On March 19, 2003, the United States declared war on Iraq. More than
70 percent of Americans supported the war. The question is why. According
to most theories of public opinion, support for this war should have been
extremely low. Many casualties were expected; most Americans anticipated a
long conflict, detrimental to the U.S. economy; and most Americans thought
war in Iraq would increase the likelihood of terrorist activities in the
United States. Large antiwar demonstrations broke out in Los Angeles, New
York, Berlin, London, Paris, and across the Middle East. The United
Nations did not authorize the action; on the contrary Russia and France,
permanent members of the Security Council, threatened to veto any measure
allowing force. Yet support for the war was high among both Republicans
and Democrats, both men and women. It was high regardless of whether or
not weapons of mass destruction (WMD) were found in Iraq, whether or not
thousands of American soldiers died, and whether or not the war continued
for more than a year. To be sure, we expected that patriotism would
generate support for the war; but previous academic accounts of the
rally-around-the-flag effect suggest that while it might boost
support,1
it could not
by itself explain why approval was as high as it was.
We suggest that the principal reason that three-quarters of the
American public supported the war was that the Bush administration
successfully convinced them that a link existed between Saddam Hussein and
terrorism generally, and between Saddam Hussein and al Qaeda specifically.
Framing the war on Iraq in this way connected it intimately with 9/11,
leading to levels of support for this war that stretched nearly as high as
the levels of support for the war in Afghanistan.2
ABC News/Washington Post Afghanistan Attack Poll
2.
President Bush never publicly blamed Saddam Hussein or Iraq for the
events of September 11, but by consistently linking Iraq with terrorism
and al Qaeda he provided the context from which such a connection could be
made. Bush also never publicly connected Saddam Hussein to Osama bin
Laden, the leader of al Qaeda.3
In fact,
mention of bin Laden dropped completely out of speeches after February
2002 and only appeared in one speech prior to that, in the September 2001
address to a joint session of Congress. White House
2001; White House 2002a.
Nevertheless, whether or not Bush connected each dot from Saddam Hussein
to bin Laden, the way language and transitions are shaped in his official
speeches almost compelled listeners to infer a connection.
First, we briefly outline the relevant academic theories explaining
American public opinion on military conflict and the role the press plays
in shaping public opinion in times of war. Next we present our
Iraq-as-war-on-terror4
We use the term
war on terror to facilitate our analysis of the president's
language; in no way do we intend to judge the legitimacy of the
terminology.
hypothesis, investigating the administration's
framing of the war through analysis of George W. Bush's speeches from
September 11, 2001, to May 1, 2003. We analyze polling data to track the
public's response to the administration's rhetoric. Before
discussing the implications of our findings, we consider alternative
explanations for the extraordinary public support and suggest why they
give us less leverage on the question.
Domestic Public Opinion in Times of Foreign
Conflict
Researchers have long recognized that Americans tend to remain
uninformed about politics, especially foreign affairs.5
However, wars
can inspire the public to become better informed and dramatically increase
their news consumption.
6 Furthermore, we know that
Americans who are uninformed take cues from opinion leaders, and that
often it is the interpretation of the events by elites, rather than the
events themselves, that help shape public opinion.
7 In other words, “[T]hose who control
the language control the argument, and those who control the argument are
more likely to successfully translate belief into policy.”
8 Issue frames affect beliefs and the relative
importance individuals attach to beliefs.
9 The media also have an opportunity to shape public opinion through
tone, content manipulation, and issue frames.10
Since much of the
media coverage of any political event prominently features quotes from
political leaders and excerpts of official speeches, such media coverage
gives public officials a second venue, beyond their direct pronouncement,
from which to propagate their message, and one in which the generally
uninformed public is likely to hear, believe, and respond.
How should the public respond to presidential rhetoric? We know that
members of a president's party are more likely than the members of
the outparty to support the president's policy in any area.11
We also know that if opposition and debate occur
among elites, this is reflected in media coverage; however, when this
conflict is muted or nonexistent, a one-sided information flow emerges,
even if citizens or foreign critics hold other viewpoints.
12
This seemed to ring true even more so after September 11, when the press
acted deferentially to government officials' frames.
13 However, while Americans are generally politically uninformed, they do
have strong views on war. Americans generally respond unfavorably to
foreign conflict in which ground troops must be committed, especially when
large numbers of American casualties are expected. In addition, they tend
to prefer multilateral efforts.14
This makes the large, unconditional support for the
war in Iraq all the more surprising.
Finally, we also know that in times of war and national crisis,
Americans “rally around the flag,” increasing their approval
of the president, often dramatically.15
This
increase occurs regardless of the nature of the foreign policy event, at
least in the short term. But support for the president is not guaranteed;
nor is it long-lasting,
16
especially in crises that cost American lives.
17 We contend
that while the rally-around-the-flag phenomenon likely played a role in
support for the war in Iraq, the levels of support for this war were so
high and so largely unconditional that spontaneous patriotism alone cannot
account for it.
