This item is under embargo and not available online per the author's request. For access information, please visit

Date of Award

Spring 5-15-2016

Author's School

Graduate School of Arts and Sciences

Author's Department

Political Science

Degree Name

Doctor of Philosophy (PhD)

Degree Type



This project applies the methods of conceptual and normative political theory to the tools and findings of social science research, and vice versa. It asks, ``How can political theory speak to the methods of social science, and how can they, in turn, inform research in political theory?'' The dissertation consists of two parts with two chapters each, along with introductory and concluding chapters. The first part asks how the methods and findings of social science might inform normative political theory. The second asks how normative theory is, and how it ought to be, implicated in social science research.

The first chapter focuses on models, which while ubiquitous in the social sciences, are at the same time undertheorized from a broader perspective. Social scientists tend to think of models as precise mathematical or symbolic formalizations. But if we broaden this picture and think of a model in a more general way -- as an idealized and abstract device for representing theories, concepts, and observations about relationships, mechanisms, and phenomena in the world -- then not only empiricists and formal theorists, but also normative theorists rely on them as central devices in their work. Indeed political theory is awash with the same sort of simplifying abstractions social scientists employ, except these abstractions are driving models meant to represent and clarify our moral and conceptual intuitions. Thus the action in both fields is on the side of which things we ought to abstract away, and which to put into a given model. This suggests that there is a deeper methodological unity across the fields of political science than is commonly thought.

The second chapter considers the debate on ideal and non-ideal theory by applying the modelling framework for theory developed in chapter one. Once we recognize that normative arguments involve abstract or idealized models, it becomes clear that theory is never fully ideal and never fully nonideal. Instead, normative arguments vary in the form and extent of their idealizations, much as formal and empirical models do. Theorists abstract away some facts about the world and leave others in. What matters is which idealizations should be included and which facts should be abstracted away. I show that arguments attempting to justify different idealizations are both crucial to normative theorizing and badly underdeveloped.

The force of a given model hangs on the attached -- too often merely implied -- argument that those things the model abstracts are occluding irrelevancies to our thinking about the subject at hand. Good models are good analogies, not merely coherent stories. My argument shows that well-justified idealizations are those that elicit normative models of higher-order intuitions and commitments -- insights about justice as it might be -- which in turn inform arguments we make through less abstract and idealized theory. This supports a principled commitment to a fine-grained division of labor in normative theory. Arguments at all levels of abstraction are in principle important for the collective attempt to theorize what justice is, and what it recommends in the world as we find it.

In the third chapter I illustrate how some of the insights of the first part can be put into practice through the application of normative analysis to empirical social scientific research, namely to the role significance standards play in quantitative empirical research in determining what it is we infer about the world. I begin with a brief, minimally technical explanation of hypothesis testing in statistics, with particular attention to the role of critical levels of p and confidence intervals. I then offer a genealogy that traces the emergence of, and offers an account of the staying power of, our received standards ($p

The fourth chapter considers normative issues in the conduct of research itself, with a focus on political field experiments. Between a potential criminal case in Montana and the Green-LaCour retraction, field experiments have had a troubled recent history in the public eye. To a degree this may be unavoidable; experimental interventions in democratic political processes are, at best, morally fraught. What I ask in this chapter is what makes them ethical or unethical. My answer, following on and responding to recent work by Ryan Davis and Eric Beerbohm, centers on the concept of manipulation. The first problem is that most studies of manipulation are concerned with its narrowly interpersonal ethical dimensions, rather that its political contexts. I employ conceptual analysis to develop an account of manipulation suitable for the analysis of manipulative interventions in democratic decisionmaking, distinguishing between political agent manipulation, and polity manipulation. I then make the case that field experiments that engage in or threaten the latter -- they attempt to manipulate legitimate political decisionmaking -- are significantly unethical interventions in what should be autonomous processes. This conclusion issues a strong (though defeasible) duty to refrain from manipulative political interference which should trouble scholars involved with such designs.


English (en)

Chair and Committee

Frank N. Lovett

Committee Members

Randy Calvert, Clarissa Hayward, Ian MacMullen, Kit Wellman,


Permanent URL:

Available for download on Friday, May 15, 2116