motivation to call police: the exploration of racial and risk averse … · 2020. 4. 24. ·...
Post on 13-Sep-2020
1 Views
Preview:
TRANSCRIPT
University of Northern IowaUNI ScholarWorks
Theses and Dissertations @ UNI Graduate College
2019
Motivation to call police: The exploration of racialand risk averse motivationAlivia Lauren ZubrodUniversity of Northern Iowa
Copyright ©2019 Alivia Lauren ZubrodFollow this and additional works at: https://scholarworks.uni.edu/etd
Let us know how access to this document benefits you
This Open Access Thesis is brought to you for free and open access by the Graduate College at UNI ScholarWorks. It has been accepted for inclusion inTheses and Dissertations @ UNI by an authorized administrator of UNI ScholarWorks. For more information, please contact scholarworks@uni.edu.
Recommended CitationZubrod, Alivia Lauren, "Motivation to call police: The exploration of racial and risk averse motivation" (2019). Theses and Dissertations@ UNI. 989.https://scholarworks.uni.edu/etd/989
Copyright by
ALIVIA LAUREN ZUBROD
2019
All Rights Reserved
MOTIVATION TO CALL POLICE: THE EXPLORATION OF RACIAL AND RISK
AVERSE MOTIVATION
An Abstract of a Thesis
Submitted
In Partial Fulfillment
of the Requirements for the Degree
Master of Arts
Alivia Lauren Zubrod
University of Northern Iowa
July 2019
ABSTRACT
When calls are made to the police, the magnitude of their impact is often overlooked.
When calls are made to the police and there is no crime, police resources, time, and
energy could be wasted (Sampson, 2002); however, when no call is made to the police
and there is a crime, human lives could be put in danger (e.g., Felson, Messner, Hoskin,
& Deane, 2002). Based on highly publicized news reports, it appears that being a racial
minority is enough motivation to call the police in some situations (e.g., napping or
humanitarian work; Griggs, 2018; Williams, 2018b). Aversive racism theory (Gaertner &
Dovidio, 1986) suggests that when racial prejudice can be rationalized to another factor
aside from race, then aversive racists may act in discriminatory ways. Thus, a person may
rationalize a call to the police based on someone yelling rather than their skin color. Risk
averse motivation (Kahneman & Tversky, 1982) suggests that individuals prefer a certain
choice compared to an uncertain choice. Thus, individuals who call the police could be
risk averse and choose to call the police to provide a sense of certainty in an uncertain
situation. In this study, I tested these two possible motivations using an ambiguous risk
scenario. Participants (N = 295) from an online data collection platform read a scenario
and reported their likelihood to call the police, whether they would call the police
(yes/no), and whether they agreed with someone else’s decision to call the police based
on the scenario. Then participants completed a risk perception scale. The race of the
perceived suspect was not influential in the reported likelihood to call the police, whether
a participant would call the police, or their agreement with someone else’s decision to
call the police; however, participants who were risk averse, as well as women and
political conservatives reported a greater likelihood to call the police, were more likely to
report that they would call the police, and agreed more with someone else’s decision to
call the police. Despite the results of the current study, there are still news reports that
suggest racial minorities are the source of motivation for calls to the police. Thus, race as
a potential motivation to call the police should be continued to be examined.
Keywords: motivation, race, aversive racism, risk averse, police
MOTIVATION TO CALL POLICE: THE EXPLORATION OF RACIAL AND RISK
AVERSE MOTIVATION
A Thesis
Submitted
In Partial Fulfillment
of the Requirements for the Degree
Master of Arts
Alivia Lauren Zubrod
University of Northern Iowa
July 2019
ii
This Study by: Alivia Lauren Zubrod
Entitled: Motivation to Call Police: The Exploration of Racial and Risk Averse
Motivation
has been approved as meeting the thesis requirement for the
Degree of Master of Arts
___________ ____________________________________________________
Date Dr. Jiuqing Cheng, Chair, Thesis Committee
___________ ____________________________________________________
Date Dr. Helen C. Harton, Thesis Committee Member
___________ ____________________________________________________
Date Dr. Matthew Makarios, Thesis Committee Member
___________ _____________________________________________________
Date Dr. Jennifer J. Waldron, Dean, Graduate College
iii
DEDICATION
For my Uncle John, I miss you every day.
iv
TABLE OF CONTENTS
PAGE
LIST OF TABLES ............................................................................................................ vii
LIST OF FIGURES ......................................................................................................... viii
CHAPTER 1. INTRODUCTION ...................................................................................... 1
Decision to Call the Police .............................................................................................. 3
Racial Differences in Calling the Police ......................................................................... 6
Racial Differences in Arrest Rates .................................................................................. 7
Racial Differences in Incarceration Rates ....................................................................... 8
Aversive Racism Theory ................................................................................................. 9
Risk Averse Motivation ................................................................................................ 11
Other Explanations: Gender and Political Orientation .................................................. 15
CHAPTER 2. CURRENT STUDY ................................................................................. 16
Research Questions.................................................................................................... 18
Exploratory Questions ............................................................................................... 20
CHAPTER 3. METHOD .................................................................................................. 22
Pretest ............................................................................................................................ 22
Participants .................................................................................................................... 23
Procedure ....................................................................................................................... 25
Variables and Measures ................................................................................................ 26
Evaluation Questions ................................................................................................. 26
Manipulation Check .................................................................................................. 27
30-Item Domain-Specific Risk-Taking Scale (DOSPERT) ...................................... 27
Demographic Questions ............................................................................................ 28
Additional End-of-Study Questions .......................................................................... 28
CHAPTER 4. RESULTS .................................................................................................. 30
Exclusion Criteria .......................................................................................................... 30
Assumptions .................................................................................................................. 31
Plan for Analysis ........................................................................................................... 32
v
Dependent Variables Examination ................................................................................ 35
Evaluations of the Scenario, Police Involvement, and the Suspect ........................... 37
Hierarchical Linear Regression: Likelihood to Call Police........................................... 42
Research Question 1 .................................................................................................. 43
Research Question 2 .................................................................................................. 43
Research Question 3 .................................................................................................. 43
Examining Gender: Linear Regression ...................................................................... 45
Examining Political Orientation: Linear Regression ................................................. 47
Logistic Regression: Call Police ................................................................................... 49
Research Question 1 .................................................................................................. 50
Research Question 2 .................................................................................................. 50
Research Question 3 .................................................................................................. 51
Examining Gender: Logistic Regression (Call Police) .............................................. 53
Examining Political Orientation: Logistic Regression (Call Police) ......................... 56
Logistic Regression: Someone Else Call Police ........................................................... 59
Research Question 1 .................................................................................................. 60
Research Question 2 .................................................................................................. 60
Research Question 3 .................................................................................................. 60
Examining Gender: Logistic Regression (Someone Else Call Police) ...................... 62
Examining Political Orientation: Logistic Regression (Someone Else Call Police) . 67
Mediation Analysis ....................................................................................................... 72
CHAPTER 5. DISCUSSION ........................................................................................... 74
Summary of Results................................................................................................... 74
Race ........................................................................................................................... 74
Risk Perception .......................................................................................................... 77
Gender ....................................................................................................................... 77
Political Orientation ................................................................................................... 78
Limitations and Future Research ............................................................................... 80
Implications ............................................................................................................... 81
vi
Conclusion ................................................................................................................. 83
REFERENCES ................................................................................................................. 84
APPENDIX A: AMBIGUOUS RISK SCENARIO ......................................................... 96
APPENDIX B: INFORMED CONSENT ......................................................................... 97
APPENDIX C: DEBRIEFING FORM ............................................................................. 99
APPENDIX D: EVALUATION QUESTIONS ............................................................. 100
APPENDIX E: MANIPULATION CHECK .................................................................. 101
APPENDIX F: 30-ITEM DOSPERT SCALE ................................................................ 102
APPENDIX G: DEMOGRAPHICS ............................................................................... 104
APPENDIX H: ADDITIONAL END-OF-STUDY QUESTIONS ................................ 106
vii
LIST OF TABLES
TABLE PAGE
1 Demographic Characteristics .................................................................................24
2 N of Cases of Data Removed from Analyses ........................................................31
3 Correlations of the Dependent Variables ...............................................................37
4 Dichotomous Choice Item Response .....................................................................37
5 Correlations of the Evaluation Questions and the Dependent Variables ...............41
6 Correlations of the Variables by Condition ...........................................................42
7 The Likelihood to Call Police ................................................................................44
8 The Influence of Gender on the Likelihood to Call Police ....................................46
9 The Influence of Political Orientation on the Likelihood to Call Police ...............48
10 Correlation Matrix: The Choice to Call Police ......................................................50
11 The Choice to Call Police ......................................................................................52
12 Correlation Matrix: Gender and the Choice to Call Police ....................................54
13 The Influence of Gender on the Choice to Call Police ..........................................55
14 Correlation Matrix: Political Orientation and the Choice to Call Police ...............57
15 The Influence of Political Orientation on the Choice to Call Police .....................58
16 Correlation Matrix: Someone Else’s Decision to Call Police ................................59
17 Someone Else’s Decision to Call Police ................................................................61
18 Correlation Matrix: Gender and Someone Else’s Decision to Call Police ............63
19 The Influence of Gender on the Agreement with Someone Else’s Decision to Call
Police......................................................................................................................65
20 Correlation Matrix: Political Orientation and the Agreement with Someone Else’s
Decision to Call Police...........................................................................................68
21 The Influence of Political Orientation on the Agreement with Someone Else’s
Decision to Call Police...........................................................................................70
viii
LIST OF FIGURES
FIGURES PAGE
1 Interaction between Gender and Risk Perception on the Agreement with Someone
Else’s Decision to Call the Police ..........................................................................66
2 Interaction between PO and Risk Perception on the Agreement with Someone
Else’s Decision to Call the Police ..........................................................................71
3 Mediation Model ....................................................................................................73
1
CHAPTER 1.
INTRODUCTION
"Who ya gonna call?"(Parker, 1984) has been asked in popular media, but perhaps
the more important and often ignored question is, "When ya gonna call?” Each year,
about 240 million calls are made to 911 (“9-1-1 Statistics,” 2017). There are numerous
reasons to call 911, such as to report a violent crime1 (45%), a serious violent crime2
(51%), or a property crime 3 (36%; Morgan & Truman, 2018). Thus, the decision to call
the police4 is more complex than just dialing numbers. When calls are made to the police
and there is no actual crime (e.g., misuse of 911, non-emergency situations), this can take
away valuable resources from individuals who need police services (Sampson, 2002);
however, not calling the police when there is a crime could potentially endanger human
life (e.g., Felson et al., 2002). Thus, the decision to call the police has social and
emotional consequences (Greenberg, Wilson, Ruback, & Mills, 1979). Further, there are
many potential factors that may influence the decision to call the police.
One potential motivation to call the police could be related to race. In 2018, an
Oregon state representative was canvassing door-to-door in her district when the police
were called because of her “strange behavior” (i.e., using her cell phone after visiting
each house; Archie & Smith, 2018). In Ohio, the police confronted a 12-year-old
landscaping entrepreneur who mistakenly mowed a lawn next to a house of a client
1 Violent crime includes rape or sexual assault, robbery, aggravated assault, simple assault, and serious
violent crime; thus, serious violent crime is a subset of violent crime (Morgan & Truman, 2018). 2 Serious violent crimes includes rape or sexual assault, robbery, and aggravated assault (Morgan &
Truman, 2018). 3 Property crime includes household burglary, theft, and motor vehicle theft (Morgan & Truman, 2018). 4 Throughout this paper, the terms call 911, report crime, and call police are used interchangeably.
2
(Williams, 2018a). The police also confronted a California Samaritan who was helping a
local homeless man (Williams, 2018b). A student napping in a dorm common space
(Griggs, 2018), customers not ordering anything at a coffee shop and leaving (McCleary
& Vera, 2018), and friends simply leaving an Airbnb also had police called on them
(Criss & Vera, 2018). Additionally, the police confronted a man walking with a duffel
bag at the University of Massachusetts at Amherst (UMass; Jaschik, 2018), a woman
looking “out of place” at Smith College (Whitford, 2018), and a man at a local dog park
(Fieldstadt, 2019). In all these incidents, the alleged police suspects were black and
innocent: A black state representative, a black child, a black Samaritan, a black student,
two black potential coffee shop customers, three black friends, an employed black man at
UMass, a black student at Smith College, and a black man who took his dog to a dog
park. From all these examples, it appears that the presence of a racial minority is
motivation enough to call the police for some people. Thus, there appears to be potential
prejudice and discrimination toward racial minorities when calls are made to the police.
The current study will examine two possible theories that may help explain why
people call the police. People may call the police because of racial bias, or people may
call the police because of the perceived risk regarding a given situation. Aversive racism
theory (Gaertner & Dovidio, 1986) suggests that individuals generally support egalitarian
values; however, aversive racists discriminate when bias can be attributed to another
factor besides race (e.g., a woman looking “out of place;” Dovidio & Gaertner, 2000;
Whitford, 2018). Alternatively, the decision to call the police could have nothing to do
with discrimination and prejudice and could be motivated by other factors, such as risk
3
averse motivation. Risk averse motivation (Kahneman & Tversky, 1982) is based off
gambling research and is characterized as human preference in decision making
(Kahneman & Tversky, 1982; Zhang, Brennan, & Lo, 2014). A risk averse decision is the
tendency to prefer the certain option compared to the uncertain option (Paulsen, Platt,
Huettel, & Brannon, 2012). Thus, in an ambiguous or perceived suspicious situation,
calling the police, regardless of race, could provide a sense of certainty. In the next
sections, I will further discuss the decision to call the police, disparities regarding the
criminal justice system, aversive racism theory, and risk averse motivation.
Decision to Call the Police
There are many reasons to call the police. Although this paper focuses on calls
made to the police in the United States for possible crime situations, there are other
reasons someone may choose to call the police (e.g., non-crime related emergencies; to
seek help from police; participation in an anti-crime program). Further, police contact is
not always initiated by a civilian, but could be police-initiated contact. For instance, of
53.5 million individuals who experienced contact with the police in 2015, about 11% of
those contacts were police initiated (Davis, Whyde, & Langton, 2018).
Of the U.S. population, 16 and older, overall contact with the police has decreased
from 26 to 21% since 2011 (Davis et al., 2018). Although there has been an overall
decrease in police and citizen contact in the United States, it is important to understand
and examine the decision to call the police.
4
Within previous psychological literature, the decision to call the police has
focused on bystander intentions inspired by the Kitty Genovese5 case. The decision to
call the police creates complex feelings, which generate a sense of conflict from both
humanitarian and fear-based emotions in individuals (Darley & Latané, 1968). Despite
the complexity of helping behavior, Latané and Darley (1970) suggest that there are a
series of decisions an individual must make before deciding to intervene and act in an
emergency situation. First, the bystander must notice that a situation is happening. Once
the bystander is aware of the event, the bystander must interpret the event as an
emergency. If the bystander concludes that something is indeed wrong, the bystander
must decide their personal responsibility in the decision to act upon the situation. If the
bystander decides to act upon the situation (i.e., help), the bystander must decide what
form of help to give (e.g., call 911, provide direct assistance). Lastly, the bystander must
decide how to implement the plan of action (e.g., use a cell phone). When a bystander is
confronted with an emergency situation, depending on the series of choices made, the
bystander will intervene or not in an emergency situation (Latané & Darley, 1970).
There are other determinants that can influence how bystanders perceive these
decisions, such as social influence. Social impact theory suggests that decision to call the
police or not could be exacerbated by the influence of other individuals based on the
strength (i.e., status of others), immediacy (i.e., closeness of others), and the number of
people there are in a given situation (Latané, 1981). Research has shown that social
5 In 1964, Kitty Genovese was stabbed to death outside of her apartment building, in a New York
residential area. Although no one came to her aid directly, it was reported later that there were calls made to
the police (Kassin, 2017). It has been reported that 38 of her neighbors heard her murder from their
windows (Darley & Latané, 1968).
5
inhibition (the presence of others influencing behavior) exists in response to emergencies
(e.g., Latané & Nida, 1981; Latané & Darley, 1970). When a bystander is in the presence
of others, helping behavior decreases. For instance, there was a significant difference
between a bystander’s helping behaviors when they were alone (70%), with a passive
confederate (7%), or with two strangers (20%) in a helping situation regarding a hurt
woman (Latané & Rodin, 1969).
There are at least three reasons why the presence of others can inhibit helping
behavior: audience inhibition, social influence, and diffusion of responsibility (Latané &
Nida, 1981). First, audience inhibition can inhibit helping behavior. The possibility of
misinterpreting a situation in the presence of others could cause a bystander
embarrassment. Thus, the presence of others can inhibit a bystander’s decision to help to
avoid possible negative appraisals from others. Second, social influence can also inhibit
helping behavior. Bystanders use others to help define a helping situation, but this fosters
both correct and incorrect interpretations. For instance, in a helping situation, if a
bystander observes that others are not helping in a given situation, the bystander might
inaccurately interpret the situation as less critical than it really is. Lastly, diffusion of
responsibility can lessen the psychological burden of the cost of not helping in a given
situation. In the presence of others, the responsibilities and costs of helping behavior or
lack thereof can be shared with others (Latané & Darley, 1970; Latané & Nida, 1981).
There are also other determinants that affect both bystanders’ and victims’
decision to call the police. Other determinants that influence the decision to call the
police could include the seriousness of the crime (e.g., injury or financial loss; Skogan,
6
1984), considerations of obligation (e.g., maintain status of being a “good citizen”
Skogan, 1984, p. 121), and culpability (e.g., own past behavior with police; Sparks,
Genn, & Dodd, 1977). There is also evidence that suggests there are racial differences in
the decision to call the police (e.g., Skogan, 1984; Davis et al., 2018).
Racial Differences in Calling the Police
Although there is a lack of psychological literature examining racially based
motivation to call the police, there is research concerning racial differences in reporting
crime to the police. In the 1980s, it was believed there were minimal racial differences in
reporting crime to the police (Skogan, 1984). For instance, in some categories (e.g.,
crimes that involved personal injury), black individuals reported crimes at higher rates
than white individuals (Skogan & Maxfield, 1981). More recently, however, evidence
suggests that white individuals contact police more than black individuals.
Overall, the U.S population consists of over 200 million individuals age 16 and
older, and about 21.1% of these individuals have had contact with the police (about 53.5
million individuals) in the past 12 months. Of this 21.1%, about 11% of individuals
(about 27 million individuals) initiated police contact, and about 7% initiated contact
with police to report a possible crime (about 16 million people; Davis et al., 2018).
Of the 16 million reports to the police regarding a possible crime, about 12
million (75%) were from white individuals, whereas about two million were from black
individuals (11%; Davis et al., 2018). Based on population estimates from the most recent
U.S. Census (2010), relative to the white U.S. population size (73%) in the United States,
white individuals are overreporting to the police for a possible crime, whereas relative to
7
the black U.S. population size (13%) black individuals are under reporting to the police
for a possible crime (U.S. Census Bureau, 2017). The lower rates of reporting by
minority populations could be connected to evidence that suggests that black individuals
are more distrustful of police (Tyler, 2005) and have lower levels of confidence in the
police (Lasley, 1994). Further, there is evidence that suggests that the police have greater
suspicion of minority populations (e.g., Alpert, MacDonald, & Dunham, 2005; Harris,
2002). These factors could influence the racial differences in the decision to contact the
police.
