The Vicarious Bases of Perceived Injustice: BASES OF PERCEIVED INJUSTICE

Abstract

Profound differences exist in how Americans from various racial and ethnic groups view police and court officials. We argue that vicarious experiences contribute to this racial and ethnic divide. Drawing on research on social communication, social network composition, and negativity biases in perception and judgment, we devise a theoretical framework to articulate why vicarious experiences magnify racial and ethnic disparities in evaluations of judicial actors. Four hypotheses are tested using original survey data from the state of Washington. Results provide strong evidence that vicarious experiences influence citizens’ evaluations of both police and courts, and they do so in a manner that widens racial divides in how those actors are perceived.

Full text

The Vicarious Bases of Perceived Injustice
Jeffery J. Mondak
University of Illinois
Jon Hurwitz
University of Pittsburgh
Mark Peffley
University of Kentucky
Paul Testa
Brown University
Abstract: Profound differences exist in how Americans from various racial and ethnic groups view police and court
officials. We argue that vicarious experiences contribute to this racial and ethnic divide. Drawing on research on social
communication, social network composition, and negativity biases in perception and judgment, we devise a theoretical
framework to articulate why vicarious experiences magnify racial and ethnic disparities in evaluations of judicial actors.
Four hypotheses are tested using original survey data from the state of Washington. Results provide strong evidence that
vicarious experiences influence citizens’ evaluations of both police and courts, and they do so in a manner that widens racial
divides in how those actors are perceived.
Replication Materials: The data, code, and any additional materials required to replicate all analyses in this arti-
cle are available on the American Journal of Political Science Dataverse within the Harvard Dataverse Network, at:
http://dx.doi.org/10.7910/DVN/QTG6FQ.
In recent years, a litany of high-profile incidents has
emphasized not only the significance of race within
the criminal justice system, but also the racial divide
in how that system is perceived. Regarding the latter, a
wealth of research has sought to explain profound dif-
ferences in how, and how favorably, African Americans
and Whites view the justice system and its actors (e.g.,
Bobo and Johnson 2004; Peffley and Hurwitz 2010; Tonry
2011; Unnever 2008; Weitzer and Tuch 2006). Although
the bases of this racial divide are multifaceted, part of
the explanation appears to be experiential. Blacks and
Whites report starkly different experiences with police
and courts, experiences that shape views of the justice
system.
While acknowledging the role of direct encounters,
we focus on the possibility that information beyond the
realm of personal experience also matters. We posit that
the racial gulf widens as a function of social interaction:
Differences in Blacks’ and Whites’ vicarious encounters
Jeffery J. Mondak is James M. Benson Chair in Public Issues and Civic Leadership, Department of Political Science, University of Illinois,
420 David Kinley Hall, 1407 W. Gregory Drive, Urbana, IL 61801 (jmondak@illinois.edu). Jon Hurwitz is Professor, Department of
Political Science, University of Pittsburgh, 4600 Wesley W. Posvar Hall, Pittsburgh, PA 15260 (hurwitz@pitt.edu). Mark Peffley is Professor,
Department of Political Science, University of Kentucky, 1615 Patterson Office Tower, Lexington, KY 40506-0027 (mpeffl@email.uky.edu).
Paul Testa is Assistant Professor, Department of Political Science, Brown University, Box 1844, 36 Prospect Street, Providence, RI 02912
(paul_testa@brown.edu).
Data for this project were gathered under our supervision with the support of the Washington State Supreme Court Minority and Justice
Commission, and the State of Washington Administrative Office of the Courts—Washington State Center for Court Research. We thank
all involved in this project, and especially Carl McCurley, for their contributions.
with the justice system exacerbate corresponding differ-
ences in how that system is perceived.
Prior scholars have noted the possible significance
of vicarious experiences on views of the justice system
and have provided preliminary evidence that such effects
are consequential. For example, Weitzer and Tuch (2006)
posed questions about respondents’ negative experiences
with the police and any similar experiences among oth-
ers in their households. Rosenbaum et al. (2005) asked
Chicagoans about both their personal experiences with
the justice system and whether they knew of anyone who
had been arrested in the past year. Results in these studies
suggested that vicarious experiences influence individu-
als’ views of the police.
Although prior works offer initial evidence regarding
the possible significance of vicarious experiences, none
provides a holistic theoretical or empirical account. Miss-
ing is a rationale for how and why vicarious experiences
should be expected to matter and, most critically, for why
American Journal of Political Science,Vol. 00, No. 0, xxxx 2017, Pp. 1–16
C2017, Midwest Political Science Association DOI: 10.1111/ajps.12297
1

2JEFFERY J. MONDAK ET AL.
these experiences may widen racial disparities in how the
criminal justice system is perceived. Prior works also have
used coarse empirical indicators. At most, data have been
obtained about only a single acquaintance rather than
about respondents’ broader networks. No information
has been reported about experiences with actors other
than the police, about gradations of positive and negative
experiences, or about the race and ethnicity of respon-
dents’ acquaintances.
We seek to study vicarious effects more systemati-
cally. We delineate a precise theoretical account of why
vicarious effects should be important, and why such ef-
fects magnify racial gaps in perceptions of judicial ac-
tors. Each of our four hypotheses enjoys support in past
research and comports with intuition regarding the na-
ture of social experience. It is in combination that our
hypotheses gain their full power. The first three hypothe-
ses provide a step-by-step case for why patterns of so-
cial interaction give rise to a widening of racial disparity
in perceptions of the justice system. The final hypothe-
sis incorporates insights from psychology to explore not
the existence of vicarious effects, but rather their mag-
nitude. To test these hypotheses, we draw on original
survey data from the state of Washington. Central to this
effort are empirical measures designed with the intent
of generating insight on whether, how, and to what ex-
tent vicarious experiences influence views of the justice
system.
Vicarious Effects
Our central empirical proposition is that vicarious expe-
riences influence people’s assessments of judicial actors in
a manner that magnifies racial and ethnic gaps in those
evaluations. This proposition necessitates a theoretical
framework addressing several matters: the information
about police and courts people encounter vicariously,
whether such information should influence evaluations,
and whether such influence contributes to racial differ-
ences in those appraisals. Our approach focuses on peo-
ple’s justice encounter networks, or circles of acquaintances
who have had experiences with judicial actors.
We employ survey batteries on which respondents are
asked to identify acquaintances who have had encoun-
ters with police and courts. A prefatory matter concerns
network composition. Homophily abounds in social net-
works (McPherson, Smith-Lovin, and Cook 2001),1and
1For examples of research on racial and ethnic homophily in a
variety of types of social networks, see Ibarra (1995), Marsden
(1988), and Shrum, Cheek, and Hunter (1988).
race and ethnicity are powerful structuring forces. We
examine the justice encounter networks of Washingto-
nians from four groups: Blacks, Whites, Latinos, and
Asians. Drawing on research on homophily, we assume
that the justice encounter networks reported by African
Americans are largely composed of African Americans,
the networks of Whites are mostly composed of Whites,
and so on. Evidence consistent with this assumption is
presented below. For now, the key matter is that our hy-
potheses presuppose homophily.
The first portion of our theoretical framework per-
tains to whether people’s experiences with police and
courts vary by race. The people we have in mind, of
course, are our survey respondents’ acquaintances, the
members of our respondents’ justice encounter net-
works. There are several reasons to expect a racial di-
vide. First, racial and ethnic gaps exist in the nature and
frequency of encounters with the justice system. Com-
paring Blacks and Whites, disparities exist in rates of ar-
rest (Walker, Spohn, and DeLone 2004), prosecution and
sentencing (Spohn 2013), and incarceration (Harrison
and Beck 2003). Washington’s racial and ethnic composi-
tion (72.1% White; 11.6% Latino; 8.2% Asian, including
Native Hawaiian and other Pacific Islanders; and 3.8%
African American) permits us to explore racial differences
beyond those between Blacks and Whites. In Washington,
Whites (59.4% of prisoners) and Asians (3.7%) are un-
derrepresented, and Latinos (12.1%) are slightly overrep-
resented, in proportion to their share of the state’s popu-
lation.2In contrast, African Americans constitute 18.7%
of the prison population, an overrepresentation of nearly
fivefold compared to their share of the state’s population.3
Second, numerous studies reveal racial profiling and dis-
crimination in police behavior, including in traffic stops
and the use of force (e.g., Engel and Swartz 2013; Jacobs
and O’Brien 1998; Wilson 1978). Third, these differences
affect perceptions. Peffley and Hurwitz (2010, 42) report
that 29.6% of Black respondents claimed to have been
treated unfairly by the police in the past 5 years, versus
only 12.7% of Whites. Together, these rationales support
our first hypothesis:
H1: Perceived negative experiences with police and
courts are most abundant among individuals’
African American acquaintances.
2Information about Washington’s prison population is taken
from the December 31, 2013, version of the Washington
State Department of Corrections Fact Card, www.doc.wa.gov/
aboutdoc/docs/msFactCard_012.pdf.
3Compared to the other 10 states in the West Census region, Wash-
ington falls near the center of the distribution both in terms of the
percent of the total and prison population that is African American
(Sakala 2014).

