About 20 years ago, psychologist Lynn Liben presented an model of the relationships among a learner, an external (i.e. not mental) representation, and those aspects of the real world represented by the representation (the “referent”). Liben notes that the learner can learn either through direct interactions with the real world or through interactions with a representation. In cases where learning is mediated through a representation, the learner has to draw on another kind of cognitive resource, which Liben calls “representational strategies.” In Liben’s work, the external representation is a map, and the “representational strategies” include things like map scale and map projection. In this white paper, Kastens extends Liben’s framework to any kind of visible representation, and in particular to data visualizations.