Exploring Types of Parent Attachment via the Clustering Modules of a New Free Statistical Software, ROP-R

Abstract

The aim of the paper is threefold: (1) to demonstrate the rich repertoire of clustering capabilities of a ROPstat and R-based new and free software, called ROP-R, by illustrating several analyses with real psychological data; (2) to show how well ROP-R works in tandem with ROPstat software in complex classification analyses; and (3) to explore some nontrivial types of parent attachment using the clustering modules of ROP-R. Four modules of ROP-R are available for performing cluster analyses (CAs), with several methods (e.g., divisive hierarchical CA, k-medoids CA, k-medians CA, model-based CA) not found in other user-friendly menu-driven software. In the paper, mother and father attachment data are used from a study with adolescents (Mirnics et al., 2021) to illustrate how the ROP-R software can be used to perform various CAs and evaluate the results using attractive graphs and useful tables. Comparing different clustering methods, it was found that both standard AHCA and k-means CA could discover a 7-type structure, which was also verified by the nonstandard k-medians CA. However, the nonstandard k-medoids CA and MBCA methods were not very effective in identifying a structure with an acceptable overall homogeneity. Nevertheless, they were able to identify some types through extremely homogeneous clusters.