Racial Differences in Arrest Rates
From the 1970s to the 2000s, arrest rates of black individuals have declined
(Travis, Western, & Redburn, 2014); however, changes in society (Blumstein &
Wallman, 2006) and trends in policies (Blumstein & Beck, 1999) can affect arrest rates.
For instance, in the 1980s, there was a large increase in the arrest of black individuals
compared to white individuals, due to black individuals predominately selling the drug,
crack (Blumstein & Wallman, 2006). Further, in the late 1980s, overall drug related arrest
rates were six times higher for black individuals compared to white individuals
(Blumstein & Wallman, 2006). In more recent years, however, drug related arrests of
black individuals were only three to four times higher than for white individuals (Travis
et al., 2014). The decrease of black drug arrests compared to white drug arrests is due to
the emphasis on total drug arrests and greater emphasis on marijuana arrests (Travis et
al., 2014).
8
More generally, white individuals (70%) are arrested at higher rates than black
individuals (27%; FBI Uniform Crime Reporting Program, 2016); however, black arrest
rates are double their actual representation in the U.S. population (Garland, Spohn, &
Wodahl, 2008). Despite experiencing lower arrest rates, black individuals still experience
racial bias. For example, black individuals experience higher rates of traffic stops
compared to white individuals (Harris, 1999; Langton & Durose, 2016; Walker, Spohn,
& DeLone, 2000). Ultimately, these stops result in more searches, tickets, and arrests for
black individuals compared to white individuals (Langton & Durose, 2016; Pierson et al.,
2019). Further, evidence suggests that racial discrimination during the early stages of the
criminal justice process (i.e., arrest, pretrial process, and sentencing) is concerning for
incarceration rates (Weich & Angulo, 2002). If there was no discrimination within the
criminal justice system following initial arrests then black and white individuals would be
imprisoned at the exact same distributions; however, there is evidence of post-arrest
discrimination in incarceration rates (Garland et al., 2008).
Racial Differences in Incarceration Rates
In the United States, there appears to be a disparity of incarceration rates between
black and white individuals (Travis et al., 2014). Historically, U.S. prisons have
contained primarily disadvantaged populations. In the 1970s, about one-third of white
males who dropped out of high school had served time in prison, whereas about two-
thirds of black males who had dropped out of high school had a prison record (Travis et
al., 2014). More recently, from 1995 to 2005, incarceration rates have decreased overall,
9
but there are still more black individuals (40%) in state or federal prisons than white
individuals (35%; Harrison & Beck, 2006).
Further, evidence from 2014 suggests that black individuals are imprisoned at six
times the rate of non-Hispanic white individuals (Travis et al., 2014). As of 2019, black
individuals still comprise about 40% of the U.S. prison systems despite accounting for
only 13% of the U.S population (Sawyer & Wagner, 2019). Ultimately, there is an
overrepresentation of racial and ethnic minorities in prison systems in relation to their
overall population (Nellis, 2016). Collectively this information suggests that race should
be evaluated in understanding calls made to the police as well as arrest and incarceration
rates.
Police officers and the criminal justice system are tools enacting the biases held
by citizens who call the police (Takei, 2018). Although people believe they call police for
justified reasons, this might not be the case. One possible explanation for why people
choose to call the police stems from racial motivation to call police using aversive racism
theory (Gaertner & Dovidio, 1986). Another possibility could be related to individual
variability in risk perception using risk averse motivation (Kahneman & Tversky, 1982).
In the next sections, I will examine these two possibilities.
Aversive Racism Theory
Aversive racism theory (Gaertner & Dovidio, 1986) is consistent with the
American Dilemma (Myrdal, 1944), which suggests that Americans are inconsistent in
their beliefs of equality for all, and their ability to treat black individuals as equal to white
individuals (Dovidio, Mann, & Gaertner, 1989; Loury, 1984). Although America is
10
founded on ideals and values of equality, there seems to be a contradiction of this belief
that is demonstrated by the nation’s history of overt signs of prejudice and discrimination
(Dovidio & Gaertner, 2004; Loury, 1984).
Although prejudice might not be as overt and explicit as it previously has been in
America’s history, prejudice has persisted and become more subtle and unconscious
(Dovidio & Gaertner, 2000). Aversive racism theory (Gaertner & Dovidio, 1986)
explores these changes from overt, explicit forms of prejudice and discrimination to the
increase in unconscious, subtle forms of prejudice and discrimination (Dovidio &
Gaertner, 2000). Even though subtle forms of prejudice might not be as explicitly
detrimental as overt forms of prejudice, they have even stronger and more damaging
repercussions (e.g., “the restriction of economic opportunity;” Gaertner, Dovidio, Nier,
Hodson, & Houlette, 2008, p. 378).
Aversive racism theory (Gaertner & Dovidio, 1986) suggests that aversive racists
generally support equalitarian values; however, aversive racists possess biased feelings,
often unconsciously. Thus, to avoid contradiction, aversive racists are likely to rationalize
their racist attitudes and beliefs to another factor besides race (Dovidio & Gaertner,
2000). For instance, white individuals did not discriminate in their recommendations of
hiring decisions when white or black candidates had clearly strong or weak candidate
qualifications. When qualifications were more ambiguous, however, white individuals
did discriminate in their hiring decisions of black candidates (Dovidio & Gaertner, 2000).
Thus, the ambiguity of the qualifications allowed participants to rationalize their biased
recommendation against a black candidate.
11
Aversive racism creates complex attitudes and feelings in individuals who hold
egalitarian intent but display and implicitly harbor racist feelings (Dovidio & Gaertner,
2000; Gaertner & Dovidio, 2005). Ultimately, this complexity creates an atmosphere of
ambivalence driven by both feelings of equality and implicit negative feelings about
black individuals (Dovidio & Gaertner, 2000; Pearson, Dovidio, & Gaertner, 2009).
Rather than expressing explicit hatred or aggression, aversive racists often feel
uneasiness, physiological arousal, discomfort, or even fear towards black individuals
(Dovidio & Gaertner, 2004; Nail, Harton, & Decker, 2003).
Following the line of aversive racism research, often individuals hold non-
prejudiced attitudes; however, without the decision to call the police clearly defined, the
decision to call the police could be rationalized by other qualities of the situation. For
example, if the situation is late at night, people are yelling, or fists are being pounded on
a table then these factors may be rationalized as the decision to call the police rather than
the race of the perceived suspect. It is possible, however, that race (i.e., black or white)
could be an unconscious factor in the decision to call the police. Thus, individuals may
discriminate in their choice to call the police.
Risk Averse Motivation
Alternatively, motivation to call the police could be related to individual variation
in levels of risk aversion. The concept of risk aversion can be traced back to Daniel
Bernoulli, a Mathematician in 1738 (Kahneman & Tversky, 1982). Based off gambling
research, Bernoulli characterized risk aversion as a characteristic of human preference
(Kahneman & Tversky, 1982; Zhang et al., 2014). Since Bernoulli, risk aversion has been
12
extensively researched in both economic and psychological literature (Zaleskiewicz,
2001). Studies often examine preferential outcomes that motivate either risk averse or
risk seeking behavior. To further understand both of these concepts:
To understand risk aversion, imagine that you are given a choice between two
options. The first is a sure gain of $80. The second is a risky prospect that offers
an 85 percent chance of winning $100 and a 15 percent chance of winning
nothing. Most people who are presented with this choice prefer the certain gain to
the gamble, in spite of the fact that the gamble has a higher "monetary
expectation" than the certain outcome…A choice is risk averse if a certain
outcome is preferred to a gamble with an equal or greater monetary expectation.
A choice is risk seeking, on the other hand, if a certain outcome is rejected in
favor of a gamble with an equal or lower monetary expectation. (Kahneman &
Tversky, 1982, p. 160)
More simply put, a risk averse decision is the tendency to prefer the certain option
compared to the uncertain option (Paulsen et al., 2012); risk seeking behavior is the
tendency to prefer the uncertain option (Kahneman & Tversky, 1982).
There are three individual characteristics that have been identified as likely
determinants of risk behavior (i.e., risk averse or risk seeking behavior): (1) risk
preferences, (2) risk perceptions, and (3) risk propensity. The first determinant, risk
preferences, refers to an individual’s overall attitude toward risk (Wen, He, & Chen,
2014). Some individuals enjoy the thrill and challenge that risk can entail and others do
13
not (Sitkin & Pablo, 1992). For example, consistently across several countries, investors’
risk preferences are risk seeking (Wen et al., 2014).
The second determinant, risk perception, is created through an individual
assessment of a situation, estimates of the extent and controllability of risks, and the
confidence in those estimates (Sitkin & Pablo, 1992). Risk perception is a personal
process and decision about health and safety (Ropeik, 2012), which incorporates thoughts
of “the probability of something bad happening” (Brown, 2014, p. A277). After an initial
assessment of a situation, individuals can assess the amount of risk present and decide to
take action based on their own level of discretion of perceived risk (Sitkin & Pablo,
1992). Often, perceived risk can result in heightened senses (i.e., sights, sounds, smells)
and can affect memories, which can then exacerbate perceived risk and fear (Brown,
2014). For instance, hearing a higher probability of the chance of a car crash fatality
increases risk perception and risk averse behavior (i.e., frequent seat belt use).
The third determinant, risk propensity, is characterized as risk taking tendencies.
Based on the risk taking tendencies of an individual, risks are evaluated in terms of taking
or avoiding risks (Sitkin & Pablo, 1992). For instance, researchers found that business
executives who were more mature (i.e., older, seniority) were consistently more risk
averse compared to individuals who were less mature (i.e., younger; MacCrimmon &
Wehrung, 1990).
Multiple factors can influence individual variability in the perception of risk and
ultimately, risk behavior, aside from the three individual characteristics: risk preferences,
risk perception, and risk propensity. Another factor that could influence risk behavior is
14
group contexts (Janis, 1972). While in a group, individuals can become victims to the
phenomenon, groupthink (Janis, 1982). Groupthink suggests that group cohesion
becomes so strong that thoughts and values become likeminded and personal doubts are
dismissed within the group (Janis, 1982). For example, there were signs and antecedents
of groupthink surrounding the decision to launch the space shuttle Challenger that
exploded seconds after its launch (Esser & Lindoerfer, 1989). Further, emotions can also
influence risk behavior (Kusev et al., 2017). For instance, some research suggests that
anxious and depressive states are connected to risk averse preferences (Miu, Heilman, &
Houser, 2008; Yuen & Lee, 2003). Hormones, such as cortisol or testosterone can also
influence risk behavior (Kusev et al., 2017). Individuals with higher levels of testosterone
tend to take greater risks (Stanton, Liening, & Schultheiss, 2011) whereas individuals
who experience chronically raised levels of cortisol tend to be more risk averse (Kusev et
al., 2017).
Collectively, the preference for risk aversion can depend on a number of factors.
Risk averse motivation could stem from influences of individual characteristics (e.g., risk
preference, risk perception, and risk propensity). Thus, some individuals are more risk
averse, and may be more willing to call the police in a possible risk situation. Further,
risk perceptions could be influenced by evidence that suggests white individuals (and
black individuals) hold negative stereotypes about the black community (Plous &
Williams, 1995), and that individuals link black men with criminality (Kleider-Offutt,
Bond, & Hegerty, 2017). These negative stereotypes can influence the perception of risk
15
in a given situation. If a risk averse individual holds a negative stereotype of a black
individual, then this could perpetuate the need to call the police in a crime context.
Other Explanations: Gender and Political Orientation
Demographic factors may also influence the decision to call the police. According
to the most recent Police-Public Contact Survey, women initiate contact with police and
report possible crime more often than men (Davis et al., 2018). Further, women express
more confidence in the police than men. Women often also hold more favorable
sentiments towards police than men because their relationship with police is often less
antagonistic, and women are more likely to make service requests to police compared to
men (Cao, Frank, & Cullen, 1996). Therefore, women may be more likely to call the
police more than men.
Historically, America has been politically divided in its support for the police
(Ekins, 2016). For instance, political conservatives tend to hold more positive attitudes
towards police than political liberals (e.g., Hindelang, 1974; Huang & Vaughn, 1996;
Zamble & Annesley, 1987). More recently, only 33% of liberals rated the police as “very
warm,” whereas 74% of conservatives rated the police as “very warm” (Fingerhut, 2017).
These differences could stem from the differentiation in values supported by liberals and
conservatives; specifically, conservatives place greater emphasis on authority, which is
represented by the police (Graham, Haidt, & Nosek, 2009). Thus, political orientation
may affect the decision to call the police.
16
CHAPTER 2.
CURRENT STUDY
There seems to be a surplus of examples of racially biased calls made to the
police (Takei, 2018). These biased calls share two common themes: (1) individuals seem
to be behaving in seemingly harmless activities (e.g., napping in a dorm common area;
Griggs, 2018), and (2) the perceived suspect is black. The current study aimed to further
understand the motivation to call the police through the lens of aversive racism theory
(Gaertner & Dovidio, 1986) and risk averse motivation (Kahneman & Tversky, 1982).
Aversive racism theory (Gaertner & Dovidio, 1986) suggests that aversive racists
discriminate when bias can be rationalized to another factor besides race. Aversive racists
generally endorse egalitarian values, which makes understanding and recognizing
aversive racism difficult (Dovidio & Gaertner, 2000). The aversive racism theory
framework could be applied to understand the motivation to call the police. For instance,
regarding the call made to police on the canvassing state representative, the caller
indicated they placed the call because of the suspicious behavior of the politician. Thus,
the caller rationalized the odd behavior of the black state representative— using her cell
phone after visiting each house— as justification to call the police (Archie & Smith,
2018).
Individuals may also be influenced by risk averse motivation. Risk averse
motivation (Kahneman & Tversky, 1982) is the tendency to prefer the certain option
compared to the uncertain option (Paulsen et al., 2012). Risk aversion can be influenced
by various determinants of an individual’s character (i.e., risk perception; Sitkin & Pablo,
17
1992). Examining risk averse motivation could be used to understand the motivation to
call the police. For instance, regardless of race, for a risk averse individual, a call to the
police in a risky situation would lead to a certain outcome of the police coming, which
could provide a sense of relief and safety.
Further, both race of a perceived suspect and risk aversion may influence the
decision to call the police. Evidence suggests that Americans still hold stereotypical,
negative views of black individuals, such as linking them to crime (Kleider-Offutt et al.,
2017). Thus, an assessment of a black individual in a possible crime situation could
influence a risk averse individual to call the police to provide a sense of certainty in a
perceived risky situation.
Additionally, demographic factors may influence the decision to call the police.
Women often have more favorable views of police (Cao et al., 1996) and initiate contact
with police more than men (Davis et al., 2018). Further, conservatives compared to
liberals also have more favorable views of the police (Hindelang, 1974; Huang &
Vaughn, 1996; Zamble & Annesley, 1987) and more confidence in the police (Stack &
Cao, 1998). Thus, women and conservatives may be more likely to call the police than
men and liberals.
The current study used an exploratory approach to examine the motivation to call
police through both aversive racism theory and risk averse motivation. To test these two
frameworks, participants were recruited from Amazon Mechanical Turk (mTurk) and
were asked to provide their evaluations and attitudes about a mock crime summary.
Participants were randomly assigned to one of two conditions that contained either a
18
black or white perceived suspect (i.e., the customer in the scenario). In each condition,
participants read a scenario about an altercation (i.e., yelling and fists pounding on the
counter) between a store clerk and a store customer at a convenience store. After reading
the scenario, participants were asked to provide their attitudes and perceptions of both the
scenario (e.g., “To what extent, if any, do you think the scenario was risky?”) and the
perceived suspect (e.g., “To what extent, if any, was the customer at the front of the store
dangerous?”), and then participants completed the manipulation check. Some of these
items were used as the dependent variables within analyses. Then participants completed
a risk perception scale (e.g., How risky is “Walking home alone at night in an unsafe area
of town.”). This scale was used to assess whether participants were risk averse or risk
seeking. After the risk perception scale, participants completed demographic measures.
Two of the demographic questions relating to gender and political orientation (assessed
three domains) were used in analyses: “What is your gender identity?” and “How would
you describe your views on current social issues, economic issues, and foreign policy?”
Lastly, participants completed an honesty check and open-ended items regarding the
intent of the study, possible reasons to exclude data, and a comment section for any
possible comments the participant had for the researcher.
Research Questions
Research questions were used rather than hypotheses based on the lack of
research in this area to form hypotheses. Consistently throughout the study, race of the
perceived suspect was operationalized as the race of the customer in the scenario (i.e.,
black or white); risk perception was operationalized as a score on the 30-Item Doman-
19
Specific Risk-Taking Scale (DOSPERT scale; Blais & Weber, 2006a); the decision to
call the police was operationalized as self-reports of both the likelihood, dichotomous
choice of the decision to call the police, and the agreement with someone else’s decision
to call the police. For my exploratory analyses, gender was operationalized with the
single item, “What is your gender identity?” and political orientation was operationalized
as a mean value of an overall score assessing three domain areas: “social issues,”
“economic issues,” and “foreign policy” (adapted from Inbar & Lammers, 2012)
1. Are participants more likely to call the police when the perceived suspect is black
compared to white? According to aversive racism theory (Gaertner & Dovidio,
1986) when bias can be rationalized to another factor besides race, then
discrimination can happen. When the perceived suspect in the scenario is black,
the scenario (e.g., yelling, fists pounding on the counter) may cause individuals to
rationalize that the decision to call the police is important and necessary.
2. Are participants who are more risk averse more likely to call the police? Based on
previous research suggesting that risk averse individuals prefer the certain option
compared to the uncertain option (Kahneman & Tversky, 1982; Paulsen et al.,
2012) then risk averse individuals would prefer the certain outcome of police help
in the choice to call the police.
3. Are participants who are risk averse more likely to call the police when the
perceived suspect is black compared to white? Evidence suggests that risk averse
individuals have various characteristics (e.g., risk perception) that are influential
in their risk behavior (Sitkin & Pablo, 1992). Further, the race of the perceived
20
suspect could be influential in the assessment of perceived risk. For instance,
when the perceived suspect in the scenario is black, the assessment of risk
perception might influence the decision to call the police due to stereotypes that
black individuals are linked to criminality (Kleider-Offutt et al., 2017). Thus,
calling the police will provide both certainty and security in an uncertain situation
for risk averse individuals.
Exploratory Questions
4. Does risk perception mediate the relationship between race of the perceived
suspect and choice to call the police? The individual assessment of risk perception
could exacerbate the influence of the race of the perceived suspect on the decision
to call the police.
5. Are there gender differences in the choice to call the police? There is evidence
that suggests women have more favorable views and higher confidence in the
police than men (Cao et al., 1996). Thus, women might be more likely to call the
police than men.
a. Is there an interaction between gender and risk perception in the choice to
call the police? Evidence suggests that women are more risk averse than
men in many areas of life (e.g., Jianakoplos & Bernasek, 1998; Spigner,
Hawkins, & Loren, 1993); thus, women who are risk averse might be
especially likely to report that they would call the police.
