ABSTRACT
Tolerance and Support for Expression: the Same Construct?
by
Julie L. Andsager
Dept. of Journalism, Box 64
Middle Tennessee State University
Murfreesboro, TN 37132
[log in to unmask]
Although the terms tolerance and support are often used almost interchangeably
in
free-expression research, they have different conceptual and operational
definitions. This
experiment, in seeking to determine which term more accurately reflects how
people think
about free speech, found a moderate correlation between tolerance and support,
which
decreased as expression became more threatening. Subjects who were less tolerant
were most
likely to voice their opinions on expression. Tolerance seemed to better
describe the
underlying construct.
Tolerance and Support for Expression: the Same Construct?
Given the interconnection of subjects included under the rubric of social
science, it is
not surprising that some research topics overlap two or more disciplines.
Similar strands of
research often appear in political science and mass communication, for example.
One
specific case is that of research into public opinion and civil liberties:
Political scientists often
use the term tolerance to refer to people's attitude toward various expressive
acts, while
communication researchers seem to prefer the word support. Communication
researchers dip
freely into the tolerance literature to extrapolate in studies of support for
civil liberties.
Tolerance, however, tends to encompass attitudes toward a broader range of civil
liberties,
and support for free speech falls under that umbrella (see this use of terms
clearly illustrated in
Gibson & Bingham, 1982; emphasis added). Confounding even this categorization,
McClosky and Brill (1983) use the phrase "[s]upport for tolerance and civil
liberties" (p. 15)
to describe the construct they seek to measure.
Although the terms are often used interchangeably in the titles, literature
reviews and
discussion sections of studies on public opinion and civil liberties, they may
well represent
quite distinct constructs. Aside from the obvious nods to their origin
mentioned above, the
words themselves hold different connotations. Tolerance seems passive, support
active.
Indeed, Webster's New Collegiate Dictionary offers disparate denotations:
tolerance is
"sympathy or indulgence for beliefs or practices differing from or conflicting
with one's own;
the act of allowing something"; support means to "to promote the interests or
cause of; to
uphold or defend as valid or right."
Previous studies on tolerance and support sometimes hint at such a distinction,
but
few if any have directly addressed it. Indeed, it appears that only Nunn (1973)
consistently
referred to both terms separately, though he did not define them. Though the
authors of one
of the classic studies on support for democratic principles were not looking for
a relationship
between activism and tolerance/support, they concluded that "many people express
undemocratic principles in response to questioning but are too apathetic to act
on their
undemocratic opinions in concrete situations" (Prothro & Grigg, 1960, pp.
293-294). Prothro
and Grigg did not ask their respondents whether they would act on their
opinions, but based
their statement on observation of society. Because not acting on such opinions
(whether one
agrees or disagrees with democratic principles, for example) would seem inherent
in
tolerance, however, it is necessary to directly ask individuals whether they
would indeed act
to uphold or repress certain rights. On the other hand, the connotation of
support suggests
that one would act on his/her opinions.
If the terms tolerance and support are not synonymous, it may be a mistake to
fully
integrate them in building theory. Certainly a theory of how individuals think
about speech
and media rights would need to incorporate a notion of tolerance/support. Which
term
more accurately and parsimoniously captures the construct underlying people's
attitudes
toward free expression? Does voicing one's opinion on a potentially threatening
act of
expression better reflect (a lack of) tolerance or support for expression? The
purpose of this
exploratory study is to attempt to better delineate these terms through an
examination of
previous literature and to measure whether exercising one's own First Amendment
rights to
quell or uphold the expressive rights of others is related to measures of
tolerance and support.
While previous research has centered on the independent variables predicting
support/tolerance, this study will focus on defining the dependent variable.
