Working Papers
2002
36 records found
State and American Indian Negotiation of Gaming Compacts: An Event Count Analysis
Boehmke, Frederick, Witner, Richard
Submitted: 2002-01-23
Keywords: american indian, gaming, event count, policy adoption
Abstract: (click to show/hide) There has been a proliferation of casino-style Indian gaming in the years since the passage of the Indian Gaming Regulatory Act in 1988. Yet little is known about the factors that influence state and Indian nations’ decisions to enter into gaming compacts. In this paper we seek to achieve two objectives. First, we seek to understand the expansion of Indian-state gaming compacts by studying how characteristics of states and Indian nations, along with spatial and temporal diffusion, affect the number of compacts negotiated. Most importantly, we focus on Indian nation’s relationships with the states; their political influence with respect to the state and the contact they have with state government. Second, we introduce an empirical model new to the study of state politics by modeling the compacting process between Indian nations and states as an event count process. The event count model allows us to explain why some states have more Indian gaming than others and how the compacting process has evolved over time.
Boehmke, Frederick, Witner, Richard
Submitted: 2002-01-23
Keywords: american indian, gaming, event count, policy adoption
Abstract: (click to show/hide) There has been a proliferation of casino-style Indian gaming in the years since the passage of the Indian Gaming Regulatory Act in 1988. Yet little is known about the factors that influence state and Indian nations’ decisions to enter into gaming compacts. In this paper we seek to achieve two objectives. First, we seek to understand the expansion of Indian-state gaming compacts by studying how characteristics of states and Indian nations, along with spatial and temporal diffusion, affect the number of compacts negotiated. Most importantly, we focus on Indian nation’s relationships with the states; their political influence with respect to the state and the contact they have with state government. Second, we introduce an empirical model new to the study of state politics by modeling the compacting process between Indian nations and states as an event count process. The event count model allows us to explain why some states have more Indian gaming than others and how the compacting process has evolved over time.
Did Illegally Counted Overseas Absentee Ballots Decide the 2000 U.S. Presidential Election?
Imai, Kosuke, King, Gary
Submitted: 2002-02-13
Keywords: 2000 U.S. Presidential Election, Ecological Inference, Bayesian Model Averaging
Abstract: (click to show/hide) Although not widely known until much later, Al Gore received 202 more votes than George W. Bush on election day in Florida. George W. Bush is president because he overcame his election day deficit with overseas absentee ballots that arrived and were counted after election day. In the final official tally, Bush received 537 more votes than Gore. These numbers are taken from the official results released by the Florida Secretary of State's office and so do not reflect overvotes, undervotes, unsuccessful litigation, butterfly ballot problems, recounts that might have been allowed but were not, or any other hypothetical divergence between voter preferences and counted votes. After the election, the New York Times conducted a six month long investigation and found that 680 of the overseas absentee ballots were illegally counted, and no partisan, pundit, or academic has publicly disagreed with their assessment. In this paper, we describe the statistical procedures we developed and implemented for the Times to ascertain whether disqualifying these 680 ballots would have changed the outcome of the election. The methods involve adding formal Bayesian model averaging procedures to King's (1997) ecological inference model. Formal Bayesian model averaging has not been used in political science but is especially useful when substantive conclusions depend heavily on apparently minor but indefensible model choices, when model generalization is not feasible, and when potential critics are more partisan than academic. We show how we derived the results for the Times so that other scholars can use these methods to make ecological inferences for other purposes. We also present a variety of new empirical results that delineate the precise conditions under which Al Gore would have been elected president, and offer new evidence of the striking effectiveness of the Republican effort to convince local election officials to count invalid ballots in Bush counties and not count them in Gore counties.
Imai, Kosuke, King, Gary
Submitted: 2002-02-13
Keywords: 2000 U.S. Presidential Election, Ecological Inference, Bayesian Model Averaging
Abstract: (click to show/hide) Although not widely known until much later, Al Gore received 202 more votes than George W. Bush on election day in Florida. George W. Bush is president because he overcame his election day deficit with overseas absentee ballots that arrived and were counted after election day. In the final official tally, Bush received 537 more votes than Gore. These numbers are taken from the official results released by the Florida Secretary of State's office and so do not reflect overvotes, undervotes, unsuccessful litigation, butterfly ballot problems, recounts that might have been allowed but were not, or any other hypothetical divergence between voter preferences and counted votes. After the election, the New York Times conducted a six month long investigation and found that 680 of the overseas absentee ballots were illegally counted, and no partisan, pundit, or academic has publicly disagreed with their assessment. In this paper, we describe the statistical procedures we developed and implemented for the Times to ascertain whether disqualifying these 680 ballots would have changed the outcome of the election. The methods involve adding formal Bayesian model averaging procedures to King's (1997) ecological inference model. Formal Bayesian model averaging has not been used in political science but is especially useful when substantive conclusions depend heavily on apparently minor but indefensible model choices, when model generalization is not feasible, and when potential critics are more partisan than academic. We show how we derived the results for the Times so that other scholars can use these methods to make ecological inferences for other purposes. We also present a variety of new empirical results that delineate the precise conditions under which Al Gore would have been elected president, and offer new evidence of the striking effectiveness of the Republican effort to convince local election officials to count invalid ballots in Bush counties and not count them in Gore counties.
Armed Conflict as a Public Health Problem
Murray, Christopher J. L., King, Gary, Lopez, Alan D., Tomijima, Niels, Krug, Etienne
Submitted: 2002-02-25
Keywords: International Conflict Data, public health, war, epidemiology
Abstract: (click to show/hide) Armed conflict is a major cause of injury and death worldwide, but we need much better methods of quantification before we can accurately assess its effect. Armed conflict between warring states and groups within states have been major causes of ill health and mortality for most of human history. Conflict obviously causes deaths and injuries on the battlefield, but also health consequences from the displacement of populations, the breakdown of health and social services, and the heightened risk of disease transmission. Despite the size of the health consequences, military conflict has not received the same attention from public health research and policy as many other causes of illness and death. In contrast, political scientists have long studied the causes of war but have primarily been interested in the decision of elite groups to go to war, not in human death and misery. We review the limited knowledge on the health consequences of conflict, suggest ways to improve measurement, and discuss the potential for risk assessment and for preventing and ameliorating the consequences of conflict.
Murray, Christopher J. L., King, Gary, Lopez, Alan D., Tomijima, Niels, Krug, Etienne
Submitted: 2002-02-25
Keywords: International Conflict Data, public health, war, epidemiology
Abstract: (click to show/hide) Armed conflict is a major cause of injury and death worldwide, but we need much better methods of quantification before we can accurately assess its effect. Armed conflict between warring states and groups within states have been major causes of ill health and mortality for most of human history. Conflict obviously causes deaths and injuries on the battlefield, but also health consequences from the displacement of populations, the breakdown of health and social services, and the heightened risk of disease transmission. Despite the size of the health consequences, military conflict has not received the same attention from public health research and policy as many other causes of illness and death. In contrast, political scientists have long studied the causes of war but have primarily been interested in the decision of elite groups to go to war, not in human death and misery. We review the limited knowledge on the health consequences of conflict, suggest ways to improve measurement, and discuss the potential for risk assessment and for preventing and ameliorating the consequences of conflict.
An Automated Information Extraction Tool For International Conflict Data with Performance as Good as Human Coders: A Rare Events Evaluation Design
King, Gary, Lowe, Will
Submitted: 2002-02-25
Keywords: rare events, Bayes, international conflict data, computer
Abstract: (click to show/hide) Despite widespread recognition that aggregated summary statistics on international conflict and cooperation miss most of the complex interactions among nations, the vast majority of scholars continue to employ annual, quarterly, or occasionally monthly observations. Daily events data, coded from some of the huge volume of news stories produced by journalists, have not been used much for the last two decades. We offer some reason to change this practice, which we feel should lead to considerably increased use of these data. We address advances in event categorization schemes and software programs that automatically produce data by ``reading'' news stories without human coders. We design a method that makes it feasible for the first time to evaluate these programs when they are applied in areas with the particular characteristics of international conflict and cooperation data, namely event categories with highly unequal prevalences, and where rare events (such as highly conflictual actions) are of special interest. We use this rare events design to evaluate one existing program, and find it to be as good as trained human coders, but obviously far less expensive to use. For large scale data collections, the program dominates human coding. Our new evaluative method should be of use in international relations, as well as more generally in the field of computational linguistics, for evaluating other automated information extraction tools. We believe that the data created by programs similar to the one we evaluated should see dramatically increased use in international relations research. To facilitate this process, we will be releasing with this article data on 4.3 million international events, covering the entire world for the last decade.
King, Gary, Lowe, Will
Submitted: 2002-02-25
Keywords: rare events, Bayes, international conflict data, computer
Abstract: (click to show/hide) Despite widespread recognition that aggregated summary statistics on international conflict and cooperation miss most of the complex interactions among nations, the vast majority of scholars continue to employ annual, quarterly, or occasionally monthly observations. Daily events data, coded from some of the huge volume of news stories produced by journalists, have not been used much for the last two decades. We offer some reason to change this practice, which we feel should lead to considerably increased use of these data. We address advances in event categorization schemes and software programs that automatically produce data by ``reading'' news stories without human coders. We design a method that makes it feasible for the first time to evaluate these programs when they are applied in areas with the particular characteristics of international conflict and cooperation data, namely event categories with highly unequal prevalences, and where rare events (such as highly conflictual actions) are of special interest. We use this rare events design to evaluate one existing program, and find it to be as good as trained human coders, but obviously far less expensive to use. For large scale data collections, the program dominates human coding. Our new evaluative method should be of use in international relations, as well as more generally in the field of computational linguistics, for evaluating other automated information extraction tools. We believe that the data created by programs similar to the one we evaluated should see dramatically increased use in international relations research. To facilitate this process, we will be releasing with this article data on 4.3 million international events, covering the entire world for the last decade.
Optimal Campaigning in Presidential Elections: The Probability of Being Florida
Stromberg, David
Submitted: 2002-03-07
Keywords: elections, political campaigns, public expenditures
Abstract: (click to show/hide) This paper delivers a precise recommendation for how presidential candidates should allocate their resources to maximize the probability of gaining a majority in the Electoral College. A two-candidate, probabilistic-voting model reveals that more resources should be devoted to states which are likely to be decisive in the electoral college and, at the same time, have very close state elections. The optimal strategies are empirically estimated using state-level opinion-polls available in September of the election year. The model's recommended campaign strategies closely resemble those used in actual campaigns. The paper also analyses how the allocation of resources would change under the alternative electoral rule of a direct national vote for president.
