American Patriotism, National Identity,
and Political Involvement
Leonie Huddy
Nadia Khatib
Stony Brook University
University of Arizona
Researchers disagree over the definition, measurement, and expected political consequences of American patriotism, a
situation that is fueled by the absence of a strong theoretical research foundation. We develop and evaluate a new measure of
national attachment that is grounded in social identity theory (Tajfel and Turner 1979), drawing on data from three distinct
sources: two studies of undergraduate students and the 1996 General Social Survey (GSS). Confirmatory factor analyses
provide clear evidence that national identity is distinct from other measures of national attachment including symbolic,
constructive, and uncritical patriotism (and nationalism). National identity has a number of other good measurement
properties when compared to existing measures: it receives equal endorsement from conservatives and liberals (unlike most
other measures which exhibit an ideological bias), develops with time spent in the United States among immigrants, and
most importantly is the only measure of national attachment to predict political interest and voter turnout in both student
and adult samples, consistent with the predictions of social identity theory. In that sense, the national identity measure
outperforms all other measures of national attachment and provides unambiguous evidence that a strong American identity
promotes civic involvement.
R
esearch on patriotism has been marred by a confusing array of terms, definitions, and expected
consequences in which patriotism is variously defined as a sense of national loyalty, a love of national symbols, specific beliefs about a country’s superiority, and
as a crucial ingredient in the development of civic ties
to a mature nation (Hurwitz and Peffley 1999; SpinnerHalev and Theiss-Morse 2003; Sullivan, Fried, and Dietz
1992). Patriotism researchers have reached some (although far from uniform) consensus that a sense of superiority and need for foreign dominance better reflect
nationalism than patriotism (De Figueiredo and Elkins
2003; Karasawa 2002; Kosterman and Feshbach 1989;
Mummendey, Klink, and Brown 2001; Sidanius et al.
1997), clarifying what patriotism is not. But this still leaves
considerable disagreement over what exactly it is.
Diverse Forms of Patriotism
There is broad agreement on the meaning of patriotism as
“a deeply felt affective attachment to the nation” (Conover
and Feldman 1987, 1) or the “degree of love for and pride
in one’s nation” (Kosterman and Feshbach 1989, 271).
More pronounced disagreement emerges, however, over
the way in which patriotism is measured. Patriotism items
are commonly tinged with political ideology in the United
States, resulting in greater apparent patriotism among political conservatives than liberals. Consider the symbolic
patriotism scale in the American National Election Studies
(ANES) which combines pride in being American with
pride in the flag and anthem (Conover and Feldman 1987;
Hurwitz and Peffley 1999; Karasawa 2002; Kosterman
and Feshbach 1989; Sidanius et al. 1997). Ideological bias
An earlier version of this paper was presented at the annual meeting of the Midwest Political Science Association. We wish to thank
the National Opinion Research Council for use of the 1996 GSS data. We also wish to thank Stanley Feldman, Jenny Boldero, Wendy
Rahn, Elizabeth Theiss-Morse, Christopher Parker, and members of the Stony Brook Political Psychology seminar, and several anonymous
reviewers for their helpful comments and insights on an earlier draft of this paper.
Leonie Huddy is professor of political science, Center for Survey Research, Stony Brook University, Stony Brook, NY 11794-4392 (631)
632-7639 (Leonie.Huddy@sunysb.edu). Nadia Khatib is a lecturer in political science, University of Arizona, 315 Social Sciences Building,
PO Box 210027, Tucson, AZ 85721-0027 (nkhatib@email.arizona.edu).
American Journal of Political Science, Vol. 51, No. 1, January 2007, Pp. 63–77
C 2007,
Midwest Political Science Association
ISSN 0092-5853
63
64
enters into the scale because liberals express some reserve
about national symbols as a lingering historical consequence of opposition to the Vietnam war, in which flag
burning became synonymous with liberal, antiwar sentiment. As a consequence, support for symbolic patriotism is stronger among conservatives than liberals, raising needed questions about its utility as a broad measure
of patriotism (Conover and Feldman 1987; Hurwitz and
Peffley 1999).1 De Figueiredo and Elkins (2003) analyze
the political effects of national pride, a concept closely
aligned with symbolic patriotism since both measures reference a sense of pride. National pride has an ideological
basis in Europe, but this link has not yet been tested in the
United States (Elkins and Sides 2003).
Schatz, Staub, and colleagues deal with ideological
asymmetry in the measurement of patriotism by distinguishing between constructive and blind patriotism. This
distinction is designed to differentiate a conservative aspect of patriotism (blind) from one that is more likely to
be endorsed by liberals or perhaps those on both sides of
the ideological divide (constructive). Blind, or what we
refer to as uncritical patriotism (to get around the normative implications of the term blind), is defined as “an
unwillingness both to criticize and accept criticism” of the
nation and is indexed by items such as “my country right
or wrong” (Schatz and Staub 1997, 231; Schatz, Staub, and
Lavine 1999). Uncritical patriotism is linked to authoritarianism which is characterized, in turn, by a tendency
to defer to authority figures and support them unconditionally. And authoritarians are typically conservative
(although not all conservatives are authoritarians) producing higher levels of uncritical patriotism among political conservatives, on average (Schatz, Staub, and Lavine
1999). Uncritical patriotism is, therefore, not a good measure of broad patriotism because it is both ideologically
divisive and closely aligned with nationalism and ethnocentrism, blurring the distinction between patriotism and
nationalism.
Schatz and Staub developed a scale of constructive
patriotism that encapsulates a form of patriotism more
acceptable to liberals. Constructive patriotism is defined
as “an attachment to country characterized by critical loyalty” and “questioning and criticism” driven by “a desire
for positive change” (Schatz, Staub, and Lavine 1999, 153).
Constructive patriots agree with items such as “I oppose
some U. S. policies because I care about my country and
want to improve it” and “I express my love for America
1
Symbolic patriotism is more typically aligned with political ideology than partisanship in the United States because both parties
vary in how vigorously they identify themselves with the flag and
other national symbols.
LEONIE HUDDY AND NADIA KHATIB
by supporting efforts at positive change.” Despite expectations that it would gain stronger endorsement from liberals, constructive patriotism is empirically unrelated to
political ideology, increasing its potential as a measure of
broad, nondivisive patriotism (Schatz, Staub, and Lavine
1999). Unfortunately, there are several problems with the
scale that limit its empirical utility.
First, items in the constructive patriotism scale are
complex, mixing a love of country with political efforts
directed at a change in the status quo. Scale items may primarily detect love of country and thus reflect a broad sense
of national attachment, but they may also confound patriotism with political interest, involvement, and activity
(Rothi, Lyons, and Chryssochoou 2005; Schatz, Staub, and
Lavine 1999). Schatz and colleagues find a clear link between political involvement and constructive patriotism,
but this further muddies the waters. Is constructive patriotism primarily a measure of patriotism? Or of political
interest and involvement? Current research fails to clearly
resolve this issue. Second, all of the agree-disagree items in
the constructive scale are worded in a positive direction,
introducing problematic acquiescence bias (Schuman and
Presser 1981). This problem is difficult to fix because the
compound nature of constructive patriotism items makes
scale items difficult to reverse. Should items be reversed
by stating that one’s efforts at change do not arise from
a love of country? Or that Americans should not work
towards positive change?