Conceptual and Operational Definitions of Tolerance and Support
Tolerance is itself a relatively amorphous term, even among political
scientists who
study it. Communication scholars have attributed our lack of knowledge about
who is
tolerant and why to the disagreement on the conceptual definition of the term
(McLeod,
Steele, Chi, & Huang, 1991). In their 1982 study on conceptualization of
political tolerance,
Gibson and Bingham noted that a rigorous definition of the term would fail to
satisfy
scholars. For the most part, however, the conceptual definition put forth by
Sullivan,
Piereson and Marcus (1979) -- that individuals will allow expression and
existence of ideas
even when they object to those ideas -- seems to be accepted by a number of
researchers. In a
later study by Sullivan and other colleagues, this rationale is summed up
nicely: "It is easy to
afford civil liberties to groups we support; it is not so easy to tolerate those
we oppose"
(Sullivan, Shamir, Walsh, & Roberts, 1985, p. 77). As I will discuss further
below, the
requirement that an individual or group is disliked is a crucial distinction
between measures
of tolerance and support.
The negative affect for certain speakers or groups required in tolerance
research may
not be as important as other variables in determining whether the expression
should be
allowed, however, which has the potential to confound interpretation. Studies
suggest that
the construct of tolerance generalizes across levels of dislike, as individuals
who tend to be
(in)tolerant of one group are also likely to feel the same way about other
groups; in other
words, tolerance of expression in general is more predictive of tolerance of a
particular group
than affect toward that group (Bobo & Licari, 1989; Gibson, 1986; McLeod, Guo,
Huang,
Rzeszut, & Voakes, 1992).
Support is typically defined as the willingness to extend protection to various
acts of
expression (Andsager, 1993; Gibson & Bingham, 1982; Wyatt, 1991). This
definition is
broader than that of tolerance in that it includes expression which may not be
objectionable
(though of course noncontroversial expression items are seldom addressed in
surveys).
According to this definition, support is also slightly more active than
tolerance, as "extending
protection" suggests a stronger attitude than "allowing." One problem with this
conceptualization, however, is that individuals themselves may perceive they
have little
control over protection of rights; that is, the abstract nature of this notion
of support may
not reflect an actual willingness to promote, defend or uphold protection for
free speech.
Though it is beyond the scope of this study to discuss correlations between
attitudes and
behavior, a distinction such as this is important in determining the most
rigorous mechanism
to drive a theory. Because tolerance is by definition passive, abstractness
does not harm its
validity as much.
While tolerance may be passive, its supposed opposite is not. Intolerance is
defined as
the willingness to repress the rights of others (Gibson, 1986), which seems to
be the antonym
for support rather than tolerance. No satisfactory opposite of support exists,
although
McLeod, Guo, Huang, Rzeszut, and Voakes (1992) used the term non-support to
refer to a
negative feeling toward free expression. However, the non-supportive prong of
the two-
factor model of support that they posited closely resembles the intolerance
Gibson describes,
in that both are politically oriented and contain an inherent element of social
control.
Although the terms tolerance and support are often used interchangeably in
research
studies, a closer examination of survey items indicates that the
operationalization of how the
public thinks about free expression tends to either reflect one
conceptualization or the other.
Many studies focus on support. For example, a typical survey item stem asks
whether some
type of right should be protected by law:
l Should this type of speech be protected by the First Amendment? (Gibson &
Bingham, 1982);
l "Do you support the right of..." (McLeod, Steele, Chi, & Huang, 1991);
l "Should a person's rights be protected by law when..." (Wyatt, 1991).
Andsager's (1993) Support for Personal Rights Scale similarly contains
permutations of the
information contained in the following example, varying the act of expression
for each item:
"An individual's right to protest for or against abortion should be protected by
law."
Other studies use items that operationally define tolerance. Stouffer's 1955
survey
serves as the basis for many of these, using items that do not ask whether
rights should be
protected, but whether an extremist "should be allowed to speak, or not?" (p.
252). The
General Social Survey has used versions of Stouffer's items, which have been
included in a
number of studies on political tolerance (e.g., Bobo & Licari, 1989), and other
researchers
have also replicated parts of Stouffer's study (Prothro & Grigg, 1960; Sullivan,
Piereson, &
Marcus, 1979). McClosky and Brill (1983) asked respondents whether a variety of
expressive
acts should be allowed, but addressed the issue of rights in one of their
response sets.