Stromberg, David
Submitted: 2002-03-07
Keywords: elections, political campaigns, public expenditures
Abstract: (click to show/hide) This paper delivers a precise recommendation for how presidential candidates should allocate their resources to maximize the probability of gaining a majority in the Electoral College. A two-candidate, probabilistic-voting model reveals that more resources should be devoted to states which are likely to be decisive in the electoral college and, at the same time, have very close state elections. The optimal strategies are empirically estimated using state-level opinion-polls available in September of the election year. The model's recommended campaign strategies closely resemble those used in actual campaigns. The paper also analyses how the allocation of resources would change under the alternative electoral rule of a direct national vote for president.
Turnout Effects on the Composition of the Electorate: A Multinomial Logit Simulation of the 2000 Presidential Election
Martinez, Michael
Submitted: 2002-03-18
Keywords: turnout, multinomial logit, simulation
Abstract: (click to show/hide) Conventional wisdom among pundits and some scholars posits that higher turnout should benefit liberal parties, since lower socioeconomic classes comprise a disproportionate share of the nonvoting population. Empirical tests of this prediction across elections have produced a wide variety of results, ranging from support for the conventional wisdom to suggestions that Republicans benefit from higher turnout to null findings. In this paper, we provide a simulation of the possible impact of increasing or decreasing turnout in a single election. Using data from the 2000 American National Election Study, we find that Gore would have benefitted slightly from higher turnout and would have been harmed slightly by lower turnout, but the overall magnitude of the effects of turnout on Gore's share of the two party vote is small. At higher levels of turnout, Democrats comprise a larger share of the electorate, but they also have a higher defection rate.
Martinez, Michael
Submitted: 2002-03-18
Keywords: turnout, multinomial logit, simulation
Abstract: (click to show/hide) Conventional wisdom among pundits and some scholars posits that higher turnout should benefit liberal parties, since lower socioeconomic classes comprise a disproportionate share of the nonvoting population. Empirical tests of this prediction across elections have produced a wide variety of results, ranging from support for the conventional wisdom to suggestions that Republicans benefit from higher turnout to null findings. In this paper, we provide a simulation of the possible impact of increasing or decreasing turnout in a single election. Using data from the 2000 American National Election Study, we find that Gore would have benefitted slightly from higher turnout and would have been harmed slightly by lower turnout, but the overall magnitude of the effects of turnout on Gore's share of the two party vote is small. At higher levels of turnout, Democrats comprise a larger share of the electorate, but they also have a higher defection rate.
Conflict and Mediation Event Observations (CAMEO): A New Event Data Framework for the Analysis of Foreign Policy Interactions
Schrodt, Philip A., Gerner, Deborah J., Abu-Jabr, Rajaa, Yilmaz, Omur
Submitted: 2002-04-01
Keywords: event data, mediation, WEIS, Middle East, Balkans, West Africa
Abstract: (click to show/hide) The Conflict and Mediation Events Observations (CAMEO) framework is a new event data coding scheme optimized for the study of third-party mediation in international disputes. We have developed and implemented this system using the TABARI automated coding program, and have generated data sets for the Balkans (1989-2002; N=69,620), Levant (1979-2002; N=146,283), and West Africa (1989-2002; N=17,468) from Reuters and Agence France Presse reports. We describe why we decided to develop a new coding system, rather than continuing to use the World Events Interaction Survey (WEIS) framework that we have used in earlier work. Our decision involved both known weaknesses in the WEIS system, and some additional problems that we have found occur when WEIS is coded using automated methods. We have addressed these problems in constructing CAMEO, as well as producing much more completed documentation than has been available for WEIS. In this paper, we make several statistical comparisons of CAMEO-coded and WEIS-coded data in the three geographical regions. When the data are aggregated to a general behavioral levelÑverbal cooperation, material cooperation, verbal conflict and material conflictÑmost of the data sets show a high correlation (r>0.90) in the number of WEIS and CAMEO events coded per month. However, as we expected, CAMEO consistently picks up a greater number of events involving material cooperation. CAMEO and WEIS show similar irregularities in the distribution of events by category. Finally, there is a very significant correlation (r>0.57) between the count of CAMEO events specifically dealing with mediation and negotiation, and a pattern-based measure of mediation we developed earlier from WEIS data. Appendices in the paper show the CAMEO coding framework and examples from the codebook.
Schrodt, Philip A., Gerner, Deborah J., Abu-Jabr, Rajaa, Yilmaz, Omur
Submitted: 2002-04-01
Keywords: event data, mediation, WEIS, Middle East, Balkans, West Africa
Abstract: (click to show/hide) The Conflict and Mediation Events Observations (CAMEO) framework is a new event data coding scheme optimized for the study of third-party mediation in international disputes. We have developed and implemented this system using the TABARI automated coding program, and have generated data sets for the Balkans (1989-2002; N=69,620), Levant (1979-2002; N=146,283), and West Africa (1989-2002; N=17,468) from Reuters and Agence France Presse reports. We describe why we decided to develop a new coding system, rather than continuing to use the World Events Interaction Survey (WEIS) framework that we have used in earlier work. Our decision involved both known weaknesses in the WEIS system, and some additional problems that we have found occur when WEIS is coded using automated methods. We have addressed these problems in constructing CAMEO, as well as producing much more completed documentation than has been available for WEIS. In this paper, we make several statistical comparisons of CAMEO-coded and WEIS-coded data in the three geographical regions. When the data are aggregated to a general behavioral levelÑverbal cooperation, material cooperation, verbal conflict and material conflictÑmost of the data sets show a high correlation (r>0.90) in the number of WEIS and CAMEO events coded per month. However, as we expected, CAMEO consistently picks up a greater number of events involving material cooperation. CAMEO and WEIS show similar irregularities in the distribution of events by category. Finally, there is a very significant correlation (r>0.57) between the count of CAMEO events specifically dealing with mediation and negotiation, and a pattern-based measure of mediation we developed earlier from WEIS data. Appendices in the paper show the CAMEO coding framework and examples from the codebook.
Moving Mountains: Bayesian Forecasting As Policy Evaluation
Brandt, Patrick T., Freeman, John R.
Submitted: 2002-04-24
Keywords: Bayesian vector autoregression, VAR, policy evaluation, conditional forecasting
Abstract: (click to show/hide) Many policy analysts fail to appreciate the dynamic, complex causal nature of political processes. We advocate a vector autoregression (VAR) based approach to policy analysis that accounts for various multivariate and dynamic elements in policy formulation and for both dynamic and specification uncertainty of parameters. The model we present is based on recent developments in Bayesian VAR modeling and forecasting. We present an example based on work in Goldstein et al. (2001) that illustrates how a full accounting of the dynamics and uncertainty in multivariate data can lead to more precise and instructive results about international mediation in Middle Eastern conflict.
Brandt, Patrick T., Freeman, John R.
Submitted: 2002-04-24
Keywords: Bayesian vector autoregression, VAR, policy evaluation, conditional forecasting
Abstract: (click to show/hide) Many policy analysts fail to appreciate the dynamic, complex causal nature of political processes. We advocate a vector autoregression (VAR) based approach to policy analysis that accounts for various multivariate and dynamic elements in policy formulation and for both dynamic and specification uncertainty of parameters. The model we present is based on recent developments in Bayesian VAR modeling and forecasting. We present an example based on work in Goldstein et al. (2001) that illustrates how a full accounting of the dynamics and uncertainty in multivariate data can lead to more precise and instructive results about international mediation in Middle Eastern conflict.
What to do When Your Hessian is Not Invertible: Alternatives to Model Respecification in Nonlinear Estimation
Gill, Jeff, King, Gary
Submitted: 2002-05-14
Keywords: Hessian, Cholesky, generalized inverse, maximum likelihood, statistical computing, importance sampling, pseudo-variance, generalized linear model, singular normal
Abstract: (click to show/hide) What should a researcher do when statistical analysis software terminates before completion with a message that the Hessian is not invertable? The standard textbook advice is to respecify the model, but this is another way of saying that the researcher should change the question being asked. Obviously, however, computer programs should not be in the business of deciding what questions are worthy of study. Although noninvertable Hessians are sometimes signals of poorly posed questions, nonsensical models, or inappropriate estimators, they also frequently occur when information about the quantities of interest does exist in the data, through the likelihood function. We explain the problem in some detail and lay out two preliminary proposals for ways of dealing with noninvertable Hessians without changing the question asked.
Gill, Jeff, King, Gary
Submitted: 2002-05-14
Keywords: Hessian, Cholesky, generalized inverse, maximum likelihood, statistical computing, importance sampling, pseudo-variance, generalized linear model, singular normal
Abstract: (click to show/hide) What should a researcher do when statistical analysis software terminates before completion with a message that the Hessian is not invertable? The standard textbook advice is to respecify the model, but this is another way of saying that the researcher should change the question being asked. Obviously, however, computer programs should not be in the business of deciding what questions are worthy of study. Although noninvertable Hessians are sometimes signals of poorly posed questions, nonsensical models, or inappropriate estimators, they also frequently occur when information about the quantities of interest does exist in the data, through the likelihood function. We explain the problem in some detail and lay out two preliminary proposals for ways of dealing with noninvertable Hessians without changing the question asked.
Logical Inconsistency in King-based Ecological Regressions
Herron, Michael C., Shotts, Kenneth W.
Submitted: 2002-07-03
Keywords: ecological inference, EI-R, consistency, second stage regressions
Abstract: (click to show/hide) The statistical procedure EI-R, in which point estimates produced by the King (1997) ecological inference technique are used as dependent variables in a linear regression, can be logically inconsistent insofar as the assumptions necessary to support EI-R's first stage (ecological inference via King's method) can be incompatible with the assumptions supporting its second stage (linear regression). In light of this problem, we derive a specification test for logical consistency of EI-R and describe options available to a researcher who confronts test rejection. We then apply our test to the implementation of EI-R in Burden and Kimball's (1998) study of ticket splitting and find that this implementation is logically inconsistent. In correcting for this problem we show that Burden and Kimball's alleged substantive results are not results at all and instead are artifacts of a self-contradictory statistical technique.
