How bias subconsciously emerges in teacher assessment
Daniel Kahneman, the Nobel prize winning economist, has demonstrated through forty years of experiments that people exhibit bias in their everyday and professional lives, mostly without being conscious of it. He describes one form of bias as an anchoring effect. Anchoring occurs when we try to think of a value for an unknown quantity before estimating that quantity. Anchoring is a natural human response, but as we are rarely conscious of it, when it emerges through assessment it can be extremely problematic.
When we assess a piece of work from a child that we know well, our bias emerges. Perhaps we know they can perform better than the piece in front of us, subconsciously prompting us to raise the mark. Even if the work is assessed anonymously, the existing evidence shows that bias is exhibited against pupils with SEN, those whose behaviour is challenging, those for whom English is an additional language, and those on Free School Meals. Assessment judgments can often be overly-lenient, overly-harsh or, indeed, can reinforce stereotypes, such as boys being perceived as better than girls at mathematics.
This doesn’t mean that teachers should abandon teacher assessment altogether but it does require an acknowledgement that reliable and unbiased assessment is a considerable challenge. To improve the quality of teacher assessments it is important to consider how to:
- improve the reliability (consistency) of assessments;
- increase the accuracy of teacher judgement
- increase the precision of inferences drawn from assessments;
- reduce systematic biases.
The key message here is not that teacher assessment can’t or shouldn’t be done; it is that teacher assessment is hard to get right, and that it requires excellent training, moderation, standardisation and quality control. Our starting point should be that great assessments are valid, reliable, purposeful and valuable, but these things are not easy to get right.
1. Kahneman, D. (2013), Thinking Fast and Slow, Penguin, London