Mautone, P. D. & Mayer, R. (2007). Cognitive aids for guiding graph comprehension. Journal of Educational Psychology, 99, 640-652.
Problem/Purpose
The researchers in this study examined how three different types of graphic organizers improved student comprehension of scientific graphs compared to students that were not exposed to graphic organizers at all. Mautone and Mayer examined the effects of signaling, concrete, and structural graphic organizers on students’ generation of causal and relational statements. Signaling organizers help students to understand the structure of graphs by using highlights, breakdowns, and other tools to convey graph structure. Concrete organizers help students to integrate graphical information through the use of images to connect variables and data in graphs to students’ existing schemata. Structural organizers help students to understand how patterns between new variables work by comparing them to similar patterns followed by other more familiar variables.
This study is of interest because it expands the understanding of how graphic organizers affect students’ comprehension and more specifically, how different kinds of graphic organizers affect different aspects of comprehension in different ways. From a theoretical perspective, this expansion provides data that may help construct a more detailed understanding of cognition. From a practical perspective, the expansion and application of graphic organizers can provide means for improving students’ comprehension of scientific graphs, especially in understanding causal and relational relationships between and among variables
Methods
Mautone and Mayer conducted three experiments.
In the first experiment, 60 students were divided into a control group and a cognitive aids group. Both groups were allowed to view self-paced PowerPoint presentations focused on four graphs derived from introductory college geography texts. The four graphs were each different in type, but all relatively complex and all depicted different aspects of common fluvial processes associated with rivers and streams and the deposits and landforms created by them. After studying the graph slide at their own pace, the control group students were prompted to write a summary of each slide for five minutes. This process was repeated for each of the four slides. The cognitive aids group in experiment one was given a modified version of the same PowerPoint presentation where the graphs were broken apart into multiple slides containing highlights, illustrations, and breakdowns that signaled the structure of the graphs. Students paged through the multiple slides for each graph at their own pace. After viewing the full graph and its preceding cognitive aids, students were then prompted to write a summary of the slide for five minutes. Student responses for both groups were scored based on the number of relational and causal statements. Separate scores were recorded for the number of relational statements and the number of causal statements. Inter-rate reliability was measured to be r=0.91 and r=0.82 and deemed to be high.
Because students in the cognitive aids group had more study time than the control group, Mautone and Mayer conducted multiple regression analyses with time and condition, finding no interaction. Analyses were also conducted using time as a covariate and no differences were found.
To separate the effects of signaling and graphic organizers, Mautone and Mayer conducted a second experiment with a 2x2 design with signaling and concrete organizers present and not present. The procedure for the second experiment was similar to experiment one except that there were four groups: control, signaling, concrete, and signaling with concrete. Scoring was done similarly to experiment one including follow-up analyses to determine that time spent studying was not a critical factor in between group differences.
Mautone and Mayer followed up with a third experiment to test the effects of structural graphic organizers that are intended to prime cognitive organizing and should increase the number of students’ relational statements and not increase the number of students’ causal statements. Again, students were divided into two groups: control and structural. The same graphs were used and the structural group viewed a self-paced PowerPoint embedded with structural cognitive aids. Student statements in the graph summaries were scored as before and time spent studying was analyzed to verify again that time was not a critical factor in the between group differences.
Results
In experiment one, Mautone and Mayer found statistically and practically significant increases in the number of causal and relational statements generated by students in the cognitive aids group. Cognitive aids group students generated an average of 5.3 relational statements (SD=2.19) while control group students generated an average of 3.6 relational statements (SD=1.99). The difference between groups was t(58)=2.96, p=0.004 and Cohen’s d was a medium-to-large 0.76. Cognitive aids group students generated an average of 1.1 causal statements (SD=0.88) while control group students generated an average of 0.4 relational statements (SD=0.62). The difference between groups was t(58)=3.55, p=0.001 and Cohen’s d was a large 0.93. Effects of time spent studying were not statistically significant.
In experiment two, Mautone and Mayer found a significant main effect of signaling graphic organizers on relational statements. Students who were in signaling groups generated an average of 5.14 relational statements (SD=1.90) while students in the other groups generated an average of 4.08 relational statements (SD=1.64). The difference between groups was F(1,98)=9.02, p=0.03 and Cohen’s d was a medium 0.60. There was no main effect for the presence of concrete graphic organizers on relational statements and there was no interaction effect between the presence of concrete and signaling graphic organizers on relational statements. Effects of time spent studying were not statistically significant.
