|
Activity Number:
|
50
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 8, 2004 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
ENAR
|
| Abstract - #301427 |
|
Title:
|
A Goodness-of-fit Test for Logistic Regression Models with Continuous Predictors
|
|
Author(s):
|
Xianjin Xie*+ and Jane Pendergast and William R. Clarke
|
|
Companies:
|
University of Iowa and University of Iowa and University of Iowa
|
|
Address:
|
1100 Arthur St., G6, Iowa City, IA, 52240,
|
|
Keywords:
|
goodness of fit ; logistic regression
|
|
Abstract:
|
When continuous predictors are present, classical Pearson and deviance goodness-of-fit tests to assess logistic model fit break down. The Hosmer-Lemeshow goodness-of-fit statistic is often used in these situations. Their procedure groups observations into G bins according to the percentiles of the estimated probabilities. It uses a Pearson chi-square statistic with G-2 degrees of freedom to compare the observed frequency of events to that expected using the model's average predicted value in each group. While simple to perform with satisfactory properties, it provides no further information on the source of any detectable lack of fit. Tsiatis (1980) proposed an alternative statistic which partitions the covariate space and uses a score statistic to test for regional effects. We propose a new method for goodness-of-fit testing which uses a very general partitioning strategy in the covariate space and is based on either a Pearson statistic or a score statistic. Properties of the proposed statistics are discussed and simulation studies comparing it to the existing tests are presented, demonstrating its usefulness in practice.
|
