# In defence of levels and grades

6 May 2015

I have heard a lot of talk about the problem of replacing National Curriculum levels, from both primary and secondary teachers. The suggestions for a new system are varied in quality (to say the least!) but I'd like to run through a few and offer my thoughts. I'll be thinking of them in a Maths context, as is my wont, but I hope this is interesting for teachers of other subjects too.

The first technique I'd like to deal with was mentioned to me by a Head of Maths working at a large comprehensive school in London, and is as follows – write out a list of the topics in maths, then put either a tick or a cross for each student in the school, according to whether they can do it or not. This approach is so ludicrously simplistic, subjective and misguided that I won't waste much time on it, other than to say that it is indicative of the difficulties teachers are having.

The second a KS3/KS4 grading system which works as follows – you compute a GCSE target based on KS2 data and required value add performance, as I'm sure all schools do, then define the student to be 'at' that grade, for their year group. You define what it means for a student to be an A*, B, C etc. in  year 7, 8, 9 etc. then the aim is for students to remain at the grade they start at, or improve. This is rife with problems. Firstly, there is subjectivity in how you define the Y7 A, Y8 A, and so on. Secondly, it makes it difficult to compare with other subjects. Thirdly, it surely leads to low expectations for students who've underperformed at KS2, more so than for other systems.

The third approach I'll discuss is sticking with NC levels until Y9, then switching to GCSE grades in Y10, using some agreed conversion, generally that a level 7 is a C. I don't agree with this approach as it relies on a spurious equivalence (between a 7 and a C) and makes it very difficult to assess progress at one of the most crucial stages in a student's education.

So, what should be done? Here's what:

• Students' GCSE targets are calculated using their KS2 data, national statistics, and the school's Value Add targets.

• Students are given a baseline paper, which is a Foundation GCSE paper. The grade boundaries (which are the official ones) are extended downwards to H3 so that all students are able to get a grade.

• Previous data is used (and iterated each year) to calculate the expected progress that each student should have made at each assessment point (end of each term) as a percentage of the total.

• Students are tested using GCSE past papers (either Foundation or Higher depending on their ability) each term and their results are compared to the grade they would be on if they were on track to make their target grade, as well as if they were on track to make progress that would exceed the value add target.

• These comparisons form the basis of our assessment of their progress.

I think this approach is objective, rigorous and well thought out. It yields meaningful data that is comparable to other subject areas. It uses a nationally assessed starting point to create the target, but compares like for like at every stage, i.e. it doesn't involve a 'leap' from NC levels to GCSE grades.

I wonder if primary schools can do the same thing. Use levels as a simple way to assess progress, using broadly the same techniques as described above, explained along the lines of 'if you did your SATS today, this is what you'd get'. Use a KS1 SATS paper as a baseline, use either KS1 or KS2 SATS papers throughout, depending on ability, and assume linear progression until the data shows you otherwise. No more teachers pretending their students can do things they can't, just formal, rigorous and reliable testing. What's not to like?!

We'd love to hear your thoughts. Get in touch with us on Twitter @createatest