‘Constructs’ are the kind of things psychology researchers try to measure and ‘instruments’ are what they try to measure them with. For instance, in my field, cognitive load is a construct and instruments used to measure it include surveys asking subjects how much mental effort they invested in a task.
For an instrument to be considered a good instrument, it must be both valid and reliable.
A reliable instrument is one where the same person, under the same conditions, will produce the same response at different points in time and where different subjects who both have the same form of the construct will produce similar results.
A valid instrument measures what it is supposed to measure. An instrument may not be valid if it either does not measure the construct in question or if that construct is itself misconceived, perhaps because it doesn’t actually exist.
Clearly, there is something of a hierarchy here. An unreliable instrument cannot really be valid.
In education, a number of researchers have proposed the construct of learning styles. The basic idea is that individuals have different preferred styles of learning, such as visual or auditory, and that if teachers cater to these styles then students will learn better. Instruments have been developed that survey students on their preferred learning styles, with the idea that teachers may make use of the results in the classroom. Unfortunately for this hypothesis, we have known for at least ten years that the construct of learning styles and the various instruments used to measure it are invalid. Nevertheless, the idea seems incredibly difficult to kill off.
One possible reason for the persistence of a belief in learning styles is that the instruments do measure something – a preference for learning in a specific way – but that this is invalid in the sense that targeting teaching to these preferences doesn’t lead to any improvements in learning. This may be because the content a teacher is attempting to teach is a more important determiner of the mode of teaching than individual preferences. This would lead to the surveys reliably measuring something, just not the construct we think they are measuring.
The potential harm of a persistent belief in learning styles is that students will be given suboptimal teaching. Students who are classified as kinaesthetic learners, for instance, may be given a degraded form of maths instruction that focuses more on the need for movement than the optimal methods for teaching the maths in question. And if these associations align with societal stereotypes then we could entrench disadvantage.
Yet, the unintended outcomes of belief in a flawed construct could be even worse and that leads us on to the Implicit Association Test (IAT).
The IAT was developed by researchers at the end of the last century. The basic idea is a simple one – if individuals are quicker to associate some traits and slower to associate other traits to a particular class of individual when compared to another group – such as men versus women or disabled versus able-bodied – then we may infer that the individual holds an implicit or unconscious bias regarding that group.
So the construct is implicit bias – a bias an individual unknowingly holds about a group of people – and the instrument is the IAT.
The evidence for the reliability and validity of the IAT is at best mixed. Even where the IAT suggests individuals hold such a bias, it is not at all clear that their behaviour towards the group they are supposedly biased against is any different as a result. One possible reason is that the test measures familiarity with societal stereotypes rather than underlying beliefs and values.
In the current climate of enhanced awareness of racism, there is a chance that teachers will introduce the IAT to the classroom. Or perhaps they will engage consultants who will do so. Teachers may then draw conclusions about implicit racial biases that we cannot be sure are valid.
This was modelled in a recent documentary for Channel 4, a UK TV channel. 11-year-old children were administered a bespoke IAT before being informed that it showed the majority of them viewed white people more favourably than non-white people. They were also put through a number of other experimental treatments such as placing them in race-based ‘affinity’ groups.
In years to come, we may see this as a very strange point in time. While pundits wring their hands, concerned that enduring constants of the educational landscape such as exams are somehow now causing unprecedented levels of stress, it goes almost without remark that, on the basis of questionable but scientific-looking assessments, children are being told that although they may not be aware of it, they are fundamentally racist.
I believe that practices such as these, alongside staples of implicit bias training such as affinity groups, will eventually come to be seen as not only divisive and counterproductive, but as a mild form of abuse.