Academy Fellow

Elections, electoral systems and public opinion are vital components of democracies. My research has made a number of contributions to our understanding of each. Surveys have a long history in the social sciences, but face considerable challenges from non-response bias and cost. As a principal investigator of large survey projects in the United States, I helped develop and test techniques for driving high rates of completion to lower-cost modes of survey administration, helping to map tradeoffs with respect to data quality and survey cost. Using administrative records linked to survey data, we were able to disentangle non-response, mis-reporting and survey activation as sources of bias in survey respondents’ self-reports of political participation, a decades-long puzzle in the study of public opinion.

As a user of surveys, I pioneered techniques for combining surveys from disparate sources, gathered over different time periods and locations, to form temporally smooth estimates of the evolution of public opinion, drawing on techniques from statistics and computer science (hidden Markov models). Any one opinion poll is limited by its sample size and coverage (in time and space), and may come accompanied with bias (sample composition, question-wording effects). Moreover, polls are hardly the only relevant source of information about public opinion; in the specific context of elections and political attitudes, there is considerable information in “free” sources of data such as previous election outcomes, Census data and so on, and the performance of polls in prior elections. My work with Bayesian statistical models lets us rigorously integrate these sources of information, and to produce estimates and uncertainty assessments. Techniques I pioneered are now used widely by scholars and practitioners, especially to track trajectories of public opinion over election campaigns and to predict election results.

Electoral systems translate preferences (expressed as votes) into seats in legislatures. Understanding the properties of different electoral systems is a central problem in political science. Like many democracies, both Australia and the United States employ single-member districts for their lower-house elections. Drawing district boundaries is not politically innocent, but measuring the partisan bias or partisan asymmetry of a districting plan is not a settled matter and controversial when litigation and the constitutionality of a districting plan is being contemplated. My work in this domain spans multiple decades, including work tracking the use of malapportionment as a tool of partisan, electoral manipulation in Australian jurisdictions. In recent years my work has crossed from the academy into the courtroom, with my analyses of districting plans in American Congressional and state legislative cases serving as the main, evidentiary component in two trials, delivering the first verdicts for plaintiffs bringing partisan gerrymandering claims in decades. My analysis and testimony (a) compared the performance of different metrics of partisan advantage encoded in district boundaries, (b) compared the magnitude of partisan gerrymandering in the plans being litigated to hundreds of districting plans used in American jurisdictions (c) estimated the durability of partisan advantage in these districting plans, distinguishing features of the plans per se from transitory, election-specific factors; (d) proposed thresholds and standards for courts, litigants and map-drawers. This work is a vivid example of rigorous, empirical, social-science serving the cause of democracy, where data and analysis help us better understand and hence strengthen democratic institutions.

The Academy of the Social Sciences in Australia (Fellow)

American Academy of Arts and Sciences (Fellow)

Society for Political Methodology (Fellow)

1. Bayesian Analysis for the Social Sciences. Wiley. Hoboken, New Jersey. 2009.
2. with Bradley T. Spahn. ‘‘Why Does the ANES Overestimate Voter Turnout?’’ Political Analysis. 2018. To appear. Pre-print.
3. ‘‘The Predictive Power of Uniform Swing’’ PS: Political Science and Politics. 2014. 47(2):317-321.
4. ‘‘The Campaign that Wasn’t: Tracking Public Opinion over the 44th Parliament and the 2016 Election Campaign.’’ (with Luke Mansillo). Anika Gauja, Peter Chen, Jennifer Curtin and Juliet Pietsch (eds). 2018. Double Disillusion: the 2018 Australian Federal Election. Australian National University Press: Canberra. pp133-156. doi.org/10.22459/DD.04.2018.06.
5. “Assessing the Current North Carolina Congressional Districting Plan” Expert report filed in support of plaintiffs in Common Cause and League of Women Voters vs Rucho, Case 1:16-cv-01026-WO-JEP (US Federal Court) 2017.