Iraq as War on Terror
The Bush administration successfully framed the war in Iraq as an
extension of its response to September 11 and the war on terror. The
administration juxtaposed allusions to Iraq with the terms terror, bin
Laden, and al Qaeda. Furthermore, there was little elite
opposition to this rhetoric, leaving the American public with a one-sided
flow of information. The more people watched television news about the war
in Iraq, the more they were exposed to the Bush administration's
rhetoric.18
Given that Colin Powell's
speech to the United Nations was the only time this link was explicitly
stated by a member of the Bush administration, hearing Colin Powell's
speech should increase an individual's likelihood of connecting Iraq
with al Qaeda, even controlling for general Iraq war news
consumption.
We propose that the connection between Iraq and
9/11 made the latter the most relevant consideration in the minds of
the American people when they thought abut the war in Iraq, which
increased support for the war.
The Iraq-as-war-on-terror framing results that we demonstrate below
are not merely statistical artifacts or functions of our methodology or
coding scheme; in fact, other scholarship and quotations in the media
suggest that Bush himself believed that Saddam Hussein and Osama bin Laden
were connected. As early as September 17, 2001, Bush said, “I
believe Iraq was involved,” when asked about the September 11
attacks.19
And according to Bob Woodward's account of
National Security Council meetings, from the moment the September 11
terrorist attacks occurred, the Bush administration framed the U.S.
response in terms of a global war on terrorism, with targets including not
just Afghanistan, but other nations as well, including Iraq. Thus it seems
that Bush sincerely believed in the Iraq-9/11 connection.
A Content Analysis of the Bush
Administration's Rhetoric
In order to measure Bush's rhetoric on the Iraq war, we analyzed
presidential speeches dealing with terrorism and/or Iraq given from
September 11, 2001, to May 1, 2003—the declared end of hostilities
in Iraq.20
See appendix A for a description
of the content analysis methodology.
We coded only presidential
speeches since the president is responsible for foreign policy, and other
members of the administration publicly take cues from him.
21 We excluded press conferences and interviews,
as we were interested in the polished language of speeches intended to be
persuasive. A list of the speeches is available at
www.princeton.edu/∼agershko/iraqwar.htm.
The language of Bush's speeches mattered. The language used to
describe a conflict allows the president to shape citizens'
interpretations of events and information.22
As Donald
Kinder states, “Frames seek to capture the essence of an issue. They
define what the problem is and how to think about it; often they suggest
what, if anything, should be done to remedy it.”
23 Whether the
prevailing frame for a conflict is World War II or Vietnam has real
consequences for how willing the public is to support it, which matters
since positive public opinion is an “essential domino” of
successful military operations.
24 In 2002 the terrorism frame
was available, believable, and understandable to a country scarred by
September 11, making the frame powerful and convincing.
What should we expect to see from the Bush rhetoric under the
Iraq-as-war-on-terror frame? Before President Bush's speech about
Iraq to the United Nations on September 12, 2002, we expect terrorism to
appear prominently in rhetoric and mentions of Iraq to be rare. However,
after that date, when the Bush administration began to make a case for
invading Iraq, we expect to see Iraq figure more prominently, together
with, instead of substituting for, references to terrorism.25
Strong emphasis on Iraq in presidential
speeches did not begin in earnest until September 12, 2002, which is why
we focus only on September 2002–May 2003 in this
section.
To test this, we created an “Iraq index,” which counted
how many times Bush used the terms Iraq and Saddam
Hussein, and a “terrorism index,” which counted use of
the terms terror, Afghanistan, Taliban, bin Laden, and al
Qaeda.26
We divided the Iraq index and
terrorism index by speech word count for a normalized score. However,
figure 2 presents the raw counts because they
are both substantively and visually easier to interpret.
After
September 11, 2002, the Iraq index outstrips the terrorism index in all
but one speech: the May 2003 speech ending the military phase of the
conflict (see
fig. 1). However, in early
speeches on Iraq in September and October 2002, the terrorism index
reached levels seen in November 2001, indicating that references to
terrorism appeared in the same magnitude in the build-up to Iraq as in the
months following September 11. The sustained dialogue on Iraq continued to
reference terror and terrorism at a relatively high intensity. What is
more striking than the sheer number of mentions of either Iraq or
terrorism is that the indices move in the same direction at the same time.
This confirms our expectations that terrorism and Iraq would appear
together in speeches.
Iraq and terrorism indices
However, the number of mentions incompletely assesses the rhetoric.
Thus we also coded term proximity, that is, whether the president referred
to Iraq/Saddam Hussein and terrorism in the same parts of speeches.
From September 12, 2002, to May 2003, the subjects of terrorism and Iraq
were intertwined on a regular basis. Of the 13 speeches given in this
period, 12 referenced terror and Iraq in the same paragraph and 10 placed
them within the same sentence. In 4 speeches, a discussion of terrorism
preceded the first mention of Iraq, giving the impression that Iraq was a
logical extension of the terrorism discussion.
The president's speech to the United Nations on September 12,
2002, is an example of this kind of structure. His remarks commenced with
a remembrance of the September 11 attacks, continued with a discussion of
the war in Afghanistan, and addressed the threat to world peace from
outlaw regimes and terrorists—all before focusing specifically on
Iraq, which was thus rhetorically linked with terrorism. Further in the
speech, Bush stated, “Iraq's government openly praised the
attacks of September the 11. And al Qaeda terrorists escaped from
Afghanistan are known to be in Iraq.”27
While
not explicitly stating that Iraq was involved in the terrorist attacks on
the United States, the president allowed the listener to come to that
conclusion. The administration gave no specific intelligence to verify the
claim that known terrorists were harbored in Iraq until Colin
Powell's speech four months later.