6. Are there political orientation differences in the choice to call the police?
Evidence suggests that conservatives have more positive feelings toward the
21
police and additionally, have a value system that better aligns with police offers
roles (i.e., authority figures) compared to liberals (Hindelang, 1974; Huang &
Vaughn, 1996; Zamble & Annesley, 1987; Graham et al., 2009). Thus,
conservatives might be more likely to call the police than liberals.
a. Is there an interaction between political orientation and risk perception in
the choice to call the police? Existential motives of threat and fear are
most often linked to political conservatism (Jost, Glaser, Kruglanski, &
Sulloway, 2003); thus, political conservatives who are risk averse might
be especially likely to report that they would call the police.
22
CHAPTER 3.
METHOD
Pretest
I initially generated four scenarios to test for use in the study. In all the scenarios,
the reader imagined they left work and stopped to get gas at a convenience store when
they witnessed an altercation between the convenience store clerk and another
convenience store customer (who was either black or white). Details, however, were
changed for each scenario based on the concept of perceived risk; thus, the scenarios
were created to represent a possible range of perceived risk. Specifically, the details
surrounding the altercation between the store clerk and the other convenience store
customer were altered in each scenario. In one scenario aspects of a robbery unfold (i.e.,
yelling, demanding of money) and a gunshot occurs, in another scenario only the aspects
of the robbery unfold (i.e., yelling, demanding of money), whereas in another scenario
the altercation involves only incoherent yelling and the customer pounding his fists on
the counter. In the last scenario, the reader takes notice of a conversation between a store
clerk and convenience store customer.
To determine which scenario represented an account of ambiguous risk, 56
undergraduate students at a Midwestern University rated each of the four scenarios using
the two items: “To what extent, if any, do you believe this situation is risky?” using a 10-
point Likert scale from, 1 (Not at all Risky) to 10 (Extremely Risky) and “To what extent,
if any, do you believe you need to call the police?” using a 10-point Likert scale from, 1
(Strongly Disagree) to 10 (Strongly agree). Ambiguity was determined by a mean value
23
of the two items (i.e., risk question and need for police question) ranging around 5.0-7.0.
These qualifications were similar to Dovidio and Gaertner’s (2000) criteria for
determining ambiguous qualifications for potential job candidates. Furthermore, an
ambiguous scenario made the appropriate decision to call or not call the police more
difficult, which aids in the exploration of discrimination. The scenario deemed
ambiguous and used in the current study (Appendix A) involved aspects of yelling
between the store clerk and customer and the customer’s fists pounding on the counter.
Participants
I conducted a power analysis to determine the number of participants needed for
this study. Due to the lack of psychological literature on the current topic, a similar topic
using aversive racism theory was used for the power analysis. The effect sizes (Pearson’s
r) for this study were based on correlations of a racial stereotyping item and two
measures of the approval for use of police force (i.e., regarding excessive police force
and reasonable police force; r = .12) and a racial stereotyping scale and antipathy scale (r
= .35; Barkan & Cohn, 1998). These two correlations were averaged to reach a small
effect size. To obtain .90 power with a two-tailed test, and small effect size of .20, I
needed a total of 259 participants using Cohen’s tables (Cohen, 1988). However, due to
possible data sample issues with mTurk data (e.g., attrition, inattention; Buhrmester,
Talaifar, & Gosling, 2018), I decided to oversample by 20 percent to account for
participants that would later need to be excluded. With a 20 percent increase, I needed a
sample size of 311 participants to examine my research questions. Increasing the sample
size by 20% assured that my original target sample size of 259 would be met. Participants
24
were all recruited from an online data collection platform, Amazon Mechanical Turk
(mTurk users).
Participants were recruited using the TurkPrime website and were compensated
$1.00 for their participation. Participants had to have a HIT approval rating of 95% or
better and have completed between 5,000 and 50,000 HITS approved to participate.
There were initially 332 participants recruited from mTurk, and after the deletions based
on the exclusion criteria there were 295 mTurk participants.
Table 1
Demographic Characteristics
Demographics Mechanical Turk
participants
Gender Identity Male 56%
Age Age 38.87 (11.59)
Race/Ethnicity White or Caucasian 74%
Black or African American 11%
Asian or Asian American 8%
Hispanic or Latinx 7%
Multiracial 1%
Education Bachelor’s degree 43%
Associate degree 21%
High School diploma or less 21%
Graduate degree 11%
(table continues)
25
Political Party
Identification
Democrat 46%
Independent 27%
Republican 21%
No affiliation 5%
Other affiliation (e.g., Libertarian) 1%
Political Orientation Very Liberal 18%
Liberal 33%
Moderate 28%
Conservative 17%
Very Conservative 4%
Note. n = 290-295. The percentages provided are based on data of participants who
met all inclusion criteria. Parenthesis indicate standard deviation.
Procedure
Participants read an electronic version of the informed consent before proceeding
to the study (Appendix B). Participants were randomly assigned to read an ambiguous
risk scenario that contained either a black or a white customer (i.e., perceived suspect),
and were required to stay on this page for 30 seconds to ensure they read the scenario.
The scenario placed the reader at a convenience store at night when an altercation occurs
between a store clerk and a store customer (i.e., perceived suspect). The altercation
involved yelling and the customer pounding his fists on the counter of the convenience
store. During the altercation between the store clerk and customer, the reader learns the
race of the customer as either a young black or white male. Following the scenario,
participants completed questions regarding their likelihood to call the police (i.e., Likert
26
scale), their choice to call the police (i.e., yes or no), and their agreement with someone
else’s decision to call the police (i.e., yes or no); perceptions of the scenario, police
involvement, and the customer in the scenario (i.e., perceived suspect); and the
manipulation check in a randomized order.
Then participants completed a risk perception scale and demographic measures.
Lastly, participants completed an honesty check and open-ended response items that
asked about the purpose of the study, reasons to exclude data, and participant comments
for the researcher. Before concluding the study, participants read the debriefing form
(Appendix C).
Variables and Measures
Evaluation Questions
The evaluation questions (author-generated; Appendix D) assessed the
perceptions and attitudes associated with the scenario and customer through seven items.
All seven items were answered by participants in randomized order. Three of the items
were used to examine the evaluations of the police, (i.e., “To what extent, if any, do you
worry that calling the police would bring you trouble?”), the customer in the scenario,
(i.e., “To what extent, if any, was the customer in the front of the store dangerous?”) and
the scenario overall, (i.e., “To what extent, if any, do you think the scenario is risky?”)
using a 10-point Likert scale from 1 (Strongly disagree) to 10 (Strongly Agree). Four
items of the evaluation questions were used as the dependent variables:
“Based on this scenario, what is the likelihood that you would call the
police?” (Likelihood question)
27
“If someone else saw the same scenario as you and they decided to call
the police, what is the likelihood that calling the police is the correct
decision?” (Likelihood question)
“Based on the scenario you previously read, would you call the police?”
(Dichotomous choice question)
“If someone else saw the same scenario as you and they decided to call
the police, do you agree or disagree with that choice?” (Dichotomous
choice question)
These items were assessed with either a 10-point Likert scale, 1 (Not at all likely)
to 10 (Extremely likely) or a dichotomous choice, “Yes” or “No.” Both likelihood
questions were combined to create an overall likelihood score, whereas the dichotomous
choice questions were examined separately within analyses. The two likelihood questions
had a reliability of ⍺ = .84, whereas the dichotomous choice questions had a reliability of
⍺ = .68.
Manipulation Check
The manipulation check included a single multiple-choice question asking the
race of the customer at the front of the store in the scenario (Appendix E).
30-Item Domain-Specific Risk-Taking Scale (DOSPERT)
The 30-item DOSPERT (Blais & Weber, 2006a; Appendix F) contains five
subscales (i.e., ethical, financial, health/safety, recreational, and social) that assess risk
perception. Using a gut level assessment, participants were asked to evaluate how risky
the action or behavior was in each item using a 7-point Likert rating scale, ranging from 1
28
(Not at all risky) to 7 (Extremely risky). Example items include: “Disagreeing with an
authority figure on a major issue” (social); “Bungee jumping off a tall bridge”
(recreational); “Drinking heavily at a social function” (health/safety); “Not returning a
wallet you found that contains $200” (ethical); "Investing 10% of your annual income in
a new business venture” (financial). Blais and Weber (2006a, 2006b) found good
construct validity and reliability of the subscales ranging from .74 to .83 (mean ⍺ = .79).
Subscale reliability in this study ranged from ⍺ = .78 to .83 (mean ⍺ = .81). In the
current study, I used the entire 30-item scale in my analyses, and the overall reliability
was ⍺ = .92. The overall 30-item DOSPERT scale score can be used as a comprehensive
form of assessment (Blais & Weber, 2006b).
Demographic Questions
The demographics questions (Appendix G) collected basic demographic
information (e.g., gender, race/ethnicity, political orientation) about each participant. For
example, “What is your race/ethnicity? Check all that apply.” The demographic items
regarding gender and political orientation were used in analyses. Gender was collected
through obtaining gender identity: “What is your gender identity?” and political
orientation was obtained from assessing three domains areas: “social issues,” economic
issues, and “foreign policy” from 1 (Very conservative) to 5 (Very liberal; adapted from
Inbar & Lammers, 2012).
Additional End-of-Study Questions
The additional end-of-study questions examined participant honesty (i.e., “How
honest were your answers throughout the study? You will receive payment regardless of
29
what you answer.”) and open-ended questions that requested participant input regarding
the study (e.g., “What do you think the current study was about?”; Appendix H).
30
CHAPTER 4.
RESULTS
Exclusion Criteria
Features were enabled within TurkPrime that aided in the exclusion process.
TurkPrime blocked mTurkers on the universal exclude workers list and blocked
suspicious geo-locations.
Data were excluded from analyses if the study was discontinued prematurely (less
than 25% of the study completed), if participants were not U.S. citizens, if the total time
to complete the study was below 90 seconds, if participants inaccurately answered the
manipulation check, if participants reported they were “not at all honest” or only “slightly
honest” during the study, and if participants failed the attention check. As an attention
check, participants reported both their age and year born. Data were removed from
further analyses if these two items (i.e., age and year born) were not within two years of
each other. Additionally, data were excluded from analyses if participants objected to
their data being used in analyses. If there were duplicate IP addresses, then the data from
the second sample was removed. Using Mahalanobis distance, univariate outliers
exceeding a z-score of ± 3.29 were removed, as defined by Mertler and Vannatta (2013).
Data from multivariate outliers were determined and removed if the Mahalanobis
distance values exceed the chi-square value of 20.515 (Table 2).
31
Table 2
N of Cases of Data Removed from Analyses
Exclusion Criteria Mechanical
Turk
participants
Discontinuing the study 15
Not a U.S. citizen 4
Timing 0
Manipulation check 11
Honesty check 4
Attention check 1
Objection to use of data 0
Duplicate IP address 0
Univariate outliers 0
Multivariate outliers 2
Total Remaining N 295
Assumptions
Before conducting analyses, frequencies, distributions, and ranges were evaluated
for each measure and variable. Violations of normality were assessed through substantial
skewness and kurtosis values. For all measures and variables, skewness and kurtosis were
near -1 and +1 (Mertler & Vannatta, 2013); thus, no variables required transformation
(Kline, 1998). Normality was confirmed by the Kolmogorov-Smirnov test statistic.
Missing values were identified and only viably treated for the 30-Item DOSPERT scale.
The eight missing data values for the 30-Item DOSPERT scale were handled the same as
32
Blais and Weber’s (2006a) standards. Therefore, sample mean values replaced the eight
missing data points based on the criteria that none of the missing values came from the
same participant. Homogeneity of variances was confirmed with Levene’s Test, and
homoscedasticity was confirmed from bivariate scatterplots and statistically using Box’s
M. Scatterplot matrixes indicated there was acceptable linearity based on the appearance
of elliptical shapes (Mertler & Vannatta, 2013).
Plan for Analysis
Before conducting analyses, I examined the relationships between the dependent
variables using bivariate correlations and frequencies. These analyses indicated which of
the items could be combined for analyses and which of the items could not be combined
for analyses. Then, in relation to the dependent variables, I examined three scenario
evaluation items (i.e., evaluations of the scenario, involvement of the police, and the
perceived dangerousness of the customer) using bivariate correlations and independent
samples t-tests. These items were examined to provide a preliminary understanding of the
dependent variables. Further, I examined the correlations among the variables by
condition (i.e., black or white perceived suspect). This provided information regarding
the relationships among my variables.
To examine my research questions, the predictors of risk perception and political
orientation were mean centered. The mean of each variable was subtracted from each
variable value and the predictors of gender and race of the customer were dummy coded
(Irwin & McClelland, 2001). First, I conducted one hierarchical linear regression using
the total likelihood to call police score as the dependent variable. The predictor variables
33
of age and gender were entered in the model first as control variables. The demographic
factors of age and gender were added to the model because evidence suggests they could
be influential in regards to the decision to call the police. Older individuals often have
increased contact with the police (Sever & Youdin, 2006), and women have more
positive (Cao et al., 1996) and frequent contact with police than men (Davis et al., 2018).
The variable gender was dummy coded (i.e., men = 1 and women = 0), and data from
participants who did not identity in these two categories were removed. Then, risk
perception and race of the customer in the scenario were added to the model. Race of the
customer was dummy coded (i.e., black = 1 and white = 0). Next, the interaction of risk
perception and race of the customer in the scenario was added to the regression model.
The interaction term was calculated by multiplying the risk perception scale and the race
of the customer in the scenario (Iacobucci, Schneider, Popovich, & Bakamitsos, 2016).
Next, to examine my exploratory questions and to avoid repercussions to power, I
ran a second linear regression including the addition of the interaction term of risk
perception and gender. The interaction term was calculated by multiplying the risk
perception scale and gender (Iacobucci et al., 2016). Only two blocks were examined for
this model. Age, gender, race of customer, and risk perception are in the first block, and
the second block contained the two interaction variables.
Then, to examine my next exploratory questions and to avoid repercussions to
power, I ran a third linear regression including the variable political orientation and the
interaction term of risk perception and political orientation. The interaction term was
calculated by multiplying the risk perception score and political orienatation (Iacobucci et
34
al., 2016). Only two blocks were examined for this model. Age, gender, race of customer,
risk perception, and political orientation are in the first block, and the second block
contained the three interaction variables.
I also conducted two hierarchical logistic regressions to assess my research
questions, using the two dichotomous choice dependent variables: the dichotomous
choice to call police item and the dichotomous choice agreement with someone else’s
decision to call the police item. These analyses are presented in sequential order to
prevent redundancy since the analyses were identical besides the dependent variable. For
these models, the variables were handled the same as in the linear regression analyses
(i.e., mean centering, dummy coded, calculation of interaction terms), but there were
different dependent variables. First, for the logistic regression model, age and gender
were added in the model first. Then risk perception and race of the customer in the
scenario were added to the models. Third, the interaction term of risk perception and the
race of the customer in the scenario were added to the models.
Next, to examine my exploratory questions and to avoid repercussions to power, I
ran a two more logistic regressions, which included the addition of the interaction term of
risk perception and gender. Only two blocks were examined for this model. Age, gender,
race of customer, and risk perception are in the first block, and the second block
contained the two interaction variables.
Lastly, to examine my next exploratory questions and to avoid repercussions to
power, I ran two more logistic regressions, including the variable political orientation and
the interaction term of risk perception and political orientation. Only two blocks were
35
examined for this model. Age, gender, race of customer, risk perception, and political
orientation are in the first block, and the second block will contain the three interaction
variables. Overall, nine regressions were conducted to aid in the examination of the
research questions. Separating the analyses assisted in interpretation and dissemination of
the results regarding the research questions.
Additionally, I conducted one exploratory mediational analysis to determine
whether risk perception mediated the relationship between the race of the customer and
the choice call to the police (i.e., single-item dichotomous choice question). For the
mediational analysis, I completed three separate regression analyses (i.e., two linear and
one multiple regression), and then completed a Sobel test (Preacher & Leonardelli, 2001)
to examine the indirect effect of path a and path b (Baron & Kenny, 1986).
Dependent Variables Examination
The four items: (1) “Based on this scenario, what is the likelihood that you would
call the police?”; (2) “If someone else saw the same scenario as you and they decided to
call the police, what is the likelihood that calling the police is the correct decision?”; (3)
“Based on the scenario you previously read, would you call the police?”; (4) “If someone
else saw the same scenario as you and they decided to call the police, do you agree or
disagree with that choice?” were used as the dependent variables in analyses. Initially, the
first two items examining (1) the likelihood to call police and (2) the likelihood to have
someone else call police were going to be combined together, and the latter two items, (3)
the dichotomous choice to call the police and (4) the dichotomous choice agreement with
someone else’s decision to call police item were also going to be combined together.
36
After a preliminary examination of the data, however, results revealed that only the two
likelihood items should be combined; thus, the two dichotomous choice questions were
assessed independently in analyses.
The two likelihood items were combined to create a total likelihood score. There
was a strong significant correlation between the two likelihood items: (1) the likelihood
to call police (M = 4.89; SD = 2.94) and (2) the likelihood to agree with someone else’s
decision to call the police (M = 6.06; SD = 2.69), r = .72, p < .001, n = 294. Previous
research has combined scales with a correlation of r = .70 (e.g., Stillman, Medvedev, &
Ferguson, 2017); thus, it was deemed appropriate to combine the two questions. The total
likelihood score was used as a dependent variable to assess the research questions.
The two dichotomous choice items were not combined to create an overall
dichotomous choice score. Although there was strong significant correlation between the
two dichotomous choice items: (3) the dichotomous choice to call police and (4) the
dichotomous choice agreement with someone else’s decision to call the police item,
r(292) = .50, p < .001 (Table 3) these items were not combined in analyses. Despite the
significant correlation of the two items, about 41% of participants changed their decision
between the choice to call the police and the agreement with someone else’s decision to
call the police (Table 4). Only 38% of participants said they would call the police
whereas 63% agreed with someone else’s decision to call the police. This change in
perception of the choice to call the police and the agreement with someone else’s
decision to call police could be possibly explained by previous literature examining
bystander influence (e.g., Latané, 1981; Latané & Darley, 1970; Latané & Nida, 1981) or
37
incorrect interpretations of the items due to the wording of the two items. Based on these
results, both dichotomous choice questions “Based on the scenario you previously read,
would you call the police?” and “If someone else saw the same scenario as you and they
decided to call the police, do you agree or disagree with that choice?” were assessed as
separate dependent variables to assess the research questions.
Table 3
Correlations of the Dependent Variables
Variables 1 2 3 M(SD)
1. Likelihood:
Call Police
4.89(2.94)
2. Likelihood:
Someone Else
.72** 6.06(2.69)
3. Dichotomous:
Call Police
.82*** .61*** NA
4. Dichotomous:
Someone Else
.49*** .58*** .52*** NA
Note. (ns = 294-295); ***: p < .001; **: p < .01; *: p < .05.
1. The likelihood to call the police item.
2. The likelihood to agree with someone else’s decision to call the police item.
3. The dichotomous choice item to call the police item.
4. The dichotomous choice agreement with someone else’s decision to call the police.
Table 4
Dichotomous Choice Item Response
Call Police (%) Someone Else Call Police (%)
111 (38%) 187 (63%)
183 (62%) 108 (37%)
Note. (ns = 294-295).