Third, the empirical measure of constructive patriotism is not clearly differentiated from that of blind patriotism, blurring the meaning of both concepts. Consider the
following example. The blind patriotism scale contains
the following reverse-worded item: “For the most part,
people who protest and demonstrate against U.S. policy
are good, upstanding, intelligent people.” This seems conceptually similar to the following item from the constructive patriotism scale: “If you love America, you should
notice its problems and work to correct them.” But the
two scales (blind and constructive) are unrelated (Schatz,
Staub, and Lavine 1999). All four patriotism scales (symbolic, national pride, uncritical, and constructive) are thus
open to criticism as general measures of national attachment. At least two measures (symbolic and uncritical patriotism and possibly national pride) are conflated with
a conservative ideology, leading to fractious debate over
who is “truly” patriotic. And a third measure (constructive patriotism) is defined as support for active political
change which could easily exclude patriotic individuals
who have no interest in altering the status quo. In our
view, a strong theoretical framework is needed to guide
the measurement of patriotism and account for its political consequences.
AMERICAN PATRIOTISM REDEFINED AS AMERICAN IDENTITY
A Social Identity Approach to
National Attachments
We draw on social identity theory to develop a theoretically grounded measure of national identity with clear
implications for intergroup behavior (Tajfel 1981; Tajfel
and Turner 1979). Social identity theory represents a rich
body of thought and findings on the origins and consequences of a strong social identity that has had growing
influence on political research in recent years (Gibson and
Gouws 2000; Sniderman, Hagendoorn, and Prior 2004). A
social identity is typically defined as an awareness of one’s
objective membership in the group and a psychological
sense of group attachment (Tajfel 1981). Consistent with
this definition, we define national identity as a subjective
or internalized sense of belonging to the nation and measure it with questions that typically assess social identities
(Huddy 2001, 2003).
Social identity theory generates numerous predictions about the consequences of national loyalties. We
explore several politically central expectations in this research. First, national identity is expected to be nonideological in flavor because it represents a pervasive sense
of subjective attachment to the nation. Feeling American
does not hinge on the endorsement of a specific political ideology (Huddy and Khatib in press). Research on
American identity among members of ethnic and racial
minority groups provides evidence of the nonideological nature of American identity (Citrin, Wong, and Duff
2001). In this research, American identity is defined as
a sense of being or feeling American and is empirically
unrelated to self-described liberalism or conservatism
(Sniderman et al 2004; Citrin, Wong, and Duff 2001;
Sidanius et al. 1997). In contrast, other forms of patriotism hinge on beliefs about the specific meaning of
American identity and can be endorsed more narrowly.
For example, uncritical patriotism is an especially contentious, ideologically tinged form of patriotism that involves unwavering support of political leaders (Schatz,
Staub, and Lavine 1999). Uncritical patriots are very likely
to identify as Americans, but not all American identifiers
endorse uncritical patriotism.
Second, a strong national identity is expected to increase political involvement. Turner and colleagues’ selfcategorization theory, an offshoot of social identity theory, predicts that individuals with a strong group identity
are most likely to conform to group norms (Turner et al.
1987; Terry, Hogg, and White 1999). Self-categorization
researchers find, for example, that the intention to practice protective health behaviors among strong group identifiers depends on whether or not these behaviors are seen
65
as the norm among fellow group members (Terry and
Hogg 1996; Terry, Hogg, and White 1999). They have
also manipulated group norms experimentally and observe stronger adherence to manipulated norms among
group identifiers (Wellen, Hogg, and Terry 1998; Terry,
Hogg, and Duck 1999). Moreover, highly identified group
members are most likely to conform to ideal or prescriptive norms (e.g., all good Americans should vote) rather
than descriptive norms (e.g., only some Americans actually vote) of group behavior and to experience more
positive emotions after conforming to them (Christensen
et al. 2004).
We extend these findings from self-categorization
theory to American identity. Acts of civic participation are
viewed by political theorists as central to national identity in democratic countries and constitute what is seen as
normative behavior for a “good” citizen (Conover, Searing, and Crewe 2004). This conception of national identity is widely endorsed in the United States. When asked
why they rate themselves as a “particularly good,” “good,”
“ordinary,” or “not so good citizen” and what they could
do to become a better citizen, a majority of Americans
mention voting in research conducted by Conover and
colleagues (2004). References to voting were more numerous among American than British participants, in
part because a higher number of Americans mention it as
something they could do to become a better citizen. But
over a third of Americans also list voting as the basis for
their current sense of citizenship. Voting thus constitutes
a prescriptive, normative component of American identity (see also Devos and Banaji 2005). As a consequence,
we expect a strong American identity to increase political
interest and involvement, including voter turnout.
An expected positive link between national identity
and political involvement stands in marked contrast to
the predicted effects of other forms of patriotism. Consider first uncritical patriotism. Schatz and colleagues
(1999) find lower not higher levels of political involvement and interest among uncritical patriots. They do
not provide a clear theoretical basis for this observed
relationship, but they do find a link between authoritarianism and uncritical patriotism that could explain the
connection. Authoritarians typically exhibit lower levels
of political information, interest, and involvement that
can be traced to their reduced scores on openness to
experience, a key personality trait (Huddy et al. 2005;
McFarland and Mathews 2005; McRae 1996; Oesterreich
2005; Peterson, Duncan, and Pang 2002). As a result, authoritarians are less likely than others to expose themselves
to new experiences or challenging information (Altmeyer
1996; Feather, Boeckman, and McKee 2001, Lavine,
Lodge, and Freitas 2005). Moreover, a general political
66
LEONIE HUDDY AND NADIA KHATIB
passivity among authoritarians is consistent with one of
their defining attributes—a greater deference to authority
figures (Adorno et al. 1950; Altmeyer 1998). Thus, we expect to find lower levels of political involvement and interest among uncritical patriots due, in part, to their greater
authoritarianism.
Constructive patriotism has been linked to participation and interest in Schatz and colleagues’ (1999) research but the theoretical reason for this link also remains
unclear. It could be a real reflection of constructive patriots’ desire to participate in American politics. It could
be a measurement artifact, as noted earlier, that arises
because political involvement is explicitly mentioned in
scale items. Or constructive patriotism could overlap with
a sense of national identity and simply convey support for
American norms of political involvement, in which case its
effect would be eclipsed by national identity. To date, there
has been no direct empirical test of the relative impact of
national identity and constructive patriotism on political
involvement. Finally, there is no empirical evidence of, or
theoretical basis for, a direct link between symbolic patriotism and political involvement or national pride and
involvement.
Research Expectations Concerning
National Identity
Several propositions derived from social identity theory are tested in this research. First, a strong national
identity should be related to but distinct from symbolic
patriotism, national pride, constructive patriotism, and
uncritical patriotism. Second, national identity should
be less ideologically divisive than other patriotism measures. Third, national identity should increase adherence to group norms which heighten political interest and involvement. In contrast, uncritical patriotism
should lower political involvement, symbolic patriotism
should have no impact on it, and the effects of constructive patriotism remain unclear. Constructive patriotism
could have a direct impact on political interest and involvement or its effects might be conveyed by national
identity.
Methods
Sample
To test our hypotheses, we draw data from three distinct sources: a 2002 student study, a 2004 student study,
and the 1996 General Social Survey (GSS). The two student surveys afford an opportunity to develop and test
a new measure of national identity, examine its impact
on various measures of political involvement and interest, and contrast its effects with competing patriotism
measures. The GSS provides a nationally representative
sample of Americans on which to test the prediction that
a strong national identity leads to higher levels of political
involvement.
Student Studies
Both student studies included responses from students
at Stony Brook University who were enrolled in political
science classes and participated in the studies either as
a course requirement or for extra credit. The first study
was conducted over a two-week period in November 2002
(N = 341); the second occurred in February, 2004 (N =
300). Students in both studies had very similar characteristics. Both samples were almost evenly split between men
and women (53% male in 2002 and 50% male in 2004)
and were equally diverse in terms of student race and ethnicity. In 2002, 58% of students were white, 19% Asian,
15% black, and 7% Latino. In 2004, the number of white
students had decreased slightly to 48%, and the number
of minority students had increased overall to include 25%
Asian, 13% black, and 10% Latino.
