Dear all,
Today at Applied Stats<http://projects.iq.harvard.edu/applied_stats> we will be hearing from Stefan Hoderlein (Harvard Economics) who will be speaking on "Consumer Demand and Welfare Estimation in a Heterogeneous Population." As per usual, the talk will be held in CGIS K354<http://map.harvard.edu/?bld=04471&level=9> at 12 noon and lunch will be served.
Abstract:
This is an overview about own recent econometric work related to the modeling of heterogeneity in applied consumer demand models. The focus will be on non-parametric random coefficient models. The main application will come from estimating gasoline demand; in particular, estimating the distribution of welfare effects of a 5% gasoline price change.
I hope to see you all there!
Tess
-----------------
Tess Wise
PhD Candidate
Harvard Department of Government
http://tesswise.com
Dear All,
Tomorrow at Applied Stats<http://projects.iq.harvard.edu/applied_stats> we will be hearing from Tyler VanderWeele, Professor of Epidemiology at the Harvard School of Public Health. Professor VanderWeele will be presenting on Surrogate Measures and Consistent Surrogates. As per usual, the talk will be held in CGIS K354<http://map.harvard.edu/?bld=04471&level=9> at 12 noon and lunch will be served.
Abstract: Surrogates which allow one to predict the effect of the treatment on an outcome from the effect of the treatment on the surrogate are of interest when it is difficult or expensive to measure the primary outcome. There have, however, been several instances of drugs that have been approved for use on the grounds of randomized trials using surrogate outcomes, that have subsequently led to public health catastrophes, costing thousands of lives. It is now clear that the use of surrogates can give rise to paradoxical situations in which the effect of the treatment on the surrogate is positive, the surrogate and outcome are strongly positively correlated, but the effect of the treatment on the outcome is negative, a phenomenon sometimes referred to as the "surrogate paradox." New results are given for consistent surrogates that extend the existing literature on sufficient conditions that ensure the surrogate paradox is not manifest. Specifically, it is shown that for the surrogate paradox to be manfiest it must be the case that either there is (i) a direct effect of treatment on the outcome not through the surrogate and in the opposite direction as that through the surrogate or (ii) confounding for the effect of the surrogate on the outcome, or (iii) a lack of transitivity so that treatment does not positively affect the surrogate for all the same individuals for which the surrogate positively affects the outcome. The results are related to several common approaches and measures for assessing surrogacy including the "proportion explained" and the Prentice criteria, the "proportion mediated", meta-analytic approaches, and principal strata effects. None of these measures or approaches entirely protect against the surrogate paradox. An attempt is made to synthesize the existing approaches and results into guidelines on avoiding the surrogate paradox and ensuring consistent surrogates.
You can also find a link to the corresponding paper here<http://projects.iq.harvard.edu/applied_stats/event/tyler-vanderweele-school…>.
Tess
-----------------
Tess Wise
PhD Candidate
Harvard Department of Government
http://tesswise.com
Dear all,
I am pleased to announce that tomorrow at Applied Stats we will hear a talk from Professor Amy Silva (Northeastern and Charles River Analytics) entitled Scalable Analysis of Behavioral Models and Decision-Making.
As per usual, the talk will be held at 12noon in CGIS K354<http://map.harvard.edu/?bld=04471&level=9> and lunch will be served.
Abstract: The ability to model, forecast, and understand the behavioral dynamics and decision-making patterns of human agents has applications in many contexts. One particularly salient domain is the field of international security where artificial intelligence models can be leveraged to analyze complex and uncertain security situations. Real world datasets can contain 10^30,000 possible behaviors—requiring efficient techniques to manage the confluence of cultural, social, economic, political, and temporal information. In this talk, I will present a probabilistic logic formalism, the Stochastic Opponent Modeling Agents (SOMA) framework, and several scalable reasoning algorithms for modeling behavioral dynamics. SOMA has been used to study the Afghan drug trade, violent ethnopolitical conflicts in the Middle East and Asia Pacific, and the terror organization Lashkar-e-Taiba. Interpreting and using these models in national security settings requires human insight into characteristic relationships of the domain as well as computational methods such to handle the overwhelming quantity of data. I will briefly discuss the Model Analyst’s Toolkit, a software tool designed to leverage both human knowledge and computational power to refine models and aid in decision-making, and the SOMA Terror Organization Portal (STOP), a prototype system that allows users throughout the national security community to analyze the behaviors of violent organizations.
Tess
-----------------
Tess Wise
PhD Candidate
Harvard Department of Government
http://tesswise.com