English Reading Placement Tests for Schools

Schools compare placement tests by length, price and the number of skills covered. None of those tells you whether the test works. A test works when the pupils it places stay where it put them, and when their teachers agree with the result two weeks later. This article explains what a school placement test should contain, how to check whether it is working, how to read a publisher's claims, and how to build your own from CEFR descriptors.

The short answer: judge it by the moves

A placement test's job is to put a pupil in the right group quickly. So the number that matters is how many pupils have to be moved afterwards. Hughes (2003) describes placement tests as identifying the stage of a teaching programme most appropriate to a learner. The most successful ones, he says, are constructed for particular situations: tailor-made, not bought off the peg. Buying a test is advisable only when the school is sure it suits its programme.

Wall, Clapham and Alderson (1994) list the questions to ask of any placement test. Does it correctly identify the pupils who most need the classes? Do the pupils feel their language was accurately measured? Is the content appropriate to the use made of the results? Is it reliable? Every question on that list can be answered inside your own school, with your own pupils.

What a good placement test for a school contains

• Levelled reading. Short texts at each level you group by, with a few questions each. A pupil's level is the highest text they handle with ease, not a total score across all of them.
• Vocabulary. A short recognition test, because it is fast to mark and gives a second reading on the pupil. How Many Words Do You Need at Each CEFR Level? gives the figures.
• A writing or speaking sample, where staff can mark it. Five sentences about a picture, or two minutes of talk, catches the pupil who reads well and cannot produce. Only include it if someone will mark it within the week.
• Items that match your groups. If you stream into A1, A2 and B1, the test must separate those three. Items at C1 cost time and tell you nothing.
• A result the same day. The test should fit one lesson and be marked before groups are set. Fulcher (1997) shows how administrative and logistic constraints shaped what a university placement test could do.

Length is not on that list. A short test whose result flows straight into a group beats a long test whose score sits in a spreadsheet.

How to check the test is working

First, the misplacement rate. Count every pupil moved to another group in the first four weeks, and record the direction. Second, teacher agreement. Two weeks after placement, ask each teacher to write a level for each pupil without looking at the test score, then compare. Bachman and Palmer (2010) frame test justification as an argument that links a score to the decision made from it and to the consequences of that decision. For a placement test, the decision is the group, and the consequence is whether the pupil can read what the group reads.

Wall, Clapham and Alderson (1994) include the test takers' own view among their questions. Add one line to the answer sheet: 'Was this test too easy, about right, or too hard?' A pupil who writes 'too easy' and lands in A1 is worth a second look.

Reliability in plain words

A test is reliable when the same pupil gets about the same result on two occasions, or on two versions of the test. A test can be reliable and still useless, if it measures the wrong thing. It cannot be useful without being reliable, because a result that would change tomorrow is not a result.

Published tests do not always show their reliability. Nilsson (2008) examined eight published informal reading inventories and found that only one gave enough reliability data to support the use of its alternate forms. Those were reading inventories for US schools, not EFL placement tests, but the lesson travels. The Council of Europe (2009) Manual is sharper. An examination with no procedure to keep markers equally strict, or successive forms equivalent, cannot make credible claims of any link to the CEFR.

So ask a publisher for the reliability figure and the sample it came from. Then run your own check: the misplacement rate.

What published tests claim, and how to read the claims

Many published tests say they are 'aligned to the CEFR' or 'report a CEFR level'. The Council of Europe (2009) Manual describes what that should mean. It sets out five sets of procedures for linking an examination to the framework: familiarisation, specification, standardisation training and benchmarking, standard setting, and validation. It states that a relationship between a test and the CEFR is not an observable fact but an assertion for which the provider must supply theoretical and empirical evidence.

The Manual also says what it does not do. It provides no label, statement of validity or accreditation; any claim of linkage is the responsibility of the institution making it. And it warns that B2 is a wide band, so two tests both 'at B2' may set their pass marks at different points inside it. So two tests can both be honestly aligned and still disagree about a pupil near a boundary.

