Monday, 30 June 2014

"What is the best place for testing in schools?" Has testing supplanted effective teaching and learning?

I am pleased to offer this further contribution to this month’s #blogsync

Since the announcement of the scrapping of levels and the changes to testing arrangements there has been copious debate and a fair amount of head scratching as to the future of testing and its place in the cycle of assessment.

Who do we assess for?

·         For the children – to inform the next stages of their learning, to inform them of their progress and what they need to do to reach the next step of their learning journey.

·         For the parents- to inform them of their children’s progress and how they can best support their offspring.

·         For ourselves- to plan the next stage of the children’s learning, to set learning targets, to judge relative progress.

We don’t assess for OFSTED, though OFSTED will be interrogating our data, books and the children in regard to their attitude to learning.

What is assessment?

There’s been much dialogue about assessment since the announcement about the scrapping of National Curriculum levels. Much of this debate has been not so much about assessment as about tracking. Whilst both are required, the current system of National Curriculum levels has been that its use as a tracking tool has seemingly replaced its purpose as an assessment tool.

To assess: to evaluate or estimate the nature, ability, or quality something.
To track: to follow the trail or movements of something.

Inevitably, the way that Ofsted works has meant that schools have been forced to use their levelled assessments to prove that they are tracking progress towards the end-of-key-stage expectations. However, in doing so we have all but removed the act of assessment from the processes of teaching and learning.

The problem with the drive towards the tracking of progress is that with each large step towards numerical data we lose the small step detail. Ultimately we lose sight of what makes a good teacher: one that knows the children in their class.

For assessment to be useful and meaningful, it needs to tell children, teachers and sometimes parents, what it is that the child can or cannot do. Levels aren’t especially helpful, since they were designed to be broad. Accordingly, sub-levels were created to try to fill the gap as indicators on a tracking scale. Remember though, sublevels were only intended originally to subdivide level 2 for a more measured ability profile at Year 2. Sublevelling higher up transpired from use of the optional tests; readers will recall that APP deemed children as low, secure or high, not C, B or A.

For assessment to work well, it needs to be directly linked to the taught curriculum. In schools, we need processes which are directly linked to the curriculum that allow teachers to judge how well their children have learned what has been taught. Current national tests (optional and statutory) do not fully serve this purpose. A child in Y5 might not have covered all of the KS2 curriculum content which could come up on a test, but a test level does not discern between that the taught and untaught, the grasped and ungrasped.

For example a child can achieve a Level 4 on National Curriculum tests, or even through APP assessment, without knowing their tables up to 10×10, despite this being a requirement of both the APP criteria and the National Curriculum attainment target. Indeed, some children who manage well in many areas of mathematics can continue to appear to be making progress in tracking records, despite never confidently knowing some key aspects of the curriculum.
In the absence of assessment, it is perfectly possible for a child to never have this key need identified. Despite all that we know about the importance of some key aspects of subjects, progress measurement relates to tracking first and foremost.

A child might well be achieving level 5, or within the expected range for his/her age, or 101 on a scaled score- but none of that information gives away the truth about whether he/she can quickly recall 6 x 7 or to apply this knowledge further.

The issue with testing

QCA optional tests were only ever designed as a summative measure, designed to be delivered at the end of a school Year. There was never any intention to correlate the standards of the 1998, 2003 and 2006 tests to each other, or to correlate the Year 3, 4 and 5 tests to each other- hence a high achieving child in Year 3 would score an indicative level 4, and a high achieving child in Year 4 would also score the same indicative level 4. The key term is ‘indicative’ here. The same issue exists with level 5 in the Year 5 tests. At best the ‘indicative’ level is a 5c (or a 4c in the Year 3 and 4 materials) but it is also worth considering the implications for KS2 tests- until recently a child achieving level 5 in Year 5 tests had nowhere higher to go in Year 6.

One further issue is the sublevelling. In some tests a margin of 5 marks could turn a ‘c’ to an ‘a’, which in terms of APS progress is 4 points, theoretically in excess of a whole Year’s progress. With an element of ‘teaching to the test’ a few ‘tricks’, such as looking for the questions where drawing a line or completing a box is the requirement, can make the apparent difference to the result. This is not representative of good teaching or learning.

One observed issue has been the splitting of the indicative upper level into sublevels which is completely inaccurate in terms of reporting. Overall testing has issues with reliability and validity, and hence there are issues with what testing actually means.

Why teaching to the test is so bad

Tests are about making a measurement: of something enormous and not clearly defined.  The term ‘domain’ describes what is available to be tested. The domain is huge; just consider how much as a primary teacher you cover in an entire academic Year. No test can ever truly measure the domain, as parts of the curriculum cannot be tested in a written form. In short, tests are at best only a sample.

Teachers need to see and use standardised tests which prompt certain responses and which are the same for each child taking them. The QCA tests are standardised, but not in relation to each other, and were only ever designed to be end of Year, summative assessments with results that were only ever intended to be indicative rather than strictly accurate.

Memorisation of things such as times tables and verb endings, is of some value, but memorisation of the wrong things like exam tricks and hints, is clearly not. If the writing test is use, teaching Year 5 a unit on persuasive writing shortly before that test will inevitably result in a skewed result. Matching teaching to the test will inevitably result in fabricated gains.  The syllabus is not the curriculum, and nor should it be the curriculum.

It is so much easier nowadays for teachers (and parents) to download past papers from the internet and work on their children’s areas of weakness. However teaching which fixates on past papers and test preparation is not teaching to the domain. It’s teaching to the sample. The improvements it generates are not likely to be genuine. Teaching to the test doesn’t replace teaching and learning.  A good Year 6 teacher will complete old SATS papers with a class, but not to teach test tricks, they to allow children the experience of working to time and under pressure.

Some schools give the same QCA test at the end of each term. This can produce feelings of familiarity which may result in overconfidence, and subsequent underperformance, or undermined confidence for the fear of repeated mistakes. In contrast a child who performs well at Christmas has nowhere else to go in terms of progress for the rest of the year.

So: What now?

Testing clearly has a place as a measure of progress with a National standard in place. Overtesting results in anxieties, for children and teachers, and potential inaccuracies.

Good teachers know their children: what they can do well, where they need support, how they can develop in the future. Rigorous and regular formative teacher assessment, monitored and moderated, is representative of good practice. Knowing that a child is working at a secure level 3, rather than saying the child is there ‘because the test says so’, is a clearer indication of a teacher with a knowledge of their class, and a firmer basis for planning next steps in learning. With levels going, these teachers are going to be in the best position to adjust to any change.