The estimation of the accuracy of an indicator in

Often, the researcher in educational testing theory, would like to have a model that simply explains the obtained scores. In order to establish if such a model is adequate, an indicator based on the eigenvalues of the empirical covariance matrix of the scores is used. This paper presents the standard methodology to construct this indicator, its bootstrap estimate of the standard error and its percentile bootstrap confidence interval. Finally, for a case study, a program that implements in Matlab the described algorithm provides the desired results.