Domain-specific prior knowledge and learning: A meta-analysis

It is often hypothesized that prior knowledge strongly predicts learning performance. It can affect learning positively mediated through some processes and negatively mediated through others. We examined the relation between prior knowledge and learning in a meta-analysis of 8776 effect sizes. The stability of individual differences, that is, the correlation between pretest and posttest knowledge, was high (r(P)(+) = .534). The predictive power of prior knowledge for learning, i.e., the correlation between pretest knowledge and normalized knowledge gains, was low (r(NG)(+) = -.059), almost normally distributed, and had a large 95% prediction interval [-.688, .621]. This strong variability falsifies general statements such as knowledge is power as well as the effect of prior knowledge is negligible. It calls for systematic research on the conditions under which prior knowledge has positive, negative, or negligible effects on learning. This requires more experiments on the processes mediating the effects of prior knowledge and thresholds for useful levels of prior knowledge.

Automated summary

The impact of prior knowledge on learning performance is not linear, with high individual differences and low predictive power. Further research is needed to explore the conditions under which prior knowledge has positive, negative, or negligible effects on learning.