Alan Stanley
Wed, 11/03/2021 - 07:22
Edited Text
DIFFERENTIAL EFFECTS OF POSITIVE MOOD
STATES ON RACIAL INGROUP PERCEPTION

by
Alan T. Pugh and Beverly S. Li

A thesis submitted in partial fulfillment of the requirements
for graduation with Honors in Psychology.

Whitman College
2014

Certificate of Approval

This is to certify that the accompanying thesis by Alan T. Pugh and Beverly S. Li has
been accepted in partial fulfillment of the requirements for graduation with Honors in
Psychology.

________________________
S. Brooke Vick

Whitman College
May 14, 2014

ii

Table of Contents
List of Illustrations...............................................................................................................v
Abstract................................................................................................................................1
Introduction..........................................................................................................................2
Emotions and Cognition..........................................................................................4
Emotions and Motivational States...........................................................................7
Social Categorization...............................................................................................8
Study Overview.....................................................................................................11
Method...............................................................................................................................12
Participants.............................................................................................................13
Pilot Tests..............................................................................................................13
Design and Procedure............................................................................................14
Measures................................................................................................................17
Results................................................................................................................................21
Discussion..........................................................................................................................29
Limitations.............................................................................................................36
Conclusion.............................................................................................................38
References..........................................................................................................................40
Tables.................................................................................................................................46
Appendix A: Collective Self-Esteem Scale.......................................................................48
Appendix B: Modified Differential Emotions Scale.........................................................50
iii

Appendix C: Rapid Categorization Task...........................................................................51
Appendix D: Perceived Attractiveness Task.....................................................................52

iv

List of Illustrations
Table 1: Mood condition x other-race relationship experience factorial analysis of
variance for perceived attractiveness of other-race faces..................................................46
Table 2: Means and simple effects for the interaction between mood and interracial
relationship experience on other-race perceived attractiveness.........................................47

v

Abstract
Positive emotions may have the ability to elicit more inclusive, flexible thinking, but
positive moods can also lead to more stereotyping behavior. We examined whether two
positive mood states, joy and contentment, had different effects on racial ingroup
perception. We hypothesized that joy would elicit a more inclusive framework whereby
participants would perceive ambiguous-/other-race faces as more a part of their racial
ingroup, whereas contentment would promote a more exclusive framework. We induced
White participants (n = 58) into one of three mood conditions (joy, contentment, neutral);
then they rated a series of monoracial and White/Asian biracial faces on a spectrum from
completely White to completely Asian. After inducing participants into a different mood
state, they rated the attractiveness of other-race and same-race faces. Our results did not
support our hypotheses likely because the mood inductions were unsuccessful. However,
those in the contentment condition (relative to the neutral condition) found other-race
faces more attractive. Although this was a significant finding, it was in the opposite
direction of our predictions. Two potential moderators, strength of racial group
identification and interracial dating experience, showed interesting trends but were not
significant. Implications of the mood, interracial dating, and group identification effects
for ingroup bias and racial prejudice are discussed.

1

Differential Effects of Positive Mood States on Racial Ingroup Perception
Most people agree that being happy feels good and may even lead to more
positive interactions with other people. Although research has supported the personal
benefits of experiencing positive emotions (e.g., increased emotional well-being and life
satisfaction), the positive emotion literature has produced ambivalent results in regards to
the effect of positive emotional experiences on intergroup relations. This study seeks to
understand and potentially resolve this discrepancy in the research.
One area of positive emotion research has pointed to the many positive
consequences associated with positive emotions. In her analysis of the adaptive benefits
of emotions, Fredrickson (2001) explained how negative emotions elicit specific action
tendencies associated with the source of the emotional reaction. For instance, the
negative emotion of fear narrows one’s focus, or thought-action repertoire, and results in
a fight-or-flight response. When it comes to positive emotions, however, there is little
research that proposes any specific adaptive benefits for feeling happy; additionally, what
little research exists neglects to differentiate between various states of positive emotion,
instead choosing to lump this state of being into one large category. As a response,
Fredrickson (2001) introduced the broaden-and-build theory of positive emotions, which
states that positive emotions serve to broaden, instead of narrow, thought-action
repertoires and also build long-term personal resources that enable people to better cope
with life stressors.
While negative emotions elicit specific responses that are immediately adaptive
for survival, positive emotions elicit general responses, such as using more broad-minded
coping strategies (e.g., infusing ordinary events with positive meaning), which is

2

correlated with increased resiliency and better emotional well-being in the long-term
(Fredrickson, 2001). It is this broadened thinking, resulting from experiences of positive
emotion, that then builds lasting personal resources and enables better coping with
adversity, which consequently leads to more experiences of positive emotion, creating an
upward spiral toward improved general life satisfaction (Cohn, Fredrickson, Brown,
Mikels, & Conway, 2009). Research in support of this theory has also found that positive
emotions promote a complex understanding of other people in the formation of new
relationships (Waugh & Fredrickson, 2006).
The theory also includes a delineation of different discrete states of positive
emotion, specifically joy, interest, contentment, pride and love, each of which broadens
cognition in unique ways (Fredrickson, 2001). The present study focuses on the
emotions of joy and contentment, because these two emotions in particular (in terms of
high and low arousal) should theoretically result in behavioral variation. Joy elicits the
general action tendencies of exploration, play and creativity while contentment is
correlated with appreciation of current life circumstances and development of dynamic
views of the self and world (Fredrickson, 2001).

Fredrickson and Branigan (2005)

conducted a study that focused on the unique broadening effects of positive emotions
(specifically, joy and contentment). The experimenters induced participants into five
different mood states and discovered that participants in the joy and contentment mood
conditions demonstrated more global biases, as opposed to local biases, on a global-local
visual processing task as well as more action urges on the Twenty Statements Test (TST)
as opposed to those in negative and neutral moods. The global-local task presented
participants with a target image made up of individual shapes arranged to create an

3

overall shape (e.g., three squares arranged in a triangle pattern) and then asked them to
select which other image was most similar to the target image. A global bias on this task
meant that participants would choose an image that took the overall shape of the target
image (e.g., three triangles arranged in a triangle pattern) as being more similar, whereas
a local bias meant that participants would choose an image that arranged the specific
elements of the target image in a different overall pattern (e.g., four squares arranged in a
square pattern) as being more similar. The TST was a task which asked participants to
complete the sentence “I would like to…” on twenty blank lines and was used to measure
quantity of action urges. The authors used these findings as evidence that positive mood
states broaden thought-action repertoires, although no significant distinction was found
between joy and contentment. Other studies using differing methodologies have found
that participants induced to feel an equivalent of joy demonstrate increased cross-cultural
empathy (Nelson, 2009) and reduced own-race bias in face recognition (Johnson &
Fredrickson, 2005).
Emotions and Cognition
One of the important ways in which emotions, both good and bad, can affect our
everyday lives is through the influence that they have on categorization. Isen and
Daubman (1984) demonstrated how positive affect impacted how participants
categorized words and colors. Participants in a positive affect condition exhibited more
flexible thinking than controls in the way they created and used different categories; for
instance, on one task participants tended to perceive generally low-prototypic examples
of a category (e.g., camel) as belonging to that category (e.g., transportation). Later
studies expanded upon this initial research on categorization of things to include the

4

influence of different emotions on social categorization. Bodenhausen, Sheppard, and
Kramer (1994) found that two distinct negative emotions – anger and sadness – had
differential impacts on social information-processing strategies. The authors discovered
that participants induced to feel angry were more likely than those induced to feel sad to
process heuristically, and thus stereotype, in a social perception task and to rely on
heuristic cues in a persuasion situation. From an evolutionary perspective, it makes sense
that people who feel sad are more inclined to use systematic thinking because sadness
indicates a problem in the environment and thus cues a problem-solving response. The
ability to think systematically is not adaptive in situations in which anger is present,
however, because anger often arises in situations of threat which trigger the need for
quick responses, not extensive contemplation of various courses of action.
Experimenters have also considered the impact of positive emotions on social
judgment. Whereas Fredrickson and colleagues have touted the beneficial effects of
positive emotions in relating to others (e.g., cross-cultural empathy, complex
understanding of others, etc), other researchers seem to have discovered a potential
downside to feeling happy. Bodenhausen, Kramer and Susser (1994) found that
participants induced into a state of positive emotion were more likely than those in a
neutral condition to use stereotypic thinking later in evaluating the guilt of other-race
defendants in a courtroom scenario. Apparently, feeling happy can have socially
negative outcomes in an applied setting. The authors posited that stereotypic judgments
were more likely to be rendered by individuals in a positive state of emotion because they
were less inclined to engage in effortful cognitive processing out of a desire to remain
happy. Importantly, the stereotypic thinking among participants in a positive mood was

