Does Big Data Really Make the Scientific Method Obsolete?
In conjunction with the 2019 International Conference of Information Systems (ICIS), the AIS Special Interest Group on Philosophy in Information Systems (AIS-SIGPHIL) will hold its 8th Concurrent SIGPHIL@ICIS Research Workshop during two evenings of the ICIS conference in Munich, Germany at the Internationales Congress Center München (ICM). Following last year’s SIGPHIL@ICIS, this year’s event continues the call for the edited series on “Advancing IS theories” by Nik Hassan and Leslie Willcocks. At the same time, the workshop provides an excellent opportunity to spend quality time with thought leaders of the IS community in an informal and friendly environment.
Workshop Presenters, Organizing Committee and Contributors (in alphabetical order)
Ahmed Abbasi, University of Virginia
Alan Dennis, Indiana University
Vasant Dhar, New York University
Varun Grover, University of Arkansas
Nik Rushdi Hassan, University of Minnesota Duluth
Alan Hevner, University of South Florida
Shirley Gregor, Australian National University
Rudy Hirschheim, Louisiana State University
Kalle Lyytinen, Case Western Reserve University
Foster Provost, New York University
Sumit Sarkar, University of Texas Dallas
Ramesh Sharda, Oklahoma State University
Leslie Willcocks, London School of Economics
Workshop Theme
The title of this workshop is paraphrased from the title of an editorial by Wired Magazine’s chief editor Chris Anderson (2008) who argued that with big data, we no longer have to settle for imperfect models, and since the scientific method relies on models from which we test hypotheses, big data has essential made the scientific method obsolete. Extending this argument, because theory is the goal of the scientific method, theory itself becomes unnecessary. Why do we need theory when big data can already help us predict? Not surprisingly this claim has attracted much attention from both industry and academia (Mayer-Schönberger and Cukier, 2013; Kitchin, 2014). Like many highly cited pieces, the Wired editorial has taken on a life of its own, as it is interpreted and reinterpreted by many to support their own stance on the topic of theory. Another article from Wired Magazine (Steadman, 2013) featured how big data predicted Osama bin Laden’s location from publicly available data without any need for models or theories. In other words, “it just needs to work: prediction trumps explanation” (Siegel, 2016, p. 90). Some of these researchers take big data research as an extreme form of empiricism to reignite long-standing debates surrounding the legitimacy of the social sciences and eagerly use big data to claim the status of the natural sciences for their own works. Big data has supposedly shifted the paradigm of research itself which previously could only take place as a result of trade-offs among generality, control and realism (Chang, Kauffman and Kwon, 2014). “Computational social science” (Lazer et al., 2009) using big data is free of those trade-offs. Others are more cautious. Mayer-Schönberger and Cukier (2013) consider preposterous the claim that generalizable rules are irrelevant. For example, they argue that the process of collecting big data itself is based on some kind of theory.
For a field that is still struggling with theory and the role of theory in its research, this debate about the relevance and irrelevance of theory vis-a-vis analytics places our researchers in a difficult position on at least two levels. First, theory for description, explanation and prediction for any area within IS – analytics included – is itself being put to question by big data. Second, theory building within and for analytics as a subfield of IS too is unclear. At the first level, can practical questions that are targeted by analytics for specific instances (Will the customer buy? Can the MRI show anomalies?) be generalized? At the second level, why do we, IS researchers, need to worry about theory in analytics when all of its theoretical foundations were already built by scholars of statistical learning theory, computer science and operations research? Only a few scholars have addressed theory for analytics directly (Shmueli and Koppius, 2011; Agarwal and Dhar, 2014; Abbasi, Sarker and Chiang, 2016; Maass et al., 2018) while the rest of the IS community remain silent. Very few studies discuss explanatory analytics theory in general or foundational analytics theories that cut across all analytics processes including data collection, data preparation and cleaning, exploratory data analysis, model building, evaluation and deployment. Anecdotal evidence suggest that the IS community may have gravitated towards the notion that theory is indeed irrelevant for business analytics. For instance, an editor for a prestigious IS conference noted: “we have witnessed explosive growth of the business analytics field in this decade, both in research and in practice. So, why is theory building mandatory for the growth and legitimacy of the field?” Such as a state of affairs is problematic, not the least because the majority of programs in business analytics (BA) in the Schools of Business around the world are led by IS scholars and researchers.
The discussion on theory in the field has left us unclear about whether or not theory does or does not play a major role and whether theory for business analytics matter at all. It may be clear to most within the IS field that our researchers are not expected to invent the next Hadoop or MapReduce, or even to write the next classification or clustering algorithm. If those technologies are not where our efforts should be expended, what exactly is the role of the IS researcher, and by extension, the practice of BA that is most relevant to IS? Is the IS researcher left with the trite and uninspiring task of researching the adoption or acceptance of big data analytics? Or can the IS researcher, as Dhar (2013) proposes, provide interesting answers to questions that we do not yet know? Or even better, as Pentland (2014) claims, we can solve macro-level problems using the micro-level big data that are being analyzed and “build a society that is better at avoiding market crashes, ethnic and religious violence, political stalemates, widespread corruption, and dangerous concentrations of power” (p. 17), all of which cannot do without solid theoretical foundations.
