Abstract
When the categorical canonical correlation analysis studied in the authors' previous paper is applied to a set of real data, it is desirable to have some idea about the reliability of the estimates obtained, scores, weights and eigenvalues. The purpose of this paper is to show by means of the δ-method that the estimates follow asymptotically a normal distribution with the true values of the parameters as the means and also to find the formulas for the asymptotic variances and covariances of the estimates. As an illustration the estimated values and their asymptotic variances are computed for Hayashi's data dealing with the label preference problem.