5

attenuated when the experimenters told the participants that they would be held directly
accountable for their decisions. This instruction was sufficient to snap participants out of
their emotional state and force them to think deeply, instead of relying on heuristics, in
their social judgments. A limitation of this study is that the researchers treated positive
mood as a unitary construct and did not identify particular states of positive emotion.
According to the broaden-and-build theory, the consequences of positive
emotions are universally positive for the individual (e.g., undoing the effects of negative
emotions, broadening cognition, fostering resiliency and personal resources) as well as
for society (e.g., increased cross-cultural empathy, reduced own-race bias). However,
other researchers have determined that positive emotions lead people to render more
stereotypic judgments which may be beneficial for individuals’ preservation of cognitive
resources (Macrae, Milne, & Bodenhausen, 1994), but can be detrimental for society in
situations where social perception plays a key role in how people are treated. How can
these two areas of research be reconciled? The first step is finding areas where more
research needs to be done as well as areas of overlap between the two research domains.
For instance, extant research on the broaden-and-build theory has not differentiated
between joy and contentment but instead has lumped them together to demonstrate that
they broaden cognition more than being in a neutral or negative mood (e.g., Fredrickson
& Branigan, 2005). Differential outcomes between these two positive mood states have
not been explored, which may indicate that positive mood states are more alike than
different negative mood states. This study seeks to explore the theoretical distinctions
between joy and contentment and, furthermore, if these differences result in behavioral
variation.

6

In trying to reconcile the divergent results on the effect of positive emotions on
cognitive processing, we found an area of overlap. Bodenhausen, Kramer et al.’s (1994)
explanation that participants did not want to use effortful thinking in their study because
they were content in their state of being is descriptively similar to Fredrickson’s (2001)
elucidation of the discrete positive state of contentment. According to Fredrickson,
contentment leads people to appreciate their current condition in life and to then integrate
this state of being into the formation of an ever-changing worldview (Fredrickson, 2001).
Interestingly, although the difference was not significant, Fredrickson and Branigan
(2005) found that participants induced to feel joy exhibited a greater global bias
(indicative of broadened cognition) as well as more action urges than did participants
induced to feel contentment. It is possible that this finding indicates a fundamental
difference in the type of thinking that takes place when feeling joyful as opposed to
feeling content. Based on Bodenhausen, Kramer et al.’s (1994) finding that positive
emotions (i.e., contentment) result in stereotypical thinking and Fredrickson’s (2001)
differing conceptions of joy and contentment, we propose that the two mood states are
phenomenologically distinct and should impact cognition in different ways.
Emotions and Motivational States
In determining how joy and contentment might elicit distinct cognitive states, it is
important to discuss the relationship between moods and motivational states on
subsequent behavior. It is clear from psychological research that emotions influence
behavior (Bodenhausen, Sheppard et al., 1994; Bodenhausen, Kramer et al., 1994; Isen &
Daubman, 1984) but Brockner and Higgins (2001) demonstrated that emotions are
connected to motivational states (and thus future behavior) and that emotional regulation

7

impacts the focus, nature, and strength of the emotional experience. These authors
showed that those who are promotion-focused in their motivation try to become their
ideal selves by thinking about opportunities and what they can gain, heightening the
salience of the (potential) positive outcomes. On the other hand, those who are
prevention-focused are motivated to become their “ought” selves by thinking about their
current state and what they can lose, heightening the salience of (potential) negative
outcomes.
In terms of future behavior, promotion-focused people are approach-oriented
while prevention-focused people are avoidant-oriented (Shah, Brazy, & Higgins, 2004).
Shah et al. determined that people in a promotion-focused framework were likely to feel
more cheerful and less dejected toward ingroup members while people in a preventionfocused framework were likely to feel less relaxed and more agitated toward outgroup
members. Furthermore, those in the promotion framing condition were more likely to
engage in approach-related behaviors toward ingroup members and nonavoidance
behaviors toward outgroup members than participants in the prevention framing
condition. Overall, people in a prevention-focused frame of mind are less inclined to
actively engage with their environment and feel more uncomfortable toward outgroups.
The creative and explorative aspects of Fredrickson’s (2001) conception of joy led us to
contend that joy causes people to become more promotion-focused while the passive,
serene aspects of contentment causes people to become more prevention-focused and
thus more motivated to maintain their current cognitive state.
Social Categorization
Considering the breadth of research on emotions and social judgment, we wanted

8

to explore the differences between joy and contentment in terms of ingroup perception,
particularly racial ingroup perception. There are a number of factors, including social
categorization, that contribute to the establishment of an ingroup. Ingroup membership
can often be based on arbitrary characteristics as shown by the minimal group paradigm
(Tajfel, Flament, Billig, & Bundy, 1971). Tajfel and colleagues (1971) randomly divided
participants into two groups based on supposed preference of a painting that was in
actuality decided by a coin toss. The participants then completed a distribution task,
where they were asked to distribute valuable resources (e.g., money or points) to other
participants who were only identified by a code number and group membership. These
researchers demonstrated that while participants showed a considerable amount of
fairness, they also showed a significant preference for their arbitrarily assigned ingroup
members by allocating more coins to them. This study showed how easily ingroups are
formed and maintained.
Within ingroups, members often show ingroup favoritism, a preference for
ingroup members over outgroup members (Aberson, Healy, & Romero, 2000). Research
has demonstrated how ingroup favoritism can negatively affect outgroup members when
certain individuals are allocated more resources or evaluated differently (Tajfel et al.,
1971), particularly in the case of race relations (Brewer, 1999). Given the significant
impact of racial prejudice and discrimination in society, we wanted to explore how
different emotions might affect racial ingroup perception because this seemed to be a
socially meaningful way to measure differences in cognition.
The other-race effect (ORE), which states that same-race faces are more easily
recognized than other-race faces (Anzures et al., 2013), is an ingroup process that has a

9

direct bearing on racial perception. The ORE refers to an ingroup familiarity bias,
whereby one can recognize ingroup members more easily and faster than outgroup
members, even if people are unfamiliar or new (Beaupré & Hess, 2006). In evolutionary
terms, this can be thought of as an ingroup advantage since being able to quickly and
accurately determine who belongs to one’s ingroup guards against improper allocation of
resources to outgroup members. Thus, individuals who are of the same race should be
more easily determined to be part of one’s ingroup than those who are not of the same
race. As a result of this ingroup familiarity, it should be more difficult for monorace
people to assess the group membership of ambiguous-race faces because the faces cannot
be easily included nor easily excluded from their ingroup.
Determining ingroup member status for racially ambiguous individuals is also
influenced by a number of factors, such as whether ambiguous-race faces are presented
with a specific monoracial label (Hourihan, Fraundorf, & Benjamin, 2013), or in a
specific context (e.g., a Black/White biracial face seen with mostly White faces will be
considered Black but seen as less Black in a Black only context) (Ito, Willadsen-Jensen,
Kaye, & Park, 2011), or if the evaluator is presented with a race stereotype prior to
making judgments (e.g., negative racial primes impede the ability to quickly categorize
racially ambiguous faces compared to positive primes) (Dickter & Kittel, 2012). From
these studies, we know that we do not categorize ambiguous-race faces easily or
automatically and as a result, other factors (such as emotions) have the potential to
impact our perceptions. Since we know that environmental factors influence ingroup
perception, we wanted to explore whether environmentally manipulated mood states can
influence the malleability of ingroups as well.

10

Study Overview
We wanted to know whether environmentally induced mood states differentially
affect perceptions of one’s racial ingroup. To accomplish this, we randomly assigned
participants into three different mood groups and then administered two different tasks to
assess perceptions of ambiguous- and other-race faces in terms of their closeness to
participants’ own racial ingroups. On the first task, in which participants rated
ambiguous-race (Asian/White biracial) faces on a spectrum from completely White to
completely Asian, we predicted that participants in the joy condition would rate the faces
as closer to their racial ingroup (relative to neutral) and that participants in the
contentment condition would rate the faces as further away from their racial ingroup
(relative to neutral).
We predicted these different outcomes in terms of the distinct motivational states
elicited by these two emotions. Specifically, those who feel joy should be more
motivated to actively pursue their ideals, which (to the extent that people value racial
equality and inclusion) would prompt them to be more inclusive than exclusive in their
ingroup category conceptions when presented with ambiguous-race faces (Brockner &
Higgins, 2001; Fredrickson & Branigan, 2005). In a contentment condition, however, in
which people are satisfied with their current life circumstances, they would be more
likely to evaluate ambiguous-race faces as part of an outgroup because they are
prevention-focused and thus more concerned about losing something that they have (i.e.,
happiness) than in changing their social category conceptions (Brockner & Higgins,
2001). They are likely to notice certain features that are dissimilar to their own and
automatically view those faces as not belonging to their racial ingroup.

