Measuring Bias and Uncertainty in Ideal Point Estimates via the Parametric Bootstrap

Keith T. Poole Center for Advanced Study in the Behavioral Sciences and Department of Political Science, University of Houston, Houston, TX 77204-3011 e-mail: kpoole{at}uh.edu Over the last 15 years a large amount of scholarship in legislativepolitics has used NOMINATE or other similar methods to constructmeasures of legislators' ideological locations. These measuresare then used in subsequent analyses. Recent work in politicalmethodology has focused on the pitfalls of using such estimatesas variables in subsequent analysis without explicitly accountingfor their uncertainty and possible bias ( Herron and Shotts2003, Political Analysis 11:44–64). This presents a problemfor those employing NOMINATE scores because estimates of theirunconditional sampling uncertainty or bias have until now beenunavailable. In this paper, we present a method of forming unconditionalstandard error estimates and bias estimates for NOMINATE scoresusing the parametric bootstrap. Standard errors are estimatedfor the 90th U.S. Senate in two dimensions. Standard errorsof first–dimension placements are in the 0.03 to 0.08range. The results are compared with those obtained using theMarkov chain Monte Carlo estimator of Clinton et al. (2002,Stanford University Working Paper). We also show how the bootstrapcan be used to construct standard errors and confidence intervalsfor auxiliary quantities of interest such as ranks and the locationof the median senator.     

Selection Bias and Continuous-Time Duration Models: Consequences and a Proposed Solution

This article analyzes the consequences of nonrandom sample selection for continuous-time duration analyses and develops a new estimator to correct for it when necessary. We conduct a series of Monte Carlo analyses that estimate common duration models as well as our proposed duration model with selection. These simulations show that ignoring sample selection issues can lead to biased parameter estimates, including the appearance of (nonexistent) duration dependence. In addition, our proposed estimator is found to be superior in root mean-square error terms when nontrivial amounts of selection are present. Finally, we provide an empirical application of our method by studying whether self-selectivity is a problem for studies of leaders' survival during and following militarized conflicts.     

Gender Bias and Selection Bias in House Elections

We demonstrate that female incumbents areof higher average candidate quality than maleincumbents. This quality difference is the result ofbarriers to entry faced by potential femalecandidates, although the observed effects of thisquality differential on vote share are partiallymasked by the fact that female incumbents are alsomore likely to be opposed or to be opposed by highquality challengers. Using data from House electionsfor 1984–1992, we estimate that the gender-baseddifferential in candidate quality yields an extra sixpercentage points of vote share for femaleincumbents.     

Binary Games as Models of Public Goods Provision*

Problems of public goods provision are categorized according to attributes of the good to be provided, and properties of the group of potential beneficiaries. It is argued that not all such problems are Prisoner's Dilemmas. Other games of interest include Chicken, the Volunteer's Dilemma, a variant of the Assurance Came and several others. Which particular model is relevant in a given context depends critically on the specific characteristics of the situation under consideration.     

Ignorance Is Bias: The Effect of Latino Losers on Models of Latino Representation

Nearly every aggregate study of minority legislative representation has observed outcomes of elections (officeholders), rather than the supply of minority candidates. Because of this, scholars have left a large amount of important data, the election losers, out of their models of minority representation. The evidence presented in this article demonstrates that voters in the United States cannot choose minority officeholders because there are rarely minority candidates on the ballot. I use state legislative candidate data from Carsey et al. ( 2008 ) and Klarner et al. ( 2012 ) to test models of Latino representation that correct for first‐stage selection bias. Once candidate self‐selection is taken into account, the probability of electing a Latino increases enormously. I then use data from 2010 to make out‐of‐sample predictions, which clearly favor the conditional model. Thus, our current understanding of Latino representation is significantly biased by ignoring the first stage of an election, a candidate's decision to run.     

Expertise and Bias in Political Communication Networks

Citizens minimize information costs by obtaining political guidance from others who have already assumed the costs of acquiring and processing political information. A problem occurs because ideal informants, typically characterized by the joint presence of political expertise and shared viewpoints, are frequently unavailable or rare within the groups where individuals are located. Hence, individuals must often look beyond their own group boundaries to find such informants. The problem is that obtaining information from individuals located beyond their own groups produces additional costs. Moreover, the availability of ideal informants varies across groups and settings, with the potential to produce (1) context‐dependent patterns of informant centrality, which in turn generate (2) varying levels of polarization among groups and (3) biases in favor of some groups at the expense of others. The article's analysis is based on a series of small‐group experiments, with aggregate implications addressed using a simple agent‐based model.     

Accuracy and Bias in Perceptions of Political Knowledge

Learning through social communication is promoted when citizens are able to identify which of their associates is likely to possess the necessary political information. This paper examines the factors that influence individuals’ evaluations of political expertise. Actual political expertise plays a large role in perceived expertise, but mistakes are made. These are largely the result of assuming that those engaged in politics must also be knowledgeable about politics. This paper uses the 1996 Indianapolis-St. Louis Study and the 2000 National Election Study to identify factors that bias levels of perceived political knowledge. The paper concludes by demonstrating that perceived expertise plays a larger role than actual expertise in the social influence process.     