Herron, Michael C., Shotts, Kenneth W.
Submitted: 2002-07-03
Keywords: ecological inference, EI-R, consistency, second stage regressions
Abstract: (click to show/hide) The statistical procedure EI-R, in which point estimates produced by the King (1997) ecological inference technique are used as dependent variables in a linear regression, can be logically inconsistent insofar as the assumptions necessary to support EI-R's first stage (ecological inference via King's method) can be incompatible with the assumptions supporting its second stage (linear regression). In light of this problem, we derive a specification test for logical consistency of EI-R and describe options available to a researcher who confronts test rejection. We then apply our test to the implementation of EI-R in Burden and Kimball's (1998) study of ticket splitting and find that this implementation is logically inconsistent. In correcting for this problem we show that Burden and Kimball's alleged substantive results are not results at all and instead are artifacts of a self-contradictory statistical technique.
The importance of statistical methodology for analyzing data from field experimentation: Evaluating voter mobilization strategies
Imai, Kosuke
Submitted: 2002-07-08
Keywords: field experiments, causal inference, instrumental variables
Abstract: (click to show/hide) We introduce a set of new Markov chain Monte Carlo algorithms for Bayesian analysis of the multinomial probit model. Our Bayesian representation of the model places a new, and possibly improper, prior distribution directly on the identifiable parameters and thus is relatively easy to interpret and use. Our algorithms, which are based on the method of marginal data augmentation, involve only draws from standard distributions and dominate other available Bayesian methods in that they are as quick to converge as the fastest methods but with a more attractive prior specification.
Imai, Kosuke
Submitted: 2002-07-08
Keywords: field experiments, causal inference, instrumental variables
Abstract: (click to show/hide) We introduce a set of new Markov chain Monte Carlo algorithms for Bayesian analysis of the multinomial probit model. Our Bayesian representation of the model places a new, and possibly improper, prior distribution directly on the identifiable parameters and thus is relatively easy to interpret and use. Our algorithms, which are based on the method of marginal data augmentation, involve only draws from standard distributions and dominate other available Bayesian methods in that they are as quick to converge as the fastest methods but with a more attractive prior specification.
Estimating Voters' Taste for Risk: Candidate Choice under Uncertainty
Berinsky, Adam, Lewis, Jeffrey B.
Submitted: 2002-07-08
Keywords: elections, risk preferences
Abstract: (click to show/hide) Recent work in political science has taken up the question of issue voting under conditions of uncertainty -- situations in which voters have imperfect information about the policy positions of candidates. Models that recognize this principle are realistic portrayals of the campaign environment, but may be limited in important respects. To date, the study of vote choice under uncertainty has made a common assumption of quadratic preferences. Such preferences implying that citizens behave in a very risk-averse manner when casting votes But this assumption is simply that; an assumption. Many other utility functions consistent with ``proximity'' voting could be chosen, including functions that imply risk neutral and risk acceptant behavior The assumption of risk-aversion is, after all, not simply a technical choice: it has important implications for how we view the process of citizen choice in elections and campaigns. If voters are risk-averse, candidates can benefit by making clear their positions on issues that they know will appeal to the electorate. Risk-averse voters therefore improve the quality of campaign discourse because candidates are punished for taking vague positions. But this scenario is only one among several possibilities. If voters are risk-neutral or risk-acceptant, candidates may have incentive to muddle the details of their policy plans and send ambiguous signals about their positions. Such a story of the campaign process may be less normatively appealing than one in which voters are risk-averse, but it might also more accurate portray the dynamics of political campaigns. We believe that the nature of risk preferences among the electorate should, therefore, be subject to greater scrutiny. In this paper, we move beyond the assumption of a particular spatial utility function and estimate voter's preferences for risk. We find that, contrary to the literature, voters are less risk averse than the quadratic model implies. Indeed, by the definition of risk preference developed in the paper, we find voters to be generally (nearly) risk-neutral and, in some cases, risk-acceptant.
Berinsky, Adam, Lewis, Jeffrey B.
Submitted: 2002-07-08
Keywords: elections, risk preferences
Abstract: (click to show/hide) Recent work in political science has taken up the question of issue voting under conditions of uncertainty -- situations in which voters have imperfect information about the policy positions of candidates. Models that recognize this principle are realistic portrayals of the campaign environment, but may be limited in important respects. To date, the study of vote choice under uncertainty has made a common assumption of quadratic preferences. Such preferences implying that citizens behave in a very risk-averse manner when casting votes But this assumption is simply that; an assumption. Many other utility functions consistent with ``proximity'' voting could be chosen, including functions that imply risk neutral and risk acceptant behavior The assumption of risk-aversion is, after all, not simply a technical choice: it has important implications for how we view the process of citizen choice in elections and campaigns. If voters are risk-averse, candidates can benefit by making clear their positions on issues that they know will appeal to the electorate. Risk-averse voters therefore improve the quality of campaign discourse because candidates are punished for taking vague positions. But this scenario is only one among several possibilities. If voters are risk-neutral or risk-acceptant, candidates may have incentive to muddle the details of their policy plans and send ambiguous signals about their positions. Such a story of the campaign process may be less normatively appealing than one in which voters are risk-averse, but it might also more accurate portray the dynamics of political campaigns. We believe that the nature of risk preferences among the electorate should, therefore, be subject to greater scrutiny. In this paper, we move beyond the assumption of a particular spatial utility function and estimate voter's preferences for risk. We find that, contrary to the literature, voters are less risk averse than the quadratic model implies. Indeed, by the definition of risk preference developed in the paper, we find voters to be generally (nearly) risk-neutral and, in some cases, risk-acceptant.
Enhancing the Validity and Cross-cultural Comparability of Survey Research
King, Gary, Murray, Christopher J. L., Salomon, Joshua A., Tandon, Ajay
Submitted: 2002-07-10
Keywords: survey research, scaling, vignettes, measurement
Abstract: (click to show/hide) We offer a new approach to writing survey questions and a new statistical model that together at least partially ameliorate two long-standing problems in survey research. The first is how to measure complicated concepts, such as freedom, health, political efficacy, pornography, etc., that researchers know how to define clearly only with reference to examples. The second problem is when different respondents interpret identical survey questions in incomparable ways, as can occur when comparing respondents in different countries speaking different languages, but it also occurs frequently with different groups in the same country. Our approach to these problems is to ask respondents for self-assessments of the concept being measured along with assessments, on the same scale, of each of several hypothetical individuals described by short vignettes. The actual (but not necessarily reported) levels for the people in the vignettes are, by the design of the survey, invariant over respondents and thus provide anchors for our statistical model to transform the self-assessments to a comparable scale. With analysis, simulations, and real surveys in several countries, we show how ignoring these problems can lead to the wrong substantive conclusions and how our approach can fix them. Our methods build on insights from application-specific research on voters and legislators in political science to produce a more general measurement device.
King, Gary, Murray, Christopher J. L., Salomon, Joshua A., Tandon, Ajay
Submitted: 2002-07-10
Keywords: survey research, scaling, vignettes, measurement
Abstract: (click to show/hide) We offer a new approach to writing survey questions and a new statistical model that together at least partially ameliorate two long-standing problems in survey research. The first is how to measure complicated concepts, such as freedom, health, political efficacy, pornography, etc., that researchers know how to define clearly only with reference to examples. The second problem is when different respondents interpret identical survey questions in incomparable ways, as can occur when comparing respondents in different countries speaking different languages, but it also occurs frequently with different groups in the same country. Our approach to these problems is to ask respondents for self-assessments of the concept being measured along with assessments, on the same scale, of each of several hypothetical individuals described by short vignettes. The actual (but not necessarily reported) levels for the people in the vignettes are, by the design of the survey, invariant over respondents and thus provide anchors for our statistical model to transform the self-assessments to a comparable scale. With analysis, simulations, and real surveys in several countries, we show how ignoring these problems can lead to the wrong substantive conclusions and how our approach can fix them. Our methods build on insights from application-specific research on voters and legislators in political science to produce a more general measurement device.
The Fruit of Jefferson's Dinner Party: Roll Call Analysis of the Compromise of 1790 with Substantive and Relational Constraints
Clinton, Joshua, Meirowitz, Adam
Submitted: 2002-07-12
Keywords: ideal point estimation, log roll, First Congress, agenda estimation
Abstract: (click to show/hide) The "Compromise of 1790" -- in which legislative gridlock in the First House (1789-1791) was supposedly resolved by a deal in which Southern states conceded to the assumption of states' Revolutionary War debt by the federal government in exchange for locating the permanent Capitol along the Potomac -- is one of the earliest and most colorful examples of log rolls in American politics. However, historians disagree on the validity or completeness of this story and this account is only directly supported by an account from Jefferson. We assess the extent to which the voting record actually supports the hypothesis that a compromise was reached sometime in mid June. Using substantive information about the roll call votes and relational information about the agenda to specify a model in which bill locations are identified we implement a Bayesian analysis (using MCMC methods). Our results do not support the traditional account of the compromise. In resolving the capital question legislators did not anticipate that assumption would carry. We also find that the final outcome was quite centrist and legislator ideal points are better explained by sectional, as opposed to ideological, theories.
Clinton, Joshua, Meirowitz, Adam
Submitted: 2002-07-12
Keywords: ideal point estimation, log roll, First Congress, agenda estimation
Abstract: (click to show/hide) The "Compromise of 1790" -- in which legislative gridlock in the First House (1789-1791) was supposedly resolved by a deal in which Southern states conceded to the assumption of states' Revolutionary War debt by the federal government in exchange for locating the permanent Capitol along the Potomac -- is one of the earliest and most colorful examples of log rolls in American politics. However, historians disagree on the validity or completeness of this story and this account is only directly supported by an account from Jefferson. We assess the extent to which the voting record actually supports the hypothesis that a compromise was reached sometime in mid June. Using substantive information about the roll call votes and relational information about the agenda to specify a model in which bill locations are identified we implement a Bayesian analysis (using MCMC methods). Our results do not support the traditional account of the compromise. In resolving the capital question legislators did not anticipate that assumption would carry. We also find that the final outcome was quite centrist and legislator ideal points are better explained by sectional, as opposed to ideological, theories.
