2013/10/15_María Álvarez Hernández (University of Vigo)

Classic Wald Method and Score Method about a linear combination of proportions

2013/10/15 – 12:00 h | María Álvarez Hernández, University of Vigo

Abstract

Statistical methods for carrying out asymptotic inferences relative to one or two independent binomial proportions are very frequent in the literature. However inferences about a linear combination of K independent proportions L=\Sigma ?ipi have had very little attention paid to them (focused exclusively on the classic Wald method), despite their great practical importance. This investigation approaches the problem from the more efficient viewpoint of the score method, gives in addition a general proof of Agresti’s heuristic method (consisting of adding a certain number of successes and failures to the original results before applying Wald’s method).

Numerical analysis indicates that the score method is the optimal method in comparison with other methods, with the additional theoretical advantage that it also verifies the desirable properties of “spatial convexity” and “parametric convexity”.