Evaluations of the Scenario, Police Involvement, and the Suspect
Before the research questions were explored, analyses were also conducted to
understand the evaluations of the scenario, police involvement, and the suspect in relation
to the dependent variables, using these three items:
38
“To what extent, if any, do you think the scenario is risky?”
“To what extent, if any, do you worry that calling the police would bring
you trouble?”
“To what extent, if any, was the customer in the front of the store
dangerous?”
There was a significant positive association between the perceived risk in the
scenario and the dichotomous choice to call the police item, r = .50, p < .001 n = 294.
There was also a significant positive correlation between the perceived risk of the
scenario and the dichotomous choice agreement with someone else’s decision to call the
police, r = .45, p < .001 n = 295. Lastly, there was a significant positive correlation
between the perceived risk of the scenario and the total likelihood to call the police score,
r = .68, p < .001, n = 294 (Table 5).
Based on these correlations, independent samples t-test were conducted using the
dichotomous choice dependent variables and the perceived risk of scenario item. First,
there was a significant difference between the decision to call the police, (M = 7.37; SD =
1.38) and to not call the police, (M = 5.26; SD = 1.97), t(292) = 9.882, p < .001, d = 1.24.
Participants rated the scenario as riskier when they decided to call the police compared to
when they did not call police. The scenario was also perceived as more risky when there
was agreement with someone else’s decision to call the police, (M = 6.75; SD = 1.73)
than disagreement with someone else’s decision to call the police, (M = 4.83; SD = 1.98),
t(293) = 8.626, p < .001, d = 1.03.
39
Considering the item, police involvement potentially bringing trouble, there was a
nonsignificant association between police involvement bringing trouble and the
dichotomous choice to call the police item, r = -.01, p = .09, n = 294. There was also a
nonsignificant association between police involvement bringing trouble and the
dichotomous choice agreement with someone else’s decision to call the police, r = -.05, p
= .41 n = 295. Lastly, there was a nonsignificant association between police involvement
bringing trouble and the total likelihood to call the police score, r = -.06, p = .30, n = 294
(Table 5). Due to the nonsignificant correlations, independent samples t-tests were not
conducted.
Concerning the item, how dangerous was the customer in the scenario, there was
a significant positive association between the perceived danger level of the customer and
the dichotomous choice to call the police item, r = .61, p < .001 n = 294. There was also
a significant positive correlation between the perceived dangerousness of the customer
and the dichotomous choice agreement with someone else’s decision to call the police, r
= .51, p < .001 n = 295. Lastly, there was a significant positive correlation between the
perceived dangerousness of the customer and the total likelihood to call the police score,
r = .747, p < .001, n = 294 (Table 5).
Based on these correlations, independent samples t-test were conducted using the
dichotomous choice dependent variables and the perceived dangerousness of the
customer item. First, there was a significant difference between the decision to call the
police, (M = 7.55; SD = 1.65) and to not call the police, (M = 4.76; SD = 1.86), t(292) = -
12.988, p < .001, d = 1.59. Participants rated the participant as more dangerous when they
40
decided to call the police compared to when they did not call the police. The customer
was also rated as more dangerous when there was agreement with someone else’s
decision to call the police, (M = 6.67; SD = 1.97) than disagreement with someone else’s
decision to call the police, (M = 4.31; SD = 1.87), t(293) = -10.084, p < .001, d = 1.23.
Collectively, these results indicate that when the scenario was perceived as riskier
there was greater reported likelihood to call the police and more agreement with someone
else’s decision to call the police. Additionally, when the customer in the scenario was
perceived as more dangerous there was greater reported likelihood to call the police and
more agreement with someone else’s decision to call the police. Thus, the assessment of
possible risk and danger were important aspects in the decision to call the police and the
agreement with someone else’s decision to call the police.
In addition to examining the evaluation questions and the dependent variables, I
examined the correlations among the variables by condition (i.e., black and white
perceived suspect). The correlations were fairly consistent in each condition (Table 6).
For instance, there are strong significant correlations among the dependent variables in
both conditions. Further, the age variable does not significantly correlate among the
dependent variables in either conditions. An interesting pattern emerges, however,
between the risk perception score and the three dependent variables. In the black
condition, the risk perception score significantly correlates with the three dependent
variables; however, in the white condition, the risk perception score only significantly
correlates with the dichotomous choice agreement with someone else’s decision to call
the police. This same pattern emerges for political orientation. In the black condition, all
41
three of the dependent variables are significantly correlated with political orientation,
whereas in the white condition, none of the dependent variables are significantly
correlated with political orientation.
Table 5
Correlations of the Evaluation Questions and the Dependent Variables
1 2 3 4 5
1. Risky
2. Trouble .03
3. Dangerous .76*** .04
4. Call Police .50*** -.01 .61***
5. Someone
Else Call Police
.45*** -.05 .51*** .52***
6. Likelihood
Score
.68*** -.60 .75*** .75*** .57***
Note. (ns = 293-295); ***: p < .001; **: p < .01; *: p < .05.
1. Risky: a higher value indicates more perceived risk.
2. Trouble: a higher value indicates more perceived trouble.
3. Dangerous: a higher value indicates more perceived danger.
4. The dichotomous choice to call police: single item.
5. The dichotomous choice agreement with someone else’s decision to call the police:
single item.
6. Total likelihood to call police score: composite score.
42
Table 6
Correlations of the Variables by Condition
1 2 3 4 5 6 7
1.Call Police .52** .77** -.12 -.06 .24** -.18*
2. Someone
Else Call
Police
.52** .60** -.10 -.06 .26** -.18*
3. Likelihood
Score
.77** .55** -.10 -.12 .29** -.26**
4. Age .04 -.03 .06 -.20* .13 -.14
5. Gender -.25 -.18* -.27** -.21** -.31** -.02
6. Risk .12 .16 .25** .17* -.31** -.10
7. PO .04 -.06 -.09 -.09 -.04 -.02
Note. Correlations above the diagonal represents the black perceived suspect condition
and the correlations below the diagonal represents the white perceived suspect condition.
ns = 140-149; **: p < .01; *: p < .05. 1. The dichotomous choice to call police item. 2.
The dichotomous choice agreement with someone else’s decision to call the police item.
3. Total likelihood to call police score. 4. Age of the participant. 5. Gender of the
participant (- female and + male). 6. Risk perception score. 7. Political orientation score
(a higher score indicates liberal ideology and a lower score indicates conservative
ideology).
Hierarchical Linear Regression: Likelihood to Call Police
First, the control variables, age and gender, were added to the model, [R2 = .040,
R2adj = .033, F(2, 285) = 5.951, p < .01]. Next, the addition of race of the customer in the
scenario and risk perception to the model added significantly, [R2 = .086, R2adj = .073,
F(4, 283) = 7.099, p < .001]. The interaction term between risk perception and race of the
customer, however, did not significantly add to the regression model, [R2 = .086, R2adj =
.070, F(5, 282) = .001, p = .97].
43
Research Question 1
First, I examined whether participants were more likely to call the police when the
customer was black compared to white with a hierarchical linear regression. As shown in
Table 7, Block 2, there was not a significant main effect of race of the customer in the
scenario influencing the likelihood to call the police, β = -.24, 95% CI [-.822, .337],
t(282) = -.824, p = .41. Participants did not report a difference in the likelihood to call
the police based on race.
Research Question 2
Next, I examined whether participants who were more risk averse were more
likely to call the police. There was a significant main effect for risk perception and
likelihood to call the police, β = .66, 95% CI [.301, 1.028], t(282) = 3.594, p < .001
(Table 7: Block 2). Participants who were risk averse were more likely to call the police.
Research Question 3
Then, I examined the interaction of risk perception and the race of the customer.
There was not a significant interaction of risk perception and race of the customer on the
likelihood to call the police, β = .01, 95% CI [-.679, .705], t(282) = .037, p = .970 (Table
7: Block 3).
44
Tab
le 7
The
Lik
elih
ood t
o C
all
Poli
ce
Note
. ***:
p <
.001;
**:
p <
.01;
*:
p <
.05. n =
288.
Gen
der
: dum
my c
oded
wit
h f
emal
e =
0 a
nd m
ale
= 1
.
Rac
e re
fers
to t
he
race
of
the
cust
om
er i
n t
he
scen
ario
; dum
my c
od
ed w
ith w
hit
e =
0 a
nd
1 =
bla
ck.
Ris
k r
efer
s to
ris
k p
erce
pti
on:
a hig
her
sco
re i
ndic
ates
ris
k a
ver
sion a
nd a
low
er s
core
indic
ates
ris
k s
eekin
g;
the
var
iable
was
mea
n c
ente
red.
Inte
ract
ion R
isk* R
ace
incl
udes
ris
k p
erce
pti
on
sca
le a
nd r
ace
of
the
cust
om
er, ca
lcula
ted b
y m
ult
iply
ing r
isk p
erce
pti
on a
nd
race
of
the
cust
om
er.
Outc
om
e
Var
iable
Blo
cks
& V
aria
ble
s
B (
SE
)
R2
R2
chan
ge
F c
han
ge
T
ole
rance
Lik
elih
ood t
o
Cal
l P
oli
ce
Blo
ck 1
.040
.040
5.9
51**
A
ge
-.01 (
.01)
.9
58
G
ender
-1
.06 (
.31)*
**
.958
B
lock
2
.0
86
.046
7.0
99***
A
ge
-.02 (
.01)
.9
48
G
ender
-.
74 (
.31)*
.882
R
ace
-.
24 (
.29)
.9
92
R
isk
.66 (
.19)*
**
.894
B
lock
3
.0
86
.000
.001
A
ge
-.02 (
.01)
.9
48
G
ender
-.
74 (
.32)*
.882
R
ace
-.
24 (
.30)
.9
92
R
isk
.66 (
.25)*
*
.4
78
In
tera
ctio
n:
Ris
k*R
ace
.01 (
.35)
.5
08
45
Overall, women compared to men were more likely to call the police, and risk
averse individuals compared to risk seeking individuals were more likely to call the
police.
Examining Gender: Linear Regression
To avoid redundancy, the control variables block including age and gender was
removed. First, the variables, age, gender, race of the customer in the scenario, and risk
perception were added to the model, [R2 = .086, R2adj = .073, F(4, 283) = 6.652, p < .001].
The addition of interaction term between risk perception and gender did not add
significantly to regression model, [R2 = .094, R2adj = .070, F(6, 281) = 1.218, p = .30]. [R2
= .040, R2adj = .033, F(2, 285) = 5.951, p < .01]
In the previous model (Table 7: Block 2), women were more likely to call the
police compared to men; thus, only the interaction term between risk perception and
gender are explored. There was not a significant interaction of risk perception and gender
on the likelihood to call the police, β = -.58, 95% CI [-1.312, .152], t(281) = -1.560, p =
.12 (Table 8: Block 2).
46
Tab
le 8
The
Infl
uen
ce o
f G
end
er o
n t
he
Lik
elih
ood t
o C
all
Poli
ce
Outc
om
e
Var
iable
Blo
cks
& V
aria
ble
s
B (
SE
)
R2
R2
chan
ge
F
chan
ge
Tole
rance
Lik
elih
ood t
o
Cal
l P
oli
ce
Blo
ck 1
.086
.086
6.6
52*
**
A
ge
-.02 (
.01)
.9
48
G
ender
-.
74 (
.31)*
.882
R
ace
-.
24 (
.29)
.9
92
R
isk
.66 (
.19)*
**
.894
B
lock
2
.0
94
.008
1.2
18
A
ge
-.02 (
.01)
.9
47
G
ender
-.
70 (
.32)*
.877
R
ace
-.
24 (
.29)
.9
92
R
isk
1.0
0 (
.34)*
*
.2
73
In
tera
ctio
n:
Ris
k*R
ace
.004 (
.35)
.4
05
In
tera
ctio
n:
Ris
k*G
ender
-.
58 (
.37)
.5
08
Note
. ***:
p <
.001;
**:
p <
.01;
*:
p <
.05. n =
288.
Gen
der
: dum
my c
oded
wit
h f
emal
e =
0 a
nd m
ale
= 1
.
Rac
e re
fers
to t
he
race
of
the
cust
om
er i
n t
he
scen
ario
; dum
my c
od
ed w
ith W
hit
e =
0 a
nd 1
= B
lack
.
Ris
k r
efer
s to
ris
k p
erce
pti
on:
a hig
her
sco
re i
ndic
ates
hig
her
ris
k a
ver
sion a
nd a
low
er s
core
indic
ates
ris
k s
eekin
g;
the
var
iable
was
mea
n c
ente
red.
Inte
ract
ion R
isk* R
ace
incl
udes
ris
k p
erce
pti
on s
cale
and r
ace
of
the
cust
om
er.
Inte
ract
ion R
isk*G
ender
incl
udes
ris
k p
erce
pti
on s
cale
and g
ender
.
47
Overall, women compared to men were more likely to call the police, and risk
averse individuals compared to risk seeking individuals were more likely to call the
police.
Examining Political Orientation: Linear Regression
To avoid redundancy, the control variables block including age and gender was
removed. First, the variables age, gender, race of the customer in the scenario, risk
perception, and political orientation were added to the model, [R2 = .123, R2adj = .108,
F(5, 282) = 6.652, p < .001]. The addition of the interaction term risk perception and
political orientation did not add significantly to the regression model, [R2 = .131, R2adj =
.106, F(8, 279) = .821, p = .48].
There was a significant main effect of political orientation, β = -.47, 95% CI [-
.737, 977], t(282) = -3.461, p < .001 (Table 9: Block 1). Liberals were less likely to call
the police compared to conservatives. There was not a significant interaction of risk
perception and political orientation on the likelihood to call the police, β = -.15, 95% CI
[-1.312, .152], t(282) = -.949, p = .34.
48
Tab
le 9
The
Infl
uen
ce o
f P
oli
tica
l O
rien
tati
on o
n t
he
Lik
elih
ood t
o C
all
Poli
ce
Note
. ***:
p <
.001;
**:
p <
.01;
*:
p <
.05. n =
288.
Gen
der
: dum
my c
oded
wit
h f
emal
e =
0 a
nd m
ale
= 1
.
Rac
e re
fers
to
the
race
of
the
cust
om
er i
n t
he
scen
ario
; dum
my c
od
ed w
ith W
hit
e =
0 a
nd 1
= B
lack
.
Ris
k r
efer
s to
ris
k p
erce
pti
on:
a hig
her
sco
re i
ndic
ates
hig
her
ris
k a
ver
sion a
nd a
low
er s
core
indic
ates
ris
k s
eekin
g;
the
var
iable
was
mea
n c
ente
red.
Inte
ract
ion R
isk* R
ace
incl
udes
ris
k p
erce
pti
on s
cale
and r
ace
of
the
cust
om
er.
Inte
ract
ion R
isk*G
ender
incl
udes
ris
k p
erce
pti
on s
cale
and g
ender
. In
tera
ctio
n R
isk*P
O i
ncl
udes
ris
k p
erce
pti
ons
scal
e an
d
poli
tica
l ori
enta
tion
.
Outc
om
e
Var
iable
Blo
cks
& V
aria
ble
s
B (
SE
)
R2
R2
chan
ge
F c
han
ge
T
ole
rance
Lik
elih
ood t
o
Cal
l P
oli
ce
Blo
ck 1
.123
.123
7.9
25*
**
A
ge
-.02 (
.01)
.9
35
G
ender
-.
83 (
.31)*
.875
R
ace
-.
02 (
.30)
.9
45
R
isk
.62 (
.18)*
**
.890
P
O
-.47 (
.14)*
**
.932
B
lock
2
.1
31
.008
.821
A
ge
-.02 (
.01)
.9
34
G
ender
-.
77 (
.31)*
*
.8
62
R
ace
-.
03 (
.30)
.9
45
R
isk
.86 (
.33)*
*
.2
65
P
O
-.46 (
.14)*
**
.918
In
tera
ctio
n:
Ris
k*R
ace
.05 (
.36)
.4
75
In
tera
ctio
n:
Ris
k*G
ender
-.
44 (
.37)
.4
00
In
tera
ctio
n:
Ris
k*P
O
-.15 (
.15)
.9
23
49
Overall, women compared to men were more likely to call the police, risk averse
participants compared to risk seeking participants were more likely to call the police, and
conservatives were more likely to call the police compared to liberals.
Logistic Regression: Call Police
Further, to assess the research questions, I also completed two logistic regressions
based on the questions: “Based on the scenario you previously read, would you call the
police?” and “If someone else saw the same scenario as you and they decided to call the
police, do you agree or disagree with that choice?” For each research question, these two
items’ results are presented in this sequential order as “Call Police” and “Someone Else
Call Police.”
Results indicated that the overall fit of the five predictors (age, gender, race, risk
perception, and the interaction term: risk*race) was questionable due to a high -2 Log
likelihood score, -2 Log likelihood = 362.974 (a perfect model would have a -2 Log
Likelihood score of 0 as defined by Mertler & Vannatta, 2013), but the -2 Log Likelihood
was statistically reliable in distinguishing between the predictors [2(5) = 13.979, p =
.02]. The model correctly classified 63.8% of the cases. Further, the correlation matrix
(Table 10) suggests that a majority of the variables are not strongly intercorrelated, which
suggests that the regression is reliable. Risk perception and the interaction between risk
and race, however, were strongly related, (r = -.66).
50
Table 10
Correlation Matrix: The Choice to Call Police
Age Gender Race Risk
Gender .19
Race -.03 .03
Risk -.08 .20 .10
Interaction: Risk*Race .02 .00 -.09 -.66
Note. n = 287. Race refers to the race of the customer in the scenario; dummy coded with
white = 0 and 1 = black.
Risk refers to risk perception: a higher score indicates risk aversion and a lower score
indicates risk seeking; the variable was mean centered.
Interaction Risk* Race includes risk perception scale and race of the customer, calculated
by multiplying risk perception and race of the customer.
Research Question 1
First, I examined whether race was a significant predictor in the choice to call the
police. As shown in Table 11, Block 2, using the Wald chi-square statistic to test the
individual regression coefficients (Peng, Lee, & Ingersoll, 2002), race was not a
significant predictor in the choice to call the police, β = -.15, p = .56. Race did not affect
the choice to call the police.
Research Question 2
Next, I examined whether risk perception was a significant predictor in the choice
to call the police. As shown in Table 11, Block 2 risk perception was a significant
predictor in the choice to call the police, β = .34, p = .04. Participants who were risk
averse were more likely to call the police.
51
Research Question 3
Then, I examined the interaction of risk perception and the race of the customer to
predict the choice to call the police. There was not a significant interaction of risk
perception and race of the customer on the choice to call the police, β = .26, p = .69
(Table 11: Block 3).
52
Tab
le 1
1
The
Cho
ice
to C
all
Poli
ce
Outc
om
e
Var
iable
Blo
cks
& V
aria
ble
s
β (
SE
β)
Wald
’s
2
df
p
e β
(Odds
Rat
io)
Cal
l P
oli
ce
Blo
ck 1
A
ge
-.01 (
.01)
1.3
6
1
.24
0.9
87
G
ender
-.
71 (
.26)*
*
7.6
5
1
.01
2.0
25
B
lock
2
A
ge
-.02 (
.01)
1.7
7
1
.18
0.9
85
G
ender
-.
