ABSTRACT
Using models for agreement was introduced in Section 4.7 of Chapter 4, where the logistic model was employed to measure the effect of two tests on the scores used to diagnose tuberculosis with a skin test. Two populations were involved and the objective is to measure the effect of the two tests on agreement and to compare the agreement between the two populations. A similar approach is taken here, where the logistic linear model is again employed, however, the approach is extended in order to examine more subtle effects of agreement between two or more raters and to incorporate factors that could have a possible infl uence on the scores of the raters and hence on the various measures that estimate agreement.