BASES OF PERCEIVED INJUSTICE 3
At least four scenarios may generate perceptions con-
sistent with Hypothesis 1: (1) criminal justice actors target
African Americans for unduly negative treatment; (2) the
valence of individuals’ experiences with police and courts
varies with socioeconomic status, which, in turn, is cor-
related with race and ethnicity; (3) experiences with the
police and courts involving serious matters are most likely
to be perceived negatively, and African Americans are dis-
proportionately likely to interact with justice system ac-
tors over serious matters; and (4) Blacks and others view
similar experiences differently, with Blacks being more
likely to perceive justice system encounters as negative.
Our theory is agnostic about these scenarios. Hypothesis 1
pertains to whether perceived negative experiences are
more abundant among African Americans, not why.
The first hypothesis accounts for the valence of so-
cially communicated information about the justice sys-
tem. If circles of acquaintances are composed randomly
with regard to race, then the race and ethnicity of our
respondents would be uncorrelated with vicarious expo-
sure to positive or negative information; in the aggregate,
all racial and ethnic groups would be exposed to similar
information. Conversely, if networks exhibit racial and
ethnic homophily (and if, as per Hypothesis 1, people’s
Black social contacts are especially likely to perceive that
they have had bad experiences with actors in the justice
system), then vicarious exposure to the justice system will
vary in valence as a function of the respondent’s race and
ethnicity:
H2: Vicarious exposure to information about nega-
tive experiences with police and courts is most
prevalent among African Americans.
Vicarious justice experiences should vary by the re-
cipient’s race and ethnicity, but should they affect subse-
quent perceptions? The role of social influence in politics
has been recognized since the efforts of the Columbia re-
searchers (e.g., Katz and Lazarsfeld 1955). Beyond its basic
social function, social communication provides efficiency
for citizens interested in better understanding aspects of
the social and political worlds (Huckfeldt and Sprague
1995). When people share information, recipients are ex-
posed to insights that hold the potential to influence their
attitudes and behaviors. Political discussion affects politi-
cal participation, vote preferences, tolerance, and political
expertise (e.g., Huckfeldt 2001; Huckfeldt and Sprague
1995; Mutz 2006; Sokhey and McClurg 2012).
Lessons from research on political discussion link
nicely with the possibility that what people learn about
their acquaintances’ encounters with police and courts
influences their evaluations of the justice system. Infor-
mation is received at no cost, typically comes from trusted
sources, and potentially provides a relevant, on-point ba-
sis for gauging the actors in question.
The norm in work on political discussion is to ask
respondents to identify and provide information about
three or more discussion partners. The researcher can
ask about conversations regarding broad matters such
as “important problems” (e.g., Marsden 1987), narrower
concerns such as politics (e.g., Huckfeldt and Sprague
1995), specific matters such as politics within the work-
place (Mutz and Mondak 2006), or single issues such as
educational policy (e.g., Schneider et al. 1997).4We ad a p t
these approaches by asking respondents to identify ac-
quaintances who have had encounters with the police or
courts. These acquaintances will provide our operational
measure of a respondent’s justice encounters network.
We focus on what people glean from acquaintances’
experiences because we expect such anecdotes to be es-
pecially influential. The relaying of anecdotes, or “story-
telling” (Mandelbaum 2013), is a central part of conver-
sation.5Such tales can be vivid, such as when a neighbor
or coworker tells us of the car accident she witnessed, the
case she heard while on jury duty, or how she received a
warning instead of a ticket from a police officer. Because
it is lively, detailed, and comes from a trusted source, sto-
rytelling has a privileged position in terms of influence.
Walsh (2007) finds that the sharing of anecdotes is a key
way people communicate in community groups, and that
“narratives of personal experience, particularly of first-
hand experience with bearing the burden of discrimina-
tion, held a special authority in these groups” (152). We
envision comparable effects when anecdotal experiences
are shared through casual exchanges:
H3: Exposure to information about acquaintances’
justice encounters influences individuals’ evalu-
ations of the justice system.
Racial homogeneity in networks, combined with the
prevalence of perceived unfavorable experiences among
African Americans, dictates that Black respondents’ vi-
carious justice experiences will be especially negative. The
efficiencies associated with social communication show
why vicarious experiences are expected to influence in-
dividuals’ assessments of police and courts. Together, the
first three hypotheses indicate why the collective effect of
vicarious justice experiences should be a widening of the
racial divide in evaluations of judicial actors.
4For discussion of the attributes of different forms of name gener-
ators, including the benefits of more targeted name generators, see
Sokhey and Djupe (2014), Klofstad et al. (2013).
5Some research suggests that this is particularly true among African
Americans; see, for example, Harris-Lacewell (2004).

4JEFFERY J. MONDAK ET AL.
Our final hypothesis addresses the magnitude of this
divide. Learning of acquaintances’ bad experiences should
be especially influential. “Bad,” according to Vohs and
Luce (2010, 736), “is stronger than good. . . . [Across a
variety of domains,] events that yield negative outcomes
have a significantly greater psychological impact than
equivalent events that yield positive outcomes.” Psycholo-
gists have generated abundant evidence that people assign
negative information more weight than positive informa-
tion (e.g., Baumeister et al. 2001). Negative impressions
and stereotypes are quicker to form and more resistant to
disconfirmation than positive ones (e.g., Skowronski and
Carlston 1987).
Negativity biases have been found to operate in the
political realm (e.g., Lau 1982; McGraw and Steenbergen
1995; Soroka 2014), including on evaluations of the
Supreme Court (Grosskopf and Mondak 1998) and on
perceptions of fairness in evaluations of judicial actors
(Tyler 1990). People may have positive encounters with
police and courts that influence subsequent evaluations,
but we expect negative experiences to exert dispropor-
tionately strong effects on how police and court officials
are evaluated:
H4: Exposure to information about acquaintances’
negative justice encounters influences individ-
uals’ evaluations of the justice system more
strongly than does information about positive
justice encounters.
Data and Methods
Data for this study are from the 2012 Justice in Wash-
ington State Survey, an Internet survey administered by
YouGov. Surveys were completed between June 14 and
July 2, 2012. The survey obtained oversamples of Black,
Asian, and Latino respondents.6The data set includes re-
sponses from 611 Whites, 320 respondents who identified
themselves as being Asian/Pacific Islander, 305 Latinos,
and 288 African Americans.
Because our tasks involve comparisons of the four
groups, we retain the oversamples in their unweighted
form—that is, group samples are not weighted to reflect
Washington’s racial and ethnic composition. However, for
all analyses, weights calculated using Census marginals
for education, age, and gender are employed within each
group to help ensure that results reflect the attitudes and
6Ten cases are omitted because respondents’ ZIP Codes on file
either are located in states other than Washington or were entered
incompletely.ZIP Codes are not available for 17.6% of respondents.
experiences of members of the respective groups.7As with
any survey, caution must be exercised in generalizing to
the larger population. A related matter is whether there is
anything unique about Washington that might skew tests
of our theoretical framework. Because Washington has
a relatively small African American population, the po-
tential for racial homophily in networks is more limited
in Washington than elsewhere. This makes Washington
a conservative test case. Conversely, relative to the state’s
population, the overrepresentation of African Americans
in Washington’s prisons slightly exceeds the national aver-
age. If incarceration rates are a proxy for negative judicial
experiences, this would produce a liberal effect with re-
spect to our hypotheses. On balance, with these effects
cutting in opposite directions, we see Washington as a
reasonable test case.
The survey employed a focused prompt to cue re-
spondents to think about acquaintances who had inter-
acted with police and courts:
Many people have friends, relatives and other
acquaintances who have had encounters with the
justice system. These can be anything involving
the police or courts, such as calling the police
for help, talking with the police after a traffic
accident, being stopped by a police officer for
questioning or for a traffic violation, being placed
under arrest, going to court as a witness in a
case, going to court to serve as a juror, or being
a party in a criminal or civil court proceeding.
How many people do you know who have had
these kinds of encounters?
To help ensure that respondents had specific acquain-
tances in mind, they were asked to type in the first name or
initials of up to three such acquaintances. Overall, 26.1%
provided zero names, 18.2% offered one, 15.2% named
two, and 40.5% listed three.8Respondents were asked up
to four follow-up questions per acquaintance: the per-
son’s race and ethnicity; whether the encounter was with
the police, courts, or both; for police encounters, how the
police had behaved, evaluated on a scale ranging from –3
7All results with respect to our hypotheses also are found in repli-
cations using the unweighted data.
8For readers familiar with research on political discussion, these
numbers may seem high. We were not surprised by these marks.
First, the prompt we used did not focus on a limited time period.
Second, we expect justice encounters to be noteworthy, providing
fodder for conversation. Third, encounters with police and courts
are sufficiently common for most people to have acquaintances
who have had such encounters. For example, over 60 million people
interact with the police each year, with the majority of these being
traffic stops (Langton and Durose 2011).

