Comparing Grouping Results Between Cluster Analysis and Q-Methodology

The primary purpose of this Student Research Poster Paper is to discuss two grouping methodologies: cluste analysis and the Q-Methodology. Each of these statistical methodologies quantitatively group similar individuals but do so in two separate ways. In cluster analysis, individuals are grouped by optimizing proximity measures. For example, the single link clustering algorithm groups individuals together that have the smallest Euclidean distance from each other. In Q-Methodology, individuals are grouped by evaluating person-to-person correlations. For example, if two individuals have a correlation of 0.78, they are likely to be grouped together whereas individuals with a correlation of 0.09 are likely to be in different groups. In this paper, we outline multiple clustering approaches, the grouping mechanism in Q-Methodology, and discuss the differences between these two approaches when grouping participants. During this discussion, we will use an example from our own engineering education research to compare grouping results from the same data set. This paper contributes to the research field by describing and utilizing a relatively unknown methodology in engineering education. It will also add to our knowledge of cluster analysis techniques and compare those algorithms to another robust grouping method.