Latent rank analysis (LCA) is a method for assessing non-sampling errors, especially measurement error of categorical data. [1] has introduced four latent rank modeling approaches, namely,
parameterization of the probabilistic model, the log-linear model, the adjusted path model, and the graph model using path graphs. These
models are interchangeable. Latent rank probabilistic models express the likelihood of a cross-ranking table of variables in terms of the conditional and marginal probabilities associated with each cell of this table. In this approach, the model parameters are estimated using the EM algorithm.
For testing the latent rank model, the chi-square statistic is introduced as a goodness-of-fit criterion. This paper uses LCA and data from a small survey
to calculate the misclassification error (a type of measurement error) of the proportion of students who failed at least one course and the misclassification error of the proportion of students who were suspended at least once.