Aside from the phrasing that directly relates to tolerance or support, the
focus of the
items in the two types of studies also differs. As alluded to above, studies in
tolerance mainly
center on the speaker, who must be suspect by definition -- a neo-Nazi, a
communist, a racist,
a homosexual or an atheist (Bobo & Licari, 1989; Stouffer, 1955) -- in order to
be tolerated.
In research on support, the expression itself is usually the focus, such as in
Wyatt's (1991)
extensive survey on free speech rights, which asked respondents about everything
from using
obscene gestures to committing espionage. Similarly, the Support for Personal
Rights Scale
includes items on an individual's protesting in public and using words that
might be offensive
to certain groups (Andsager, 1993). In a departure from this practice, McLeod
and his
colleagues (1991, 1992) combined tolerance's focus on the speaker with the "do
you support
the right" question. Gibson and Bingham (1982) contend that support for free
speech is
based heavily on the type of speech in question; this argument suggests that
even the speaker-
based tolerance items are being answered primarily in terms of the things
respondents expect
that a communist might say in his/her speech.
These phrasing peculiarities suggest the likely presence of artifacts in the
data on
public opinion on free expression. Exactly where those artifacts occur may be
difficult to
detect, however. As McClosky and Brill (1983) note, the source of the speech in
question
and the phrasing of the items in terms of support or tolerance may interact, at
least in the
respondent's mind:
Respondents who agree that radicals or atheists have the same general right to
speak as other Americans may nevertheless be unwilling to allow them to address
a school assembly or to hold a public meeting in the town hall. In these cases,
respondents may feel that they are being asked not merely to tolerate the
expression
of opinions by unpopular groups, but to abet them by granting them access to a
public forum. (p. 80; emphasis in original)
Clearly, a disparity exists in the question wording in these two sets of
surveys.
Restricting democratic rights is hardly the same as not allowing someone to
speak --
particularly someone whose words are potentially harmful. Even subtle
differences such as
these can strongly affect responses (Babbie, 1995). Others have already argued
that it is
nearly impossible to track public opinion on free expression over time because
the data have
been gathered from surveys that focus on timely topics (Andsager, 1993; Erskine,
1970). The
more glaring problem, however, lies in the attempt to examine those trends using
the
previous studies of tolerance and support almost without regard to the
underlying differences
in their question wording. Before we can move further toward building a theory
of public
support or tolerance for free expression, we need to determine which we are
really studying.
Nonetheless, tolerance and support are obviously related, if separate,
constructs.
They address similar issues relative to free speech, which suggests that they
might be closely
related, especially considering that the same demographic variables such as sex
and political
orientation have been consistently found to predict both. Previous research has
found that
men are typically more supportive/tolerant than women (Miller, Andsager, &
Wyatt, 1992;
Nunn, 1973; Stouffer, 1955), and those who identify themselves as liberal tend
to be more
supportive/tolerant than self-reported moderates and conservatives (Bobo &
Licari, 1989;
McLeod, Guo, Huang, Rzeszut, & Voakes, 1992; McLeod, Steele, Chi, & Huang, 1991;
Miller, Andsager, & Wyatt, 1992). The differences in operational definition
discussed above,
however, should prevent tolerance and support from being highly correlated.
Therefore,
H1. Tolerance and support will be moderately correlated.
H2. Sex will be related to both support and tolerance in the same manner, such
that men will be more supportive/tolerant than women.
H3. Political orientation will be related to both support and tolerance in the
same
way, such that liberals will be more supportive/tolerant than moderates and
conservatives.
Responses to both types of items depend to a large extent on the level of fear
or
dislike associated with the expression in question (Gibson, 1982). If the level
of fear
associated with the expression increases, the level of tolerance of the
expression should
decrease somewhat, although the individuals who most strongly support or
tolerate
expression should not be swayed by potential consequences. As noted above,
asking
someone to support a disliked individual's rights is quite different than
allowing her to speak,
particularly in one's own community as is traditional in tolerance scenarios.
