The Effects Of Racial Animus On A Black Presidential
Candidate: Using Google Search Data To Find What Surveys
Miss
Seth Stephens-Davidowitz
June 9, 2012
http://www.people.fas.harvard.edu/~sstephen/papers/RacialAnimusAndVotingSethStephensDavidowitz.pdf

[moderator: the full paper is found at the above link]

Introduction

Does racial animus cost a black candidate a substantial
number of votes in contemporary America? The most recent
review of the literature is inconclusive: "Despite
considerable effort by numerous researchers over several
decades, there is still no widely accepted answer as to
whether or not prejudice against blacks remains a potent
factor within American politics" (Huddy and Feldman,
2009).1

There are two major reasons this question has been of
such enduring interest to scholars: first, it helps us
understand the extent of contemporary prejudice;2
second, it informs us on the determinants of voting.3
There is one major reason the question has proven so
difficult: individuals' tendency to withhold socially
unacceptable attitudes, such as negative feelings
towards blacks, from surveys (Tourangeau and Ting, 2007;
Berinsky, 1999; Berinsky, 2002; Gilens et al., 1998;
Kuklinski et al., 1997).

This paper uses non-survey-based methodology. I use a
novel data source to proxy an area's racial animus:
Google search queries that include racially charged
language. I compare the proxy to an area's change in
Democratic vote shares from the 2004 all-white
presidential election to the 2008 biracial presidential
election. This empirical specification is most similar
to that of Mas and Moretti (2009). They use a survey
measure of support for a law banning interracial
marriage from the General Social Survey (GSS) as their
state-level proxy for racial attitudes. They do not find
evidence that racial attitudes affected Barack Obama's
2008 vote share.

Google data, evidence suggests, are unlikely to suffer
from major social censoring: Google searchers are online
and likely alone, both of which make it easier to
express socially taboo thoughts (Kreuter et al., 2009).
Furthermore, individuals say they are forthcoming with
Google (Conti and Sobiesk, 2007). The large number of
searches for pornography and sensitive health
information adds additional evidence that Google
searchers express interests not easily elicited by other
means. Relative to measures from the GSS, Google-based
measures are also meaningfully available at a finer
geographic level, use more recent data, and aggregate
information from much larger samples.4

The baseline proxy that I use is the percentage of an
area's total Google searches from 2004-2007 that
included the word "nigger" or "niggers." I choose the
most salient word to constrain data-mining.5 I do not
include data after 2007 to avoid capturing reverse
causation, with dislike for Obama causing individuals to
use racially charged language on Google.6 My regression
analysis includes 196 of 210 media markets, encompassing
more than 99 percent of American voters.

The epithet is a common term used on Google. During the
period 2004-2007, there were roughly the same number of
Google searches that included the word "nigger(s)" as
there were Google searches that included words and
phrases such as "migraine(s)," "economist," "sweater,"
"Daily Show," and "Lakers." (Google data are case-
insensitive.) The most common searches including the
epithet (such as "nigger jokes" and "I hate niggers")
return websites with derogatory material about African-
Americans. The top hits for the top racially charged
searches are nearly all textbook examples of
antilocution, a majority group's sharing stereotype-
based jokes using coarse language outside a minority
group's presence. This was determined as the first and
crucial stage of prejudice in Allport's (1979) classic
treatise. From 2004-2007, the searches were most popular
in West Virginia; upstate New York; rural Illinois;
eastern Ohio; southern Mississippi; western
Pennsylvania; and southern Oklahoma.

I find that racially charged search is a large and
robust negative predictor of Obama's vote share. A one
standard deviation increase in an area's racially
charged search is associated with a 1.5 percentage point
decrease in Obama's vote share, controlling for John
Kerry's vote share.7 The statistical significance and
large magnitude are robust to controls for changes in
unemployment rates; home-state candidate preference;
Census division fixed effects; prior trends in
presidential voting; changes in Democratic House vote
shares; swing state status; and demographic controls.
The estimated effect is somewhat larger when adding
controls for an area's Google search volume for other
terms that are moderately correlated with search volume
for "nigger" but are not evidence for racial animus. In
particular, I control for searches including other terms
for African-Americans ("African American" and "nigga,"
the alternate spelling used in nearly all rap songs that
include the word) and profane language.

