A proposal for examining and reporting stepwise regressions
- 1 May 1991
- journal article
- research article
- Published by Wiley in Statistics in Medicine
- Vol. 10 (5), 711-715
- https://doi.org/10.1002/sim.4780100505
Abstract
A common mode of regression analysis is a stepwise regression. One of the problems associated with the stepwise algorithm is that the data analyst may wrongly identify the selected variables as the important variables. We propose a simple table as an aid in examining and reporting stepwise regression analyses. The key to our proposal is recognition of ‘close alternatives’, that is, those variables excluded from selection by the inclusion of another variable. This presentation should encourage the analyst and reader to think in terms of sets of variables that, together, represent factors associated with the outcome.Keywords
This publication has 4 references indexed in Scilit:
- Data SplittingThe American Statistician, 1990
- Factors associated with obtaining nipple aspirate fluid: Analysis of 1428 women and literature reviewBreast Cancer Research and Treatment, 1990
- Frequency of Selecting Noise Variables in Subset Regression Analysis: A Simulation StudyThe American Statistician, 1987
- Cross-Validation, the Jackknife, and the Bootstrap: Excess Error Estimation in Forward Logistic RegressionJournal of the American Statistical Association, 1986