PhD

Economic Sociology Seminars

Seminars take place on Wednesdays 4:00-5:30 p.m in room E62-450 and also over Zoom unless otherwise noted. Please contact Jessica Lipsey (jessi71@mit.edu) for additional details, or if you wish to be added to the mailing list to receive updates. 

Spring 2024

  • February 14, 2024

    James Chu, Assistant Professor of Sociology, Columbia University

    How Beneficiaries Become Sources of Normative Control

     Organizations can motivate and coordinate work by socializing members to internalize organizational values. Existing theories posit that organizations achieve normative control through encapsulation, wherein peers and managers are primary sources of members’ socialization. Drawing on ethnographic data from a not-for-profit school, I show how an external actor—beneficiaries—can become a source of normative control. I develop a multistage process that explains how teachers were socialized by parents, specifically by hearing these parent beneficiaries narrate their needs; engaging in collective storytelling about beneficiaries; experiencing episodic shaming centered on how teachers’ daily performance met (or did not meet) beneficiaries’ needs; and receiving validation from beneficiaries. Because these sequential stages establish beneficiaries as sources of control through social interactions set in specific times and places, and establish shared emotional states among organizational members, I theorize that these stages compose a ritual of integration. Although teachers initially arrived at the school with heterogeneous values, this ritual led many of them to internalize the organizational value of self-sacrifice. Teachers who were unmoved by parents’ stories or came to see parents as exploitative did not internalize this value, and they tended to exit the organization. This study reveals how normative control can arise not only through socialization from in-group members but also from ritual interactions with and about beneficiaries.

  • February 28, 2024

    Luis Flores, Harvard Kennedy School, Harvard University

    The Crisis of Asset Homeownership: Remaking Landed Economic Security in Neoliberal America

    It has become common to interpret the transformation of household life in contemporary America through the lens of “financialization.” However, in important ways, the postwar home had long been “financialized”—reduced to exchange value rather than the sites of productive or commercial life that anchored older notions of landed independence. Drawing on two historical case studies, I propose a fuller theory of what changed since the 1980s, involving the shifting foundations of homes as sources of economic security. I trace divergent and conflicting efforts to turn homes into liquid and short-term resources by supporting the deregulation of second mortgages (for home equity loans) and for the relaxation of zoning restrictions on accessory units in single-family homes (enabling rental incomes) in the 1980s. Advocates for an emerging “liquid homeownership” envisioned a shift away from a regime I term “asset homeownership.” Between the 1930s and 1980s, mortgage regulations and zoning rules built a dam around the single-family home, protecting it from devaluation but “locking in” home equity, while restricting home-based commerce. Asset homeownership promised postwar owners access to a long-term, stable, but ultimately illiquid asset, that instead offered old-age security and access to place-based privileges, from tax benefits to quality school. In the late 1970s, intersecting crises of inflation, old age housing, and the rise of immigrant populations, propelled divergent strategies to transform homes into short-term and liquid assets through mortgage and zoning deregulation. In both cases, homeowners and regulators clashed over the desirability of protections that previously were the foundations of asset homeownership. Meanwhile for policymakers, “liquid homeownership” offered a third way between public policy and incentives for private initiative to address the multiple crises of the 1970s.

  • March 13, 2024

    Pat Reilly, Assistant Professor in the Organizational Behaviour and Human Resources Division of University of British Columbia’s Sauder School of Business

    Point Break? The Process and Impact of Collaborative Breakdowns in Creative Work

    Popular images of creative leaders often depict them as difficult to work with and yet, worth the struggle because of their ability to innovate: the “creative jerk.” Leading creative group work requires, by design, managing differences in perspective, vision, and ideas – creating some level of friction within the group, while also managing the collaboration so that the differences sparked by creativity do not lead to impasses. In this multimethod study, we examine what happens when the creative process breaks down and leader can no longer work with the group: a collaborative breakdown. We address two questions: How do creative leaders manage disagreements to avoid collaborative breakdowns? What are the consequences of collaborative breakdowns for creative leaders’ careers? In a qualitative study of Hollywood insiders, we uncover two complimentary processes: collaborative plasticity, how creative leaders attempt to keep a collaboration moving forward, and creative vitrification, how incompatibilities between ideas harden, leading to “collaborative breakdowns.” We then examine the long-term career impact for creative leaders associated with a creative breakdown using an archival sample of Hollywood directors leaving projects due to “creative differences.” Empirical evidence reveals that creative leaders leaving projects experience drop offs in future employability and creativity success.

  • March 20, 2024

    Paul Leonardi, Ph.D., Department Chair and Duca Family Professor of Technology Management, UC Santa Barbara.

    Black Box or Pandora’s Box? How Decisions About Organizing Shape Participation and Inclusion in the Development and Use of AI Models

    A growing body of research and commentary on the proliferation of AI in organizations bemoans the high levels of opacity associated with most machine learning models and calls for models whose inner-workings are more transparent to both developers and users. In pursuit of this goal, we explored how organizations could encourage more participation and inclusion of diverse groups in the development and use of AI models at the fine-tuning stage. Our year-long field study of two metropolitan planning organizations in the Western United States revealed very different choices about how to organize around the development and implementation of AI models. The data did indeed reveal how an organization could make decisions that simultaneously broadened participation in the development of AI models and do so in a way that allowed new participants to actually see into the black box. But surprisingly, the data also showed that when members of key groups were able to contribute directly to the model’s development and could understand what it was doing and why (all qualities of transparency), they pushed for their own interests at the expense of community goals, debated ancillary topics, and stymied the very process for which the model was deployed. These findings caution us to consider how prying the lid off of the black box of AI models can create a Pandora’s box of problems for decision making and planning. We argue that models cannot simply be “made” more transparent because opacity is not a quality inherent in an AI model, but an emergent process that arises from people’s experience actively constructing its working.

  • April 10, 2024

    Chengwei Liu, Associate Professor of Strategy and Behavioral Science, ESMT Berlin

    Rank Reversals and the Limits of Learning from Successes and Failures: Empirical Evidence and Public Predictions from Billboard Singles, Auto Racing, and Firm Growth

    Conventional wisdom and theories of learning often endorse a “higher-is-better” principle. Individuals and organizations repeat actions associated with past success and avoid those linked to failures, assuming that superior current performances predict superior future performances. In contrast, prior studies identified a phenomenon of “rank reversals,” where superior performers obtain worse future performances than their worse-ranked counterparts when the impact of luck overwhelms that of agency. Analyzing 429,902 actor performances across three domains—Billboard Hot Singles, auto racing, and firm growth—we find consistent evidence for rank reversals around exceptionally high or low performances. In a subsequent prediction task involving 642 participants sourced from Prolific, we find that while the majority acknowledged the impact of luck as their predictions reflect regression-to-the-mean, they failed to predict rank reversals. Notably, participants with domain-specific knowledge were even more prone to this prediction bias, especially when evaluating individuals (e.g., musicians or drivers) as opposed to firms (e.g., record labels, public firms). An exception emerges among those with economics training: around half of them accurately predicted rank reversals in high-growth firms, albeit often for flawed reasoning as anticipated by the fundamental attribution error. Our findings challenge prevailing assumptions about the monotonic relationship between current and future performance, revealing the limits of learning from past successes or failures. This oversight can result in systematic errors in decision-making processes, such as hiring and resource allocation, thereby creating untapped strategic opportunities for those who appreciate the nuanced impact of luck.

  • May 1, 2024

    Robert Freeland, Professor of Sociology, University of Wisconsin

  • May 8, 2024

    Carly Knight, Assistant Professor, Department of Sociology, New York University

  • May 15, 2024

    TBD