State-Level Opinions from National Surveys: Poststratification using Hierarchical Logistic Regression
Park, David K., Gelman, Andrew, Bafumi, Joseph
Submitted: 2002-07-12
Keywords: Bayesian Inference, Hierarchical, Logit, Poststratification, Public Opinion, States, Elections
Abstract: (click to show/hide) Previous researchers have pooled national surveys in order to construct state-level opinions. However, in order to overcome the small n problem for less populous states, they have aggregated a decade or more of national surveys to construct their measures. For example, Erikson, Wright and McIver (1993) pooled 122 national surveys conducted over 13 years to produce state-level partisan and ideology estimates. Brace, Sims-Butler, Arceneaux, and Johnson (2002) pooled 22 surveys over a 25-year period to produce state-level opinions on a number of specific issues. We construct a hierarchical logistic regression model for the mean of a binary response variable conditional on poststratification cells. This approach combines the modeling approach often used in small-area estimation with the population information used in poststratification (see Gelman and Little 1997). We produce state-level estimates pooling seven national surveys conducted over a nine-day period. We first apply the method to a set of U.S pre-election polls, poststratified by state, region, as well as the usual demographic variables and evaluate the model by comparing it to state-level election outcomes. We then produce state-level partisan and ideology estimates by comparing it to Erikson, Wright and McIver's estimates.
Park, David K., Gelman, Andrew, Bafumi, Joseph
Submitted: 2002-07-12
Keywords: Bayesian Inference, Hierarchical, Logit, Poststratification, Public Opinion, States, Elections
Abstract: (click to show/hide) Previous researchers have pooled national surveys in order to construct state-level opinions. However, in order to overcome the small n problem for less populous states, they have aggregated a decade or more of national surveys to construct their measures. For example, Erikson, Wright and McIver (1993) pooled 122 national surveys conducted over 13 years to produce state-level partisan and ideology estimates. Brace, Sims-Butler, Arceneaux, and Johnson (2002) pooled 22 surveys over a 25-year period to produce state-level opinions on a number of specific issues. We construct a hierarchical logistic regression model for the mean of a binary response variable conditional on poststratification cells. This approach combines the modeling approach often used in small-area estimation with the population information used in poststratification (see Gelman and Little 1997). We produce state-level estimates pooling seven national surveys conducted over a nine-day period. We first apply the method to a set of U.S pre-election polls, poststratified by state, region, as well as the usual demographic variables and evaluate the model by comparing it to state-level election outcomes. We then produce state-level partisan and ideology estimates by comparing it to Erikson, Wright and McIver's estimates.
Robust Estimation and Outlier Detection for Overdispersed Multinomial Models of Count Data, with an Application to the Elian Effect in Florida
Mebane, Walter R., Sekhon, Jasjeet
Submitted: 2002-07-12
Keywords: robust estimation, overdispersed multinomial regression
Abstract: (click to show/hide) We develop a robust estimation method for regression models for vectors of counts (overdispersed multinomial models). The method requires only that the model is good for most---not all---of the observed data, and it identifies outliers. A Monte Carlo sampling experiment shows that the robust method can produce consistent parameter estimates and correct statistical inferences even when ten percent of the data are generated by a significantly different process, where nonrobust maximum likelihood estimation fails. We analyze Florida county vote data from the 2000 presidential election, considering votes for five categories of presidential candidates (Buchanan, Nader, Gore, Bush and ``other''), focusing on Cuban-Americans' reactions to the Elian Gonzalez affair. We replicate results regarding Buchanan's vote in Palm Beach County. We use Census tract data within Miami-Dade County to confirm the need to take the Cuban-American population explicitly into account. The analysis illustrates how the robust method can support triangulation to verify whether a regression specification is adequate.
Mebane, Walter R., Sekhon, Jasjeet
Submitted: 2002-07-12
Keywords: robust estimation, overdispersed multinomial regression
Abstract: (click to show/hide) We develop a robust estimation method for regression models for vectors of counts (overdispersed multinomial models). The method requires only that the model is good for most---not all---of the observed data, and it identifies outliers. A Monte Carlo sampling experiment shows that the robust method can produce consistent parameter estimates and correct statistical inferences even when ten percent of the data are generated by a significantly different process, where nonrobust maximum likelihood estimation fails. We analyze Florida county vote data from the 2000 presidential election, considering votes for five categories of presidential candidates (Buchanan, Nader, Gore, Bush and ``other''), focusing on Cuban-Americans' reactions to the Elian Gonzalez affair. We replicate results regarding Buchanan's vote in Palm Beach County. We use Census tract data within Miami-Dade County to confirm the need to take the Cuban-American population explicitly into account. The analysis illustrates how the robust method can support triangulation to verify whether a regression specification is adequate.
Tactical Coalition Voting
Morton, Becky, McCuen, Brian
Submitted: 2002-07-12
Keywords: strategic voting, proportional representation, coalition bargaining
Abstract: (click to show/hide) Most research on voting in proportional representation electoral systems assumes that voters either choose sincerely for their most preferred parties or strategically if threshold constraints mean their party has little chance of winning a seat. Voters are assumed to ignore possible coalition implications of their choices. However, formal models of coalition formation in PR systems, such as Austen-Smith and Banks (1988), assume voters care about the ultimate coalition formation in the parliament and vote strategically in order to affect that coalition formation process, which we call "tactical coalition voting." In this paper, we experimentally evaluate the extent voters in a PR system engage in tactical coalition voting. We find significant evidence that voters, even those non experienced with PR systems, do choose strategically to affect post election coalitions.
Morton, Becky, McCuen, Brian
Submitted: 2002-07-12
Keywords: strategic voting, proportional representation, coalition bargaining
Abstract: (click to show/hide) Most research on voting in proportional representation electoral systems assumes that voters either choose sincerely for their most preferred parties or strategically if threshold constraints mean their party has little chance of winning a seat. Voters are assumed to ignore possible coalition implications of their choices. However, formal models of coalition formation in PR systems, such as Austen-Smith and Banks (1988), assume voters care about the ultimate coalition formation in the parliament and vote strategically in order to affect that coalition formation process, which we call "tactical coalition voting." In this paper, we experimentally evaluate the extent voters in a PR system engage in tactical coalition voting. We find significant evidence that voters, even those non experienced with PR systems, do choose strategically to affect post election coalitions.
A Monte Carlo Analysis for Recurrent Events Data
Box-Steffensmeier, Janet M., De Boef, Suzanna
Submitted: 2002-07-13
Keywords: survival analysis, repeated events, heterogeneity, event dependence, simulations
Abstract: (click to show/hide) Scholars have long known that multiple events data, which occur when subjects experience more than one event, cause a problem when analyzed without taking into consideration the correlation among the events. In particular there has not been a solution about the best way to model the common occurrence of repeated events, where the subject experiences the same type of event more than once. Many event history model variations based on the Cox proportional hazards model have been proposed for the analysis of repeated events and it is well known that these models give different results (Clayton 1994; Lin 1994; Gao and Zhou 1997; Klein and Moeschberger 1997; Therneau and Hamilton 1997; Wei and Glidden 1997; Box-Steffensmeier and Zorn 1999; Hosmer and Lemeshow 1999; Kelly and Lim 2000). Our paper focuses on the two main alternatives for modeling repeated events data, variance corrected and frailty (also referred to as random effects) approaches, and examines the consequences these different choices have for understanding the interrelationship between dynamic processes in multivariate models, which will be useful across disciplines. Within political science, the statistical work resulting from this project will help resolve some important theoretical and policy debates about political dynamics, such as the liberal peace, by commenting on the reliability of the different modeling strategies used to test those theories and applying those models. Specifically, the results of the project will help assess whether one of the two primary approaches is better able to account for within-subject correlation. We evaluate the various modeling strategies using Monte Carlo evidence to determine whether and under what conditions alternative modeling strategies for repeated events are appropriate. The question as to the best modeling strategy for repeated events data is an important one. Our understanding of political processes, as in all studies, depends on the quality of the inferences we can draw from our models. There is currently little guidance about which approach or model is appropriate and so, not surprisingly, we see analysts unsure of the best way to analyze their data. Given the dramatic substantive differences that result from using the different models and approaches, this is a problem that will be of interest across research communities.
Box-Steffensmeier, Janet M., De Boef, Suzanna
Submitted: 2002-07-13
Keywords: survival analysis, repeated events, heterogeneity, event dependence, simulations
Abstract: (click to show/hide) Scholars have long known that multiple events data, which occur when subjects experience more than one event, cause a problem when analyzed without taking into consideration the correlation among the events. In particular there has not been a solution about the best way to model the common occurrence of repeated events, where the subject experiences the same type of event more than once. Many event history model variations based on the Cox proportional hazards model have been proposed for the analysis of repeated events and it is well known that these models give different results (Clayton 1994; Lin 1994; Gao and Zhou 1997; Klein and Moeschberger 1997; Therneau and Hamilton 1997; Wei and Glidden 1997; Box-Steffensmeier and Zorn 1999; Hosmer and Lemeshow 1999; Kelly and Lim 2000). Our paper focuses on the two main alternatives for modeling repeated events data, variance corrected and frailty (also referred to as random effects) approaches, and examines the consequences these different choices have for understanding the interrelationship between dynamic processes in multivariate models, which will be useful across disciplines. Within political science, the statistical work resulting from this project will help resolve some important theoretical and policy debates about political dynamics, such as the liberal peace, by commenting on the reliability of the different modeling strategies used to test those theories and applying those models. Specifically, the results of the project will help assess whether one of the two primary approaches is better able to account for within-subject correlation. We evaluate the various modeling strategies using Monte Carlo evidence to determine whether and under what conditions alternative modeling strategies for repeated events are appropriate. The question as to the best modeling strategy for repeated events data is an important one. Our understanding of political processes, as in all studies, depends on the quality of the inferences we can draw from our models. There is currently little guidance about which approach or model is appropriate and so, not surprisingly, we see analysts unsure of the best way to analyze their data. Given the dramatic substantive differences that result from using the different models and approaches, this is a problem that will be of interest across research communities.
Degeneracy and Inference for Social Networks
Handcock, Mark S.
