PubMed Health⌕ Search
Home/Archive/PubMed

PubMed · 11315021

Efficiency robust tests for survival or ordered categorical data.

B Freidlin·M J Podgor·J L Gastwirth
View original source ↗Export metadata ↓

Abstract

The selection of a single method of analysis is problematic when the data could have been generated by one of several possible models. We examine the properties of two tests designed to have high power over a range of models. The first one, the maximum efficiency robust test (MERT), uses the linear combination of the optimal statistics for each model that maximizes the minimum efficiency. The second procedure, called the MX, uses the maximum of the optimal statistics. Both approaches yield efficiency robust procedures for survival analysis and ordinal categorical data. Guidelines for choosing between them are provided.

Explore related subjects

BiometryData Interpretation, StatisticalHumansLinear ModelsRandomized Controlled Trials as TopicSurvival AnalysisJournal ArticleResearch Support, U.S. Gov't, Non-P.H.S.

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

B Freidlin, M J Podgor, J L Gastwirth. 1999. Efficiency robust tests for survival or ordered categorical data.. https://doi.org/10.1111/j.0006-341x.1999.00883.x

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related citations

PubMed2001

Effect of pH on the heat resistance of spores comparison of two models.

All published models describing the effect of pH on the heat resistance of spores can be regarded either as a linear first degree equation or a linear second degree equation. This work aimed to compare both models from three sets of published data for, Clostridium sporogenes and Bacillus stearothermophilus, respectively. The relative quality of fit of each model with respect to the other depends on the species, the strain and the heating temperature. Parameter estimation was more reliable for the second degree model than for of the simple first degree equation. However, in the case of acidic foodstuffs, predictions obtained from the second degree model are more sensitive toward errors of parameter values. The second degree model is better from the point of view of safety at most frequent ranges of pH of foods. Moreover, for Clostridium botulinum the goodness of fit of this model is clearly higher than that of the first degree equation. If this observation is confirmed by further work, the second degree model in application of standard calculations of heat processes of foods would be preferred.

Biometry↗
PubMed2001

Development and validation of a combined temperature, water activity, pH model for bacterial growth rate of Lactobacillus curvatus.

A model was established to predict growth rate as a function of temperature, pH and water activity. The model is based on two, earlier developed models, one for growth rate as a function of temperature and water activity and the other for growth rate as a function of temperature and pH. Based on the assumption that combinatory effects between pH and water activity do not exist, the two models were multiplied to produce one overall model. The overall model was then fitted to data sets measured earlier, and the parameters of the model were determined. A new data set with values for controlling variables outside the range of the earlier developed model was then used to validate the overall model statistically. The model was well able to extrapolate outside the measured data range. Finally, the model was updated with all measured data. No significant changes in the parameters were found. The approach followed underpins the gamma concept, since in the gamma concept it is assumed that the effects of controlling variables can be multiplied, and cardinal parameters are not a function of other variables (temperature, pH, and water activity).

Biometry↗