1. Lauren Bryant
10/11/14
PLAP 4150
Measuring the Effect of a Speaker’s Race on Agreement with the Speaker’s Political
Arguments
Question:
My group decided to research the question of whether a non-white speaker
racializes otherwise non-racial political issues, affecting how people perceive the political
argument that the speaker is presenting.
Theoretical Antecedents:
Our question was informed by Dangerous Frames: How Ideas About Race and Gender
Shape Public Opinion by Nicholas Winter. Dangerous Frames discusses how people form
schemas about race and gender, which can be used to form understandings of many
political issues (Winter 2008, 2-3). For instance, Winter describes how Hillary Clinton was
a “gendered image” during the health care debate during the 1990s. Winter says, “[...] As
head of the administration’s task force Clinton ‘violate[d] the traditional separation of the
masculine sphere and the feminine domestic sphere that ha[d] previously defined the role
of First Lady. [...] Moreover, her role put a woman in charge of reforming the traditionally
male-controlled health care industry.’” (Winter 2008, 124). The health care debate, while
not explicitly about gender, became gendered simply because someone of the minority
gender was the leader of the debate. This informed our research question because we
wanted to know if a person’s race had a similar effect as gender. That is, we wanted to
2. know if an issue not explicitly related to race could become racialized because someone of
a minority race was leading the discussion.
Furthermore, our research question was informed by Greenwald and Banaji’s
Implicit Social Cognition: Attitudes, Self-Esteem, and Stereotypes. Greenwald and Banaji
assert that people can be influenced by their implicitly held stereotypes and may not even
know that they hold those stereotypes or that they are influenced by them. Greenwald and
Banaji say, “[...] Considerable evidence now supports the view that social behavior often
operates in an implicit or unconscious fashion. The identifying feature of implicit cognition
is that past experience influences judgment in a fashion not introspectively known by the
actor. [...] The theorized ordinariness of implicit stereotyping is consistent with recent
findings of discrimination by people who explicitly disavow prejudice” (Greenwald and
Banaji 1995, 4). The idea that people act on their implicit cognitions without realizing that
they have those implicit biases or that they are acting on them influenced us to think that
activation of peoples’ racial schemas would probably cause them to act on implicitly held
beliefs about race without realizing that’s what they were doing. The ideas of implicit
cognition and activation of racial schemas presented by Greenwald, Banaji, and Winter
were the antecedents that informed the development of our research question.
Hypothesis:
Our hypothesis was that the race of the speaker would activate racial schemas in
participants and cause them to have different opinions about the same political issue
depending on the race of the speaker. Specifically, we thought that racial conservatives
would show less agreement with the statements attributed to non-white speakers and
more agreement with statements attributed to white speakers. Furthermore, we thought
3. that racial liberals would show similar levels of agreement between statements attributed
to white and nonwhite speakers.
Procedure:
To test our hypothesis, we developed a survey that asked the participants to read an
article and answer questions about their agreement with a law about motorcycles being
allowed to legally run red lights that was attributed to people of different races. There
were five treatment groups. Each treatment group read the same statement about the
motorcycle law attributed to a person of a different race. The races of the speakers (our
independent variable) were White, Black, Hispanic, Middle Eastern, and Asian. Our control
was the White speaker treatment group. We used photos of men who were about the same
age and socioeconomic status so that gender, age, and socioeconomic status would not
affect our results. We also looked up common names for each race to give the speakers to
further activate peoples’ racial schemas. The participants were asked to rate their level of
agreement with the motorcycle law and the persuasiveness of the speaker’s arguments
(our dependent variables).