In addition to being ethnically diverse, the university’s student body attracts a large number of immigrants.
In 2002, 23% of participants were first-generation immigrants (born outside the United States), and 25% were second generation (born in the United States but both parents
born outside the country). The 2004 sample is similarly
diverse, including 26% first-generation and 24% secondgeneration immigrants. Among immigrant students, time
in the United States ranged from a few months to 23 years
with a median of 12 years in 2002 and 13 years in 2004.
The bulk of immigrants reported their background as either white (35% in 2002 and 31% in 2004) or Asian (32%
in 2002 and 36% in 2004). The majority of participants
in both studies are between the ages of 18 and 22 (median
age = 20). We capitalize on the immigrant nature of the
sample to examine the impact of immigrant status on the
strength of different national attachments.
1996 General Social Survey (GSS)
The 1996 General Social Survey (GSS) includes 2,904
respondents drawn from all U.S. residents. The sample
is split into two halves, with each split half randomly
67
AMERICAN PATRIOTISM REDEFINED AS AMERICAN IDENTITY
assigned to one of three questionnaire “ballots” resulting
in six different conditions. The survey included a special
national identity module developed by the International
Social Survey Programme (ISSP). Roughly half (1,367)
of all individuals answered questions related to patriotism and national identity, of whom 1,285 are citizens.
Our effective subsample is further reduced to 680 citizens who were asked a number of added national identity
questions, including a critical component of the national
identity scale. We refer to this throughout as “the GSS subsample.” This GSS subsample included 283 (42%) male,
397 (58%) female, 96 (14%) black, and 555 (82%) white
respondents. Their median age is 44 with an average of
13 years of education.
Analytic Strategy
We employ structural equation modeling (SEM) to investigate the structure of national attachments and their
relationships with various aspects of political involvement. There has been a proliferation of national attachment measures, many of which are interrelated and depend on similar questions or question formats. The use
of SEM identifies the underlying structure of opinions
by incorporating well specified measurement models in
place of traditional summed scales or mean scores and
controls for measurement error. When dealing with Likert agree-disagree response items, all items are loaded
on both a common method covariance or style factor (with all loadings set to 1) and a substantive factor
to correct for acquiescence bias (with the exception of
the constructive patriotism scales in which all items are
in the same direction and cannot be loaded on a style
factor).
All analyses are conducted using maximumlikelihood estimation, except for two models with discrete
dependent variables which were estimated with weighted
least squares and noted in the text.2 We report robust standard errors for all parameter estimates. All models were
estimated using the Mplus® statistical package, which provides more robust estimation for limited and categorical
latent dependent variables and more efficiently handles
missing data (imputed using Maximum Likelihood) than
other SEM programs (Muthén & Muthén 2001).
2
All items containing four or more response options are estimated
assuming a continuous latent trait. To avoid violating the assumption of multivariate normality, we also ran models (for items that
had less than 10 categories) treating the items as categorical (using
robust WLS with polychoric correlations) and found comparable
results. We report results from the continuous models for their
relative ease in interpretation, except where noted in the text.
Structure of National Attachments
The structure of national attachments was assessed in all
three data sets. The 2002 and 2004 student studies have
nearly identical measures of national identity and symbolic patriotism; the measures of uncritical patriotism
were more divergent; and a measure of constructive patriotism was confined to the 2004 study. National Identity was assessed by four questions in 2004.3 The symbolic
patriotism scale was formed from two items adapted from
the NES patriotism scale.4 Constructive patriotism was assessed by four agree-disagree items from the Schatz scale.
Uncritical patriotism was comprised of seven items from
Schatz, Staub, and Lavine’s (1999) blind patriotism scale.5
Table 1 includes the wording of all national attachment
items in the 2004 student study.
Levels of national identity were high in both student studies, with over 65% of students referring to
Americans as “we” most or all of the time and reporting that “American” was a term that described them very
or somewhat well. Levels of national identity are also high
in the GSS sample with over 80% of all respondents rating
themselves at 6 or above on a 0 to 10 scale on the importance of being American. Student support for symbolic
patriotism and constructive patriotism was similarly high
with relatively few reporting that the flag and anthem did
not make them feel good or that they did not support efforts at positive change. In contrast, uncritical patriotism
and nationalism were less pervasive. Sixty-one percent of
students disagreed in 2002 with the statement that they
would “support their country right or wrong,” and only
38% of the GSS subsample agreed somewhat or strongly
with the statement that “the world would be a better place
if other people were more like Americans.”
3
In 2002, respondents were asked “When you hear non-Americans
criticizing Americans to what extent do you feel personally criticized?” This item loaded on both the national identity and uncritical
patriotism scales and replaced the 2004 item on identity importance
in the national identity scale.
4
In the 2002 symbolic patriotism scale, an item was included that
asked respondents “How proud do you feel when you hear the
national anthem,” to replace the item on positive feelings about the
anthem.
5
The 2002 uncritical patriotism scale shared two items with the
2004 scale (protesters are good, and there is too much criticism of
the United States). It included one positive and one negative worded
item from the Schatz et al. (1999) blind patriotism scale: “If another
country disagreed with the an important United States policy that
I knew little about, I would not necessarily support my country’s
position” and “I would support my country right or wrong.” It
also included a third new item that asked “How angry does it make
you feel, if at all, when you hear someone criticizing the United
States,” adapted from the NES patriotism scale but which scaled
with uncritical patriotism.
68
LEONIE HUDDY AND NADIA KHATIB
TABLE 1 Factor Structure of National Attachment Items: Student Study, 2004
National
Identity
Symbolic
Patriotism
How important is being American to you?
1.00 (.21)
To what extent do you see yourself as a typical
.94 (.20)
American?
How well does the term American describe you?
.65 (.14)
When talking about Americans how often do you
.86 (.18)
say “we” instead of “they”?
How good does it make you feel when you see the
1.00 (.24)
American flag flying?
How good does it make you feel when you hear
.97 (.246)
the national anthem?
People should work hard to move this country in
a positive direction.
If I criticize the United States, I do so out of love
of country.
I oppose some U.S. policies because I care about
my country and want to improve it.
I express my attachment to America by
supporting efforts at positive change.
I support my country’s leaders even if I disagree
with their actions?
People who do not wholeheartedly support
American should live elsewhere.
For the most part, people who protest and
demonstrate against US policy are good,
upstanding, intelligent people.
The United States is virtually always right.†
I support US policies for the very reason that they
are the policies of my country.
There is too much criticism of the US in the world,
and we as its citizens should not criticize it.
I believe that US policies are almost always the
morally correct ones.
Correlations between Factors
National Identity
.744
Symbolic
Constructive
Constructive
Patriotism
Uncritical
Patriotism
1.00 (.08)
2.19 (.17)
2.50 (.19)
1.80 (.13)
1.00 (.20)
.84 (.17)
.56 (.11)
.94 (.19)
1.17 (.23)
1.13 (.22)
1.11 (.22)
.218
.265
.512
.558
−.197
Note: N = 300, One item-loading for each factor is constrained to 1.00 for identification. All items are rescaled to vary between 0 and 1.
Cells contain unstandardized factor loadings with standardized estimates in parentheses and standardized factor correlations. All factor
loadings and correlations are significant at the 1% level.
†
Indicates a reversed item.
The confirmatory factor structure for the 2004 student study is presented in Table 1. The three factors common to both the 2002 and 2004 student studies—national
identity, symbolic patriotism, and uncritical patriotism—
have almost identical structure with minor differences
noted in the text. The factor structure for the GSS identity subsample is included in Table 2.