Therefore,
H4. As the potential for negative consequences from an expressive act
increases,
the level of tolerance for that act will decrease.
H5. As the potential for negative consequences from an expressive act
increases,
the relationship between support and tolerance will decrease.
Speaking Out and Support/Tolerance
Given the differences between support and tolerance in their dictionary,
conceptual
and operational definitions, the notion of the two as separate but related
constructs would
seem to have face validity. In each of these definitions, the crucial
distinction appears to be
the active element underlying support and the passive one in tolerance. Those
who are more
likely to say they support expressive rights should, theoretically, be more
likely to actively
defend those rights than their less supportive counterparts.
Surprisingly, little research has directly examined this type of activism and
its relation
to tolerance/support. McClosky and Brill (1983) included a section titled "What
Would You
Do If..." in their study, which asked respondents to choose one of five
behavioral responses
to a series of civil liberties items. Responses ranged from doing nothing to
joining
organizations and protesting. The responses also incorporated reasons for the
behavior,
resulting in double-barreled choices. Thus, respondents could not simply do
nothing because
they did not feel strongly, for example. For each of the items directly related
to freedom of
expression, at least 20 percent of the mass public sample replied that they were
undecided
about what they would do. It is impossible to determine whether the reported
indecision is
an artifact of the response set or if it reflects the situational nature of the
survey items.
Political participation may be considered akin to speaking out in support of
one's
position on expressive rights, and it has been measured in several previous
studies. Political
participation typically includes expressive behaviors such as circulating
petitions, speaking
out at public meetings or contacting elected officials (McLeod, Steele, Chi, &
Huang, 1991;
Wyatt, 1991). McLeod and his colleagues found that political participation was
significantly,
positively related to support for free speech, both in the abstract and in a
concrete scenario.
Similar behaviors that are more closely related to free expression itself --
contacting the media
or protesting -- should also be related to support of expression.
Because previous research has not directly measured tolerance/support and the
likelihood of speaking out to support one's position, it is unwise to attempt to
predict such
relationships. Thus, the second portion of this study will examine whether and
how
tolerance, support for free speech, speaking out on one's position on a
tolerance situation,
and exercising one's expressive rights in general are related.
Method
Measures
In the attempt to better define tolerance and support, internal validity was a
key
concern, so we developed an experiment in which half the subjects would need
higher levels
of tolerance than the other half. Because Gibson (1982) noted that fear of
violence resulting
from an act of expression might cause some people to be much less tolerant than
they might
normally be, violence resulting from a demonstration was manipulated in the two
conditions. The KKK is a group traditionally used in tolerance studies, and
given the area
where most of the subjects grew up, the threat of the KKK marching in their
community is
realistic.
An actual news brief about a peaceful Ku Klux Klan demonstration held in
October
1994 in Annapolis, Md., served as a stimulus in one condition of the experiment.
According
to the brief, no one was injured and only one man was arrested (for throwing an
egg at the
demonstrators). In the second condition, the news brief was altered so that the
march
became violent: 17 people were arrested and six taken to the hospital with
injuries after a
protester began pelting Klansmen with rocks, and the Klansmen responded with
their fists
and protest signs. For both conditions, the stimuli were the same length, and
each had a two-
line headline.
After reading one of the news stories, subjects answered a set of traditional
tolerance
items (Stouffer, 1955) on whether they would allow the Klan to march in their
community,
whether a KKK member should be allowed to speak in their community, write a
newspaper
column, or have a book in the public library on white supremacy. The four
tolerance items
were summed into a reliable additive index (coefficient alpha=.84). In
addition, subjects were
asked to estimate how likely they would be to speak out in support of their
positions on the
tolerance items; these items were summed into an index. Subjects also completed
the 13-item
Support for Personal Rights Scale (Andsager, 1993), which was subsequently
formed into an
additive index (coefficient alpha=.80). Using traditional tolerance and
support batteries
allowed for direct comparison of responses to the two. Finally, subjects were
asked whether,
during the last five years, they had written a letter to the editor, complained
to an editor
about content, complained to a television station about content, or participated
in a public
protest. These items were also summed.
