Comparing Analytic Techniques to Classify Dyadic Relationships: An Example Using Siblings

• Published in: Journal of marriage and family 2006-12, Vol.68 (5), p.1370-1382
• Main Authors: Whiteman, Shawn D., Loken, Eric
• Format: Article
• Language: English
• Subjects: A priori knowledge; Adolescents; Algorithms; Classification; Cluster Analysis; Comparative studies; Correlations; Dyad; Dyadic relations; dyadic/couple data; Dyadics; Dyads; Evaluation Methods; Families & family life; Family (Sociological Unit); Family relations; Family studies; History, theory and methodology; Interpersonal relations; Interpersonal Relationship; Intimacy; Longitudinal Studies; Methodology; Methodology (Data Analysis); Modeling; Multivariate Analysis; Of General Interest; Parametric models; Parent Child Relationship; Personal relationships; Research methodology; Researchers; Sample size; Sexuality. Marriage. Family relations; sibling relations; Sibling Relationship; Siblings; Social research; Sociology; Sociology of the family. Age groups; Statistical Analysis; Taxonomy; Teenagers
• ISSN: 0022-2445; 1741-3737
• DOI: 10.1111/j.1741-3737.2006.00333.x

Several methods exist for typologizing dyadic relationships. The present study reviews the utility of methods for typologizing relationships, exploring the strengths and weaknesses of 3 different approaches, and applying them to the study of adolescent sibling relationships. The 3 methods examined are a priori methods, cluster analysis, and mixture models. Each approach was used to analyze the same data taken from a cross section of a longitudinal study of 384 families with adolescent children. Clustering appeared to favor a 2-type solution whereas the mixture model approach supported 4 types. The 3 strategies are discussed in terms of their statistical rigor and utility in the study of close relationships.