Submitted: 2002-07-15
Keywords: Random graph models, log-linear network model, Markov fields, Markov Chain Monte Carlo
Abstract: (click to show/hide) Networks are a form of "relational data". Relational data arise in many social science fields and graph models are a natural approach to representing the structure of these relations. This framework has many applications including, for example, the structure of social networks, the behavior of epidemics, the interconnectedness of the WWW, and long-distance telephone calling patterns. We review stochastic models for such graphs, with particular focus on sexual and drug use networks. Commonly used Markov models were introduced by Frank and Strauss (1986) and were derived from developments in spatial statistics (Besag 1974). These models recognize the complex dependencies within relational data structures. To date, the use of graph models for networks has been limited by three interrelated factors: the complexity of realistic models, lack of use of simulation studies, and a poor understanding of the properties of inferential methods. In this talk we discuss these factors and the degeneracy of commonly promoted models. We also review the role of Markov Chain Monte Carlo (MCMC) algorithms for simulation and likelihood-based inference. These ideas are applied to a sexual relations network from Colorado Springs with the objective of understanding the social determinants of HIV spread. In this talk we focus on stochastic models for such graphs that can be used to represent the structural characteristics of the networks. In our applications, the nodes usually represent people, and the edges represent a specified relationship between the people.
Handcock, Mark S.
Submitted: 2002-07-15
Keywords: Random graph models, log-linear network model, Markov fields, Markov Chain Monte Carlo
Abstract: (click to show/hide) Networks are a form of "relational data". Relational data arise in many social science fields and graph models are a natural approach to representing the structure of these relations. This framework has many applications including, for example, the structure of social networks, the behavior of epidemics, the interconnectedness of the WWW, and long-distance telephone calling patterns. We review stochastic models for such graphs, with particular focus on sexual and drug use networks. Commonly used Markov models were introduced by Frank and Strauss (1986) and were derived from developments in spatial statistics (Besag 1974). These models recognize the complex dependencies within relational data structures. To date, the use of graph models for networks has been limited by three interrelated factors: the complexity of realistic models, lack of use of simulation studies, and a poor understanding of the properties of inferential methods. In this talk we discuss these factors and the degeneracy of commonly promoted models. We also review the role of Markov Chain Monte Carlo (MCMC) algorithms for simulation and likelihood-based inference. These ideas are applied to a sexual relations network from Colorado Springs with the objective of understanding the social determinants of HIV spread. In this talk we focus on stochastic models for such graphs that can be used to represent the structural characteristics of the networks. In our applications, the nodes usually represent people, and the edges represent a specified relationship between the people.
Connecting Interest Groups and Congress: A New Approach to Understanding Interest Group Success
Victor, Jennifer Nicoll
Submitted: 2002-07-16
Keywords: Interest Groups, Congress, Multiple Imputation, Bayesian Information Criterion, Ordinal Probit, Non-nested Models, Legislative Context
Abstract: (click to show/hide) The primary challenge in explaining interest group legislative success in Congress has been methodological. The discipline requires at least two critical elements to make progress on this important question. First, we need a theory that accounts for the highly interactive spatial game between interest groups and legislators. Second, the discipline needs an empirical model that associates interest groups and their activities with specific congressional bills. In this project I begin to contribute to our understanding of the complex relationship between interest groups and Congress. I develop a theory of group success that is based upon the strategies in which groups engage, the groups' organizational capacity, and the strategic context of legislation. I predict that groups will tailor their activities (and strategically spend their resources) in Congress based upon two critical factors: whether the group supports or opposes the legislation, and the legislative environment for the bill. To test this model I develop a unique sampling procedure and survey design. I use legislative hearings to generate a sample of groups that are associated with specific issues and survey them about their activities on those issues. Then, I associate each group's issue with a specific bill in Congress. I then track the bill to discern its final status. I create a dependent variable of interest group success that is based on the group's position (favor or oppose) and the final status of the bill. This sampling procedure and dependent variable allow me to make inferences about group behavior over specific legislative proposals. I develop independent variables of group activity, group organizational capacity, and legislative context from the survey instrument and objective information about the bills. To fill in gaps in the survey data set, I use a multiple imputation method that generates plausible values based on given distributions of data. I estimate two models-one for groups in favor of legislation, and one for opposition groups. The ordinal probit models generally support the theoretical expectations. In sum, I find that groups can best expend their resources in pursuit of rules that advantage their position rather than fighting for bill content.
Victor, Jennifer Nicoll
Submitted: 2002-07-16
Keywords: Interest Groups, Congress, Multiple Imputation, Bayesian Information Criterion, Ordinal Probit, Non-nested Models, Legislative Context
Abstract: (click to show/hide) The primary challenge in explaining interest group legislative success in Congress has been methodological. The discipline requires at least two critical elements to make progress on this important question. First, we need a theory that accounts for the highly interactive spatial game between interest groups and legislators. Second, the discipline needs an empirical model that associates interest groups and their activities with specific congressional bills. In this project I begin to contribute to our understanding of the complex relationship between interest groups and Congress. I develop a theory of group success that is based upon the strategies in which groups engage, the groups' organizational capacity, and the strategic context of legislation. I predict that groups will tailor their activities (and strategically spend their resources) in Congress based upon two critical factors: whether the group supports or opposes the legislation, and the legislative environment for the bill. To test this model I develop a unique sampling procedure and survey design. I use legislative hearings to generate a sample of groups that are associated with specific issues and survey them about their activities on those issues. Then, I associate each group's issue with a specific bill in Congress. I then track the bill to discern its final status. I create a dependent variable of interest group success that is based on the group's position (favor or oppose) and the final status of the bill. This sampling procedure and dependent variable allow me to make inferences about group behavior over specific legislative proposals. I develop independent variables of group activity, group organizational capacity, and legislative context from the survey instrument and objective information about the bills. To fill in gaps in the survey data set, I use a multiple imputation method that generates plausible values based on given distributions of data. I estimate two models-one for groups in favor of legislation, and one for opposition groups. The ordinal probit models generally support the theoretical expectations. In sum, I find that groups can best expend their resources in pursuit of rules that advantage their position rather than fighting for bill content.
Models of Causal Inference: Going Beyond the Neyman-Rubin-Holland Theory
Brady, Henry E.
Submitted: 2002-07-17
Keywords: causality, inference, models
Abstract: (click to show/hide) This paper explores various statistical and philosophical theories of causality, including the Neyman-Rubin-Holland (NRH) theory that is widely used in statistics. Although the NRH theory is increasingly well-known in the social sciences, philosophical theories are mostly unknown, partly because philosophers so seldom concern themselves with the everyday problems of practicing scientists, much less social scientists. Part of my goal is to bring these theories to a wider audience. I argue that the NRH theory is sometimes too "thin" a theory for reliable causal inferences (especially when it is naively extrapolated to all research situations without an understanding of the difficulty of supporting its assumptions) and that it would be better to develop a "thick" theory of causality that requires researchers to verify a number of different conditions before claiming causal relationships. Ideally, I argue, researchers should verify that any purported causal relationship satisfies the Humean conditions of the temporal precedence of causes before effects and the constant conjunction, or association, between causes and their effects. In addition, if a cause is observed in conjunction with its effect, then it should also satisfy the counterfactual condition that when the cause is absent in the same situation, it leads to the absence of the effect as well. Researchers should strive to find observable substitutes for this counterfactual situation in order to verify this condition. Causes should also be actively manipulated (and manipulatable) in ways that produce effects before claims are made about a causal relationship. Causes that cannot be manipulated or whose manipulation cannot be definitively described should be discounted. Finally, cause-effect relationships should be describable in terms of micro-mechanisms, typically employing theories at a lower level than the cause-effect relationship itself, that help to explain the details of why a particular cause leads to a particular effect. These mechanisms should themselves produce hypotheses that can be tested with data. I justify these conditions by an eclectic appeal to the philosophical and statistical literatures which have developed theories of causality based upon them, and I argue that methodologists should not be dogmatic about the "correct" theory of causality. Instead, methodologists should consider any philosophical position about causality to be useful if its ideas can improve our inferences. I also justify these conditions by showing that the very popular Neyman-Rubin-Holland theory implicitly requires some of these conditions, even though the theory is not always explicit about them. And where these conditions go beyond the Neyman-Rubin-Holland theory, there are good reasons to argue for their consideration in making causal claims. Ultimately, I suggest that we have been too cavalier about causality, and we must think harder about research design and theory in order to develop better causal claims.
Brady, Henry E.
Submitted: 2002-07-17
Keywords: causality, inference, models
Abstract: (click to show/hide) This paper explores various statistical and philosophical theories of causality, including the Neyman-Rubin-Holland (NRH) theory that is widely used in statistics. Although the NRH theory is increasingly well-known in the social sciences, philosophical theories are mostly unknown, partly because philosophers so seldom concern themselves with the everyday problems of practicing scientists, much less social scientists. Part of my goal is to bring these theories to a wider audience. I argue that the NRH theory is sometimes too "thin" a theory for reliable causal inferences (especially when it is naively extrapolated to all research situations without an understanding of the difficulty of supporting its assumptions) and that it would be better to develop a "thick" theory of causality that requires researchers to verify a number of different conditions before claiming causal relationships. Ideally, I argue, researchers should verify that any purported causal relationship satisfies the Humean conditions of the temporal precedence of causes before effects and the constant conjunction, or association, between causes and their effects. In addition, if a cause is observed in conjunction with its effect, then it should also satisfy the counterfactual condition that when the cause is absent in the same situation, it leads to the absence of the effect as well. Researchers should strive to find observable substitutes for this counterfactual situation in order to verify this condition. Causes should also be actively manipulated (and manipulatable) in ways that produce effects before claims are made about a causal relationship. Causes that cannot be manipulated or whose manipulation cannot be definitively described should be discounted. Finally, cause-effect relationships should be describable in terms of micro-mechanisms, typically employing theories at a lower level than the cause-effect relationship itself, that help to explain the details of why a particular cause leads to a particular effect. These mechanisms should themselves produce hypotheses that can be tested with data. I justify these conditions by an eclectic appeal to the philosophical and statistical literatures which have developed theories of causality based upon them, and I argue that methodologists should not be dogmatic about the "correct" theory of causality. Instead, methodologists should consider any philosophical position about causality to be useful if its ideas can improve our inferences. I also justify these conditions by showing that the very popular Neyman-Rubin-Holland theory implicitly requires some of these conditions, even though the theory is not always explicit about them. And where these conditions go beyond the Neyman-Rubin-Holland theory, there are good reasons to argue for their consideration in making causal claims. Ultimately, I suggest that we have been too cavalier about causality, and we must think harder about research design and theory in order to develop better causal claims.