Another notable construction in Bush's speeches is the
juxtaposition of Iraq/Saddam Hussein with September 11, as seen in the
above quotation. Seven of 13 speeches from September 2002 to May 2003,
place September 11 and Iraq in the same paragraph, while four speeches
place them in the same sentence. Three times in this period, Bush speeches
proposed a hypothetical situation in which the September 11 hijackers were
armed with WMD provided by the Iraqi government. For example, in a March
8, 2003, radio address the president said, “The attacks of September
11, 2001 showed what the enemies of America did with four airplanes. We
will not wait to see what terrorists or terror states could do with
weapons of mass destruction.”28
Bush proposes that the
“enemies” of the United States are indistinguishable and
allied by the use of terror; furthermore, “terrorists” and
“terrorist states” have designs to use WMD and must be dealt
with in the same manner. This combination provided the logical basis for
attacking Iraq; the regime was linked to al Qaeda and therefore to
September 11.
The Iraq-as-war-on-terror frame is also apparent in the transitions in
the state of the union address of 2003. Following a long discussion of
domestic issues, the president turned to foreign affairs. After mentioning
the successes in the war against terrorism, the speech introduced the
topic of outlaw regimes, including the two other “axis of
evil” countries, Iran and North Korea, priming the connection of
Iraq and the fight against terrorism. Thus, there was a direct transition
from September 11, to outlaw regimes in the context of the war on terror,
to WMD, and finally, to Iraq. The speech also offered another example of
the metaphorical use of September 11 hypothetical scenario:
Before September the 11, many in the world believed that
Saddam Hussein could be contained. But chemical agents, lethal viruses and
shadowy terrorist networks are not easily contained. Imagine those 19
hijackers with other weapons and other plans—this time armed by
Saddam Hussein. It would take one vial, one canister, one crate slipped
into this country to bring a day of horror like none we have ever known.
We will do everything in our power to make sure that that day never
comes.29
This scenario occurred after discussions of the war on terror,
Iraq's failure to disarm, and after a description of Saddam Hussein
as a “brutal dictator, with a history of reckless aggression, with
ties to terrorism.”30
At the end of the address, the
president promised to bring food and medicine to Iraqi citizens in case of
war, as troops did in Afghanistan, furthering the parallels between the
two conflicts. Thus the Iraq war was framed by the Bush administration as
part of a larger war on terrorism. What other frames did the public
hear?
Wither the Opposition?
The public's acceptance of the administration's argument was
considerably aided by the fact that it had little competition. According
to John Zaller's theory of public opinion, when elite discourse
favors a given policy, it produces a one-sided “mainstream
pattern” in which the most aware members of the public subscribe
strongly to the elite consensus.31
In the case of two-sided
information flows, the public will polarize along ideological lines, and
the most politically aware members will respond to elites within their own
party.
32 To evaluate information flow during the months preceding the war in
Iraq, we analyzed New York Times coverage of major Bush speeches
from September 11, 2002, to May 1, 2003, for the two days following each
of the speeches considered above.33
Due to
the large amount of coverage of the Iraq War, we limited the analysis to
only front-page stories under the assumption that these are the most read
and therefore most influential stories in a newspaper.
The
Times sets the tone and agenda for other newspapers and also
influences opinion leaders on foreign affairs.
34 The paper's editorial stand tends to be
liberal, a stance that may pervade the
Times' news coverage
generally through the hiring of journalists, the placement of quotes and
frames, and the choice of which facts to report.
35 Thus we
expected news coverage of the Bush speeches to provide Americans with an
alternative frame for understanding the issues by, for example, publishing
interviews with opposition sources and using critical language to describe
Bush's speeches.
To some extent, this was the case. Of the 35 newspaper stories coded,
17 were neutral, nine had a slightly positive tone, and nine were
negative.36
See appendix A for a fuller
account of the coding of tone.
The fact that as many stories were
positive as negative may be surprising. However, since this time period
encompasses the beginning of a military conflict, the lack of bias is less
remarkable since the media also tend to rally around the flag in wartime.
Six of the nine positive stories appeared after March 19, 2003, the first
day of military action.
Who is quoted affects the frame and tone of newspaper stories. Thirty
of the 35 stories quoted at least one member of the Bush administration,
with 21 quoting Bush directly. In contrast, only 12 stories quoted
Democrats, and 14 stories quoted criticism from members of the United
Nations.37
There was an overlap of five
stories that quoted both Democrats and critics of U.S. policy from UN
members other than the United States.
The debate over Iraq may
have been muted within the Democratic camp because of the timing of the
issue right before the midterm elections in November;
38 Alison Mitchell, “Democrats, Wary of War in Iraq, Aso
Worry about Battling Bush,” New York Times, September 14,
2002, A1.
the Democrats may have treaded lightly on Iraq so as to
not look weak on security issues. After Bush's speech to the United
Nations, Senate Minority Leader Tom Daschle said, “It would be
unfortunate if people drew from that [asking questions] that we
were opposed to what the president's doing.”
39 Alison Mitchell, “Democrats, Wary of War in Iraq, Aso
Worry about Battling Bush,” New York Times, September 14,
2002, A1.