11

On the second task, in which participants rated the attractiveness of other-race
faces, we predicted that participants in the joy condition would rate the faces as more
attractive (relative to neutral) while participants in the contentment condition would rate
the faces as less attractive (relative to neutral). This second task was a way to further test
our main hypothesis that joy elicits an inclusive ingroup perception while contentment
elicits an exclusive one. An important quality of ingroup membership is that ingroup
members are afforded certain benefits, such as empathy (Gutsell & Inzlicht, 2012),
recognizability (Anzures et al., 2013) and perceived attractiveness (Folkes, 1982).
Furthermore, an expansion of ingroup membership to others who are not automatically
considered a part of that group would mean sharing those benefits with them. This would
mean that evaluating people of another race as attractive would be an indicator that they
were viewed more as members of the evaluator’s ingroup since we know that people who
are perceived to be similar to ourselves are also seen as more attractive (Folkes, 1982).
Thus, perceived attractiveness of other-race faces should reflect a more inclusive ingroup
mindset.
Insofar as joy ought to elicit a promotion-focused framework, which is associated
with feeling more relaxed and less agitated toward outgroup members (Shah et al., 2004),
people who experience joy should find outgroup members more attractive than people
who experience contentment. Those who feel content, on the other hand, should find
outgroup members less attractive than those who feel joyful because these participants
are more likely to be prevention-focused, which is associated with feeling less relaxed
and more agitated toward outgroup members (Shah et al., 2004).
Method

12

Participants
We recruited 58 White participants (29 men and 29 women) from a small, liberal
arts college in the Pacific Northwest. Their ages ranged from 18 - 22 (M = 19.97, SD =
1.24). Among the participants, 31 percent (n = 18) were first-years, 25.9 percent (n = 15)
were sophomores, 17.2 percent (n = 10) were juniors, and 25.9 percent (n = 15) were
seniors. Additionally, 16 participants reported having dated people of another race than
their own. We randomly assigned participants into one of three different experimental
groups (joy, contentment, and neutral) per dependent variable with roughly 20
participants per cell.
Pilot Tests
We conducted two pilot tests to determine 1) which videos to use in order to
induce participants into the desired mood states, and 2) which faces to use based on the
extent to which they were found to be average in attractiveness and near the biracial
location on a White-to-Asian racial spectrum scale. We collected responses from 30
people (15 different people for each pilot test) who reflected the sample population of our
study in terms of age and education. None of the pilot test participants took part in the
actual study. The participants filled out the modified Differential Emotions Scale
(mDES) for each video they watched, rating the intensity of each emotion item on a 5point scale in which 0 = “not at all” and 4 = “extremely”. Of the 12 videos that we pilot
tested, we decided to use a video of red pandas for joy (M joy = 2.64, M contentment =
1.07), a video of nature scenes for contentment (M contentment = 2.77, M joy = 0.97) and
a video of birds eating at a feeder for neutral (M total = 0.50). The neutral video that we
selected was determined to elicit the least amount of emotions (either positive or

13

negative) of any video. We also pilot tested 67 different White and Asian faces and then
deleted 15 to make a total of 52 faces. The six ambiguous-race faces that we excluded
were rated as furthest away from the center on the racial spectrum that ranged from 1
(completely White) to 7 (completely Asian) (average difference from the center for the
six excluded faces = 1.31). The nine monorace faces that we excluded were rated as the
most unattractive on a scale from 1 (not at all attractive) to 7 (extremely attractive) (M
attractiveness of excluded faces = 2.48, M attractiveness of included faces = 2.83).
Design and Procedure
We used a one-way between-groups experimental design. Our independent
variable was mood with three different levels: joy, contentment and neutral. We
measured two dependent variables. The first was racial ingroup inclusion ratings of
ambiguous-race faces and the second was attractiveness ratings of other-race faces.
Experience with interracial romantic relationships and strength of racial ingroup
identification were primary moderators. The dependent variable tasks and the emotion
videos were counterbalanced across trials.
Prior to the study, participants completed a pre-screening survey that was sent out
to student listservs and students in psychology courses that signed up for our study that
asked basic demographic questions (i.e., age, race, class year, major, hometown) in
addition to some filler questions to de-emphasize the importance of the racial portion of
this pre-screening process. For example, one of the filler questions was “How often do
you watch YouTube videos?” We then screened for individuals who self-identified as
Asian American or White/Caucasian, as these were the racial groups we were interested
in including in our study.

14

Participants arrived at the lab in groups of one to four people. They sat at a table
with a desktop computer in front of them. Each station was separated by a divider so
participants could not see each other’s screens. One experimenter, either a White man or
an Asian American woman, welcomed the participants to the lab and instructed them to
read the electronic informed consent form and let them know they were free to ask any
questions. The effect of the experimenter’s race was not considered because the presence
of varying numbers of participants during any given experimental session and the fact
that sometimes White and Asian participants were tested together meant that the effect of
the experimenter’s race could not be isolated. Nevertheless, it is certainly possible that
social desirability effects could have played a role when White participants were in the
presence of the Asian female experimenter, but since the tasks were not explicitly about
evaluating racial attitudes, we are dubious about this possibility.

Participants were told

on the informed consent form that the study was about media and its impact on decisionmaking and judgment. After participants gave us their consent electronically, they were
then presented with a demographics questionnaire in which we were specifically
interested in the race with which they identified. We also included a Collective SelfEsteem Scale (CSES) out of concern that people who strongly identify with their own
race would be more likely to have a restricted view of racial ingroup inclusion, regardless
of their mood. The CSES assesses strength of racial identity. We presented the CSES at
the beginning of the study because we wanted to make participants reflect on their own
race so that later during the face judgment tasks, participants would rate the faces
according to the perceived proximity to their own personal racial ingroup. The
participants were allowed an unlimited amount of time to complete these items. Upon

15

completion, the participants were randomly assigned into one of three mood manipulation
conditions: joy, contentment, or neutral. Those in the joy condition were subsequently
presented with a short video of playing red pandas; those in the contentment condition
were presented with a short video of nature scenes; and those in the neutral condition
were presented with a short video of birds eating.
After the mood manipulation, the participants were instructed to complete a rapid
categorization task in which participants rated racially ambiguous faces (in addition to
some same-race and other-race faces as controls) on a computer, using a numerical scale
from 1 (completely White) to 7 (completely Asian). The experimenters told the
participants that they must try to complete the categorization task quickly on the pretense
that completion time was factored into their overall scores on the task. After completing
the task, participants filled out the modified Differential Emotions Scale (mDES) as a
manipulation check to determine what emotions they were feeling when they were
watching the video. This scale ensured that the mood manipulation video actually
induced the participants into the desired mood. The mDES was administered after the
task because being asked to reflect on one’s emotional state would have potentially
reduced the capacity to actually feel the emotion. Furthermore, we wanted as little delay
time as possible between the mood manipulation and the task due to the fact that
emotions tend to fade rapidly. The mDES ended the first portion of the study.
We were concerned about the transitory effects of the mood inductions and that
the mood induction would no longer be present by the second task. We resolved this
issue by re-inducing participants into new mood states after a five-minute break. They
were told that the next component of the study involved cognitive tasks. Participants

16

then completed two filler tasks during this five-minute break to ensure that any lingering
effects from the first mood induction were mitigated. The first task was decoding
anagrams, and the second one was naming as many U.S. states as possible with the aid of
a blank map. Each task lasted two and a half minutes. After the filler tasks, participants
were shown one of the three mood manipulation videos (joy, contentment, or neutral)
from the first part of the study. We ensured that participants saw a different video than
the one they saw at the beginning of the study.
Participants next looked at a series of other-race faces and same-race faces and
rated the attractiveness of the faces on a computer, using a numerical scale from 1 (not at
all attractive) to 7 (extremely attractive). The experimenter told participants that they
must try to complete the perceived attractiveness task quickly on the pretense that
completion time was factored into their overall scores on the task. We asked the
participants to complete both tasks quickly because we wanted to encourage subjective,
automatic processing as opposed to more cerebral, objective processing. This instruction
was consistent with our intent to make the face judgments more personal and relevant to
the participants since we wanted to assess personal ingroup perceptions. The participants
then completed the mDES as they did in the first half of the study. Completing this scale
ended the study. The participants were debriefed, asked to sign a data release form,
compensated, and thanked.
Measures
Demographics. Participants were asked to report on basic personal information
regarding age, race, education, and visual ability. We were primarily interested in their
self-reported racial identity as well as whether they had ever dated a person of another

