Comparison of Component Principal Analysis and Latent Class Analysis be the Example of Labour Values

Latent variables are complex concepts, which require indirect measurement on the basis of observable variables. There are numerous methods for working with latent variables. The choice of certain method bears on the type of observed and latent variables. Given sociologists most frequently use observable variables with ordinal scale, the research is dedicated to the comparison of categorical principal component analysis (CatPCA) and latent class analysis (LCA), which are aimed to work with categorical original data. The fundamental difference between these methods is the consideration of latent variables as discrete (LCA) or continuous (CatPCA). The goal of the research is to determine if the work values of generations X and Y in Russia are discrete or continuous. The empirical object of the study is generation X born in the period 1961 – 1981 and generation Y (1982 – 2005) in Russia. Generations were distinguished according to the theory of American sociologists Strauss and Howe. The results of two statistical methods and its further comparison on the basis of objective (statistical) and subjective (interpretation) criteria show that CatPCA satisfies requirements better than LCA. The possible reason may lie in the fact that LCA treats latent variables as discrete and one-dimensional; therefore, each respondent is characterized by only one value. Whereas, latent variables are not mutually exclusive in CatPCA, thus every person can have several values. Overall, CatPCA extracted three work values such as the value of the interesting job and flexible schedule, the value of financial benefits, the value of secure employment. As a result of statistical tests, it can be concluded that there are no differences in work values of generations X and Y in Russia. It is worthwhile to mention that the theory of Strauss and Howe cannot be fully extrapolated on Russian society. The possible reason is the social and cultural differences in the history of two nations.