The Democratic acquiescence to Bush policy closed off
one portion of the information flow that could have influenced the
American public.
While some debated the policy of war, there was little debate within
the Times' news coverage on the framing of the conflict in
terms of terrorism. The press plays a selectively critical role; if the
country supports the person making statements, journalists are less likely
to challenge him or her.40
This would
appear especially true for President Bush, whose approval ratings hovered
between 60 and 70 percent from September to December 2002, declined
slightly at the beginning of 2003, and rose again in March 2003.
41 After the September 2002 speech to the United
Nations, a news analysis by Patrick Tyler did state that “Mr. Bush
made no case today that Mr. Hussein's government in Baghdad was
connected in any way to the terrorists who plotted the hijackings and
assault on the United States.”
42 Patrick Tyler, “A New Face in the Sights,”
New York Times, September 13, 2002, A1.
Also, a story
that appeared in late January 2003 quoted Deputy Secretary of State
Richard Armitage as saying that the administration had built some of the
case against Iraq on ambiguous information.
43 James Dao, “US may give the UN data on Iraqi
labs,” New York Times, January 30, 2003, A1.
However, this admission was made as a prelude to Colin Powell's
February 2003 speech at the United Nations and was less damaging to the
Bush rhetoric since it originated from inside the administration. These
kinds of criticisms appeared in nine of 35 stories, not large in relative
terms, considering that all of the stories coded directly concerned
speeches given by the president.
The paucity of quotes from nonelite opponents, including members of
peace movements, is striking. Quotes from Iraqi leaders and citizens
appeared sparsely, in only four stories, while only one article quoted an
average American citizen. The small number of stories coded does not allow
us to make definitive conclusions about media criticism, but it suggests
that the press accepted the administration's framing and policy. It
appears that the information flow remained one-sided for the months
preceding the Iraq war. The administration's frame was not countered
by intense, sustained criticism by either the press or the Democratic
Party.44
Zaller's theory proposes that
public attitudes toward major issues are a response to the relative
intensity of competing political communications on those issues. See Zaller 1992, 210.
How the Public Responded to Bush's
Rhetoric
Since the first polling done about Iraq took place on September 11,
2002, in figure 2 we plot the Iraq and terror
indices beginning at that point and the corresponding support for the war
beginning the same date. As noted above, the Iraq and terror indices move
in tandem; moreover, public support frequently is higher following
increases in the Iraq and terror indices.
Bush's rhetoric and public opinion
Throughout this period, support for the war never dipped below 55
percent. As we mentioned earlier, this is surprising given that several
factors should have resulted in low levels of support for the war: on
average, 55 percent of Americans expected the war to be long; 44 percent
expected large numbers of casualties; 50 percent thought the war would
adversely affect the U.S. economy; 62 percent thought that the war
increased the short-term risk of terrorism at home; 76 percent thought
that the war increased the short-term risk of chemical or bioterrorism at
home; and 35 percent thought that the war would result in the
reinstatement of the draft.45
Based on
averages across all polls where question was asked; percentages vary
slightly by question wording and polling organization. Sources: ABC
News/Washington Post Poll, April 16, 2003, N=504, margin of
error [MoE] ± 4.5; Los Angeles Times Poll Study
484, April 2–3, 2003, N=745, MoE ± 4; CNN/USA
Today/Gallup Poll, March 29–30, 2003, N=1012, MoE ±
3; Newsweek Poll, March 27–28, 2003, N=1004, MoE ±
3; CBS News Poll, March 26–27, 2003, N=868, MoE ± 3; ABC
News/Washington Post Poll, March 20, 2003, N=506, MoE
± 4.5; ABC News/Washington Post Poll, March 17, 2003,
N=510, MoE ± 4.5; CBS News Poll, March 15–16, 2003, N=1049,
MoE ± 3.
Yet support for the war was very high.
46 Based on data from ABC News/Washington
Post polls from the following dates: March 23, 2003, March 27, 2003,
April 3, 2003, April 2–6, 2003, April 9, 2003, and April 16, 2003.
ABC News/Washington Post polls taken from August 29, 2002 to
March 20, 2003 all reflect levels of support over 55 percent. Dates of
these polls are: August 29, 2002, September 12–14, 2002, October
2–6, 2002, December 2002, January 16–20, 2003, January 27,
2003, January 28, 2003, January 30–February 1, 2003, February 5,
2003, February 6–9, 2003, February 19–23, 2003, March
5–9, 2003, March 20, 2003. Other polls from other sources such as
Los Angeles Times, Newsweek, FOX News/Opinion Dynamics, Pew
Research Center for the People and the Press, CNN/USA
Today/Gallup yielded similar results, although the exact
percentage of those approving of the war varies slightly by question
wording, polling organization, and survey date.
But Americans did not merely support the war, they did so rather
strongly. More than 40 percent felt that those opposed to the war should
not be allowed to speak out or hold protest marches or rallies because it
might hurt the war effort.47
Percentage
varies by question wording. Average across five polls is 41.0 percent,
with nonresponses and “don't knows” excluded from
analysis. Los Angeles Times Poll Study 481, January
30–February 2, 2003, N=1385, MoE ± 3; Los Angeles
Times Poll Study 482, February 7–8, 2003, reinterview of 809
respondents from Study 481, MoE ± 3; ABC News/Washington
Post poll March 27, 2003; ABC News/Washington Post poll
March 22, 2003; ABC News/Washington Post poll March 20, 2003;
CBS News Poll, March 26-27.