17

race before. Participants answered either yes or no when answering the question if they
had ever dated someone of another race. Participants’ experience of having intimate
relationships with people of another race was a potential moderator because it may have
indicated that their ingroup was already more inclusive to begin with and/or that their
ingroup was more flexible due to this experience.
Collective Self-Esteem Scale (Luhtanen & Crocker, 1992). The Collective
Self-Esteem Scale (CSES) measures individual’s social identity based on membership to
ascribed groups (e.g., ethnicity, gender, etc.), and we wanted to focus on racial identity.
Measuring participants’ collective self-esteem demonstrated how they felt about and to
what extent they identified with their racial group. Those high in collective self-esteem
are more likely to engage in group-serving biases to maintain and protect their group
image as a function of ingroup favoritism (Tajfel & Turner, 1986). It is possible, then,
that the stronger this identification, the less susceptible people are to outside influences
(i.e., emotions) in their perception of stimuli relevant to this social category. In other
words, their personally held beliefs would overpower experimentally manipulated
conditions. It is important to keep in mind that strong ingroup identification is not
equivalent to negative views of outgroups (Brewer, 1999). Thus, participants in our
study who strongly identified with their race were not necessarily racially biased in the
sense that they held prejudicial views of racial outgroups.
The CSES originally had four subscales (Membership, Private, Public, and
Identity) with four items per scale, but we were primarily interested in the Identity
subscale (α = .76) to measure how much our participants identified with their race. We
presented participants with a modified version of the CSES scale in which they saw a

18

total of 10 (out of 16) item measures. All four item measures from the Identity subscale
were used, and two each from the other three subscales were included in this modified
CSES. Only the items from the Identity subscale were analyzed, but we wanted to
include the other items to disguise our main focus. The Identity subscale included items
like, “The racial group I belong to is unimportant to my sense of what kind of a person I
am” and “In general, belonging to my race is an important part of my self-image.”
Participants rated their responses on a 7-point scale (1 = strongly disagree, 7 = strongly
agree). See Appendix A.
Videos (emotion prime). We pilot tested short videos which were intended to
induce a particular mood state, either joy, contentment, or neutral. Joy was induced
through a 90 s video of two baby red pandas frolicking in the snow to the tune of fun,
upbeat music. We induced feelings of contentment with a 90 s nature video which
showed slow pans of vast landscapes from around the world with calming music in the
background. Lastly, we produced a neutral mood with a 3 min video of birds eating at a
bird feeder with no sounds, other than the birds chirping, in the background. The
participants saw two of the three videos during the duration of their participation, one
before the rapid categorization task (see below), and one before the perceived
attractiveness task (see below).
Modified Differential Emotions Scale (mDES). This questionnaire was a
manipulation check used to determine what types of emotions were elicited while
watching the mood manipulation videos. This scale is effective at capturing the temporal
frame of a study by focusing on a particular time when the participant was experiencing
certain emotions. The questionnaire asks about peak emotional experiences (“greatest

19

amount you’ve experienced each of the following feelings”) because findings indicate
that people are better at recalling peak experiences rather than determining an average
level of emotion across different time periods (Fredrickson & Kahneman, 1993, as cited
in Fredrickson, in press). The scale has 20 items with emotions presented in groups of
threes. We only presented 15 of the 20 possible items to prevent any potential fatigue
effects. Participants in our study rated the intensity with which they experienced each
emotion triad while watching the video, using a numerical scale in which 0 = “not at all”
and 4 = “extremely”. For example, the first question read, “What is the most amused,
fun-loving or silly you felt?”
We also created three subscales for joy, contentment and neutral based on specific
items on the mDES. The joy subscale was formed from two items on the scale which
assessed feelings of joy and amusement (r = .68, p < .001). The contentment subscale
was formed from two items on the scale which assessed feelings of awe and serenity (r =
.39, p = .002). Lastly, the neutral subscale was formed by averaging the sum of all 15
items (α = .84). An item that measured joy was, “What is the most joyful, glad, or happy
you felt?” and an item that measured contentment was, “What is the most serene, content,
or peaceful you felt?” See Appendix B.
Rapid categorization task. The participants were instructed that they were about
to be presented with faces that they must categorize on a scale from 1 (completely White)
to 7 (completely Asian). Rapid categorization of racially ambiguous faces has precedent
in the literature (Hugenberg & Bodenhausen, 2004; Peery & Bodenhausen, 2008), but
participants in previous studies engaged in dual-categorization, meaning that they had to
decide between two options (e.g., Black/not White or White/not Black). Based on the

20

finding that as faces became increasingly ambiguous, participants tended to rate them as
belonging to the opposite race category (i.e., a biracial White/Asian face being
categorized as Asian) (Freeman, Pauker, Apfelbaum, & Ambady, 2010), we decided to
have participants rate the faces on a scale of ambiguity to measure to what degree
participants thought a face belonged to a certain racial category.
Participants saw a total of 20 different ambiguous-race faces, as well as 6
monoracial faces of each race as a control, for the duration of this task. The monorace
White faces were drawn from the Productive Aging Lab Face Database developed by
Minear and Park (2004) and the monorace Asian faces were drawn from CAS-PEAL
Face Database developed by Gao et al. (2008). We mixed the various monorace faces to
create 50% Asian, 50% White mixed-race faces using FantaMorph technology (Version
5; Abrosoft Co., Beijing, China). Faces were presented in the middle of the screen, in
black-and-white photos, and remained on the screen until the participants made a
decision. See Appendix C.
Perceived attractiveness task. Participants were presented with a series of
images of same-race and other-race monoracial faces (i.e., completely Asian faces and
completely White faces). They saw 10 White faces and 10 Asian faces that were
monoracial versions of the morphed faces used in the rapid categorization task. They
then rated attractiveness on a scale from 1 (not at all attractive) to 7 (extremely
attractive). Similar to the rapid categorization task, the faces were presented in the
middle of the screen, in black-and-white photos, and remained on screen until the
participants made a decision. See Appendix D.
Preliminary Analyses

Results

21

We collected responses from both White and Asian participants, but only included
analyses of the White participants because the number of Asian participants was very
small (n = 12) and they were unevenly distributed among the conditions. Of the 60
White participants, two were excluded from the analyses because they expressed high
levels of suspicion and atypical reactions to the experimental stimuli.
We used a factorial ANOVA (mood condition X sex) across the dependent
variables and found nonsignificant sex differences for the categorization of ambiguousrace faces, F(1, 52) = 1.76, p = .19 and for perceived attractiveness of Asian faces, F(1,
52) = 0.44, p = .51.
We randomly assigned participants to one of two different dependent variable
order groups, one of which experienced the rapid categorization task followed by the
perceived attractiveness task and vice versa. We also randomly assigned participants to
one of six different mood video order groups, which presented two of the three possible
mood videos in every possible order. We ran a one-way between-groups ANOVA to test
for order effects of the two dependent variables as well as for the three mood videos. We
found no significant dependent variable order effects on ratings of perceived
attractiveness of Asian faces, F(1, 57) = 0.84, p = .37 or on ratings of ambiguous-race
face categorization, F(1, 57) = 0.00, p = .97. We also did not find mood video order
effects on perceived attractiveness of Asian faces, F(5, 57) = 0.26, p = .94 or on rapid
categorization of ambiguous-race faces, F(5, 57) = 0.25, p = .94.
Manipulation Checks
We used the three subscales for joy, contentment and neutral to assess the
effectiveness of the mood videos in capturing the intended emotional state. As a side

22

note, the participants saw a scale from 0 to 4, but responses were recorded and analyzed
on a 1 to 5 scale. The mDES does not include any specific items that assess neutral mood
but one large-scale validation study (Schaefer, Nils, Sanchez, & Philippot, 2010)
determined the neutrality of film clips by ensuring that they elicited the least amount of
arousal and the least amount of positive and negative affect relative to other emotioninducing film clips. Therefore, we defined neutrality as the absence of emotion, either
positive or negative, which is why we computed a composite of each of the 15 emotion
items on our modified DES and determined that the closer the score was to 1 on the scale
(i.e., no emotional response), the more neutral the score.
We conducted one-sample t tests to determine if the mood videos elicited
particular emotions that were significantly greater than the midpoint (i.e., 3) on the 5point mDES scale. These tests determined that the contentment video elicited an average
contentment score (M = 2.82, SD = 0.72) that was not significantly different from the
midpoint on the scale, t(57) = -1.86, p = .07, while the joy video elicited an average joy
score (M = 2.50, SD = 0.71) that was significantly less than the midpoint on the scale,
t(57) = -5.29, p < .001. Nevertheless, these positive emotions scores were significantly
greater than the minimum point (i.e., 1) on the scale, t(57) = 16.05, p < .001 for joy, t(57)
= 19.18, p < .001 for contentment. Additionally, the neutral video elicited emotional
responses that were significantly greater (M = 2.02, SD = 0.40) than the minimum point
on the scale, t(57) = 19.29, p < .001.
We conducted three one-way between-group ANOVAs to test the effect of each
mood video on mDES joy scores, contentment scores and neutral scores for the mood
videos presented before the rapid categorization task. Unexpectedly, there was no