The expected four-factor model—national identity,
symbolic patriotism, constructive patriotism, and uncritical patriotism—was a very good fit to the data in the 2004
69
AMERICAN PATRIOTISM REDEFINED AS AMERICAN IDENTITY
TABLE 2 Factor Structure of National Attachment Items: 1996 General Social Survey
National ID National Pride Nationalism
How important is being an American to you, where 0 is
1.00 (0.13)
not at all important and 10 is the most important thing
in your life?
To begin, we have some questions about where you live:
1.24 (0.16)
your neighborhood or village, your town or city, your
county, and so on. How close do you feel to America?
Some people say the following things are important for
1.41 (0.18)
being truly American. Other say they are not important.
How important do you think each of the following is:
- to feel American?
Are you proud of the way democracy works here?
Are you proud of economic achievements here?
Are you proud of your country’s science and technology
achievements?
Are you proud of your country’s history?
Are you proud of your country’s fair and equal treatment
of all groups in society?
Are you proud of your country’s achievements in arts and
literature?
Are you proud of your country’s social security system?
How strongly do you agree or disagree with the following:
America is a better country than most others?
How strongly do you agree or disagree with the
following: The world would be better if more people
from other countries were like Americans?
Correlations between Factors
National Identity
–
National Pride
1.00 (.17)
.94 (.15)
.69 (.11)
.80 (.13)
1.02 (.17)
.58 (.10)
.90 (.15)
1.00 (0.18)
.83 (0.15)
.61
.65
.46
Note: N = 678. One item-loading for each factor is constrained to 1.00 for identification. All items are rescaled to vary
between 0 and 1. Cells contain unstandardized factor loadings with standardized estimates in parentheses. All factor
loadings and correlations are significant at the 1% level.
student study. The ratio of the chi-squared value to the
degrees of freedom was below 2, an acceptably small value.
The RMSEA of .057 indicated a good fit and the normed
fit indices are also high: CFI = .943 and TLI = .932.6 The
four different national attachments are thus distinct. But
they are also related. National identity was strongly tied to
symbolic patriotism in both student studies, with a standardized factor correlation of .74 in 2004 and .68 in 2002.
National identity and uncritical patriotism were also posi-
6
Values of RMSEA (root mean squared error of approximation)
less than .10 indicate a good fit of the model to the data and
values less than .05 indicate a very good fit (Browne and Cudek
1993). The three-factor model was also a good fit to the 2002 data;
RMSEA = .049, CFI = .98 and TLI = .97.
tively related although the relationships were not as strong
(r = .51 in 2004 and r = .56 in 2002).
In contrast to other national attachments, constructive patriotism stood out as a distinct concept. The
standardized factor correlation between constructive patriotism and symbolic patriotism was a modest .27. Constructive patriotism was weakly and negatively correlated
with uncritical patriotism, consistent with Schatz and colleagues’ (1999) findings. And there was only a weak link
between national identity and constructive patriotism
(r = .22), suggesting that the two are unlikely to convey the
same political effects. Overall, the modest link between
constructive patriotism and other national attachment
measures reinforces earlier concerns about whether the
scale measures patriotism or detects political involvement
70
LEONIE HUDDY AND NADIA KHATIB
instead. Finally, despite strong ties among three of the four
national attachments, the four-factor model provided a
much better fit to the data than a one or two-factor model.
The GSS subsample includes measures of national
identity, national pride, and nationalism. National identity is assessed in the GSS with three items. National pride,
analogous to symbolic patriotism, is assessed with seven
items from de Figueiredo and Elkin’s (2003) scale.7 Nationalism, which is typically linked to uncritical patriotism, was assessed with two items. The wording of all three
item sets is included in Table 2. Three clear factors emerged
in analysis of the GSS data. All three identity items load
on one factor, all seven national pride items load on a second, and both nationalism items load on a distinct third
factor.8 National identity and national pride are positively
related (r = .61), and national pride is positively related
to nationalism (r = .65). National pride and nationalism
were more weakly related, although still closely aligned
(r = .46). Moreover, the three-factor model was a very
good fit to the data. The ratio of the chi-squared value to
the degrees of freedom was 3.47, an acceptably low value.
The RMSEA of .06 indicated a good fit and the normed
fit indices are reasonably high: CFI = .92 and TLI = .90.
The GSS analysis thus confirms that national identity is
distinct from, but positively related to, nationalism and
national pride. Once again, the three-factor model was a
better fit to the data than a single or two factor model.
The Origins of National Identity,
Patriotism, and Nationalism
Social identity theory predicts that national identity will
be less influenced by political ideology than other forms
of patriotism and nationalism. We test this connection by
regressing national attachments on self-identified ideology and party identification, measured on a 5-point scale
in the student studies and 7 points in the GSS. Analyses
also include a series of demographic controls: participant
race (dummy variables for Black, White, other in the GSS
and White, Black, Latino, and Asian in the student study),
first (born outside the United States), second (both par-
ents born outside the United States), and third immigrant
generation (both parents born in the United States), and
the percent of an immigrant’s life spent in the United
States (student studies only). Authoritarianism was included in GSS analyses and is measured as a latent construct indicated by two items: how strongly respondents
favor spanking as a form of discipline for children and
how important it is for children to obey.9
The political and demographic determinants of national attachments were estimated in regression analyses
conducted using Mplus® in which national attachments
are modeled as latent constructs, and a single model was
estimated for each of the three studies. Overall, the three
models demonstrated very good fit. In the student studies, the ratio of the chi-squared value to the degrees of
freedom was 1.59 (2002), and 1.79 (2004). The RMSEA
was .042 in 2002 and .051 in 2004, indicating a good fit in
both years. And the normed fit indices are suitably high:
CFI = .96 (2002) and .91 (2004), and TLI = .94 (2002)
and .90 (2004). The GSS model also performed reasonably well. The ratio of the chi-squared value to the degrees
of freedom was 2.72, with an RMSEA of .025, CFI = .861,
and TLI = .814. Findings for the student studies are presented in Table 3 and for the GSS in Table 4.
Consistent with a distinction in social identity theory between identity and its meaning, national identity
is less ideological than symbolic, constructive, or uncritical patriotism. National identity was unrelated to
political ideology or partisanship in both student studies. In contrast, conservatives scored consistently higher
than liberals on uncritical patriotism in both studies. Unexpectedly, partisanship also affected uncritical patriotism in the 2004 student study, perhaps reflecting the
Republican administration’s recent emphasis on patriotic
loyalty in connection to the Iraq war. The effects of ideology on symbolic patriotism are more mixed. Ideology
strongly shaped symbolic patriotism in the 2002 student
study but had no effect in 2004. Finally, ideology also influenced constructive patriotism but it was liberals, not
conservatives, who scored more highly on the scale. While
past research has found no link between ideology and constructive patriotism, stronger liberal endorsement of scale
items is consistent with the items’ emphasis on progressive
change.
7
Two items were dropped from de Figueiredo and Elkins (2003)
scale. Feeling ashamed of one’s country did not load on the national
pride factor in our analyses. And feeling close to America loaded
on the national identity not national pride factor.
8
We also modeled two error correlations—one between two of
the national identity items (closeness and feeling American, which
had similar response options) and one between pride in the country’s history and pride in the country’s fair and equal treatment of
groups.
9
GSS respondents ranked the following values in order of their
importance “for a child to learn to prepare him or her for life”: to
obey, be popular or well-liked, think for himself or herself, work
hard, and help others when they need help. A dummy variable was
created for those who ranked “obey” above all else. They were also
asked how strongly they agreed or disagreed with the following
statement: “It is sometimes necessary to discipline a child with a
good, hard spanking.”
