Text Mining Systematic Reviews: New Directions for Qualitative Research Synthesis

Text Mining Systematic Reviews: New Directions for Qualitative Research Synthesis

This live webinar has already taken place. Scroll down to view the video recording.

In this TI Methods Speaker Series webinar, Dr. Caspar J. Van Lissa, Associate Professor at Tilburg University (Netherlands), discussed pioneering work on Text Mining Systematic Reviews (TMSR) by presenting two studies that used different TMSR pipelines to extract a latent nomological network from the literature in a particular subfield, and reflected on general considerations regarding the use of text mining in research synthesis, based on his experience as Associate Editor of Research Synthesis Methods.

TITLE: Text Mining Systematic Reviews: New Directions for Qualitative Research Synthesis

WHEN: Wednesday, July 31st, 2024 at 12:00 noon PDT [convert to your local time]

WHERE: free online webinar.

SPEAKER: Dr. Caspar J. Van Lissa, Associate Professor of Social Data Science, Department of Methodology & Statistics, Tilburg University, and Chair of the Open Science Community Tilburg, Netherlands.


About the topic: As the number of publications in most fields continues to grow exponentially, it becomes increasingly unfeasible for scholars to remain informed about the entire literature. Moreover, narrative reviews are subjective and susceptible to a variety of biases, including confirmation bias. Innovations in the machine learning domain of text mining can be used to synthesize the burgeoning literature in a relatively objective and scalable manner. This presentation discusses pioneering work on Text Mining Systematic Reviews (TMSR, Van Lissa 2022). TMSR is an umbrella term for quantitatively aided qualitative research synthesis methods that use machine learning to extract knowledge from published scientific literature. I present two studies that used different TMSR pipelines to extract a latent nomological network from the literature in a particular subfield. These networks can serve as a useful starting point for theory development, help researchers find their bearings in the literature, and identify knowledge gaps. Furthermore, I present one application of TMSR that set out to identify causal claims in the literature. In fields where causal assumptions are rarely made explicit (especially social science), extracting the latent causal network from the published literature may advance a more explicit discussion about causality and formal theory. To conclude, I reflect on general considerations regarding the use of text mining in research synthesis, based on my experience as associate editor of Research Synthesis Methods.


About the speaker: Dr. Caspar J. Van Lissa is Associate Professor of Social Data Science at Tilburg University, the Netherlands. He serves as chair of the local Open Science Community, and is Associate Editor for Research Synthesis Methods and statistical consulting editor for Child Development. His primary research line uses machine learning for rigorous exploration to complement blind spots in theory. His secondary research line centers on machine learning-based evidence synthesis: summarizing existing knowledge, e.g. through systematic reviews and meta-analysis.


About the TI Methods Speaker Series: The TI Methods Speaker Series are offered free of charge and everyone is welcome. The event is usually held at noon on the last Wednesday of each month via Zoom videoconference. The presentations are recorded and the video recordings are posted online. Click here to view the list of talks offered in 2024.

No Comments

Post A Comment