On the Impact of Variable Selection in Fitting Regression Equations
Lecture notes in economics and mathematical systemsPublished 1 January 1988
Darcy A. Freedman, William Navidi, Stephen C. Peters
Citations50
Generate an AI Snapshot to get a quick, structured summary of this paper.
Study Snapshot
ObjectiveStudy objective
MethodsResearch methodology
PopulationPopulation studied
Sample sizeSample sizes
OutcomesStudy outcomes here
ResultsStudy results comes here
LimitationsResearch study limitations comes here
A concise AI-generated summary of the paper will appear here once you click Generate AI Snapshot.
Abstract
Consider developing a regression model in a context where substantive theory is weak. Search procedures are often used to develop the equation: eg, fitting the equation, dropping insignificant variables, and refitting. As is well known, this can seriously distort the conventional goodness-of-fit statistics. Furthermore, the bootstrap and jackknife may not help in high-dimensional cases.
Keywords
Computer ScienceMathematics
The Annals of StatisticsBootstrap Methods: Another Look at the Jackknife
17,446 Citations1979B. Efron
Society for Industrial and Applied Mathematics eBooksThe Jackknife, the Bootstrap and Other Resampling Plans
7,903 Citations1982Bradley Efron
The Annals of StatisticsSome Asymptotic Theory for the Bootstrap
1,640 Citations1981Peter J. Bickel, David A. Freedman
Econometrics: Alchemy or Science?
627 Citations2000David F. Hendry
Interpreting and Using Regression
594 Citations1982Christopher H. Achen
Journal of the American Statistical AssociationBootstrapping a Regression Equation: Some Empirical Results
400 Citations1984David A. Freedman, Stephen C. Peters
Journal of Educational StatisticsAs Others See Us: A Case Study in Path Analysis
310 Citations1987Darcy A. Freedman
Journal of the American Statistical AssociationCohort Analysis in Social Research: Beyond the Identification Problem.
252 Citations1986Norval D. Glenn, William M. Mason +1 more
Journal of the American Statistical AssociationCross-Validation, the Jackknife, and the Bootstrap: Excess Error Estimation in Forward Logistic Regression
236 Citations1986Gail Gong
Three estimates of the excess error are considered: cross-validation, the jackknife, and the bootstrap, for a moderately complicated prediction rule, involving a variable-selection procedure based on forward logistic regression.
Journal of Business and Economic StatisticsBootstrapping an Econometric Model: Some Empirical Results
169 Citations1984David A. Freedman, Stephen C. Peters
Journal of the American Statistical AssociationHow Many Variables Should Be Entered in a Regression Equation?
155 Citations1983Leo Breiman, Darcy A. Freedman
The optimal number of regressors is determined to minimize mean squared prediction error and is shown to be a small fraction of the number of data points and the Sp criterion provides an asymptotically optimal rule for theNumber of variables to enter.
Journal of Business and Economic StatisticsForecasting Accuracy of Alternative Techniques: A Comparison of U.S. Macroeconomic Forecasts
153 Citations1986Stephen K. McNees
This article compares the accuracy of these VAR forecasts with that of several prominent forecasters and provides an interpretation of this evidence and speculations on the potential of these two approaches to economic modeling and forecasting.
Journal of the American Statistical AssociationBootstrapping a Regression Equation: Some Empirical Results
104 Citations1984David A. Freedman, Stephen C. Peters
Statistics and the Scientific Method
82 Citations1985David A. Freedman
Regression models have not been so useful in the social sciences, and such models are contrasted with successful mathematical models in the natural sciences, including Kepler’s three laws of motion for the planets.
Journal of Personality and Social PsychologySpecious causal attributions in the social sciences: The reformulated stepping-stone theory of heroin use as exemplar.
80 Citations1983Diana Baumrind
It is argued that O'Donnell and Clayton's version of the regularity model is not sufficient to establish cause and that the planning of social interventions both presumes and requires a generative rather than a regularity causal model.
RePEc: Research Papers in EconomicsAn Analysis of Annual and Multiperiod Qtrly Forecasts of Aggregate Income, Output, and the Price Level.
62 CitationsVictor Zarnowitz
Statistical ScienceRegression Models for Adjusting the 1980 Census
51 Citations1986Darcy A. Freedman, William Navidi
Journal of Business and Economic StatisticsSome Notes on the Bootstrap in Regression Problems
46 Citations1984Stephen C. Peters, Darcy A. Freedman
This section will include short papers and notes on topics of interest to readers, including measurement problems, revisions of major business and economic time series, forecasting experiments, computer programs, experience with applications of new methods, and comments on previously published papers.
Journal of the American Statistical AssociationCross-Validation, the Jackknife, and the Bootstrap: Excess Error Estimation in Forward Logistic Regression
44 Citations1986Gail Gong
Journal of Business and Economic StatisticsOn Energy Policy Models
41 Citations1983David A. Freedman, Thomas J. Rothenberg +1 more
Journal of the American Statistical AssociationHow Many Variables Should be Entered in a Regression Equation?
37 Citations1983Leo Breiman, Darcy A. Freedman
Journal of ForecastingUsing the bootstrap to evaluate forecasting equations
32 Citations1985Stephen C. Peters, David A. Freedman
The Journal of BusinessSome Pitfalls in Large Econometric Models: A Case Study
20 Citations1981David A. Freedman
UC BerkeleyEconometrics and the law: a case study in the proof of antitrust damages
4 CitationsDaggett, R. S., Author, Freedman, D., Author