1 A SYNOPSIS OF A 2018 AHRD FOCUS SESSION HELD IN RICHMOND, VIRGINIA ON FEBRUARY 16, 2018 Topic Modeling: A Text Mining Technique for HRD Researchers David L. Passmore, Penn State, dlp@psu.edu Rose M. Baker, University of North Texas, rose.baker@unt.edu The AHRD Session: HRD scholars are exposed to massive amounts of text data available to harvest that could guide theory, policy, and practice in HRD, yet few HRD scholars are prepared to exploit these text data in research. Topic modeling processes natural language to uncover themes by analyzing patterns of words in a collection of documents. Topic modeling theory has evolved from such diverse fields as computer science, artificial intelligence, statistics, and linguistics. The aim of the AHRD FOCUS session was to provide for HRD researchers a non- mathematical introduction to topic modeling, a machine learning and natural language processing tool, using a common and popular analytic approach, Latent Dirichlet Allocation (LDA) The session was structured around the following themes: • The crushing force of information; • Mining of information in text; • Use of machine learning and natural language processing; • Unsupervised machine learning about topics in text — LDA; and • Promising applications of LDA in HRD research. A slide deck that supported a 2017 AHRD paper presentation about topic modeling is available from Chae, Al-Khadhuri, Passmore, Baker, & Turner (2017). The slide deck used to support the 2018 AHRD FOCUS session is downloadable from Passmore and Baker (2018). Need, Perspective, and Approach for Topic Modeling: Need The almost universal digitalization of swelling amounts of information is making the problem of staying up-to-date with cumulative knowledge growth acute in many areas of scholarship. Consider that global data amounted to 4 zettabytes in 2013 (Container Eyes, 2013, para 5) and, if these data double annually as anticipated, this figure will surge to 44 zettabytes by 2020 (Turner, Reinsel, Gantz, & Minton, 2014, p. 2). One million minutes of video every second could cross global networks by 2020 (“The Zettabyte Era,” 2016, p. 3). The “Internet of Things” (IoT) — that is, the networking and data sharing of devices as varied as cars, toys, airplanes, dishwashers, turbines, and dog collars—included approximately 10 billion units in 2015. The IoT is forecasted to involve 34 billion devices in 2020 (Greenough, 2016, para 5). Maintaining currency with the state of knowledge and practices will strain resources and imagination as more and different data flow faster and more widely than ever before. Text data are among the most common data. Gentzkow, Kelly, and Taddy (2017) observed that “An ever-increasing share of human interaction, communication, and