71
AMERICAN PATRIOTISM REDEFINED AS AMERICAN IDENTITY
TABLE 3 Determinants of National Attachments:
Student Study, 2004 (N = 300)
National
ID
∗∗
Constructive
Patriotism
Symbolic
Patriotism
∗∗
Student Study, 2002 (N = 341)
Uncritical
Patriotism
Percent of life in .003 (.001) .000 (.000)
.003 (.001) .000 (.001)
the U.S.
First generation −.076 (.044) −.011 (.018)
.000 (.052) −.046 (.044)
Second
−.078∗ (.034) −.002 (.014) −.055 (.040) −.043 (.033)
generation
Age
.003 (.003)
.002 (.001) .005 (.004) −.002 (.003)
Female
−.027 (.024) −.023∗∗ (.011) −.061∗ (.028) −.001 (.024)
Black
−.109∗∗ (.037) −.031∗ (.016) −.158∗∗ (.043) −.075∗ (.037)
Latino
−.064 (.041) −.008 (.017) −.078 (.048) −.017 (.040)
Asian
−.004 (.034)
.013 (.014) −.005 (.040) −.024 (.034)
Party (Strong
.083 (.050) −.029 (.020)
.044 (.058)
.130∗∗ (.050)
Republican)
Ideology
−.029 (.053) −.060∗∗ (.023) .085 (.062)
.164∗∗ (.053)
(Strong
Conservative)
National ID
Symbolic
Patriotism
Uncritical
Patriotism
.003∗∗ (.001) .003∗∗ (.001) .002∗ (.001)
−.007 (.032) −.032 (.068) −.001 (.051)
−.028 (.019) −.115∗∗ (.039) −.063∗ (.029)
.004 (.002)
.007 (.004) −.003 (.003)
−.009 (.015) −.051 (.031)
.023 (.023)
−.064∗∗ (.024) −.193∗∗ (.048) −.093∗∗ (.036)
−.047 (.031)
.031 (.064) −.061 (.047)
−.021 (.022) −.138∗∗ (.046) −.039 (.034)
.004 (.031) −.013 (.065) −.042 (.048)
.043 (.036)
.356∗∗ (.074) .345∗∗ (.062)
Note: Cells contain unstandardized parameter estimates with standard errors in parentheses. All variables vary from 0 to 1 except age,
which is measured in years, and percent of life in the United States, which is a percentage. ∗ p < .05; ∗∗ p < .001.
The GSS included a measure of authoritarianism,
providing a direct test of the relationship between authoritarianism and nationalism. Political ideology has no
direct impact on nationalism in the GSS data. But authoritarianism does. This is somewhat surprising given the
strong connection between ideology and uncritical patriotism in the student data. But findings are consistent with
earlier theoretical expectations that authoritarianism explains the link between ideology and uncritical patriotism
(and by extension, nationalism). The impact of authoritarianism on nationalism is sizeable and helps to explain
why nationalism (and uncritical patriotism) is associated
with lower levels of political involvement. Authoritarians typically support strong leaders and are intolerant of
active political dissent; together both forces diminish active political participation (Altemeyer 1988; Feldman and
Stenner 1997). At odds with our expectations, national
pride was not ideological in nature and was no stronger
among conservatives than liberals.
When taken together, data from both student studies
and the GSS indicate that national identity was thoroughly
nonideological in nature, consistent with social identity
theory. In contrast, symbolic, constructive, and uncritical
patriotism, and nationalism were more strongly endorsed
by either liberals or conservatives (or authoritarians in the
case of nationalism). National pride was the only other
measure of national attachment that was not ideological
in nature.
One other notable aspect of national identity is the
ease with which it develops among immigrants. Not surprisingly, immigrants who had spent longer in the United
States more strongly endorsed all forms of national attachment except constructive patriotism in the student
studies. When national identity was added as a predictor of symbolic and uncritical patriotism to the models depicted in Table 3 (in analyses not shown here), it
fully mediated the impact of percent of life spent in the
United States on both dependent variables. This suggests
that national identity develops rapidly among young new
immigrants and intensifies other forms of national attachment as a consequence. Of course, these mediational
analyses imply that national identity is causally prior to
symbolic and uncritical patriotism, a casual order that
is not well established. National identity remained unaffected by subsequent immigration generation in the 2002
student study but was weaker among second- than thirdgeneration students in 2004. This is an unexpected finding
that was also observed for symbolic and uncritical patriotism. Immigrant generation had no effect on national
identity, national pride, or nationalism in the GSS data.
Finally, there are several other trends in Tables 3 and
4 that deserve mention. In both student studies, blacks
72
LEONIE HUDDY AND NADIA KHATIB
TABLE 4 Determinants of National
Attachments: 1996 General Social
Survey
First generation
Second
generation
Age
Years of
Education
White
Black
Female
Party (Strong
Republican)
Ideology (Strong
Conservative)
Authoritarianism
National
ID
National
Pride
Nationalism
−.041
(.033)
.009
(.015)
.002∗∗
(.000)
−.008∗∗
(.003)
−.021
(.031)
−.041
(.035)
−.018
(.013)
−.019
(.021)
.057
(.033)
.070
(.038)
−.003
(.037)
.029
(.018)
.004∗∗
(.000)
−.002
(.003)
−.051
(.038)
−.094∗
(.041)
−.031∗
(.015)
.030
(.025)
.022
(.040)
−.019
(.047)
−.020
(.038)
−.003
(.019)
.002∗∗
(.000)
−.011∗∗
(.003)
.064
(.039)
.035
(.043)
−.035∗
(.016)
.001
(.042)
.078
(.042)
.112∗
(.052)
Note: N = 676. Cells contain unstandardized parameter estimates
with standard errors in parentheses. Authoritarianism is estimated
as a latent variable consisting of two items. All variables vary from
0 to 1 except age and education, which are both measured in years.
∗
p < .05; ∗∗ p < .001.
expressed lower levels of all forms of national attachment, and Asians expressed lower levels of symbolic but
not uncritical patriotism or national identity. Blacks also
expressed lower levels of national pride but not national
identity or nationalism in the GSS.10 Women in the GSS
subsample were somewhat less nationalistic and scored
lower than men on the national pride scale, but did not
differ in strength of national identity or other forms of
patriotism (including uncritical patriotism) in the student studies. Older adults were more likely than younger
individuals to hold a strong identity, express greater pride,
and endorse nationalism in the GSS data. And better educated individuals held a weaker national identity than
those with less education and scored lower on the nationalism scale. The latter finding raises questions as to
10
Lower levels of pride among blacks disappear once the two items
that refer to pride in the equal treatment of groups, and the country’s
history are removed from the national pride factor in analyses not
shown here.
whether higher levels of education are associated with
diminished national identity because better educated individuals are less nationalistic (and nationalism and national identity are related). There was some support for
this hypothesis in added mediational analyses (not shown
here) in which nationalism fully mediated the effects of
education on national identity. But once again, such findings need to be treated with caution given uncertainty as
to the true causal relationship between national identity
and different forms of patriotism and nationalism.
Political Involvement
Political Attention and Knowledge
We now turn to the positive impact of national identity
on political engagement. Social identity theory predicts
higher levels of political involvement among strong national identifiers because of their greater adherence to
group norms. In contrast, other forms of patriotism are
not expected to promote political involvement. Political
involvement was assessed as attention to politics, knowledge of current events, and voter turnout. We consider
political attention and knowledge first. The 2002 student
study focused on political attention to and knowledge
of events in the Middle East; the 2004 student study assessed attention to electoral politics. Students were asked
two questions in 2002: how closely they were “following
news stories about events in Iraq” and how much thought
they had given in the past week or two “to a possible
war with Iraq.” In the same 2002 study, students’ political
knowledge concerning the Middle East was assessed with
five factual questions (asking respondents to name one
country bordering Iraq, the Iraqi capital city, the Middle
Eastern TV network broadcasting statements by Osama
bin Laden, the ruling Iraqi political party, and the ethnic
group in northern Iraq), treated as dichotomous indicators of a continuous underlying latent knowledge score.
