Why Do People Accept Public Policies They Oppose? Testing Legitimacy Theory with a Survey-Based Experiment

• Date: 01 June 2005
• Author: James L. Gibson,Gregory A. Caldeira,Lester Kenyatta Spence
• Published date: 01 June 2005
• DOI: 10.1177/106591290505800201
• Subject Matter: Articles

Why Do People Accept Public Policies They Oppose?
Testing Legitimacy Theory with a
Survey-Based Experiment
JAMES L. GIBSON, WASHINGTON UNIVERSITY IN ST. LOUIS
GREGORY A. CALDEIRA, OHIO STATE UNIVERSITY
LESTER KENYATTA SPENCE, WASHINGTON UNIVERSITY IN ST. LOUIS
The orthodox answer to the question posed in the title of this article is that the legitimacy of institutions has
something to do with acquiescence to unwelcome public policy decisions. We investigate that conventional
wisdom using an experiment embedded within a representative national sample in the United States. We test
hypotheses concerning not only the effect of institutional legitimacy on acquiescence, but also the influence of
partisanship, the rule of law, and simple instrumentalism on willingness to accept an objectionable policy deci-
sion. Our analyses reveal that legitimacy does matter for acquiescence, and that the Supreme Court is more
effective at converting its legitimacy into acceptance than is Congress. Yet, many important puzzles emerge
from the data (e.g., partisanship is not influential), so we conclude that Legitimacy Theory still requires much
additional empirical inquiry.
Over the past decade or so, scholars of the judicial its “political capital” and worry about the efficacy of the insti-
process have devoted considerable attention to
tution in future clashes, but the Court seems to have enough
studying the so-called legitimacy-conferring capac-
legitimacy to “get away with” its ruling and make it stick.
ity of courts. The most common research design involves
This conflict will no doubt enter our textbooks as a proto-
presenting people with a policy with which they disagree
typical example of the efficacy of judicial legitimacy.
and then trying to ascertain whether a court ruling increases
There is of course more to Bush v. Gore than this simple
the probability of acceptance, acquiescence, and compli-
caricature (for a more detailed analysis see Gibson, Caldeira,
ance. The evidence seems to be that courts have the power
and Spence 2003). Nonetheless, the dispute provides a nat-
to “legitimize” some types of decisions, under some circum-
ural context within which to reexamine the question of judi-
stances (Gibson 1989; Gibson and Caldeira 2003).
cial legitimacy. That is our purpose in this article. In partic-
This process seems to be well represented by the presi-
ular, we report the results of a national survey conducted in
dential election of 2000 and the decision of the Supreme
January 2001 on the legitimacy of the Supreme Court and
Court in Bush v. Gore (121 S Ct 525 (2000)). A simple view
its ability to generate acceptance of an unpopular Court
of the connection between this case and extant theory is easy
ruling. Our survey included an experimental vignette, based
to construct. A bitter political controversy arises. The dispute
on a hypothetical scenario. Although hypothetical, the sce-
bounces around various institutions, with no definitive reso-
nario was not fanciful; instead, at the time of the interview,
lution. Finally, the Supreme Court intervenes and makes a
it was highly plausible since it posited that the Supreme
decision. Some grumble about the Court’s opinion, but polit-
Court or Congress makes a decision in a dispute over count-
ical elites call for the ruling to be “respected,” much if not
ing the ballots in Florida after the presidential election had
most of the mass public accepts the outcome, and the con-
ended.1 We take advantage of the plausibility of this sce-
troversy subsides. Scholars then talk of the Court expending
nario to test a series of theoretically informed hypotheses
about institutional legitimacy and acquiescence. Specifically,
we consider whether the Court has more power than Con-
gress to generate acceptance, whether the processes and
NOTE: This article is a revised version of a paper delivered at the Confer-
methods of institutional decisionmaking contribute to citi-
ence on Experimental Methods, The Center for Basic Research in
the Social Sciences, Harvard University, Cambridge, MA, May 11-
zens’ willingness to acquiesce to a decision, and whether
12, 2001. This project would not have been possible without the
legitimacy has any efficacy when juxtaposed against strong
support of the Weidenbaum Center on the Economy, Govern-
instrumental considerations. This article is not an analysis of
ment, and Public Policy, Washington University in St. Louis, and
Bush v. Gore. It does, however, use the election dispute to
The Ford Foundation (Grant Number 1015-0840). We are espe-
reconsider the essential nature of judicial legitimacy and its
cially indebted to Steve Smith, Director of the Weidenbaum
Center, for his support of this work. We appreciate the comments
consequences for politics and the political system.
of Steve Ansolabehere and Jim Alt on an earlier version of this arti-
cle. We are also thankful for the research assistance of Christine
1
Lamberson.
The realism of the scenario presented to the respondents imposes some
important limits on the generalizability of these findings. We address this
Political Research Quarterly, Vol. 58, No. 2 (June 2005): pp. 187-201
issue in the concluding section of this article.
187

