OfQual has just released guidance on the centre assessment. Predicted grades are to be submitted as a ranking system based on teacher assessment. The only coursework/NEA/RA required by exam boards will be in Science and English.

There are issues with teacher assessment and predicted grades being biased against certain groups. The process of bias is not because teachers are racist, classist, etc. All of us are prone to creating a stereotype and prototype models. I will not go into the literature or the data in this piece, but here is an article that I wrote previously, which succinctly shows the evidence and its impact.

Let me reiterate that:

This is not about teachers and leaders deliberately under-assessing pupils.

What can we do? As schools are required to submit teacher ranking and predicted grades without further assessment; Both measures are problematic in themselves and will be detrimental to some of the following groups of pupils.

Black pupils of all categories
Pakistani
Bangladeshi
FSM
etc.

What to do?

As a former Curriculum and Standard Assistant Principal and data geek I’m interested in the processes involved the intricacies of getting this right. First, use a system which allows you to see all of your data together, that is module tests, mock exams, etc. in one place with a grade boundary analysis (from exam boards).

Leaders and teachers should look across the sheet and make a judgement based on the data you have in front of you.

Then the next step in the process is to apply (initially) a systematic process of adjustment, i.e. the same amount of improvement in all pupils – in the three months till the exams all pupils would’ve made two grades progress.

And we are done.

No, we are not.

All exams and the teacher assessments are prone to bias – we have all had that annoying pupil who pulls ‘it’ out of the bag on results day.

Here are some of the factors we need to consider to ameliorate the impact of underassessment.

Behaviour
Race
Gender
Socio-economic background.
EAL
as well as other others

We are lucky that today’s schools are data-rich havens. I would advocate looking at the data you have on predicted grades with actual grades from previous years and complete a forensic analysis the accuracy of prediction against the factors.

At the department level, this could inform legacies of under predicting, i.e. History always underpredict but deliver on the day. After all, we don’t have the benefit of the ‘day’. On the teacher level, it may highlight a stereotype model for over prediction, i.e. Chinese girls are regularly over predicted in science.

There are still issues with teacher assessment, but we must at least try and get this right for the pupils we serve. Now, it is up to school leaders to make decisions of whether you adjust individual pupils grades predicted grades or consider this in the ranking process. If you require support as always, please do get in touch.

If you are still confused about the idea of teacher bias click here for the research on the topic.