TRENDS: Toward a Separate Ethics of Political Field Experiments

• Date: 01 September 2019
• DOI: 10.1177/1065912919835970
• Published date: 01 September 2019
• Author: Gregory Whitfield
• Subject Matter: Articles

835970PRQXXX10.1177/1065912919835970Political Research QuarterlyWhitfield
research-article2019
Article
Political Research Quarterly
2019, Vol. 72(3) 527 –538
TRENDS: Toward a Separate Ethics of
© 2019 University of Utah
Article reuse guidelines:
sagepub.com/journals-permissions
Political Field Experiments
https://doi.org/10.1177/1065912919835970
DOI: 10.1177/1065912919835970
journals.sagepub.com/home/prq
Gregory Whitfield1
Abstract
In this article, I develop a critical view of the development and state of research ethics in political science. The central
problem is that political scientists have inappropriately followed the lead of clinical biomedical research ethics in
thinking about their own designs. Specifically I argue that the focus on institutional and group decision-making contexts
distinctive to political research presents normative problems not well-addressed by clinical biomedical approaches.
First, I make the case that research ethics as it has been conceived won’t capture all that might be wrong in political
research designs because some of the potential harms/wrongs will be to political norms and institutions and thus will
violate political (although not individual ethical) rights/values/and so on. Second, I rebut the challenge that principles of
justice and equipoise standard to biomedical research ethics might be suitable for political research. And third, I argue
that political theorists and philosophers must involve themselves in empirical political science research ethics if we are
to effectively communicate the stakes of these research designs to practitioners, consumers, funders, and editors who
remain steeped in the norms of biomedical research ethics.
Keywords
research ethics, field experiments, democracy, legitimacy
Introduction
new set of ethical challenges requiring us to think care-
fully about the values implicated by this new and dynamic
Quantitative political science has long been well-described
set of practices.
as the analysis of data gathered through observation.
We might think that the ethics of these treatments are
Political scientists collect such data without any inten-
straightforward—that research ethics boards, charged
tional or significant intervention in the lives or institu-
with validating the ethics of complicated clinical trials
tional contexts of subjects. Ethically, these methods for
that directly affect the medical and psychological well-
data collection are for the most part either trivial or already
being of subjects, are well-equipped to assess these less
well addressed (e.g., Michalos 1991; Murphy and invasive and lower-stakes political studies. We might
Dingwall 2001; Parker 2007). But the discipline has
even think that the impacts of political science experi-
recently seen a significant move toward nonobservational
ments do not require the sort of regulation that medicine
research. Laboratory and especially field experiments are
does—that free enquiry trumps because the stakes are
now an important and growing part of the discipline
low. But I argue that this position gets the situation back-
(Druckman et al., 2011; Morton and Williams 2010; Teele
ward. The stakes should not appear so low when we take
2014). These studies are different in kind from observa-
seriously the values impacted. Political field experiments
tional work in that they involve direct interventions in
risk significant effects to political behaviors and institu-
people’s lives: researchers are actively treating subjects in
tional orders that have normative standing that biomedi-
some way in order to measure their responses, in field
cal principles miss. Exclusive employment to the values
contexts often without subjects’ prior knowledge.
Scientifically, experiments are extraordinarily valuable,
the logic of control group testing allowing scholars to
1University College London, UK
identify causal relationships with a confidence missing
from other more conjectural methods. But at the same
Corresponding Author:
Gregory Whitfield, Politics & International Relations, University of
time, these methods affect subjects in ways political scien-
Edinburgh, 21 George Square, Edinburgh EH8 9LF, UK.
tists rarely have in the past. We are thus presented with a
Email: g.whitfield@ed.ac.uk

---

528
Political Research Quarterly 72(3)
and principles developed for biomedical clinical research
merely ethical—impacts of these experiments are worth
leaves researchers, review boards, and disciplinary thinking about. But Desposato’s empirical approach to
authorities without the language to talk about the ethical
ethics points to intuitions about values without reflec-
weight of these effects. As political scientists we must
tively developing principles for taking them seriously. In
attend also to these distinctively political dimensions if
order to go beyond self-protection triangulated using
we are to have confidence in the permissibility of our
public beliefs about experiments, political scientists
field experiments.1
need systematic normative theorizing about experiments.
To that end, I here seek to go some ways toward an
We need supplementary principles that organize relevant
ethical accounting of new political science methods. My
political values in order to be confident that we take seri-
goal is the framing of a research agenda on the ethics of
ously the ethical dimensions of our new methods. I
political science methods using treatments (either con-
believe experiments can meet these value challenges,
sented to beforehand or not) on subjects, where those
and formulating those principles is the goal of the broad
subjects can be taken to include both individual citizens
research agenda I frame in this article.
and officials, as well as political institutions like courts,
The article proceeds in four sections. First, I make the
legislatures, and the like. That agenda broadens the exist-
case that research ethics as it has been conceived won’t
ing literature on the ethics of political science experimen-
capture all that might be wrong in political research
tal methods by challenging the dominance of value
designs because some of their potential harms/wrongs will
frameworks imported from clinical and biomedical be to political norms and institutions and thus will violate
research contexts and by showing that these frameworks
political (although not ethical) values. As a corollary, I
are insufficient in capturing the range of values impli-
argue that political theorists and philosophers must
cated in political science research.
involve themselves in this sort of work if we are to effec-
The central problem with much prior work on political
tively communicate the stakes of these research designs to
science research ethics is its uncritical acceptance of the
practitioners, consumers, funders, and editors who remain
existing principles guiding research ethics. Dawn Langan
steeped in the norms of biomedical research ethics.
Teele’s chapter on experimental research ethics is a recent
Which norms in particular is a question I address in
example. She discusses researchers’ ethical duties toward
the second section. Though there may be others also
their subjects using the Belmont Report—an important
worth considering, I discuss two norms in biomedical
document in setting out ethical wrongs to clinical sub-
research ethics that tempt social scientists with ready-
jects in research contexts—with no discussion of further
made analytical frameworks but fail to address the rele-
values or principles appropriate to our discipline (Teele
vant normative challenges of their research designs. First,
2014, ch5). Similarly, the editorial team for the Journal of
I address the conception of harms/wrongs in the theories
Experimental Political Science, writing in the inaugural
of justice that bioethicists employ in thinking about
volume of that journal, endorse the U.S. federal govern-
human subject research. While they employ a conception
ment’s Common Rule as the best available basis for
that takes seriously individual harms and certain diffuse
addressing ethical issues in experimental research harms, I argue this conception is insufficient for political
(Morton and Tucker 2014, 102).2
contexts because it will fail to account for distinctively
Now I have no doubt that these frameworks are useful
political values related to political legitimacy and collec-
and well-established bases for thinking about the ethics
tive decision making.
of political science experiments. Indeed their application
The second norm I’ll discuss is the biomedical research
throughout Scott Desposato’s (2015) volume on the sub-
principle of clinical equipoise. This principle restricts
ject helpfully frames issues ranging from local review to
biomedical researchers to conducting clinical trials where
religion. But their as-is application threatens to obscure
subjects in each arm of the trial (receiving the experimen-
many of the most significant issues unique to political
tal treatment under study or receiving some comparison/
research. Desposato’s recent article on subjects’ and
control class of treatment) must be equally likely to
scholars’ ethical beliefs about experiments goes further
achieve positive outcomes—that is, we cannot give the
to thinking about the political ethics of these experi-
control group a sugar pill if we have evidence that the
ments when he addresses the...