MODELS OF VARYING PARAMETRIC FORM IN CASE-REFERENT STUDIES
- 1 January 1982
- journal article
- research article
- Published by Oxford University Press (OUP) in American Journal of Epidemiology
- Vol. 115 (1), 129-137
- https://doi.org/10.1093/oxfordjournals.aje.a113267
Abstract
The analysis of case-referent data can be based on a consideration of case and referent counts in various exposure categories as the realizations of a set of binomial processes. After appropriate modeling of the binomial parameters, a joint likelihood function can be formed and maximized to obtain estimates of the parameters constituting the model elements. The procedure has been applied to the problem of additive and multiplicative models of disease incidence rates, as encountered in case-referent studies. Likelihood ratios can be used to compare models with equal numbers of parameters. These ratios do not lead to significance tests, but to estimates of the relative degree of corroboration of different hypotheses by the data at hand.Keywords
This publication has 11 references indexed in Scilit:
- Dose-response in case-control studies.Journal of Epidemiology and Community Health, 1980
- Additive excess risk model for epidemiologic interaction in retrospective studiesJournal of Chronic Diseases, 1980
- Logistic disease incidence models and case-control studiesBiometrika, 1979
- Some results on the estimation of logistic models based on retrospective dataBiometrika, 1979
- INTERACTION IN EPIDEMIOLOGIC STUDIESAmerican Journal of Epidemiology, 1978
- THE USE OF LOGISTIC REGRESSION FOR MODELLING RISK FACTORS: WITH APPLICATION TO NON-MELANOMA SKIN CANCERAmerican Journal of Epidemiology, 1978
- Retrospective Studies and Failure Time ModelsBiometrika, 1978
- The Estimation and Interpretation of Attributable Risk in Health ResearchBiometrics, 1976
- Use of the Logistic Model in Retrospective StudiesBiometrics, 1976
- Dynamic Risk Analysis in Retrospective Matched Pair Studies of DiseaseBiometrics, 1962