A common causal identification strategy in political science is selection on observables. This strategy assumes one observes a set of covariates that is, after statistical adjustment, sufficient to make treatment status as-if random. Under adjustment methods such as matching or inverse probability weighting, coefficients for control variables are treated as nuisance parameters and are not directly estimated. This is in direct contrast to regression approaches where estimated parameters are obtained for all covariates. Analysts often find it tempting to give a causal interpretation to all the parameters in such regression models—indeed, such interpretations are often central to the proposed research design. In this paper, we ask when we can justify interpreting two or more coefficients in a regression model as causal parameters. We demonstrate that analysts must appeal to causal identification assumptions to give estimates causal interpretations. Under selection on observables, this task is complicated by the fact that more than one causal effect might be identified. We show how causal graphs provide a framework for clearly delineating which effects are presumed to be identified and thus merit a causal interpretation, and which are not. We conclude with a set of recommendations for how researchers should interpret estimates from regression models when causal inference is the goal.

The level of legislative turnover in a polity can have significant political consequences. Low turnover may increase the number of legislators who are out of touch with constituents, while high turnover can limit a legislature’s ability to fulfill its duties. Focusing on separation of powers arrangements, a factor overlooked by previous studies, we identify institutional conditions that affect turnover. When the executive and legislature are equally responsible for budgetary outcomes, we argue, this creates ambiguous contexts, leaving voters more likely to re-elect incumbents, thereby lowering turnover. We test our predictions using US state-level data.

Existing arguments across political science posit that parties in government use domestic and international institutions to lock in their own policy preferences by tying the hands of successors. I demonstrate that these arguments contrast with the assumption of office-seeking parties and therefore portray an incomplete picture of the incentives of governments. The paper emphasizes the trade-off between implementing policy preferences, on the one hand, and exploiting partisan differences for electoral success, on the other hand: locking in a policy takes an issue off the table, but it also undermines a party’s ability to leverage differences to the opposition in elections. Because office-seeking parties need to take into account these electoral consequences, they have a disincentive to tie their successors’ hands. I advance this argument in the context of the establishment of independent central banks, provide empirical evidence, and suggest implications for the literature on international institutions.

While party splits are a relatively frequent phenomenon in many new and established democracies, the systematic empirical research on electoral competition after schisms is limited. The analysis of more than 200 splits across 25 European countries in the post-war period addresses this gap in the research. The study shows that the vote shares of rump and splinter parties in the first election after fission are related to their membership strength and the share of splinter legislators. This relationship is present in both Western Europe and Central and Eastern Europe. Additionally, economic growth affects the support of rump parties that hold government office while party system fragmentation and party funding regulations correlate with the electoral performance of splinter parties.

Local politicians are often expected to mobilize voters on behalf of copartisan candidates for national office. Yet this requirement is difficult to enforce because the effort of local politicians cannot be easily monitored and the promise of rewards in exchange for help is not fully credible. Using a formal model, we show that the incentives of local politicians to mobilize voters on behalf of their party depend on the proportion of copartisan officials in a district. Having many copartisan officials means that the party is more likely to capture the district, but the effort of each local politician is less likely either to be noticed by higher-level officials or to make a difference on the election outcome, thus discouraging lower-level officials from exerting effort. We validate these claims with data from federal elections in Mexico between 2000 and 2012. In line with the argument, the results show that political parties fail to draw great mobilization advantages from simultaneously controlling multiple offices.

Although politicians’ personal attributes are an important component of elections and representation, few studies have rigorously investigated which attributes are most relevant in shaping voters’ preferences for politicians, or whether these preferences vary across different electoral system contexts. We investigate these questions with a conjoint survey experiment using the case of Japan’s mixed-member bicameral system. We find that the attributes preferred by voters are not entirely consistent with the observed attributes of actual politicians. Moreover, voters’ preferences do not vary when asked to consider representation under different electoral system contexts, whereas the observed attributes of politicians do vary across these contexts. These findings point to the role of factors beyond voters’ sincere preferences, such as parties’ recruitment strategies, the effect of electoral rules on the salience of the personal vote, and the availability of different types of politicians, in determining the nature of representation.

Most agree that models of binary time-series-cross-sectional data in political science often possess unobserved unit-level heterogeneity. Despite this, there is no clear consensus on how best to account for these potential unit effects, with many of the issues confronted seemingly misunderstood. For example, one oft-discussed concern with rare events data is the elimination of no-event units from the sample when estimating fixed effects models. Many argue that this is a reason to eschew fixed effects in favor of pooled or random effects models. We revisit this issue and clarify that the main concern with fixed effects models of rare events data is not inaccurate or inefficient coefficient estimation, but instead biased marginal effects. In short, only evaluating event-experiencing units gives an inaccurate estimate of the baseline risk, yielding inaccurate (often inflated) estimates of predictor effects. As a solution, we propose a penalized maximum likelihood fixed effects (PML-FE) estimator, which retains the complete sample by providing finite estimates of the fixed effects for each unit. We explore the small sample performance of PML-FE versus common alternatives via Monte Carlo simulations, evaluating the accuracy of both parameter and effects estimates. Finally, we illustrate our method with a model of civil war onset.

