In my studies as a doctoral student of political science at the University of Wisconsin-Madison I have focused on three areas of research:
- Local electoral environments
- Statistical modeling in the social sciences
- Applied policy analysis
Additionally, I am currently writing a dissertation examining the response of voters, candidates, and elected officials in school districts and on school boards to the passage of Wisconsin Act 10 and the 2011-13 biennial budget, which greatly changed the relationship between school boards and their employees, particularly unionized employees.
I seek out projects that help policy-makers better understand the mechanisms and outcomes of their decisions. I have especially focused on moving beyond the point-estimate to explore heterogeneity in parameters of interest and attempt to understand why such variation exists in the impact of a policy or program, and how future work might support this with evidence. In policy making decisions, effect heterogeneity is an important part of the benefit-cost calculation.
My dissertation explores the question of whether or not large and salient policy shocks at the state level can cause voter, candidate, and board member behavior to change in school districts. Previous research has indicated that board elections are sleepy affairs with few candidates, few choices, and few voters--except in exceptional cases. This dissertation will leverage recent political events in the state of Wisconsin to evaluate if a state level policy shock can increase political activity in local elections of school boards. The project leverages a number of unique datasets to paint a more comprehensive picture of school district politics including election results for school boards for all school districts in Wisconsin, budgetary decisions made by school boards, partisanship of school district electorates, and changes in the preferences of the school district electorate in each of Wisconsin's school districts.
For more information you can read the dissertation proposal abstract here.
And, you can read the full proposal here.
Local electoral environments
In addition to my dissertation, I have done work on the conditions that influence school district property tax levy decisions. In Wisconsin districts operate under strict revenue limits which set a ceiling of spending. My work looks at the financial, political, and social conditions that lead districts to raise less than the maximum allowed from local property taxes (under levy) or to exceed it by going to a referenda.
To date this work has employed a number of administrative record datasets and approaches to statistical modeling including two-stage least squares, selection models, Bayesian binary response models, and latent growth curves.
- A working paper: Understanding Property Tax Levies in an Era of Reduced State and Federal School Aid
- A conference presentation at Association of Education Finance and Policy (AEFP) and the National Tax Association (NTA): The Propspects of School Property Tax Revenue in the Wake of the 'Great Recession'
- A conference paper at AEFP: "The Political Economy of Voter Support for School Property Taxation"
- Two reports to the Wisconsin Association of School Boards
- Andrew Reschovsky, UW-Madison Emeritus Professor
- Lindsay Amiel, UW-Madison
Statistical modeling in the social sciences
A key aspect of my research has been identifying the best tools for identifying empirical support for social science questions. My coursework experience includes rigorous training in linear models and many derivatives of linear modeling such as multi-level modeling, hierarchical linear modeling, generalized linear models, and advanced econometrics estimators like panel, geospatial, and geographically weighted regressions. I have focused on gaining as much training as possible in order to identify the appropriate solution to the data and the question at hand.
The most important aspect of this training is understanding the limitations of the methods above both in theory and in practice. Understanding how causal inference is aided or hindered by methods such as these is crucial.
- A working paper on geospatial statistical models in political science
- A working paper on the importance of treatment fidelity in understanding effect heterogeneity
Applied policy analysis
Applied policy analysis that is rigorous, reliable, and informs policy requires a bit of opportunism and a bit of luck. Leveraging the state longitudinal data system SLDS at the Wisconsin Department of Public Instruction has allowed some programs and policies to be evaluated with more causal rigor than was previously possible. Having access to both the training in causal inference, policy makers with pressing questions, and Wisconsin's educational administrative records has allowed me to explore the ability of policy analysis to provide sufficiently useful and rigorous evaluations of program effectiveness.
To date this work has centered around state policy in Wisconsin surrounding English-Language Learners (ELL), students in K-12 public schools for whom English is not their first language. Leveraging student fixed-effects models, value-added models, regression discontinuity designs, and more recently Bayesian multilevel models, I have explored the impact that being offered certain instructional programs have on future student outcomes, the effect of exiting ELL specific services on academic achievement, and the degree of variation in these impacts across different contexts within the state.
- Two presentations to stakeholders at DPI
- A report to the DPI policy and budget team
- Under review at AERJ: Is Bilingual Better? Using Statewide Longitudinal Data to Analyze the Effect of Bilingual-Bilcultural Education Programs on English Language Profciencyand Reading and Mathematics Achievement
- Under review at JPAM: English Language Learner Reclassification Policy: Patterns and Effects
- A publicly available presentation: Bilingual-Bicultural Programs in Wisconsin
- Deven Carlson, Assistant Professor at University of Oklahoma
- Sara Dahill-Brown, Assistant Professor at Wake Forest
- Dominique Bradley, UW-Madison