Turnout and Representation Bias in Post-communist Europe

Electoral participation has been declining in post-Soviet Europe as in almost all of the established democracies. Patterns of electoral abstention in Russia, Belarus and Ukraine reflect those in other countries, but show particularly strong effects for older age. Not only do older electors vote more often, they also have distinctive views on matters of public policy, particularly on the economy but also on the Soviet system, strong leadership and hypothetical membership of the European Union. These differences are diminished but nonetheless generally remain statistically significant even when socio-economic controls are introduced. These differences may be seen as a 'representation bias' that advantages particular sections of the electorate and the views with which they are associated. The particular forms that are taken by this bias in post-communist societies may be transitory, but here as elsewhere lower levels of turnout will continue to impart a significant bias to the extent to which some views rather than others are articulated within the political process.     

Bias in OMB's economic forecasts and budget proposals

The Office of Management and Budget (OMB) annually submits the President's budget for the U.S. government to the Congress. The economic forecasts and revenue and outlay proposals contained in the budget have been criticized as biased, especially during the 1980s. Tests for bias in one year ahead proposals for the 1963–89 period show no bias in economic forecasts and revenue estimates, but substantial bias in outlay proposals. Most outlay proposals are consistently less than actual outcomes, accounting for underprediction of the Federal deficit. OMB outlay proposals appear to be influenced by politics. Republican administrations show more significant proposal biases, with defense proposals higher and domestic outlay proposals lower than outcomes. The Office of Management and Budget consistently understates the deficits by resorting to the most optimistic economic assumptions it can credibly — and now sometimes even incredibly — employ. — Senator James SasserWall Street Journal, 25 January 1990     

Using Auxiliary Data to Estimate Selection Bias Models, with an Application to Interest Group Use of the Direct Initiative Process

Recent work in survey research has made progress in estimatingmodels involving selection bias in a particularly difficultcircumstance—all nonrespondents are unit nonresponders,meaning that no data are available for them. These models arereasonably successful in circumstances where the dependent variableof interest is continuous, but they are less practical empiricallywhen it is latent and only discrete outcomes or choices areobserved. I develop a method in this article to estimate thesemodels that is much more practical in terms of estimation. Themodel uses a small amount of auxiliary information to estimatethe selection equation parameters, which are then held fixedwhile estimating the equation of interest parameters in a maximum-likelihoodsetting. After presenting Monte Carlo analyses to support themodel, I apply the technique to a substantive problem: Whichinterest groups are likely to to be involved in support of potentialinitiatives to achieve their policy goals?     

Bias in Conditional and Unconditional Fixed Effects Logit Estimation

Fixed-effects logit models can be useful in panel data analysis,when N units have been observed for T time periods. There aretwo main estimators for such models: unconditional maximum likelihoodand conditional maximum likelihood. Judged on asymptotic properties,the conditional estimator is superior. However, the unconditionalestimator holds several practical advantages, and thereforeI sought to determine whether its use could be justified onthe basis of finite-sample properties. In a series of MonteCarlo experiments for T < 20, I found a negligible amountof bias in both estimators when T 16, suggesting that a researchercan safely use either estimator under such conditions. WhenT < 16, the conditional estimator continued to have a verysmall amount of bias, but the unconditional estimator developedmore bias as T decreased.     

Analysing Welfare State Variations: The Merits and Limitations of Models Based on the Residual–Institutional Distinction*

A key distinction in models of social policy and typologies of welfare provision has been a differentiation between residual and institutional types. However, despite the gradual elaboration of models based on this distinction, there has been little effort to apply them in empirical comparative analysis or to assess the strengths and weaknesses of the models. This article initially examines two main approaches in identifying types of welfare states and major welfare state variations. It subsequently applies several dimensions of variation posited by the residual and institutional models in a comparison of the United States, the United Kingdom, the Netherlands and Sweden. This comparison forms the point of departure for evaluating the merits and limitations of models based on the residual-institutional distinction.     

Parties in Elections, Parties in Government, and Partisan Bias

Adam Meirowitz Department of Politics, Princeton University, Princeton, NJ 08544 e-mail: ameirowi{at}princeton.edu Thomas Romer Department of Politics and Woodrow Wilson School of Public and International Affairs, Princeton University, Princeton, NJ 08544 e-mail: romer{at}princeton.edu Political parties are active when citizens choose among candidatesin elections and when winning candidates choose among policyalternatives in government. But the inextricably linked institutions,incentives, and behavior that determine these multistage choicesare substantively complex and analytically unwieldy, particularlyif modeled explicitly and considered in total, from citizenpreferences through government outcomes. To strike a balancebetween complexity and tractability, we modify standard spatialmodels of electoral competition and governmental policy-makingto study how components of partisanship—such as candidateplatform separation in elections, party ID-based voting, nationalpartisan tides, and party-disciplined behavior in the legislature—arerelated to policy outcomes. We define partisan bias as the distancebetween the following two points in a conventional choice space:the ideal point of the median voter in the median legislativedistrict and the policy outcome selected by the elected legislature.The study reveals that none of the party-in-electorate conditionsis capable of producing partisan bias independently. Specifiedcombinations of conditions, however, can significantly increasethe bias and/or the variance of policy outcomes, sometimes insubtle ways.     

