Those interested in political phenomena such as voting have found random utility models, originally developed for decisions such as transportation choice, especially attractive, as the underlying model can yield a statistical model with a few simple, realistic assumptions. Unfortunately, such models have proven difficult to apply to situations with more than two votes and three alternatives or an unknown cutpoint. Additionally, as we show, standard applications of such models to voting, while producing consistent parameter estimates, yield standard errors that are too small and, due to a failure to employ all relevant theoretical information, biased ideal point estimates. We specify a general model applicable to any number of votes and alternatives, with correct standard errors and unbiased ideal point estimates. We apply this model to a number of cases studied by previous scholars involving legislative voting over the minimum wage: (1) when there are two votes and two known cutpoints (K. Krehbiel and D. Rivers, American Journal of Political Science, 1988, 32, 1151–1174); (2) when there are three votes and three known cutpoints (J. Wilkerson, American Journal of Political Science, 1991, 35, 613–623); and (3) when there are three votes but where one cutpoint is unknown given a lack of knowledge about the impact of a policy (J. Wilkerson, American Journal of Political Science, 1991, 35, 613–623) or the possibility of sophisticated voting (C. Volden, Journal of Politics, 1998, 60, 149–173). We show that in various contexts our analysis improves on existing methods, yielding consistent and efficient ideal point estimates and a better-fitting model with improved predictive accuracy.

In public opinion research, response latency is a measure of attitude accessibility, which is the ease or swiftness with which an attitude comes to mind when a respondent is presented with a survey question. Attitude accessibility represents the strength of the association in memory between an attitude object and an evaluation of the object. Recent research shows that attitude accessibility, as measured by response latency, casts light on a wide range of phenomena of public opinion and political behavior. We discuss response latency methodology for survey research and advocate the use of latent response latency timers (which are invisible both to respondents and interviewers) as a low cost, low-maintenance alternative to traditional methods of measuring response latency in public opinion surveys. We show that with appropriate model specification latent response latency timers may provide a suitable alternative to the more complicated and expensive interviewer-activated timers.

In an earlier issue of this journal I brought attention to the fact that estimates of voter turnout in U.S. presidential elections from the National Election Study (NES) series have been increasingly biased. Although researchers had already noted that the NES overestimated turnout, I was concerned with the growing severity of the problem. While admitting that other factors were at work, my explanation centered on the representativeness of surveys, in particular that selection bias in the sample is correlated with the likelihood of voting (Burden 2000). Martinez (2003) and McDonald (2003) offer three possible additions to my argument. First, panel effects are responsible for particularly egregious discrepancies in a few presidential elections, particularly in the 1996 survey. Second, official turnout statistics that rely on the Voting Age Population (VAP) are themselves biased and lack perfect comparability with the NES. Third, the degree of misreporting might also depend on actual voter turnout.

Since Herron and Shotts (2003a; hereinafter HS), Adolph and King (2003; hereinafter AK), and Herron and Shotts (2003b; hereinafter HS2), the four of us have iterated many more times, learned a great deal, and arrived at a consensus on this issue. This paper describes our joint recommendations for how to run second-stage ecological regressions, and provides detailed analyses to back up our claims.

Numerical issues matter in statistical analysis. Small errors occur when numbers are translated from paper and pencil into the binary world of computers. Surprisingly, these errors may be propagated and magnified through binary calculations, eventually producing statistical estimates far from the truth. In this replication and extension article, we look at one method of verifying the accuracy of statistical estimates by running these same data and models on multiple statistical packages. We find that for two published articles, Nagler (1994, American Journal of Political Science 38:230-255) and Alvarez and Brehm (1995, American Journal of Political Science 39:1055-1089), results are dependent on the statistical package used. In the course of our replications, we uncover other pitfalls that may prevent accurate replication, and make recommendations to ensure the ability for future researchers to replicate results.

This article shows that highly correlated measures can produce different results. We identify a democratization model from the literature and test it in more than 120 countries from 1951 to 1992. Then, we check whether the results are robust regarding measures of democracy, time periods, and levels of development. The findings show that measures do matter: Whereas some of the findings are robust, most of them are not. This explains, in part, why the debates on democracy have continued rather than been resolved. More important, it underscores the need for more careful use of measures and further testing to increase confidence in the findings. Scholars in comparative politics are drawn increasingly to large-N statistical analyses, often using data sets collected by others. As in any field, we show how they must be careful in choosing the most appropriate measures for their studies, without assuming that any correlated measure will do.

