Archivio istituzionale della Ricerca - Bocconi
Not a member yet
30225 research outputs found
Sort by
Partially factorized variational inference for high-dimensional mixed models
No full textWhile generalized linear mixed models are a fundamental tool in applied statistics, many
specifications, such as those involving categorical factors with many levels or interaction terms,
can be computationally challenging to estimate due to the need to compute or approximate 15
high-dimensional integrals. Variational inference is a popular way to perform such computations,
especially in the Bayesian context. However, naive use of such methods can provide unreliable
uncertainty quantification. We show that this is indeed the case for mixed models, proving that
standard mean-field variational inference dramatically underestimates posterior uncertainty in
high-dimensions. We then show how appropriately relaxing the mean-field assumption leads to 20
methods whose uncertainty quantification does not deteriorate in high-dimensions, and whose
total computational cost scales linearly with the number of parameters and observations. Our
theoretical and numerical results focus on mixed models with Gaussian or binomial likelihoods,
and rely on connections to random graph theory to obtain sharp high-dimensional asymptotic
analysis.We also provide generic results, which are of independent interest, relating the accuracy 25
of variational inference to the convergence rate of the corresponding coordinate ascent algorithm
that is used to find it. Our proposed methodology is implemented in the R package vglmer.
Numerical results with simulated and real data examples illustrate the favourable computation
cost versus accuracy trade-off of our approach compared to various alternatives
Influencer activism: insights for effective partnership with brands and organizations
Get PDFInfluencers can effectively promote products and brands but are also leading personalities who might inspire others to support polarizing and/or prosocial issues (e.g., against gender‐based violence, social inequality, and racism). This research analyses the impact of influencer activism on perceived authenticity and prosocial behaviors, focusing on collaborations with brands and nonprofit organizations. Drawing on social influence theory and signaling theory, two key factors are examined: the influencerpartner congruity, and the influencer‐sociopolitical issue alignment. The research consists of a preliminary study, four experimental studies on behavioral intentions, two experimental studies on proxies of actual behavior, and a content analysis on Instagram comments. The results suggest that congruity and alignment significantly enhance perceived authenticity, which positively impacts attitudes toward the influencer and increases the intention to engage in prosocial behaviors. The article offers further insight into individual engagement in actual prosocial behavior, such as seeking information on an issue, subscribing to newsletters, and signing petitions. The research underscores the importance of selecting congruent influencers, whose values and advocacy efforts are consistent with the brand's or organization's characteristics, and aligned with relevant issues, to enhance perceived authenticity, foster genuine engagement and effectively drive prosocial behaviors through effective partnerships
Teacher personality and the perceived socioeconomic gap in student outcomes
No full textWe randomly assign student profiles to teachers and elicit teachers’ beliefs about the student’s likelihood of success in alternative high school tracks. We document a large and statistically significant gradient in teachers’ beliefs about students’ high school prospects with respect to students’ socioeconomic background (SEB), ceteris paribus. We find that this gradient varies with teacher’s personality, a hard-to-observe and understudied teacher trait. Specifically, higher levels of teacher’s extraversion and openness are associated with a steeper negative SEB gradient in teachers’ beliefs about students’ success prospects in an academic track. Conversely, more conscientious and agreeable teachers assign to low-SEB students, on average, a higher probability of success in a vocational track. We discuss some policy implications of our findings
Integrating local market operations into transmission investment: a tri-level optimization approach
Get PDFThe rise of Local Energy Markets (LEMs) and increasing local flexibility present a key research question: How do local flexibility and LEM operations impact merchant-regulated transmission investments? This paper introduces a novel tri-level framework to integrate local market dynamics into transmission investment decisions. The framework models the sequential operations of the WSM and LEMs, adhering to their respective network constraints, and includes a regulatory mechanism that incentivizes profit-driven Transmission Companies (Transcos) to make social welfare maximizing investments while accounting for local refinement costs. The tri-level optimization problem is asymptotically approximated by a mixed-integer second-order cone programming problem. Our findings from three case studies reveal that the provision of local flexibility substantially reduces reliance on conventional energy generation supplies. Additionally, transmission investment decisions are influenced by the levels of flexible generation and consumers, while adhering to network constraints. Moreover, the tri-level model enhances Transcos’ awareness of the sequential interactions between the WSM and LEMs, enabling them to make investment strategies that are responsive to the changing dynamics of local markets
