NotesFAQContact Us
Collection
Advanced
Search Tips
ERIC Number: ED407425
Record Type: Non-Journal
Publication Date: 1997-Jan
Pages: 10
Abstractor: N/A
Reference Count: N/A
ISBN: N/A
ISSN: N/A
Stepwise Analyses Should Never Be Used by Researchers.
Lockridge, Jewel
Researchers persist in using stepwise regression in spite of problems with this approach. As noted by B. Thompson (1995), three problems accompany the use of stepwise applications. The first is that computer packages may use incorrect degrees of freedom in their computations, resulting in a greater likelihood of obtaining a spurious statistical significance. In the Statistical Package for the Social Sciences, although all the predictor variables explained in the analysis are examined for the initial step, the computer package only shows the degree of freedom corresponding to one predictor variable. Secondly, stepwise methods do not identify the best variable set of a given size correctly. Finally, stepwise methods tend to capitalize on sampling error and tend to produce results that are not replicable. This problem is caused by the uniqueness of sample data and the fact that sampling error in a given sample is not likely to occur in another sample. Researchers should consider and select other available methods for research. (Contains one table and five references.) (SLD)
Publication Type: Reports - Evaluative; Speeches/Meeting Papers
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A