We develop a model of legislative policymaking in which individual legislators are concerned with both policy and reelection. Legislators' preferences are private information, and they have two means of communicating their preferences to voters. First, they each have a “party label” that credibly identifies an interval within which their ideal points must lie. Second, their roll call votes may convey additional information about their preferences. Each legislator must therefore tailor his or her votes to his or her district in order to forestall a reelection challenge from the opposing party. In equilibrium, nonsincere voting records will occur mostly in moderate districts, where extreme incumbents are vulnerable to challenges from relatively centrist candidates. In those districts, the most extreme legislators may even choose to vote sincerely and retire rather than compile a moderate voting record. Thus, both roll call scores and candidate types will be responsive to district type. An empirical test of shifts in roll call scores of retiring House members in moderate districts confirms these findings.

In a 2000 article in American Journal of Political Science, James Snyder and Tim Groseclose develop and apply an innovative method for detecting and estimating the frequency and magnitude of party influence in congressional roll call voting. This paper presents a framework for assessing the coefficient that the authors interpret as “party influence.” The analysis reveals that, and shows why, the coefficient manifests two troublesome characteristics. The coefficient cannot discriminate between disparate types of party influence because the mapping between types of partison influence and signs of the coefficient is not one-to-one. Similarly, the coefficient has a responsiveness problem because a marginal increase in one party's influence can cause the estimate of the coefficient to increase, decrease, or remain constant. Because the literature on parties in Congress emphasizes majority-party strength, the inability of the coefficient to isolate party-specific effects is a serious drawback in the ongoing hunt for genuine party discipline.

The causal arguments of modern, positive political science often imply complex interactions among multiple explanatory factors. In one example from comparative and international political economy (C&IPE), sharing of monetary-policymaking control between partially autonomous central banks and politically responsive governments yields inflation as a convex combination of the rates that would have held under full-government and full-central-bank policy control. The anti-inflationary effect of central bank independence (autonomy plus conservatism: CBA) therefore depends on all political—economic variables to which central banks and governments respond differently, and, vice versa, CBA mutes the inflation effects of all such factors. Extending that logic of shared policy control to open political economies: insofar as domestic monetary authorities fix exchange rates, they effectively delegate inflation control to foreign (peg-currency) authorities, and policymakers in small and financially exposed economies must match domestic inflation to foreign (global) rates to avoid massive exchange-rate pressures. Thus, analogously, the domestic-inflation effects of fixed exchange rates and of monetary exposure depend on each other and on many other institutional and structural aspects of the domestic and foreign political economies, and, vice versa, the effects of all domestic and foreign political—economic conditions depend on degrees of exchange-rate fixity and financial openness at home and abroad. This article shows how to model such complexly interactive hypotheses empirically compactly and substantively meaningfully, and demonstrates the postwar inflation records of 21 developed democracies to favor such specifications decidedly over standard linear-additive or linear-interactive alternatives. The concluding sections discuss specific results and implications and then suggest several more potential applications of this general approach to further instances of shared policy control and other substantive contexts that induce the multiple, complex interactions characteristic of modern, positive political science in general and C&IPE especially.

This note replicates and extends Paolino's discussion (2001, Political Analysis 9:325–346) on the estimation of models with beta-distributed dependent variables in two ways. First, it introduces an easy-to-use program for estimating the model using Stata. Second, it presents a Bayesian estimator for beta variables based on Paolino's model. Results are compared to those in the original article.

Keith Krehbiel (2003) provides a thoughtful note that explains why the estimation strategy we proposed in Snyder and Groseclose (2000) cannot distinguish between majority-party influence and minority-party influence theories. We agree with Krehbiel's major points. In this note, we explain our disagreement with some of his minor points, and we suggest an additional assumption under which our method can distinguish between majority-party influence and minority-party influence.

Groseclose and Snyder's revised method (2003) for assessing relative party influence deserves high praise for its creativity and explicitness. Two limitations should also be noted, however. First, the method seems likely to overstate majority-party influence. Second, even if taken at face value, the finding of asymmetric party influence is weak when compared with a baseline model.