Estimation of Evolutionary Processes
Honaker, James
Submitted: 2002-07-18
Keywords: evolution, replicator, dynamics, compositional, ECM
Abstract: (click to show/hide) Evolutionary game theory has accumulated an enormous body of theoretical work and even some proposed substantive applications. However, current empirical work has shown little evidence of evolutionary models matching field or experimental data. We argue that this is in part because estimation has been of overly restrictive models that make unwarranted assumptions either on the matrix of fitnesses in the evolutionary game, or more often, on the rule mapping the selection process. These heavy assumptions facilitate easy estimation procedures, but cripple the ability for evolutionary models to describe the data and for the researcher to reveal from the data the true quantities of interest in the evolutionary model. We demonstrate an EM based algorithm capable of estimating both the matrix of fitnesses and the selection mechanism, and apply this to experimental data. We show that the evolutionary model fits the experimental data progressively better as the assumptions of the evolutionary model are incorporated into the experiment. We also show that the model we propose can be used as a flexible estimator for deducing flows over compositional variables across time, and compare it to the more typical compositional model of Aitchison (1986).
Honaker, James
Submitted: 2002-07-18
Keywords: evolution, replicator, dynamics, compositional, ECM
Abstract: (click to show/hide) Evolutionary game theory has accumulated an enormous body of theoretical work and even some proposed substantive applications. However, current empirical work has shown little evidence of evolutionary models matching field or experimental data. We argue that this is in part because estimation has been of overly restrictive models that make unwarranted assumptions either on the matrix of fitnesses in the evolutionary game, or more often, on the rule mapping the selection process. These heavy assumptions facilitate easy estimation procedures, but cripple the ability for evolutionary models to describe the data and for the researcher to reveal from the data the true quantities of interest in the evolutionary model. We demonstrate an EM based algorithm capable of estimating both the matrix of fitnesses and the selection mechanism, and apply this to experimental data. We show that the evolutionary model fits the experimental data progressively better as the assumptions of the evolutionary model are incorporated into the experiment. We also show that the model we propose can be used as a flexible estimator for deducing flows over compositional variables across time, and compare it to the more typical compositional model of Aitchison (1986).
Diffusion or Confusion? Modeling Policy Diffusion with Discrete Event History Data
Buckley, Jack
Submitted: 2002-07-24
Keywords: policy, diffusion, survival, event, history
Abstract: (click to show/hide) No abstract provided.
Buckley, Jack
Submitted: 2002-07-24
Keywords: policy, diffusion, survival, event, history
Abstract: (click to show/hide) No abstract provided.
Taking the State Space Seriously: The Dynamic Linear Model and Bayesian Time Series Analysis
Buckley, Jack
Submitted: 2002-08-02
Keywords: time series, bayesian, state space, think tanks, business
Abstract: (click to show/hide) No abstract submitted.
Buckley, Jack
Submitted: 2002-08-02
Keywords: time series, bayesian, state space, think tanks, business
Abstract: (click to show/hide) No abstract submitted.
Rational Voting
Gelman, Andrew, Kaplan, Noah, Edlin, Aaron
Submitted: 2002-08-02
Keywords: elections, rational choice, sociotropic voting, turnout
Abstract: (click to show/hide) By separating the assumptions of ``rationality'' and ``selfishness,'' we show that it can be rational to vote if one is motivated by the effects of the election on society as a whole. For voters with ``social'' preferences the expected utility of voting is approximately independent of the size of the electorate, suggesting that rational voter turnouts can be substantial even in large elections. Less important elections are predicted to have lower turnout, but a feedback mechanism keeps turnout at a reasonable level under a wide range of conditions. We show how this feedback mechanism distinguishes voting from other free-rider problems. Our theory is consistent with several empirical findings in political science, including survey results that suggest that people vote based on perceived social benefit, the positive relation between turnout and (anticipated) closeness of the election, other forms of political participation, and declining response rates in opinion polls. Since our ''social'' theory of rational voting is instrumental, it creates a rich foundation to study {\em how} people vote as well as why. A rational person should make voting decisions almost entirely based on perceived social benefits of the election outcome.
Gelman, Andrew, Kaplan, Noah, Edlin, Aaron
Submitted: 2002-08-02
Keywords: elections, rational choice, sociotropic voting, turnout
Abstract: (click to show/hide) By separating the assumptions of ``rationality'' and ``selfishness,'' we show that it can be rational to vote if one is motivated by the effects of the election on society as a whole. For voters with ``social'' preferences the expected utility of voting is approximately independent of the size of the electorate, suggesting that rational voter turnouts can be substantial even in large elections. Less important elections are predicted to have lower turnout, but a feedback mechanism keeps turnout at a reasonable level under a wide range of conditions. We show how this feedback mechanism distinguishes voting from other free-rider problems. Our theory is consistent with several empirical findings in political science, including survey results that suggest that people vote based on perceived social benefit, the positive relation between turnout and (anticipated) closeness of the election, other forms of political participation, and declining response rates in opinion polls. Since our ''social'' theory of rational voting is instrumental, it creates a rich foundation to study {\em how} people vote as well as why. A rational person should make voting decisions almost entirely based on perceived social benefits of the election outcome.
Have Turnout Effects Really Declined? Testing the Partisan Implications of Marginal Voters
Gill, Jeff, Martinez, Michael
Submitted: 2002-08-09
Keywords: voting, turnout, partisan effects, simulation, multinomial logit
Abstract: (click to show/hide) In this paper, we review the theoretical foundations of the debate about whether higher election turnout advantages left parties, suggest a method of assessing the effects of turnout within a single election, and provide evidence from four U.S. elections that the partisan effects of turnout are contingent on the strength and polarity of the short-term forces. Our methodological approach to addressing whether the Democrats would have benefited from higher turnout (and whether the Republicans would have benefited from lower turnout) in a given election is to employ a new type of simulation based on multinomial logit estimates of the choices made by individual citizens. Our substantive approach is similar to Lacy and Burden (1999), in that we posit that U.S. citizens have three unordered choices in each election: vote Democratic, vote Republican, or abstain. We first estimate vote choice (including the abstention category) as an unordered multinomial logit function of standard variables associated with both candidate preference and the likelihood of voting. From that estimation, we derive probabilities for each respondent's selection of each of the three choices (abstain, vote Democratic, or vote Republican). From those probabilities, we simulate several levels of turnout. Higher turnout is simulated by progressively adding to the pool of voters actual abstainers with the lowest probability of abstaining of those remaining in the pool of abstainers. Whereas lower turnout is simulated by progressively subtracting from the electorate actual voters with the highest probability of abstaining. Our results across the four elections provide partial support for both the conventional SES-based model and the alternative defection-based model, though neither model's predictions are completely borne out empirically. As predicted by the conventional model, we find that the electorate has a greater Democratic tilt at higher levels of turnout, although that relationship has significantly weakened over time.
Gill, Jeff, Martinez, Michael
Submitted: 2002-08-09
Keywords: voting, turnout, partisan effects, simulation, multinomial logit
Abstract: (click to show/hide) In this paper, we review the theoretical foundations of the debate about whether higher election turnout advantages left parties, suggest a method of assessing the effects of turnout within a single election, and provide evidence from four U.S. elections that the partisan effects of turnout are contingent on the strength and polarity of the short-term forces. Our methodological approach to addressing whether the Democrats would have benefited from higher turnout (and whether the Republicans would have benefited from lower turnout) in a given election is to employ a new type of simulation based on multinomial logit estimates of the choices made by individual citizens. Our substantive approach is similar to Lacy and Burden (1999), in that we posit that U.S. citizens have three unordered choices in each election: vote Democratic, vote Republican, or abstain. We first estimate vote choice (including the abstention category) as an unordered multinomial logit function of standard variables associated with both candidate preference and the likelihood of voting. From that estimation, we derive probabilities for each respondent's selection of each of the three choices (abstain, vote Democratic, or vote Republican). From those probabilities, we simulate several levels of turnout. Higher turnout is simulated by progressively adding to the pool of voters actual abstainers with the lowest probability of abstaining of those remaining in the pool of abstainers. Whereas lower turnout is simulated by progressively subtracting from the electorate actual voters with the highest probability of abstaining. Our results across the four elections provide partial support for both the conventional SES-based model and the alternative defection-based model, though neither model's predictions are completely borne out empirically. As predicted by the conventional model, we find that the electorate has a greater Democratic tilt at higher levels of turnout, although that relationship has significantly weakened over time.
Electoral Outcomes, Economic Expectations and the 'Ethic of Self-Reliance'
Glasgow, Garrett, Weber, Roberto
Submitted: 2002-08-18
Keywords: economic individualism, election outcomes, economic expectations
Abstract: (click to show/hide) This paper examines how election outcomes affect individual economic expectations. In particular, we are interested in how differences in economic individualism change the relationship between election outcomes and individual expectations for personal economic well-being. We hypothesize that economic individualists will not link electoral outcomes to expectations for their personal economic well-being, while individuals who are not economic individualists will link the two. We confirm this hypothesis empirically using a postelection survey from the 1994 German Bundestag election.
Glasgow, Garrett, Weber, Roberto
Submitted: 2002-08-18
Keywords: economic individualism, election outcomes, economic expectations
Abstract: (click to show/hide) This paper examines how election outcomes affect individual economic expectations. In particular, we are interested in how differences in economic individualism change the relationship between election outcomes and individual expectations for personal economic well-being. We hypothesize that economic individualists will not link electoral outcomes to expectations for their personal economic well-being, while individuals who are not economic individualists will link the two. We confirm this hypothesis empirically using a postelection survey from the 1994 German Bundestag election.
Forecasts and Contingencies: From Methodology to Policy
Schrodt, Philip A.