55 (
.27)*
4.2
4
1
.04
1.7
33
R
ace
-.
15 (
.25)
.335
1
.56
1.1
57
R
isk
.34 (
.16)*
4.4
0
1
.04
1.4
08
B
lock
3
A
ge
-.02 (
.01)
.011
1
.19
0.9
85
G
ender
-.
55 (
.28)*
.2
67
1
.04
1.7
33
R
ace
-.
16 (
.26)
.253
1
.52
1.1
78
R
isk
.22(.
30)
.216
1
.31
1.2
48
In
tera
ctio
n:
Ris
k*R
ace
.26 (
.31)
.314
1
.41
1.2
97
Note
. ***:
p <
.001;
**:
p <
.01;
*:
p <
.05. n =
287.
Gen
der
: dum
my c
oded
wit
h f
emal
e =
0 a
nd m
ale
= 1
.
Rac
e re
fers
to t
he
race
of
the
cust
om
er i
n t
he
scen
ario
; dum
my c
od
ed w
ith w
hit
e =
0 a
nd 1
= b
lack
.
Ris
k r
efer
s to
ris
k p
erce
pti
on:
a hig
her
sco
re i
ndic
ates
ris
k a
ver
sion a
nd a
low
er s
core
indic
ates
ris
k s
eekin
g;
the
var
iable
was
mea
n c
ente
red.
Inte
ract
ion R
isk* R
ace
incl
udes
ris
k p
erce
pti
on
sca
le a
nd r
ace
of
the
cust
om
er, ca
lcu
late
d b
y m
ult
iply
ing r
isk p
erce
pti
on a
nd
race
of
the
cust
om
er.
53
Overall, the model predicted that risk averse individuals compared to risk seeking
individuals chose to call the police more and women compared to men chose to call the
police more.
Examining Gender: Logistic Regression (Call Police)
Results indicated that the overall fit of the six predictors (age, gender, race of the
customer, risk perception, the interaction term: risk*race, and the interaction term:
risk*gender) was questionable due to a high -2 Log likelihood score, -2 Log likelihood =
359.847, but the -2 Log likelihood was statistically reliable in distinguishing between the
predictors [2(6) = 17.106, p = .01]. The model correctly classified 63.4% of the cases.
Further, the correlation matrix (Table 12) suggests that a majority of the variables are not
strongly intercorrelated, which suggests that the regression is reliable. Risk perception
and the interaction between risk and race (r = -.48) and the interaction between risk and
gender (r = -.69), however, does have a strong correlation.
54
Table 12
Correlation Matrix: Gender and the Choice to Call Police
Age Gender Race Risk Interaction:
Risk*Race
Gender .18
Race -.03 .02
Risk -.10 .28 .07
Interaction:
Risk*Race
.02 -.15 -.06 -.48
Interaction:
Risk*Gender
.06 .01 -.02 -.69 .02
Note. n = 287. Race refers to the race of the customer in the scenario; dummy coded with
white = 0 and 1 = black.
Risk refers to risk perception: a higher score indicates risk aversion and a lower score
indicates risk seeking; the variable was mean centered.
Interaction Risk* Race includes risk perception scale and race of the customer, calculated
by multiplying risk perception and race of the customer.
Interaction Risk*Gender includes risk perception scale and gender, calculated by
multiplying risk perception and race of the customer.
In a previous model (Table 11: Block 2), women were more likely to call the
police compared to men; thus, only the interaction term between risk perception and
gender are explored. There was not a significant interaction of risk perception and gender
on the likelihood to call the police, β = -.59, p = .08 (Table 13: Block 2).
55
Tab
le 1
3
The
Infl
uen
ce o
f G
end
er o
n t
he
Choic
e to
Call
Poli
ce
Outc
om
e
Var
iable
Blo
cks
& V
aria
ble
s
β (
SE
β)
Wald
’s
2
df
p
e β
(Odds
Rat
io)
Cal
l P
oli
ce
Blo
ck 1
A
ge
-.02 (
.01)
1.7
7
1
.18
0.9
85
G
ender
-.
55 (
.27)
*
4.2
4
1
.04
1.7
33
R
ace
-.
15 (
.25)
.34
1
.56
1.1
57
R
isk
.34 (
.16)
*
4.4
0
1
.04
1.4
08
B
lock
2
A
ge
-.02 (
.01)
1.9
63
1
.16
0.9
84
G
ender
-.
49 (
.28)
3.0
52
1
.08
1.6
26
R
ace
-.
16 (
.25)
.375
1
.54
1.1
68
R
isk
.57 (
.30)
3.5
42
1
.06
1.7
69
In
tera
ctio
n:
Ris
k*R
ace
.24 (
.31)
.593
1
.44
1.2
72
In
tera
ctio
n:
Ris
k*G
ender
-.
59 (
.34)
3.0
25
1
.08
0.5
54
Note
. ***:
p <
.001;
**:
p <
.01;
*:
p <
.05. n =
287.
Gen
der
: dum
my c
oded
wit
h f
emal
e =
0 a
nd m
ale
= 1
.
Rac
e re
fers
to t
he
race
of
the
cust
om
er i
n t
he
scen
ario
; dum
my c
od
ed w
ith W
hit
e =
0 a
nd 1
= B
lack
.
Ris
k r
efer
s to
ris
k p
erce
pti
on:
a hig
her
sco
re i
ndic
ates
hig
her
ris
k a
ver
sion a
nd a
low
er s
core
indic
ates
ris
k s
eekin
g;
the
var
iable
was
mea
n c
ente
red.
Inte
ract
ion R
isk* R
ace
incl
udes
ris
k p
erce
pti
on s
cale
and r
ace
of
the
cust
om
er.
Inte
ract
ion R
isk*G
ender
incl
udes
ris
k p
erce
pti
on s
cale
and g
ender
.
56
Overall, the model predicted that women compared to men chose to call the police
more and risk averse individuals compared to risk seeking individuals chose to call the
police more.
Examining Political Orientation: Logistic Regression (Call Police)
Results indicated that the overall fit of the eight predictors (age, gender, race of the
customer, PO, risk perception, the interaction term: risk*race, the interaction term:
risk*gender, and the interaction term: risk*PO) was questionable (-2 Log likelihood =
357.978; Mertler & Vannatta, 2013), but the -2 Log likelihood was statistically reliable in
distinguishing between the predictors [2(8) = 18.974, p = .015]. The model correctly
classified 64.1% of the cases. Further, the correlation matrix (Table 14) suggests that a
majority of the variables are not strongly intercorrelated, which suggests that the
regression is reliable. Risk perception and the interaction between risk and race (r = -.50)
and the interaction between risk and gender (r = -.69), however, does have a strong
correlation.
57
Table 14
Correlation Matrix: Political Orientation and the Choice to Call Police
Age Gender Race Risk PO Interaction:
Risk*Race
Interaction:
Risk*Gender
Gender .19
Race -.05 -.00
Risk -.09 -.26 .05
PO .12 -.11 -.21 .07
Interaction:
Risk*Race
.02 -.02 -.06 -.50 .04
Interaction:
Risk*Gender
.05 -.14 .00 -.69 -.07 .03
Interaction:
Risk*PO
-.00 -.05 .03 .14 -.00 -.21 -.04
Note. n = 287. Race refers to the race of the customer in the scenario; dummy coded with
white = 0 and 1 = black.
Risk refers to risk perception: a higher score indicates risk aversion and a lower score
indicates risk seeking; the variable was mean centered. PO refers to political orientation:
a higher score indicates liberal ideology and a lower score conservative ideology.
Interaction Risk* Race includes risk perception scale and race of the customer, calculated
by multiplying risk perception and race of the customer.
Interaction Risk*Gender includes risk perception scale and gender, calculated by
multiplying risk perception and race of the customer. Interaction Risk*PO includes risk
perception scale and political orientation, calculated by multiplying risk perception and
race of the customer.
There was not a significant main effect of political orientation, β = -.17, p = .16
(Table 15: Block 1). There were no political orientation differences in the choice to call
the police. There was not a significant interaction of risk perception and political
orientation on the likelihood to call the police, β = -.09, p = .52.
.
58
Tab
le 1
5
The
Infl
uen
ce o
f P
oli
tica
l O
rien
tati
on o
n t
he
Choic
e to
Call
Poli
ce
Note
. ***:
p <
.001;
**:
p <
.01;
*:
p <
.05. n =
287.
Gen
der
: dum
my c
oded
wit
h f
emal
e =
0 a
nd m
ale
= 1
.
Rac
e re
fers
to t
he
race
of
the
cust
om
er i
n t
he
scen
ario
; dum
my c
od
ed w
ith W
hit
e =
0 a
nd 1
= B
lack
.
Ris
k r
efer
s to
ris
k p
erce
pti
on:
a hig
her
sco
re i
ndic
ates
hig
her
ris
k a
ver
sion a
nd a
low
er s
core
indic
ates
ris
k s
eekin
g;
the
var
iable
was
mea
n c
ente
red.
PO
(poli
tica
l ori
enta
tion)
refe
rs t
o p
oli
tica
l ori
enta
tion v
aria
ble
that
was
co
mpute
d f
rom
4 i
tem
s an
d w
as m
ean c
ente
red.
Inte
ract
ion R
isk* R
ace
incl
udes
ris
k p
erce
pti
on s
cale
and r
ace
of
the
cust
om
er.
Inte
ract
ion R
isk*G
ender
incl
udes
ris
k p
erce
pti
on s
cale
and g
ender
.
Inte
ract
ion R
isk*P
O i
ncl
udes
ris
k p
erce
pti
on s
cale
and p
oli
tica
l ori
enta
tio
n.
Outc
om
e
Var
iable
Blo
cks
& V
aria
ble
s
β (
SE
β)
Wald
’s
2
df
p
e β
(Odds
Rat
io)
Cal
l P
oli
ce
Blo
ck 1
A
ge
-.02 (
.01)
2.1
74
1
.14
0.9
83
G
ender
-.
58 (
.27)*
4.6
94
1
.03
1.7
92
R
ace
-.
07 (
.26)
.075
1
.78
1.0
73
R
isk
.33 (
.16)*
3.9
48
1
.05
1.3
86
P
O
-.17 (
.12)
1.9
44
1
.16
0.8
47
B
lock
2
A
ge
-.02 (
.01)
2.2
89
1
.13
0.9
83
G
ender
-.
51 (
.28)
3.3
10
1
.07
1.6
65
R
ace
-.
10 (
.26)
.140
1
.71
1.1
02
R
isk
.51 (
.31)
2.8
01
1
.09
1.6
71
P
O
-.15 (
.12)
1.4
27
1
.23
0.8
65
In
tera
ctio
n:
Ris
k*R
ace
.27 (
.32)
.709
1
.40
1.3
09
In
tera
ctio
n:
Ris
k*G
ender
-.
55 (
.34)
2.5
84
1
.11
0.5
79
In
tera
ctio
n:
Ris
k*P
O
-.09 (
.14)
.416
1
.52
0.9
14
59
Overall, women chose to call the police more than men and risk averse
individuals chose to call the police more than risk seeking individuals.
Logistic Regression: Someone Else Call Police
Results indicated that the overall fit of the five predictors (age, gender, race, risk
perception, and the interaction term: risk*race) was questionable (-2 Log likelihood =
364.142), but the -2 Log likelihood was statistically reliable in distinguishing between the
predictors [2(5) = 15.882, p < .001]. The model correctly classified 65.3% of the cases.
Further, the correlation matrix (Table 16) suggests that a majority of the variables are not
strongly intercorrelated, which suggests that the regression is reliable. Risk perception
and the interaction between risk and race (r = -.67), however, does have a strong
correlation.
Further, the correlation matrix (Table 14) suggests that a majority of the variables are
not strongly intercorrelated, which suggests that the regression is reliable. Risk
perception and the interaction between risk and race (r = -.50) and the interaction
between risk and gender (r = -.69), however, does have a strong correlation.
Table 16
Correlation Matrix: Someone Else’s Decision to Call Police
Age Gender Race Risk
Gender .19
Race -.04 .01
Risk -.11 .19 .01
Interaction: Risk*Race .01 .01 .10 -.67
Note. n = 288. Race refers to the race of the customer in the scenario; dummy coded with
white = 0 and 1 = black.
Risk refers to risk perception: a higher score indicates risk aversion and a lower score
indicates risk seeking; the variable was mean centered.
Interaction Risk* Race includes risk perception scale and race of the customer, calculated
by multiplying risk perception and race of the customer.
60
Research Question 1
First, I assessed whether race was a significant predictor in the agreement with
someone else’s decision to call the police. As shown in Table 17, Block 2, again, using
the Wald chi-square statistic to test the individual regression coefficients (Peng et al.,
2002), race was not a significant predictor in the agreement with someone else’s decision
to call the police, β = -.02, p = .93. Race did not affect the agreement with someone else’s
decision to call the police.
Research Question 2
Next, I examined whether risk perception was a significant predictor in the
agreement with someone else’s decision to call the police. Risk perception was a
significant predictor in the agreement with someone else’s decision to call the police, (β =
.47, p = .003; Table 17: Block 2). Risk averse individuals were more likely to agree with
someone else’s decision to call the police.
Research Question 3
Lastly, for my main analyses, I examined the interaction of risk perception and the
race of the customer to predict the agreement with someone else’s decision to call the
police. There was not a significant interaction of risk perception and the race of the
customer regarding the agreement with someone else’s decision to call the police, (β =
.21, p = .49; Table 17: Block 3).
61
Tab
le 1
7
Som
eone
Els
e’s
Dec
isio
n t
o C
all
Poli
ce
Outc
om
e
Var
iable
Blo
cks
& V
aria
ble
s
β (
SE
β)
Wald
’s
2
df
p
e
β
(Odds
Rat
io)
Som
eone
Els
e
Cal
l P
oli
ce
Blo
ck 1
A
ge
-.02 (
.01)
2.3
2
1
.13
0.9
84
G
ender
-.
59 (
.26)*
5.1
6
1
.02
1.7
97
B
lock
2
A
ge
-.02 (
.01)
3.3
1
1
.07
0.9
79
G
ender
-.
38 (
.28)
1.9
4
1
.16
1.5
50
R
ace
-.
02 (
.26)
.008
1
.93
1.0
78
R
isk
.47 (
.16)*
*
8.5
9
1
.00
1.5
85
B
lock
3
A
ge
-.02 (
.01)
3.2
9
1
.07
0.9
80
G
ender
-.
38 (
.27)
1.9
5
1
.16
1.4
59
R
ace
-.
01 (
.25)
.00
1
.98
1.0
06
R
isk
.37 (
.22)
2.8
5
1
.09
1.4
41
In
tera
ctio
n:
Ris
k*R
ace
.21 (
.30)
.48
1
.49
1.2
34
Note
. ***:
p <
.001;
**:
p <
.01;
*:
p <
.05. n =
288.
Gen
der
: dum
my c
oded
wit
h f
emal
e =
0 a
nd m
ale
= 1
.
Rac
e re
fers
to t
he
race
of
the
cust
om
er i
n t
he
scen
ario
; dum
my c
od
ed w
ith w
hit
e =
0 a
nd
1 =
bla
ck.
Ris
k r
efer
s to
ris
k p
erce
pti
on:
a hig
her
sco
re i
ndic
ates
ris
k a
ver
sion a
nd a
low
er s
core
indic
ates
ris
k s
eekin
g;
the
var
iable
was
mea
n c
ente
red.
Inte
ract
ion R
isk*
Rac
e in
cludes
ris
k p
erce
pti
on
sca
le a
nd r
ace
of
the
cust
om
er, ca
lcula
ted b
y m
ult
iply
ing r
isk p
erce
pti
on a
nd
race
of
the
cust
om
er.
62
Overall, the model predicted that risk averse individuals compared to risk seeking
individuals agreed more with someone else’s decision to call the police, and that women
compared to men agreed more with someone else’s decision to call the police.
Examining Gender: Logistic Regression (Someone Else Call Police)
Results indicated that the overall fit of the six predictors (age, gender, race of the
customer, risk perception, the interaction term: risk*race, and the interaction term:
risk*gender) was questionable due to a high -2 Log likelihood score, -2 Log likelihood =
358.405, but the -2 Log likelihood was statistically reliable in distinguishing between the
predictors [2(6) = 21.619, p < .001]. The model correctly classified 66.7% of the cases.
Further, the correlation matrix (Table 18) suggests that a majority of the variables are not
strongly intercorrelated, which suggests that the regression is reliable. Risk perception
and the interaction between risk and race (r = -.50) and the interaction between risk and
gender (r = -.71), however, does have a strong correlation.
63
Table 18
Correlation Matrix: Gender and Someone Else’s Decision to Call Police
Age Gender Race Risk Interaction:
Risk*Race
Age
Gender -.20
Race -.04 .01
Risk -.12 .12 -.01
Interaction:
Risk*Race
.01 .00 .12 -.48
Interaction:
Risk*Gender
.06 -.01 -.01 -.71 .00
Note. n = 288. Race refers to the race of the customer in the scenario; dummy coded with
white = 0 and 1 = black.
Risk refers to risk perception: a higher score indicates risk aversion and a lower score
indicates risk seeking; the variable was mean centered.
Interaction Risk* Race includes risk perception scale and race of the customer, calculated
by multiplying risk perception and race of the customer.
Interaction Risk*Gender includes risk perception scale and gender, calculated by
multiplying risk perception and race of the customer.
In a previous model (Table 17: Block 1), women were more likely to call the
police compared to men; thus, only the interaction term between risk perception and
gender are explored. There was a significant interaction of risk perception and gender on
the agreement with someone else’s decision to call the police, β = -.59, p = .08 (Table 19:
Block 2). The interaction effects were examined through using a median split to create a
risk seeking and risk averse group. This method has been adopted in other studies (e.g.,
Peters, & Bjalkebring, 2015; Peters et al., 2009; Peters, Sol Hart, Tusler, & Fraenkel,
2014). Risk averse women compared to risk seeking women were more likely to agree
with someone else’s decision to call the police, whereas men who were risk seeking and
64
risk averse agreed with someone else’s decision to call the police similarly. Further, risk
averse women agreed more with someone else’s decision to call the police than risk
averse men overall. (Figure 1).
65
Tab
le 1
9
The
Infl
uen
ce o
f G
end
er o
n t
he
Agre
emen
t w
ith S
om
eone
Els
e’s
Dec
isio
n t
o C
all
Poli
ce
Outc
om
e
Var
iable
Blo
cks
& V
aria
ble
s
β (
SE
β)
Wald
’s
2
df
p
e β
(Odds
Rat
io)
Som
eone
Els
e
Cal
l P
oli
ce
Blo
ck 1
A
ge
-.02 (
.01)
3.3
1
1
.07
0.9
80
G
ender
-.
38 (
.27)
1.9
4
1
.16
0.6
86
R
ace
-.
02 (
.25)
.008
1
.93
0.9
77
R
isk
.47 (
.16)
**
8.5
9
1
.00
1.5
97
B
lock
2
A
ge
-.02 (
.01)
3.6
0
1
.07
0.9
57
G
ender
.3
8 (
.27)
1.8
8
1
.16
1.4
56
R
ace
.0
0 (
.25)
.00
1
.98
0.9
98
R
isk
.86 (
.22)
7.4
9
1
.09
2.3
59
In
tera
ctio
n:
Ris
k*R
ace
.20 (
.31)
.427
1
.52
1.2
27
In
tera
ctio
n:
Ris
k*G
ender
-.