Sex and political orientation information was also gathered because these
variables
have been consistently related to tolerance and support in previous studies.
Sample
The experiment was conducted in Spring 1995 using 179 undergraduate students
enrolled in the introductory mass communication course at a medium-sized
Southeastern
university. The sample consisted of 106 men and 72 women. In terms of
self-reported
political orientation, 49 were liberal, 60 moderate, 32 conservative, and 37
said they did not
know.
Subjects participated in the study during their regular class period. They
were
randomly assigned to one of the two experimental conditions based on the
questionnaire set
they received. The subjects were told they were participating in studies for
two different
professors to reduce the obtrusiveness of the tolerance and support measures.
(Another set of
questions unrelated to expression was included in the packet to separate the
support and
tolerance measures, and the fonts were different to substantiate the cover
story.) After all
subjects had completed the questionnaire packet, they were debriefed and
thanked.
Data were entered into an SPSS for Windows program for subsequent analyses.
Results
To investigate the relationship between tolerance and support, a Pearson
correlation
coefficient was calculated between the two indexes. The first hypothesis had
suggested a
moderate correlation between tolerance and support, because the constructs are
clearly too
related to be only lowly correlated, but the item phrasing is too distinct for
the correlation to
be high. The data supported this hypothesis with a correlation of .48 (p<.001).
The second and third hypotheses posited that men would be more
supportive/tolerant than women and liberals would be more supportive/tolerant
than
moderates and conservatives. These propositions were tested with a repeated
measures
MANOVA, using sex and political orientation as the independent variables and the
standardized z-scores (mean=0; variance=1) of the summed tolerance and support
measures
as dependent variables, resulting in a 2X3X2 design. The test supported both
hypotheses, as
sex was a significant between-subjects effect (F1,173=7.89; p<.01) but men and
women did not
differentiate between tolerance and support in the within-subjects test. As
predicted, men
were more tolerant (mean=.148; SD=1.00) and more supportive (mean=.151; SD=.96)
of
expression than women (tolerance mean=-.215, SD=.97; support mean=-.202,
SD=1.02).
(See means displayed in Table 1.) Both sexes scored slightly higher on the SPRS
than on the
tolerance measure.
TABLE 1 ABOUT HERE
The third hypothesis, predicting that liberals would be more
tolerant/supportive of
expression than conservatives was also supported (F3,173=4.79; p<.01
between-subjects
effect). As shown by the means in Table 1, subjects who identified themselves
as liberal were
far more supportive (mean=.369; SD=1.17) and tolerant (mean=.217; SD=.90) than
moderates, conservatives and don't knows. Those who did not know their
political
orientation were by far the least tolerant/supportive of expression (tolerance
mean=-.357,
SD=1.10; support mean=-.334, SD=.90). Unlike sex, however, political
orientation was not
consistently related to tolerance and support. Liberals and don't knows -- like
men and
women -- scored higher on the support measure than on tolerance, but moderates
and
conservatives scored higher on the tolerance index than on the SPRS. These
within-subjects
effects were insignificant.
As the potential for negative consequences from an expressive act increases,
the level
of tolerance for that act was predicted to decrease. This fourth hypothesis was
not wholly
supported, according to a t-test using treatment group as the grouping variable
and tolerance
of the KKK march as the dependent variable. The peaceful-march group was
slightly more
likely to tolerate the march (mean=4.27; SD= 2.12) than the violent-march group
(mean=4.12; SD=1.93), but the difference was not significant.