23

significant effect of the mood videos on mDES joy scores, F(2, 57) = 1.08, p = .35.
Those who watched the joy video (M = 2.75, SD = 1.30) did not feel more joyous than
those who watched the contentment video (M = 2.72, SD = 0.89), p = .92, or those who
watched the neutral mood video (M = 1.96, SD = 0.52), p = .34. Contrary to our
predictions, there was also no significant effect of the mood videos on mDES
contentment scores, F(2, 57) = 2.50, p = .12. Those who watched the contentment video
(M = 3.34, SD = 0.91) did not feel significantly more content than those who watched the
joy video (M = 2.61, SD = 1.00), p = .12, or those who watched the neutral video (M =
2.24, SD = 0.47), p = .27, though the means were trending in the right direction.
Unfortunately, the positive mood manipulating videos presented before the rapid
categorization task were not effective in inducing participants to feel distinct positive
emotions.
We then conducted another three one-way between-group ANOVAs to test the
effect of each mood video on mDES joy scores, contentment scores and neutral scores for
the mood videos presented before the other-race perceived attractiveness task. There was
no significant effect of the mood videos on mDES neutral scores, F(2, 57) = 1.78, p =
.18. Participants who watched the neutral video (M = 1.83, SD = 0.42) did not feel more
neutral than those in the joy condition (M = 2.30, SD = 0.91) or those in the contentment
condition (M = 2.75, SD = 0.79), although the means were trending in the right direction.
Levene’s statistic for the Test of Homogeneity of Variances showed that the averages for
the feelings of joy and contentment on the second mDES were significant (p = 0.04 and p
< .001, respectively) and the assumption of homogeneity was not met. Kruskal-Wallis
tests were used instead of ANOVAs, and they showed no effect of the joy and

24

contentment videos on feelings of joy and contentment (p = .15 and p = .77,
respectively). The joy video did not elicit significantly more feelings of joy than the
other mood videos and the contentment video did not elicit more feelings of contentment
than the other mood videos. Ultimately, the positive mood manipulating videos
presented before the other-race perceived attractiveness task were also not effective in
inducing participants to feel distinct positive emotions. We considered separating the
different subscale responses and comparing those who scored high on either positive
emotion to those who scored low, regardless of the mood video groups. However, we
determined that there was not enough variation in the subscale scores (likely due to the
small number of items which made up the subscales) to meaningfully separate
participants based on their scores.
Rapid Categorization Task
Although we determined that the mood induction videos were not especially
effective at eliciting distinctive mood states, we still ran analyses based on these different
mood conditions since some of the emotion ratings were trending in the right direction.
Additionally, there could have been differences between the mood conditions that we
were not able to assess or even differences in emotion that the mDES failed to capture.
We used a one-way between groups ANOVA to measure the effect of the mood
conditions on people’s tendency to perceive ambiguous-race faces as more a part of their
racial ingroup by rating them as more White. We expected that those in the joy condition
would rate ambiguous-race faces as closer to the “White” end of the racial spectrum on
the scale from 1 (completely White) to 7 (completely Asian). We also expected those in
the contentment condition would rate ambiguous-race faces as closer to the “Asian” end

25

of the racial spectrum. On average, participants rated the White faces as White (M =
1.57, SD = 0.46), Asian faces as Asian (M = 6.03, SD = 0.50), and ambiguous-race faces
as fairly biracial (M = 3.80, SD = 0.62). Contrary to our predictions, there was no
difference between mood conditions in participants’ responses on their perception of
ambiguous-race faces, F(2, 55) = 0.34, p = .71, η2 = .01. Those who watched the joy
video (M = 3.71, SD = 0.71) did not differ in their perception of ambiguous-race faces
from those who watched the contentment video (M = 3.79, SD = 0.73), p = .91, or those
who watched the neutral video (M = 3.88, SD = 0.43), p = .69. Those who watched the
contentment video (M = 3.79, SD = 0.73) also did not differ significantly in their
responses compared to those in the neutral mood condition (M = 3.88, SD = 0.43), p =
.91.
Perceived Attractiveness Task
A one-way between groups ANOVA was used to measure the effect of the mood
conditions on perceptions of other-race face attractiveness, a measure that is theoretically
linked to racial ingroup perception. We expected those in the joy condition would rate
other-race faces as more attractive than those who in the neutral condition. We also
predicted that that those in the contentment condition would rate other-race faces as being
less attractive than those who felt neutral. On average, participants rated White faces as
average in attractiveness (M = 3.92, SD = 0.63) and Asian faces as average in
attractiveness as well (M = 3.79, SD = 0.70). There was a significant difference between
mood conditions in participants’ perceptions of other-race faces’ attractiveness, F(2, 55)
= 3.52, p = .04, η2 = 0.11. Tukey post-hoc comparisons between the three video groups
showed that participants who watched the contentment video rated other-race faces as

26

significantly more attractive (M = 4.10, SD = 0.68) than those who watched the neutral
mood video (M = 3.52, SD = 0.83), p = .03. However, there was no significant difference
in attractiveness ratings when comparing those in the contentment condition (M = 4.10,
SD = 0.68) with those in the joy condition (M = 3.80, SD = 0.45), p = .35. Comparisons
between those in the joy and neutral conditions also did not reveal a significant difference
in attractiveness ratings, p = .41. This indicated that those who watched the contentment
video perceived other-race faces as more attractive than those who watched the neutral
video but not compared to those who watched the joy video.
Considering the significant effect of contentment on other-race perceived
attractiveness (compared to neutral), we also ran a one-way between groups ANOVA to
measure the effect of the different mood conditions on same-race (White) perceived
attractiveness. In terms of perceived attractiveness of White faces, participants in the
various mood conditions did not differ in their scores, F(2, 55) = 1.69, p = .19.
Participants in the contentment condition (M = 4.14, SD = 0.46) did not find same-race
faces significantly more attractive than participants in the joy condition (M = 3.83, SD =
0.57) or participants in the neutral condition (M = 3.82, SD = 0.77). We then ran a paired
samples t test to test for a significant difference between average other-race perceived
attractiveness scores (M = 3.79, SD = 0.70) and same-race perceived attractiveness scores
(M = 3.92, SD = 0.63) and found no significant difference, t(57) = 1.30, p = .20.
Participants did not rate same-race faces as significantly more attractive than other-race
faces.
Moderating Variables

27

In addition, we tested for the effects of two potentially moderating variables:
interracial dating experience and racial ingroup identity strength. We ran an independent
samples t test and found no significant effects of having been in a personally significant
relationship with a person of a different race on the perceived attractiveness of Asian
faces, t(56) = 0.96, p =.34, nor on the categorization of ambiguous-race faces, t(56) =
0.73, p =.47. However, we then ran a 2 (interracial relationship experience: yes vs. no) x
3 (mood: joy, contentment, neutral) ANOVA to test whether relationship experience
interacted with the mood conditions in terms of perceived attractiveness. As shown in
Table 1, we found a marginal interaction between relationship experience and the mood
conditions on the perceived attractiveness of Asian faces measure, F(2, 52) = 2.70, p =
.08. Pairwise comparisons reveal that among those participants who had interracial
relationship experience, those in the joy condition rated Asian faces as significantly more
attractive (M = 4.07, SD = 0.35) than those in the neutral condition (M = 3.08, SD =
1.22), p = .02, and those in the contentment condition also rated Asian faces as
significantly more attractive (M = 4.44, SD = 0.43) than those in the neutral mood
condition, p = .003. However, it is important to recognize that cell totals for the three
mood groups among the participants who had interracial relationship experience were
quite low (joy = 7, contentment = 5, neutral = 4). See Table 2 for comparisons of means
and simple effects.
In terms of racial ingroup identity strength (M = 3.39, SD = 1.20), we did find one
significant relationship. We found a significant negative correlation between ratings on
the Identity subscale of the Collective Self-Esteem Scale (CSES) and the perceived
attractiveness of Asian faces, r = -.34, p = .01. Regression analysis indicated that group

28

identity ratings significantly predicted perceived attractiveness scores, t(56) = -2.69, p =
.01, whereby higher group identification was related to lower perceived attractiveness
ratings of other-race (Asian) faces. Analyses also indicated that 11.4 percent of the
variance in the perceived attractiveness of Asian faces could be explained by group
identity ratings. We did not find any significant correlations between group identification
and ratings of ambiguous-race faces.
We used a hierarchical linear regression to test for the effect of the mood
conditions and group identification on perceived attractiveness to see if group
identification moderated the effect of the mood conditions on perceptions of
attractiveness of Asian faces. Group identification and dummy codes for the joy and
neutral mood conditions were entered in Step 1, and the interactions between group
identification and mood were entered into Step 2. The overall model accounted for 29.1
percent of variance for the perceived attractiveness task, but the overall interactions were
not significant as they accounted for only 6.3 percent of the variance, R² = 0.29, F(2, 52)
= 2.31, p = .11. Group identification did not interact with the mood conditions to affect
participants’ perceptions of Asian faces’ attractiveness overall.
Discussion
Our study aimed to examine the differential effects of specific positive moods on
racial ingroup perception. We expected that participants induced to feel joy would
engage in more flexible thinking and see ambiguous-race faces and other-race faces as
more a part of their racial ingroup than those in a neutral mood state, while those induced
to feel contentment would engage in less flexible thinking, and be more exclusive in their

