|simon jackman||61 entries|
|research||last updated: March 27, 2014|
Uniform swing — the assumption that all states move by the same amount in a given election — is now a better approximation to state-level election outcomes in U.S.~presidential elections than at any time since 1944. This implies that reasonable forecasts of state-level election outcomes can be made with swing estimates from national-level forecasting models. I assess the performance of state-level forecasts made this way, comparing them to forecasts based on poll-averaging models (such as my own). If a good forecast of the national swing is available, it is possible — but not easy — to improve upon the state-level forecasts generated by applying uniform swing.
To appear, PS: Political Science and Politics. 47(2).
Poll movements shape the betting markets, especially as the election draws close, and especially when the polls suggest the election might be closer than previously thought. Ahead of the 2013 Australian Federal election, Newspoll and Nielsen seem especially important "market movers", probably due to their long-standing brand power and their association with newspapers in multiple media markets. Betting markets react quickly to changes in the polls, but not instantaneously. Poll movements take at least 48 hours to be digested by the betting markets, suggesting that secondary media reports of the polls are important (e.g., evening TV news reports of poll results from that morning’s newspapers).
In a lop-sided election like 2013, we can expect that betting markets will react to little or no change in the polls,"catching up" with the polls in the final week of the campaign. That is, we should not to be surprised to see political betting markets reacting to polls. Rather, we should expect a somewhat subtle interplay between the two, as shown in the analyses presented here.
To appear in Carol Johnson and John Warhurst (eds), The 2013 Australian Federal Election: Australian National University Press: Canberra.
with John Ahlquist and Kenneth Mayer.
State legislatures around the United States have entertained–and passed–laws requiring voters to present various forms of state-issued identification in order to cast ballots. Proponents argue that such laws protect the integrity of the electoral process, sometimes claiming that fraudulent voting is widespread. We report the results of a survey list experiment fielded immediately after the 2012 US general election designed to measure the prevalence of one specific type of voter fraud most relevant to voter ID laws: voter impersonation. We find no evidence of voter impersonation, even in the states most contested in the Presidential or statewide campaigns. We also find that states with strict voter ID laws and states with same-day voter registration are no different from others in the (non) existence of voter impersonation. To address possible "lower bound" problems with our conclusions we run both parallel and subsequent experiments to calibrate our findings. These ancillary list experiments indicate that the lower bound on the population reporting voter impersonation is nearly identical with the proportion of the population reporting abduction by extraterrestrials. Based on this evidence, strict voter ID requirements address a problem that did not exist in the 2012 US election. Effort designed to improve American election infrastructure and security would be better directed toward other initiatives. Invited for revise-and-resubmit.
with Larry Bartels
We propose a mathematical framework for modeling opinion change using large-scale longitudinal data sets. Our framework encompasses two varieties of Bayesian learning theory as well as Mannheim’s theory of generational responses to political events. The basic assumptions underlying the model are (1) that historical periods are characterized by shocks to existing political opinions, and (2) that individuals of different ages may attach different weights to those political shocks. Political generations emerge endogenously from these basic assumptions: the political views of identifiable birth cohorts differ, and evolve distinctively through time, due to the interaction of age-specific weights with period-specific shocks. We employ this model to examine generational changes in party identification using survey data from the 1952–2008 American National Election Studies. Electoral Studies 33:7-18.
with Shanto Iyengar, Solomon Messing, Nicholas Valentino, Toril Aalberg, Raymond Duch, Kyu S. Hahn, Stuart Soroka, Allison Harell and Tetsuro Kobayashi
Public Opinion Quarterly. 77(3):641-655.
with Lynn Vavreck
Would the outcome of the 2008 U.S. presidential election have been different if Barack Obama had not been the Democratic nominee? In this paper we analyze 33 head-to-head match-ups, some real and some hypothetical, to understand more precisely how candidate traits and election-level characteristics affect election outcomes. We find that "old-fashioned" racial stereotyping is uniquely important in decisions about Obama in 2008, relative to its role in past elections or in 2008 choices substituting Clinton or Edwards for Obama. Similarly, "new" or symbolic racism is an exceptionally important predictor of vote choice when Obama in the choice set, and it gains in importance over the year leading up to the election. The Democrats would have won the 2008 election regardless of who they nominated, but the average Democratic party nominee from the last 16 years, and either of Edwards or Clinton, would have done better against McCain than Obama, although in Clinton's case, not by much.