Edited Book Series: Advancing IS Theories
This struggle for theory is the theme for this year’s SIGPHIL@ICIS workshop, focusing on theory in business analytics and supporting the goals of a planned series of volumes on information systems (IS) titled: “Advancing Information Systems Theories.” The goal of this series of volume is to advance IS research beyond borrowed legitimization and derivative research towards fresh and original research that naturally comes from its own theories – information system theories. The first volume on the process of IS theorizing is in the final stages of review and near publication. The second volume concerns efforts that approach theories – what Weick (1995) calls “interim struggles.” This volume comes out of the realization that the process of theorizing can be long and arduous and like all great things, will not be built in a day, much less in an edited volume. So, although they may not be called theories with a capital “T,” they nevertheless approximate theory and should not be dismissed. They may be called “principles,” “propositions,” “models,” “paradigms,” “concepts,” “frameworks” or what have you. They are the products of theorizing and are precursors to strong theory, and as long as they are fresh and original, they go a long way in advancing IS theories. A demonstrative list of chapters for Vol. II is provided below:
Introduction: The products of IS theorizing (Hassan, Mathiassen & Lowry) 1
The prospects of theory for business analytics 20
A review of information theory in information systems (McKinney) 40
Design principles in design science (Gregor and Hevner) 60
IS Concepts: Declaring IS to the world 80
Mapping an IS research framework 100
Models and contexts of discovery in IS 120
IS constructs and variables 140
A collection of IS propositions 160
Program Schedule
Sunday, Dec 15, 2019 (Location: ICM Room 21A)
7:30pm-7:40pm: Introductions by Nik Hassan: The goals for the workshop and the notion of products of theorizing
7:40pm – 8:10pm First Plenary Keynote by Varun Grover and Kalle Lyytinen on “Big data and the changing role of theory”
8:10pm-8:40pm Second Keynote by Rudy Hirschheim on “The attack on understanding: How big data and theory has led us astray” followed by Q&A
10 min Coffee Break
9:00-9:30pm Third Keynote by Alan Hevner and (via Zoom) Shirley Gregor on “The Essential Role of Design Science Research for the Effective Interplay of Analytics and Theory”
9:30pm-10:00pm Panel response to Keynotes: Panelists Sumit Sarkar and Ramesh Sharda
Monday, Dec 16, 2019 (Location: ICM Room 11a/b)
Dinner 5:30-7:00pm (Venue: Bayrische Stube located in H4 Hotel, a few minutes walk from ICM. Click here for menu))
7:00pm-7:15pm: Brief introduction by Leslie Willcocks: Where are the IS Theories in Analytics?
7:15pm-7:50pm Third Keynote, Ahmed Abbasi on “The Pendulum has Swung: From Big Data Hubris to AI Hubris” with Discussion and Q&A
7:50-8:30pm Zoom Guest Speakers Vasant Dhar and Foster Provost on “Prospects of Theory with Big Data Analytics”
10 min Coffee Break
8:45pm-9:30pm Workshop wrap-up discussion by Alan Dennis, Rudy Hirschheim and Leslie Willcocks
References
Abbasi, A., S. Sarker and R. Chiang. (2016). “Big data research in information systems: Toward an inclusive research agenda.” Journal of the Association for Information Systems, 17(2), i–xxxii.
Agarwal, R. and V. Dhar. (2014). “Big data, data science, and analytics: The opportunity and challenge for IS research.” Information Systems Research, 25(3), 443–448.
Anderson, C. (2008). “The end of theory.” Wired, 16(7), 71.
Chang, R. M., R. J. Kauffman and Y. Kwon. (2014). “Understanding the paradigm shift to computational social science in the presence of big data.” Decision Support Systems, 63, 67–80.
Dhar, V. (2013). “Data science and prediction.” Communications of the ACM, 56(12), 64–73.
Kitchin, R. (2014). The Data Revolution: Big Data, Open Data, Data Infrastructures & Their Consequences. Thousand Oaks, CA: SAGE Publications.
Lazer, D., A. Pentland, L. Adamic, S. Aral, A.-L. Barabási, D. Brewer, … M. Van Alstyne. (2009). “Computational social science.” Science, 323(5915), 721–723.
Maass, W., J. Parsons, S. Purao, V. C. Storey and C. Woo. (2018). “Data-driven meets theory-driven research in the era of big data: Opportunities and challenges for information systems research.” Journal of the Association for Information Systems, 19(12), 1253–1273.
Mayer-Schönberger, V. and K. Cukier. (2013). Big Data: A Revolution That Will Transform How We Live, Work, and Think. New York: Houghton Mifflin Harcourt.
Pentland, A. (2014). Social Physics: How Good Ideas Spread — the Lessons from a New Science. New York: Penguin Press.
Shmueli, G. and O. Koppius. (2011). “Predictive analytics in information systems research.” MIS Quarterly, 35(3), 553–572.
Siegel, E. (2016). Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die. New York: Wiley.
Steadman, I. (2013). “Big data, language and the death of the theorist (Wired UK).” Retrieved from http://www.wired.co.uk/news/archive/2013-01/25/big-data-end-of-theory (visited on September 27, 2019)
Weick, K. E. (1995). “What theory is not, theorizing Is.” Administrative Science Quarterly, 40(3), 385–390.
Special Interest Group on Philosophy (SIGPhil) in Information Systems 