Working in Data Mines: Conducting Multiple Analyses on Qualitative Data Sets

摘要

This paper has two purposes: to introduce the idea of mining qualitative data to new engineering education researchers, and to provide useful examples for reference. While it is common to see large quantitative data sets being mined for new findings, large qualitative data sets (whether interviews or observations) are often only used for one research agenda. Qualitative data sets-like their quantitative counterparts-are rich enough in information to support secondary analyses, and researchers should consider them as viable sources to support multiple investigative agendas. There are multiple reasons for re-using qualitative data sets. Large qualitative data sets require significant time and resources for data collection and transcription. Particularly for faculty that face limited funding, and graduate students that face limited timelines for their theses and dissertations, pre-existing large qualitative data sets are valuable resources that can reduce the time and resources necessary for traditional qualitative studies. However, secondary analysis of qualitative data (i.e., analysis of an existing qualitative data set to answer new research questions) comes with challenges. For example, the purpose of the new analysis can be quite different from that of the original study, which can raise questions of alignment of the data with the research questions and the data analysis methods. In this paper, we present an example of how researchers used one set of qualitative data to support multiple research agendas. First, we describe the large longitudinal qualitative data set, including the data collection methods and the original research questions guiding the study. Subsequently, we discuss how three graduate students separately used this qualitative data set in completion of their dissertations. Specifically, two graduate students used the data for secondary analyses to ask new questions of the data and another graduate student used the data to pilot an observation protocol as part of her dissertation study. We discuss how those graduate students determined that the data set was viable for their individual research agendas, and outline some guidelines that others could use in doing so. Lastly, we conclude with the implications particular in designing a qualitative study around secondary data analysis. We believe that the information in this paper is valuable to graduate students and new faculty considering new research avenues with limited resources, in an effort to maximize the usefulness of previously dedicated resources.