ERIC Number: ED316583
Record Type: Non-Journal
Publication Date: 1990-Jan
Three Reasons Why Stepwise Regression Methods Should Not Be Used by Researchers.
L. A. Marascuilo and R. C. Serlin (1988) note that stepwise regression is a method used frequently in social science research. C. Huberty (1989) characterizes such applications as being "common". In support of this latter statement, a review of dissertations by B. Thompson (1988) demonstrated that dissertation students frequently use stepwise methods in their data analyses. Unfortunately, stepwise regression can lead to serious misinterpretations of results. Three fundamental problems with traditional stepwise regression analysis are discussed in this paper. Firstly, most researchers who use "canned" computer programs do not use the correct degrees of freedom when evaluating changes in explained variance. Secondly, there is a tendency for some researchers to incorrectly interpret order of selection and entry of the "g" predictor variables as defining the "best" subset of predictors of size "g". Thirdly, some researchers incorrectly consult order of entry information to evaluate the importance of various predictor variables. A small data set is used to illustrate the difficulties with stepwise methods, and it is argued that stepwise methods should not be used by researchers when the desired outcome is accurate variable selection or ordering of variables with respect to importance. Three data tables are included. (Author/TJH)
Publication Type: Information Analyses; Reports - Evaluative; Speeches/Meeting Papers
Education Level: N/A
Authoring Institution: N/A