The amount of attention paid to, and closely following
news about, the 2004 presidential election served as two
indicators of political attention in 2004. All models are estimated using Mplus® . Analyses for attention and knowledge are presented in the first three columns of Table 5.
Consider first the impact of national identity on political attentiveness. Consistent with social identity theory,
holding a strong national identity increased attentiveness
to events in Iraq in 2002. The same pattern is observed
for attentiveness to the presidential election in 2004, and
knowledge of Iraq in 2002. In all three instances, national
identity has a large, positive effect on political attention
and knowledge. In contrast, constructive patriotism has
73
AMERICAN PATRIOTISM REDEFINED AS AMERICAN IDENTITY
TABLE 5 National Attachments and Political Involvement
Student Study
National Identity
National Pride
Symbolic Patriotism
Uncritical Patriotism
Nationalism
Constructive Patriotism
Percent of life in the United States
First generation
Second generation
Age
Years of Education
Female
White
Black
Latino
Asian
Party (Strong Republican)
Ideology (Strong Conservative)
Authoritarianism
N
1996 GSS
Attention to
Politics
(2002)
Attention to
Politics
(2004)
Middle East
Knowledge
(2002)
.463∗∗ (.195)
–
−.019 (.100)
−.295∗ (.146)
–
–
−.001 (.001)
−.037 (.057)
−.028 (.033)
.006 (.004)
–
−.072∗∗ (.029)
–
−.072 (.040)
−.092 (.055)
−.122∗∗ (.041)
.023 (.054)
.026 (.071)
–
341
.381∗∗ (.127)
–
−.086 (.100)
−.346∗∗ (.110)
–
.422 (.248)
−.001 (.001)
.007 (.045)
.088∗ (.035)
.006∗ (.003)
–
−.060∗ (.025)
–
−.098∗ (.038)
−.003 (.042)
−.133∗∗ (.042)
−.064 (.050)
.119∗ (.059)
–
300
1.265∗∗ (.516)
–
−.086 (.326)
−1.880∗∗ (.494)
–
–
−.008 (.004)
−.389 (.251)
−.044 (.137)
.014 (.018)
–
−.561∗∗ (.123)
–
−.247 (.164)
−.169 (.246)
−.162 (.164)
−.133 (.203)
.522∗ (.257)
–
309
Vote
(2002)
Vote
1.676∗ (.809)
3.248∗∗ (1.07)
–
−.077 (.635)
−.066 (.572)
–
−1.579∗ (.734)
–
–
−2.298∗∗ (.718)
–
–
.018 (.012)
–
.062 (.448)
.817 (.439)
.063 (.186)
.172 (.156)
∗∗
.071 (.019)
.024∗∗ (.004)
–
.134∗∗ (.022)
.104 (.156)
.179 (.122)
–
−.039 (.352)
−.258 (.237)
.294 (.384)
−.368 (.296)
–
−.542∗ (.253)
–
−.431 (.318)
.225 (.221)
.719 (.415)
−.014 (.293)
–
.232 (.447)
309
640
Note: Cells contain unstandardized parameter estimates with standard errors in parentheses. Middle East knowledge and voting are
estimated as categorical traits, latent in the case of knowledge and observed in the case of voting. All variables vary from 0 to 1 except age,
years of education (both measured in years), and percent of life in the United States.
∗
p < .05, ∗∗ p < .001.
no effect on political involvement with or without national identity in the analyses. This contradicts past evidence and suggests a weak relationship overall between
constructive patriotism and political involvement. Symbolic patriotism has no independent effect on political
attention or knowledge.
Uncritical patriotism is the only other national attachment to significantly affect political involvement, and
its effects are negative as expected. Uncritical patriots are
less likely to attend to the Iraq war or campaign politics
and know less about the Middle-East. Moreover, the positive effects of national identity and the negative effects
of uncritical patriotism are intertwined (since the two are
strongly positively related yet have powerfully opposing
political effects). The strong positive effects of national
identity hinge on controlling for uncritical patriotism and
disappear in a similar equation that predicts either attentiveness or knowledge (not shown here). This suggests
that American identity heightens political engagement,
but only after controlling for individuals who are reluctant to criticize the government and its political leaders
(and who also have a strong national identity).
Voter Turnout
Voter turnout was assessed in the 2002 study with a question asking students whether they had voted in “the elections this November for Congress and other offices?”
Turnout in the 1992 presidential election was assessed in
the 1996 GSS.11 Voter turnout is regressed onto national
attachments and other demographic controls for both the
GSS and 2002 student study, in probit models estimated
11
The GSS question was “In 1992, you remember that Clinton
ran for President on the Democratic ticket against Bush for the
Republicans and Perot as an Independent. Do you remember for
sure whether or not you voted in that election”?
74
in Mplus® via weighted least squares.12 National identity
predicted voter turnout in both studies, as seen in the last
two columns of Table 5. The coefficient for national identity is large, positive, and significant in the GSS. It is also
significant in the student data although the effect is not
quite as large. Both uncritical patriotism and nationalism
depress voter turnout, as expected.
The power of national identity to enhance and uncritical patriotism or nationalism to depress voter turnout is
made apparent by examining differences in the predicted
probability of voting at different levels of each type of national attachment. Predicted probabilities are calculated
from the equations presented in Table 5 for individuals
at the 25th (low) and 75th (high) percentiles of national
identity and uncritical patriotism or nationalism. All
turnout probabilities are estimated for third-generation
American, white males who are partisan independents
and political moderates, and score at the mean on all
other independent variables. Consider first the impact
of national identity in the student study. For individuals
who score at the mean on uncritical patriotism, predicted
turnout ranges from a low of .53 among weak national
identifiers to a high of .70 among stronger identifiers.
Levels of turnout are higher in the GSS sample because
the sample is older and confined to citizens. In the GSS,
turnout ranges from a low of .86 among weak identifiers
to a high of .99 among strong identifiers.
In the student study, uncritical patriotism and nationalism have roughly the same effect on voter turnout
as national identity, but in the opposite direction. Thus
for students at the mean on national identity strength,
the predicted probability of voting ranges from a high of
.70 for those low in uncritical patriotism to a low of .48
among those who are high. Likewise, turnout ranges from
a predicted probability of .99 for Americans in the GSS
subsample low in nationalism to .82 among those high in
nationalism and at the mean on national identity.
Other factors play an expected role in driving voter
turnout, political interest, and knowledge. Older and better educated Americans were more likely to report having voted in the 1992 presidential election. Age increased
turnout and attention to politics among students in 2002.
Women were less attentive and knowledgeable about Iraq,
but no less likely to have voted than men in either the GSS
or the student study. Asian students were less attentive to
politics and less likely to have voted in the 2002 election,
independently of immigrant status. Finally, conservative
students were more knowledgeable about Iraq in 2002
12
All models in Table 5 have decent model fit, although the worst
fit occurs for the vote model, which still falls within acceptable fit
limits (RMSEA below .1).
LEONIE HUDDY AND NADIA KHATIB
and more attentive to politics in 2004, although this effect
emerged only with the inclusion of uncritical patriotism.
Once uncritical patriotism is removed, conservatives are
no more knowledgeable than liberals.
Conclusion
This research underscores the contribution of social identity theory to the study of patriotism, a field of inquiry
that has typically lacked a coherent theoretical focus. According to the theory, a social identity develops readily
among members of a salient group and does not depend
on a common set of beliefs or shared outlook (Tajfel and
Turner 1979; Turner et al. 1987). This prediction is borne
out in the current study in which national identity was
consistently unrelated to political ideology. In contrast, at
least three of the other four forms of assessed patriotism
exhibit some degree of ideological bias.
