Expert Blind Spot

What it is

Nathan, Koedinger and Alibali introduced the term "expert blind spot" in a paper for the Third International Conference on Cognitive Science, held in Beijing in 2001. Their title states the claim: Expert Blind Spot: When Content Knowledge Eclipses Pedagogical Content Knowledge. The argument is that deep knowledge of a subject can backfire in teaching. An expert tends to assume that learners find abstract, formal representations as accessible as the expert does. That conflicts with how students actually develop, where informal reasoning often comes first and formal reasoning is built on it.

The authors drew on mathematics and on language arts, and concluded that teachers should use assessment practices that test their own assumptions about how students develop. The recommendation matters because it treats the problem as one of calibration. An expert who checks a prediction against real student answers can correct it. An expert who relies on introspection cannot, because the introspection is the source of the error.

The expert blind spot is an application of the curse of knowledge to the classroom. The curse of knowledge, named by Camerer, Loewenstein and Weber in 1989, is the general difficulty of imagining not knowing something. The blind spot is that difficulty at work in the specific task of judging what a learner will find hard. See Curse of Knowledge.

In effect

In research: ranking algebra problems

The central study had 105 school teachers rank a set of algebra problems from easiest to hardest for beginning algebra students. The problems varied by type (arithmetic or algebraic), by verbal form (story or word equation) and by symbolic format. The comparison data came from two samples of ninth-graders who had completed a year of algebra, 76 and 171 students, from earlier work by Koedinger and Nathan (1999). The students solved equations correctly less than 30 per cent of the time and verbal problems more than 50 per cent of the time, in both samples.

The 39 high school teachers judged verbal problems harder than symbolic ones and rated algebra story problems hardest, on the reasoning that students would find pure mathematics easiest. Their ranking was unrelated to the students' performance (a rank correlation of zero). The 30 middle school teachers predicted that story and word problems would be easiest, and their ranking matched the students' performance well (a rank correlation of 0.73). Knowing more algebra did not help the high school teachers see the problems as a beginner does.

In everyday life: choosing what to teach

The same paper reports a comparison from language arts, based on Grossman (1990), of six secondary English teachers who were all strong in literature and grammar. Three had formal teacher education and three did not. Asked to plan a hypothetical ninth-grade literature class, the teachers without teacher education chose about 72 per cent canonical texts and favoured a text-centred, analytical approach. The teachers with teacher education chose about 78 per cent adolescent literature and 22 per cent canonical texts, and focused on connecting the material to students' experiences and stage of development. The sample is six teachers, so it illustrates the idea and does not test it.

In design: instructions written by experts

The blind spot is not confined to classrooms. Pamela Hinds's 1999 study of expert prediction found, according to secondary accounts, that experts asked to predict how long novices would need to learn a new phone task, and what would go wrong, were too optimistic about both. The abstract was not available for this page, so the result is reported in direction only. For anyone writing a manual, the lesson is that the steps the writer finds trivially obvious are the ones most likely to be missing, and that the writer's own rereading will not reveal the gaps.

What it does not say

The evidence is thinner than the idea's reputation suggests. The founding paper is a conference proceeding, and the English teaching comparison rests on six teachers. A later study of pre-service and in-service teachers in an engineering-inspired training programme, by Anthony Petrosino and Prateek Shekhar (International Journal of Engineering Education, 2018), had participants rank six algebraic representation problems by difficulty for beginning students. It found no significant differences between pre-service and in-service teachers, and the authors read this as showing that teaching experience does not change how teachers judge difficulty. That is consistent with a persistent blind spot, but it is a ranking task with one subject area, and its figures are not reported here.

The studies do not show that every expert misjudges every learner, nor how large the effect is in general. The curse of knowledge more broadly is well supported in direction and disputed in size, since a 2014 replication of a false-belief version found less than half the original effect. The expert blind spot should be read the same way.

Who developed it

Mitchell Nathan, Kenneth Koedinger and Martha Alibali, in 2001. Nathan's university page sits at the University of Wisconsin-Madison. See Mitchell Nathan.

---