Building your own test from CEFR descriptors

Hughes (2003) favours the tailor-made test, and the materials for one are public. The Companion Volume (Council of Europe, 2020) describes an A1 reader as understanding very short, simple texts a single phrase at a time. An A2 reader understands short, simple texts on familiar, concrete matters. A B1 reader reads straightforward factual texts on subjects of interest with satisfactory comprehension. Those three sentences are a test specification.

1. Decide the groups you place into. Three levels need three texts.
2. Write or choose one text of 80 to 120 words that fits each descriptor. How Is a Story Levelled to the CEFR? shows how to check a text's words and grammar.
3. Write five questions per text: facts, order, a word in context, and one 'why'. Mix multiple choice and short answers.
4. Add a twenty-item vocabulary section, with words drawn from each level's texts.
5. Add one short writing prompt, if someone will mark it.
6. Trial the test on current pupils whose level their teachers already know. Set the cut scores where those pupils' results split cleanly.
7. Run it with new entrants, count the moves in the first month, and adjust the cut scores for next year.

The Council of Europe (2009) Manual's own method for cut scores is more formal, with panels of judges and statistical checks. A school placing pupils into three groups needs less: a trial, a record and a yearly correction. How to Assess a Student's English Reading Level gives a worked A2 text and five questions to start from.

Testing young beginners

A child who cannot yet read cannot sit a reading placement test. The Companion Volume (Council of Europe, 2020) adds a Pre-A1 reading descriptor for this stage: recognising familiar words accompanied by pictures, such as a picture book using familiar vocabulary. It also points to a separate collation of descriptors for learners aged 7 to 10 and 11 to 15.

For these pupils, placement relies on listening and vocabulary recognition rather than passages. Say a word and have the child point to the picture. Show a word card and have the child find the picture. Read a very short story aloud and ask what happened, in the child's first language if needed. Record what the child did. Form groups from that record and review them often.

Using the results well

A placement result is only useful if it maps onto something: a group, a pack, a shelf of readers. Decide that map before test day. Pupils in the A2 band go to this group and start on these stories; pupils at the boundary go to the lower group with a date to review.

• Test new entrants on arrival, not at the next termly round.
• Re-check existing pupils lightly once or twice a term, from quiz scores, finished stories and a new text.
• Keep the September test paper. The same test next September lets you compare cohorts; a new test does not.
• Record the test, the date and the result beside the level, so a receiving teacher can see how the level was reached.

How to Measure CEFR Progress by Cohort shows how placement records become a progress report by the end of the year.

How iRead handles this

In iRead, a pack is a CEFR-levelled set of stories, and every story, quiz and game is levelled from A1 to C2. A school picks a global pack by level, theme and values, copies it into the school in one click, and schedules sessions. School leaders see, by class, grade and term, the levels reached, words retained and stories completed. A school can run a free pilot with one class for four to six weeks, with dashboards for attendance and retention.

Key takeaway

Choose or build a short test that separates the groups you actually have. Map every result to a group before test day. Judge the test each year by how many pupils you had to move.

Frequently asked questions

Sources

1. Hughes, A. (2003). Testing for Language Teachers (2nd ed.). Cambridge University Press, Cambridge. doi.org/10.1017/CBO9780511732980
2. Wall, D., Clapham, C., & Alderson, J. C. (1994). Evaluating a placement test. Language Testing, 11(3), 321-344. doi.org/10.1177/026553229401100305
3. Fulcher, G. (1997). An English language placement test: Issues in reliability and validity. Language Testing, 14(2), 113-139. doi.org/10.1177/026553229701400201
4. Bachman, L. F., & Palmer, A. S. (2010). Language Assessment in Practice: Developing language assessments and justifying their use in the real world. Oxford University Press, Oxford. elt.oup.com/catalogue/items/global/linguistics/9780194422932?cc=global&selLanguage=en
5. Nilsson, N. L. (2008). A critical analysis of eight informal reading inventories. The Reading Teacher, 61(7), 526-536. doi.org/10.1598/RT.61.7.2
6. Council of Europe (2009). Relating language examinations to the Common European Framework of Reference for Languages: Learning, teaching, assessment (CEFR). A manual. Language Policy Division, Council of Europe, Strasbourg. rm.coe.int/1680667a2d
7. 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