Enhancing card sorting dendrograms through the holistic analysis of distance methods and linkage criteria
EntityUAM. Departamento de Ingeniería Informática
PublisherUser Experience Professionals Association (UXPA)
CitationJournal of Usability Studies 16.2 (2021): 73–90
Funded byThis work was partially supported by the Spanish Government (grant number RTI2018-095255-B-I00) and the Madrid Research Council (grant number P2018/TCS-4314)
ProjectGobierno de España. RTI2018-095255-B-I00; Comunidad de Madrid. P2018/TCS-4314/Forte-CM
SubjectsInformation architecture; Card sorting; Quantitative analysis; Agglomerative clustering; Distance method; Linkage criterion; Dendrogram; Informática
Rights© The authors/ACM
Esta obra está bajo una Licencia Creative Commons Atribución 4.0 Internacional.
Card sorting has been widely used in information architecture to analyze and improve web content and navigation. This is an intuitive and cost-effective technique also useful in user research and evaluation. However, while the implementation of sorting tasks comprises a constructive and easy-toaccomplish process, the quantitative analysis of resulting card-sorting data can be a challenge for non-skilled evaluators. Several tools exist to support sorting tasks and data analysis, but still some users utilize custom spreadsheets or statistical packages in order to enhance analysis and obtain more expressive and comparable results. One of the most utilized diagrams for analyzing card-sorting results is the dendrogram, also known as a tree diagram, which is commonly based on an agglomerative clustering representation depicting groupings of related cards. However, several issues have to be considered by evaluators in order to produce meaningful dendrograms for decisionmaking. In fact, the distance method and the linkage criterion greatly influence the final dendrogram obtained. In this paper, an analysis on distance methods and linkage criteria for obtaining suitable dendrograms is proposed. The main aim is to guide evaluators and usability engineers to produce appropriate dendrograms based on available cardsorting data. In this sense, the provided clues can be widely applied to any card-sorting domain and sample size, improving card-sorting analysis by comparing different solutions through goodness indicators. Analyses applied to a publicly available dataset indicate that the results are highly dependent of the data type, so the right selection of both distance method and linkage criterion is essential for obtaining a suitable dendrogram and correctly interpreting the results
Google Scholar:Macías Iglesias, J. A.
This item appears in the following Collection(s)
Showing items related by title, author, creator and subject.