Political scientists are often interested in estimating causal effects. Identification of causal estimates with observational data invariably requires strong untestable assumptions. Here, we outline a number of the assumptions used in the extant empirical literature. We argue that these assumptions require careful evaluation within the context of specific applications. To that end, we present an empirical case study on the effect of Election Day registration on turnout. We show how different identification assumptions lead to different answers, and that many of the standard assumptions used are implausible. Specifically, we show that EDR likely had little effect on turnout in any state including Minnesota and Wisconsin. We conclude with an argument for stronger research designs.

We study the conditions under which estimating the incumbency advantage using a regression discontinuity (RD) design recovers the personal incumbency advantage in a two-party system. Expanding on the interpretation proposed by Lee (2008), who introduced RD as a method for estimating the party incumbency advantage, we develop a simple model that leads to unbiased estimates of the personal incumbency advantage and yields the surprising result that the RD effect double counts the personal incumbency advantage. Obtaining this result is possible under certain conditions, which we argue hold in the U.S. House when the analysis is restricted to open seats. We apply our model to analyze non-southern open-seat U.S. House elections between 1968 and 2008, where we estimate a personal incumbency advantage of about 7 percentage points. We also explore the estimation of the incumbency advantage beyond the limited RD conditions where knife-edge electoral shifts create the leverage for causal inference. 1 1

Randomization of treatment assignment is the gold-standard of causal inference in statistics: it is the only sure way to control all confounding. Of the options available when experiments

The role of majorities in public procurement:

The Simons Papers in Security and Development are edited and published at the School for

promote behavior that is consistent with institutional purposes. Evidence for this proposition is based almost entirely on studies that compare the behavior of states that have and have not ratified treaties. This paper evaluates instances in which some member states temporarily experience increased entanglement with an IO because they or their nationals serve in a position of authority. Unlike selection into IOs, selection into positions of authority is often governed by a common, observable, and partially exogenous process. I exploit exogenous exit, random assignment to different term lengths, and competitive elections in three contexts: the International Criminal Court (ICC), the UN Human Rights Commission (UNHRC), and the UN Security Council (UNSC). The evidence implicates that acquiring a position of authority can make states more willing to reject U.S. advances to sign non-surrender agreements, adopt domestic legislation that changes the penal code (ICC case), ratify legally binding treaties (UNHRC case), and contribute to peacekeeping missions (UNSC case). On the other hand, there is no evidence that UN institutions successfully select more cooperative states for positions of authority. Similar research designs can gainfully be employed to identify the causal effects of other forms of institutional participation.

For comments and suggestions on earlier versions of this article, we are grateful to Marko Klašnja, Isabela Mares, and Rocío Titunik, as well as to participants at a Workshop on

In many countries, important features of municipal government (such as the elect-oral system, mayors ’ salaries, and the number of councillors) depend on whether the municipality is above or below arbitrary population thresholds. Several papers have used a regression discontinuity design (RDD) to measure the effects of these threshold-based policies on political and economic outcomes. Using evidence from France, Germany, and Italy, we highlight two common pitfalls that arise in exploit-ing population-based policies (compound treatment and sorting) and we provide guidance for detecting and addressing these pitfalls. Even when these problems are present, population-threshold RDD may be the best available research design for studying the effects of certain policies and political institutions.

resen wom edera comm Pres ntly more likely to sentation than are ayandDuflo, 2004). and scholars have alwomen to office tes onballots holds ps in participation (see Dolan, 2006a and next section for review). Specifically, female candidacies andofficeholdinghave longbeen argued both to increase political empowerment among women in the mass public, leading more women to vote, and to demonstrate to political elites, voters, and women them-selves what women can accomplish in office, leading more women to run. Recent findings from India’s unique policy experiments have crucially bolstered these arguments by 2012 State Politics and Policy Conference for helpful feedback. I further acknowledge the National Science Foundation Graduate Research Fellowship Program for support. I thank Dan Butler for his assistance in collecting the election results and Stacey Chen and Cameron Rotblat for

Many areas of political science focus on causal questions. Evidence from statistical analyses is often used to make the case for causal relationships. While statistical analyses can help establish causal relationships, it can also provide strong evidence of causality where none exists. In this essay, I provide an overview of the statistics of causal inference. Instead of focusing on specific statistical methods, such as matching, I focus more on the assumptions needed to give statistical estimates a causal interpretation. Such assumptions are often referred to as identification assumptions, and these assumptions are critical to any statistical analysis about causal effects. I outline a wide range of identification assumptions and highlight the design-based approach to causal inference. I conclude with an overview of statistical methods that are frequently used for causal inference.