Multidimensional scaling modelling approach to latent profile analysis in psychological research

Abstract

Because profile analysis is widely used in studying types of people, we propose an alternative technique for such analysis in this article. As an application of the multidimensional scaling (MDS) model, MDS profile analysis is proposed as an approach for studying both group and/or individual profile patterns. This approach requires one to think of MDS solutions as profiles. The MDS profile analysis approach re‐parameterizes the linear latent variable model in such a way that the latent variables can be interpreted in terms of profile patterns rather than factors. It is an exploratory profile analysis method to identify major patterns among psychological variables and can serve as the basis for further study of correlates and/or predictors of profiles and other background and external variables. In the article, we have outlined the procedure of MDS profile analysis and discussed the issues that are related to parameter estimation and interpretation of the results. Such a discussion is presented in the context of studying psychosocial adjustment of adolescents in the US. The results of MDS profile analysis indicated that two psychosocial profiles could be derived from the data. Profile 1 represents a prototypical adolescent with high scores on Body Image and low scores on the Vocational‐Educational (Voc‐Ed) Goal scale. Profile 2 represents a prototypical adolescent with high scores on the Peer Relationships scale and low scores on the Family Relationships scale. These profiles suggested several types of adolescents with different psychosocial adjustment orientations, which may be linked to different subgroups of adolescents in the population. In addition, these profiles are also associated with the outcomes of mental health, which implies profiles' potential utility when studying relationships among different variables. Finally, this article discusses some advantages and disadvantages of this approach in psychological research.