BASES OF PERCEIVED INJUSTICE 5
(very unfair and disrespectful) to 3 (very fair and respect-
ful); and, for court encounters, how court officials had
behaved, using the same scale. We take the average across
respondents’ acquaintances to produce final scales of re-
spondents’ justice encounter networks for the police and
courts. These data provide our measures of the valence of
police and court encounters, both for individual network
members (Hypothesis 1) and for each respondent’s net-
work (Hypotheses 2, 3, and 4). Among respondents who
named at least one acquaintance with a police encounter,
the network mean is 0.06 (s.d. =1.79, N =1,062). The
corresponding mean for court networks is 0.41 (s.d. =
1.69, N =958).
The analyses testing Hypotheses 3 and 4 focus on two
dependent variables, evaluations of police and evaluations
of courts. These use data from four survey questions in-
cluded early in the instrument. For police, respondents
were asked, “Based on what you have heard or your own
experience, how often would you say the police gener-
ally treat all people with respect?” and “About how often
would you say that the police make fair, impartial (un-
biased) decisions in the cases they deal with?” The court
battery begins, “Here are some questions about the crim-
inal courts in the U.S. that deal with crimes such as house
burglary and physical assault.” Except for a reference to
courts, the first item has wording identical to the first po-
lice question, whereas the second asks, “How often do you
think the courts make fair and impartial decisions based
on the evidence available to them?” All items used 6-point
metrics (0 to 5). Summing the police questions yields a 0
(“never” on both items) to 10 (“always” on both items)
scale (mean =4.80; s.d. =2.18; r =0.69). A comparable
scale is formed using the two court items (mean =5.24;
s.d. =2.18; r =0.70).
Although our account emphasizes vicarious experi-
ences, personal experiences also are expected to be conse-
quential. To measure personal experiences with the police,
respondents were asked how many times they had been
treated disrespectfully by a police officer, and how many
times they were treated unfairly by the police because of
their race or ethnic background. Parallel questions were
used to measure experience with criminal courts. Re-
sponses range from 0 (never) to 4 (seven or more times).
Descriptive statistics are as follows: police unfair, mean =
0.87, s.d. =1.06; police disrespectful, mean =0.58, s.d. =
1.01; courts unfair, mean =0.32, s.d. =0.80; courts dis-
respectful, mean =0.31, s.d. =0.74. The means differ
substantially as a function of the race and ethnicity of
respondents: on 11 of 12 contrasts, African American re-
spondents report significantly worse personal experiences
than do others; on seven of eight tests, Latinos report sig-
nificantly worse experiences than Whites and Asians; on
both measures of disrespectful treatment, Asians report
worse experiences than Whites.
Multivariate models include the following as con-
trols: age (mean =45.03; s.d. =16.59); gender (1 =
female, 0 =male; 49.6% of respondents are female); ed-
ucation (0 =less than high school degree to 5 =post-
graduate; mean =2.22; s.d. =1.44); marital status (1 =
married, 0 =other; 50% of respondents are married); in-
come (an ordinal categorization with a maximum value
of 16; mean =5.96; s.d. =3.04); employment status (1 =
employed full time, 0 =other; 41.1% of respondents are
employed full time); importance of religion (0 =not at
all important to respondent, 3 =very important; mean =
1.72; s.d. =1.15); religious attendance (0 =never, 5 =
more than once a week; mean =1.77; s.d. =1.66); justice
encounter network size (0 to 3 acquaintances; mean =
1.70; s.d. =1.24); and party identification (1 =strong
Democrat, 7 =strong Republican; mean =3.47; s.d. =
1.93).9We also include quartile markers for population
density.
Exploring Vicarious Effects
Our theoretical framework builds on the assumption that
justice encounter networks will exhibit racial and ethnic
homophily. Networks, which could include up to three ac-
quaintances, average approximately the same size (1.72 to
1.84) for African Americans, Whites, and Latinos, and are
slightly smaller (1.44) for Asians. Homophily is suggested
if networks for each racial and ethnic group are domi-
nated by fellow members of that group. Strong evidence
is found. For Whites, 84.9% of network members are
White. For Asians, 55.4% of network members are Asian.
Fellow Latinos account for 47.4% of Latino respondents’
network members. Homophily is most noteworthy for
African Americans. Although accounting for under 4%
of Washington’s population, 74.5% of Black respondents’
network members are also African Americans.
Acquaintances’ Encounters with Police
and Courts
Our first hypothesis posits that the experiences of African
American network members will be perceived as hav-
ing been more negative than the experiences of Whites,
9Respondents who did not answer the party identification and in-
come items are assigned the scales’ middle (PID) or mean (income)
values. For both PID and income, indicator variables are included
to parse out any effects associated with nonresponse.

