Academic journal article The Qualitative Report

Using Clustering as a Tool: Mixed Methods in Qualitative Data Analysis

Academic journal article The Qualitative Report

Using Clustering as a Tool: Mixed Methods in Qualitative Data Analysis

Article excerpt

Tools and Techniques for the Analysis of Qualitative Data

Qualitative research produces a great volume of data that can be challenging, and at times overwhelming, to analyze. While literature on qualitative data collection methods and theoretical approaches for their analysis is increasingly rich, a gap still remains on works that offer detailed accounts on data analysis techniques (Attride-Stirling, 2001; Fielding, 2005). This gap is maintained in part by the recognition that different theoretical approaches demand different analytic approaches to qualitative research, and also by the still existent guarded acceptance (and in some cases rejection) of tools, particularly computer-based ones or those traditionally used in quantitative approaches, that although providing an appearance of clarity can misconstrue qualitative research findings (Gilbert, Jackson, & di Gregorio, 2014; James, 2013).

However, quantitative analytical tools can be helpful in qualitative analysis, as long as there is clarity on what can and cannot be done with them. Literature that details the use of specific mixed techniques in qualitative analysis is important to avoid misusing them, or misrepresenting their findings. Additionally, clear explanations provide transparency to qualitative data analysis processes, which are sometimes regarded as obscure and lacking trustworthiness by scholars and practitioners not fully familiar with qualitative methodologies (Barusch, Gringeri, & George, 2011; Miller & Crabtree, 1994).

Cluster Analysis in Qualitative Research

In this article I discuss cluster analysis as an exploratory tool to support the identification of associations within qualitative data. While not appropriate for all qualitative projects, cluster analysis can be helpful in identifying patterns where numerous cases are studied. As the number of elements and facets considered for each case increase, so does the complexity of finding associations between them. When this happens, even the analytic tools provided for this purpose by commonly used software tools for qualitative analysis (e.g., coding co-occurrences, matrixes or Boolean searches) can fall short, with the related risk of reaching conclusions that overly reflect the researcher's preconceptions. Some qualitative analysis software (NVivo, QDA Miner, ANTRHOPAC) have included cluster analysis tools, mainly as text-mining tools (Silver & Lewins, 2014). However, its use remains limited or underreported, perhaps due to their limitations in the number of items allowed (ANTHROPAC), the techniques and measures available (NVivo), or the array of data used to determine the formation of clusters.

The use of cluster analysis with qualitative data has already been discussed and documented (Guest & McLellan, 2003; MacQueen et al., 2001), but remains underused. It is possible that this is partially due to an obscurity regarding how to apply this tool to qualitative data, and a need for detailed but accessible explanations of the basic considerations and steps required from qualitative researchers interested in clustering as a tool. Through the detailed presentation of a case study using content cluster analysis in a qualitative project, I explain in this article the main elements of this technique as it applies to qualitative data, including the steps and considerations required to perform clustering and interpret its results. Because cluster analysis is an exploratory tool meant to support the analysis of qualitative data, I also discuss when and how to return to the full qualitative data.

The Case Study: The Latino Grievances Project

The case study discussed through this article comes from a research project on grievance management among Latino immigrants to southwestern Pennsylvania, approved by the University of Pittsburgh IRB. Between 2007 and 2009 I collected qualitative data from twenty-one in-depth interviews with Latino immigrants. I was exploring how these immigrants dealt with perceived grievances, including how and when they felt aggrieved, what options they recognized as available, and how they decided what to do. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.