Slides from a presentation to a workshop sponsored by the by the University of Queensland's Institute for Social Science Research (ISSR) and the Social Research Centre. Melbourne, Australia.
Slides from a presentation to the ANES Board meeting, Stanford University, Stanford, California, March 18, 2011.
with Lynn Vavreck
Journal of Elections, Public Opinion and Parties. 20(2):153-186
with Lynn Vavreck (UCLA)
This document replicates the data analysis reported in "Cosmopolitanism", which appears in the Paul Sniderman and Ben Highton edited collection. In fact, the R in this Sweave document actually produces the graphs and tables that appear in our essay, inter alia.
with Lynn Vavreck (UCLA)
Born to a black father from Kenya and a white mother from Kansas, Barack Hussein Obama was raised at various times in Honolulu and Jakarta by his mother, his father, an Indonesian stepfather, and his maternal grandparents. Like many U.S. presidents, Obama attended private schools and worked in a law firm; yet Obama is different from most Americans, and not just because he is black. Even his name suggests something different: he is neither parochial nor local, but a product of different cultures. Using panel data collected over Obama's candidacy for the presidency, we explore the role of cosmopolitanism in shaping vote intentions during the Democratic primary and the general election. Despite once being an important concept in mid-century political science and sociology, cosmopolitanism is largely absent from contemporary research on public opinion and behavior. We introduce a new operationalization of the concept and show that in the face of numerous controls, it has strong effects on vote choice both in the Democratic primary and the general election. Cosmopolitanism may constitute a new dimension in the study of race and politics, especially given the increasingly diverse pool of candidates entering American politics.
Analysis of Green support by polling place, 2010 Australian Federal election, examining the high spatial concentration of the high levels of Green support.
Entry for the Sage Encyclopedia of Political Science
Slides from a presentation to a conference sponsored by the American Political Science Association on "Democracy Audits and Governmental Indicators", University of California, Berkeley, October 30-31, 2009.
with Lynn Vavreck
Despite Barack Obama's momentum in the early phase of the Democratic primaries and caucuses, the process of selecting a nominee took longer than usual. Eight candidates were winnowed to two, but the remaining two were competitive until the last primary was held. Further, when it appeared Barack Obama would win the nomination, substantial numbers of Hillary Clinton supporters vowed to defect from the party and support the opposition in the general election. Momentum, it seems, got stuck. How do we explain this unusual string of events? Why were some people unwilling to join the Obama bandwagon once he emerged as a viable front-runner --- and utlimately, the nominee? In this paper we bring unique panel data to bear on questions about primary vote choice. While it is known that attitudes about race predict vote choice in partisan contests, we demonstrate that these attitudes interact with the race of the candidate to explain voters' choices and transitions among candidates even in an intra-party contest.
with Joshua Clinton
A comparison of the Bayesian quadratic-normal approach to the analysis of roll call data with W-NOMINATE. Legislative Studies Quarterly, V34(4):593-621.
with Lynn Vavreck
Presentation in "Presidential Politics: Race, Class, Faith & Gender in the 2008 Election", Monday, November 10, 2008
Report on surveys fielded in Australia and the United States ahead of the 2008 U.S. Presidential election. The Australian fieldwork was commissioned by the United States Studies Centre at the University of Sydney; the U.S. fieldwork is part of the Cooperative Campaign Analysis Project (Lynn Vavreck of UCLA and I are the principal investigators).
Oxford Handbook of Political Methodology, Henry E. Brady, David Collier and Janet Box-Steffensmeier (eds). 2008. Oxford University Press: Oxford, UK. pp119-151.
with Joshua Clinton
We reassess the claim that Senator Barack Obama is the most liberal senator in 2007. National Journal said the same thing about John Kerry in 2003. Presentation to the Department of Political Science, University of Rochester
with Shawn Treier
We apply formal, statistical measurement models to the Polity indicators, used widely in studies of international relations to measure democracy. In so doing, we make explicit the hitherto implicit assumptions underlying scales built using the Polity indicators. Modeling democracy as a latent variable allows us to assess the ``noise'' (measurement error) in the resulting measure. We show that this measurement error is considerable, and has substantive consequences when using a measure of democracy as an independent variable in cross-national statistical analysis. Our analysis suggests that skepticism as to the precision of the Polity democracy scale is well-founded, and that many researchers have been overly sanguine about the properties of the Polity democracy scale in applied statistical work. American Journal of Political Science, 2008, V52(1):201-217.
