Marking System for Aptis Advanced (Pro-ELT Project)

The Aptis Advanced test is a specially localised variant of the Aptis General test, which has been adapted to respond to the Ministry’s requirement for differentiation between C1 and C2 levels of performance. Aptis Advanced was designed, implemented and monitored closely throughout by the UK Aptis Examiner Management team and the UK Aptis Language Assessment Research team, with senior UK examiners specially trained to use the rating scale designed for this purpose.

The current Aptis General test provides reporting up to an overall C level, but does not distinguish finer levels of performance at the highest level. In order to distinguish C1 and C2 level performances, a double-rating system has been instituted. All test takers receive a final scaled score for each component and an initial CEFR level designation for these skills, using the same scoring and CEFR cut-offs employed for the Aptis General test. Those scaled scores and CEFR cut-offs are set through extensive piloting and standard-setting projects employing the methodology recommended by the Council of Europe for linking examinations to the CEFR. For the Ministry of Education & Higher Learning’s Pro-ELT project, a second stage of evaluation was built into the quality assurance system to enable the finer-grained, high-level distinctions the Ministry requires.

In the second stage of evaluation, test takers scoring at the top end of the scored scale (C level candidates) are re-rated by specially trained senior UK examiners. These examiners use a rating scale specifically designed for this purpose which allows for finer-grained distinctions at the C1 and C2 level. In a small number of cases this means that test takers with the same final scaled score, when it is at the top end of the score scale, will sometimes receive different component CEFR level designations if their performance is determined to be C1 or C2. In the same way, a small percentage of test takers will find that the finer-grained evaluation of their performance will result in a slightly lower component CEFR designation when a test taker is borderline B2/C.

What is the difference between the Aptis Advanced test used in the Malaysian Ministry of Education & Higher Learning’s Pro-ELT project and an Aptis General test?

The version of the test developed for the project can distinguish between candidates at a high level in a way that an Aptis General test cannot i.e. it can tell the difference between a C1 and C2 level. Aptis General can tell if a candidate is better than B2 i.e. a C, but cannot differentiate between C1 and C2.

How is the Aptis Advanced test marked differently?

A double–marking system for the Writing and Speaking test components is employed. Initially, every test is marked once by an examiner and given a final score. This score is then converted into an initial CEFR level using cut-off scores i.e. the barrier between one CEFR level and another. A specially trained Senior Examiner then marks all the tests that have been awarded a CEFR level at C a second time to differentiate performance at a high level. The Senior Examiner will award a final CEFR level for higher level candidates. This can be C1 or C2, or in a very small number of case a B2.

Is it possible for 2 candidates to get the same component score but different component CEFR levels?

Yes. For example, 2 candidates may receive a final score of 48/50 for e.g. Writing, which, using the cut-off score system will place them both in the C CEFR band. However, this score is very close to the border between B2 and C, and so when the Senior Examiner re-marks their test, one candidate may be awarded a B2 and one candidate a C1. This system of double-marking makes the test very accurate for the point of view of CEFR banding at a high level.

How often will this happen?

It will be quite common for C1 and C2 candidates to have the same mark, because the original cut-off scores did not differentiate between the two. However, it will be much rarer for B2 and C1 candidates to have the same mark. However, it will happen occasionally. We calculate that this happened in approximately 1% of cases in the Pro-ELT testing.

How did the British Council decide on the cut-off scores in the first place?

Our CEFR cut-off scores are set through extensive piloting and standard-setting projects employing the methodology recommended by the Council of Europe for linking examinations to the CEFR.