The government has announced proposals for a new assessment system for primary schools, including making key stage 1 Sats non-statutory, reform of the EYFS Profile, and "a new teacher-mediated assessment in reception, developed with the profession" for the purpose of measuring progress later in school.
They are consulting on these changes, which follow the recommendations of the NAHT’s independent Assessment Review Group Report earlier this year. While attempts to reduce the amount of high-stakes testing in primary schools will be welcomed by many, there is a cloud to this silver lining in the form of the return of the baseline assessment.
These proposals follow the failed policy of baseline assessment in Reception, introduced in 2015, which saw 15,000 primary schools conduct an assessment with four- and five-year-olds which produced no useful data, but cost an estimated £4.5 million. After thousands of teachers had assessed their Reception pupils in the first weeks of school, they were told that baseline scores were not useful because the different forms of assessment provided by different companies were incompatible.
Our research with the NUT and ATL, which involved a nationwide survey of teachers and in-depth interviews with Reception teachers and headteachers during that six-week assessment phase in autumn 2015, found a number of serious problems relating to baseline assessment. These problems remain whenever the baseline is assessed – at the beginning of Reception or at the end.
The 2015 version was an attempt to reduce the complexity and contradiction of four- and five-year-old children to a single number. The plan was to use this number to predict children’s progress across seven years of primary schooling, and compare this prediction with their Sats results at the end of key stage 2. However inaccurate, the aim was to hold a primary school to account by measuring the "value-added" to each child. Baseline assessment provided the input information to this inappropriate input/output model of education; no wonder then that teachers responding to our survey described it as treating children like sausages in a factory.
The first problem with a progress measure system is that it is dependent on the idea that you can establish with some certainty what a child can or cannot do at a given point, reduce this to a single number, and use that number as a point of comparison years later. At this age children demonstrate what they can do in their play in diverse and imaginative ways, but a baseline assessment offers only crude judgements, made sometimes through observation and sometimes through tests. Whatever form it takes, and even if developed ‘with the profession’ it is impossible to decide with certainty what a child of four or five can or cannot do, because they are affected by so many other contextual factors.
For example, teachers told us baseline scores were affected by whether children speak English as a first language, their age (some Reception pupils are only just 4, others are 5), familiarity with a tablet, their confidence and relationship with the teacher, and even how tired or hungry they were. Some teachers saw baseline as damaging children’s wellbeing, confidence and self-esteem, as they were upset when given difficult tasks. So, rather than being an exact science, the baseline outcomes were an often negative and inaccurate approximation of what children could do.
This leads us to the second problem with any kind of baseline assessment: the setting of low expectations. Here the inaccurate information initially put into baseline and the invisibility of the mathematical algorithms used to produce a single figure have potentially harmful effects.
There is a danger that if you label a child with a score at age 4 or 5, and then measure their progress seven years later, you engage in a model of education which defines children as either having "ability" or not – which tends to stick with them as they travel through the system. This model engages, explicitly or not, in the business of predicting whether has a child has potential. Baseline, if used in this way, has the potential to become a self-fulfilling prophecy, even though it is based on inaccurate data.
Finally, there is a real temptation in any form of value added system for teachers to manipulate of results. This has already been found in relation to the EYFS Profile assessment at age 5, where teachers describe being told to keep results low so that the "story" of attainment as children go up the school is more attractive to Ofsted. In our research on baseline, headteachers described this temptation to "limit the damage" within a system they saw as perpetually critical. So, even if you could produce an accurate, reliable test, it would be affected by its role as a performance measure.
These dangers of baseline will not go away with the announcement of the new proposals, because they are based on the same flawed principles and logic as the original policy. Like many in the primary and early years education world, we are surprised that the government has attempted to bring back a discredited policy that is inappropriate for young children and potentially damaging. An effective coalition of education groups came together in 2015-16 as the Better Without Baseline campaign; we hope the government will listen again to those who know about children and education and abandon their plans.
Alice Bradbury and Guy Roberts-Holmes are researchers at UCL Institute of Education in London