Maximizing the reliability of marketing measures under budget constraints

Abstract

Generalizability theory provides a framework for examining the dependability of measurements in
marketing research. When limited resources are available determining the optimal number of observations
to use in a marketing measurement design that will maximize reliability is not a simple task. This
paper presents a method for determining the optimal number of observations to use in fully-crossed,
univariate and multivariate two- and three- facet measurement designs when resource constraints are
imposed.