Past Events

  • 2023 Sep 28

    Leo Burstzyn (Alesina Seminar)

    4:30pm to 5:45pm

    Location: 

    Littauer 301 (Hansen-Mason Lounge)

    Today's Speaker

    Leo Burstzyn (UChicago), "When Product Markets Become Collective Traps: The Case of Social Media"

    Abstract

    Many products generate negative consumption spillovers to potential non-users. One may experience disutility when one is not consuming the product, while others are; for example, if consuming the product gives social status, or if not consuming it leads to social isolation. We show that, in the presence of such spillovers, standard measures that take the aggregate consumption as given fail to appropriately capture welfare. We propose a new methodology to measure welfare that allows for these consumption spillovers, which we apply to estimate the consumer surplus of two popular social media platforms, TikTok and Instagram. In large-scale, incentivized experiments with college students we show that, while the standard welfare measure suggests large and positive surplus, our measure accounting for consumption spillovers indicates negative surplus, with a large share of active users even deriving negative utility from the platforms. We also shed light on the drivers of consumption spillovers to non-users in the case of social media, and show that, in this setting, the "fear of missing out'' plays an important role. Our framework and estimates highlight the possibility of product market traps, where large shares of consumers are trapped in an inefficient equilibrium and would prefer the product not to exist.
    ... Read more about Leo Burstzyn (Alesina Seminar)

  • 2023 Sep 27

    Tyler Simko (Workshop in Applied Statistics)

    12:00pm to 1:30pm

    Location: 

    CGIS Knafel, room K354 or Online via Zoom

    This Week's Speaker

    Tyler Simko (Department of Government), "School Desegregation by Redrawing District Boundaries: Evidence from New Jersey"

    Abstract

    Schools in the United States remain heavily segregated by race and income. Previous work demonstrates districts can reduce segregation between their schools with policies like redrawing attendance zones. Yet, the promise of such policies in many areas is limited by the fact that most school segregation occurs between school districts, and not between schools in the same district. I adapt Markov Chain...

    Read more about Tyler Simko (Workshop in Applied Statistics)
  • 2023 Sep 27

    Dataverse Open Office Hours

    Repeats every week every Wednesday until Wed Aug 28 2024 except Wed Dec 27 2023, Wed Mar 13 2024, Wed Jun 19 2024.
    11:00am to 1:00pm

    Location: 

    Virtual via Zoom

    Weekly office hours are open to Harvard researchers and staff to provide support for Dataverse 6.0. Demo of 6.0 will begin promptly at 11am.

    Open Hours: Wednesdays, 11AM - 1PM

    RSVP required to: support@dataverse.org

    For any questions on how to share your data with Dataverse, contact: support@dataverse.org

     

  • 2023 Sep 26

    Tyler Simko (APRW)

    12:00pm to 1:30pm

    Location: 

    CGIS Knafel, room K354 or Online via Zoom

    Speaker

    Tyler Simko (Dept of Government), "School Desegregation by Redrawing District Boundaries"

    Abstract

    Schools in the United States remain heavily segregated by race and income. Previous work demonstrates districts can reduce segregation between their schools with policies like redrawing attendance zones. Yet, the promise of such policies in many areas is limited by the fact that most school segregation occurs between school districts, and not between schools in the same district. I adapt Markov Chain Monte Carlo (MCMC) algorithms from political...

    Read more about Tyler Simko (APRW)
  • 2023 Sep 21

    Francesco Squintani (Alesina Seminar)

    4:30pm to 9:45pm

    Location: 

    Littauer 301 (Hansen-Mason Lounge)

    Today's Speaker

    Francesco Squintani (University of Warwick), "The Choice of Political Advisers,” (w/Hyungmin Park)

    Abstract

    We study the choice of multiple advisers, balancing loyalty, competence, and diverse perspectives. A leader can consult one or both of two advisers. One has views that align closely with the leader's, but her information is imprecise or correlated with the leader's. The other is more biased but has independent and more precise information. We find that the leader consults the most informed expert, adding the other adviser only if the additional communication does not hinder truthful advice from the former. If the leader consults both advisers, we paradoxically find that increasing the more biased expert's bias causes the dismissal of the other adviser. Hence, information trumps political proximity, when seeking advice. Exactly the opposite happens when it comes to delegating decision-making. The leader may choose to delegate, but only to the less biased adviser. The analysis is generalized to the case where the most informed adviser is not necessarily more biased. Reducing the probability that the more informed adviser is also more biased leads to hiring also the other expert. The leader may delegate to the uncertain-bias adviser, although she is possibly more biased, because she is better at aggregating information.... Read more about Francesco Squintani (Alesina Seminar)

  • 2023 Sep 20

    Introduction to Data Management & FAIR

    1:00pm to 2:00pm

    Location: 

    Online via Zoom

    Registration Required - click here to register

    Presenter: Katie Mika

    Join us for an in-depth exploration of advanced data management and the FAIR principles within the context of Dataverse repositories. This webinar goes beyond the basics, offering a comprehensive overview of how the FAIR principles - Findability, Accessibility, Interoperability, and...

    Read more about Introduction to Data Management & FAIR
  • 2023 Sep 20

    Larry Han (Workshop in Applied Statistics)

    12:00pm to 1:30pm

    Location: 

    CGIS Knafel, room K354 or Online via Zoom

    This Week's Speaker

    Larry Han (Northeastern), "Federated transfer learning for robust inference of causal effects"

    Abstract

    Federated or multi-site studies have distinct advantages over single-site studies, including increased generalizability, the ability to study underrepresented populations, and the opportunity to study rare exposures and outcomes. However, these studies are challenging due to the need to preserve the privacy of each individual's data and the heterogeneity in their covariate distributions. We propose a novel federated approach to...

    Read more about Larry Han (Workshop in Applied Statistics)
  • 2023 Sep 20

    Dataverse Open Office Hours

    11:00am to 1:00pm

    Location: 

    Virtual via Zoom

    Weekly office hours are open to Harvard researchers and staff to provide support for Dataverse 6.0. Demo of 6.0 will begin promptly at 11am.

    Open Hours: Wednesdays, 11AM - 1PM

    RSVP required to: support@dataverse.org

    For any questions on how to share your data with Dataverse, contact: support@dataverse.org

     

Pages