---

188
POLITICAL RESEARCH QUARTERLY
Our findings are at once simple and complicated. For
of political institutions at bringing about compliance with or
instance, we do indeed discover that the Supreme Court is
acceptance of their unpopular decisions. Consequently, the
better able than Congress to translate its legitimacy into
experiment begins by establishing an “objection precondi-
acquiescence, and that the Court has more institutional
tion.” The respondents were first asked: “Some people want
legitimacy than Congress. The Court profits especially when
all of the ballots in Florida to be destroyed to eliminate
it grounds its decisions in law; Congress is more effective
doubts about which candidate won the presidential election
when it bases policy on fairness rather than law. In general,
in Florida. Others want the ballots to be available to be
decisionmaking procedures have few consequences for
recounted. What is your view—should the ballots be
acquiescence, while instrumental calculations clearly con-
destroyed or made available to be recounted?”5 Those who
tribute to willingness to accept an unfavorable and unde-
favored destroying the ballots were told a vignette that
sired outcome. Although many puzzles remain, our analysis
began: “Suppose that later this year, a group of citizens
points to the crucial importance of legitimacy when politi-
comes forward demanding that a special governmental
cal institutions are called upon to make highly controversial
panel be established to take custody of the ballots from
and divisive policy decisions.
Florida and make them available to be counted” and con-
cluded that: “The ballots are then made available to be
A THEORETICAL FRAMEWORK FOR
counted.” Contrariwise, those who initially favored count-
ANALYZING ACQUIESCENCE
ing the ballots were told: “Suppose that later this year, a
group of citizens comes forward demanding that a special
Why do people accept institutional decisions with which
governmental panel be established to take custody of the
they disagree? We consider three broad sets of determinants
ballots from Florida and prevent them from being counted.
of willingness of citizens to “accept” a court decision,2 to
. . . The ballots are then impounded so that they cannot be
“respect” it, to cease opposition and get on with politics: (1)
counted.” Thus, the conclusion in the vignette is disagree-
institutional legitimacy; (2) procedural assessments; and (3)
able to each of the respondents, thereby satisfying the objec-
a major rival hypothesis to Legitimacy Theory, instrumental
tion precondition necessary for testing for the effects of
determinants of compliance. We represented each of these
legitimacy. A slim majority (50.0 percent) supported count-
factors in an experimental vignette presented to a random
ing the ballots; 35.5 percent opposed doing so.6
sample of the American mass public (see Appendix A for the
Four manipulations were incorporated into the vignette.
survey details).
Table 1 summarizes the structure of the experiment, as well
The Legitimacy Vignette employed four manipulations,
as the individual hypotheses we test.7
in a 2222 fully crossed factorial design. This portion
of the interview began with a short story about the demand
Institutional Legitimacy
of a group of citizens for a governmental panel to take cus-
tody of the ballots from Florida. In the stories, four dichoto-
Diffuse support—a synonym for “legitimacy”—refers to “a
mous characteristics were manipulated, resulting in sixteen
reservoir of favorable attitudes or good will that helps mem-
versions of the vignette.3 Embedding an experiment within
bers to accept or tolerate outputs to which they are opposed
a nationally representative survey yields the substantial
or the effects of which they see as damaging to their wants”
advantages of both internal and external validity.4
(Easton 1965: 273). The most useful way to conceptualize
The purpose of this experiment is to assess hypotheses
diffuse support, we argue, is to think of it as institutional
concerning the circumstances contributing to the effectiveness
loyalty—support not contingent upon satisfaction with the
immediate outputs of the institution. Easton’s apt phrase—
2
a “reservoir of good will”—captures the idea that...