29

racial ingroup perception. We found that our results did not support our main
hypotheses, but there were some interesting trends.
Rapid Categorization Task
We determined that there were no differences in perceptions of ambiguous-race
face across the three mood conditions, meaning that there was no significant effect of
mood on racial perception. Although it is possible that our moods have no bearing on
how we perceive race, we think it is equally possible that the specific limitations of our
experimental study contributed to this result. A number of participants commented on
the fake or bizarre nature of some of the biracial faces in the rapid categorization task due
to particular effects of the morphing technology. A few went so far as to say that they
were more likely to think a face was completely biracial if they noticed this strange
quality about it. Additionally, in the interest of reducing any fatigue effects, we
presented participants with only 20 ambiguous-race faces instead of 30, a number which
has precedent in the literature (Ito et al., 2011). However, the chief issue with this
dependent measure likely has to do with the mood videos that were presented before the
task in an attempt to induce specific positive mood states. Our analyses demonstrated
that neither joy, contentment, nor neutral scores significantly differed among participants
in each of the mood conditions, potentially meaning that the mood videos did not
effectively put participants into a particular mood state more than another. The fact that
we cannot confidently contend that the independent variable was sufficiently manipulated
calls into question the validity of the results of the rapid categorization task.
Perceived Attractiveness Task

30

Our analyses indicated that there was a significant difference in perceived
attractiveness of other-race (Asian) faces between participants in the contentment
condition and participants in the neutral condition, whereby those in the contentment
condition rated Asian faces as significantly more attractive than those in the neutral
condition. This difference was in the opposite direction of our predictions but it is still
interesting to note that some quality of the different mood conditions affected perceived
attractiveness. Cullen and Newell (2012) studied the role of cross-sensory information
on the perceived attractiveness of static face images and found that emotional content,
especially content positive in nature (e.g., laughter), that was presented along with a face
image positively influenced perceived attractiveness of the face. Based on this research,
perhaps the contentment video was better than the joy video at eliciting the particular
positive mood state that has been shown to influence perceived attractiveness.
Additionally, we found that the different mood groups differentially affected
participants’ perceptions of same-race and other-race face attractiveness. Analyses
indicated that participants in the contentment condition found other-race, but not samerace, faces significantly more attractive than those in the neutral condition. In other
words, the contentment condition affected participants’ perceptions of other-race faces’
attractiveness, but did not affect their perceptions of same-race faces. Considering the
fact that overall attractiveness ratings did not differ between other-race and same-race
faces, we surmise that this result must have something to do with the effect of the mood
video and not an effect of the faces. Although we know that people perceive similar
levels of variation in attractiveness across racial groups (Bernstein, Lin, & McClellan,
1982), it is unclear if the way in which people perceive attractiveness is different when

31

evaluating members of one’s racial ingroup versus evaluating members of racial
outgroups. It is possible that perceptions of the attractiveness of people of our own race
are more fixed, but when it comes to people of another race, our perceptions of their
attractiveness are more malleable and can be influenced by mood effects. This is an area
of research that future studies could work to address.
Importantly, however, our results seemed to indicate that none of the mood
manipulation videos effectively manipulated mood, meaning that people who saw the joy
video did not feel significantly more joyous than people in the other conditions and
people who saw the contentment video did not feel significantly more content than
people in the other conditions (although these means were in the right direction). The
same pattern was true for neutral scores on the neutral video. The fact that a significant
difference was found between two mood groups (contentment and neutral) that did not
effectively manipulate participant mood may indicate that other factors could be
accounting for this difference. One factor that could have contributed to this difference
was the type of music that accompanied the contentment video. A number of participants
remarked afterwards that the music in the contentment video was distinctly “Asian”
sounding. North (2012) showed that auditory stimuli can influence our perception of
taste, so there is a possibility that auditory stimuli can influence the perception of our
other senses. Thus, it is possible that when participants were presented with other-race
(Asian) faces, they made an easier connection back to that video due to the music and it
was this aspect of the video (not so much the mood) that influenced perceived
attractiveness responses. Although it remains unclear what exactly differentiated the
contentment and neutral mood groups if it was not mood (although there could be a

32

difference in mood that the mDES failed to capture), analyses on two potential
moderating variables (see below) indicated that other factors could easily be involved.
Moderating Variables
We tested for the effect of interracial dating experience on the two dependent
variables. We found no main effect of interracial relationship experience for either
dependent variable but we did find a marginally significant interaction effect of
relationship experience and the mood conditions on perceived attractiveness of other-race
faces. Specifically, we found that among participants who claimed to have had an
interracial relationship experience, those in the joy and contentment conditions found
other-race faces significantly more attractive than those in the neutral condition.
Assuming that participants in the positive mood conditions felt more positive emotions
than those in the neutral condition (and the numbers were trending in this direction), this
finding is interesting because it indicates that having intimate interracial experiences has
the potential to alter perceived attractiveness of other-race individuals, but only if in a
positive mood.
This finding is consistent with research that found that college interracial dating
experience predicted lower ingroup bias ratings by the end of their fourth year of college
(Levin, Taylor, & Caudle, 2007). Ingroup bias was measured on a 7-point affect scale
which asked about participants’ feelings toward various racial/ethnic groups from 1 (very
negatively) to 7 (very positively). Participants who had interracial dating experience
reported lower ingroup bias ratings, meaning that they felt more positively toward otherrace groups than did people who did not have interracial dating experience. Based on this
research that found that interracial dating experience predicts positive emotions toward

33

racial outgroups, it is possible that this relationship works in the other direction, meaning
that positive emotions in general could lead people with interracial dating experience to
think in more racially inclusive terms and thus feel more positively toward racial
outgroups (which could manifest itself in higher ratings of perceived attractiveness). It
could be that when in a positive mood, people are more inclined to draw upon diverse
experiences when forming perceptions of others compared to when they are in a neutral
mood. The bi-directionality of this relationship would be a good topic for future studies.
Although we cannot speak to differences between positive mood states, this result
does appear to support broaden-and-build theory (Fredrickson, 2001) generally because
people who received positive mood inductions demonstrated broadened cognition. The
caveat of our results is that being in a positive mood state is not sufficient, necessarily, to
broaden cognition, but that it is being in a positive mood in conjunction with having had
a life experience that is relevant to the social perception at hand that predicts such mental
expansion. Importantly, we did not ask participants if they had had a significant
relationship with someone of an Asian race, but just of another race in general, because
we presented our measures to participants of two different races to examine other-race
perception more generally. It would be interesting to see what patterns emerged if we
isolated participants who had had an other-race relationship experience with the specific
other race measured in the study. Given the finding that having a significant relationship
with an other-race individual is linked to greater liking of racial outgroups in general
(Levin et al., 2007), we imagine that a stronger moderating effect of interracial dating
experience on the relationship between mood and perceived attractiveness would emerge
if we only analyzed those White participants who had relationship experience with Asian

34

people. We would be interested to discover if interracial dating experience could even
affect racial perception on its own beyond any interactions with mood.
We also tested for the potential moderating factor of racial ingroup identification
on the effect of the mood conditions on perceived attractiveness of other-race faces.
Although there was a significant negative correlation between group identification and
other-race perceived attractiveness, regression analyses determined that group
identification did not effectively moderate the relationship between the mood conditions
and perceived attractiveness. These results can perhaps be understood in the context of
the findings by Wilton, Sanchez and Giamo (in press) on the moderating effect of racial
identification on the relationship between ambiguous-race face exposure and intergroup
similarity ratings. Participants in their study were shown either White, Asian or
White/Asian biracial faces with a corresponding racial label and were then asked to rate
their beliefs on the similarity between Whites and Asians in terms of attitudes, beliefs and
behaviors. The authors found that biracial label exposure (as opposed to monoracial
labels) predicted greater intergroup similarity ratings and that racial identification
moderated this relationship, whereby Whites who were low in racial identification
reported higher intergroup similarity ratings compared to Whites who were high in racial
identification.
The results of Wilton et al. (in press) indicate that being exposed to faces that blur
dichotomous racial lines leads to greater perceived similarity between races (i.e., ingroup
expansion) but that being highly identified with one’s racial ingroup disrupts this
relationship, presumably because these people are more motivated to engage in ingroup
bias processes to protect collective self-esteem. These findings lend credence to the

