Testing odds-ratio equality for several diseases
WHITTEMORE, ALICE S.; HALPERN, JERRY; Department of Health Research and Policy, Stanford University School of MedicineStanford, California 94305, U.S.A.
Журнал:
Biometrika
Дата:
1989
Аннотация:
We present a test for the hypothesis that odds-ratios are constant across two or more diseases, using matched case-control data. The test requires less computing than the usual conditional likelihood score test (Prentice, 1979) because it excludes the control covariates. We give a practical algorithm to compute the test statistic and discuss its power.
190.8Кб