Utility of Profile Analysis via Multidimensional Scaling in R for the Study of Person Response Profiles in Cross-Sectional and Longitudinal Data

• Published in: Tutorials in quantitative methods for psychology Vol. 20; no. 3; pp. 230 - 247
• Main Authors: Kim, Se-Kang, Kim, Donghoh
• Format: Journal Article
• Language: English
• Published: Université d'Ottawa 01.09.2024
• Subjects: bootstrap; cross-sectional and longitudinal data; multidimensional scaling; profile analysis
• ISSN: 1913-4126
• DOI: 10.20982/tqmp.20.3.p230

Profile analysis using multidimensional scaling (PAMS) attempts to identify core profiles, which are the central response patterns from a group of person profiles. From a micro perspective, we can use the core profile information to assess the strengths and weaknesses of individual traits measured by subscales, whereas from a macro perspective, we can use it to understand the average strengths and weaknesses of all responses in the sample. Initially designed for a proprietary purpose, PAMS's use by the general public remains restricted. Here, we introduce a public version of PAMS in R, enhanced with inferential capabilities that enable researchers to assess the statistical significance of core profile coordinates during interpretation. We analyzed the same cross-sectional data using cluster analysis and latent profile analysis and compared the results to amplify the PAMS utility over the other methods because it holds more information. We also demonstrated the utility of PAMS through the analysis of longitudinal data. We extracted three core profiles from both cross-sectional and longitudinal data sets and then described how to interpret core profile coordinates for each data set, in addition to limitations and suggestions for potential PAMS in R users.