Specificity and learning with graphics

Presents a method for characterising a constrained set of representational media for the purpose of making an effective instructional choice between alternatives within a context. An interpretation of the specificity principle is made and developed into a quantitative metric which, it is argued, if the principle were sufficient, would enable the prediction of learned outcomes. This metric is applied to two quite different graphical systems for demonstrating formal reasoning. During the application of the metric to two systems, a discovery is made in the form of an intermediate representation. Outcomes of the metric application are treated as hypothetical results, the significant negation of which, in empirical study, would demand revision of the principle. Empirical work using interactive graphic representational systems for both Venn and Euler diagrams returned results of both a qualitative and quantitative nature, showing that, although many systems improve students´ performance and help to de-bias most reasoning errors for most subjects, systems with greater specificity did not yield better learning outcomes-indeed quite the opposite. Even across different learner groups with and without advanced maths training (in itself a significant variable), the effects of specificity appear to degrade learning outcomes