The fifth hypothesis was also related to the correlation between support and
tolerance, predicting an inverse relationship between the potential for negative
consequences
resulting from an expressive act and the correlation between support and
tolerance. That is,
as expression becomes more threatening, those who support expressive rights in
general will
be less likely to tolerate the concrete and potentially dangerous expression in
question. This
hypothesis was supported in part, as the correlation between tolerance of the
KKK march
itself and support was lower in the violent-march group (r=.44; p<.001) than in
the peaceful-
march group (r=.51; p<.001), but the difference was not significant. The trend
was also true
for the general tolerance index and support, with r=.46 in the violent-march
group and
r=.50 in the peaceful-march group. Both correlations were significant at the
p<.001 level,
but they were not significantly different.
Speaking out
The second portion of the analysis attempted to examine the relationships among
tolerance, support, speaking out in a tolerance situation and exercising one's
expressive rights
in general. Pearson correlations between each pair of indexes were obtained to
explore these
relationships. As shown in Table 2, exercising expressive rights in general was
lowly but
positively related to both tolerance (r=.16; p<.05) and support of expressive
rights (r=.19;
p<.05). A similar correlation appeared between exercising expressive rights in
general and
speaking out in the tolerance scenario (r=.15; p<.05).
TABLE 2 ABOUT HERE
The correlation between tolerance and speaking out in the tolerance situation
was of
the same magnitude, but negative (r=-.16; p<.05). In other words, those who
were least
likely to allow the KKK to march, speak, etc., were the most likely to speak out
in favor of
their position, while subjects who were more tolerant were less likely to speak
out. To
better examine the negative correlation, the data were divided by treatment and
correlations
calculated between the tolerance and speaking out measures in each condition.
Because the
correlations between tolerance and support had varied somewhat by treatment, it
seemed
reasonable to expect that this correlation might also be affected by the
perceived level of fear
associated with the KKK. Subjects in the violent-march condition who were less
tolerant of
the KKK were slightly more likely to speak out on their positions (r=-.20) than
those in the
peaceful-march condition (r=-.11).
Discussion
Although this study is exploratory in nature, the findings provide some clues
to a
distinction between tolerance and support for free speech. They suggest that
the constructs
of tolerance and support are related, but not closely enough to be used
interchangeably in
theory-building. A moderate correlation existed between the two in this study,
perhaps due
to the differences in their operational definitions. The item phrasing in
tolerance asks
individuals whether they would allow a Ku Klux Klansman, for example, to speak
or march
in their community, while support for expression asks if they support the right
on another
(unidentified) individual to express him/herself in a more specific way.
The way tolerance items are traditionally phrased -- as they were in this study
--
requires respondents to guess at the possible consequences of an expressive act
because they
focus on the speaker, not the type of expression in question. Personal
experience could
conceivably affect such an estimation, introducing potential artifacts into
tolerance research.
This experiment attempted to reduce that threat to internal validity by
providing subjects
with an example of the consequences of expressive acts -- one violent march, one
peaceful
march. Gibson and Bingham (1982) in particular had suggested that the more
fearful the
perceived consequences of an expressive act, the less tolerant individuals would
be of that act.
Indeed, in their study, support of speech designed to incite an audience to
violence was
somewhat lower than that of speech that might incite an audience to violence
(emphasis in
original). In something as offensive to many as a KKK march, however, it is not
always clear
in advance whether violence will occur, and if it does, whether it was by
design. Thus, items
such as these arguably remove too much of the estimation involved in determining
tolerance,
such that responses are not as generalizable to how people will behave when
faced with the
possibility of a Klan march in their town.
Gibson and Bingham's suggestion was not completely supported by this
experiment,
however, as subjects who read about a violent KKK march in another city before
responding
to tolerance items about the KKK in their own community were only slightly less
tolerant of
the Klan than those who read about a peaceful march. Perhaps tolerance of a
potentially
threatening act of expression, such as a KKK march, relies more on individual
personality
characteristics than on the estimation of the results of the act.