35

marginally significant moderating effect of racial group identification on the effect of the
mood conditions on the perception of other-race faces’ attractiveness in our study.
Although the independent variable manipulations were different in our study, we assessed
the equivalent of intergroup similarity with our other-race perceived attractiveness task
since we know that perceived attractiveness is indicative of psychological closeness and
similarity (Folkes, 1982). It is likely that the differences in manipulations between these
two studies accounted for the differences in the moderating effect of racial group
identification. Nevertheless, our study appears to support the literature that group
identification impacts ingroup perception and biases.
Limitations
There were a few limitations in our study. We decided to recruit Asian American
participants since they represented the greatest minority student population and we
wanted to test the generalizability of our findings across different races. Unfortunately,
the participation rate of Asian Americans was not adequate enough so we were unable to
get a better understanding of how certain positive emotions may have differential effects
on people based on their race. Future studies can hopefully explore this avenue further
by effectively recruiting a more diverse sample of participants. Although we know that
all racial/ethnic groups (not just White people) engage in ingroup bias processes (e.g.,
Lee, 1993), we would be interested to find out if particular mood manipulation
techniques are more effective for certain racial groups, if group identification is
qualitatively different for minority racial groups than the majority group and if racial
minority individuals’ perceived attractiveness of White faces would differ from White’s

36

perception of other-race faces due to a difference in cultural salience between White and
racial minority individuals.
Another limitation was that our positive mood videos did not seem to successfully
induce the specific emotions we were trying to manipulate. This could be due to the fact
that we placed our manipulation check after the dependent variables, and at that point, the
positive or neutral mood had dissipated. However, research has shown that the
presentation of the manipulation check preceding the dependent variable dilutes the effect
of the manipulation (Brown & Brown, 2011; Fredrickson & Branigan, 2005; Kühnen,
2010). They are essentially “unprimed” (Sparrow & Wegner, 2006). Future research can
hopefully find a balance between the timing of the manipulation and the presentation of
the dependent variable, and the manipulation check.
Although the length of the clips that we used has precedent in the literature
(Fredrickson & Branigan, 2005), it is possible that they were too short to successfully
induce a particular emotion that remained through the duration of our dependent
measures and subsequent manipulation check questionnaire. It is also possible that our
dependent variable tasks induced strong emotions (e.g., stress, guilt) that interfered with
participants’ ability to accurately recall the emotions they felt during the mood
manipulation videos. Research into the effect of music on emotions has found that
people feel more positively toward familiar music (Pereira et al., 2011) and that specific
structural elements of background music (e.g., harmony, tempo, dynamics, rhythm) affect
emotional responses (Alpert & Alpert, 1990). Although subjective reactions to the music
used in our video clips included the fact that it was congruent with the video content and
appropriate for the specific positive emotion attempting to be elicited, measuring for

37

music familiarity and/or manipulating specific structural elements of the music could
greatly improve the videos’ effectiveness at manipulating mood.
Conclusion
Although our results did not match our predictions, we think that there is a strong
theoretical foundation for a measurable difference between specific positive mood states
in terms of social perception. Given the fact that people in positive moods have been
shown to demonstrate flexible, broadened and inclusive cognition (Fredrickson, 2001;
Fredrickson & Branigan, 2005; Isen & Daubman, 1984; Waugh & Fredrickson, 2006) as
well as narrow, heuristic judgments (Bodenhausen, Kramer et al., 1994; Macrae, Milne,
& Bodenhausen, 1994), it is clear that there is a gap in the literature related to
understanding the effects of positive emotions. With a larger, more diverse participant
sample and effectively piloted mood manipulation videos, we contend that different
positive moods should have an impact on our perception of others.
Given the detrimental effects of discrimination and racial bias (Brewer, 1999;
Cole, Kemeny, & Taylor, 1997; Jamieson, Koslov, Nock, & Mendes, 2013), it is
important to understand how to combat these unconscious processes. We all engage in
automatic ingroup biases that preferentially favor those who we perceive as being a part
of our ingroup (Aberson, Healy, & Romero, 2000; Beaupré & Hess, 2006; Gutsell &
Inzlicht, 2012; Folkes, 1982; Tajfel et al., 1971) but if we know how to expand who gets
included in our ingroups, then these same ingroup advantages could be applied to people
who would normally get excluded. This process would have important implications in
real-world settings, such as in corporations, schools and courtrooms. If we know, for
example, that jurors are inherently biased against outgroup members but that being in a

38

certain positive mood effectively broadens ingroup perception, then more efforts could be
made to ensure that the jurors are in that mood state when rendering their judgments.
Ultimately, if we can learn how positive emotions lead to more inclusive thinking or if
specific types of positive emotions differentially impact social perception, then we will
be one step closer to understanding how to reduce racial bias and discrimination.

39

References
Aberson, C. L., Healy, M. R., & Romero, V. L. (2000). Ingroup bias and self-esteem: A
meta-analysis. Personality and Social Psychology Review, 4, 157-173.
Alpert, J. I., & Alpert, M. I. (1990). Music influences on mood and purchase intentions.
Psychology and Marketing, 7(2), 109-133.
Anzures, G., Quinn, P. C., Pascalis, O., Slater, A. M., Tanaka, J. W., & Lee, K. (2013).
Developmental origins of the other-race effect. Current Directions in
Psychological Science, 22, 173-178.
Beaupré, M. G., & Hess, U. (2006). An in-group advantage for confidence in emotion
recognition judgments: The moderating effect of familiarity with the expressions
of out- group members. Personality and Social Psychology Bulletin, 32, 16-26.
Bernstein, I. H., Lin, T., & McClellan, P. (1982). Cross- vs. within-racial judgments of
attractiveness. Perception and Psychophysics, 32(6), 495-503.
Bodenhausen, G. V., Kramer, G. P., & Süsser, K. (1994). Happiness and stereotypic
thinking in social judgment. Journal of Personality and Social Psychology, 66,
621-632.
Bodenhausen, G. V., Sheppard, L., & Kramer, G. P. (1994). Negative affection and
social perception: The differential impact of anger and sadness. European
Journal of Social Psychology, 24, 45-62.
Brewer, M. B. (1999). The psychology of prejudice: Ingroup love or outgroup hate?
Journal of Social Issues, 55, 429-444.
Brockner, J., & Higgins, E. T. (2001). Regulatory focus theory: Its implications for the
study of emotions in the workplace. Organizational Behavior and Human

40

Decision Processes, 86, 35-66.
Brown, J. D., & Brown, M. A. (2011). Self-reflection and feelings of self-worth: When
Rosenberg meets Heisenberg. Journal of Experimental Social Psychology, 47,
1269-1275.
Cohn, M. A., Fredrickson, B. L., Brown, S. L., Mikels, J. A., & Conway, A. M. (2009).
Happiness unpacked: Positive emotions increase life satisfaction by building
resilience. Emotion, 9, 361-368.
Cole, S. W., Kemeny, M. E., & Taylor, S. E. (1997). Social identity and physical health:
Accelerated HIV progression in rejection-sensitive gay men. Journal of
Personality and Social Psychology, 72(2), 320-335.
Cullen, B., & Newell, F. (2012). Your face looks funny: The role of emotion on
perceived attractiveness of face images. Perception, 41, ECVP Abstract
Supplement, 106.
Dickter, C. L., & Kittel, J. A. (2012). The effect of stereotypical primes on the neural
processing of racially ambiguous faces. Social Neuroscience, 7, 622-631.
Folkes, V. S. (1982). Forming relationships and the matching hypothesis. Personality
and Social Psychology, 8, 631-636.
Fredrickson, B. L. (2001). The role of positive emotions in positive psychology: The
broaden-and-build theory of positive emotions. American Psychologist, 56, 218
226.
Fredrickson, B. L. (2013). Positive emotions broaden and build. In E. A. Plant & P. G.
Devine (eds.), Advances in experimental social psychology (Vol. 47) (pp. 1-53).
Burlington: Academic Press.