6JEFFERY J. MONDAK ET AL.
TABLE 1 Acquaintances’ Encounters with Police and Courts
Acquaintances’ Race/Ethnicity
White Asian African American Latino
A. Encounters with police
(–3 =most negative to 3 =most positive)
Percent positive encounters (scale values of 1 to 3) 55.3 47.7 25.3 32.2
Percent neutral encounters (0) 10.0 24.9 13.3 12.9
Percent negative encounters (–1 to –3) 34.7 27.4 61.4 54.9
Mean 0.50 0.49 −0.80 −0.49
Standard deviation 2.04 1.71 1.94 1.91
Number of acquaintances 1,150 228 527 294
B. Encounters with courts
Percent positive encounters (1 to 3) 59.7 59.2 30.9 36.9
Percent neutral encounters (0) 13.2 22.8 16.3 19.4
Percent negative encounters (–1 to –3) 27.1 18.0 52.8 43.7
Mean 0.71 1.00 −0.46 −0.13
Standard deviation 1.89 1.67 1.97 1.82
Number of acquaintances 995 172 405 227
Source: 2012 Justice in Washington State Survey.
Note: A total of 2,644 acquaintances were identified. An acquaintance can appear in both sections of this table if that person’s interactions
involved both the police and the courts; 102 acquaintances identified by respondents are omitted because they were either identified as
being Native American (37) or their race and ethnicity were not provided (65). All means for African Americans are significantly lower
than means for the other groups, and all means for Latinos are lower than means for Whites and Asians; the latter two groups do not differ
significantly.
Latinos, and Asians. Two points warrant emphasis. First,
the focus is the race and ethnicity of the network mem-
ber, not the respondent. Second, data reflect how ac-
quaintances’ experiences were seen by respondents; we
donothavereportsfromthenetworkmembersthem-
selves. This use of respondent reports is not problematic.
People’s views of their social contacts tend to be accurate.
Huckfeldt and Sprague (1995) show this in the context of
perceptions of presidential vote preferences. We assume
that justice encounters are vivid and salient to most peo-
ple, and thus should be at least as likely to be perceived
accurately. When speaking with a friend, we likely will get
the right take on whether she had a good or bad day in
court. Another key point noted by Huckfeldtand Sprague
(1995) is that perception underlies influence. If Mary per-
ceives that Max was mistreated by the police, Max most
likely would agree. After all, it probably was something
Max said that led Mary to perceivethat he had been t reated
poorly. However, even if Max disagrees, Mary’s perception
ofhisexperiencestillmayinfluenceherevaluationsofpo-
lice. Lastly, the most likely complication with respondent
reports is projection—that respondents extrapolate from
their own experiences when gauging the experiences of
their acquaintances—but all models include controls for
personal experience, meaning effects in the multivariate
models represent influence over and above any impact of
projection.
Respondents named 2,644 acquaintances who had
had justice encounters. Analyses pertain to 2,253 en-
counters with the police and 1,830 with courts. To test
Hypothesis 1, the data are aggregated within racial and
ethnic categories. We omit 102 who were not identified as
being Black, White, Asian, or Latino. Descriptive statis-
tics are reported in Table 1. For each group and target
of encounter (police or courts), we report the percentage
of experiences identified as positive, neutral, and nega-
tive. Beneath these are the mean and standard deviation
for the 7-point scale. For instance, for network members
who are White, 55.3% of police encounters are positive,
10.0% neutral, and 34.7% negative, with a mean on the
7-point scale of 0.50. On average, police experiences are
positive for Whites and Asians, negative for Latinos, and
sharply negative for Blacks. Court experiences are more
positive than police experiences, but the patterns between
the racial groups are identical for police and courts, and
the average experiences of Blacks are more than one full
point more negative than those of Whites and Asians.
It is when seen together with evidence of network
homophily that the data in Table 1 gain their full
significance. Due to network homogeneity, a racial gap

BASES OF PERCEIVED INJUSTICE 7
necessarily emerges with respect to vicarious experiences.
For White and Asian respondents, some 85% of their
encounters are with individuals who are themselves
Whites and Asians—and who generally have had positive
experiences. Conversely, for Latinos, over 60% of
network members are Latino or African American. For
Blacks, just under 80% of network members are African
American or Latino. Hence, vicarious justice encounters
most often convey moderately negative information for
Latinos, and exceedingly negative information for Blacks.
The Valence of Vicarious Justice Encounters
Our second hypothesis holds that social exposure to infor-
mation about negative justice experiences is most preva-
lent among African Americans, a virtual truism now that
we have support for our first hypothesis. Still, it will be
useful to devise precise estimates of how positive or neg-
ative the information is that respondents are exposed to
through social contact. To help ensure that we can be
confident that vicarious justice encounters truly vary by
race and ethnicity, these estimates should be derived from
multivariate analyses that also account for other relevant
attributes and predispositions.
Network measures are devised separately for police
and court encounters by taking the average of the re-
ported encounters of respondents’ acquaintances with
the police and courts. We control for total number of
acquaintances with justice encounters.10 At question is
the extent to which these values vary across groups, and
especially whether racial and ethnic differences persist in
multivariate specifications.
Results from six regression equations are reported in
Table 2. The first models for police and courts include only
indicator variables for race and ethnicity, with African
American as the excluded category; thus, these test for
differences in group means. The second model for each
dependent variable adds our control variables. Inclusion
of these controls enables determination of whether ap-
parent racial and ethnic differences in vicarious exposure
to information about the justice system trace partly to re-
spondents’ other attributes. Finally, we estimate a model
for each outcome including a control for the respondent’s
personal experiences with the police and courts.
For police encounters, Whites, Latinos, and Asians
report being exposed to more positive information from
10For now, we focus only on respondents who named acquain-
tances. Below, when the network measures are used as independent
variables, we include all respondents; those who did not name ac-
quaintances who had had police or court encounters receive values
of zero on the network measures.
their acquaintances than are African Americans (and
Whites and Asians to more positive information than
Latinos), providing support for Hypothesis 2. The pat-
tern across the four racial and ethnic groups is similar
irrespective of whether controls are included. Results for
court-related networks are similar, although the gap be-
tween African Americans and Latinos is now negligible.
These differences remain, even when controlling for re-
spondents’ own experiences.
What people learn from their acquaintances’ encoun-
ters with police and courts differs as a function of the race
and ethnicity of the individuals themselves. Results in
Table 2 reveal the magnitude of these effects and establish
that racial and ethnic disparities in vicarious justice ex-
periences persist after introduction of controls for a host
of other factors, providing strong support for the second
hypothesis.
Evaluations of Judicial Actors
Hypothesis 3 posits that information from people’s social
networks influences evaluations of police and courts. We
test this with focus on assessments of how often police and
courts treat people fairly and respectfully. Results of eight
ordinary least squares (OLS) regression models are re-
ported in Table 3. For both the police dependent variable
(first four models) and the court measure, the first model
includes the indicator variables for race and ethnicity,
along with our battery of control variables. The second
model adds measures of whether respondents themselves
have been treated disrespectfully and unfairly. Our cen-
tral measures, the network averages, are added in the
third model, and the network variables are retained in
Model 4 while the measures of personal experience are
omitted.
We note three results from the police models. First,
the coefficients for the network variable are substantively
impressive. The dependent variable ranges in value from
0 to 10 and has a standard deviation of 2.18. A shift across
the full range of the network measure corresponds with
a predicted 2.40-point swing in assessments of the po-
lice. Second, personal and vicarious experiences account
for a considerable portion of the variation in evaluations
of the police and courts. For instance, adding the expe-
riential variables to the baseline model more than dou-
bles the adjusted R2. Third, personal and vicarious ex-
periences also account for much of the initial racial and
ethnic variation seen in the first model. The coefficients
for the variables differentiating African Americans from
other respondents all drop once we add personal experi-
encesinthesecondmodel,andtheydropfurtherupon

8JEFFERY J. MONDAK ET AL.
TABLE 2 Determinants of Vicarious Exposure to Positive vs. Negative Information about Police
and Courts
Valence of Police Encounter
Networks
Valence of Court Encounter
Networks
Baseline
With
Controls
With
Personal
Contact Baseline
With
Controls
With
Personal
Contact
Constant −0.62 −0.76 0.18 −0.09 −0.44 −0.21
(0.12) (0.31) (0.31) (0.12) (0.32) (0.32)
White 0.99 0.85 0.61 0.68 0.50 0.37
(0.14) (0.16) (0.16) (0.15) (0.16) (0.16)
Asian 1.03 0.96 0.85 0.89 0.77 0.66
(0.17) (0.18) (0.18) (0.17) (0.18) (0.18)
Latino 0.49 0.43 0.32 0.27 0.23 0.20
(0.16) (0.17) (0.17) (0.17) (0.18) (0.17)
Age −0.00 −0.00 0.00 0.00
(0.00) (0.00) (0.00) (0.00)
Gender −0.06 −0.35 0.13 0.07
(0.11) (0.11) (0.11) (0.11)
Education −0.03 −0.04 −0.04 −0.06
(0.04) (0.04) (0.04) (0.04)
Marital status 0.26 0.08 0.12 −0.00
(0.12) (0.12) (0.12) (0.12)
Employment status −0.27 −0.25 −0.11 −0.18
(0.12) (0.11) (0.12) (0.12)
Importance of religion 0.03 0.04 0.01 0.05
(0.07) (0.07) (0.07) (0.07)
Religious attendance 0.03 0.07 0.04 0.04
(0.05) (0.05) (0.05) (0.05)
Party identification 0.02 0.00 0.03 0.05
(0.03) (0.03) (0.03) (0.03)
Don’t know or missing on party ID −0.15 −0.12 −0.43 −0.33
(0.21) (0.20) (0.21) (0.21)
Income 0.05 0.03 0.07 0.08
(0.02) (0.02) (0.02) (0.02)
Missing or prefer not to say on income 0.53 0.41 0.35 0.27
(0.20) (0.19) (0.20) (0.20)
Population: 2nd quartile 0.02 −0.04 0.02 −0.00
(0.17) (0.16) (0.17) (0.17)
Population: 3rd quartile −0.26 −0.21 −0.01 0.03
(0.16) (0.16) (0.16) (0.16)
Population: 4th quartile 0.00 −0.03 −0.01 −0.00
(0.16) (0.16) (0.16) (0.16)
Missing ZIP Code −0.32 −0.45 −0.24 −0.23
(0.19) (0.18) (0.19) (0.19)
Justice encounter network size −0.03 0.05 −0.09 −0.06
(0.07) (0.06) (0.07) (0.07)
(Continued)

