On Sep 22, 2022, at 4:01 PM, Jodi Schneider <[log in to unmask]> wrote:
> There is some work comparing topic modeling to qualitative analysis. Here
> are two papers I'd suggest:
>
> Baumer, Eric P. S., David Mimno, Shion Guha, Emily Quan, and Geri K. Gay.
> “Comparing Grounded Theory and Topic Modeling: Extreme Divergence or
> Unlikely Convergence?” Journal of the Association for Information Science
> and Technology 68, no. 6 (2017): 1397–1410.
> https://doi.org/10.1002/asi.23786
>
> Roberts, Margaret E., Brandon M. Stewart, Dustin Tingley, Christopher
> Lucas, Jetson Leder-Luis, Shana Kushner Gadarian, Bethany Albertson, and
> David G. Rand. “Structural Topic Models for Open-Ended Survey Responses.”
> American Journal of Political Science 58, no. 4 (2014): 1064–82.
> https://doi.org/10.1111/ajps.12103
Jodi, thank you for bringing these to our attention.
I'd like to double suggest the article by Roberts because it outlines a technique I use often. More specifically, I often use MALLET to topic model a corpus, and this results in a sort of document-term matrix. This matrix can be augmented with metadata values such as date, author, genre, gender, nationality, etc. The matrix can then be pivoted and visualized thus illustrating how different topics correspond to different metadata values. Using this technique one can plot the ebb & flow of ideas ("topics") over time, across genders, between nationalities, etc. For example, I might topic model works by Longfellow, Emerson, Thoreau, Austen, and Melville. Using the technique outlined above and alluded to in the Roberts article, one can then illustrate how these authors compare & contrast.
--
Eric Lease Morgan
Navari Family Center for Digital Scholarship
University of Notre Dame