with Matthew S. Levendusky and Jeremy C. Pope
Many empirical studies of American politics, particularly legislative politics, are vitally dependent on measures of the partisanship of a district. We develop a measurement model for this quantity, estimating how Democratic (or, conversely, Republican) districts are in the absence of a election-specific, short-term forces, such as national-level swings specific to particular elections, incumbency advantage, and home-state effects in presidential elections. We estimate the model using readily available data: electoral returns and district-level demographic characteristics. We estimate the model with five decades of data (1952-2000), and describe how the distribution of district partisanship has changed over time, in response to population movements and redistricting, particularly via the creation of majority-minority districts. We validate the measure with analysis of Congressional roll call data, and show how to enrich this measure using other available indicators of district partisanship, such as survey data. Journal of Politics. July 2008. V70(3):736-753.
with Peter Brent
We examine recent trends in electoral enrollment in Australia, which suggest that since the 2004 Federal election, enrollments have fallen as a share of the eligible population. Recent months have seen a marked increase in enrolment, but the disparity remains. We suggest some reasons as to why this has occured. Democratic Audit of Australia
In this short note I draw attention to R's abilities for data processing, and in particular, R's ``Perl-like'' abilities for parsing text. The application involves reading data from the web site of the California Secretary of State containing "real-time" updates of results from the 2006 congressional elections. In the zip file I provide a PDF document describing the R code that reads and processes the data, along with the R command file itself. The Political Methodologist. V14(2):11-15.
with Matthew S. Levendusky and Jeremy C. Pope
We've been getting numerous requests for our estimates of district-level partisanship. Here we provide (1) estimates of district partisanship, one set per decade; (2) 2000 samples from the joint posterior density of district-level partisanship, one file per decade. Our partisanship scale runs in a Republican to Democratic direction (negative to positive), and is normalized to have mean zero and standard deviation one across districts. The zip file contains comma-delimited ascii files, with a header row. 37MB zip file.
Slides from a presentation made at the Social Science Methodology Conference, hosted by the Australian Consoritium for Social and Political Research, University of Sydney, Dec 10-13, 2006
Data, R code, detailed instructions on the analysis appearing "The Most Liberal Senator?" PS: Political Science and Politics, 2004 V37(3): 805-811. The zip file contains a PDF file documenting the analysis, and two data sets (one an R binary object, the other a comma-delimited file) containing the roll call matrix from the U.S. Senate for National Journal's 62 "key votes" for 2003.
with Alex Tahk and Christina Maimone; contributions from Achim Zeileis and Jim Fearon.
R classes and methods for the Bayesian analysis of roll call data (item-response models); elementary Bayesian statistics; maximum likelihood estimation of zero-inflated and hurdle models for count data; seats-votes curves; utility functions. Version 0.73. Available from the Comprehensive R Archive Network (CRAN) Package manual available for download here (below).
with Paul M. Sniderman
Can citizens learn from talking politics with one another? To bring out the logic of deliberation, we focus on a simplified model of political discussion: a one-exchange argument. Our model rests on three conditions, all commonly satisfied in real life: (1) that only two alternatives are open for choice – support or opposition to a policy; (2) that as political sophistication increases, so too does the probability that citizens will choose the policy alternative more consonant with their most thoroughly considered view of the matter; and (3) that arguments on opposing sides of an issue are of equal quality. Taking advantage of a specially designed experiment embedded in a large public opinion survey in France, we find that the proportion of citizens choosing policy alternatives consonant with their more general ideological orientations does not increase over the course of our experiment. In the aggregate, we find that deliberation leads at least as many people to ideologically inconsistent positions as it helps people find their way to ideologically consistent positions. In this sense, we find that deliberation is for naught.