79 (
.34)*
5.4
55
1
.02
0.4
55
Note
. ***:
p <
.001;
**:
p <
.01;
*:
p <
.05.
Gen
der
: dum
my c
oded
wit
h f
emal
e =
0 a
nd m
ale
= 1
.
Rac
e re
fers
to t
he
race
of
the
cust
om
er i
n t
he
scen
ario
; dum
my c
od
ed w
ith W
hit
e =
0 a
nd 1
= B
lack
.
Ris
k r
efer
s to
ris
k p
erce
pti
on:
a hig
her
sco
re i
ndic
ates
hig
her
ris
k a
ver
sion a
nd a
low
er s
core
indic
ates
ris
k s
eekin
g;
the
var
iable
was
mea
n c
ente
red.
Inte
ract
ion R
isk*R
ace
incl
udes
ris
k p
erce
pti
on a
nd r
ace
of
the
cust
om
er.
Inte
ract
ion R
isk*G
ender
incl
udes
ris
k p
erce
pti
on s
cale
and g
ender
.
66
Interaction between Gender and Risk Perception on the Agreement with Someone Else’s
Decision to Call the Police
Figure 1. Bars indicate standard errors.
67
Overall, the model showed that risk averse individuals compared to risk seeking
individuals agreed more with someone else’s decision to call the police. Additionally,
women who were risk averse compared to risk seeking agreed more with someone else’s
decision to call the police. Further, risk averse and risk seeking men agreed with someone
else’s decision to call the police similarly.
Examining Political Orientation: Logistic Regression (Someone Else Call Police)
Results indicated that the overall fit of the eight predictors (age, gender, race of
the customer, PO, risk perception, the interaction term: risk*race, the interaction term:
risk*gender, and the interaction term: risk*PO) was questionable (-2 Log likelihood =
357.978; Mertler & Vannatta, 2013), but the -2 Log likelihood was statistically reliable in
distinguishing between the predictors [2(8) = 18.974, p = .015]. The model correctly
classified 64.1% of the cases. Further, the correlation matrix (Table 20) suggests that a
majority of the variables are not strongly intercorrelated, which suggests that the
regression is reliable. Risk perception and the interaction between risk and race (r = -.47)
and the interaction between risk and gender (r = -.70), however, does have a strong
correlation.
68
Table 20
Correlation Matrix: Political Orientation and the Agreement with Someone Else’s
Decision to Call Police
Age Gender Race Risk PO Interaction:
Risk*Race
Interaction:
Risk*Gender
Age
Gender .20
Race -.07 -.04
Risk -.11 .10 .01
PO .14 .13 -.24 .01
Interaction:
Risk*Race
.03 -.01 .15 -.47 -.05
Interaction:
Risk*Gender
.04 .01 -.03 -.70 -.04 -.02
Interaction:
Risk*PO
.02 -.04 -.06 -.01 -.26 -.25 .06
Note. n = 288. Race refers to the race of the customer in the scenario; dummy coded with
white = 0 and 1 = black.
Risk refers to risk perception: a higher score indicates risk aversion and a lower score
indicates risk seeking; the variable was mean centered. PO refers to political orientation:
a higher score indicates liberal ideology and a lower score conservative ideology.
Interaction Risk* Race includes risk perception scale and race of the customer, calculated
by multiplying risk perception and race of the customer.
Interaction Risk*Gender includes risk perception scale and gender, calculated by
multiplying risk perception and race of the customer. Interaction Risk*PO includes risk
perception scale and political orientation, calculated by multiplying risk perception and
race of the customer.
There was a significant main effect of political orientation, β = -.29, p = .02
(Table 21: Block 1). Liberals were less likely to agree with someone else’s decision to
call the police compared to conservatives. There was also a significant interaction of risk
perception and political orientation on the agreement with someone else’s decision to call
69
the police, β = -.34, p < .001. The interaction effects were examined through using a
median split to create a risk seeking and risk averse group. This method has been adopted
in other studies (e.g., Peters, & Bjalkebring, 2015; Peters et al., 2009; Peters et al., 2014).
Political orientation was split into two separate categories: political conservative and
political liberal. Participants who reported to be “Very Conservative” or “Conservative
were considered conservative, whereas participants who reported to be “Liberal” or
“Very Liberal” were considered liberal. This method has been adopted in other studies
(e.g., Champion, 1987; Cutshall & Adams, 1983; Wilbanks & Kim, 1984). Conservatives
who were risk averse were more likely to agree with someone else’s decision to call the
police than risk seeking conservatives, whereas for liberals, risk perceptions did not affect
agreement with someone else’s decision to call the police similarly (Figure 2).
70
Tab
le 2
1
The
Infl
uen
ce o
f P
oli
tica
l O
rien
tati
on o
n t
he
Agre
emen
t w
ith S
om
eone
Els
e’s
Dec
isio
n t
o C
all
Poli
ce
Outc
om
e
Var
iable
Blo
cks
& V
aria
ble
s
β (
SE
β)
Wald
’s
2
df
p
e β
(Odds
Rat
io)
Cal
l P
oli
ce
Blo
ck 1
A
ge
-.02 (
.01)*
4.3
86
1
.04
0.9
77
G
ender
-.
44 (
.28)
2.6
16
1
.11
0.6
41
R
ace
.1
1 (
.26)
.189
1
.66
1.1
20
R
isk
.45 (
.16)*
*
7.8
53
1
.01
1.5
75
P
O
-.29 (
.12)*
5.4
16
1
.02
0.7
52
B
lock
2
A
ge
-.03 (
.01)*
4.4
90
1
.03
0.9
76
G
ender
-.
41 (
.28)
2.1
58
1
.14
0.6
62
R
ace
.1
6 (
.27)
.349
1
.56
1.1
73
R
isk
.79 (
.32)*
6.0
19
1
.01
2.1
94
P
O
-.33 (
.13)*
6.2
53
1
.01
0.7
20
In
tera
ctio
n:
Ris
k*R
ace
.38 (
.33)
1.2
93
1
.26
1.4
58
In
tera
ctio
n:
Ris
k*G
ender
-.
74 (
.35)*
4.6
36
1
.03
0.4
76
In
tera
ctio
n:
Ris
k*P
O
-.34 (
.15)*
*
4.9
66
1
.03
0.7
14
Note
. ***:
p <
.001;
**:
p <
.01;
*:
p <
.05. n =
288.
Gen
der
: dum
my c
oded
wit
h f
emal
e =
0 a
nd m
ale
= 1
.
Rac
e re
fers
to t
he
race
of
the
cust
om
er i
n t
he
scen
ario
; dum
my c
od
ed w
ith W
hit
e =
0 a
nd 1
= B
lack
.
Ris
k r
efer
s to
ris
k p
erce
pti
on:
a hig
her
sco
re i
ndic
ates
hig
her
ris
k a
ver
sion a
nd a
low
er s
core
indic
ates
ris
k s
eekin
g;
the
var
iable
was
mea
n c
ente
red.
PO
(poli
tica
l ori
enta
tion)
refe
rs t
o p
oli
tica
l ori
enta
tion v
aria
ble
that
was
com
pute
d f
rom
4 i
tem
s
and w
as m
ean
cen
tere
d.
A h
igher
sco
re i
ndic
ates
lib
eral
ideo
log
y a
nd a
low
er s
core
indic
ates
conse
rvat
ive
ideo
log
y.
Inte
ract
ion R
isk* R
ace
incl
udes
ris
k p
erce
pti
on s
cale
and r
ace
of
the
cust
om
er.
Inte
ract
ion R
isk*G
ender
incl
udes
ris
k p
erce
pti
on s
cale
and g
ender
.
Inte
ract
ion R
isk*P
O i
ncl
udes
ris
k p
erce
pti
on s
cale
and p
oli
tica
l ori
enta
tio
n.
71
Interaction between PO and Risk Perception on the Agreement with Someone Else’s
Decision to Call the Police
Figure 2. Bars indicate standard errors.
Overall, risk averse participants agreed more with someone else’s decision to call
the police than risk seeking individuals and conservatives agreed more with someone
else’s decision to call the police than liberals. Additionally, risk averse women agreed
more with someone else’s decision to call the police than risk seeking women, whereas
there were no differences based on risk preferences for men. Conservatives agreed more
72
with someone else’s decision to call the police than those who were risk seeking, whereas
there were no differences based on risk preferences for liberals.
Mediation Analysis
I examined whether risk perception mediated the relationship between race of
customer and choice to call the police (i.e., dichotomous variable). I conducted four steps
to examine mediation (Baron & Kenny, 1986). First, the regression of race of the
customer on the choice to call the police, ignoring the mediator, was not significant, b =
.02, t(293) = .373, p = .71. Next, the regression of the race of the suspect on the mediator,
risk perception, was also not significant, b = .13, t(293) = 1.340, p = .18. Then the
mediation process showed that risk perception, controlling for the race of the customer,
was significant, b = .12, t(293) = 3.658, p < .001. Then analyses revealed that controlling
for risk perception, race of suspect was not a significant predictor of the choice to call the
police, b = -1.54, t(293) = .094, p = .93. A Sobel test (Preacher & Leonardelli, 2001) was
conducted and found no mediation in the model (z = 1.25, p = .21; Figure 3). In
summary, these four steps illustrated that there was no evidence of mediation within the
model.
73
Mediation Model
Figure 3. *** p < .001. Race of Customer = independent variable; Risk Perception =
mediator; Choice to Call Police = Dependent Variable. n = 294.
Risk
Perception
Choice to
Call Police
Race of
Customer
a = .13 b = .12***
c = .02 (c’ = .12)
74
CHAPTER 5.
DISCUSSION
Summary of Results
Race of the customer (i.e., black or white) in the scenario did not influence the
likelihood to call the police, whether participants would call the police, or agreement with
someone else’s choice to call the police. Regardless of race, individuals who were risk
averse reported a higher likelihood to call the police, reported more often that they would
call the police, and agreed more with someone else’s decision to call the police. Taken
together, the race of the perceived suspect and the assessment of perceived risk (i.e., risk
aversion) did not influence the likelihood to call the police, whether participants would
call the police, or their agreement with someone else’s choice to call the police.
Women reported a higher likelihood to call the police, reported more often that they
would call the police, and agreed more with someone else’s decision to call the police
than men. Risk averse women, compared to risk seeking women, agreed more with
someone else’s decision to call the police. Also, conservatives reported a higher
likelihood to call the police, reported more often that they would call the police, and
agreed more with someone else’s decision to call the police than liberals. Risk averse
conservatives, compared to risk seeking conservatives, agreed more with someone else’s
decision to call the police.
Race
Across all analyses, race of the perceived suspect (i.e., the customer in the
scenario) was not influential in the reported likelihood to call the police, whether
75
participants would call the police, and their agreement with someone else’s choice to call
the police. News reports of calls made to the police (e.g., a student napping in a dorm
common area; Griggs, 2018) could be isolated incidents given the amount of 911 calls
received (an estimated 240 million calls made to 911 each year; “9-1-1 Statistics,” 2017).
Publicized news reports of black individuals having the police called on them could
create an availability heuristic. An availability heuristic suggests that a person judges the
occurrence of events by the ease of retrieval of relevant examples (Tversky & Kahneman,
1973). For instance, videos taken by the student who was napping in a dorm common
space received over a million views (one of the videos a million and a half views) and
was mentioned on the popular show, The Daily Show with Trevor Noah (Noah, 2018).
Perhaps, due to these publicized reports, the frequency of calls to police on black
individuals for seemingly harmless activities comes to mind easily; however, this
availability heuristic could be biasing the reality of the frequency of these calls.
One explanation for the ineffectiveness of race to influence the reported
likelihood to call the police, whether participants would call the police, and their
agreement with someone else’s choice to call the police could be that the sample used
might not have had biased attitudes toward black individuals. Although mTurk
populations are similar to the demographic characteristics of aversive racists (e.g., liberal,
well-educated, and white; Dovidio & Gaertner, 2000; Huff & Tingley, 2015; Levay,
Freese, & Druckman, 2016; Nail et al., 2003), there was no evidence for aversive racism
in the current study. Thus, the current sample might not have had biased attitudes toward
76
black individuals. The study sample was only 74% white, and racial/ethnic minorities
may not hold the same biases against black individuals.
Another possibility is that the customer being identified briefly as a young black
or white male through a text description might not have been enough to influence the
likelihood to call the police, whether participants would call the police, and their
agreement with someone else’s choice to call the police. The young black male
stereotype of delinquency and criminality (Steffensmeier, Ulmer, & Kramer, 1998) might
not have been aroused in the scenario. The use of images to represent the customer in the
scenario might have been more effective in representing the demographic characteristics
(i.e., race) of the customer. For instance, judges see and interact with a defendant before a
decision is made regarding incarceration and sentencing. Judges, like U.S. citizens, can
fall victim to stereotypes, perceiving young, black men as more dangerous, involved in
street life, and less reformable (Daly, 1994) Thus, simply reading a scenario might not
have aroused real-life feelings and attitudes of a perceived suspect (i.e., the customer in
the scenario).
Despite race remaining a highly charged topic in society (Suchet, 2004), these
results suggest that in this study, race was not influential in the reported likelihood to call
the police, whether participants would call the police, and their agreement with someone
else’s choice to call the police. Further, the race of the perceived suspect and risk
perception (i.e., risk aversion) did not influence the likelihood to call the police, whether
participants would call the police, or their agreement with someone else’s choice to call
the police.
77
Risk Perception
Risk averse participants compared to risk seeking participants had a higher reported
likelihood to call the police, would call the police more, and agreed more with someone
else’s decision to call the police. This finding highlights that risk perception fosters an
individual assessment of a situation (Sitkin & Pablo, 1992). Prior evidence has revealed
that risk perception invokes two methods for assessment: risk as feelings (instinctive
reactions to danger) and risk as an analysis (logical reactions to danger). Risk is often
evaluated through intuitive feelings, automatically, and handled quickly (i.e., risk as
feelings; Slovic & Peters, 2006). Thus, an evaluation based on feelings, intuition, and
instinctive reactions to danger further supports how individualistic the assessment of risk
perception is in a risk situation. In a possible crime situation, calling the police could
provide a sense of security in a perceived uncertain situation. Thus, calling the police
might not be a decision that is based solely on another person (i.e., offender), but on the
personal assessment of the situation and fear of the caller. Ultimately, the decision to call
the police is personal, and is based on a subjective evaluation of a situation.
Gender
Women compared to men had a higher reported likelihood to call the police,
reported more often that they would call the police, and agreed more with someone else’s
decision to call the police. This finding is consistent with previous research (Eith &
Durose, 2011). Women compared to men are more likely to contact the police, report a
crime, disturbances, and suspicious activity to police (Davis et al., 2018). Further, women
might have more contact with police because they have higher confidence in the police
78
than men (Apple & O’Brien, 1983; Cao et al., 1996). This might be because women have
different types of relationships with police than men. For women, contact with the police
is often less hostile, whereas for men, hostile contact (e.g., excessive police action) with
police is more common (Cao et al., 1996; Davis et al., 2018). Generally, compared to
men, women have more positive police contact and relationships, which could be
influential in their choices to contact police.
Women, compared to men, are more risk averse in numerous areas of life:
financial decision making (Jianakoplos & Bernasek, 1998), drug and alcohol choices
(Spigner et al., 1993), and gambling decisions (Levin, Synder, & Chapman, 2010).
Further, compared to men, women have a greater fear of crime (Stanko, 1995). Thus, it is
logical to extrapolate that having the police called brings a sense of certainty in an
uncertain possible crime situation for risk averse women. For instance, in the current
study, risk averse women agreed more with someone else’s decision to call the police.
Women being more risk averse might agree more with someone else’s decision to call the
police because it provides another sense of security and certainty in an uncertain
situation.
Political Orientation
Conservatives compared to liberals had a higher reported likelihood to call the
police, reported more often that they would call the police, and agreed more with
someone else’s decision to call the police. Conservatives, compared to liberals, typically
report having a high confidence in the police (Congressional Research Service, 2018;
79
Stack & Cao, 1998). This could be because political conservatives believe police
authority conveys conservative ideologies (Frimer, Gaucher, & Schaefer, 2014).
Further, research using the Moral Foundations Questionnaire (Graham et al.,
2011) suggests that liberals and conservatives have different considerations for moral
foundations (i.e., loyalty, purity, authority, fairness, and harm). In regard to the original
five foundations, political conservatives are more likely to endorse the binding
foundations (i.e., loyalty, purity, and authority), whereas political liberals are more likely
to endorse the individualizing foundations (i.e., fairness and harm; Graham et al., 2009).
The binding foundation of authority foundation invokes appeals to obedience and respect,
which is often observed in institutions of authority, such as a police department (Clifford,
Iyengar, Cabeza, & Sinnott-Armstrong, 2015). Thus, when making the decision to call
the police, political conservatives could exercise their moral beliefs by showing respect to
the laws and rules of an institutional authority.
Additionally, this study suggests that risk averse conservatives, compared to risk
seeking conservatives, agreed more with someone else’s decision to call the police. High
levels of confidence in the police could possibly explain why risk averse conservatives
are more likely to agree with someone else’s decision to call the police. Evidence has
shown that conservatives are more sensitive to fear (e.g., risk averse; Jost et al., 2003),
and with high confidence in the police (e.g., Congressional Research Service, 2018; Stack
& Cao, 1998), if someone made a call to police that might provide a sense of security and
certainty.
80
Limitations and Future Research
One potential limitation within this study is the scenario. The scenario was
author-generated and although it was pretested to address this possible limitation, it was
not perfect. The scenario used only two different combinations of race, age, and gender
(i.e., young black or white male) though a text-only description. Since the customer in the
scenario was not visible, characteristics of the customer could have been highlighted
more. For instance, other characteristics could have been added to the description of the
customer, such as wearing a hooded sweatshirt, wearing a suit, having tattoos, or baggy
pants. The racial characteristics of the customer could also have been highlighted with
use of pictures or other visual aids. Furthermore, manipulations of the race of the store
clerk might influence the motivation to call the police. Collectively, future research
should alter aspects of the scenario (e.g., using pictures), the suspect (e.g., wearing baggy
pants), and the actors in the scenario (e.g., a black store clerk) to examine their influence
on the motivation to call the police.
There are also some demographic limitations of the mTurk population. Although
mTurk provides more diversity than undergraduate populations, the demographic
variables on mTurk are often different than the U.S. populations (Ross, Irani, Silberman,
Zaldivar, & Tomlinson, 2010). For example, the mTurk population often has lower
income, higher education, and consists of smaller percentages of minority populations
compared to the national average (Huff & Tingley, 2015). Future research should explore
different samples and continue to examine mTurk populations to examine motivation to
81
call the police. Examining different sample populations will allow there to be greater
external validity of understanding the motivation to call the police.
