The Myth of Grade Inflation: Why English GCSE A-Cs Should be Higher

Henry Stewart's picture
 0
This week's TES has a comprehensive article by William Stewart on GCSE grading, asking if exam results will ever rise again. This explains the introduction of "comparable outcomes", that grades are not just based upon actual student performance in the exam but on benchmarking to that cohort's previous Key Stage 2 results.

Does this mean that there is no possibility of secondary schools adding more value? The TES quotes Ofqual's rules as saying the comparable outcome technique should only be used if there has been "no substantial improvement in the quality of teaching and learning". Has there been such an improvement? Ofsted's 2012 report states that 70% of schools are now Good or Outstanding, compared to just 64% five years ago. That seems a pretty significant improvement.

So on Ofqual's own terms there is an argument for grades to be allowed to rise. But there is also a strong argument in terms of "comparable outcomes" themselves. The general assumption is that English GCSE results have increased hugely and the fall in GCSE results last year, under comparable outcomes, should have been based on bringing them into line with the KS2 figures to which they are being benchmarked.

Has there been GCSE Grade Deflation?



In fact the comparison of English KS2 and GCSE results indicates grade deflation is more likely than inflation:

* In the twenty years from 1993 to 2012, English GCSE A-Cs rose only from 57% to 64%, a rise of just 7 % pts
* Yet, from their introduction in 1995, KS2 level 4s (and above) rose from 49% to 82%, a rise of 33 % pts

Comparison Chart 2

It could be argued that the sharp rise in the early years of KS2 SATs was due to schools getting to know the tests and getting better at helping their children to pass them. So let's take just the last five years, and look at how English GCSE results would have risen if they had kept in line with KS2 results:

Comparison Chart 5

Using comparable outcomes, last year's English GCSE results should have risen not fallen. Taking 2008 as the base, they are now 3% below where they should be, based on benchmarking against KS2 results, with a further rise due this year.

Will GCSE Grades Rise this Year?



It is clear that there was substantial pressure on Ofqual and the exam boards to reduce the numbers passing their GCSE. As the TES noted, the fall in results is seen by government as "an unquestionably good thing" - despite its consequence of thousands of students finding it harder to continue their education as they might wish.

It may be that there will be similar pressure on Ofqual this year. However using Ofqual's own terms, providing the actual exam papers justify it, there should be a rise in the number of students achieving English GCSE A-C grades this year:

1) Ofsted ratings indicate there have been significant improvement in the quality of teaching and learning in schools over the last few years

2) The comparable outcomes approach of benchmarking against previous performance would lead to a rise of 1% (if kept in line with last year), 3% (if kept in line with 2011) or 4% (if kept in line with 2008).

I look forward to an independent Ofqual and an independent set of exam boards resisting political pressure and allowing the number of students achieving GCSE grades to rise - and students to be allowed to get the grade they deserve.

 

Sources & Data Notes:

GCSE results (all subjects) - from Guardian datablog
KS2 results (English, Maths) - DfE data: scroll down to EXCEL (national tables)

This analysis refers only to English. English A-C GCSE grades rose from 57% to 64% between 1993 and 2012, Maths A-Cs went from 46% to 58%. However there may have been grade inflation in the non-core subjects, where A-Cs went from 42% to 69%.

The DfE KS2 figures are rounded to the nearest % pt. As any successful GCSE Maths student could explain a rise from 79% to 80%, as in 2007 (the year that 2012 GCSE students took their KS2 tests), could mean a rise as small as 79.4% to 79.5% or as large as 78.5% to 80.4%.

 

 
Share on Twitter

Add new comment

Already a member? Click here to log in before you comment. Or register with us.