After having the participants read the statement and answer the questions about
their level of agreement and the persuasiveness of the arguments, they answered questions
about their racial predispostitions using the Racial Resentment scale used in Dangerous
Frames. The questions involved in the Racial Resentment scale included: 1. Over the past
few years, blacks have gotten less than they deserve. 2. Irish, Italian, Jewish, and many
other minorities overcame prejudice and worked their way up. Blacks should do the same
without any special favors. 3. It's really a matter of some people not trying hard enough; if
blacks would only try harder they could be just as well off as whites. 4. Generations of
4. slavery and discrimination have created conditions that make it difficult for blacks to work
their way out of the lower class. These questions were rated on a scale from “strongly
disagree” to “strongly agree.” We coded the answers from -2 (strongly disagree) to 2
(strongly agree) for the questions for which answering in agreement was racially liberal
and reversed the coding for questions for which answering in agreement was racially
conservative. Therefore, in the results, the higher the score, the more racially liberal the
participant was. We also gathered demographic data including party affiliation, age,
race/ethnicity, socioeconomic status, country of origin, and gender.
So that we did not activate peoples’ racial schemas before taking the survey, we told
participants that the survey was studying the effects of ideology on public opinion of this
motorcycle law.
We sent out the survey to our networks on social media and to groups we are
involved with at UVa. to gather participants. We had a total of 137 subjects complete the
survey, most of whom were UVa. students.
Findings:
Agreement with the motorcycle law did not change significantly based on the race of
the speaker for either racial liberals or racial conservatives. However, as we predicted, the
difference in agreement based on race was lower for racial liberals than for racial
conservatives (0.45 for racial liberals compared to 0.66 for racial conservatives on a scale
from 1 to 5 with 1 being “strongly oppose” and 5 being “strongly support”). It is also
interesting to note that the Middle Eastern speaker and the Black speaker received the
lowest levels of agreement (average of 2.5 and 3 respectively) among racial conservatives
(which we expected), and the White and Hispanic speakers received the highest levels of
5. agreement (average of 3.16 for both). We would expect that the White speaker would have
the highest level of agreement among racial conservatives, but the equally high level of
agreement with the Hispanic speaker surprised us. These differences of agreement based
on race, however, are not statistically significant. Among racial liberals, agreement with
the Black speaker and the Asian speaker were slightly higher than the control (the White
speaker), however, the differences aren’t statistically significant.
Racial conservatives also generally agreed with the proposed motorcycle law
slightly more than racial liberals (average of 3.00 versus 2.47 respectively on a scale of 1 to
5 with 1 being “strongly oppose” and 5 being “strongly support”). We did not expect this
result. Probably the most striking finding we noticed from our data was that racial liberals
were much more likely than racial conservatives to identify as Democrat (73.81%
compared to 14.29% respectively) and racial conservatives were much more likely than
racial liberals to identify as Republican (45.71% compared to 7.14% respectively).
Furthermore, the average age of racial conservatives was significantly higher than that of
racial liberals (35.6 versus 24.8 respectively). The findings about the political affiliation
and age of participants were expected.
Internal Validity:
We did have a fairly large sample size, which supports our internal validity,
however, we had a disproportionately low number of people in the Black speaker
treatment group, which could have affected our results. Furthermore, our racial
resentment scale only addressed white and black race relations, so it may not be very
generalizable to racial resentment of the other races that we tested. Also, the survey froze
for many people so we have a large number of incomplete survey responses. This could
6. have impacted our results because if someone was in one treatment group before the
survey froze and was in another treatment group after they restarted the survey, they
would probably realize that the person pictured wasn’t really the author of the statement,
and possibly figure out what we were really trying to measure.
Furthermore, while we tried to choose a political issue that people wouldn’t have
much prior knowledge or opinions of, it is possible that some people—particularly, racial
conservatives who were disproportionately Republican—may have had preexisting
opinions on the subject, which could have affected our results because race would be less
important than the actual policy in question, or the participant could know that the person
we attributed the statement to was not really the President of Motorcycle Nuts of America.
Another factor that could have affected our internal validity is that we used real
congressmen for the photos in the survey. While we tried to pick obscure politicians that
most UVa. students would not already be familiar with, it is possible that some participants
recognized them and realized they weren’t the author of the statement. Furthermore, the
way that we distributed the survey meant that we had mostly white, educated, people in
their early to mid twenties, but a few outliers that were much older or of very different
demographics. This could have affected the internal validity of our results as well.
