By Ben Durbin
Friday 8 January 2016
When you walk along a coastal path close to the cliff edge, it’s hard to concentrate on anything other than the sheer drop below. It’s more difficult to hold a conversation and positively risky to look too far ahead for fear you might slip. This situation becomes all the more exaggerated for someone who does slip towards the edge, scrabbling to regain their footing.
I offer this scenario as it may help to illustrate why threshold measures as part of a performance or accountability framework are not a good idea. A threshold is a single, fixed level of performance, below which the individual or institution is deemed to be underperforming. One recent example from the education sector is the C/D borderline at Key Stage 4. Pupils obtaining a D grade have not been considered to have achieved a ‘good’ GCSE, and schools have been measured according to the number of pupils clearing this threshold in five of their subjects, including English and Maths.
The problem is that threshold measures act like cliff edges…
…creating perverse incentives and disproportionate levels of stress for anyone close to the edge. This can lead to odd behaviour, short-termism, and ultimately negative consequences. For example, it encourages schools to focus their efforts on pupils close to the C/D border, to the detriment of their higher or lower attaining peers. And even for these pupils who might be expected to benefit, it encourages schools to place undue pressure on them to achieve a specific grade in a specific exam, rather than considering their overall educational needs.
These issues were recognised by the Department for Education (DfE), which sought to address them through the new Progress 8 performance measure. This focuses on the progress made by pupils rather than their absolute levels of attainment, and includes pupils across the attainment spectrum. A school helping a pupil to achieve a D rather than an E, or an A* rather than an A will be rewarded just as much under the new measure as helping a pupil achieve a C rather than a D.
This is why it is surprising that DfE’s proposed new approach to tackling underperforming schools is based around a set of threshold measures. The issue is particularly acute with the proposed primary school measures, presented in a DfE consultation document published last autumn, and I’ll focus on these for the remainder of this post.
Hundreds of primary schools are teetering on the brink
A primary school will be classified as ‘coasting’ if it fails to clear thresholds relating to pupil progress and absolute levels of attainment in maths, reading and writing for three years in a row. These schools will become the subject of their Regional Schools Commissioner’s (RSC) attention, potentially leading to additional support or to intervention. The nature of this support is also considered by the consultation, a topic we’ve explored in another recent post.
In NFER’s response to the consultation, we highlighted a number of risks.
- The ‘three years in a row’ condition risks a scenario whereby schools with two years in a row below the coasting threshold invest excessive effort in ensuring the upcoming year 6 or year 11 cohort perform above the threshold, at the expense of other pupils in the school. This increases the chances that in the following year the school will drop back below the threshold, and possibly get trapped in a cycle of focusing on the threshold at the expense of a longer-term, whole-school approach to school improvement.
- To qualify as coasting, a primary school has to fall below twelve distinct thresholds (one each for reading, writing, maths and overall progress in each of three years). If it clears any one of these twelve thresholds it will not qualify as coasting, despite still appearing substantively like any other coasting school.
- A school just below the coasting threshold in maths for example, but well below the threshold for writing, would be incentivised by the regime to focus on their moderate maths performance rather than their poor writing results.
We undertook an analysis of the numbers of schools likely to be classified as coasting based on recent performance. This identified 496 mainstream primary schools that would have been deemed coasting (i.e. having slipped off the metaphorical cliff edge). A further 754 schools only cleared one of the twelve thresholds (and hence would be walking perilously close to the drop).
A steep decline is preferable to a sheer drop
So what’s the alternative? In our consultation response we suggested an approach which would mitigate some of the risks associated with the Government’s proposals. Rather than a binary ‘coasting or not’ classification based on schools failing to clear thresholds for three years in a row, schools could be assigned a score out of five (say), recording how many of the past five years they have cleared the threshold.
Those schools scoring four (i.e. dropping the below the threshold in just one of the four years) might just be placed on ‘watch’ by the RSC, and schools scoring two or three are likely to be roughly those identified as coasting under the current proposals. This approach would also better align the definition itself with the nuance of the RSCs’ potential responses.
We are suggesting that school performance is considered on more of a continuum, rather than the ‘all or nothing’ approach of a single threshold. Challenging underperforming schools with a steep performance curve, rather than an unassailable cliff edge, is more likely to result in better outcomes for all their pupils.