Abstract: Latent class analysis is a statistical technique used to identify subtypes of related cases. In the method, statistic model is built according to relationship between categorical observed and latent variables, and then manifest performances of each latent classes are respectively described based on the related parameters in the model. The analytical approach is mainly used in (1) classification of psychological or behavioral traits; (2) controlling the systematical error of cognitive experiments caused by individual differences; (3) assessing accuracy of clinical psychology diagnostics; (4) analyzing difficulty, discrimination, reliability and structure of psychology measurement. Both strong and weak points are considered, which indicates that, in respect of future development in the psychometric application, the method will be combined with other theoretical frameworks of measurement and also applied to longitudinal and multilevel data. Moreover, the application still sees room for improvement in its methodological technique.

Key words: latent class, cluster model, conditional probability, latent class probability