Implications
This study examined an exploratory approach to understand the motivation to call
the police through aversive racism theory (Gaertner & Dovidio, 1986) and risk averse
motivation (Kahneman & Tversky, 1982). According to aversive racism theory, an
aversive racist would have potentially rationalized a call to police in an ambiguous
situation if the customer was black compared to white. According to risk averse
motivation, after an assessment of the situation, a risk averse individual would have
potentially been motivated to call the police to provide a sense of certainty in an
uncertain situation. The race of the customer in the scenario (i.e., perceived suspect) did
not influence the reported likelihood to call the police, whether participants would call
the police, or their agreement with someone else’s choice to call the police. Risk averse
participants, however, reported a higher likelihood to call the police, would call the police
more, and agreed more with someone else’s decision to call the police than risk seeking
participants. The current study adds to a preliminary start of the examination of aversive
racism theory and risk averse motivation in relation to decision to call the police. Further,
the current study draws attention to the limitation that there is no available information
regarding the demographic characteristics of police callers.
Practical implications include the need to educate 911 dispatch operators and 911
callers to enhance the use of police resources and time. Depending on the assessment of a
situation, there are varying degrees of what constitutes a need for police resources. With
82
an execution of questions asked to 911 callers, however, police resources could be
potentially saved from a situation where these resources are not needed.
Further, 911 education is not a required subject within United States school
systems. It is important that children learn at a young age the resources that calling police
provides. A 911 education could help ensure appropriate use of this privilege for both
children and adults. Although perceived risk is subjective, a general education on right-
and-wrong instances to use 911 is informative on ensuring accurate and professional use.
For example, on Nextdoor.com, which is used as platform for neighborhood contact on
various topics (i.e., finding a local plumber, neighborhood events), reports of a
“suspicious black man” were surfacing. To reduce this racial profiling, rather than
instantly being able to post “suspicious black man,” on the site, Nextdoor users were
asked to slow down in their observation, to identity actual suspicious behavior, and then
be specific about what the person looks like to avoid putting all black people in the same
category. Slowing down and thinking about what people were posting helped reduce
racial profiling by 75% on the website (Eberhardt, 2019). Extrapolating this concept,
slowing down and taking notice of the details of an event that could motivate a call to the
police could help individuals recognize potentially biased calls or unnecessary calls of
when police resources are truly not needed.
In the current study, the race of a perceived suspect (i.e., the customer in the
scenario) did not influence the decision to call the police, but race as a potential influence
in the decision to call the police should not be ignored from further exploration. There are
many disparities that exist between minority and majority groups, such as arrest and
83
incarceration rates. Thus, it is necessary to explore and understand why these disparities
exist to find solutions.
Conclusion
In this study, it is evident that the race of a suspect involved in a possible crime
was not influential, whereas the individual variability in risk perception did affect the
reported likelihood to call the police, whether participants would call the police, and their
agreement with someone else’s decision to call the police. Being a woman and a
conservative also affected the reported likelihood to call the police, whether participants
would call the police, and their agreement with someone else’s decision to call the police.
The overall finding of race not affecting the reported likelihood to call the police,
whether participants would call the police, and their agreement with someone else’s
decision to call the police does not dismiss that future research should still examine and
explore race as a motivation to call the police. News reports continue to demonstrate that
unjust calls are being made to police on black individuals. As these unjust calls continue,
the motivation to call the police should be explored. Racial injustices, discrimination,
biases, and disparities exist; thus, these issues need to be continually studied.
84
REFERENCES
Alpert, G. P., MacDonald, J. M., & Dunham, R. G. (2005). Police suspicion and
discretionary decision-making during citizen stops. Criminology, 43, 407–434.
doi:10.1111/j.0011-1348.2005.00012.x
Apple, N., & O’Brien, D. (1983). Neighborhood racial composition and residents’
evaluation of police performance. Journal of Police Science and Administration,
11, 76-84. Retrieved from
https://www.ncjrs.gov/App/abstractdb/AbstractDBDetails.aspx?id=89715
Archie, A., & Smith, E. (2018, July). Women calls the police on a black representative
campaigning in Oregon. Retrieved from
https://www.cnn.com/2018/07/04/us/oregon police-called-on-black-
representative-trnd/index.html
Barkan, S. E., & Cohn, S. F. (1998). Racial prejudice and support by whites for police
use of force: A research note. Justice Quarterly, 15, 743–753.
doi:10.1080/07418829800093971
Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable distinction in
social psychological research: Conceptual, strategic, and statistical considerations.
Journal of Personality and Social Psychology, 51, 1173–1182. doi:10.1037/0022-
3514.51.6.1173
Blais, A.-R., & Weber, E. U. (2006a). A domain-specific risk-taking (DOSPERT) scale
for adult populations. Judgment and Decision Making, 1, 33-47. Retrieved from
http://journal.sjdm.org/06005/jdm06005.htm
Blais, A.-R., & Weber, E. U. (2006b). Testing invariance in risk taking: A comparison
between Anglophone and Francophone groups. Série Scientifique, 1-25. doi:
10.1037/e518572013-183
Blumstein, A., & Beck, A. J. (1999). Population growth in U. S. prisons, 1980-1996.
Crime and Justice, 26, 17–61. doi:10.1086/449294
Blumstein, A., & Wallman, J. (2006). The crime drop and beyond. Annual Review of Law
and Social Science, 2, 125-146. Retrieved from
https://doi.org/10.1146/annurev.lawsocsci.2.081805.110011
Brown, V. J. (2014). Risk perception: It’s personal. Environmental Health Perspective,
122,A276-A279. doi: 10.1289/ehp.122-A276. doi:10.1289/ehp.122-a276
85
Buhrmester, M. D., Talaifar, S., & Gosling, S. D. (2018). An evaluation of amazon’s
mechanical turk, its rapid rise, and its effective use. Perspectives on
Psychological Science, 13, 149–154. doi:10.1177/1745691617706516
Cao, L., Frank, J., & Cullen, F. T. (1996). Race, community context and confidence in
the police. American Journal of Police, 15, 3–22.
doi:10.1108/07358549610116536
Champion, D. J. (1987). Elderly felons and sentencing severity: Interregional variations
in leniency and sentencing trends. Criminal Justice Review, 12, 7–14.
doi:10.1177/073401688701200203
Clifford, S., Iyengar, V., Cabeza, R., & Sinnott-Armstrong, W. (2015). Moral
foundations vignettes: A standardized stimulus database of scenarios based on
moral foundations theory. Behavior Research Methods, 47, 1178-1198.
doi: 10.3758/s13428-014-0551-2
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.).
Hillsdale, NJ: Lawrence Earlbaum Associates.
Congressional Research Service. (2018). Public trust and law enforcement— A
discussion for policymakers (R43904). Retrieved from
https://fas.org/sgp/crs/misc/R43904.pdf
Criss, D., & Vera, A. (2018, May). Three black people checked out of their Airbnb rental.
Then someone called the police on them. Retrieved from
https://www.cnn.com/2018/05/07/us/airbnb-police-called-trnd/index.html
Cutshall, C. R., & Adams, K. (1983). Responding to older offenders: Age selectivity in
the processing of shoplifters. Criminal Justice Review, 8, 1–8.
doi:10.1177/073401688300800202
Daly, K. (1994). Gender, Crime, and Punishment. New Haven, Connecticut: Yale
University Press.
Darley, J. M., & Latané, B. (1968). Bystander intervention in emergencies: Diffusion of
responsibility. Journal of Personality and Social Psychology, 8, 377–383.
doi:10.1037/h0025589
Davis, E., Whyde, A., & Langton, L. (2018). Contacts between police and the public,
2015. Bureau of Justice Statistics, NCJ 251145, 1-33. Retrieved from
https://www.bjs.gov/content/pub/pdf/cpp15.pdf
86
Dovidio, J. F., & Gaertner, S. L. (2000). Aversive racism and selection decisions: 1989
and 1999. Psychological Science, 11, 315-319. doi: 10.1111/1467-9280.00262
Dovidio, J. F., & Gaertner, S. L. (2004). Aversive racism. In M. Zanna (Ed.), Advances in
Experimental Social Psychology, 36, (pp. 1-52).
doi:10.1016/s00652601(04)36001-6
Dovidio, J. F., Mann, J., & Gaertner, S. L. (1989). Resistance to affirmative action: The
implications of aversive racism. In F. A. Blanchard & F. J. Crosby (Eds.),
Affirmative Action in Perspective (pp. 83–102). doi:10.1007/978-1-4613-9639-
0_7
Eberhardt, J. L. (2019). Biased: Uncovering the hidden prejudice that shape what we see,
think, and do. New York, NY: Viking.
Eith, C., & Durose, M. R. (2011). Contacts between police and the public, 2008. Bureau
of Justice Statistics, NCJ 234599, 2-27. Retrieved from
https://www.bjs.gov/content/pub/pdf/cpp08.pdf
Ekins, E. (2016). Policing in America: Understanding public attitudes toward the police.
Results from a national survey: Results from the Cato Institute 2016 Criminal
Justice Survey. Cato Institute. Retrieved from https://www.cato.org/survey-
reports/policing-america
Esser, J. K., & Lindoerfer, J. S. (1989). Groupthink and the space shuttle Challenger
accident: Toward a quantitative case analysis. Journal of Behavioral Decision
Making, 2, 167-177.
doi: 10.1002/bdm.3960020304
FBI Uniform Crime Reporting Program. (2016). Persons arrested, 2016 crime in the
United States. Retrieved from https://ucr.fbi.gov/crime-in-the-u.s/2016/crime-in-
the-u.s.-2016/tables/table-21
Felson, R. B., Messner, S. F., Hoskin, A. W., & Deane, G. (2002). Reasons for reporting
and not reporting domestic violence to the police. Criminology, 40, 617
648.doi:10.1111/j.1745-9125.2002.tb00968.x
Fieldstadt, E. (2019, March). White women calls police on black man after his dog
‘assaults’ hers. Retrieved from https://www.nbcnews.com/news/us-news/white-
woman calls-police-black-man-after-his-dog-assaults-n97805
Fingerhut, H. (2017). Deep racial, partisan divisions in Americans’ views of police
officers. Washington, DC: Pew Research Center.
87
Frimer, J. A., Gaucher, D., & Schaefer, N. K. (2014). Political conservatives’ affinity for
obedience to authority is loyal, not blind. Personality and Social Psychology
Bulletin, 40, 1205–1214. doi:10.1177/0146167214538672
Gaertner, S. L., & Dovidio, J. F. (1986). The aversive form of racism. In J.F. Dovidio &
S.L. Gaertner (Eds.), Prejudice, discrimination, and racism (pp. 61.61-89).
Orlando, FL: Academic Press.
Gaertner, S. L., & Dovidio, J. F. (2005). Understanding and addressing contemporary
racism: From aversive racism to the common ingroup identity model. Journal of
Social Issues, 61, 615–639. doi:10.1111/j.1540-4560.2005.00424.x
Gaertner, S. L., Dovidio, J. F., Nier, J., Hodson, G., & Houlette, M. A. (2008). Aversive
racism: Bias without intention. In L.B. Nielsen, & R.L. Nelson (Eds.), Handbook
of employment discrimination research, (pp. 373-393). New York, New York:
Springer.
Garland, B. E., Spohn, C., & Wodahl, E. J. (2008). Racial disproportionality in the
American prison population: Using the Blumstein Method to address the critical
race and justice issue of the 21st century. Justice Policy Journal, 5, 2-42.
Retrieved from
http://www.cjcj.org/uploads/cjcj/documents/racial_disproportionality.pdf
Graham, J., Haidt, J., & Nosek, B. A. (2009). Liberals and conservatives rely on different
sets of moral foundations. Journal of Personality and Social Psychology, 96,
1029-1046. doi:10.1037/a0015141
Graham, J., Nosek, B. A., Haidt, J., Iyer, R., Koleva, S., & Ditto, P. H. (2011). Mapping
the moral domain. Journal of Personality and Social Psychology, 101, 366-385.
doi:10.1037/a0021847
Greenberg, M. S., Wilson, C. E., Ruback, R. B., & Mills, M. K. (1979). Social and
emotional determinants of victim crime reporting. Social Psychology Quarterly,
42, 364.doi:10.2307/3033806
Griggs, B. (2018, May). A black Yale graduate student took a nap in her dorm’s common
room. So a white student called police. Retrieved from
https://www.cnn.com/2018/05/09/us/yale-student-napping-black-trnd/index.html
Harris, D. A. (1999). The stories, the statistics and the law: Why ‘driving while Black’
matters. University of Minnesota Law Review, 84, 265-326. Retrieved from
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=199508
Harris, D. A. (2002). Profiles in injustice. New York: The New Press.
88
Harrison, P. M., & Beck, A. J. (2006). Prisoners in 2005 (Report No. NCJ215092).
Washington, DC: U.S. Department of Justice.
Hindelang, M. J. (1974). Public opinion regarding crime, criminal justice, and related
topics. Journal of Research in Crime and Delinquency, 11, 101–116.
doi:10.1177/002242787401100202
Huang, W., & Vaughn, M. S. (1996). Support and confidence: Public attitudes toward the
police. In T. Flanagan & D. Longmire (Eds.), Americans view crime and justice:
A national public opinion survey (pp. 31–45). Thousand Oaks, CA: Sage
Huff, C., & Tingley, D. (2015). “Who are these people?” Evaluating the demographic
characteristics and political preferences of MTurk survey respondents. Research
and Politics, 2, 205316801560464. doi:10.1177/2053168015604648
Iacobucci, D., Schneider, M. J., Popovich, D. L., & Bakamitsos, G. A. (2016). Mean
centering, multicollinearity, and moderators in multiple regression: The
reconciliation redux. Behavior Research Methods, 49, 403–404.
doi:10.3758/s13428-016-0827-9
Inbar, Y., & Lammers, J. (2012). Political diversity in social and personality psychology.
Perspectives on Psychological Science, 7, 496-503. doi:10.2139/ssrn.2002636
Irwin, J. R., & McClelland, G. H. (2001). Misleading heuristics and moderated multiple
regression models. Journal of Marketing Research, 38, 100–109.
doi:10.1509/jmkr.38.1.100.18835
Janis, I. L. (1972). Victims of groupthink. Boston, MA: Houghton Mifflin.
Janis, I. L. (1982) Groupthink. Boston: Houghton Mifflin
Jaschik, S. (2018, September). Walking on campus… While Black. Retrieved from
https://www.insidehighered.com/news/2018/09/18/incident-umass-latest-which-
calls-campus-police-suggest-racial-profiling
Jianakoplos, N. A., & Bernasek, A. (1998). Are women more risk averse? Economic
Inquiry, 36, 620-630. doi:10.1111/j.1465-7295.1998.tb01740.x
Jost, J. T., Glaser, J., Kruglanski, A. W., & Sulloway, F. J. (2003). Political conservatism
as motivated social cognition. Psychological Bulletin, 129, 339 –375. doi:
10.4324/9781315175867-5
Kahneman, D., & Tversky, A. (1982). The psychology of preferences. Scientific
American, 246, 160-173. Retrieved from https://www.jstor.org/stable/24966506
89
Kassin, S. M. (2017). The killing of Kitty Genovese: What else does this case tell us?
Perspectives on Psychological Science, 12, 374-381. doi:
10.1177/1745691616679465
Kleider-Offutt, H. M., Bond, A. D., & Hegerty, S. E. A. (2017). Black stereotypical
features: When a face type can get you in trouble. Current Directions in
Psychological Science, 26, 28–33. doi:10.1177/0963721416667916
Kline, R. B. (1998). Principles and practice of structural equation modeling. New York:
The Guilford Press.
Kusev, P., Purser, H., Heilman, R., Cooke, A. J., Van Schaik, P., Baranova, V., … Ayton,
P. (2017). Understanding risky behavior: The influence of cognitive, emotional
and hormonal factors on decision-making under risk. Frontiers in Psychology, 8,
1-10. doi:10.3389/fpsyg.2017.00102
Langton, L., & Durose, M. (2016). Police behavior during traffic and street stops, 2011.
Bureau of Justice Statistics, NCJ 242937, 1-21. Retrieved from
https://www.bjs.gov/content/pub/pdf/pbtss11.pdf
Lasley, J. R. (1994). The impact of the Rodney King incident on citizen attitudes toward
police. Policing and Society, 3, 245–255. doi:10.1080/10439463.1994.9964673
Latané, B. (1981). The psychology of social impact. American Psychologist, 36, 343-356.
http://dx.doi.org/10.1037/0003-066X.36.4.343
Latané, B., & Darley, J. M. (1970). The unresponsive bystander: Why doesn’t he help?
New York: Appleton-Centruy Crofts.
Latané, B., & Nida, S. (1981). Ten years of research on group size and helping.
Psychological Bulletin, 89, 308–324. doi:10.1037/0033-2909.89.2.308
Latané, B., & Rodin, J. (1969). A lady in distress: Inhibiting effects of friends and
strangers on bystander intervention. Journal of Experimental Social Psychology,
5, 189-202. doi:10.1016/0022-1031(69)90046-8
Levay, K. E., Freese, J., & Druckman, J. N. (2016). The Demographic and Political
Composition of Mechanical Turk Samples. SAGE Open, 6, 215824401663643.
doi:10.1177/2158244016636433
Levin, I. P., Synder, M. A., & Chapman, D. P. (2010). The interaction of experiential and
situational factors and gender in a simulated risky decision-making task. The
Journal of Psychology Interdisciplinary and Applied, 122. 173-181. doi:
10.1080/00223980.1988.9712703
90
Loury, G. (1984). A new American dilemma. The New Republic, 13-18. Retrieved from
https://www.brown.edu/Departments/Economics/Faculty/Glenn_Loury/louryhom
epage/The%20New%20Republic%20Articles/A%20New%20American%20Dile
mma%20(1984).pdf
MacCrimmon, K. R., & Wehrung, D. A. (1990). Characteristics of risk taking executives.
Management Science, 36, 422–435. doi:10.1287/mnsc.36.4.422
McCleary, K., & Vera, A. (2018, April). A video of black men being arrested at
Starbucks. Three very different reactions. Retrieved from
https://www.cnn.com/2018/04/14/us/philadelphia-police-starbucks-
arrests/index.html
Mertler, C. A., & Vannatta, R. A. (2013). Advanced and multivariate statistical methods.
Los Angeles, CA: Pyrczak.
Miu, A. C., Heilman, R. M., & Houser, D. (2008). Anxiety impairs decision-making:
Psychophysiological evidence from an Iowa Gambling Task. Biological
Psychology, 77, 353–358. doi:10.1016/j.biopsycho.2007.11.010
Morgan, R. E., & Truman, J. L. (2018). Criminal Victimization, 2017 (Report No. NCJ
252472). Washington, DC: U.S. Department of Justice.
Myrdal, G. (1944). An American dilemma:The negro problem and modern democracy.
New York: Harper.
Nail, P. R., Harton, H. C., & Decker, B. P. (2003). Political orientation and modern
versus aversive racism: Tests of Dovidio and Gaertner's (1998) integrated model.
Journal of Personality and Social Psychology, 84, 754-770. doi:10.1037/0022-
3514.84.4.754
Nellis, A. (2016). The color of justice: Racial and ethnic disparity in state prisons. The
Sentencing Project. Retrieved from http://www.sentencingproject.org/wp-
content/uploads/2016/06/The-Color-of-Justice-Racial-and-Ethnic-Disparity-in-
State-Prisons.pdf
9-1-1 statistics. (2017, December). 9-1-1 call volume. NENA: The 9-1-1 Association.