Journal of Politics. 2006. V68(2): 272-283 On-line appendix with measurement details available, below.
with Neal Beck and Howard Rosenthal
We use a Bayesian dynamic linear model to track approval for George W.Bush over time. Our analysis deals with several issues that have been usually addressed separately in the extant literature. First, our analysis uses polling data collected at a higher frequency than is typical, using over 1,100 published national polls, and data on macro-economic conditions collected at the weekly level. By combining this much poll information, we are much better poised to examine the public's reactions to events over shorter time scales than can the typical analysis of approval that utilizes monthly or quarterly approval. Second, our statistical modeling explicitly deals with the sampling error of these polls, as well as the possibility of bias in the polls due to house effects. Indeed, quite aside from the question of ``what drives approval?'', there is considerable interest in the extent to which polling organizations systematically diverge from one another in assessing approval for the president. These bias parameters are not only necessary parts of any realistic model of approval that utilizes data from multiple polling organizations, but easily estimated via the Bayesian dynamic linear model.
slides from a talk given to the annual meeting of the American Association for the Advancement of Science, St Louis, Missouri, February 20, 2006
slides from a talk to the Method of Analysis in the Social Sciences Colloquium, Stanford University, March 7, 2006
Poll results vary over the course of a campaign election and across polling
organisations, making it difficult to track genuine changes in voter support. I
present a statistical model that tracks changes in voter support over time by
pooling the polls, and corrects for variation across polling organisations due to
biases known as ‘house effects’. The result is a less biased and more precise
estimate of vote intentions than is possible from any one poll alone. I use five
series of polls fielded over the 2004 Australian federal election campaign
(ACNielsen, the ANU/ninemsn online poll, Galaxy, Newspoll, and Roy
Morgan) to generate daily estimates of the Coalition’s share of two-party
preferred (2PP) and first preference vote intentions. Over the course of the
campaign there is about a 4 percentage point swing to the Coalition in first
preference vote share (and a smaller swing in 2PP terms), that begins prior to
the formal announcement of the election, but is complete shortly after the
leader debates. The ANU/ninemsn online poll and Morgan are found to have
large and statistically significant biases, while, generally, the three phone polls
have small and/or statistically insignificant biases, with ACNielsen and (in
particular) Galaxy performing quite well in 2004. Australian Journal of Political Science. 2005. V40(4): 499-517. Replication archive and technical appendix available below.
in Mortgage Nation: The 2004 Australian Election. Marian Simms and John Warhurst (eds). 2005. Perth, Western Australia: API Network/Curtin University of Technology. pp335-347.
slides from two workshop talks given at Pennsylvania State University, April 29-30, 2005
slides from a presentation to the Voteworld conference, Institute for Governmental Studies, University of California, Berkeley, February 26, 2005
slides from a talk given at the Seminar in Bayesian Inference in Econometrics and Statistics, Washington University in St Louis, August 1, 2005
multiple regression analysis of rates of informality in the 2004 Australian House of Representatives election; divisional level data; key predictors are non-English speaking at home, ballot length (and the interaction of the two), along with tertiary education and an indicator for divisions in jurisdictions with optional preferential voting in their legislative elections.
Op-ed piece comparing the two conservative victories in the 2004 U.S. and Australian elections; submitted to Dissent (www.dissent.com.au)
Annual Reviews of Political Science. 2004. 7:483-505.
with Joshua Clinton and Doug Rivers
PS: Political Science and Politics. 2004. 37(3):805-811.
Political Analysis 2004 12(4): 400-424.
with Joshua Clinton and Doug Rivers
American Political Science Review. 98(2):355-370.
The Political Methodologist. 2004. 12(2):6-11
Sage Encyclopedia of Social Science Research Methods
with D. Sunshine Hillygus
American Journal of Political Science. 2003. 47(4):583-596.
The Political Methodologist. 2003. 11(2):20-22.
The Cambridge Handbook of the Social Sciences in Australia. 2002. pp266-286.
a contribution to the International Encyclopedia of the Social and Behavioral Sciences (Elsevier)
Slides from a presentation to the Annual Meeting of the Australasian Political Studies Association, Brisbane, September 2001.
Political Analysis. 2001. 9(3):227-241.
Political Analysis. 2000. 8(4):307-332.
American Journal of Political Science. 2000. 44(2):375-404.
Electoral Studies. 1999. 18:29-48.
Australian Journal of Political Science. 1998. 33(2): 167-186.
with Nathaniel Beck
American Journal of Political Science. 1998. 33:167-186.
Current Affairs Bulletin V73(3, October/November): 23-26
Australian Journal of Political Science 30:347-55
with Gary Marks
Australian Journal of Political Science 29:277-91
British Journal of Political Science 24:319-57
Australian Journal of Political Science 27: 434-448