Future research should address this issue more directly. It is possible that
the Klan's
regular (and usually peaceful) appearances in the Southeast made it a less
threatening group
for subjects in this study to tolerate than an unfamiliar entity. Obviously,
the use of a
student sample also limits the generalizability of these findings. Because
previous studies
indicate that the most (in)tolerant people do not differentiate based on their
liking or
disliking (Bobo & Licari, 1989; Gibson, 1986; McLeod, Guo, Huang, Rzeszut, &
Voakes,
1992), the stimulus group may not make a huge difference in responses, but
familiarity
and/or personal experience might well affect tolerance levels and should be
considered in
future work.
Nonetheless, the moderate correlation between tolerance and support also
indicates
that some individuals adhere more strongly to the democratic principles inherent
in both.
This study found, as have numerous previous studies, that men tend to be more
tolerant and
supportive of expressive rights than women, and liberals tend to be more
tolerant and
supportive of expressive rights than moderates and conservatives. Extensive
explanations for
these trends have been offered in previous research (Miller, Andsager, & Wyatt,
1992) and
will not be addressed here. What this experiment adds to previous research on
the topic is
the direct comparison of measures of tolerance and support; our findings
indicate that the
same sex and political orientation characteristics are significant predictors of
both. This
reinforces the idea that the same underlying construct informs both support and
tolerance,
and those who cling to this construct most tightly are unswayed by peculiarities
in item
phrasing when it comes to expressing their support or tolerance for free speech.
The changes in the correlations between tolerance and support by treatment in
this
experiment also support this contention. When the KKK march was portrayed as
violent,
the correlation between tolerance and support was slightly lower than it was in
the peaceful-
march condition. As expression becomes threatening, then, supporters of
expressive rights in
general will be somewhat less likely to tolerate the concrete act. It appears
that a hardcore
group of individuals -- who tend to be male and identify themselves as liberal
-- are the most
tolerant of expression. Because these findings suggest that it is perhaps
easier and less risky to
be supportive of some expression in a general sense than to allow it in a
specific, localized
and potentially threatening setting, the term tolerance may be a better measure
of how strong
individuals' attitudes toward free speech are.
This study set out to determine which term better captured the construct
underlying
the way people think about free expression. Given the active-passive
distinction inherent in
both the conceptual and operational definitions of the terms tolerance and
support, it is
necessary to now consider the findings on speaking out.
Despite the fact that none of the correlations among speaking out and either
tolerance or support were even moderate, two findings in these data suggest that
tolerance is
the more accurate term to use in describing the construct in question. First,
if the opposite of
tolerance is intolerance -- the willingness to repress the rights of others
(Gibson, 1986) -- the
correlations produced by these data support both the active connotation of
intolerance and
the passive connotation of tolerance. That is, subjects in the study who were
most likely to
allow the KKK to march, speak, etc., were not nearly as likely to voice their
opinions as were
those most vehemently opposed to the Klan's expression. Secondly, the
definition of support
includes the willingness to protect the rights of others, which suggests the
possibility of a
positive correlation between support for expressive rights and voicing one's
opinion on KKK
expression. While a low, positive correlation occurred between support and
exercising one's
expressive rights in general, the correlation between support and speaking out
regarding the
Klan was practically nonexistent.
Taken together, the estimation of this behavioral component and the declining
relationship between support and tolerance as expression became potentially
threatening
suggest that tolerance may better describe the construct that informs people's
attitudes
toward free speech. This does not mean that the traditional measures of
tolerance are the
most parsimonious way of tapping into the underlying construct, nor does it
reduce the
usefulness of measuring support. As discussed above, each has its merits and
flaws. Future
research should, however, attempt to better refine and perhaps find ways to
merge the two
types of measures. More research also needs to be conducted using a variety of
attitudinal
and behavioral measures to determine the likelihood of speaking out to support
one's stance
on the expression of others. After all, if individuals who tolerate free
expression do not
defend it, those who are intolerant may prevail.
References
Andsager, J.L. (1993, August). Support for media and personal expressive rights:
development of
parallel scales. Paper presented at the annual convention of the Association
for Education in
Journalism and Mass Communication, Kansas City, MO.