BASES OF PERCEIVED INJUSTICE 9
TABLE 2Continued
Valence of Police Encounter
Networks
Valence of Court Encounter
Networks
Baseline
With
Controls
With
Personal
Contact Baseline
With
Controls
With
Personal
Contact
Respondent treated disrespectfully −0.11 −0.20
(0.07) (0.11)
Respondent treated unfairly −0.46 −0.32
(0.07) (0.10)
Adjusted R20.05 0.07 0.16 0.03 0.05 0.10
N 1,083 1,066 1,045 980 965 950
Source: 2012 Justice in Washington State Survey.
Note: Cell entries are OLS regression coefficients with standard errors in parentheses. The dependent variables range in value from –3 to 3
and capture the average valence, at the network level, of vicarious experiences with police and courts. Cases are included here only if the
network includes at least one member. For racial and ethnic classifications, African American is the omitted category.
introduction of the network measure in the third model.
Racial and ethnic gaps in evaluations of the police are
rooted largely in differences in individuals’ personal and
vicarious experiences.
Of these three points, the first two also are seen in
the court models. However, the residual racial and eth-
nic gaps shrink only modestly in these models. Even af-
ter accounting for personal and vicarious experiences,
Blacks remain significantly more critical of courts than
are Whites, Asians, and Latinos.11
The results in Table 3 provide evidence that vicarious
justice experiences matter for appraisals of both police
and courts, contributing to racial and ethnic differences
in those judgments. We will gauge the substantive mag-
nitude of these effects momentarily. First, the possibility
of a negativity bias must be explored.
Positive vs. Negative Vicarious Justice
Encounters
Our fourth hypothesis posits that a negativity bias exists
when people form evaluations of police and court offi-
cials. To test this, we estimate variants of the models in
11The coefficients for the vicarious experience variables indicate
that vicarious effects exist beyond the influence of projection based
on respondents’ own experiences, as the models control for per-
sonal experience. Still, it remains possible that the bulk of would-be
vicarious effects are rooted in projection, and the 0.40 coefficients
in Table 3 merely represent what remains once we control for
personal experience. To test this, the fourth models for each depen-
dent variable exclude personal experience. If the network variables
mostly capture projection, their coefficients should skyrocket once
the personal experience variables are removed. As can be seen, this
does not occur.
Table 3, but with the average network valence measures
split into separate positive and negative scales, and the
network size measure decomposed into separate counts
of the number of respondents’ acquaintances with pos-
itive, negative, or neutral experiences. For each of the
valence measures, values range from 0 (no network expo-
sure to information with that valence) to 3 (a network of
acquaintances with uniformly extreme negative or pos-
itive experiences). Network size measures range from 0
(no acquaintances were identified) to 3. The coefficients
on the valence and network size measures should pro-
duce opposite signs, with positive effects emanating from
positive encounters and negative effects resulting from
perceived unfairness and disrespect. The hypothesis is
supported if the absolute values of the coefficients for
negative encounters exceed those for positive encounters.
Both models in Table 4 reveal negativity effects. This
is especially evident when looking at the network size vari-
ables. Adding another acquaintance with a positive expe-
rience with the police or courts appears to have no effect
on respondents’ evaluations (although receiving consis-
tently positive information about these encounters does
appear to move attitudes), whereas the effects of adding
another acquaintance with a negative experience are large
in both models. Negative vicarious experiences strongly
influence individuals’ assessments of judicial actors, and
Blacks have the most negative vicarious experiences.
To summarize the impact of vicarious experiences on
racial gaps in evaluations of police and courts, we calcu-
late predicted values on the dependent variables under
four scenarios. In each, estimates draw on coefficients in
Table 4, with all variables except race and ethnicity, and
personal and vicarious experiences, held constant at mean

10 JEFFERY J. MONDAK ET AL.
TABLE 3 Impact of Vicarious Justice Encounters on Evaluations of Police and Courts
Evaluations of Police Evaluations of Courts
Constant 2.73 3.99 4.07 3.28 3.74 4.30 4.59 4.16
(0.28) (0.29) (0.29) (0.27) (0.28) (0.29) (0.29) (0.28)
Respondent treated disrespectfully −0.15 −0.13 −0.45 −0.39
(0.08) (0.08) (0.11) (0.11)
Respondent treated unfairly −0.44 −0.27 −0.15 −0.04
(0.07) (0.07) (0.10) (0.10)
Vicarious experience 0.40 0.46 0.41 0.44
(0.04) (0.03) (0.04) (0.04)
Justice encounter network size −0.11 −0.15 −0.13 −0.13
(0.04) (0.04) (0.04) (0.04)
White 0.86 0.51 0.33 0.54 0.97 0.78 0.68 0.82
(0.16) (0.17) (0.16) (0.15) (0.17) (0.17) (0.16) (0.16)
Asian 0.68 0.41 0.14 0.27 1.10 0.97 0.74 0.83
(0.18) (0.18) (0.17) (0.17) (0.18) (0.18) (0.18) (0.18)
Latino 0.85 0.59 0.49 0.67 0.86 0.75 0.70 0.79
(0.18) (0.18) (0.17) (0.17) (0.18) (0.18) (0.18) (0.18)
Age 0.01 0.01 0.01 0.01 0.01 0.00 0.00 0.00
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
Gender −0.25 −0.48 −0.35 −0.21 −0.13 −0.26 −0.25 −0.15
(0.11) (0.11) (0.11) (0.11) (0.11) (0.11) (0.11) (0.11)
Education 0.03 0.01 0.03 0.05 0.02 −0.01 0.01 0.04
(0.04) (0.04) (0.04) (0.04) (0.04) (0.04) (0.04) (0.04)
Marital status 0.19 0.13 0.12 0.13 0.18 0.10 0.11 0.16
(0.12) (0.12) (0.11) (0.11) (0.12) (0.12) (0.12) (0.12)
Employment status −0.02 0.02 0.08 0.07 0.27 0.22 0.28 0.32
(0.12) (0.12) (0.11) (0.11) (0.12) (0.12) (0.12) (0.12)
Importance of religion 0.14 0.11 0.11 0.13 0.18 0.20 0.20 0.18
(0.07) (0.07) (0.07) (0.07) (0.07) (0.07) (0.07) (0.07)
Church attendance 0.05 0.08 0.05 0.03 −0.01 0.01 −0.01 −0.02
(0.05) (0.05) (0.04) (0.04) (0.05) (0.05) (0.05) (0.05)
Party identification 0.21 0.19 0.19 0.20 0.04 0.05 0.03 0.04
(0.03) (0.03) (0.03) (0.03) (0.03) (0.03) (0.03) (0.03)
Don’t know or missing on party ID 0.38 0.37 0.40 0.42 0.07 0.10 0.17 0.16
(0.21) (0.20) (0.19) (0.19) (0.21) (0.21) (0.20) (0.20)
Income 0.01 −0.01 −0.02 −0.00 0.00 0.00 −0.02 −0.02
(0.02) (0.02) (0.02) (0.02) (0.02) (0.02) (0.02) (0.02)
Missing or prefer not to say on income −0.03 −0.17 −0.31 −0.23 −0.31 −0.35 −0.44 −0.43
(0.19) (0.19) (0.18) (0.18) (0.20) (0.20) (0.19) (0.19)
Population: 2nd quartile −0.28 −0.36 −0.34 −0.29 −0.02 −0.11 −0.10 −0.02
(0.17) (0.17) (0.16) (0.16) (0.17) (0.17) (0.17) (0.17)
Population: 3rd quartile −0.18 −0.21 −0.15 −0.09 −0.08 −0.09 −0.08 −0.07
(0.17) (0.16) (0.16) (0.16) (0.17) (0.17) (0.16) (0.17)
Population: 4th quartile −0.11 −0.17 −0.15 −0.10 −0.33 −0.35 −0.33 −0.31
(0.17) (0.16) (0.16) (0.16) (0.17) (0.17) (0.16) (0.16)
Missing ZIP Code −0.38 −0.46 −0.32 −0.25 −0.34 −0.32 −0.21 −0.21
(0.19) (0.18) (0.17) (0.18) (0.19) (0.19) (0.18) (0.18)
(Continued)

