The Effects of Generalizations as Cues on the Learning of Information from Graphs

Abstract

This study investigated the effects of generalizing cues on the learning of material presented in bar graphs. Upper level college stu dents were randomly assigned to either a control or an experimental condition. The experimental group was asked to inspect five graphs to determine the validity of two generalizations used as cues, and the control group was directed merely to notice the results of the experiments. All subjects were posttested using four different types of items designed to measure general inspection of the graph. Results showed that the control group scored significantly higher than the experimental group on all but one of the subtests. Success ful performance with the cues did not result in higher scores on the posttest. GRAPHS ARE OFTEN USED as supplements in sci ence books, and it is generally assumed that material in graphs helps to clarify points made in the regular text. However, research on the effectiveness of graphs (8,9) has shown that they often add little to the understanding of written material and may under certain conditions (10) have a deleterious effect on learning from the text. Research has also shown that supplementary questions and objectives can act as aids to learning from prose mate rials (2,3,4,7). The effectiveness of supplementary cues has been attributed to their ability to induce searching or attending behaviors in the reader. Such cues also help to focus the learners' attention on important aspects of the text. Most studies focusing upon the use of such cues have utilized prose materials, but one study (6) found that questions can also be used to increase the amount learned from graphs. In general, studies utilizing supplemental cues have focused on knowledge-level, recall-type learning rather than on higher level abilities. Two studies are exceptions to this trend. One (5) investigated the effect of analysis and evaluation level questions (1) on written social stud ies materials. This study found that higher level questions resulted in significantly higher performance on an evaluation-level subtest than a similar condition employ ing knowledge-level questions. Further, no differences were found on a knowledge-level test, which should osten sibly favor the group given knowledge-level questions. In a second study involving higher cognitive types of learning (11), high school seniors were presented written materials concerning psychological principles. The treatment con sisted of different types of supplementary questions. One group was asked to name the psychologist associated with a given principle ; a second group was asked to provide an example of the principle from the text; a third group was asked to provide new or original examples of the princi ple; and a control group merely read the text. The group that provided new examples scored significantly better on a posttest than all other groups. The results were explained in the following way: The authors believe that answering application questions facilitates later performance by encour aging students to process the content of the instruc tion more thoroughly, in fact to transform it, in the effort to apply it in a new situation (11: 393). The present study combined the previous lines of re search by measuring the effects of higher cognitive level cues on learning from graphical material. It was hypothe sized that judging the validity of graph-related generaliza tions would result in more thorough processing of the information in the graph and better consequent learning.