Babbie, E. (1995). The practice of social research (7th ed.) (Belmont, CA:
Wadsworth).
Bobo, L., & Licari, F.C. (1989). Education and political tolerance: testing the
effects of cognitive
sophistication and target group affect. Public Opinion Quarterly, 53, 285-308.
Erskine, H. (1970). The polls: freedom of speech. Public Opinion Quarterly, 34,
483-496.
Gibson, J.L. (1986). Pluralistic intolerance in America: a reconsideration.
American Politics
Quarterly, 14, 267-293.
Gibson, J.L., & Bingham, R.D. (1982). On the conceptualization and measurement
of political
tolerance. The American Political Science Review, 76, 603-620.
McClosky, H., & Brill, A. (1983). Dimensions of tolerance: what Americans
believe about civil
liberties (New York: Russell Sage).
McLeod, J.M., Guo, Z., Huang, K., Rzeszut, A.K., & Voakes, P.S. (1992, August).
Evaluating models
of public support for First Amendment rights. Paper presented at the annual
convention of
the Association for Education in Journalism and Mass Communication, Montreal.
McLeod, J.M., Steele, C.A., Chi, H., & Huang, K. (1991, August). Understanding
political tolerance:
alternative models of public support. Paper presented at the annual convention
of the
Association for Education in Journalism and Mass Communication, Boston.
Miller, M.M., Andsager, J.L., & Wyatt, R.O. (1992, August). How gender and
select demographics
relate to support for expressive rights. Paper presented at the annual
convention of the
Association for Education in Journalism and Mass Communication, Montreal.
Nunn, C.Z. (1973). Support of civil liberties among college students. Social
Problems, 20, 300-318.
Prothro, J.W., & Grigg, C.M. (1960). Fundamental principles of democracy: bases
of agreement and
disagreement. Journal of Politics, 22, 276-294.
Stouffer, S.A. (1955). Communism, conformity, and civil liberties (Garden City,
NY: Doubleday).
Sullivan, J.L., Piereson, J.E., & Marcus, G.E. (1979). An alternative
conceptualization of political
tolerance: illusory increases 1950s-1970s. American Political Science Review,
73, 781-794.
Sullivan, J.L., Shamir, M., Walsh, P., & Roberts, N.S. (1985). Political
tolerance in context: support
for unpopular minorities in Israel, New Zealand, and the United States
(Boulder, CO:
Westview).
Wyatt, R.O. (1991). Free expression and the American public: a survey
commemorating the 200th
anniversary of the First Amendment (Washington, DC: American Society of
Newspaper
Editors).
Table 1. Mean z-scores on tolerance and support indexes by sex and political
orientation.
Tolerance
Support
Sex*
Men
.148
.151
Women
-.215
-.202
Political orientation
Liberal
.218*
.369*
Moderate
.031
-.011
Conservative
.037
-.136
Don't know
-.358*
-.334*
Overall
-.002
.007
*Group means are significantly different at p<.05.
Table 2. Correlation matrix of speaking out, tolerance and support indexes.
1
2
3
4
Tolerance
1.00
Support
.48**
1.00
Speaking out on tolerance
-.16*
.00
1.00
Exercising expressive rights
in general
.16*
.19*
.15*
1.00
______
*Correlation is significant at p<.05.
**Correlation is significant at p<.001.
The distinction between abstract and concrete is an important one in examining
support for free
speech. As previous research has indicated, nearly everyone agrees with
abstract ideals of free speech
and free press, but as the situations in question become more concrete and
specific, the level of
support often drops dramatically (see review in Andsager, 1993).
This is a paraphrase used above a table of means by those authors. Gibson and
Bingham's (1982) mail
survey instructions read, "For each of the following types of speech, please
indicate whether you
believe it should definitely be protected, probably not be protected, or
definitely not be protected by
the First Amendment" (p. 608).
Half the subjects read the news story and completed the tolerance items first;
the other half
completed the SPRS first. Statistical tests revealed no significant difference
in scores due to
questionnaire order.
|