External Validity:
Due to our convenience sampling (Shuttleworth 2009), the population of our sample
was mostly young, white, and well educated, which is not very generalizable to the rest of
the nation. Furthermore, we had no control over where and when people took the survey
so our ecological validity (Shuttleworth 2009) could have been affected.
7. What I Would Do Differently:
If I were to do this study again, I would do a few things differently. First, I would
probably use three treatment groups: White, Black, and Barack Obama. The original
research question started out asking specifically if acceptance of Barack Obama’s policies
was affected by his race, and it transformed into a more general question about race
affecting acceptance of political arguments. I would return to the more specific question
and I would get rid of the Hispanic, Asian, and Middle Eastern treatment groups because I
would expect the results to be strongest between White and Black races, and because the
racial resentment scale only measures White and Black race relations. It would be
interesting to compare the results of the White and Black treatment groups to the Barack
Obama group because this could potentially show whether race is a factor in participants’
acceptance of political messages, or whether people just base their acceptance off of their
opinions of Barack Obama as a person.
Also, if I were to do this study again, I would try to get a much larger and more
diverse sample size so that the results were more generalizable to the population as a
whole. I would also reverse the way we coded the Racial Resentment scale so that a more
positive number meant more racial resentment.
I might also try to pick a different issue than the motorcycle law, if were to do this
study again. While we thought the motorcycle law would be obscure enough to most
people that they wouldn’t have strong opinions about it, there was a slight trend of racial
conservatives (who mostly identified as Republican) favoring the law more than racial
liberals (who mostly identified as Democrat) suggesting that there may be some
partisanship to this issue. Anecdotally, I know from talking to a participant after taking the
8. survey that this person already had a strong view about the motorcycle law and was one of
the more conservative participants. The motorcycle law evokes some aspect of the idea of
“small government” or preference of the individual over public safety, which may resonate
more with conservatives. However, it may be that all issues, no matter how obscure, have
some amount of partisanship to them that is unavoidable. At least in the case of the
motorcycle law, difference in support between racial conservatives and racial liberals was
minimal and probably not even statistically significant, so it might be one of the better
issues to use in a study like this.
Implications:
The main implication of this study is that further research needs to be done on this
question. The study needs to be replicated with a much larger and much more diver se
population to make any generalizations. There were some interesting trends, such as racial
conservatives agreeing mostly with the White and Hispanic speakers and least with the
Black and Middle Eastern speakers as well as the larger difference in agreement by race for
racial conservatives than for racial liberals, however, these trends were not statistically
significant. If further research were done and these trends turned out to be statistically
significant, the implication would be fairly close to our hypothesis that racial conservatives’
acceptance of political messages would be negatively affected by a nonwhite speaker and
that racial liberals’ acceptance would be less affected by a nonwhite speaker. This could
have implications for how minorities should attempt to gain support for their political
arguments from racial conservatives. The fact that racial conservatives agreed equally as
much with the Hispanic speaker as with the White speaker would have to be further
investigated. However, I doubt that if this study were replicated with a larger, more
9. diverse population with statistically significant results, that this finding would still exist.
Overall, we found some interesting trends, but the study would have to be improved and
replicated to prove any findings and make relevant implications.
Works Cited:
Greenwald, Anthony G., and Mahzarin R. Banaji. 1995. “Implicit Social Cognition: Attitudes,
Self-Esteem, and Stereotypes.” Psychological Review 102 (1):4-27.
Martyn Shuttleworth (Sep 16, 2009). Population Validity. Retrieved Dec 11, 2014 from
Explorable.com: https://explorable.com/population-validity
Martyn Shuttleworth (Mar 19, 2009). Ecological Validity. Retrieved Dec 11, 2014 from
Explorable.com: https://explorable.com/ecological-validity
Winter, Nicholas J. G. 2008. Dangerous Frames: How Ideas About Race and Gender Shape
Public Opinion. Chicago: University of Chicago Press.