Edward E. Brent
Ph.D., University of Minnesota
Qualitative and quantitative methods, expert systems, social interaction, social psychology, sociology of computing
My research program has two recurring themes: using computing technologies to practice research and reason sociologically, and examining the nature and mechanisms of social interaction in work settings with relationships of power. I am fascinated by the process of social interaction and ways in which computers can help us study that interaction. How do people interact in various social circumstances? How does the process unfold over time? How are power and privilege exerted in face-to-face encounters? How can computers be used to track and categorize interactions for large data sets, and how can statistical and computational programs be used to help analyze them?
Social interaction was the focus of the first phase of my research with my dissertation and a post-doctoral fellowship leading to studies of social interactions in relationships of authority, first between police and civilians (Sykes & Brent, 1983), and later between doctors and patients. This early work focused on the process of social interaction as measured by systematic observation codes and modeled with stochastic discrete-state and discrete-time Markov models. This work helped us understand the asymmetries of social interaction as it unfolds over time among superiors and subordinates.
But this work was limited by the costs of coding large amounts of interaction data and the complexity of social interaction left many questions unanswered by mathematical and statistical models (a “Big Data” problem in sociology). This led to examining new applications of computing to social science research, including several articles, a second book, (Brent and Anderson, 1990), an edited book (Blank, McCartney, and Brent, 1989), and a special issue of Social Science Computer Review (Brent, 2004).
Subsequent research funded by NIMH addressed this issue by developing and testing computer programs using artificial intelligence strategies to automatically code transcripts of doctor-patient interactions. We had substantial success, coding some dimensions with over 80% accuracy and identified promising strategies for further increasing the validity of the automatic coding process. This work generated several paper presentations and an article in the Journal of Mathematical Sociology. Subsequent research employs a range of natural language understanding strategies for the more general task of coding data in qualitative analysis in Qualrus™ and Veyor® where accuracy measured by percent agreement routinely is in the mid- to high-90s.
Later research build on the early NIMH-funded research using NSF funding in a project with Howard Becker to use natural language understanding and artificial intelligence strategies to partially automate coding qualitative data (Qualrus™) and totally automate the coding of text information from many sources including the web (Veyor®). By combining automated coding using NLU strategies with expert systems to assess how well learning objectives are met, SAGrader™ and a Reviseabl® provide students with immediate feedback on writing. This creates a learning environment in which students employ iterative revision based on feedback to learn by writing. These programs are used at Mizzou in several departments as well as other colleges throughout the US and several countries.
A sabbatical and second post-doctoral fellowship in medical informatics in 1983 gave me the opportunity to focus on artificial intelligence. Expert systems became tremendously popular in late 1980s and early 1990s. With colleagues, I developed several expert systems in the social sciences, including a number to help social scientists design research (Methodologist’s Toolchest™ (Brent and Thompson, 1988-2000). At MU we created the multidisciplinary Laboratory for Applied Expert Systems Research (LAESR), staffed by faculty from Sociology, Medicine, Veterinary Medicine, Anthropology, and Computer Science. With two anthropologists in the Lab, we published a book on expert systems (Benfer, Brent, and Furbee, 1991). In other work we applied expert systems to modeling and simulating social interaction in a program based on Goffman’s dramaturgical theory (ERVING™). A much more ambitious SocActor™ project used expert systems to simulate many aspects of adolescent behavior in a model that will model school performance, entry into the workforce, sexual behavior, alcohol and drug use, and family interactions. Here students interact with other social actors simulated by intelligent agents whose behavior illustrates sociological concepts, theories, and empirical findings, all in a virtual reality (VR) environment.
While early federal funding for social science research helped launch this work, it soon became apparent that federal funding alone would be insufficient to collect and analyze the “Big Data” required. As a result I became a recidivist entrepreneur by founding or co-founding six non-profit and for-profit startups including Idea Works, Inc. the initial software company using expert systems and artificial intelligence to create many of these programs, a startup that provides crowd-funding for social entrepreneurial projects (FundProvo) and a nonprofit NGO (Yang-Ward Foundation) to provide assistance to single women in rural Nepal. This work led to the Career Achievement Award from the Section on Sociological Computing in the American Sociological Association, and the Faculty Entrepreneur Award from the University of Missouri in 2008.