In addition, 55 percent supported the
war even without the United Nations' authorization, even if the
United Nations actively opposed the action. Support was not conditional on
the United States uncovering WMD in Iraq; more than three-quarters of
Americans endorsed the war regardless of whether WMD were ever found.
48 Average over 4 polls is 64.1 percent, with
non-responses and “don't knows” excluded from analysis.
Sources: Los Angeles Times Poll Study 481; Los Angeles
Times Poll Study 482; CNN/USA Today/Gallup Poll
March 22–23, 2003, N=1020, MoE± 3; CBS News/New York
Times Poll March 22, 2003.
This high, unconditional support for the war in Iraq is even more
surprising when we compare it to the war in Afghanistan. In October 2001,
the eve of the conflict in Afghanistan, 80 percent of Americans approved
of sending troops there,49
ABC
News/Washington Post Afghanistan Attack Poll 2, October 2001
[computer file], Horsham, PA: Taylor Nelson Sofres Intersearch
[producer], Ann Arbor, MI: Inter-university Consortium for
Political and Social Research [distributor], 2002,
http://webapp.icpsr.umich.edu/cocoon/ICPSR-STUDY/03299.xml.
roughly the same number who supported sending troops to Iraq in March
2003. The conflicts differed, however, in that military action in
Afghanistan was supported by our allies, and not expected to be long, to
result in large numbers of American casualties, and to adversely affect
the U.S. economy. Yet both wars were supported strongly. What the two
conflicts had in common was being rhetorically connected to the powerful
9/11 frame.
The data show the impact of this connection. For example, after
hearing Bush's 2003 state of the union address, discussed above, 9
percent of Americans changed from not supporting to supporting the war in
Iraq and, strikingly, credit Bush's speech as the reason for their
transition. Indeed, in one survey taken just days after the address, when
asked which statement best described their views about the Iraq war, 9
percent of respondents thought that the statement best describing their
feelings before and after the speech was: “Before the speech I did
not favor U.S. military action against Iraq, but the speech changed my
mind and I favor it now.”50
Los Angeles Times Poll Study 481 (emphasis added).
Many of the polls cited in preceding endnotes show an increase in support
for the war of approximately the same magnitude following the state of the
union address.
This gain was highest among Democrats, who were
previously least supportive of the war.
More generally, we expect those who regularly heard the Bush
administration's rhetoric to be more likely to perceive a strong
connection between Saddam Hussein and terrorism, and thus more likely to
support the war in Iraq. Figure 3 shows that in
fact the more closely a respondent followed news coverage about Iraq, the
more likely she was to support invasion, all else held constant.51
PIPA-Knowledge Networks Poll: Americans on
Iraq and the UN Inspections II, February 12–18, 2003, N=3163 (N=2186
adults nationwide were sampled randomly, with an additional 977 over
sampled from California, Florida, Illinois, New York, and Texas, with the
full sample weighted so that it is correctly proportional to the
geographic distribution of the US population) MoE ± 2.6. Figure 4 shows the predicted probabilities that result
from a logistic regression of whether or not the respondent supported
invading Iraq (1 if yes, 0 if no) on attentiveness to news coverage about
Iraq (1 if least attentive, … , 4 if most attentive), party
identification dummy variables (independents were the base case), whether
or not the respondent believed Iraq was responsible for 9/11 (1 if
yes, 0 if no), and whether or not the respondent heard Powell's
speech to the United Nations (1 if yes, 0 if no). For the predicted
probabilities, whether or not the respondent believed Iraq was responsible
for 9/11 was set equal to 1; when this variable is set to zero, all
predicted probabilities are lower, but the differences in the probability
of supporting the war in Iraq between low and high attentiveness to news
coverage, viewing of the Powell speech versus not viewing the speech, and
so forth, remain nearly identical to those displayed in figure 4.
This was true for both Democrats
and Republicans. A respondent who closely followed the news about Iraq was
on average 15 percent more likely to support the war in Iraq than a
respondent who did not follow the news about Iraq at all.
52 The 15 percent figure was calculated as
follows. First, 4 differences were calculated: (1) the predicted
probability of supporting the war for Democrats who heard Powell's
speech and followed news about the Iraq War closely minus the predicted
probability of supporting the war for Democrats who heard Powell's
speech and did not follow news about the Iraq War closely; (2) the
predicted probability of supporting the war for Republicans who heard
Powell's speech and followed news about the Iraq War closely minus
the predicted probability of supporting the war for Republicans who heard
Powell's speech and did not follow news about the Iraq War closely;
(3) the predicted probability of supporting the war for Democrats who did
not hear Powell's speech and followed news about the Iraq War closely
minus the predicted probability of supporting the war for Democrats who
did not hear Powell's speech and did not follow news about the Iraq
War closely; (4) the predicted probability of supporting the war for
Republicans who did not hear Powell's speech and followed news about
the Iraq War closely minus the predicted probability of supporting the war
for Republicans who did not hear Powell's speech and did not follow
news about the Iraq War closely. These four differences were averaged, and
the result was 15.
