The process of QDA in the era of Generative-AI
Christina Silver1 , Trena Paulus2
1 CAQDAS Networking Project
2 East Tennessee State University
The use of AI tools in qualitative analysis has a history of more than 25 years, with the first real assistance – beyond managing data and text-based searching and auto-coding – coming at the end of the 1990s with unsupervised machine learning features such as topic extraction using clustering (e.g. WordStat, 1999) and concept mapping (e.g. Leximancer, 2000), and supervised machine learning features such as code suggestions (e.g. Qualrus, 2002) and collaborative and adjudicated human and machine classifiers (e.g. DiscoverText, 2009). Since then many other CAQDAS-packages have implemented various AI-informed features based on different applications of natural language processing and machine learning technologies, such as sentiment analysis, theme extraction, and topic modeling (e.g. ATLAS.ti, MAXQDA, QDA Miner and NVivo).
However, the rise of Generative-AI and in particular the release of OpenAI’s ChatGPT in November 2022 made these technologies accessible and user-friendly in ways and with impact not previously seen. Suddenly awareness of Large Language Models (LLMs) and their capabilities became the topic of discussion on mainstream media channels, among the general public and within the qualitative community of practice. The integration of these new technological capabilities into the qualitative analysis space began rapidly, with several established CAQDAS-packages beginning to incorporate functionality based on GPTs within months – for example ATLAS.ti (from March 2023), MAXQDA (from April 2023), Qual Coder (from December 2023), and NVivo (from September 2024). This is in addition to automated transcription which has significantly improved with the enhanced capabilities of models, and several CAQDAS-packages have added this functionality recently, including MAXQDA, Transana and Quirkos.
Online Panel Discussion | See the full program
Tracing Filaments: The multidimensional, transdisciplinary potentials of postformal research
Tricia Kress1, Eric Karahalis2, Kelly Bare3
1 Department of Education Molloy University, USA
2 Department of Philosophy Suffolk County Community College, USA
3 Master of Public Leadership Program, University of San Francisco, USA
While the field of qualitative research has moved away from neopositivism toward constructivist, postmodern, critical and participatory directions, the discourse on what qualifies as “research” validated by universities and funding agencies typically reinscribes neopositivism as the foundation for knowledge-seeking (Tobin & Kincheloe, 2009). As such, the utility of qualitative research for asking new questions and generating new insights into social phenomena may be undermined as researchers themselves find comfort and validation in the well-worn paths of positivist discourse typically reinforced by their disciplines and workplaces. As well-worn paths may lead to predetermined destinations, research questions and findings may become circular and repetitive, and across generations, scholars attempt to find answers to repeating questions that seem resistant to resolution.
This panel presentation introduces postformal research as a potential escape route from the circular logic of neopositivist discourse. Kincheloe (2008) described positivist research with the acronym “FIDUROD.” Drawn from Enlightenment epistemology, positivist research is rigid and unchanging (Formal), assumes the world is static (Intractable), is not grounded in context (Decontextualized), imagines itself applicable to all phenomena and contexts (Universalistic), focuses only on factors that can be measured (Reductionistic), and espouses there are singular underlying discoverable truths (One-Dimensional). Alternatively, postformal thinking seeks out complexity and difference, giving rise to further questions about humanity and the social world to home in on and make sense of entangled issues that may reproduce social inequality.
Grounded in critical, postmodern and feminist theories and in conversation with Indigenous knowledges, postformal research disrupts linear, reductionist thinking that confines researchers to the well-worn paths of neopositivism (Kress & Lake, 2021).
Online Panel Discussion | See the full program