2012-2013 Spring

Henning Silber, Professor, Gottingen University, Germany

April 1, 2013

Accuracy of relations between variables across probability and non-probability samples

ABSTRACT: In recent years internet surveys have become increasingly popular. While face-to-face and telephone surveys are commonly based on probability sampling methods, internet samples, on the other hand, often use non-probability sampling methods. Some researchers have pointed out that surveys based on non-probability sampling methods produce data of less quality (e. g., Faas and Schoen 2006; Yeager et al. 2011) . Particularly, it has been shown that distributions of key variables vary across non-probability internet samples. However, relations between variables were expected to be more stable (e. g., Schnell 1997). Our research compares three different kinds of relations between variables across surveys based on probability and non-probability sampling methods. First, we examine bivariate relationships, second we examine experiments with different question forms, and third we examine more complex relationships with more than two variables. All three kinds of relationships were found to be more stable using probability sampling methods.

Justin Grimmer, Assistant Professor of Political Science, Stanford University

April 8, 2013

The Impression of Influence: How Legislator Communication and Government Spending Cultivate a Personal Vote

Justin Grimmer

Steven Kull, Faculty, School of Public Policy, University of Maryland

April 15, 2013

Giving the Public a Greater Voice in Washington

ABSTRACT: Is there anything that can help move past the polarization and gridlock of Washington today? Harking back to the ideas of Alexander Hamilton, Steven Kull will make the case that new methods of online public consultation‹by which respondents are presented information and arguments,as well as given tools for making nuanced policy prescriptions‹have the potential for creating a new political force. He will present examples of such public consultation in regard to the Federal budget, and then share an ambitious new project for taking these methods to scale in Washington DC.

Steven Kull
Christos Markridis, PhD Candidate, Management Science and Engineering, Stanford University

April 22, 2013

A Unified Theory for Dynamical Environmental Systems: Economic Conditions, Institutions, Social Attitudes in the Interactive Agency Model

ABSTRACT: This paper introduces the Interactive Agency Model (IAM) as a tool for linking micro and macro levels of economic and sociological thinking in climate, environmental, and energy (C2E) systems. Individuals are treated as dynamic programmers whose preferences are a function of three lenses: economic conditions, social attitudes, and institutional structure. Whereas current approaches in this literature focus either on stylized, aggregate macroeconomic drivers or highly individual-specific psychological behaviors, IAM lends itself to modeling the unique interactions among the three criteria in a systems-level framework. This article introduces the conceptual theory and tests it in a simplified econometric setting in order to communicate the key insights. Results suggest that these interactive elements are both statistically and economically relevant for not only theoretical work, but also policy consideration.

Christos Markridis
Jasjeet Sekhon, Professor of Political Science and Statistics, University of California Berkeley

April 29, 2013

From SATE to PATT: Combining Experimental with Observational Studies to Estimate Population Treatment Effects

ABSTRACT: Randomized Controlled Trials (RCTs) can provide unbiased estimates of sample average treatment effects. However, a common concern is that RCTs often fail to provide unbiased estimates of population average treatment effects. We derive the assumptions for identifying population average treatment effects from RCTs. We provide a set of placebo tests, which formally follow from the identifying assumptions, that can assess whether the assumptions hold. We offer new research designs for estimating population effects that uses observational studies to adjust the RCT data. One design does not require a selection on observables assumption. We apply our approach to a cost-effectiveness analysis, and applications to other settings such comparing the estimates of survey experiments from different samples are discussed.

Jasjeet Sekhon
Gabor Simonovits, PhD Candidate, Political Science, Stanford University

June 3, 2013

One for the proponents, another for opponents: The consequences of modes of explanations for policy

ABSTRACT: Qualitative researchers highlight the importance of the explanations that politicians offer for the actions they take. The right justification may increase the support of voters, even if they oppose the action itself. In this paper we analyze the political consequences of different "styles" of rhetorics politicians use and seek to explain why different types of justifications are employed by different speakers to different audiences. We delineate two common kinds of justifications: those that are based on evidence (thus are falsifiable) and those that rely on value. Using this typology, we conducted a survey experiment in which voters evaluated fictional candidates who took and explained their positions on a proposed increase in the income tax rates for wealthy Americans. We find evidence suggesting that while arguments based on evidence (as opposed to values) help politicians when communicating to an audience with opposing opinions, but hurts them when they talk to people who agree with their positions. Moreover, our results suggest that the effect of justifications are mediated by perceived distance of the candidate's positions from the respondent's and the respondent' certainty in the candidate's position.

Gabor Simonovits