Clearly, repeatedly hearing the Bush rhetoric
affected the public's views on the war.
Probability of support for Iraq war by attentiveness to Iraq war news
coverage
Colin Powell's speech to the United Nations on February 5, 2003,
had an additional effect on public support. His speech provided more
evidence than any other official administration speech about the links
between Iraq and al Qaeda. The data show that, even controlling for
general media consumption, those who reported hearing Powell's speech
were more likely to support the war in Iraq, as shown in figure 3.53
We realize
the biases inherent in self-reporting; however, we expect that the
probability of over-reporting is uncorrelated with the other variables in
the model. Additionally, the proportion of respondents who reported
hearing Powell's speech in PIPA-Knowledge Networks Poll February
12–18, 2003, is very similar to that reported in Los Angeles
Times Poll Study 482.
This was true for both Republicans and
Democrats. The data also allow us to examine whether hearing Powell's
speech is merely correlated with increased support for the war, or whether
it caused it. Ten percent of Americans changed from not supporting to
supporting the war in Iraq after hearing the speech, and they credited the
speech with changing their views. Once again, 10 percent of respondents
thought that the statement best describing their feelings before and after
Powell's speech was: “Before the speech I did not favor U.S.
military action against Iraq, but
the speech changed my mind and
I favor it now.”
54 Los Angeles
Times Poll Study 482 (emphasis added).
This gain was again
highest among Democrats.
The reason for the shift in opinion is clear; there was a 30-point
jump in the number of Americans who felt convinced of a link between
Saddam Hussein and al Qaeda following Powell's speech to the United
Nations. Indeed, in the two days following Bush's state of the union
address in 2003, respondents were asked whether or not enough evidence had
been provided to show that Saddam Hussein and al Qaeda were connected;
only 38 percent of respondents said yes. However, when those same
respondents were reinterviewed following Powell's speech to the
United Nations, 68 percent agreed.55
Los
Angeles Times Poll Study 481 and Los Angeles Times Poll
Study 482.
Nearly the entire 30-point gain came from Democrats.
Since the state of the union address came just five days before
Powell's speech, it is unlikely that any other event caused the
shift.
Additional polls corroborated our conclusions. To pick just one
additional example, respondents who thought al Qaeda was the most
important threat facing the United States were more likely to support the
invasion of Iraq than those who thought al Qaeda was America's
second, third, or fourth most important problem.56
PIPA-Knowledge Networks Poll: Americans on Iraq after the UN
Resolution, November 19–December 1, 2002, N=1106, MoE ±
3.
Surprisingly, respondents who considered al Qaeda to be
America's most important threat were more likely to favor invading
Iraq than those who thought
Iraq was America's most
important threat.
57 By “better
predictor,” we mean that the partial correlation between the Iraq
war variable and the importance of the al Qaeda problem variable was
higher than the partial correlation between the Iraq war variable and the
importance of the Iraq problem variable. Moreover, in a regression of war
support on the importance of the al Qaeda problem, the importance of the
al Qaeda problem was a statistically significant predictor of war support
in a two-tailed test at the 5 percent level, whereas in a regression of
war support on the importance of the Iraq problem, the importance of the
Iraq problem was not a statistically significant predictor of war support
in a two-tailed test at the 5 percent level.
Was It WMD?
One of the stated goals of the war was to rid Iraq of WMD. In a speech
to the United Nations on September 12, 2002, and again in his 2003 state
of the union address, Bush claimed that Saddam Hussein's refusal to
account for and destroy banned weapons compelled the United States to
disarm him.58
“A Decade of Deception and Defiance,”
the document prepared by the White House for the United Nations, lists 16
Security Council resolutions that Iraq failed to comply with.
59 U.N. weapons inspectors had, on multiple occasions,
accused Hussein of deliberately deceiving inspectors in order to hide
evidence of WMD. The United Nations made it clear to the public that there
was reason to suspect that Hussein had undeclared WMD, and codified this
with the passage of Resolution 1441 on November 8, 2002.
Could the public's support for the Iraq war have originated from
a belief that Hussein had WMD and must be disarmed in accordance with
United Nations directives? This is highly improbable, given the public
opinion data. If the public supported the war because they believed that
he had WMD, then we would expect support for the war to be higher
contingent on finding the weapons. But 82 percent of Americans said they
would support the war, whether or not any WMD were found in Iraq,60
Los Angeles Times Poll Study
484.
and 64 percent said “Saddam Hussein should be removed
from power in Iraq, regardless of whether U.N. inspectors find evidence of
weapons of mass destruction.”
61 Based
on 4 poll average, with a range from 40.4 to 83.0 percent. Sources:
Los Angeles Times Poll Study 481; Los Angeles Times Poll
Study 482; CNN/USA Today/Gallup Poll, March 22–23,
2003; ABC News/Washington Post Poll, March 20,
2003.
We do not see WMD as a prominent frame in Bush's rhetoric. While
WMD are certainly part of the oratory, they are a much smaller percentage
than discussions of terrorism. Bush began discussing WMD at the same point
he began discussing Iraq: the January 29, 2002, state of the union
address. Bush consistently referred to “weapons of mass
destruction” (or alternatively “weapons of mass murder”)
from January 2002 to May 2003. Of 22 speeches coded in this time period,
17 referenced WMD at least once, with a mean of 2.3 mentions per speech.