41

Fredrickson, B. L., & Branigan, C. (2005). Positive emotions broaden the scope of
attention and thought-action repertoires. Cognition and Emotion, 19, 313-332.
Freeman, J. B., Pauker, K., Apfelbaum, E. P., & Ambady, N. (2010). Continuous
dynamics in the real-time perception of race. Journal of Experimental Social
Psychology, 46(1), 179-185.
Gao, W., Cao, B., Shan, S., Chen, X., Zhou, D., Zhang, X., & Zhao, D. (2008). The
CAS-PEAL large-scale Chinese face database and baseline evaluations. IEEE
Transactions on System Man, and Cybernetics (Part A), 38(1), 149-161.
Gutsell, J. N., & Inzlicht, M. (2012). Intergroup differences in the sharing of emotive
states: Neural evidence of an empathy gap. Social Cognitive Affective
Neuroscience, 7, 596-603.
Hourihan, K. L., Fraundorf, S. H., & Benjamin, A. S. (2013). Same faces, different
labels: Generating the cross-race effect in face memory with social category
information. Memory and Cognition, 41, 1021-1031.
Hugenberg, K., & Bodenhausen, G. V. (2004). Ambiguity in social categorization: The
role of prejudice and facial affect in race categorization. Psychological Science,
15(5), 342-345.
Isen, A. M., & Daubman, K. A. (1984). The influence of affect on categorization.
Journal of Personality and Social Psychology, 47, 1206-1217.
Ito, T. A., Willadsen-Jensen, E. C., Kaye, J. T., & Park, B. (2011). Contextual variation
in automatic evaluative bias to racially-ambiguous faces. Journal of
Experimental Social Psychology, 47, 818-823.

42

Jamieson, J. P., Koslov, K., Nock, M. K., & Mendes, W. B. (2013). Experiencing
discrimination increases risk taking. Psychological Science, 24(2), 131-139.
Johnson, K. J., & Fredrickson, B. L. (2005). “We all look the same to me:” Positive
emotions eliminate the own-race bias in face recognition. Psychological Science,
16, 875-881.
Kühnen, U. (2010). Manipulation checks as manipulation: Another look at the ease-of
retrieval heuristic. Personality and Social Psychology Bulletin, 36(1), 47-58.
Lee, Y. T. (1993). Ingroup preference and homogeneity among African American and
Chinese American students. The Journal of Social Psychology, 133(2), 225-235.
Levin, S., Taylor, P. L., & Caudle, E. (2007). Interethnic and interracial dating in
college: A longitudinal study. Journal of Social and Personal Relationships,
24(3), 323-341.
Luhtanen, R., & Crocker, J. (1992). A collective self-esteem scale: Self-evaluation of
one’s social identity. Personality and Social Psychology Bulletin, 18, 302-318.
Macrae, C. N., Milne, A. B., & Bodenhausen, G. V. (1994). Stereotypes as energy
saving devices: A peek inside the cognitive toolbox. Journal of Personality and
Social Psychology, 66, 37-47.
Minear, M., & Park, D. C. (2004). A lifespan database of adult facial stimuli. Behavior
Research Methods, Instruments, and Computers, 36, 630-633.
Nelson, D. W. (2009). Feeling good and open-minded: The impact of positive affect on
cross cultural empathetic responding. The Journal of Positive Psychology, 4(1),
53-63.
North, A. C. (2012). The effect of background music on the taste of wine. The British

43

Psychology Society, 103, 293-301.
Peery, D., & Bodenhausen, G. V. (2008). Black + white = black: Hypodescent in
reflexive categorization of racially ambiguous faces. Psychological Science,
19(10), 973-977.
Pereira, C. S., Teixeira, J., Figueiredo, P., Xavier, J., Castro, S. L., & Brattico, E. (2011).
Music and emotions in the brain: Familiarity matters. PLoS ONE, 6(11), e27241.
doi: 10.1371/journal.pone.0027241
Schaefer, A., Nils, F., Sanchez, X., & Philippot, P. (2010). Assessing the effectiveness
of a large database of emotion-eliciting films: A new tool for emotion researchers.
Cognition and Emotion, 24(7), 1153-1172.
Shah, J. Y., Brazy, P. C., & Higgins, E. T. (2004). Promoting us or preventing them:
Regulatory focus and manifestations of intergroup bias. Personality and Social
Psychology Bulletin, 30, 433-446.
Sparrow, B., & Wegner, D.M. (2006). Unpriming: The deactivation of thoughts through
expression. Journal of Personality and Social Psychology, 91(6), 1009-1019.
Tajfel, H., Flament, C., Billig, M. G., & Bundy, R. F. (1971). Social categorization and
intergroup behaviour. European Journal of Social Psychology, 1, 149-177.
Tajfel, H., & Turner, J. C. (1986). The social identity theory of intergroup behaviour. In
S. Worchel & W. G. Austin (eds.), Psychology of intergroup relations (pp. 7–24).
Chicago, IL: Nelson-Hall.
Waugh, C. E., & Fredrickson, B. L. (2006). Nice to know you: Positive emotions, selfother overlap, and complex understanding in the formation of new relationship.
Journal of Positive Psychology, 1, 93-106.

44

Wilton, L., Sanchez, D. T., & Giamo, L. (In press). Seeing similarity or distance? Racial
identification moderates intergroup perception after biracial exposure. Social
Psychology.

45

Table 1
Mood Condition x Other-Race Relationship Experience Factorial Analysis of Variance
for Perceived Attractiveness of Other-Race Faces
Source

Df

F

p

(A) Mood condition

2

5.78

.005

(B) Relationship

1

0.33

.566

A x B (interaction)

2

2.70

.077

Error (within groups)

52

46

Table 2
Means and Simple Effects for the Interaction between Mood and Interracial Relationship
Experience on Other-Race Perceived Attractiveness
Joy

Mood Condition Groups
Contentment
Neutral Simple Effects: F
df (2, 52)

Relationship
Experience

4.07b
(0.35)

4.44b
(0.43)

3.08a
(1.23)

5.09**

No Relationship
Experience

3.65a
(0.44)

3.97a
(0.73)

3.63a
(0.71)

1.15

Simple Effects: F
df (1, 52)

1.92

1.87

2.31

Note. ** = p ≤ .01. Standard deviations appear in parentheses below means. Means with
differing subscripts within rows are significantly different at the p ≤ .05 level based on
post-hoc pairwise comparisons.

47

APPENDIX A
Collective Self-Esteem Scale (CSES)
INSTRUCTIONS: We are all members of different social groups or social categories.
Some of such social groups or categories pertain to gender, race, religion, nationality,
ethnicity, and socioeconomic class. We would like you to consider your membership in
your racial group, and respond to the following statements on the basis of how you feel
about that group and your memberships in them. There are no right or wrong answers to
any of these statements; we are interested in your honest reactions and opinions. Please
read each statement carefully, and respond by using the following scale from 1 to 7:

Disagree

Disagree
Somewhat

Neutral

Agree
Somewhat

Agree

Strongly
Agree

I often regret that I
belong to the racial
group I do.

1

2

3

4

5

6

7

2.

Overall, my racial
group is considered
good by others.

1

2

3

4

5

6

7

3.

Overall, my racial
group membership has
very little to do with
how I feel about
myself.

1

2

3

4

5

6

7

4.

I feel I don't have
much to offer to the
racial group I belong
to.

1

2

3

4

5

6

7

5.

In general, I'm glad to
be a member of the
racial group I belong
to.

1

2

3

4

5

6

7

6.

The racial group I
belong to is an

1

2

3

4

5

6

7

Strongly
Disagree

1.

48

important reflection of
who I am.
7.

I am a cooperative
participant in the
racial group I belong
to.

1

2

3

4

5

6

7

8.

The racial group I
belong to is an
unimportant to my
sense of what kind of
a person I am.

1

2

3

4

5

6

7

9.

I feel good about the
racial group I belong
to.

1

2

3

4

5

6

7

10.

In general, belonging
to a racial group is an
important part of my
self image.

1

2

3

4

5

6

7

49

APPENDIX B
modified Differential Emotions Scale (mDES)
Instructions: Please think back to how you felt during the video that you watched. Using
the 0-4 scale below, indicate the greatest amount that you've experienced each of the
following feelings.
Not at all
0

A little bit
1

Moderately
2

Quite a bit
3

Extremely
4

____ 1. What is the most amused, fun-loving, or silly you felt?
____ 2. What is the most angry, irritated, or annoyed you felt?
____ 3. What is the most awe, wonder, or amazement you felt?
____ 4. What is the most contemptuous, scornful, or disdainful you felt?
____ 5. What is the most embarrassed, self-conscious, or blushing you felt?
____ 6. What is the most grateful, appreciative, or thankful you felt?
____ 7. What is the most hopeful, optimistic, or encouraged you felt?
____ 8. What is the most inspired, uplifted, or elevated you felt?
____ 9. What is the most interested, alert, or curious you felt?
____ 10. What is the most joyful, glad, or happy you felt?
____ 11. What is the most love, closeness, or trust you felt?
____ 12. What is the most proud, confident, or self-assured you felt?
____ 13. What is the most sad, downhearted, or unhappy you felt?
____ 14. What is the most serene, content, or peaceful you felt?
____ 15. What is the most stressed, nervous, or overwhelmed you felt?

50

APPENDIX C
Rapid Categorization Task

Circle a number below
1
2
Completely
White

3

4
Biracial

5

51

6

7
Completely
Asian

APPENDIX D
Perceived Attractiveness Task

Circle a number below
1
Not at all
attractive

2

3

4

5

52

6

7
Extremely
attractive