Bias in the formulation of local government budget problems

Government budgets are premised on forecasts of revenues and expenditures. These forecasts are subject to both stochastic error and strategic manipulation. Circumstantial evidence in the budgeting literature and in the popular media suggest that government officials routinely bias the forecasts underlying budgets. The research reported here asked three primary questions: To what extent are budget forecasts systematically biased? Why? (Are fiscal and electoral variables systematically related to the magnitude and direction of the biases?) What political and ethical difference do the biases make? From the literature and an analysis of the incentives facing politicians and bureaucrats, we developed hypotheses about budget biases. These hypotheses were tested using time series data for the City of Pittsburgh, Pennsylvania (1941–1983); the City of San Diego, California (1950–1982); and the Pittsburgh (Pennsylvania) School District (1946–1983). In these locales over the periods examined, budgets were systematically pessimistic; revenues were underestimated and expenditures were overestimated. The fiscal and electoral factors hypothesized to account for this pessimism are, however, very mixed in their ability to explain the biases.     

Interest groups in the media: Bias and diversity over time

A prominent presence in the news media is important for interest groups. This article investigates the development in the diversity of interest group media attention over time. The analysis draws on a dataset of 19,000 group appearances in the Danish news media in the period 1984–2003. It demonstrates how diversity has risen continually over time, leading to a media agenda less dominated by labour and business and more by public interest groups and sectional groups. This development is related to the increasing political importance of the news media and the decline in group integration in public decision‐making processes. The article also shows how the development of group appearances is closely related to changes in media attention towards different policy areas.     

Bias in Conditional and Unconditional Fixed Effects Logit Estimation: A Correction

In a recent paper published in this journal, Katz (2001) comparesthe bias in conditional and unconditional fixed effects logitestimation using Monte Carlo Simulation. This note shows thatwhile Katz's (2001) specification has "wrong" fixed effects(in the sense that the fixed effects are the same for all individuals),his conclusions still hold if I correct his specification (sothat the fixed effects do differ over individuals). This notealso illustrates the danger, when using logit, of includingdummies when no fixed effects are present.     

Design Thinking in Policymaking Processes: Opportunities and Challenges

Design thinking has the potential to improve problem definition and mechanism design in policymaking processes. By promoting greater understanding of how citizens experience government services, design thinking can support public managers who desire to enhance public value. In Australia, as elsewhere, design thinking currently remains separated from mainstream policymaking efforts. This article clarifies the essence of design thinking and its applicability to policy development. Five design thinking strategies are discussed, all of which have lengthy histories as social science methodologies. They are (1) environmental scanning, (2) participant observation, (3) open‐to‐learning conversations, (4) mapping, and (5) sensemaking. Recent examples from Australia and New Zealand are used to illustrate how these strategies have been incorporated into policymaking efforts. The article concludes by considering how design thinking might be more broadly applied in policymaking, and the training and resourcing requirements that would entail.     

Comparing Implementation Processes in Sweden and the United States

This article attempts a systematic comparison of how policy is implemented in Sweden and the United States. We suggest that differences might be explained by examining policymaking processes, policy content, and governmental structure. The comparison reveals differences in 'incentive structure'. Both Sweden and the U.S. have responded to the complexity and uncertainty that accompanies increased demands on central government by adopting 'softer' mechanisms for implementation. The effect of these shifts in both systems is to devolve important decisions to lower levels of government. In Sweden, this devolution is still accompanied by strong incentives for national policymakers to undertake implementation, although the incentives may not be as strong as they were in the past. In the U.S., it is accompanied by a radical weakening of incentives to carry out implementation and increasing instability in policy content and relations among key actors.     

Participation Bias, Durable Opinion Shifts and Sabotage through Withdrawal in Citizens' Juries

We analyse a citizens' jury experiment held in Dublin on the 'hot-button' issue of waste incineration. The jury was a random sample drawn from the 2002 Irish Election Study (IES), and we compare jurors to non-jurors in the 2002 IES, and in 2003 and 2004 panel studies. Large opinion shifts were observed in a representative jury; these shifts remained observable nine months later. However, as a direct consequence of our choice of a contested issue, the jurors' verdict was framed by the unwillingness of key stakeholders to participate. Stakeholders who expect to be on the 'losing' side, including public representatives and officials, may obstruct and delegitimise citizens' juries by withholding participation. This informal gatekeeping power undermines the potential of citizens' juries to empower citizens.