Calendrier du 28 mai 2024
PSI-PSE (Petit Séminaire Informel de la Paris School of Economics) Seminar
Du 28/05/2024 de 17:00 à 18:00
R1-14
ANDRé Loris (PSE)
Land Conversion and Species Extinction
Econometrics Seminar
Du 28/05/2024 de 14:45 à 16:00
Sciences Po, room H405
MOLINARI Francesca (Cornell University)
Inference for an Algorithmic Fariness-Accuracy Frontier
écrit avec Co-author: Yiqi Liu
Decision-making processes increasingly rely on the use of algorithms. Yet, algorithms' predictive ability frequently exhibit systematic variation across subgroups of the population. While both fairness and accuracy are desirable properties of an algorithm, they often come at the cost of one another. What should a fairness-minded policymaker do then, when confronted with finite data? In this paper, we provide a consistent estimator for a theoretical fairness-accuracy frontier put forward by Liang, Lu and Mu (2023) and propose inference methods to test hypotheses that have received much attention in the fairness literature, such as (i) whether fully excluding a covariate from use in training the algorithm is optimal and (ii) whether there are less discriminatory alternatives to an existing algorithm. We also provide an estimator for the distance between a given algorithm and the fairest point on the frontier, and characterize its asymptotic distribution. We leverage the fact that the fairness-accuracy frontier is part of the boundary of a convex set that can be fully represented by its support function. We show that the estimated support function converges to a tight Gaussian process as the sample size increases, and then express policy-relevant hypotheses as restrictions on the support function to construct valid test statistics.
Texte intégral
GPET Seminar
Du 28/05/2024 de 13:45 à 17:30
R1-13
• 13:40 -14:20 : Gísli Gylfason : From Tweets to the Streets : Twitter and Extremist Protests in the United States
• 13:20-15:00 : Vitaliia Eliseeva : Failing to forge the New Soviet Woman: Long-term effect of WW2-induced sex ratios on family formation.
Break
• 15:15-15:55 : Maximilian Dörfler : The Political Consequences of Adverse Labor Market Shocks
• 15:55-16:35 : Ninon Moreau-Kastler Multiplicative difference-in-differences, staggered treatment and proportional treatment effect.
Break
• 16:50-17h30 : Heddie Moreno : export survival of African firms
Applied Economics Lunch Seminar
Du 28/05/2024 de 12:30 à 13:30
Salle R2.21
VAN EFFENTERRE Clémentine( University of Toronto)
AMER Abdelrahman(University of Toronto)
CRAIG Ashley(Australian National University)
Decoding Gender Bias: The Role of Personal Interaction
Subjective performance evaluation is an important part of hiring and promotion decisions. We combine experiments with administrative data to understand what drives gender bias in such evaluations in the technology industry. Our results highlight the role of personal interaction. Leveraging 60,000 mock video interviews on a platform for software engineers, we find that average ratings for code quality and problem solving are 12 percent of a standard deviation lower for women than men. Half of these gaps remain unexplained when we control for automated measures of coding performance. To test for statistical and taste-based bias, we analyze two field experiments. Our first experiment shows that providing evaluators with automated performance measures does not reduce the gender gap. Our second experiment removed video interaction, and compared blind to non-blind evaluations. No gender gap is present in either case. These results rule out traditional economic models of discrimination. Instead, we show that gender gaps widen with extended personal interaction, and are larger for evaluators educated in regions where implicit association test scores are higher.