Also in experiment two, Mautone and Mayer found a significant main effect of concrete graphic organizers on causal statements. Students who were in concrete groups generated an average of 1.02 causal statements (SD=0.96) while students in the other groups generated an average of 0.65 causal statements (SD=0.90). The difference between groups was F(1,98)=3.86, p=0.05 and Cohen’s d was a small-to-medium 0.39. There was no main effect for the presence of signaling graphic organizers on causal statements and there was no interaction effect between the presence of concrete and signaling graphic organizers on causal statements. Effects of time spent studying were not statistically significant.
In experiment three, Mautone and Mayer found statistically and practically significant increases in the number of relational statements generated by students in the structural graphic organizers group. Structural group students generated an average of 4.92 relational statements (SD=1.78) while control group students generated an average of 3.00 relational statements (SD=1.69). The difference between groups was t(45)=3.78, p<0.001 and Cohen’s d was a large 1.10. There was no significant difference between groups in the number of causal statements generated. Effects of time spent studying were not statistically significant.
Discussion
Mautone and Mayer conducted three experiments that demonstrate how different types of graphic organizers can improve students’ understanding of scientific graphs in different ways.
The differences in the three types of graphic organizers are not entirely obvious at first, but Mautone and Mayer provide clear descriptions and clarifying examples to explain the types and how they are intended to work. The signaling organizers help students to understand how variables are represented on graphs and how the different pieces of graphs are put together. Especially valuable in understanding signaling organizers is the idea of presenting partial graphs so that students can understand how fundamental parts of a graph are structured relative to each other before additional graph elements are added. Concrete organizers help students related the information on graphs to concepts and knowledge they already possess by using pictures, photographs, and diagrams to illustrate variables and their different values.
Structural organizers are perhaps the most sophisticated because they help students to understand the relationships between new variables by using other relationships the students already understand to help students understand the new relationship. Mautone and Mayer described a clever way to explain curvilinear relationships by using age and driving ability. Extremely young and old drivers have higher rates of accidents that those drivers who are closer to middle age. This structural organizer is used to then explain the curvilinear relationship between river water velocity and particle size with respect to erosion. Brilliant!
Mautone and Mayer addressed the major limitation of this study in their post-experiment analyses on the effects of time. In all three experiments, time was found to not be a statistically significant factor in between group differences.
Critique
Mautone and Mayer use clear language and provide concrete examples that are especially useful for distinguishing the three types of graphic organizers. Tables are clear, well organized and well placed within the text. The descriptions of methodology are detailed and thorough without being repetitive or redundant. Further, the follow-up analyses of time as an independent variable are thoughtful and are summarized clearly and concisely. All pertinent statistical data (means, standard deviations, effect sizes, p values and statistics) are provided clearly. The use of three experiments also provides evidence that the results are replicable, a commonly overlooked aspect of education research studies.
Theoretical, Methodological, and Practical Implications
A major practical implication of this article is the potential to use structural and concrete graphic organizers to promote students’ comprehension of scientific graphs. Structural organizers should help students to understand how variables are related. It would be interesting to conduct further studies to see if extended use of structural graphic organizers led to successful completion of far transfer tasks where new variables that have similar relationships can be understood without introduction using the graphic organizers.
As well, there is a possibility that repeated use of concrete graphic organizers may promote causal understanding and integration of new knowledge into existing schemata. Again, there may be a way that repeated use of concrete organizers could lead to far transfer ability and more generalized use of integration in novel situations without explicit use of concrete organizers. More research on these possible outcomes is needed.
Mautone and Mayer’s study extends and adds to theoretical understanding of cognition and graphic organizers by providing additional levels of detail. Signaling graphic organizers that focus on structure may help students to better understand relations in graphs and between variables. Concrete graphic organizers may help students to better understand causal relations in graphs and between variables. Structural graphic organizers can help students to better understand more complex relations between variables, such as curvilinear relationships. Graphic organizers may include both concrete and signaling design principles and thereby statistically increase students understanding of relations and integration of new knowledge as shown by increased numbers of accurate causal and relational statements. Mautone and Mayer’s work indicates that structure and integration may be independent cognitive processes.
Mautone and Mayer used a three-experiment approach with self-paced study time and timed responses. The use of replication and variation within and between the three experiments allows their method to demonstrate that their results with graphic organizers are reproducible and to examine details of graphic organizer and cognition theory through variations in the graphic organizers studied. Other research studies may benefit by following similar replication and variation strategies where multiple experiments are combined into one study. Pilot tests are commonly used in studies, but Mautone and Mayer went much further by using three full-scale experiments and not a mini-experiment as an antecedent to the main experiment.
