The reliability of economic textbook readability indexes as a measure of cognitive gain : a comparative analysis.

No Thumbnail Available
Deel, Rebecca
Journal Title
Journal ISSN
Volume Title
Middle Tennessee State University
This study sought to establish whether readability and cognitive gain are related in principles of economics textbooks, and thus, whether readability indexes are effective measures of text earnability. The research was conducted at Middle Tennessee State University, Murfreesboro, Tennessee, during the 1995 spring semester. Five principles of macroeconomics classes were utilized involving 81 students.
In each class, students were given three readings either from a textbook judged to be difficult or from one judged to be easy to read in an alternating pattern. Before and after the three readings each student was tested on three topics treated in similar fashion in both books. The first or narrative topic covered exchanged rates while the second and third topics were quantitative and graphical and involved cost analysis and monopoly profit maximization. Information gain then was measured by the difference in the pre- and posttest mean scores.
Statistical significance of the mean score difference was indicated by utilizing the standardized t-distribution test. When the statistical test was applied, no significant differences were found in cognitive gain for the books' narrative and graphical sections. This finding suggests that readability indexes are not indicative of potential information gain for textbook material presented in the narrative or graphical form. However, in the quantitative section, a statistically significant difference in information gain emerged, suggesting that readability indexes are indicative of possible cognitive gain.
A least squares regression model was also developed to explore the interaction between student demographic characteristics, readability, and cognitive gain. Of six demographic variables included in the model, only student sex and class standing were statistically significant. Textbook readability index value proved not to be statistically significant as an explanatory variable.