Current Projects and Working Papers

All titles are preliminary and comments are welcome.

Mouse over titles for abstracts when available.

Applied Statistics

Randomization Inference and/or Matching for Causal and Statistical Inference

"The randomization mode of statistical inference" A perspective on randomization inference in experiments with a methodological proposal for the use of regression in adjustment for experiments. (with Costas Panagopoulos).

Making Effects Manifest in Randomized Experiments Experimentalists desire precise estimates of treatment effects and nearly always care about how treatment effects may differ across subgroups. After data collection, concern may focus on random imbalance between treatment groups on substantively important variables. Pursuit of these three goals --- enhanced precision, understanding treatment effect heterogeneity, and imbalance adjustment --- requires background information about experimental units. For example, one may group similar observations on the basis of such variables and then assign treatment within those blocks. Use of covariates after data have been collected raises extra concerns and requires special justification. For example standard regression tables only approximate the statistical inference that experimentalists desire. The standard linear model may also mislead via extrapolation. After providing some general background about how covariates may, in principle, enable pursuit of precision and statistical adjustment, this paper presents two alternative approaches to covariance adjustment: one using modern matching techniques and another using the linear model --- both use randomization as the basis for statistical inference. (to appear in the Cambridge Handbook of Experimental Political Science). The compendium, or reproduction archive, for the paper is available as a package for the R statistical computing environment. See this little tutorial file for detailed instructions about how to download and use it.

"Probability of What?": A Randomization-based Method for Hypothesis Tests and Confidence Intervals about Treatment Effects This drafty working paper provides my perspective on randomization-based inference for randomized experiments. (with Costas Panagopoulos).

Attributing Effects to A Cluster Randomized Get-Out-The-Vote Campaign. (with Ben Hansen ). University of Michigan, Dept of Statistics Technical Report # 448. October 2006. (pdf version here)

Attributing Effects to A Cluster Randomized Get-Out-The-Vote Campaign: An Application of Randomization Inference Using Full Matching Statistical analysis requires a probability model: commonly, a model for the dependence of outcomes Y on confounders X and a potentially causal variable Z. When the goal of the analysis is to infer Z’s effects on Y, this requirement introduces an element of circularity: in order to decide how Z affects Y, the analyst first determines, speculatively, the manner of Y ’s dependence on Z and other variables. This paper takes a statistical perspective that avoids such cir- cles, permitting analysis of Z’s effects on Y even as the statistician remains entirely agnostic about the conditional distribution of Y given X and Z, or perhaps even denies that such a distribution exists. Our assumptions instead pertain to the conditional distribution Z|X, and the role of speculation in set- tling them is reduced by the existence of random assignment of Z in a field experiment as well as by poststratification, testing for overt bias before accept- ing a poststratification, and optimal full matching. Such beginnings pave the way for “randomization inference”, an approach which, despite a long history in the analysis of designed experiments, is relatively new to political science and to other fields in which experimental data are rarely available.

The approach applies to both experiments and observational studies. We illustrate this by applying it to analyze A. Gerber and D. Green’s New Haven Vote 98 campaign. Conceived as both a get-out-the-vote campaign and a field experiment in political participation, the study assigned households to treat- ment and desired to estimate the effect of treatment on the individuals nested within the households. We estimate the number of voters who would not have voted had the campaign not prompted them to — that is, the total number of votes attributable to the interventions of the campaigners — while taking into account the non-independence of observations within households, non-random compliance, and missing responses. Both our statistical inferences about these attributable effects and the stratification and matching that precede them rely on quite recent developments from statistics; our matching, in particular, has novel features of potentially wide applicability. Our broad findings resemble those of the original analysis by Gerber and Green (2000).
(with Ben Hansen ) prepared for presentation at the Political Methodology meetings, July 2005

Attributing Effects to a Get-Out-The-Vote Campaign Using Full Matching and Randomization Inference (with Ben Hansen ) prepared for presentation at the MPSA meetings, April 2005

Miscellaneous Fun Stuff

Cycling Involvements: Frequency Domain Time Series Analysis and Political Participation in the USA This paper shows that decomposing a time-series into periodic components can provide po- litically useful information about the shape of aggregate political participation in the United States. Specifically, it provides statistical tests for the periodicity of the aggregate time series of political participation and explains how this decomposition and associated tests work. Between 1973 and 1994 there appears to be an annual cycle in the reporting of political participation by respondents to a series of polls conducted by Gallup 10 times per year. This seasonality has been noted by in one other publication, by Rosenstone and Hansen (1993), but was explained as tied to a summer political cycle. In this article I suggest that this discovery has more to do with annual cycles in the composition of the Gallup sample than politics. I am currently trying to obtain detailed information on the monthly mail volume into and out of Congress. With this information, I will be able to test more directly if, despite the changes in sample composition of the Gallup polls, the political participation of Americans ought to be see as an "output" of Congressional mobilization or an "input" or in what way the flow of participation into Congress is related to the flow of mobilization out of it. Very drafty and still waiting for data on monthly mail to and from Congress. This paper contains a basic description of some frequency domain time series analysis as applied to a political science topic.

