STATS IN PARIS - STATISTICS & ECONOMETRICS OF NETWORKS
Network analysis has emerged as a key technique in modern statistics and econometrics to model complex interactions between individuals or complex structures. The mathematical theory used to model networks are closely related to the probabilistic models underlying random graph theory. Evolution of random graphs or random networks, detecting particular structure in large networks (clusters), inference and sampling have given rise to very recent developments in statistics and econometrics. The development of game theory has shown in the late 70's how radically social interactions can affect economic outcomes. On the empirical side, several studies have recently demonstrated and quantified the importance of such interactions in, for instance, school performance, adoption and consumption, engagement in criminal activity, drug abuse, or smoking cessation. An important challenge for many such studies, however, is the nature of traditional microeconomic data used: individuals are usually drawn independently of their social networks, often preventing the possibility to construct such networks. Even when complete networks are observed, assessing the importance of social interactions is not straightforward. For instance, as emphasized by Manski in the early 90’s, it may be impossible to separate the roles of group fixed effect (e.g., the teacher quality when studying school performance of students) from peer effects, i.e. a problem of identification. While several recent studies have made important progress in identifying network effects, it is fair to say that many challenges remain open. Applications and statistical challenges from sociology, marketing and biology will also be presented to provide a large view of the field.
|Event dates:||2013-11-18 – 2013-11-22|
|Organizer:||Centre de Recherche en Economie et Statistique, Ecole Nationale de la Statistique et de l'Administration Economique CREST-ENSAE HEC Paris (École des hautes études commerciales Paris) Centre for Microdata Methods and Practice CEMMAP Université de Cergy-Pontoise Université Paris Ouest|
|Conference venue:||Paris, ENSAE ParisTech|
|Classification:||C0 - Mathematical and Quantitative Methods. General|
|Event type:||Konferenzen, Tagungen; Conferences|