ERIC Number: ED322167
Record Type: RIE
Publication Date: 1989-Nov
Testing Interaction Effects without Discarding Variance.
Lopez, Kay A.
Analysis of variance (ANOVA) and multiple regression are two of the most commonly used methods of data analysis in behavioral science research. Although ANOVA was intended for use with experimental designs, educational researchers have used ANOVA extensively in aptitude-treatment interaction (ATI) research. This practice tends to make researchers prone to categorizing continuous variables. Categorizing variables in a non-experimental design using an ANOVA analysis produces in both investigators and their audience/clients the illusion that the investigator has experimental control over the independent variable. Multiple regression prevents the researcher from being lulled into this trap by allowing variables to retain their highest level of scale and requiring that the researcher examine data from several different perspectives. A hypothetical data set is presented to illustrate the superiority of multiple regression in analyzing data from ATI experiments. Multiple regression not only does all that ANOVA can, it also provides more explanatory power. Six data tables are included. (TJH)
Publication Type: Reports - Evaluative; Speeches/Meeting Papers
Education Level: N/A
Authoring Institution: N/A
Note: Paper presented at the Annual Meeting of the Mid-South Educational Research Association (Little Rock, AR, November 8-10, 1989).