In this same period, however, the average number of references to
terrorism was 12.2 per speech (see fig. 4).
Clearly, then, terror, not WMD, was the primary rhetorical frame.
Iraq index, terrorism index, and mentions of WMD
A speech on October 7, 2002, titled “Outlining the Iraqi
Threat,” referred to WMD seven times, the most of any address;
however, the same speech alluded to terror or terrorism 35 times, five
times as often. Both Iraq and Saddam Hussein were placed in the same
sentence as the word terrorism, and the president stated explicitly that
associations existed between the Iraqi state and terrorism. This same
pattern is apparent in the other speeches. Only once, in the March 19,
2003, address that declared the beginning of the war, did WMD appear more
than discussion of terrorism. The president employed the term WMD once in
that speech, never mentioning terrorism.
Alternative Theories
There are other possible explanations for Americans' support of
the Iraq war. Perhaps support for the Iraq war was just a by-product of
Bush's high approval ratings.62
Would the public have willingly
supported any policy Bush pursued? This theory is not well supported by
the data. Although Bush's approval levels were high, his approval
levels for handling the economy, for example, were substantially lower
than those of his handling of terrorism or foreign policy.
63 In fact, fewer than half of all Americans approved
of Bush's handling of the economy by December 2002.
64 Thus
the public was not unconditionally supportive of all Bush's policies;
rather, the Iraq war seems to be a special case.
A second possibility is that high levels of support for the war were
due to the rally-around-the-flag effect. We calculate, in fact, that a
rally effect explains approximately 9 percent of support for the war.65
Nine percent is the difference in the average
approval for the war between October 2002 and March 7, 2003, and March 8,
2003, through the declared end of hostilities in Iraq on May 1,
2003.
We can see in
figure 2 that there
is a large jump in support on March 8, 2003, which coincides with a
statement by President Bush in a radio address to the nation making it
clear that Iraq's failure to comply with United Nations'
directives would result in military action by the United States. We
conjecture that this jump represents the rally effect.
Prior to March 8, 2003, none of the conditions necessary for a rally
effect were present. According to John Mueller, in order to qualify as a
rally event, an event must be international, directly involve the United
States in general and the president in particular, and be “specific,
dramatic and sharply focused.”66
Clearly the beginning of the
Iraq war meets these criteria, but none of these conditions were present
in October 2002, when support for the war was already strong. From October
2002 to March 2003, however, the linkages between Iraq and 9/11 were
made repeatedly. So while it was too early for Americans to be rallying
around the flag to support a conflict that was far from inevitable in
October 2002, it was not too early to respond to Bush's rhetoric
linking Iraq with 9/11. The 9 percent jump in approval of the war in
March 2003, on the eve of the conflict, however, does meet Mueller's
conditions. In sum, while a rally effect might explain the boost in war
support prior to the actual onset of the conflict, it is not the entire,
or even the primary, reason that public support for the war was so high
and unconditional in late 2002 and early 2003.
Another possibility is that the public supported the Iraq war in 2003
because they felt that the 1991 Iraq war was never completely finished. In
1991 support for the war was high, and Americans' dislike of Saddam
Hussein registered long after the war's end. Perhaps the American
public had a decade-plus-long hankering to “finish” the
earlier war. Again, this is not borne out by the data. If Americans
supported the 2003 war for this reason, then we should have seen high
support from the first time the president mentioned the idea in a speech.
Bush first hinted at the possibility of war with Iraq in the 2002 State of
the Union address, yet support for the 2003 war did not reach high, almost
unconditional levels until about a year later. Indeed, support does not
begin to increase until the war with Iraq is closely and repeatedly tied
to the war on terror.
A final alternative explanation for public support of the Iraq war is
that Americans wanted to punish Muslims for 9/11. Xenophobia has been
a repeated flaw in American history; however, the data strongly suggest
that this was not a factor. Whether or not a respondent had a favorable or
unfavorable opinion about Islam and whether or not a respondent reported
being generally suspicious of people of Arab descent made him or her no
more likely to support the war with Afghanistan.67
ABC News/Washington Post Afganistan Attack Poll
2. Several regressions and Chi-square tests bore out these results; this
analysis can be obtained from the authors upon request. Questions
addressing these xenophobic concerns were not available for data on the
Iraq war. After about November 2001, these sorts of questions disappear
from public opinion polls.
Since the war with Afghanistan was
even more closely tied to 9/11, we would expect that if Americans were
harboring resentment towards people of Islamic descent because of
9/11, this would be apparent in the Afghanistan data. The fact that
this sentiment did not appear then makes it unlikely that the American
people would suddenly develop such resentment in the following year.
Responsible Use of Powerful Frames
Knowing that the Iraq-as-war-on-terror frame effectively rallied
support for the war effort, what can we say about the normative
implications for democratic theory? First, we might be concerned with the
lack of deliberation among elites and the media in the months preceding
the war.The deliberative spectrum, normally constrained by the two
moderate political parties, was in this instance further constrained by
the lack of dissent among elites. Mass opinion relies on dialogues of the
political elite, and the fact that the framing of the war and the policy
itself were scarcely discussed facilitated high levels of support for an
unprecedented foreign policy of unilateral preemption.68
This is not
to say that Americans would not have supported the war in the face of
Democratic or media opposition, but rather that their conversion might not
have been so complete nor so quick. Antiwar protesters were making
themselves heard on city streets; with more elite deliberation, more mass
deliberation might have followed.