Retrieved from https://www.nena.org/page/911Statistics
Noah, T. [The Daily Show with Trevor Noah]. (2018, May 18). Police called on sleeping
Black student & Trump meets hostages from North Korea
[Video file]. Retrieved from https://www.youtube.com/watch?v=Z_x0zaX4Z_0
91
Parker, R., Jr. (1984). Ghostbusters [Recorded by R. Parker Jr.]. On Ghostbusters:
Original soundtrack album [CD]. Brooklyn, New York: Arista.
Paulsen, D. J., Platt, M. L., Huettel, S. A., & Brannon, E. M. (2012). From risk-seeking to
risk-averse: The development of economic risk preference from childhood to
adulthood. Frontiers in Psychology, 3. doi:10.3389/fpsyg.2012.00313
Pearson, A. R., Dovidio, J. F., & Gaertner, S. L. (2009). The nature of contemporary
prejudice: Insights from aversive racism. Social and Personality Psychology
Compass, 3, 314–338. doi:10.1111/j.1751-9004.2009.00183.x
Peng, C.-Y. J., Lee, K. L., & Ingersoll, G. M. (2002). An introduction to logistic
regression analysis and reporting. The Journal of Educational Research, 96, 3–14.
doi:10.1080/00220670209598786
Peters, E., & Bjalkebring, P. (2015). Multiple numeric competencies: When a number is
not just a number. Journal of Personality and Social Psychology, 108, 802-822.
doi: 10.1037/pspp0000019
Peters, E., Dieckmann, N. F., Västfjäll, D., Mertz, C. K., Slovic, P., & Hibbard, J. H.
(2009). Bringing meaning to numbers: The impact of evaluative categories on
decisions. Journal of Experimental Psychology: Applied, 15, 213-227.
doi:10.1037/a0016978
Peters, E., Sol Hart, P., Tusler, M., & Fraenkel, L. (2014). Numbers matter to informed
patient choices: A randomized design across age and numeracy levels. Medical
Decision Making, 34, 430–442. doi:10.1177/0272989x13511705
Pierson, E., Simoiu, C., Overgoor, J., Corbett-Davies, S., Jenson, D., Shoemaker, A., …
Goel, S. (2019). A large-scale analysis of racial disparities in the police stops
across the United States. Stanford Computational Policy Lab, 1-10. Retrieved
from https://5harad.com/papers/100M-stops.pdf
Plous, S., & Williams, T. (1995). Racial stereotypes from the days of American slavery:
A continuing legacy1. Journal of Applied Social Psychology, 25, 795–817.
doi:10.1111/j.1559-1816.1995.tb01776.x
Preacher, K. J., & Leonardelli, G. J. (2001, March). Calculation for the Sobel test: An
interactive calculation tool for mediation tests [Computer software]. Retrieved
from http://quantpsy.org
Ropeik, D. (2012). The Perception Gap: Recognizing and managing the risks that arise
when we get risk wrong. Food and Chemical Toxicology, 50, 1222–1225.
doi:10.1016/j.fct.2012.02.015
92
Ross, J., Irani, L., Silberman, M. S., Zaldivar, A., & Tomlinson, B. (2010, April). Who re
the crowdworkers? Shifting demographics in Mechanical Turk." Paper presented
at the meeting of Human Factors in Computing Systems – Proceedings, Atlanta,
Georgia.
Sampson, R. (2002). Misuse and abuse of 911. Problem-Oriented Guides for Police
Series Guide No. 19, U.S. Department of Justice, Office of Community Oriented
Policing Services. Retrieved from
http://www.popcenter.org/problems/pdfs/Misuse_and_Abuse_of_911.pdf
Sawyer, W., & Wagner, P. (2019). Mass incarceration: The whole pie 2019. Prison
Policy Initiative. Retrieved from
https://www.prisonpolicy.org/reports/pie2019.html
Sever, B., & Youdin, R. (2006). Police knowledge of older populations: The impact of
training, experience and education. Professional Issues in Criminal Justice: A
Professional Journal, 1, 35-54. Retrieved from
https://kucampus.kaplan.edu/documentstore/docs09/pdf/picj/vol1/issue2/Police_K
nowled ge_of_Older_Populations.pdf
Sitkin, S. B., & Pablo, A. L. (1992). Reconceptualizing the determinants of risk behavior.
The Academy of Management Review, 17, 9-38. doi:10.2307/258646
Skogan, W. G. (1984). Reporting crimes to the police: The status of world research.
Journal of Research in Crime and Delinquency, 21, 113–137.
doi:10.1177/0022427884021002003
Skogan, W. G., & Maxfield, M. G. (1981). Coping with crime: Individual and
neighborhood reactions. Beverly Hills, CA: Sage.
Slovic, P., & Peters, E. (2006). Risk perception and affect. Pscyhological Science, 15,
322-325. Retrieved from https://journals.sagepub.com/doi/pdf/10.1111/j.1467-
8721.2006.00461.x
Sparks, R., Genn, H., & Dodd, D. (1977). Surveying victims: A study of the
measurement of criminal victimization, perceptions of crime, and attitudes to
criminal justice. New York: John Wiley.
Spigner, C., Hawkins, W. E., & Loren, W. (1993). Gender differences in perception of
risk associated with alcohol and drug use among college students. Women and
Health, 20, 87–97. doi:10.1300/j013v20n01_06
93
Stack, S. J., & Cao, L. (1998). Political conservatism and confidence in the police: A
comparative analysis. Journal of Crime and Justice, 21, 71-76. doi:
10.1080/0735648x.1998.9721066
Stanko, E. A. (1995). Women, crime, and fear. The Annals of the American Academy of
Political and Social Science, 539. 46-58. Retrieved from
https://www.jstor.org/stable/pdf/1048395.pdf
Stanton, S. J., Liening, S. H., & Schultheiss, O. C. (2011). Testosterone is positively
associated with risk taking in the Iowa Gambling Task. Hormones and Behavior,
59, 252–256. doi:10.1016/j.yhbeh.2010.12.003
Steffensmeier, D., Ulmer, J., & Kramer, J. (1998). The interaction of race, gender, and
age in criminal sentencing: The punishment cost of being young, black, and male.
Criminology, 36, 763-798. doi:10.1111/j.1745-9125.1998.tb01265.x
Stillman, P. E., Medvedev, D., & Ferguson, M. J. (2017). Resisting temptation: Tracking
how self-control conflicts are successfully resolved in real time. Psychological
Science, 28, 1240–1258. doi:10.1177/0956797617705386
Suchet, M. (2004). A relational encounter with race. Psychoanalytic Dialogues, 14, 423-
438. doi:10.1080/10481881409348796
Takei, C. (2018, June). How police can stop being weaponized by bias-motivated 911
calls. Retrieved from https://www.aclu.org/blog/racial-justice/race-and-criminal-
justice/how-police-can-stop-being-weaponized-bias-motivated
Travis, J., Western, B., & Redburn, S. (2014). The growth of incarceration in the United
States: Exploring causes and consequences. Washington, D. C.: The National
Academies Press.
Tversky, A., & Kahneman, D. (1973). Availability: A heuristic for judging frequency and
probability. Cognitive Psychology, 5, 207-232. Retrieved from:
https://msu.edu/~ema/803/Ch11-JDM/2/TverskyKahneman73.pdf
2017 Hate Crime Statistics. (2018). Uniform Crime Report, Hate Crime Statistics, 2017.
U.S. Department of Justice—Federal Bureau of Investigation. Retrieved from
https://ucr.fbi.gov/hate-crime/2017/topic-pages/incidents-and-offenses
Tyler, T. R. (2005). Policing in black and white: Ethnic group differences in trust and
confidence in the police. Police Quarterly, 8, 322–342.
doi:10.1177/1098611104271105
94
U.S. Census Bureau. (2017). 2013-2017 American Community Survey 5-Year Estimates.
American FactFinder. Retrieved from
https://factfinder.census.gov/faces/tableservices/jsf/pages/productview.xhtml?pid
=ACS_17_5YR_CP05&prodType=table
Walker, S., Spohn, C., & DeLone, M. (2000). The color of justice: Race, ethnicity, and
crime in America. Belmont, CA: Wadsworth.
Weich, R. H., & Angulo, C. T. (2002). Racial disparities in the American criminal
justice system. In D.M. Piche, W.L. Taylor & R.A. Reed (Eds.), Rights at risk:
Equality in an age of terrorism (pp. 185-218). Washington, DC: Citizens’
Commission on Civil Rights.
Wen, F., He, Z., & Chen, X. (2014). Investors’ Risk Preference Characteristics and
Conditional Skewness. Mathematical Problems in Engineering, 2014, 1–14.
doi:10.1155/2014/814965
Whitford, E. (2018, August). Police called on Black student eating lunch. Retrieved from
https://www.insidehighered.com/quicktakes/2018/08/03/police-called-black-
student-eating-lunch
Wilbanks, W., & Kim, P. (1984). Elderly criminals. New York, NY: University Press of
America.
Williams, D. (2018a, July). Neighbor calls police on a 12-year-old boy for mowing the
wrong lawn. Retrieved from https://www.cnn.com/2018/07/01/us/police-called-
lawn-mowing-boy-trnd/index.html
Williams, D. (2018b, August). Women says supermarket called police on her while she
was helping a homeless man. Retrieved from
https://www.cnn.com/2018/08/01/us/police-called-on-good-samaritan-
trnd/index.html
Yuen, K. S., & Lee, T. M. (2003). Could mood state affect risk-taking decisions? Journal
of Affective Disorders, 75, 11–18. doi:10.1016/s0165-0327(02)00022-8
Zaleskiewicz, T. (2001). Beyond Risk seeking and risk aversion: Personality and the dual
nature of economic risk taking. European Journal of Personality, 15, S105-S122.
doi: 10.1002/per.426
95
Zamble, E., & Annesley, P. (1987). Some determinants of public attitudes toward the
police. Journal of Police Science and Administration, 15, 285-290. Retrieved
from
https://www.researchgate.net/publication/232570464_Some_Determinants_of_Pu
blic_Attitudes_toward_the_Police
Zhang, R., Brennan, T. J., & Lo, A. W. (2014). The origin of risk aversion. Proceedings
of the National Academy of Sciences of the United States of America, 111, 17777-
17782. doi: 10.1073/pnas.1406755111
96
APPENDIX A: AMBIGUOUS RISK SCENARIO
Ambiguous Risk Scenario
It is 11:00 p.m. on a Friday night, and you just got off work. Once you get into
your car to go home you realize you need gas. Before going home, you decide to stop at
your local gas station. While your car begins in fill with gas, you decide to go inside the
gas station for some late-night snacks.
You’re in the back of the store looking at snack options in the freezer section,
when you hear people talking at the counter at the front of the store. The talking gets
louder and you realize the two individuals who were talking begin to yell, and you hear
fists pound on the counter. You decide to take a look at the two people yelling and you
slowly peek around the aisle in which you are hiding, and you see the profile of a young,
*(Black/White)* male.
97
APPENDIX B: INFORMED CONSENT
Informed Consent
UNIVERSITY OF NORTHERN IOWA
HUMAN PARTICIPANTS REVIEW INFORMED CONSENT
Project Title: Perceptions of Crime
Name of Investigator(s): Alivia L. Zubrod & Jiuqing Cheng, PhD
Invitation to Participate: You are invited to participate in a research project conducted
through the University of Northern Iowa. The University requires that you give your
agreement to participate in this project. The following information is provided to help
you make an informed decision about whether or not to participate.
Nature and Purpose: This study investigates evaluations and attitudes of a mock crime
summary.
Explanation of Procedures: As a participant in this study, you will be asked to read a
mock crime summary. You will be asked to evaluate the crime scenario and the suspect
presented from the summary. You will also be asked to report your perceptions of risk
and police. In addition, you will be asked to complete demographic attitude questions
and demographic measures. This study is expected to last approximately 20-30 minutes.
You may discontinue involvement in the study at any time.
Discomfort and Risks: There is minimal anticipated risk involved with participating in
this study. You may feel slightly uncomfortable answering some of the questions about
risk perception and your evaluation of the crime summary.
Benefits and Compensation: You will be compensated $1.00 for participating. Your
participation in this study will help us improve our understanding of the public’s
perception of crime.
Confidentiality: All data will be kept confidential; Worker ID’s will be deleted after
completion. Your responses will be encrypted when sent over the internet. Although
your confidentiality will be maintained to the degree permitted by the technology used,
no guarantees can be made regarding the interception of data by third parties when that
data is sent over the internet. Summarized findings with no identifying information may
be published in an academic journal or presented at a scholarly conference. Data with no
IP addresses/other identifiers may also be available for others to view on an open data
site (i.e., open science framework). These data may be used for additional studies.
Right to Refuse or Withdraw: Your participation is voluntary. You are free to
withdraw from participation at any time or to choose not to participate at all, and by
doing so, you will not be penalized or lose benefits to which you are otherwise entitled.
Questions: For questions about the study or desire information in the future regarding
your participation or the study generally, you can contact the project investigators,
Alivia L. Zubrod at zubroda@uni.edu or the project investigator’s faculty advisor
Jiuqing Cheng, Ph.D. at the Department of Psychology, University of Northern Iowa at
jiuqing.cheng@uni.edu. You can also contact the IRB Administrator, University of
Northern Iowa, at anita.gordon@uni.edu for questions about rights of research
participants and the participant review process.
Agreement: Registering for the study and clicking on the arrow below indicates that I
98
am fully aware of the nature and extent of my participation in this project as stated
above and the possible risks arising from it. I hereby agree to participate in this project. I
am 17 years of age or older.
99
APPENDIX C: DEBRIEFING FORM
Debriefing Form
Thank you for participating in the study entitled “Perceptions of Crime.” As I said
at the beginning, I am examining how people interpret evaluations of a mock crime
scenario. However, within that, I am also looking at if race and severity of risk in
scenario influence more calls to police. For instance, perceiving a high risk or weak risk
scenario the choice to call or not call police might be easy; however, with an ambiguous
scenario (in terms of risk) other factors such as race might provide as a reason to call
police. Your answers regarding your racial attitudes will be coupled with your choice to
call police. If anyone asks you what the study was about, you can honestly say that it was
about evaluations of a mock crime scenario. Please do not mention to others who
complete the study that the study looks at race a well, as that might influence how they
respond.
Worker Code: ZLAERD 3496
If you have any questions about the research protocol, theory, or results, you may
contact the Primary Researcher, Alivia L. Zubrod at zubroda@uni.edu.
Once more, thank you for your participation. We could not do our research
without you!
100
APPENDIX D: EVALUATION QUESTIONS
Evaluation Questions
Based on this scenario, what is the likelihood that you would call police?
1 (Not at all likely) (Extremely likely) 10
If someone else saw the same scenario as you and they decided to call the police, what is
the likelihood that calling the police is the correct decision?
1 (Not at all likely) (Extremely likely) 10
To what extent, if any, was the customer at the front of the store dangerous?
1 (Strongly disagree) (Strongly agree) 10
To what extent, if any, do you think the scenario is risky?
1 (Strongly disagree) (Strongly agree) 10
To what extent, if any, do you worry that calling the police would bring you trouble?
1 (Strongly disagree) (Strongly agree) 10
Based on the scenario you previously read, would you call the police?
Yes
No
If someone else saw the same scenario as you and they decided to call the police, do you
agree or disagree with that choice?
Yes
No
101
APPENDIX E: MANIPULATION CHECK
Manipulation Check
What was the race of the customer at the front of the store in the scenario?
Black
White
102
APPENDIX F: 30-ITEM DOSPERT SCALE
For each of the following statements, please indicate how risky you perceive each
situation. Provide a rating from Not at all Risky to Extremely Risky, using the
following scale.
Not at all risky – Slightly risky – Somewhat risky – Moderately risky – Risky – Very
risky – Extremely risky
1. Admitting that your tastes are different from those of a friend. (S)
2. Going camping in the wilderness. (R)
3. Betting a day’s income at the horse races. (F/G)
4. Investing 10% of your annual income in a moderate growth diversified fund. (F/I)
5. Drinking heavily at a social function. (H/S)
6. Taking some questionable deductions on your income tax return. (E)
7. Disagreeing with an authority figure on a major issue. (S)
8. Betting a day’s income at a high-stake poker game. (F/G)
9. Having an affair with a married man/woman. (E)
10. Passing off somebody else’s work as your own. (E)
11. Going down a ski run that is beyond your ability. (R)
12. Investing 5% of your annual income in a very speculative stock. (F/I)
13. Going whitewater rafting at high water in the spring. (R)
14. Betting a day’s income on the outcome of a sporting event. (F/G)
15. Engaging in unprotected sex. (H/S)
16. Revealing a friend’s secret to someone else. (E)
17. Driving a car without wearing a seat belt. (H/S)
18. Investing 10% of your annual income in a new business venture. (F/I)
19. Taking a skydiving class. (R)
20. Riding a motorcycle without a helmet. (H/S)
21. Choosing a career that you truly enjoy over a more secure one. (S)
22. Speaking your mind about an unpopular issue in a meeting at work. (S)
23. Sunbathing without sunscreen. (H/S)
24. Bungee jumping off a tall bridge. (R)
25. Piloting a small plane. (R)
26. Walking home alone at night in an unsafe area of town. (H/S)
27. Moving to a city far away from your extended family. (S)
28. Starting a new career in your mid-thirties. (S)
103
29. Leaving your young children alone at home while running an errand. (E)
30. Not returning a wallet you found that contains $200. (E)
Note. E = Ethical, F = Financial, H/S = Health/Safety, R = Recreational, and S = Social.
104
APPENDIX G: DEMOGRAPHICS
What is your gender identity?
Male
Female
Gender non-binary
Genderfluid
Genderqueer
Prefer not to answer
Not listed: __________________________________________________
What is your age? (Drop down menu)
What is your race/ethnicity? Check all that apply.
o Alaska Native
o American Indian/Native American
o Asian or Asian American
o Black or African American
o Hispanic or Latinx
o Pacific Islander
o White or Caucasian
o Prefer not to answer
o Not listed: __________________________________________________
Are you a US citizen?
Yes
No
What year were you born?
__________________________________________________
What political party do you identify, if any?
Democrat
Republican
Independent
No Affiliation
Not listed: __________________________________________________
How would you describe your...
105
Very
Conservative
Conservative Moderate Liberal Very
Liberal
Political
Orientation? o o o o o
Views on
current social
issues?
o o o o o
View on
foreign
policy?
o o o o o
Views on
economics? o o o o o
What is the highest degree or level of school you have completed? If currently enrolled,
mark the previous grade or highest degree received.
Less than high school
High School
Associate’s degree
Bachelor’s degree
Graduate degree
Not listed: _________________________________________________
106
APPENDIX H: ADDITIONAL END-OF-STUDY QUESTIONS
Additional End-of-Study Questions
How honest were your answers throughout the study? You will receive payment
regardless of what you answer.
Not honest at all
Slightly honest
Moderately honest
Extremely honest
You will still receive payment, but is there any reason we should not include your data?
__________________________________________________
__________________________________________________
What do you think the current study was about?
__________________________________________________
__________________________________________________
Do you have any comments for the researcher?
__________________________________________________
__________________________________________________
top related