I often play a collaborative computer game with my family called Knowledge is Power. The basic premise of the game is that you have to answer general knowledge questions to gain points. But built into the game are lots of ways to obstruct your opponents. So, even if the adults in the room have more background knowledge than younger players, there is balance in the gameplay because of all kinds of other variables.

I was recently reminded of this game when I read a new meta-review by Simonsmeier and colleagues (2021). In it, they look at the “Knowledge is Power (KiP) hypothesis”, which builds on a much earlier tradition of centring knowledge in learning, for example from psychologist David Ausubel, who argued back in the 1960s that “the most important single factor influencing learning is what the learner knows already”.

Simonsmeier and colleagues suggest that a simplistic version of the hypothesis would assume that (relevant) prior knowledge always aids learning. However, a more plausible version would be that incorrect prior knowledge can be a barrier to learning and correct prior knowledge can be conducive to learning.

Before they begin the meta-analysis, the authors describe some of the positive and negative effects of prior knowledge.

On the positive side, prior knowledge can direct students’ attention, support them to interpret and encode new information, facilitate text comprehension and help them to evaluate the credibility of sources.

On the other hand, prior knowledge can sometimes include misconceptions. And when correct knowledge is incomplete, this can lead students to draw incorrect conclusions. There is also the risk of the “Einstellung effect” (a perceptual bias where learners with high prior knowledge are biased to certain types of information, preventing them from finding newer or better solutions) and negative transfer (in which correct knowledge in one domain can hamper learning in another).

Far from being universally effective, prior knowledge brings with it a combination of positive and negative effects. Perhaps, then, we need to help students reach some threshold of productive prior knowledge that will ensure that the knowledge they have is useful.

But the authors of the research go further than this. They also look at the body of literature on the topic, exploring 493 studies from 47 countries, with findings from 685 independent samples and 126,050 participants in total. Together they represent 8,776 effect sizes. Publication years range from 1965 to now, with 68 per cent published within the past 10 years.

The analysis finds that although there is a strong relationship between pre-test and post-test knowledge, the predictive power of prior knowledge for learning is low.

The authors say that “this strong variability falsifies general statements such as ‘knowledge is power’ or ‘prior knowledge is negligible’”. In other words, whether prior knowledge has positive, negative or negligible effects on learning really depends on the conditions under which learning takes place.

Right now, we don’t yet know enough about those conditions to be able to fully untangle the mediating effects of prior knowledge and determine useful levels of prior knowledge.

I guess what this tells us is that knowledge is indeed power but it’s a power that takes the form of a double-edged sword: it can be an excellent tool to support further learning but it can also hinder it.

Christian Bokhove is associate professor in mathematics education at the University of Southampton and a specialist in research methodologies

This article originally appeared in the 2 July 2021 issue under the headline “Why knowledge isn’t always power”