BASES OF PERCEIVED INJUSTICE 11
TABLE 3Continued
Evaluations of Police Evaluations of Courts
Adjusted R2.11 .16 .23 .21 .05 .08 .14 .12
N 1,479 1,442 1,442 1,479 1,473 1,443 1,443 1,473
Source: 2012 Justice in Washington State Survey.
Note: Cell entries are OLS regression coefficients with standard errors in parentheses. The dependent variables range in value from 0 to 10
and represent the extent to which respondents assess police and courts.
TABLE 4 Negativity Effects and the Impact of Vicarious Justice Encounters on Evaluations of Police
and Courts
Police Courts
Constant 4.04 (0.29) 4.57 (0.29)
Respondent treated disrespectfully −0.14 (0.08) −0.41 (0.11)
Respondent treated unfairly −0.25 (0.07) −0.02 (0.10)
Vicarious experience—positive encounters 0.26 (0.08) 0.21 (0.09)
Vicarious experience—negative encounters −0.33 (0.09) −0.27 (0.11)
Network size—positive encounters 0.01 (0.10) 0.06 (0.11)
Network size—negative encounters −0.21 (0.10) −0.29 (0.13)
Network size—neutral encounters 0.03 (0.06) −0.10 (0.06)
White 0.32 (0.16) 0.67 (0.16)
Asian 0.11 (0.17) 0.73 (0.18)
Latino 0.50 (0.17) 0.70 (0.18)
Age 0.01 (0.00) 0.00 (0.00)
Gender −0.36 (0.11) −0.25 (0.11)
Education 0.02 (0.04) 0.01 (0.04)
Marital status 0.12 (0.11) 0.11 (0.12)
Employment status 0.11 (0.11) 0.28 (0.12)
Importance of religion 0.11 (0.07) 0.20 (0.07)
Church attendance 0.05 (0.04) −0.01 (0.05)
Party identification 0.19 (0.03) 0.04 (0.03)
Don’t know or missing on party ID 0.39 (0.19) 0.18 (0.20)
Income −0.02 (0.02) −0.02 (0.02)
Missing or prefer not to say on income −0.29 (0.18) −0.41 (0.19)
Population: 2nd quartile −0.34 (0.16) −0.12 (0.17)
Population: 3rd quartile −0.14 (0.16) −0.07 (0.16)
Population: 4th quartile −0.15 (0.16) −0.33 (0.16)
Missing ZIP Code −0.33 (0.18) −0.22 (0.19)
Adjusted R2.23 .14
N 1,442 1,443
Source: 2012 Justice in Washington State Survey.
Note: Cell entries are OLS regression coefficients with standard errors in parentheses. The dependent variables are the same as in Table 3.
or modal values. In the first scenario, we assume that re-
spondents have had no personal or vicarious encounters.
The second scenario adds in personal experiences. We do
so using the racial and ethnic group averages on the two
personal experience measures, with these averages mul-
tiplied by the appropriate coefficients from Table 4. For
example, for the variable regarding instances in which
the respondent has been treated unfairly by the police,
the mean for Whites is 0.71, versus 1.28 for Blacks. Mul-
tiplication of the group means by the –0.25 coefficient
in Table 4 yields an average reduction in evaluations of
police of 0.18 for Whites and 0.32 for Blacks. The third
scenario adds vicarious experiences but assumes the re-
spondent has had no personal experiences. Here, we also

12 JEFFERY J. MONDAK ET AL.
FIGURE 1 Estimated Effects of Personal and Vicarious Experiences on
Respondents’ Evaluations of Police and Courts
Latinos
Asians
Whites
Blacks ●
●
●
●
456
Evaluations of Police
Latinos
Asians
Whites
Blacks
456
Evaluations of Courts
Latinos
Asians
Whites
Blacks
456
Latinos
Asians
Whites
Blacks
456
Latinos
Asians
Whites
Blacks
456
Latinos
Asians
Whites
Blacks
456
Latinos
Asians
Whites
Blacks
456
Latinos
Asians
Whites
Blacks
456
A. No personal or vicarious experiences
B. Personal experiences only
C. Vicarious experiences only
D. Personal and vicarious experiences
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
Note: Estimates are derived from regression results reported in Table 4. For personal and vicarious
effects, estimates are calculated using the average valence of personal and vicarious experiences
for members of the respective racial and ethnic groups. Vicarious experiences reflect both average
network valence and size. Higher values indicate more positive evaluations. Scales on the dependent
variables range from 0 to 10.
use the racial and ethnic group averages, this time for
the positive and negative network measures, drawing on
only those cases in which respondents named at least one
acquaintance reporting a police or court encounter. The
final scenario assumes respondents have had both per-
sonal and vicarious experiences equal to the average for
their respective racial and ethnic groups.
Estimates are depicted in Figure 1. In the upper
left section of the figure, Blacks’ views of the police are
modestly more critical than those of other respondents
when personal and vicarious experiences are absent; the
Black-White difference is 0.32 points. When personal
experience is added in the second section, the gaps widen,
including an increase to 0.62 points when Blacks and
Whites are contrasted. Similarly large disparities emerge
in the third section, where estimates are for respondents
with no personal experience with police, but with vicar-
ious experiences typical of those for members of their
respective racial and ethnic groups. Here, the Black-White
difference is 0.63 points, suggesting vicarious experiences
contribute as much to the Black-White divide in perceptions
of the police as do personal experiences.Inthefinalsection,
we account for both personal and vicarious experiences.
The Black-White disparity is now 0.94 points, nearly