Political Behavior

The Shape of Political Participation Although 50 years of excellent scholarship have taught us a great deal about how political participation varies across individuals within one point in time, scholars do not know much about how political participation changes over time within the lives of individuals. By focusing predominantly on the preconditions for participation, the literature has largely ignored the precipitants of it. In this paper, I endeavor to show what political participation looks like if we think of it as a process evolving year-by-year across the lives of ordinary people. The new description offered here provides some evidence that challenges the basis for extant theories of why individuals participate in politics. The purpose of this paper is not to offer new theories or frameworks for understanding, but merely to offer a new vision of what political participation is; a vision which differs from and, I hope, complements that currently assumed by scholars in this field; a vision that, I hope, spurs new theories and new modes of research in this area.

A Framework for Studying the Dynamics of Political Participation. Political action is driven by events. Although the effects of particularly dramatic events on social movements is well documented, the effects of events, quotidian or exceptional, on the behavior of individuals are significantly less well understood. This paper proposes a framework for understanding how a moment of political action may occur in the life of an ordinary person. It synthesizes past literature and theories that explain variation among people at a single point in time on the basis of largely time-constant attributes of people and elaborates on this literature to suggest when we might expect the poor and disadvantaged to surmount such resource, skills, and status barriers to get involved in politics. Furthermore, this framework suggests a way for future syntheses, theory-building, and empirical studies to coordinate such that all of our disparate findings about political participation cumulate more effectively. An Appendix for this working paper.

Civic engagement of the World War II generation. (with Suzanne Mettler and Theda Skocpol)

Words to Math: Representing Theoretical Concepts Mathematically --- The Case of "Symbolic Racism" (with Cara Wong )

The True Story of the Literary Digest Poll of 1936 (with Adam Berinsky )

Threat, Mobilization, and Participation: The Impact of Crossburnings on Political Behavior in North Carolina

Black Threat, Christian Threat, Black Context, Christian Context

Unpacking Power Threat Theory (with Cara Wong )

top

Published Papers

top

Journal Articles

Attributing Effects to a Cluster Randomized Get-Out-The-Vote Campaign (with Ben Hansen ) Journal of the American Statistical Association 2009.
The compendium, or reproduction archive, for the paper has been published in my dataverse ( or feel free to use a direct link to the study)
(For more about reproduction archives in general see Gentleman, R and D. Lang. 2004. "Statistical Analyses and Reproducible Research." )

Covariate Balance in Simple, Stratified and Clustered Comparative Studies (with Ben Hansen ) Statistical Science 2008, Vol. 23, No. 2, 219-236. A typo on page 4 exists such that the variance of d should be Var(d) = (m/(mtmc))s2.

Politics across Generations: Family Transmission Reexamined (with Kent Jennings and Laura Stoker). The Journal of Politics 2009, Vol 71, No. 3, 782-799.

EDA for HLM: Visualization when Probabilistic Inference Fails (with Katherine Drake). 2005. Political Analysis . August 2005 Advance Access Online. in print Fall 2005. The Sweave source file and datasets used to produce this article have been published in my dataverse ( or feel free to use a direct link to the study)

"Using R to Keep it Simple: Exploring Structure in Multilevel Datasets" Fall 2004. The Political Methodologist. The Sweave file and datasets used to produce this article can be downloaded as a zipped archive here

Analyzing the 2000 National Election Study (with Nancy Burns, Michael Ensley, and Don Kinder). 2005. Political Analysis . 13(1):109-111.

"Does Moving Disrupt Campaign Activity?" August 2004. Political Psychology. 21(4):525-543.

"Designing Multi-level Studies: Sampling Voters and Electoral Contexts" (with Laura Stoker). 2002. Electoral Studies 21(2):235-267. This file includes the correct Figure 3 (published as an Erratum in the next issue of Electoral Studies).
(Figure 1 in color -- PDF, about .3MB)
Supplementary results and simulation programs are available on this page .

"Issues in Analyzing Data from the Dual-Mode 2000 American National Election Study" (with Michael Ensley). NES Technical Report #64. (April 2003).

"Black Threat and Christian Fundamentalist Threat: A National Election Study 1997 Pilot Study Report" 1997 NES Pilot Study Report

NES Pilot Study Efforts to Measure Values and Predispositions. NES Technical Report #18. (February 1995)

top

Dissertation

An overview and summary of my dissertation here
top

Some of the material on this page is based upon work supported by the National Science Foundation under Grant Numbers SES-0753164 and SES-0753168 and previously by a National Science Foundation Graduate Research Fellowship 1994--1999. Any opinions, findings and conclusions or recomendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation (NSF).