Submitted: 2002-08-19
Keywords: forecast, foreign policy, international relations, future
Abstract: (click to show/hide) A "folk criticism" in political science maintains that the discipline should confine its efforts to explanation and avoid venturing down the dark, dirty, and dangerous path to forecasting and prediction. I argue that not only is this position inconsistent with the experiences of other sciences, but in fact the questions involved in making robust and valid predictions invoke many core methodological issues in political analysis. Those issues include, among others, the question of the level of predictability in political behavior, the problem of case selection in small-N situations, and the various alternative models that could be used to formalize predictions. This essay focuses on the problem of forecasting in international politics, and concludes by noting some of the problems of institutional culture -- bureaucratic and academic -- that have inhibited greater use of systematic forecasting methods in foreign policy.
Schrodt, Philip A.
Submitted: 2002-08-19
Keywords: forecast, foreign policy, international relations, future
Abstract: (click to show/hide) A "folk criticism" in political science maintains that the discipline should confine its efforts to explanation and avoid venturing down the dark, dirty, and dangerous path to forecasting and prediction. I argue that not only is this position inconsistent with the experiences of other sciences, but in fact the questions involved in making robust and valid predictions invoke many core methodological issues in political analysis. Those issues include, among others, the question of the level of predictability in political behavior, the problem of case selection in small-N situations, and the various alternative models that could be used to formalize predictions. This essay focuses on the problem of forecasting in international politics, and concludes by noting some of the problems of institutional culture -- bureaucratic and academic -- that have inhibited greater use of systematic forecasting methods in foreign policy.
An Integrated Perspective on Party Platforms and Electoral Choice
Elff, Martin
Submitted: 2002-08-19
Keywords: electoral behavior, party platforms, party manifestos, ideology, social cleavages, class voting, religious voting, comparative politics, principal curves, generalized additive models, dimensional analysis, discrete choice
Abstract: (click to show/hide) There are several perspectives on voting behavior that usually constitute separate strands of research: the impact of social background on vote choice, the relation between policy positions of parties and policy preferences of voters, and the effect of party platforms on the electoral success of parties. Although they all apply to the same entities, that is, to voters and parties, these different perspectives seem to have divergent implications. Thus we are in need of a way to reconcile these perspectives. The empirical results presented in this paper suggest a way what such a reconciliation should look like. They could be summarized as follows: In party platforms, several ideological dimensions can be distinguished that are connected with different cleavages in the Lispet-Rokkan sense. Second, it is shown that individuals from different social groups differ in the way they evaluate party platforms and choose among parties. Third, the way these individuals evaluate party platforms conforms to spatial notions of voting. Fourth, a general pattern of platform evaluation established on the base of pooled data of several countries accounts to a large degree for differences between levels of religious voting in these countries.
Elff, Martin
Submitted: 2002-08-19
Keywords: electoral behavior, party platforms, party manifestos, ideology, social cleavages, class voting, religious voting, comparative politics, principal curves, generalized additive models, dimensional analysis, discrete choice
Abstract: (click to show/hide) There are several perspectives on voting behavior that usually constitute separate strands of research: the impact of social background on vote choice, the relation between policy positions of parties and policy preferences of voters, and the effect of party platforms on the electoral success of parties. Although they all apply to the same entities, that is, to voters and parties, these different perspectives seem to have divergent implications. Thus we are in need of a way to reconcile these perspectives. The empirical results presented in this paper suggest a way what such a reconciliation should look like. They could be summarized as follows: In party platforms, several ideological dimensions can be distinguished that are connected with different cleavages in the Lispet-Rokkan sense. Second, it is shown that individuals from different social groups differ in the way they evaluate party platforms and choose among parties. Third, the way these individuals evaluate party platforms conforms to spatial notions of voting. Fourth, a general pattern of platform evaluation established on the base of pooled data of several countries accounts to a large degree for differences between levels of religious voting in these countries.
A Bayesian analysis of the multinomial probit model using marginal data augmentation
Imai, Kosuke, van Dyk, David A.
Submitted: 2002-08-21
Keywords: Bayesian analysis, Data augmentation, Prior distributions, Probit models, Rate of convergence
Abstract: (click to show/hide) We introduce a set of new Markov chain Monte Carlo algorithms for Bayesian analysis of the multinomial probit model. Our Bayesian representation of the model places a new, and possibly improper, prior distribution directly on the identifiable parameters and thus is relatively easy to interpret and use. Our algorithms, which are based on the method of marginal data augmentation, involve only draws from standard distributions and dominate other available Bayesian methods in that they are as quick to converge as the fastest methods but with a more attractive prior specification.
Imai, Kosuke, van Dyk, David A.
Submitted: 2002-08-21
Keywords: Bayesian analysis, Data augmentation, Prior distributions, Probit models, Rate of convergence
Abstract: (click to show/hide) We introduce a set of new Markov chain Monte Carlo algorithms for Bayesian analysis of the multinomial probit model. Our Bayesian representation of the model places a new, and possibly improper, prior distribution directly on the identifiable parameters and thus is relatively easy to interpret and use. Our algorithms, which are based on the method of marginal data augmentation, involve only draws from standard distributions and dominate other available Bayesian methods in that they are as quick to converge as the fastest methods but with a more attractive prior specification.
The Ordinary Election of Adolf Hitler: A Modern Voting Behavior Approach
King, Gary, Rosen, Ori, Wagner, Alexander F.
Submitted: 2002-08-23
Keywords: Voting Behavior, Ecological Inference, Elections
Abstract: (click to show/hide) How did free and fair democratic elections lead to the extrordinarily anti-democratic Nazi Party winning control of the Weimar Republic? The profound implications of this question have led scholars to make the Weimar elections the most studied elections in history and ``who voted for Hitler'' the single most asked question in elections research. Yet, despite this overwhelming attention, mostly from historians, the Nazi voting literature has treated these elections as largely unique events and thus comparison with other democratic elections as mostly irrelevant. The literature has also ignored most voting behavior theory and research in political science, and it has only rarely used modern statistical methods. In this paper, we adapt existing political science theories and new methods and find that many of the explanations offered in the Nazi voting literature, while probably correct, do not distinguish this election from almost any other, occuring in any country. For example, the prevailing explanation in the literature, that the Nazis were a ``catch all party'' because most social groups shifted in their favor by roughly the same amount, is a characteristic of the vast majority of election swings in every democracy, and so does not provide a useful explanation. We also show that a standard ``retrospective voting'' account of Nazi voting fits the distinctive aspects of this election well, once we recognize that the voters who were most hurt by the economic depression and hence most likely to oppose the government fall into two separate groups that have divergent interests. Those who were unemployed or at high risk of becomming unemployed shifted to the Communists, whose platform was designed to appeal mainly to this group, whereas the working poor, those at low risk of unemployment but still poor because of the economy (such as self-employed shop keepers and professionals, domestic workers, and helping family members), shifted disproportionately towards the Nazis, and accounted for most of the unusual dynamics of this election. The consequences of the election of Hitler were extraordinary, but the voting behavior that led to it was not.
King, Gary, Rosen, Ori, Wagner, Alexander F.
Submitted: 2002-08-23
Keywords: Voting Behavior, Ecological Inference, Elections
Abstract: (click to show/hide) How did free and fair democratic elections lead to the extrordinarily anti-democratic Nazi Party winning control of the Weimar Republic? The profound implications of this question have led scholars to make the Weimar elections the most studied elections in history and ``who voted for Hitler'' the single most asked question in elections research. Yet, despite this overwhelming attention, mostly from historians, the Nazi voting literature has treated these elections as largely unique events and thus comparison with other democratic elections as mostly irrelevant. The literature has also ignored most voting behavior theory and research in political science, and it has only rarely used modern statistical methods. In this paper, we adapt existing political science theories and new methods and find that many of the explanations offered in the Nazi voting literature, while probably correct, do not distinguish this election from almost any other, occuring in any country. For example, the prevailing explanation in the literature, that the Nazis were a ``catch all party'' because most social groups shifted in their favor by roughly the same amount, is a characteristic of the vast majority of election swings in every democracy, and so does not provide a useful explanation. We also show that a standard ``retrospective voting'' account of Nazi voting fits the distinctive aspects of this election well, once we recognize that the voters who were most hurt by the economic depression and hence most likely to oppose the government fall into two separate groups that have divergent interests. Those who were unemployed or at high risk of becomming unemployed shifted to the Communists, whose platform was designed to appeal mainly to this group, whereas the working poor, those at low risk of unemployment but still poor because of the economy (such as self-employed shop keepers and professionals, domestic workers, and helping family members), shifted disproportionately towards the Nazis, and accounted for most of the unusual dynamics of this election. The consequences of the election of Hitler were extraordinary, but the voting behavior that led to it was not.
Is Abortion A Wedge Issue for Latino Voters?
Abrajano, Marisa A., Nagler, Jonathan, Alvarez, R. Michael
Submitted: 2002-09-02
Keywords: elections, voters, abortion, issues, latino, ethnic
Abstract: (click to show/hide) In 2000 both major parties courted the growing Latino vote. Republicans hoped to benefit among this group based on the party's pro-life position and the belief that Latinos tend to be ideologically conservative, and that Latinos, in general, are Catholic. We present evidence indicating that this strategy of appealing to Latinos based on George Bush's pro-life stance garnered him fewer votes from the Latino electorate than Republican strategists hoped. While our results confirm that abortion is influential on vote choice at the individual level, abortion's impact at the aggregate level is smaller. When we say abortion is influential at the individual level, we mean that an individual voter is affected by the candidate's position on abortion. The `effect' of abortion we talk about in this case is the change in the probability of a voter choosing Bush (or Gore) if the voter were to change his or her position on abortion while the candidates' positions on abortion remained fixed. However, at the aggregate level we are looking at what would happen if one or the other of the candidates changed his position on abortion. A change in Bush's position would affect all voters. However, abortion's relatively small aggregate level of influence when compared to its impact at the individual level is due to the fact that such a change of position by a candidate would cause him to win some Latino votes based on his abortion stance, and at the same time it would also cause him to lose Latino votes from those who have the opposite view of abortion. As such, when these Latino votes are aggregated, the overall impact of abortion on the total vote is minimal, because the two effects tend to cancel each other out. Our findings are the first we are aware of to measure this overall impact of abortion, though several previous studies (Abramowitz 1995, Alvarez and Nagler 1995 and 1998) have demonstrated the importance of abortion at the individual level. We expect our findings to be applicable to the entire electorate, not just Latinos.