BASES OF PERCEIVED INJUSTICE 13
triple the gap for respondents with neither personal nor
vicarious experiences. A roughly similar pattern is seen
for opinions about courts in the estimates on the right side
of Figure 1: Vicarious experiences again exert as much
or more impact as personal experiences, and the Black-
White gap in the bottom section reaches 1.09 points.
It is noteworthy that baseline estimates—racial gaps
for respondents with neither personal nor vicarious
experiences—are larger for assessments of courts than
the police. Once personal and vicarious differences are
set aside, African Americans are only modestly more crit-
ical of the police than are individuals from the other
groups. Experiences also matter for views of the courts,
but estimates in the upper right section of Figure 1 sug-
gest that racial disparities run deeper, with roots beyond
the experiences of individuals and their acquaintances.
Inferring Social Influence
We have assumed that any correspondence between our
network measures and respondents’ views of police and
courts reflects the causal influence of the former on
the latter. In actuality, the definitive establishment of
social influence is not possible with observational data
(Huckfeldt et al. 2013, 674–78). In the present case, the
observed widening of the racial gaps in perceptions of po-
lice and courts stems from three key forces: (1) people’s
African American acquaintances are disproportionately
likely to be perceived as having had negative experiences
with police and courts, (2) African American respondents
are disproportionately likely to have African American
acquaintances, and (3) what people perceive about their
acquaintances’ experiences influences their own views of
police and courts. Compared to many instances of pos-
sible social influence, we see a stronger prima facie case
here, as we are aware of no meaningful controversy per-
taining to any of these claims. Nonetheless, it remains
conceivable that the racial gap in vicarious experiences is
somehow illusory. One possibility is that African Ameri-
can respondents see the justice system as discriminatory
and thus assume that their African American acquain-
tances have been treated unfairly and disrespectfully by
police and court officials. A related possibility is that re-
spondents who hold negative views of the justice system
find it easiest to call to mind acquaintances who have had
bad experiences.
The first scenario presupposes that Black respondents
exaggerate the poor experiences of their Black friends. An
indirect test is possible by considering only the views of
Whites, Asians, and Latinos. If these respondents report
that their African American acquaintances are treated
disproportionately unfairly by the police and courts, their
views would lend credence to those reported by Black re-
spondents. Non-Black respondents named 153 Black ac-
quaintances with police encounters, and 117 with court
encounters. Of these, 61% and 54%, respectively, were
reported to be negative. In contrast, among these respon-
dents’ White and Asian acquaintances,12 the results are
32% negative for police encounters and 26% negative for
court encounters. Hence, there is consensus among all
respondents that African American acquaintances have
been treated especially poorly by judicial actors. Results
with respect to a widening of the racial gap are not the
consequence of Black respondents seeing a different re-
ality than Whites, Asians, and Latinos. Instead, because
of homophily, Black respondents vicariously experience a
different reality.
The second scenario is that respondents who view the
justice system negatively find it relatively easy to call to
mind acquaintances who have had negative experiences.
Here, rather than social influence, our results would signal
an accessibility effect. An indirect test is again possible.
If we account for the possibility that responses on our
dependent variables drive what respondents recall, then
we can test whether any signs remain of a racial gap in
acquaintances’ experiences. To do this, we focus only on
respondents who reported neutral or positive opinions
on our dependent measures, the evaluations of police
and courts. Among these respondents, 40% report that
their Black acquaintances had negative experiences with
the police, and 34% had bad experiences with the courts.
For White and Asian acquaintances, the results are, re-
spectively, 23% and 22% for police experiences and 21%
and 15% for court experiences. The experiences of re-
spondents’ African American acquaintances are dispro-
portionately negative, even among respondents whose
personal views of the police and courts are favorable.
As a final test, we make use of four 101-point scales
that ask respondents how fair the justice system is to
each of the four racial and ethnic groups under consid-
eration. Using these data, we selected only those respon-
dents who feel that Blacks are not treated more unfairly by
the justice system than all other groups. We then exam-
ine the reported experiences of these respondents’ Black
and non-Black network members. These respondents re-
port that 53% of their Black network members’ police
experiences were negative, and that 49% of their court
experiences were negative. The corresponding figures for
the remainder of network members are 36% for police
12We omit Latino acquaintances. In Table 1, the experiences of
Latino acquaintances were more positive than those of Blacks, but
considerably more negative than those of Whites and Asians.

14 JEFFERY J. MONDAK ET AL.
and 29% for courts. Hence, even among respondents
who are not prone to assume that African Americans are
treated more unfairly by the justice system than are mem-
bers of all other racial and ethnic groups, the reported
experiences of these respondents’ Black network mem-
bers are more negative than the reported experience of
their other network members.
We believe the present case is highly compelling. No
matter how we disaggregate our data, perceived nega-
tive encounters are most common among respondents’
African American associates. The most plausible inter-
pretation is that, in actuality, Blacks are disproportion-
ately likely to be treated unfairly and disrespectfully in
their encounters with police and court officials. If there is
any degree of racial and ethnic homogeneity in people’s
networks, and if respondents are not oblivious to the ex-
periences of their acquaintances, then the consequence of
vicarious experience will be a widening of the racial gap
in perceptions of the justice system.
Conclusions
The racial divide in perceptions of police and courts con-
ceivablycouldbetracedtonumerousfactors.Inthisstudy,
we have emphasized the role of vicarious experience. In
this account, a chief reason many African Americans ex-
press characteristically negative views of police and courts
is the information they have obtained about their ac-
quaintances’ experiences. The encounters we have high-
lighted concern matters of process. If people understand
their acquaintances have been treated unfairly or disre-
spectfully by justice officials, they will be more critical of
those actors.
Three points warrant emphasis. First, although we
have shown that our framework helps in understanding
racial divides in views of police and courts, race plays no
direct role in the judgmental process we have outlined.
Our third hypothesis posits that exposure to information
about acquaintances’ experiences influences individuals’
evaluations of the justice system. Race plays no neces-
sary role in this hypothesis. This same process could
be seen in a racially homogeneous context.13 Vicarious
effects contribute to racial disparities in views of police
13The key point is that what differs among racial and ethnic groups
is the valence of the information to which they are exposed, not the
processes by which information is used to guide subsequent judg-
ments. We have no basis to hypothesize process differences (e.g.,
that Whites care about personal experiences but not vicarious ones,
whereas Latinos care only about vicarious experiences). However,
as a post hoc check, we ran models separately by group. No note-
worthy differences emerged with respect to our hypotheses. For
and courts because of two prior dynamics. The first is
racial and ethnic homophily in network composition.
Second, perceived experiences with police and courts dif-
fer dramatically as a function of race and ethnicity. Across
all racial and ethnic groups, our respondents report that
their African American acquaintances receive the worst
treatment in the justice system.14 Given racial homo-
geneity in network composition, it follows that African
American respondents are vicariously exposed to dispro-
portionately more negative information than are Whites,
Latinos, and Asians.
Asecondpointwewishtoemphasizeisthatthe
present results likely understate the experiential bases of
evaluations of judicial actors. Our measures of vicari-
ous experiences provide information only on immedi-
ate acquaintances, and only about three individuals. The
measures omit information acquired secondhand. They
may capture what Max told Mary about Max’s experi-
ence, but not what Max told Mary about his cousin’s
experience, and not what Mary learned from the me-
dia. Also, over 40% of respondents named three acquain-
tances who had had justice encounters, suggesting that
future work should seek to represent network content
even more broadly.
The third point that warrants emphasis is that the
present results, while providing support for our hypothe-
ses, also raise new questions. Perhaps most importantly,
the results suggest that personal and vicarious experiences
account for more of the racial gap in evaluations of the
police than the gap in assessments of courts. The police
models exhibited greater explanatory power, and residual
racial and ethnic gaps—the remaining effects of the in-
dicator variables for race and ethnicity, after accounting
for personal and vicarious experiences—are modest. To
be sure, experiences matter for assessments of courts. But
the skepticism toward courts expressed by African Amer-
ican respondents on our survey apparently also signals a
deeper and more fundamental mistrust, one seemingly
emanating from beyond the realm of the immediate ex-
periences of respondents and their acquaintances.
instance, in Table 3, the network measures produce coefficients of
0.40 and 0.41, respectively, in the police and court models. When
estimated separately by group, the police network coefficients are
0.41 (Whites), 0.32 (Asians), 0.37 (Blacks), and 0.36 (Latinos); the
court coefficients are 0.44 (Blacks), 0.44 (Latinos), 0.33 (Whites),
and 0.53 (Asians).
14As we emphasized earlier, the perceived poor treatment of African
Americans within the justice system may reflect actual unfairness
and disrespect on the basis of race, but other scenarios, ones rooted
in factors such as the severity of the incidents in question, also may
underlie these perceptions. Put differently, both our theoretical
account and our empirical analyses start with what is perceived,
not with the forces fueling these perceptions.

