The asymptotic relative efficiency of score tests in a generalized linear model with surrogate covariates (Q3799530)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 4066193
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | The asymptotic relative efficiency of score tests in a generalized linear model with surrogate covariates |
scientific article; zbMATH DE number 4066193 |
Statements
The asymptotic relative efficiency of score tests in a generalized linear model with surrogate covariates (English)
0 references
1988
0 references
generalized linear model
0 references
epidemiology
0 references
errors-in-variables model
0 references
surrogate independent variables
0 references
optimal score test
0 references
asymptotic relative efficiency
0 references
0.8846970796585083
0 references
0.8280528783798218
0 references
0.8000733852386475
0 references
0.7876588702201843
0 references
The mathematical problem treated in this paper arose in epidemiology where exposure variables often cannot be measured directly. Suppose that a sample is taken at fixed values for u and z, \((u_ i,z_ i)\) \((i=1,...,n)\), and that \(\{y_ i,w_ i\}\) are independently distributed according to the probability functions \(f(y_ i/x_ i)f(w_ i/u_ i,z_ i)\). This corresponds to the errors-in-variables model of regression analysis. Such models can be treated satisfactorily in the linear case, but epidemiological models usually should be nonlinear. Furthermore, the traditional additive measurement error model cannot adequately represent many surrogate relationships. In this paper the effect of using surrogate independent variables on the efficient score statistic is considered; this leads to the avoidance of the difficulties inherent in the estimation problem. For the null hypothesis of no association, a simple and flexible procedure is proposed to calculate the optimal score test. The asymptotic relative efficiency of this test to the test based upon the true exposures is derived.
0 references