Abrajano, Marisa A., Nagler, Jonathan, Alvarez, R. Michael
Submitted: 2002-09-02
Keywords: elections, voters, abortion, issues, latino, ethnic
Abstract: (click to show/hide) In 2000 both major parties courted the growing Latino vote. Republicans hoped to benefit among this group based on the party's pro-life position and the belief that Latinos tend to be ideologically conservative, and that Latinos, in general, are Catholic. We present evidence indicating that this strategy of appealing to Latinos based on George Bush's pro-life stance garnered him fewer votes from the Latino electorate than Republican strategists hoped. While our results confirm that abortion is influential on vote choice at the individual level, abortion's impact at the aggregate level is smaller. When we say abortion is influential at the individual level, we mean that an individual voter is affected by the candidate's position on abortion. The `effect' of abortion we talk about in this case is the change in the probability of a voter choosing Bush (or Gore) if the voter were to change his or her position on abortion while the candidates' positions on abortion remained fixed. However, at the aggregate level we are looking at what would happen if one or the other of the candidates changed his position on abortion. A change in Bush's position would affect all voters. However, abortion's relatively small aggregate level of influence when compared to its impact at the individual level is due to the fact that such a change of position by a candidate would cause him to win some Latino votes based on his abortion stance, and at the same time it would also cause him to lose Latino votes from those who have the opposite view of abortion. As such, when these Latino votes are aggregated, the overall impact of abortion on the total vote is minimal, because the two effects tend to cancel each other out. Our findings are the first we are aware of to measure this overall impact of abortion, though several previous studies (Abramowitz 1995, Alvarez and Nagler 1995 and 1998) have demonstrated the importance of abortion at the individual level. We expect our findings to be applicable to the entire electorate, not just Latinos.
Designing Tests of the Supreme Court and the Separation of Powers
Sala, Brian R., Spriggs II, James F.
Submitted: 2002-09-13
Keywords: spatial voting theory, strategic behavior, Supreme Court
Abstract: (click to show/hide) While "rational choice" models of Supreme Court decision making have enhanced our appreciation for the separation of powers built into the Madisonian Constitutional design, convincing empirical support for a separation-of-powers (SOP) constraint on justices' behavior has been elusive. We apply a standard spatial voting model to identify circumstances in which "Attitudinalist" and SOP predictions about justices' behavior diverge. Our reconsideration of the theory indicates that prior efforts to test quantitatively the two models have been biased by having included cases for which the two models' predictions do not differ. While our more focused test offers a fairer test of the SOP constraint, the results strongly reject the SOP model. Nonetheless, our analysis provides leverage on this issue by: (1) delineating and executing necessary research design protocols for crafting a critical test of the SOP model; and (2) rejecting the two exogenously fixed alternative SOP model and suggesting avenues for future research.
Sala, Brian R., Spriggs II, James F.
Submitted: 2002-09-13
Keywords: spatial voting theory, strategic behavior, Supreme Court
Abstract: (click to show/hide) While "rational choice" models of Supreme Court decision making have enhanced our appreciation for the separation of powers built into the Madisonian Constitutional design, convincing empirical support for a separation-of-powers (SOP) constraint on justices' behavior has been elusive. We apply a standard spatial voting model to identify circumstances in which "Attitudinalist" and SOP predictions about justices' behavior diverge. Our reconsideration of the theory indicates that prior efforts to test quantitatively the two models have been biased by having included cases for which the two models' predictions do not differ. While our more focused test offers a fairer test of the SOP constraint, the results strongly reject the SOP model. Nonetheless, our analysis provides leverage on this issue by: (1) delineating and executing necessary research design protocols for crafting a critical test of the SOP model; and (2) rejecting the two exogenously fixed alternative SOP model and suggesting avenues for future research.
The Binomial-Beta Hierarchical Model for Ecological Inference: Methodological Issues and Fast Implementation via the ECM Algorithm
de Mattos, Rogerio S., Veiga, Alvaro
Submitted: 2002-10-17
Keywords: ecological inference, hierarchical models, binomial-beta distribution, ECM Algorithm
Abstract: (click to show/hide) The binomial-beta hierarchical model from King, Rosen, and Tanner (1999) is a recent contribution to ecological inference. Developed for the 2x2 tables case and from a bayesian perspective, the model is featured by the compounding of binomial and beta distributions into a hierarchical structure. From a sample of aggregate observations, inference with this model can be made regarding values of unobservable disaggregate variables. The paper reviews this EI model with two purposes: First, a faster approach to use it in practice, based on explicit modeling of the disaggregate data generation process along with posterior maximization implemented via the ECM algorithm, is proposed and illustrated with an application to a real dataset; second, limitations concerning the use of marginal posteriors for binomial probabilities as the vehicle of inference (basically, the failure to respect the accounting identity) instead of the predictive distributions for the disaggregate proportions are pointed. In the concluding section, principles for EI model building in general and directions for further research are suggested.
de Mattos, Rogerio S., Veiga, Alvaro
Submitted: 2002-10-17
Keywords: ecological inference, hierarchical models, binomial-beta distribution, ECM Algorithm
Abstract: (click to show/hide) The binomial-beta hierarchical model from King, Rosen, and Tanner (1999) is a recent contribution to ecological inference. Developed for the 2x2 tables case and from a bayesian perspective, the model is featured by the compounding of binomial and beta distributions into a hierarchical structure. From a sample of aggregate observations, inference with this model can be made regarding values of unobservable disaggregate variables. The paper reviews this EI model with two purposes: First, a faster approach to use it in practice, based on explicit modeling of the disaggregate data generation process along with posterior maximization implemented via the ECM algorithm, is proposed and illustrated with an application to a real dataset; second, limitations concerning the use of marginal posteriors for binomial probabilities as the vehicle of inference (basically, the failure to respect the accounting identity) instead of the predictive distributions for the disaggregate proportions are pointed. In the concluding section, principles for EI model building in general and directions for further research are suggested.
Standard Voting Power Indexes Don't Work: An Empirical Analysis
Gelman, Andrew, Katz, Jonathan, Bafumi, Joseph
Submitted: 2002-11-02
Keywords: Banzhaf index, decisive vote, elections, electoral college, Shapley value, voting power
Abstract: (click to show/hide) Voting power indexes such as that of Banzhaf (1965) are derived, explicitly or implicitly, from the assumption that all votes are equally likely (i.e., random voting). That assumption can be generalized to hold that the probability of a vote being decisive in a jurisdiction with $n$ voters is proportional to $1/\sqrt{n}$. We test---and reject---this hypothesis empirically, using data from several different U.S. and European elections. We find that the probability of a decisive vote is approximately proportional to $1/n$. The random voting model (or its generalization, the square-root rule) overestimates the probability of close elections in larger jurisdictions. As a result, classical voting power indexes make voters in large jurisdictions appear more powerful than they really are. The most important political implication of our result is that proportionally weighted voting systems (that is, each jurisdiction gets a number of votes proportional to $n$) are basically fair. This contradicts the claim in the voting power literature that weights should be approximately proportional to $\sqrt{n}$.
Gelman, Andrew, Katz, Jonathan, Bafumi, Joseph
Submitted: 2002-11-02
Keywords: Banzhaf index, decisive vote, elections, electoral college, Shapley value, voting power
Abstract: (click to show/hide) Voting power indexes such as that of Banzhaf (1965) are derived, explicitly or implicitly, from the assumption that all votes are equally likely (i.e., random voting). That assumption can be generalized to hold that the probability of a vote being decisive in a jurisdiction with $n$ voters is proportional to $1/\sqrt{n}$. We test---and reject---this hypothesis empirically, using data from several different U.S. and European elections. We find that the probability of a decisive vote is approximately proportional to $1/n$. The random voting model (or its generalization, the square-root rule) overestimates the probability of close elections in larger jurisdictions. As a result, classical voting power indexes make voters in large jurisdictions appear more powerful than they really are. The most important political implication of our result is that proportionally weighted voting systems (that is, each jurisdiction gets a number of votes proportional to $n$) are basically fair. This contradicts the claim in the voting power literature that weights should be approximately proportional to $\sqrt{n}$.
Causal inference with general treatment regimes: Generalizing the propensity score
Imai, Kosuke, van Dyk, David A.
Submitted: 2002-11-18
Keywords: causal inference, income, medical expenditure, non-random treatment, observational studies, schooling, smoking, subclassification
Abstract: (click to show/hide) In this article, we develop the theoretical properties of the propensity function which is a generalization of the propensity score of Rosenbaum and Rubin (1983). Methods based on the propensity score have long been used for causal inference in observational studies; they are easy to use and can effectively reduce the bias caused by non-random treatment assignment. Although treatment regimes are often not binary in practice, the propensity score methods are generally confined to binary treatment scenarios. Two possible exceptions were suggested by Joffe and Rosenbaum (1999) and Imbens (2000) for ordinal and categorical treatments, respectively. In this article, we develop theory and methods which encompass all of these techniques and widen their applicability by allowing for arbitrary treatment regimes. We illustrate our propensity function methods by applying them to two data sets; we estimate the effect of smoking on medical expenditure and the effect of schooling on wages. We also conduct Monte Carlo experiments to investigate the performance of our methods.
Imai, Kosuke, van Dyk, David A.
Submitted: 2002-11-18
Keywords: causal inference, income, medical expenditure, non-random treatment, observational studies, schooling, smoking, subclassification
Abstract: (click to show/hide) In this article, we develop the theoretical properties of the propensity function which is a generalization of the propensity score of Rosenbaum and Rubin (1983). Methods based on the propensity score have long been used for causal inference in observational studies; they are easy to use and can effectively reduce the bias caused by non-random treatment assignment. Although treatment regimes are often not binary in practice, the propensity score methods are generally confined to binary treatment scenarios. Two possible exceptions were suggested by Joffe and Rosenbaum (1999) and Imbens (2000) for ordinal and categorical treatments, respectively. In this article, we develop theory and methods which encompass all of these techniques and widen their applicability by allowing for arbitrary treatment regimes. We illustrate our propensity function methods by applying them to two data sets; we estimate the effect of smoking on medical expenditure and the effect of schooling on wages. We also conduct Monte Carlo experiments to investigate the performance of our methods.