Social identity theory also predicts greater adherence
to group norms among strong group identifiers, translating into greater civic involvement among strong national
identifiers in the case of American identity. This prediction received impressive support in our data. Americans
with a strong national identity paid more attention to
politics, knew more about current events, and were more
likely to vote. Moreover the connection between national
attachment and civic engagement is not predicted by any
other approach to patriotism. The connection between
patriotism and civic involvement has been empirically
tested by Schatz and colleagues (1999). But findings from
the current research make clear that national identity is
the only form of national attachment to positively predict
political involvement. Overall, social identity theory provides guidance on the measurement of national attachments, produces a nonideological measure that evades
contention over the meaning of American patriotism,
and generates testable and empirically substantiated predictions that underscore its contribution to research on
patriotism.
Social identity theory generates a number of other
predictions in addition to those tested in the current research. For example, some researchers have concluded
that patriotism is unrelated to outgroup animosity (De
Figueiredo and Elkins 2003; Kosterman and Feshbach
1989). But this conclusion is unlikely to hold during periods of international conflict. According to social identity
theory, ingroup and outgroup sentiments are typically
distinct because groups evade direct competition for social prestige by recognizing distinct positive ingroup and
outgroup characteristics. This possibility is eliminated
75
AMERICAN PATRIOTISM REDEFINED AS AMERICAN IDENTITY
under conditions of direct outgroup threat which produce outgroup hostility among strong ingroup identifiers (Branscombe et al. 1999; Duckitt 2003; Huddy 2003;
Tajfel and Turner 1979). Social identity theory thus posits
a link between patriotism, nationalism, and outgroup animosity under conditions of national threat. The theory
generates other relevant predictions such as intensified
national identity, more positive feelings towards conationals heightened, and conformity to national traits
when one is in the minority (during overseas travel, for
example; Haslam et al. 1992; Huddy 2003; Mullen, Brown,
and Smith 1992; Spears et al. 1997; Turner et al. 1987).
The positive connection between national identity
and political involvement mirrors the past political effects of constructive patriotism. But our research demonstrates that the effects of national identity are far more
robust. Indeed, we uncovered a series of problems with
the constructive patriotism scale. First, it was relatively
unrelated to national identity and other measures of patriotism, suggesting that it is not part of the same broad
concept. Second, it did not predict political interest even
when national identity was removed from the analysis,
raising questions about its predictive power. And third,
it was more strongly endorsed by liberals than conservatives, demonstrating ideological bias and weakening its
claim as a broad measure of patriotism. Our findings
differ markedly from those reported by Schatz and colleagues, and we can only speculate as to why. We included
a measurement model and more extensive controls than
Schatz, Staub, and Lavine (1999) in our analysis. But ultimately, further research is needed to fully resolve these
discrepancies.
The current findings not only highlight the positive
impact of national identity on political involvement; they
also reveal the negative effects of uncritical patriotism. Despite being labeled patriots, students who scored highly
on uncritical patriotism paid less attention to news about
Iraq, were less knowledgeable about events there, and were
less likely to have voted in 2002. The same holds for strong
nationalists who were less likely than others to have voted
in the 1992 presidential election in GSS data. Uncritical
patriotism and nationalism are more ideological than national identity because they are grounded in authoritarianism (as seen in this study for nationalism and in Schatz,
Staub, and Lavine 1999, for uncritical patriotism). And the
link between uncritical patriotism and authoritarianism
helps to account for the negative effects of uncritical patriotism on political involvement because authoritarians
are more likely than others to abnegate decision-making
powers to their leaders.
Despite its popularity as a measure of patriotism
within political science, symbolic patriotism emerged as
the least effective measure of national attachment in this
research. It is ideological in flavor (in one of our studies
and in other research; see Hurwitz and Peffley 1999). It
also overlaps heavily with national identity and uncritical
patriotism. In fact, one of the standard NES symbolic patriotism items—reacting angrily to criticism of the United
States—loaded on the uncritical patriotism construct in
this research. Moreover, symbolic patriotism has no independent effect on measures of political engagement. We
suggest that future research abandon this concept in favor of its two components: a more ideologically neutral
measure of national identity and the politically powerful
concept of uncritical patriotism. The latter two measures
provide a potent demonstration of the divergent effects
of different forms of national attachments on political
involvement that are masked by single measures such as
symbolic patriotism.
In sum, our research highlights the value of a social
identity approach to the study of national attachments.
Past research on patriotism and nationalism has produced
a variety of scales and concepts and has often been characterized by conflicting and contradictory terminology and
empirical measurement. The introduction of social identity theory adds a strong theoretical focus to this research
and generates national identity, an effective measure of
national attachment. In the end, a national identity scale
may prove to be an important (and nonideological) addition to research on the study of political participation
very generally. The introduction of national identity as
a measure of national attachment may also help to address Hurwitz and Peffley’s (1999) concern that ideological measures of patriotism do not tap broad support for
the nation but rather capture more narrow ideological
allegiances. Overall, a social identity approach sharpens
research on patriotism at a time when Americans are confronted with a proliferation of claims about patriotism’s
virtues and vices.
References
Adorno, Theodor, Else Frenkel-Brunswick, Daniel J. Levinson,
and R. Nevitt Sanford. 1950. The Authoritarian Personality.
New York: Harper.
Altemeyer, Bob. 1988. Enemies of Freedom. San Francisco: Jossey
Bass.
Altemeyer, Bob. 1996. The Authoritarian Specter.
Cambridge: Harvard.
Altemeyer, Bob. 1998. “The Other ‘Authoritarian Personality.’”
Advances in Experimental Social Psychology 30(1):48–92.
Branscombe, Nyla R., Naomi Ellemers, Russell Spears, and
Bertjan Doosje. 1999. “The Context and Content of Social
76
Identity Threat.” In Social identity: Context, Commitment,
Content, ed. Naomi Ellemers, Russell Spears, and Bertjan
Doosje. Oxford: Blackwell, pp. 35–58.
Browne, Michael W., and Robert Cudek. 1993. “Alternative Ways
of Assessing Model Fit.” In Testing Structural Equation Models, ed. Kenneth A. Bollen and J. Scott Long. Newbury Park,
CA: Sage, pp. 136–62.
Christensen, P. Niels, Hank Rothberger, Wendy Wood, and
David C. Matz. 2004. “Social Norms and Identity Relevance:
A Motivational Approach to Normative Behavior.” Personality and Social Psychology Bulletin 30(10):1295–1309.
Citrin, Jack, Cara Wong, and Brian Duff. (2001). “The Meaning
of American National Identity: Patterns of Ethnic Conflict
and Consensus.” In Social Identity, Intergroup Conflict, and
Conflict Reduction, ed. Richard D. Ashmore and Lee Jussim.
London: Oxford University Press, pp. 71–100.
Conover, Pamela J., and Stanley Feldman. 1987. “Memo to NES
Board of Overseers Regarding ‘Measuring Patriotism and
Nationalism’.” http://www.icpsr.umich.edu.
Conover, Pamela J., Donald D. Searing, and Ivor Crewe. 2004.
“The Elusive Ideal of Equal Citizenship: Political Theory and
Political Psychology in the United States and Great Britain.”
The Journal of Politics 66(4):1036–1068.
De Figueiredo, Rui J. P., and Zachary Elkins. 2003. “Are Patriots Bigots? An Inquiry into the Vices of In-Group Pride.”
American Journal of Political Science 47(1):171–88.
Devos, Thierry, and Mahzarin R. Banaji. 2005. “American
= White?” Journal of Personality and Social Psychology
88(3):447–66.
Duckitt, John. 2003. “Prejudice and Intergroup Hostility.”
In Oxford Handbook of Political Psychology, ed. David O.
