date. Do not convert it.
- Place the pupil with your own check: a short levelled text and a few questions, as in How to Assess a Student's English Reading Level.
- If the figure is a GSE score, read the CEFR band from Pearson's table and compare it with your check. If they differ, trust the check, and note the difference.
- If the figure is a Lexile measure, use it only to compare texts: a 600L book is easier than an 800L book by sentence length and word frequency. Then check the unknown words for your class, as in How to Match Students to the Right Graded Readers.
- Report in one system, the CEFR, and keep the other figure in the pupil's file for the next school or exam that asks for it.
Classroom example (an illustration, not a case study)
A boy of twelve transfers in with a report that says 720L. The English coordinator writes the figure in his file and gives him the school's reading check: a short A2 story, then a short B1 story, each with five questions. He manages the A2 story and about half of the B1 questions. She places him at A2+, records 'Lexile 720L (previous school, May)' beside the level, and reviews both after the first term.
If your school is moving from one system to another, run both side by side for one cohort for a term before switching the reports. You will learn how the two relate in your own population, which no published table can tell you. English Reading Placement Tests for Schools compares the tools you could use for the check.
Which scale for which job?
- Reporting to parents, boards and the next school: the CEFR. It is the shared language of exams and coursebooks, and a parent can read a can-do statement.
- Choosing between two English texts for first-language readers: a Lexile measure, if the texts have one. It compares sentence length and word frequency and nothing else.
- Choosing a text for a learner: a CEFR-levelled text, checked for unknown words. How Is a Story Levelled to the CEFR? explains what a careful levelling checks.
- Showing progress inside a level: plus levels and can-do checklists, or a GSE score if your coursebook or test already gives one.
- Estimating vocabulary size: none of the three. See How Many Words Do You Need at Each CEFR Level?.
How iRead handles this
iRead levels every story, quiz and game to the CEFR, from A1 to C2. Vocabulary and quizzes are mapped to CEFR bands, so a B1 story means the same thing whichever pack it comes from. Teachers see retention per word, per game and per class. School leaders see, by class, grade and term, the levels reached, the words retained and the stories completed.
The core principle: Use each scale for what it measures. The CEFR describes the learner, a Lexile measure describes the text, and the GSE is Pearson's finer ruler over the CEFR. None of them converts exactly into another.
See it in iRead: iRead levels every story, game and quiz from A1 to C2, so progress is reported in one system that parents, exams and coursebooks share.
See CEFR inside iReadKey takeaway
Record incoming Lexile or GSE figures, place pupils with your own reading check, report in CEFR levels, and use a Lexile measure only to compare texts, never to place a learner.
Frequently asked questions
Can you convert a Lexile measure to a CEFR level?
Not exactly. A Lexile measure is built from sentence length and word frequency (Lennon & Burdick, 2004), while a CEFR level describes what a learner can do. The MetaMetrics documents cited here give no CEFR alignment. Treat any Lexile-to-CEFR table you meet as an estimate, and place pupils with your own reading check.
What is the difference between GSE and CEFR?
The GSE is a numeric scale from 10 to 90 that Pearson has aligned to the CEFR bands (de Jong, Mayor & Hayes, 2016). The CEFR is the Council of Europe's description of ability in six levels. The GSE adds finer steps inside those levels; the alignment is Pearson's own.
What GSE score is B1?
In Pearson's own alignment, B1 runs from GSE 43 to 50 and B1+ from 51 to 58 (de Jong, Mayor & Hayes, 2016). Pearson defines being at a level as having at least a 50 percent chance of performing the activities at that level, so a score of 43 is a threshold, not a comfortable B1.
Does a Lexile measure work for English language learners?
It measures the text the same way for any reader: sentence length and word frequency. What it cannot know is which words a learner has met. Hu and Nation (2000) found that learners needed to know about 98 percent of a story's words to read it unaided, so check unknown words for your class rather than relying on the number.
Sources
- Council of Europe (2020). Common European Framework of Reference for Languages: Learning, teaching, assessment. Companion volume. Council of Europe Publishing, Strasbourg. rm.coe.int/common-european-framework-of-reference-for-languages-learning-teaching/16809ea0d4
- Lennon, C., & Burdick, H. (2004). The Lexile Framework as an approach for reading measurement and success. MetaMetrics white paper, April 2004. cdn.lexile.com/m/uploads/whitepapers/Lexile-Reading-Measurement-and-Success-0504_MetaMetricsWhitepaper.pdf
- Stenner, A. J., Burdick, H., Sanford, E. E., & Burdick, D. S. (2006). How accurate are Lexile text measures?. Journal of Applied Measurement, 7(3), 307-322. pubmed.ncbi.nlm.nih.gov/16807496/
- de Jong, J. H. A. L., Mayor, M., & Hayes, C. (2016). Developing Global Scale of English Learning Objectives aligned to the Common European Framework. Global Scale of English Research Series, Pearson, London. www.pearson.com/content/dam/one-dot-com/one-dot-com/pearson-languages/en-gb/pdfs/gse/gse-resources/gse-research-reports/developing-gse-learning-objectives-aligned-to-common-european-framework.pdf
- Crossley, S. A., Greenfield, J., & McNamara, D. S. (2008). Assessing text readability using cognitively based indices. TESOL Quarterly, 42(3), 475-493. doi.org/10.1002/j.1545-7249.2008.tb00142.x
- Natova, I. (2021). Estimating CEFR reading comprehension text complexity. The Language Learning Journal, 49(6), 699-710. doi.org/10.1080/09571736.2019.1665088
- Hu, M. H.-C., & Nation, P. (2000). Unknown vocabulary density and reading comprehension. Reading in a Foreign Language, 13(1), 403-430. doi.org/10.64152/10125/66973