The results imply that, relative to the most racially
tolerant areas in the United States, prejudice cost
Obama between 3.1 percentage points and 5.0 percentage
points of the national popular vote. This implies racial
animus gave Obama's opponent roughly the equivalent of a
home-state advantage country-wide. The cost of racial
prejudice was not decisive in the 2008 election. But a
four percentage point loss by the winning candidate
would have changed the popular vote winner in the
majority of post-war presidential elections.8

I argue that any votes Obama gained due to his race in
the general election were not nearly enough to outweigh
the cost of racial animus, meaning race was a large net
negative for Obama. Back-of-the-envelope calculations
suggest Obama gained at most only about one percentage
point of the popular vote from increased African-
American support. The effect was limited by African-
Americans constituting less than 13 percent of the
population and overwhelmingly supporting every
Democratic candidate. Evidence from other research, as
well as some new analysis in this paper, suggest that
few white voters swung in Obama's favor in the general
election due to his race.9 A large cost of race in the
general election is consistent with some scholars'
estimates that, in light of the immensely unpopular
incumbent Republican president, Obama substantially
underperformed in the 2008 general election (Lewis-Beck
et al., 2010; Tesler and Sears, 2010). It also can
explain why white male Democratic candidates
consistently outperformed Obama in hypothetical general
election polls (Jackman and Vavreck, 2011). And it can
explain why House Democrats' vote gains from 2004 to
2008 were significantly larger than Obama's gain
relative to Kerry.

The main contributions of this paper are threefold:
First, I offer new evidence that racial attitudes remain
a potent factor against African-Americans, nationwide,
in modern American politics. This suggests the null
result in Mas and Moretti (2009) was due to limitations
in the GSS proxies for racial attitudes. The results are
larger than those of most studies using individual-level
survey data (e.g., Piston, 2010; Schaffner, 2011; Pasek
et al., 2010). In addition, my main results rely on
administrative, rather than reported, vote data; some
scholars argue misreporting is a significant concern
with reported vote data (Atkeson, 1999; Wright, 1993;
Ansolabehere and Hersh, 2011).

Second, the new data source for area-level proxies of
racial attitudes may be useful to other researchers.10
Researchers studying the causes or consequences of an
area's racial attitudes previously have used decades of
aggregated GSS data on views towards interracial
marriage or similar issues to obtain such proxies (e.g.,
Alesina et al., 2001; Alesina and La Ferrara, 2002;
Charles and Guryan, 2008; Cutler et al., 1999; Card et
al., 2008). The Google data add to greatly improved
individual-level proxies from list experiments and
implicit attitude tests as tools for scholars studying
racial attitudes.11

The third, and probably most important, contribution is
methodological: I show that Google search data can yield
new evidence on a question complicated by social
desirability bias. This builds on a nascent literature
finding promise in Google data. Previous papers using
Google search data have tended to focus on its timing
advantage. Since Google makes its data available the
next day, while many agencies take weeks, Google can
yield quicker information on health (Ginsberg et al.,
2009; Seifter et al., 2010); demand (Varian and Choi,
2010); and jobs (Askitas and Zimmermann, 2009).12 In
addition, previous work has tended to report
correlations between Google data and existing proxies
from alternative data sources rather than find new
evidence on an empirical question. Scheitle (2011), for
example, notes correlations between Google searches on a
variety of topics, though not racial animus, and
existing measures. This paper shows clearly that Google
search query data can do more than correlate with
existing proxies; on socially sensitive topics, they can
give better data and open new research on old questions.
Researchers might also use Google data to understand the
causes and consequences of animus towards other
groups.13 In addition, the Conclusion lists topics of
interest across the social sciences, from Bound et al.
(2001), on socially sensitive topics in which research
has been similarly hampered and may be open to a similar
methodology as that of this paper.

The remainder of this paper is organized as follows.
Section II discusses the new Google proxy for an area's
racial animus. Section III introduces the empirical
specification and results on the effects of racial
animus in an election with a black candidate. Section IV
interprets the magnitude of the effects, comparing them
both to other research on the 2008 election and to
research on other factors found to influence voting.
Section V concludes.

___________________________________________

Portside aims to provide material of interest to people
on the left that will help them to interpret the world
and to change it.

Submit via email: [log in to unmask]

Submit via the Web: http://portside.org/submittous3

Frequently asked questions: http://portside.org/faq

Sub/Unsub: http://portside.org/subscribe-and-unsubscribe

Search Portside archives: http://portside.org/archive

Contribute to Portside: https://portside.org/donate