Skip to main content

13 May 2019

Teachers predict pupil success just as well as exam scores

New research from King’s College London finds that teacher assessments are equally as reliable as standardised exams at predicting educational success.


The researchers say their findings, published today in the Journal of Child Psychology and Psychiatryquestion whether the benefits of standardised exams outweigh the costs.

Teacher assessments were found to correlate strongly with exam scores across English, mathematics and science from age 7-14, with both measures equally as powerful at predicting later exam success. Teacher assessments predicted around 90% of the differences between pupils in exam performance at GCSE and A-level.

Co-lead researcher Dr Kaili Rimfeld, from the Institute of Psychiatry, Psychology & Neuroscience (IoPPN), said: ‘We have shown for the first time that teacher assessments predict GCSE and A-level results just as well as earlier exam scores. The fact that exam scores correlate so highly with the teacher assessments raises questions about the value of the testing culture that characterises compulsory education in the UK.’

Co-lead researcher Dr Margherita Malanchini, from the IoPPN and the University of Texas at Austin, said: ‘While testing can stimulate both pupils and teachers to focus their efforts, high-stakes exams may shift the educational experience away from learning towards exam performance. For these reasons, we suggest that teacher assessments could be relied on for monitoring progress, instead of exam scores, in particular during earlier school years.’

Previous research has looked at how either exam scores or teacher assessments predict educational success, but not compared the two. The researchers were able to make the comparison by linking data from over 5,000 twin pairs in the Twins Early Development Study (TEDS) with teacher assessments and exam scores in the National Pupil Database.

Previous research from King’s College London has established that genetic factors are the major influence on exam results for GCSEs and A-levels. Using data from TEDS, the researchers showed a strong genetic correlation between teacher assessments and exam scores, confirming that both measures were identifying the same pupils and largely measuring the same ability.

Teachers in the UK are required to assess their pupils until age 14, and children sit standardised exams throughout school education in the UK, including SATS at age 7 and 11, GCSEs at age 16 and A-levels at age 18.

Dr Rimfeld said: ‘We are not arguing against testing in general, or that teachers should increase their workloads by adding further assessments. On the contrary, we have demonstrated that current methods of teacher assessment are powerful predictors of success, allowing schools to reduce testing and still monitor progress effectively.’

Dr Malanchini said: ‘Our results should inform the debate about testing during both primary and secondary education. Trusting teachers to implement the curriculum and monitor progress could benefit the wellbeing of pupils and teachers and help to bring joy back to the classroom.’

Previous research has suggested high-stakes exams can impact teachers’ morale as well as pupils’ wellbeing and mental health. The researchers are now looking to study the links between school experiences and mental health among young people.

The study is published online