Sears, Leonie Huddy, and Robert Jervis. New York: Oxford
University Press, pp. 559–600.
Elkins, Zachary, and John Sides. 2003. “The Foundations of
National Identity: Individual, State-Level, and Temporal factors.” Presented at the annual meeting of the Midwest Political Science Association.
Feather, Norman T., Robert J. Boeckmann, and Ian R. McKee.
2001. “Reactions to an Offense in Relation to Authoritarianism, Knowledge about Risk, and Freedom of Action.” European Journal of Social Psychology 31(2):109–26.
Feldman, Stanley, and Karen Stenner. 1997. “Perceived Threat
and Authoritarianism.” Political Psychology 18(4):741–70.
Gibson, James L., and Amanda Gouws. 2000. “Social Identities and Political Intolerance: Linkages Within the South
African Mass Public.” American Journal of Political Science
44(2):272–86.
Haslam, S. Alexander, John C. Turner, Penelope J. Oakes, Craig
McGarty, and Brett K. Hayes. 1992. “Context-Dependent
Variation in Social Stereotyping 1: The Effects of Intergroup Relations as Mediated by Social Change and Frame of
Reference.” European Journal of Social Psychology 22(1):3–
20.
Huddy, Leonie. 2001. “From Social to Political Identity: A
Critical Examination of Social Identity Theory.” Political
Psychology 22(1):127–56.
Huddy, Leonie. 2003. “Group Membership, Ingroup Loyalty,
and Political Cohesion.” In Handbook of Political Psychology,
ed. David O. Sears, Leonie Huddy, and Robert Jervis. New
York: Oxford University Press, pp. 511–58.
LEONIE HUDDY AND NADIA KHATIB
Huddy, Leonie, Stanley Feldman, Charles Taber, and Gallya
Lahav. 2005. “Threat, Anxiety, and Support of AntiTerrorism Policies.” American Journal of Political Science
49(3):610–25.
Huddy, Leonie and Nadia Khatib (in press). Social Identity
Strength and Meaning: A New Look at American Patriotism. In Yoshinobu Araki and Minoru Karasawa (Eds.),
Inter-Group Behavior and Political Psychology, Tokyo: Brain
Publishing Company.
Hurwitz, Jon, and Mark Peffley. 1999. “International Attitudes.”
In Measures of Political Attitudes, ed. John P. Robinson,
Phillip R. Shaver, and Lawrence S. Wrightsman. San Diego:
Academic Press, pp. 533–90.
Karasawa, Minoru. 2002. “Patriotism, Nationalism, and Internationalism among Japanese Citizens: An Etic-Emic Approach.” Political Psychology 23(4):645–66.
Kosterman, Rick, and Seymour Feshbach. 1989. “Toward a
Measure of Patriotic and Nationalistic Attitudes.” Political
Psychology 10(2):257–74.
Lavine, Howard, Milton Lodge, and Kate Freitas. 2005. Threat,
Authoritarianism, and Selective Exposure to Information.
Political Psychology, 26(2):219–44.
McFarland, Sam, and Melissa Mathews. 2005. “Who Cares about
Human Rights?” Political Psychology 26(3):365–85.
McCrae, Robert R. 1996. “Social Consequences of Experiential
Openness.” Psychological Bulletin 120(3):323–37.
Mullen, Brian, Rupert Brown, and Colleen Smith. 1992. “Ingroup Bias as a Function of Salience, Relevance, and Status: An Integration.” European Journal of Social Psychology
22(2):103–22.
Mummendey, Amelie, Andreas Klink, and Rupert Brown. 2001.
“Nationalism and Patriotism: National Identification and
Out-Group Rejection.” British Journal of Social Psychology
40(1):159–72.
Muthén, Linda K., and Bengt O. Muthén. 2001. Statistical Analysis with Latent Variables. Los Angeles: Muthén & Muthén.
Osterreich, Detlef. 2005. “Flight into Security: A New Approach
and Measure of Authoritarian Personality.” Political Psychology 26(3):275–97.
Petersen, Bill E., Lauren E. Duncan, and Joyce S. Pang. 2002.
“Authoritarianism and Political Impoverishment: Deficits
in Knowledge and Civic Disinterest.” Political Psychology
23(1):97–112.
Rothi, Despina M., Evantia Lyons, and Xenia Chryssochoou.
2005. “National Attachment and Patriotism in a European
Nation: A British Study.” Political Psychology 26(1):135–55.
Schatz, Robert T., and Ervin Staub. 1997. “Manifestations of
Blind and Constructive Patriotism: Personality Correlates
and Individual-Group Relations.” In Patriotism: In the Lives
of Individuals and Nations, ed. Daniel Bar-Tal and Ervin
Staub. Chicago: Nelson-Hall, pp. 229–45.
Schatz, Robert T., Ervin Staub, and Howard Lavine. 1999. “On
the Varieties of National Attachment: Blind versus Constructive Patriotism.” Political Psychology 20(1):151–74.
Schuman, Howard, and Stanley Presser. 1981. Questions and
Answers in Attitude Surveys: Experiments in Question Form,
Wording, and Context. Thousand Oaks: Sage.
Sidanius, Jim, Seymour Feshbach, Shana Levin, and Felicia
Pratto. 1997. “The Interface between Ethnic and National
Attachment.” Public Opinion Quarterly 61(1):103–33.
AMERICAN PATRIOTISM REDEFINED AS AMERICAN IDENTITY
Sniderman, Paul M., Louk Hagendoorn, and Markus Prior.
2004. “Predisposing Factors and Situational Triggers: Exclusionary reactions to Immigrant Minorities.” American Political Science Review 98(1):35–49.
Spears, Russell, Penelope J. Oakes, Naomi Ellemers, and
S. Alexander Haslam. 1997. The Social Psychology of Stereotyping and Group Life. Oxford: Blackwell.
Spinner-Halev, Jeff, and Elizabeth Theiss-Morse. 2003. “National identity and Self-Esteem.” Perspectives on Politics
1(3):515–632.
Sullivan, John L., Amy Fried, and Mary G. Dietz. 1992.
“Patriotism, Politics, and the Presidential Election of
1988.” American Journal of Political Science 36(2):200–
34.
Tajfel, Henri. (1981). Human groups and Social Categories.
Cambridge: Cambridge University Press.
Tajfel, Henri, and John C. Turner. 1979. “An Integrative Theory
of Intergroup Conflict.” In The Social Psychology of Intergroup
Relations, ed. W. G. Austin and Stephen Worchel. Monterey:
Brooks/Cole, pp. 33–47.
77
Terry, Deborah J., and Michael A. Hogg. 1996. “Group Norms
and the Attitude-Behavior Relationship: A Role for Group
Identification.” Personality and Social Psychology Bulletin
22(8):776–93.
Terry, Deborah J., Michael A. Hogg, and Julie M. Duck. 1999.
“Group Membership, Social Identity and Attitudes.” In Social
Identity and Cognition, ed. Dominic Abrams and Michael A.
Hogg. Oxford: Blackwell Publishers, pp. 280–314.
Terry, Deborah J., Michael A. Hogg, and Katharine M. White.
1999. “The Theory of Planned Behavior: Self-Identity, Social Identity, and Group Norms.” British Journal of Social
Psychology 38(3):225–44.
Turner, John C., Michael A. Hogg, Penelope J. Oakes, Stephen
D. Reicher, and Margaret S. Wetherell. 1987. Rediscovering
the Social Group: A Self-Categorization Theory. Oxford: Basil
Blackwell.
Wellen, Jackie M., Michael A. Hogg, and Deborah J. Terry.
“Group Norms and Attitude-Behavior Consistency: The
Role of Group Salience and Mood.” Group Dynamics
2(1):48–56.