Multivariable modeling with cubic regression splines: A principled approach

Patrick Royston () and Willi Sauerbrei
Additional contact information
Patrick Royston: Cancer Group, MRC Clinical Trials Unit
Willi Sauerbrei: University Medical Center, Freiburg

Stata Journal, 2007, vol. 7, issue 1, pages 45-70

Abstract: Spline functions provide a useful and flexible basis for modeling re- lationships with continuous predictors. However, to limit instability and provide sensible regression models in the multivariable setting, a principled approach to model selection and function estimation is important. Here the multivariable frac- tional polynomials approach to model building is transferred to regression splines. The essential features are specifying a maximum acceptable complexity for each continuous function and applying a closed-test approach to each continuous pre- dictor to simplify the model where possible. Important adjuncts are an initial choice of scale for continuous predictors (linear or logarithmic), which often helps one to generate realistic, parsimonious final models; a goodness-of-fit test for a parametric function of a predictor; and a preliminary predictor transformation to improve robustness. Copyright 2007 by StataCorp LP.

Keywords: mvrs; uvrs; splinegen; multivariable analysis; continuous predictor; regression spline; model building; goodness of fit; choice of scale (search for similar items in EconPapers)
Date: 2007

Downloads: (external link)
http://www.stata.com/bookstore/sjj.html link to individual issue purchase
http://www.stata-journal.com/software/sj7-1/st0120/ (text/html)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: http://EconPapers.repec.org/RePEc:tsj:stataj:v:7:y:2007:i:1:p:45-70

Ordering information: This journal article can be ordered from
http://www.stata-journal.com/subscription.html

Access Statistics for this article

Stata Journal is edited by H. Joseph Newton and Nicholas J. Cox

More articles in Stata Journal from StataCorp LP
Series data maintained by Christopher F. Baum ().