by Mark Zbaracki (Guest Author)
The questions of impact raised here recalled Jim March’s longstanding claim: “I am not now, nor have I ever been, relevant” (March 2006, p. 83). In The Roots, Rituals, and Rhetorics of Change, he and Mie Augier point out that our standard measures of relevance are utilitarian. That is, we believe our teaching and scholarship should satisfy the needs and desires of both students and their employers. But they also point out that relevance is ambiguous, hard to measure, and complex. Moreover, the relevance that we pursue is often myopic across both space and time. Many of these posts wrestle with the problems ambiguity and myopia introduce. Who decides what is relevant or what utility? How do we choose standards for relevance? And standards of relevance change over time, so relevant when?
While March may have repeatedly protested that he was not interested in being relevant, his ideas also repeatedly returned to the problems of … well, relevance! In one of my favorite articles (March 1972), and in his book An Introduction to Models in the Social Sciences (Lave and March 1975), he wrestled with how we consider the social implications of our models— “speculations,” as he called them. He argued that our models may be very accurate and yet endorse unattractive behaviors. That theme nagged at me as I thought about Jeff Pfeffer’s approach to power and politics. Pfeffer (2010; 2015) repeatedly points out that people garner power quite predictably through some very unseemly means: flattery, insincerity, rule-breaking, dishonesty, and deception, among others. To understand power, he argues, we must drop our “just-world” assumptions because pursuing power may have consequences, we consider unjust. What, then, do we do with our concerns for justice?
March’s point was that predictive validity is but one way to evaluate our models. A speculation can be evaluated by three different criteria: truth, beauty, and justice. My colleagues and I take on these three criteria in a recent essay (Zbaracki, et al. 2021) in Research in the Sociology of Organizations dedicated to the work of Jim March. Truth is our standard metric for a model’s value but models do more than help us predict and control; they ask us to consider and reconsider our social worlds. Consider beauty. It beckons us; it calls us to consider something beyond our current understanding. It is more than “interesting”: it is “a guess, a suspicion, a dim awareness” of something beyond our current understanding (Nehamas 2000, p. 5). Finally, justice asks about the kind of world we want. Most models follow the utilitarian logic of economics and focus on the technical problem of resource exchanges that maximize happiness. But justice can also be treated as a political problem focused on how power and resources are allocated and how preferences are aggregated across actors. Or justice can be treated as a moral problem, focusing on the virtues necessary to build a good society.
We do not advocate replacing truth with beauty or justice. Rather, we suggest taking a wider view of the worlds we inhabit and how our work influences those worlds. Our essay points out how our desire to produce social science models to serve the good of society can create some dangerous traps. When we seek to make organizations more responsive to social problems, we may find that managers and leaders use our ideas in unanticipated ways. To avoid those traps, it may be helpful to consider relevance in two questions. First, how do we consider the truth value of our models? Our utilitarian approaches to relevance focus on models for prediction and control. But we can also build models simply for understanding. Second, how do we attend to beauty and justice? Incorporating beauty and justice can help us choose the aspects of social reality that we will consider, bring artistry to our models of that social reality, and help us accept worlds in which prediction and control do not work as we expect. It is an approach to social science that is humbler, because it depends on patient inquiry rather than on great designs. But that humble approach has its own form of beauty.
References
Augier, M., & March, J. G. (2011). The roots, rituals, and rhetorics of change: North American business schools after the Second World War. Stanford, CA: Stanford University Press.
Lave, C. A., & March, J. G. (1975). An introduction to models in the social sciences. New York, NY: Harper & Row.
March, J. G. (1972). Model bias in social action. Review of Educational Research, 42, 413–429.
March, J. G. (2006). Ideas as art. Harvard Business Review, 84(10), 82–89.
Nehamas, A. (2000). An essay on beauty. The Threepenny Review, 80(4), 4–7.
Pfeffer, J. S. (2010). Power: Why some people have it–and others don’t. New York, NY: Harper Collins.
Pfeffer, J. S. (2015). Leadership BS: Fixing workplaces and careers one truth at a time. New York, NY: Harper Collins.
Zbaracki, M.J., Watkiss, L., McAlpine, C., Barg, J. (2021). Truth, beauty, and justice in models of social action. Research in the Sociology of Organizations. 76, 159–177
Aspiring to both beauty and justice as additional criteria for research seems a good way to broaden our conceptualization of collective impact and to contemplate how to go about actually having impact. Justice, for instance, requires going beyond financial performance to outcomes that are important to society, and in doing so, requires us to conceptualize our impact in long-term ways that might influence policy, as Anita McGahan illustrates: https://mackinstitute.wharton.upenn.edu/2023/what-is-collective-impact-really/.
Beauty is perhaps a bit more complicated, as “standards” of beauty are difficult to pin down. When people think about their work, however, they often include conceptualizations of beauty. Engineers hate kluge-y fixes to problems, computer programmers admire the elegance of their colleagues’ code, and technicians pride themselves on their workmanship. What does this mean for management scholars? For models, as noted here, beauty may mean bringing artistry to incorporating variables. For ethnographers, prose that vividly evokes a situation and theory that reflects people’s experience of organizations might be more salient aspects of beauty. Regardless of the specifics, using beauty as a criteria for impact means articulating a standard for good work, linked to our research practices, that resonates with our community.
Mark’s thoughtful post led me to realize that often when academics refer to “relevance” they mean the relevance to non-academics of the theoretical propositions or empirical findings of academic research. However, there is a different type of relevance that is sometimes overlooked. Recently, I was describing some research questions investigated by a colleague, and someone asked me where those questions came from. I answered that the questions came from the real world. It strikes me that research on questions that are sparked by what we observe in the real world is relevant by definition. Whether or not the ideas or findings that we come up with are immediately applicable in the real world is an altogether different issue, and we should not confound whether managers or policy makers find a study useful with the relevance of the study.
Mark’s comments remind us that performance measures, whether for firms struggling to navigate double bottom lines or social scientists trying to do work of value, tend to be complex and multi-dimensional. Our current moment with advanced machine learning models that offer the prospect of high levels of predictive accuracy pose an important instance of the challenge of what performance gradients add most effectively to “collective impact”. While the capacity to predict, whether what applicant might make the preferred hire or how individuals are likely to respond to a government policy, is clearly valuable, a different, though not unrelated agenda of social science is understanding. Prediction pushes us to ever more granular measures and complex composites of those measures. Understanding invites simplification and abstraction. Such characterizations suffer some loss in accuracy as bases of explanation in any given instance, but they allow us to build broader and potentially robust explanations. Further, understanding is critical in motivating action. To tell a product development team that this is the prototype that the algorithm picked is likely less inspirational than a vision and logic that has meaning to the development team. Following March and Zbaracki, we need to be mindful of the diverse ways in which research might be impactful.