Scholars often use voting data to estimate central bankers’ policy preferences but consensus voting is commonplace. To get around this, we combine topic-based text analysis and scaling methods to generate theoretically motivated comparative measures of central bank preferences on the US Federal Open Market Committee (FOMC) leading up to the financial crisis in a way that does not depend on voting behavior. We apply these measures to a number of applications in the literature. For example, we find that FOMC members that are Federal Reserve Bank Presidents from districts experiencing higher unemployment are also more likely to emphasize unemployment in their speech. We also confirm that committee members on schedule to vote are more likely to express consensus opinion than their off schedule voting counterparts.

Measuring networks in the field—usually by asking individuals systematically about their networks–entails complex design choices, with large consequences for the resulting data. Because observations in a network are interconnected, well-established practices from non-network survey settings can lead researchers astray. Despite the increasing focus on networks in political science, little guidance is available for researchers facing high-stakes decisions when designing a study to elicit networks. This paper serves as a practical guide. It offers a simple framework for constructing a network theory, illuminates tradeoffs like measuring more nodes versus more ties per node or asking for names versus selections from a list, and proposes a new technique for cleaning relational data.

Researchers in political science often estimate linear models of continuous outcomes using least squares. While it is well known that least-squares estimates are sensitive to single, unusual data points, this knowledge has not led to careful practices when using least-squares estimators. Using statistical theory and Monte Carlo simulations, we highlight the importance of using more robust estimators along with variable transformations. We also discuss several approaches to detect, summarize, and communicate the influence of particular data points.

This article evaluates the reliability of sensitivity tests. Using Monte Carlo methods we show that, first, the definition of robustness exerts a large influence on the robustness of variables. Second and more importantly, our results also demonstrate that inferences based on sensitivity tests are most likely to be valid if determinants and confounders are almost uncorrelated and if the variables included in the true model exert a strong influence on outcomes. Third, no definition of robustness reliably avoids both false positives and false negatives. We find that for a wide variety of data-generating processes, rarely used definitions of robustness perform better than the frequently used model averaging rule suggested by Sala-i-Martin. Fourth, our results also suggest that Leamer’s extreme bounds analysis and Bayesian model averaging are extremely unlikely to generate false positives. Thus, if based on these inferential criteria a variable is robust, it is almost certain to belong into the empirical model. Fifth and finally, we also show that researchers should avoid drawing inferences based on lack of robustness.

To understand the relationship between place and politics, we must measure both political attitudes and the ways in which place is represented in the minds of individuals. In this paper, we assess a new measure of mental representation of geography, in which survey respondents draw their own local communities on maps and describe them. We use a panel study in Canada to present evidence that these maps are both valid and reliable measures of a personally relevant geographic area, laying the measurement groundwork for the growing number of studies using this technology. We hope to set efforts to measure “place” for the study of context and politics on firmer footing. Our validity assessments show that individuals are thinking about people and places with which they have regular contact when asked to draw their communities. Our reliability assessments show that people can draw more or less the same map twice, even when the exercise is repeated months later. Finally, we provide evidence that the concept of community is a tangible consideration in the minds of ordinary citizens and is not simply a normative aspiration or motivation.

Existing research shows that survey respondents are sensitive to the source of information about political corruption and respond more strongly to information from more credible sources. This behavior occurs more frequently among the politically sophisticated. In a nation-wide survey in Argentina, we successfully replicate results originally found in a study in Brazil. In addition, we examine whether citizens process information about corruption differently depending on their partisan identities. At odds with our initial expectations, we find that copartisans, opposition partisans, and other/non-partisans distinguish between information sources in very similar ways. These results suggest that even though partisanship affects baseline assessments of political candidates, citizens of all types are sensitive to the credibility of information they receive about political corruption.

Until recently, researchers who wanted to examine the determinants of state respect for most specific negative rights (i.e., physical integrity and empowerment rights) needed to rely on data from the CIRI or the Political Terror Scale (PTS). The new Varieties of Democracy (V-Dem) dataset offers scholars a potential alternative to the individual human rights variables from CIRI. We analyze a set of key Cingranelli–Richards (CIRI) Human Rights Data Project and V-Dem negative rights indicators, finding unusual and unexpectedly large patterns of disagreement between the two sets. First, we discuss the new V-Dem dataset by comparing it to the disaggregated CIRI indicators, discussing the history of each project, and describing its empirical domain. Second, we identify a set of disaggregated human rights measures that are similar across the two datasets and discuss each project’s measurement approach. Third, we examine how these measures compare to each other empirically, showing that they diverge considerably across both time and space. These findings point to several important directions for future work, such as how conceptual approaches and measurement strategies affect rights scores. For the time being, our findings suggest that researchers should think carefully about using the measures as substitutes.

Survey research shows that voters know very little about politics. Left unexamined is whether political knowledge is a distinct type of knowledge when compared to other subjects. We seek to establish a baseline by putting political knowledge into a broader context. We show that our respondents’ knowledge about politics is similar in construction to their knowledge about other subjects, such as shopping, sports, popular culture, geography, economics, and the rules of the road. We conclude that knowledge of politics largely resides on the same dimension as other knowledge topics, implying that knowledge of politics is not a unique construct.