How to place new language students: a working method for schools

Placing a new student looks like a small decision. It takes a few minutes and nobody remembers it a month later, unless it was wrong. A student placed too high stops speaking in class; one placed too low is gone by the third week. This guide sets out a method that a school of any size can run, with or without software.

What placement has to achieve

The aim is not to find the student’s exact level. It is to find the group in which they can follow most of what happens and still meet something new every lesson. Those are two conditions, and a good placement satisfies both. It follows that the right group depends on your timetable as much as on the student: if you run two B1 groups, the question is which of the two, and “B1” alone does not answer it.

Three sources of information

No single instrument sees the whole student. A sound placement draws on three:

• A test of grammar and vocabulary. Fast, objective, and good at ranking a large intake. It says little about how the student speaks.
• A short conversation. Two or three minutes are enough to hear fluency, pronunciation and confidence, which no written test shows.
• The student’s history and purpose. How long since they last studied, what they need the language for, and how many hours a week they can give.

The order matters. Do the test first and at home, so that the conversation at the front desk starts from a result instead of from zero.

A process in five steps

1. Test before the first visit. Send the test when the student enrols or enquires. Every day between enquiry and placement is a day in which they can change their mind.
2. Sort by level and score. List the intake by result. Most students fall clearly inside a level and need no further thought.
3. Mark the borderline cases. A student whose score sits at the edge of a level, or whose margin reaches into the next one, goes on a short list.
4. Talk to the short list. Spend your teachers’ time where it changes the decision: on the borderline students, not on all of them.
5. Check after the first week. Ask each teacher for the names of students who seem misplaced, and move them at once. A move in week one is routine; in week four it is a complaint.

Mistakes worth avoiding

Placing by years of study

“I did four years at school” describes a timetable, not a level. Two students with the same four years can be three levels apart.

Trusting a number more than it deserves

Every test result has a margin. A student who scores just under a boundary is not measurably different from one who scores just over it. If your test does not tell you its margin, assume it is wider than you would like, and see how length and precision are linked.

One paper for everyone

A fixed test must cover every level, so most of its questions are too easy or too hard for any one student. Beginners are discouraged before they start, and advanced students are bored. An adaptive test asks each student only questions near their own level.

Letting the test decide alone

A grammar test can place a careful reader above their speaking level, and a fluent speaker with shaky grammar below it. The conversation is the correction.

The borderline student

When the evidence is evenly balanced, three questions usually settle it:

• How do they speak? If speaking is stronger than the test suggests, the higher group; if weaker, the lower.
• How recently did they study? Someone returning after years regains ground quickly; lean higher.
• How much time do they have? A student with two evenings a week and a demanding job is better served by a group in which they are comfortable.

A checklist for your next intake

• The test goes out automatically at enrolment, not on placement day.
• Results are sorted by level and by score inside the level.
• Borderline students are listed and spoken to; the others are placed directly.
• Teachers report misplaced students at the end of week one.
• Moves are made in week two at the latest.

NGPT was built around this process: the test is sent by link, the result gives a level, a score and its margin, and the list can be exported for timetabling.