1. Target Grades. A thread as I've seen a fair few tweets about them recently. Buckle in. Once more into the breach dear friends..

2. Target grades generated for pupils are usually based on averages. These averages are based on a wide spread of GCSE grades obtained by children who got similar scores in the primary examinations in English and Maths they take at the end of Year 6.

3. Target grades are usually based on mean averages. It means the target grade is not what most children get in that subject. It is not the mode. There is a wide spread of grades and most children do not get their target grade. Some get higher than it and some get lower than it.

4. With large numbers of children the data can be useful because these inconsistences are ironed out a bit. If a whole cohort is achieving an average grade higher than the average ‘target’ grade then can have some confidence that this particular subject is performing 'well.'

5. None of this tells us anything about the reasons why a subject or cohort is ‘over’ or ‘under’ performing. The reasons could be teaching but they could also be to do with wider problems in the school or be affected by pupil demographics. More on this now:

6. Cohort and national headline data hides variation between different demographic groups. E g – nationally at GCSE girls out erform boys. An all girls school is more likely to achieve higher grades from the same starting points (Year 6 exams) than an all boys school will.

7. Pupils with EAL also – again only on average though – perform better than non EAL pupils against their ‘target’ grades often because many of these did not perform as highly as others in Year 6 exams simply because when they took them they had not yet developed their English.

8. We could spend hours delving into demographic differences but it’s not the point of this thread and ultimately it wouldn’t get us anywhere. How one individual child performs in their final exams is enormously complex (even infinitely) because the world is enormously complex.

9. All this complexity can’t be reduced to a generic target grade that's means anything.
Well actually it can but doing so creates damaging distortion.
Here’s a couple of examples:

10. An EAL pupil who only performed comparatively poorly on Year 6 exams because they had not yet developed their English will get an inappropriately low ‘target grade.’ This could make them complacent or lead them to believe they have lower capabilities than they really do.

11. A child who was out of school for whatever reason for months or years after taking their Year 6 exams is far less likely to achieve the ‘average’ of what pupils who got the same score did. Their ‘target grade’ will be just really demoralising

12. We could use big data as a starting point with individuals to talk about their chances of achieving certain grades. This is how the FFT began using big data with individuals – a mutually agreed ‘expected grade.’ But this is a lot of fuss and why bother? More on this in bit.

13. If generating individual target grades made any sense (it didn’t) in the past it makes less sense now. Covid-19 related shutdowns and periods of remote learning are going to create even more noise in the system with averages even less representative of any one individual.

14. If nothing else we should understand ‘target grades’ are based on data pulled down from group and we can’t really tell anything about an individual with any confidence.

15. So what do we use instead? Well first of all why should we assume we have to use anything instead? For most of the history of education we haven’t felt the need to say whether a child is ‘on track’ towards anything. It’s far from clear this does any good anyway.

16. Secondly also worth remembering these ‘target grades’ are performance goals and in themselves offer no support for improvement. Bluntly – if the children given any ‘target grade’ could achieve that target grade they probably would be doing so already.

17. Much better are what Sejits, Locke and Latham call ‘learning goals’. This means breaking down the components of improvement and encouraging pupils to work on these. If they get better at these things then they will improve and their grades will go up.

18. This does not mean big data offers nothing. Helpful if used as originally intended – working out if big groups are achieving or underachieving. But this doesn’t tell us anything about why. There’s more work to do and not just finding someone to praise, shout at or fire.

19. I wrote about all this ages ago too: https://bennewmark.wordpress.com/2017/09/10/why-target-grades-miss-the-mark/

20. Fin.