My current research draws on many aspects of my earlier work to examine writing as a social process. While there are many books and articles on HOW to write, there is surprisingly little research into how people actually DO write. Even then, most work ignores the social context of writing. My approach examines the broad social constraints on writing to understand social issues such as cheating, multitasking, and procrastinating. Specifically I am studying iterative writing with feedback, which can be conceptualized as a curiously constrained form of social interaction (iterative revision with feedback) within relationships of power (instructor and student), task requirements (learning objectives), taking place within a constrained time frame with limited motivation (procrastination) in the face of competing tasks (multitasking), and subject to behavioral norms (academic integrity). This work examines a database of student submissions and revisions to essay assignments from over 10,000 students in introductory sociology courses. This work has received funding from the Research Council and the Research Board and further external funding is being sought.
- Expert Systems
- Introduction to Sociology
- Research Methodology
- Seminar In Multivariate Analysis Techniques
*Brent, Edward. Artificial Intelligence, Expert Systems & the Internet. 2007. In Handbook of Online Research Methods, Pp. 452-468 in 2nd Ed., pp. 20 2014. Sage Publications, London, UK, Eds. Fielding, Nigel; Blank, Grant; Lee, Ray. Also 2nd Edition, 2017.
*Brent, Edward and Lewis, J. Scott. 2013. Learn Sociology. Jones & Bartlett Learning. Burlington, MA.
*Atkisson, Curtis and Edward Brent. (2011). “A Rational Framework for Student Interactions with Collaborative Educational Systems.” In Technology-Enhanced Systems and Adaptation Methods for Collaborative Learning Support. Springer-Verlag.
*Brent, Edward and Curtis Atkisson. 2011. “Accounting for Cheating: An Evolving Theory and Emergent Themes.” Research in Higher Education. 52(6).
Atkisson, Curtis, Colin Monaghan, and Edward Brent. 2010. “Using Computational Techniques to Fill the Gap between Qualitative Data Analysis and Text Analytics.” Kwalon, Journal for Qualitative Research.
Brent, Edward, Theodore Carnahan, Pawel Slusarz. SAGrader™: A Computerized Essay Grading Program. Idea Works, Inc. Columbia, Missouri. 2005.
Brent, Edward (Editor) Special Issue on Sociology and Computing. Social Science Computer Review. 2004.
*Brent, Edward and Pawel Slusarz, "’Feeling the Beat’: Intelligent Coding Advice from Meta-Knowledge in Qualitative Research” Social Science Computer Review. 2002.
Brent, Edward and Alan Thompson. 2000. Qualrus: An Intelligent Program for Qualitative Analysis. Idea Works, Inc.
Brent, Edward. Connections: Interactive Sociology. 2000, Wadsworth. Version 2.
*Brent, Edward and Alan Thompson. “Modeling Social Life with Intelligent Autonomous Agents,” Social Science Computer Review.” 1999. 17(3):313-322.
Brent, Edward. 1998. "Simulating Social Interaction in a Virtual Reality Setting: Problems and Prospects." Pp. 148-169 in Henry Millsom S. (Ed.), Information Technology in the Social Sciences." Oxford: Blackwell.
*Brent, Edward. The Role of Expert Systems in Sociological Research and Thinking. Special Issue of Sociological Methods & Research, Cathleen Karley (Editor). 25:1 (August), 1996. Pp. 31-59.
*Bainbridge, William Sims, Brent, Edward E., Carley, Kathleen, Heise, David R., Macy, Michael W., Markovsky, Barry, and Skvoretz, John. Artificial Social Intelligence. Annual Review of Sociology. 1994. (20)407-436.
Brent, Edward and Alan Thompson. 1992. Methodologist's Toolchest, Version 2. Sage Publications, Inc.
*Benfer, Robert A., Edward Brent, and Louanna Furbee, Expert Systems, Sage, 1991.
*Brent, Edward and Ronald Anderson, Computer Applications in the Social Sciences. Random House, 1990.
*Blank, Grant, James McCartney, and Edward Brent (Eds.), New Technology in Sociology: Practical Applications in Research and Work. Transaction Publishers, 1989.
*Sykes, Richard and Edward Brent, Policing: A Social Behaviorist Perspective, Rutgers University Press, 1983.