8th January 2014 9:00
By Blue Tutors
We recently reported on a news article which raised concerns about the accuracy of predicted A Level grades, suggesting that teachers aren’t great at assessing how well a student is likely to do in their final exams. While there are no arguments about whether inaccurate predicted grades create problems or not, it does seem a little unfair to criticise teachers too much; it’s not an easy task, and not an exact science.
Teachers are much better at predicting whether a student will achieve an A* or A at A Level (predicting those grades with a 60% accuracy compared to 48% overall), and this makes a lot of sense because someone on course for the top grade can only drop down, they can’t improve, so a teacher’s prediction isn’t affected by the chance that the student will exceed expectations. It’s important to point out the fragility of A Level grades though; students are assessed relative to each other, and as a result the grade boundaries can be very narrow which means that a slightly below par performance in even one exam can take a student from a solid C to a D.
Something often overlooked is that teachers aren’t assessing each student based purely on their work so far, there’s also the consideration about whether a student will rise to the occasion towards the end of his/her A Levels. Some students crack under the pressure, and some improve dramatically in the last few weeks before the final exams. When a teacher has so many students, it’s difficult to know them all well enough to predict in which of these two camps the student will be.
The worry about the seeming over-prediction is understandable; we don’t want to think that teachers are deliberately predicting higher grades in an attempt to improve a student’s chances of being offered a good university place. However, there is another way to look at the over-prediction, which is that most teachers are genuinely optimistic about their students’ chances, and can see the potential in them all. I have to say, I think we’re better off with a country of teachers like that, rather than pessimists, who have no faith in the students they are teaching.
Having said all of this, the biggest problem with criticising teachers for their predictions is that it shows little understanding of the way in which exam results are distributed. Assuming a normal distribution around a student’s predicted grade, and a standard deviation of one grade, which seems reasonable, we would only expect predicted grades to be correct 38% of the time. They are correct 48% of the time, so maybe teachers are actually predicting grades better than expected.