eCite Digital Repository

Goodness-of-fit statistics for log-link regression models

Citation

Quinn, SJ and Hosmer, DW and Blizzard, CL, Goodness-of-fit statistics for log-link regression models, Journal of Statistical Computation and Simulation, 85, (12) pp. 2533-2545. ISSN 0094-9655 (2015) [Refereed Article]

Copyright Statement

Copyright 2014 Taylor & Francis

DOI: doi:10.1080/00949655.2014.940953

Abstract

The use of log binomial regression, regression on binary outcomes using a log link, is becoming increasingly popular because it provides estimates of relative risk. However, little work has been done on model evaluation. We used simulations to compare the performance of five goodness-of-fit statistics applied to different models in a log binomial setting, namely the Hosmer–Lemeshow, the normalized Pearson chi-square, the normalized unweighted sum of squares, Le Cessie and van Howelingen's statistic based on smoothed residuals and the Hjort–Hosmer test. The normalized Pearson chi-square was unsuitable as the rejection rate depended also on the range of predicted probabilities. The Le Cessie and van Howelingen's test statistic had poor sampling properties when evaluating a correct model and was also considered to be unsuitable in this context. The performance of the remaining three statistics was comparable in most simulations. However, using real data the Hjort–Hosmer outperformed the other two statistics.

Item Details

Item Type:Refereed Article
Keywords:log binomial, log link, risk ratio, relative risk
Research Division:Mathematical Sciences
Research Group:Statistics
Research Field:Statistics not elsewhere classified
Objective Division:Health
Objective Group:Other Health
Objective Field:Health not elsewhere classified
Author:Blizzard, CL (Associate Professor Leigh Blizzard)
ID Code:98192
Year Published:2015 (online first 2014)
Funding Support:National Health and Medical Research Council (490000)
Deposited By:Menzies Institute for Medical Research
Deposited On:2015-02-04
Last Modified:2017-11-01
Downloads:0

Repository Staff Only: item control page