We might worry about whether the use of the 9/11 frame can lead to
unconditional public support for any international policy initiated by the
president, especially given that use of such a powerful frame may stifle
both deliberation and accountability. One implication of our analysis is
that this frame ought to be used sparingly and responsibly by all
political elites in both parties, and as citizens, we ought to be vigilant
in our examination of the framing of political conflicts. Pictures of the
burning World Trade Center on September 11, 2001, evoke deep emotions,
which elites can potentially use to increase support for conflict.
Political elites of both parties must not improperly use these images to
win the support of the American people for particular policy choices.
Epilogue
Not much has changed since we completed our paper in June 2003. The
United States military continues to fight in Iraq, and American citizens
continue to support the effort, even if cracks in that support are
appearing as American casualties increase. A poll released in April
200469
revealed that 20 percent of Americans still believe
Iraq is responsible for September 11 and that this misperception is still
a significant predictor of whether or not Americans support the ongoing
conflict in Iraq. In addition, according to a poll taken a few months
before the 2004 election, those who believe that Iraq provided support to
al Qaeda were nearly five times more likely to report an anticipated vote
for Bush in November.
None of this is surprising, given that Iraq and al Qaeda continue to
be linked in the Bush administration rhetoric. On September 14, 2003, in
an interview on “Meet the Press,” Vice President Cheney said,
“If we're successful in Iraq … then we will have struck
a major blow right at the heart of the base, if you will, the geographic
base of the terrorists who had us under assault now for many years, but
most especially on 9/11.”70
Vice
President Richard Cheney, Meet the Press, September 14, 2003,
quoted in Walter Pincus and Dana Milbank, “Al Qaeda-Hussein link is
dismissed,” Washington Post, June 17, 2004, A1.
In
a speech to first responders in Wisconsin on March 30, 2004, President
Bush discussed 9/11 and the war on terror immediately preceding a
defense of the decision to go to war in Iraq:
September the 11th taught us another lesson, that this nation
must always deal with threats before they fully materialize…. I
want to remind you that I saw a threat. I looked at the intelligence and
saw a threat in Iraq…. September the 11th taught us a lesson.71
Given this ongoing entanglement of Iraq and September 11, it is
understandable that the public continues to hear and accept erroneous
linkages. In a PIPA/Knowledge Networks poll taken in June 2003, 71
percent of respondents once again said that the Bush administration had
implied a link between Iraq under Saddam Hussein and 9/11.72
Nevertheless, with continued attacks on American
troops and the September 11 Commission finding no evidence of a
cooperative link between Iraq and al Qaeda prior to the launch of the Iraq
war, we wonder whether the American public will continue to respond to the
Iraq-as-war-on-terror frame with high levels of support for this ongoing
conflict.
Appendix A: Speeches and Newspaper
Codings
Our speech coding scheme considered the length of the speech as well
as frequency of the terms Iraq, terror, Saddam Hussein, Afghanistan,
Taliban, al Qaeda, Osama bin Laden, weapons of mass destruction, rogue
nations, and evil. We coded whether the terms
terror and Iraq fell in the same sentence or paragraph,
and whether the term terror preceded the first mention of Iraq.
We also coded whether the speech placed terror and Saddam Hussein in the
same sentence or paragraph, and whether 9/11 and Saddam Hussein
and/or Iraq were referenced together. The speeches range from the
assurances given to the country on September 11, to weekly radio addresses
and speeches to interest groups, and from nationally televised prime-time
speeches to addresses given to small groups of Republican supporters. The
source was the White House Web site (www.whitehouse.gov), which posts all
public talks given by the president.
Additionally, we coded articles from page A-1 of the New York
Times for the two days following each presidential speech about Iraq.
There are 35 articles from January 2002 (after the state of the union
address) to May 1, 2003. The newspaper coding scheme included information
about the date, length, and author of each article. Each article was coded
with 0/1 dichotomous variables to reflect its being
positive/accepting, negative/critical, or neutral toward the Bush
rhetoric. This decision was made based on our assessment of the language
used by the article's authors. We also coded for whether articles
accepted the Iraq-al Qaeda connection and Iraq as part of the war on
terror. We coded who was quoted as well as the overall tone of the
quotation on a scale of 1–5, with 1 being the most positive. A story
that used language critical or skeptical of the Bush rhetoric or that did
not accept the Iraq-al Qaeda tie or the connection between Iraq and the
war on terror was coded as negative or slightly negative in tone. For
example, a story by Patrick Tyler stated, “Bush made no case today
that Mr. Hussein's government was connected in any way to the
terrorists who plotted the hijackings and assault on the United
States”; it quoted Bush, but also Kofi Anan and Democrats, and was
coded as slightly negative. Stories that interviewed only or almost solely
administration officials, used positive language about the Bush speech,
and accepted the Iraq-al Qaeda link or did not discuss the validity of the
link were coded as positive or very positive. Stories with balanced
sources (some Democrats, some Republicans, and some foreign leaders) and
balanced language were coded as neutral. Final decisions about tone were
based upon the judgment of the tenor of the article based upon the
language used by the author.