Procuring medical devices: the price effect of mergers among orthopedic prostheses producers
Get PDFThis paper quantifies the price effects of a merger between two major producers of
orthopedic prostheses. It shows that, in the public procurement markets where these
products are purchased, the effect of the merger hinges on the characteristics of both
the procurement design and the organizational structure of the buyers
How to organize in turbulence: arrangements and pathways for robust governance
No full textRobustness has recently taken center stage as an emerging paradigm to cope with turbulence and “build back better” toward new normalcy. Existing literature has shown how robust governance, with its mix of flexible adaptation and proactive innovation, is well-suited to addressing turbulence. However, there remains a gap in understanding the empirical variations within robust governance arrangements. In this article, we address three questions: how (1) structures, (2) coordination mechanisms, and (3) leadership are designed and unfold in robust governance. Through a qualitative approach grounded in case studies, interviews, and archival data, we provide evidence from six Italian regions, examining how they addressed the challenges of the COVID-19 vaccination campaign. Results enable the formulation of propositions about organizational arrangements in robust governance, in addition to suggesting competing pathways for flexible adaptation and proactive innovation
Hybrid managers in an evolving healthcare: does gender matter?
No full textIntroduction: Hybrid managers have the potential to respond to the need for more integrated, responsive and accountable healthcare. Scholars have studied the antecedents of hybridization, but the role of gender has been neglected. Therefore, we study whether and how gender impacts on the way in which medical professionals exercise their managerial role. Methods: We adopted a qualitative approach in order to gain an in-depth understanding of the specificities of women hybrids. Data was collected through semi-structured interviews, focusing on hybrids in Italy in the field of neurology. Results: We found that women hybrids show specific abilities and motivations, but they also encounter a specific lack of opportunities. Women hybrid managers appear well positioned to foster the evolution of professionalism, but healthcare organizations should implement policies and practices to effectively support them. Conclusion: While existing research has treated hybrid managers as a homogenous group, we underline the specificities of women hybrids. They can support the evolution of healthcare organizations towards logics of service integration, user centricity, and staff engagement. Therefore, our findings have important theoretical and practical implications for health policy and management
Manifolds, Random Matrices and Spectral Gaps: The geometric phases of generative diffusion
No full textWomen in economics: The role of gendered references at entry in the profession
No full textWe study the presence and the extent of gender differences in reference letters for graduate students in economics and finance, and how they relate to early labor market outcomes. To these ends, we build a novel rich dataset and combine Natural Language Processing techniques with standard regression analysis. We find that men are described more often as standout and women as grindstone, i.e., hardworking and diligent; these differences are mainly driven by male letter writers, especially more senior ones. We then show that the former (latter) characteristics relate positively (negatively) with various subsequent career outcomes and that women obtain lower positive (marginally larger negative) returns from being described as standout (grindstone). We argue that, taken together, this evidence is consistent with the presence of implicit gender stereotypes as driving the observed differences in the way candidates are described
Between Stick and Carrot: Unpacking Schauer's Account of Legal Coercion
No full textIn The Force of Law, Fred Schauer challenges the view that sanctions are philosophically
marginal to law, and proposes to replace conceptual analysis with empirical research
in jurisprudence. This article defends a complementary role for conceptual analysis in legal
theory. By comparing the threat of punishment and the promise of reward, it shows how conceptual
inquiry can generate explanatory hypotheses about the features of legal phenomena,
and clarify the normative structure of legal regulation