Using Multiple Regression to Predict Achievement: A Reply to Klingele and Warrick; Penny and Bond

Publisher Website
Google Scholar
Abstract
This response critiques Klingele and Warrick's (1990) use of stepwise regression to understand the effect that independent variables have on measures of achievement. Some researchers argue that a more appropriate technique would be the use of standardized regression coefficients (beta weights) generated in nonstepwise multiple-regression models. This critique also challenges the assertion in Penny and Bond's (1991) reply that ethnicity should not be included in models of achievement. Controlling for the effect of ethnicity, a nonmanipulable variable, in multiple-regression modeling allows researchers to determine the unique effect on achievement of those variables that educators can manipulate.

This publication has 2 references indexed in Scilit: