Empirical analyses in social science frequently confront quantitative data that are clustered or grouped. To account for group-level variation and improve model fit, researchers will commonly specify either a fixed- or random-effects model. But current advice on which approach should be preferred, and under what conditions, remains vague and sometimes contradictory. This study performs a series of Monte Carlo simulations to evaluate the total error due to bias and variance in the inferences of each model, for typical sizes and types of datasets encountered in applied research. The results offer a typology of dataset characteristics to help researchers choose a preferred model.

A major focus of judicial politics research has been the extent to which ideological divergence between the Court and Congress can explain variation in Supreme Court decision making. However, conflicting theoretical and empirical findings have given rise to a significant discrepancy in the scholarship. Building on evidence from interviews with Supreme Court justices and former law clerks, I develop a formal model of judicial‐congressional relations that incorporates judicial preferences for institutional legitimacy and the role of public opinion in congressional hostility towards the Supreme Court. An original dataset identifying all Court‐curbing legislation proposed between 1877 and 2006 is then used to assess the influence of congressional hostility on the Court's use of judicial review. The evidence indicates that public discontent with the Court, as mediated through congressional hostility, creates an incentive for the Court to exercise self‐restraint. When Congress is hostile, the Court uses judicial review to invalidate Acts of Congress less frequently than when Congress is not hostile towards the Court.

Item response theory models for roll-call voting data provide political scientists with parsimonious descriptions of political actors' relative preferences. However, models using only voting data tend to obscure variation in preferences across different issues due to identification and labeling problems that arise in multidimensional scaling models. We propose a new approach to using sources of metadata about votes to estimate the degree to which those votes are about common issues. We demonstrate our approach with votes and opinion texts from the U.S. Supreme Court, using Latent Dirichlet Allocation to discover the extent to which different issues were at stake in different cases and estimating justice preferences within each of those issues. This approach can be applied using a variety of unsupervised and supervised topic models for text, community detection models for networks, or any other tool capable of generating discrete or mixture categorization of subject matter from relevant vote-specific metadata.Word Count: 8,291

Microsimulation methods are used to identify the contribution of tax and benefit reforms to the significant growth in UK income inequality since 1979. The total effect turns out to depend crucially on the counterfactual against which the reforms are assessed: compared with the alternative of pure price‐indexation, the total effect of reform is small; by contrast, compared with a counterfactual in which benefits rose in line with national income (historically the case before 1979), the effect is substantial – approximately half the total rise in income inequality is explained. The impact of reforms on inequality has varied significantly over time: income tax cuts in the late 1970s and late 1980s increased inequality; direct tax rises in the early 1980s and 1990s, together with increases in means‐tested benefits in the late 1990s, reduced it. The robustness of the results to sampling variation and to the measure of inequality used is also investigated.

No abstract

Most U.S. state supreme court justices face elections or reappointment by elected officials, and research suggests that judicial campaigns have come to resemble those for other offices. We develop predictions on how selection systems should affect judicial decisions and test these predictions on an extensive dataset of death penalty decisions by state courts of last resort. Specifically, the data include over 12,000 decisions on over 2000 capital punishment cases decided between 1980 and 2006 in systems with partisan, nonpartisan, or retention elections or with reappointment. As predicted, the findings suggest that judges face the greatest pressure to uphold capital sentences in systems with nonpartisan ballots. Also as predicted, judges respond similarly to public opinion in systems with partisan elections or reappointment. Finally, the results indicate that the plebiscitary influences on judicial behavior emerge only after interest groups began achieving success at targeting justices for their decisions.

O ne-dimensional spatial models have come to inform much theorizing and research on the U.S. Supreme Court. However, we argue that judicial preferences vary considerably across areas of the law, and that limitations in our ability to measure those preferences have constrained the set of questions scholars pursue. We introduce a new approach, which makes use of information about substantive similarity among cases, to estimate judicial preferences that vary across substantive legal issues and over time. We show that a model allowing preferences to vary over substantive issues as well as over time is a significantly better predictor of judicial behavior than one that only allows preferences to vary over time. We find that judicial preferences are not reducible to simple left-right ideology and, as a consequence, there is substantial variation in the identity of the median justice across areas of the law during all periods of the modern court. These results suggest a need to reconsider empirical and theoretical research that hinges on the existence of a single pivotal median justice.

It is well known that the public often relies on cues or heuristics when forming opinions. At the same time, leading theories of opinion formation about the Supreme Court see such support as relatively fixed. Using a series of survey experiments, we find source cues significantly influence the public's support for the Court, including the extent to which individuals believe the Court should be independent from the elected branches. Specifically, we find partisan source cues play a significant role in shaping public opinion regarding life tenure for the justices and the extent to which the Court should have the final say in constitutional matters-individuals are less likely to support court-curbing measures when informed that elites from the opposite party have proposed them than when such measures are endorsed by either a neutral source or members of their own party. We also find a strong connection between specific support for particular decisions and the degree