BASES OF PERCEIVED INJUSTICE 15
Moving forward, the nuances of justice encounter
networks require greater scrutiny. The specific effects
identified in this study are far from the final word on
the vicarious bases of justice perceptions. In the political
discussion literature, early findings of discussant influ-
ence on the vote choice gave rise to a wealth of research
regarding different types of dyadic relationships, the fre-
quency of interaction between discussion partners, the
significance of exposure to crosscutting views, and so on.
That history offers a model for future research on the
social bases of evaluations of judicial actors.15 What we
know thus far is that vicarious experiences influence peo-
ple’s appraisals, and in a manner that accentuates racial
gaps in those assessments. This initial evidence points to
the likely value of more nuanced explorations regarding
what specific sorts of information about police and courts
are communicated, and what characteristics of discussant
relationships moderate this form of social influence.
References
Baumeister, Roy F., Ellen Bratslavsky, Catrin Finkenauer, and
Kathleen D. Vohs. 2001. “Bad Is Stronger Than Good.”
Review of General Psychology 5(4): 323–70.
Bobo, Lawrence D.,and Devon Johnson. 2004. “A Taste for Pun-
ishment: Black and White Americans’ Views on the Death
Penalty and the War on Drugs.” Du Bois Rev iew1(1): 151–80.
Engel, Robin S., and Kristin Swartz. 2013. “Race, Crime, and
Policing.” In The Oxford Handbook of Ethnicity, Crime, and
Immigration, ed. Sandra M. Bucerius and Michael Tonry.
New York: Oxford University Press, 135–65.
Grosskopf, Anke, and Jeffery J. Mondak. 1998. “Do Attitudes
toward Specific Supreme Court Decisions Matter? The Im-
pact of Webs t e r and Te x a s v. J o hnson on Public Confidence
in the Supreme Court.” Political Research Quarterly 51(3):
633–54.
Harris-Lacewell, Melissa V. 2004. Barbershops, Bibles, and BET:
Everyday Talk and Black Political Thought.Princeton,NJ:
Princeton University Press.
Harrison, Paige M., and Allen J. Beck. 2003. Prisoners in 2002.
Washington, DC: U.S. Department of Justice, Bureau of Jus-
tice Statistics.
15Our data permit some more specific tests inspired by past work
on political discussion. One test pertains to the intimacy of dyadic
relationships. If first-named acquaintances are more closely linked
to respondents and if intimacy matters for the strength of social
influence, then network effects should be strongest for first-named
acquaintances. We find evidence consistent with this for courts, but
not for police. A related possibility is that group cohesiveness mag-
nifies social influence. In this scenario, network members from the
respondent’s own racial or ethnic group should be more influential
than members from other groups. In contrast with this hypothesis,
our tests yielded strong evidence that network members from all
racial and ethnic groups are equally influential.
Huckfeldt, Robert. 2001. “The Social Communication of Polit-
ical Expertise.” American Journal of Political Science 45(2):
425–38.
Huckfeldt, Robert, Jeffery J. Mondak, Matthew Hayes, Matthew
T. Pietryka, and Jack Reilly. 2013. “Networks, Interdepen-
dence, and Social Influence in Politics.” In The Oxford
Handbook of Political Psychology, 2nd ed., ed. Leonie Huddy,
David O. Sears, and Jack S. Levy. New York: Oxford Univer-
sity Press, 662–98.
Huckfeldt, Robert, and John Sprague. 1995. Citizens, Politics,
and Social Communication: Information and Influence in
an Election Campaign. New York: Cambridge University
Press.
Ibarra, Herminia. 1995. “Race, Opportunity and Diversity of
Social Circles in Managerial Networks.” Academy of Man-
agement Journal 38(3): 673–703.
Jacobs, David, and Robert M. O’Brien. 1998. “The Determi-
nants of Deadly Force: A Structural Analysis of Police Vio-
lence 1.” American Journal of Sociology 103(4): 837–62.
Katz, Elihu, and Paul F. Lazarsfeld. 1955. Personal Influence: The
Part Played by People in the Flow of Mass Communications.
Glencoe, IL: Free Press.
Klofstad, Casey A., Anand Edward Sokhey, and Scott D.
McClurg. 2013. “Disagreeing about Disagreement: How
Conflict in Social Networks Affects Political Behav-
ior.”American Journal of Political Science 57(1): 120–34.
Langton, Lynn, and Matthew Durose. 2011. Police Behavior dur-
ing Traffic and Street Stops, 2011. Washington, DC: U.S. De-
partment of Justice, Bureau of Justice Statistics.
Lau, Richard R. 1982. “Negativity in Political Perception.” Po-
litical Behavior 4(4): 353–77.
Mandelbaum, Jenny. 2013. “Storytelling in Conversation.” In
The Handbook of Conversation Analysis, ed. Jack Sidnell and
Tanya Stivers. Chichester, UK: John Wiley & Sons, 492–507.
Marsden, Peter V. 1987. “Core Discussion Networks of Ameri-
cans.” American Sociological Review 52(1): 122–31.
Marsden, Peter V. 1988. “Homogeneity in Confiding Relations.”
Social Networks 10(1): 57–76.
McGraw, Kathleen M., and Marco Steenbergen. 1995. “Pictures
in the Head: Memory Representations of Political Candi-
dates.” In Political Judgment: Structure and Processes,ed.
Milton Lodge and Kathleen M. McGraw. Ann Arbor: Uni-
versity of Michigan Press, 15–41.
McPherson, Miller, Lynn Smith-Lovin, and James M. Cook.
2001. “Birds of a Feather: Homophily in Social Networks.”
Annual Review of Sociology 27(1): 415–44.
Mutz, Diana C. 2006. Hearing the Other Side: Deliberative versus
Participatory Democracy. New York: Cambridge University
Press.
Mutz, Diana C., and Jeffery J. Mondak. 2006. “The Workplace
as a Context for Cross-Cutting Political Discourse.” Journal
of Politics 68(1): 140–55.
Peffley, Mark, and Jon Hurwitz. 2010. Justice in America: The
Separate Realities of Blacks and Whites.NewYork:Cambridge
University Press.
Rosenbaum, Dennis P., Amie M. Schuck, Sandra K. Costello,
Darnell F. Hawkins, and Marianne K. Ring. 2005. “Attitudes
toward the Police: The Effects of Direct and Vicarious Expe-
rience.” Police Quarterly 8(3): 343–65.

16 JEFFERY J. MONDAK ET AL.
Sakala, Leah. 2014. “Breaking Down Mass Incarceration
in the 2010 Census: State-by-State Incarceration Rates
by Race/Ethnicity.” Prison Policy Initiative. http://www.
prisonpolicy.org/reports/rates.html.
Schneider, Mark, Paul Teske, Melissa Marschall, and Chris-
tine Roch. 1997. “Networks to Nowhere: Segregation and
Stratification in Networks of Information about Schools.”
American Journal of Political Science 41(4): 1201–23.
Shrum, Wesley, Neil H. Cheek Jr., and Saundra MacD.
Hunter. 1988. “Friendship in School: Gender and Racial
Homophily.” Sociology of Education 61(4): 227–39.
Skowronski, John J., and Donald E. Carlston. 1989. “Negativity
and Extremity Biases in Impression Formation: A Review of
Explanations.” Psychological Bulletin 106(1): 131–42.
Sokhey, Anand E., and Paul A. Djupe. 2014 “Name Generation
in Interpersonal Political NetworkData: Results from a Series
of Experiments.” Social Networks 36: 147–161.
Sokhey, Anand E., and Scott D. McClurg. 2012. “Social Net-
works and Correct Voting.” Journal of Politics 74(3): 751–64.
Soroka, Stuart N. 2014. Negativity in Democratic Politics: Causes
and Consequences. New York: Cambridge University Press.
Spohn, Cassia. 2013. “Racial Disparities in Prosecution, Sen-
tencing, and Punishment.” In The Oxford Handbook of Eth-
nicity, Crime, and Immigration, ed. Sandra M. Bucerius and
Michael Tonry. New York: Oxford University Press, 166–93.
Tonry, Michael. 2011. Punishing Race: A Continuing American
Dilemma. New York: Oxford University Press.
Tyler, Tom R. 1990. WhyPeopleObeytheLaw. New Haven, CT:
Yale University Press.
Unnever, James D. 2008. “Two Worlds Far Apart: Black-White
Differences in Beliefs about Why African-American Men
Are Disproportionately Imprisoned.” Criminology 46(2):
511–38.
Vohs, Kathleen D., and Mary Frances Luce. 2010. “Judgment
and Decision Making.” In Advanced Social Psychology: The
StateoftheScience, ed. Roy F. Baumeister and Eli J. Finkel.
New York: Oxford University Press, 733–56.
Walker, Samuel, Cassia Spohn, and Miriam DeLone. 2004. The
Color of Justice: Race, Ethnicity, and Crime in America.3rd
ed. Belmont, CA: Thomson Learning.
Walsh, Katherine Cramer. 2007. Talking about Race: Community
Dialogues and the Politics of Difference. Chicago: University
of Chicago Press.
Weitzer, Ronald, and Steven A. Tuch. 2006. Race and Policing
in America: Conflict and Reform. New York: Cambridge Uni-
versity Press.
Wilson, James Q. 1978. Varieties of Police Behavior. Cambridge,
MA: Harvard University Press.
Supporting Information
Additional Supporting Information may be found in the
online version of this article at the publisher’s website:
Tab l e A 1: Descriptive Statistics