Abstract
This dissertation examined how useful analogies are for teaching scientific concepts in astronomy and biology, relative to expressing those same concepts in a traditional literal form.
This comparison — analogical versus literal expression — was made in the context of variations in (1) how explicitly the idea of a scientific analogy was explained to subjects, (2) how much cognitive effort subjects were required to expend during learning, and (3) the format of the learning material (i.e., verbal statements either accompanied or unaccompanied by pictures). The highest level of learning, as measured both by performance on a multiple-choice test and by delayed cued recall of statements, obtained for those subjects who received three components simultaneously: the scientific analogies, an explanation of analogy, and a depiction of verbal statements in accompanying pictures (Experiment 1). The multiple-choice questions were directed toward either basic, detail-level knowledge or abstract, inference-level knowledge. Subjects scored higher on basic-level than inference-level questions, but this difference was significantly greater for concepts expressed literally than for those expressed analogically; this finding suggests that analogies foster inference-level thinking. Subjects' performance was hindered when they were required to generate the familiar domain of the analogy, either through drawing (Experiment 4) or writing (Experiment 2). But subjects displayed remarkable comprehension of concepts when they were required to map the familiar domain onto the unfamiliar domain (Experiment 6); here the difference between basic-level and inference-level knowledge evaporated completely. Males typically outperformed females on all measures; this effect appeared to be based on the males1 greater background education in the physical sciences. Little relationship was found between imagery ability and learning. The results of this dissertation are interpreted within a new framework of analogical learning.
