The purpose of the study is to analyze the abstracts in the psychometric research field in recent 20 years. Over 1,500 abstracts of the journals were extracted from three major journals; Psychometrika, Journal of Educational Measurement, and Journal of Educational and Behavioral Statistics and analyzed them using Latent Dirichlet allocation topic modeling approach. The results suggested the six extracted topics as followings: Factor analysis, Bayesian estimation & SEM, IRT, general psychometric words, general academic words, and weight/scale/regression and agreement analysis. We examined how the research topics have changed and found out if there are topics that are preferred by journals and implications are provided.
키워드 : Latent Dirichlet Allocation, psychometric journals, abstract, topic modeling