University of Southern California Dissertations and Theses

Jilted and tilted: an exploration of post-rejection response and introduction of a novel experimental paradigm

(USC Thesis Other)

Jilted and tilted: an exploration of post-rejection response and introduction of a novel experimental paradigm

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content JILTED AND TILTED:
AN EXPLORATION OF POST-REJECTION RESPONSE AND INTRODUCTION OF A NOVEL
EXPERIMENTAL PARADIGM
by
Aili Qiao
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PSYCHOLOGY)
August 2024
Copyright 2024 Aili Qiao

ii
Dedication
To my two cats – Coffee and Oatmilk – who reject me every day.

iii
Acknowledgements
I would first and foremost like to thank my advisor, Professor Stephen J. Read. You offered so
much more than just your expertise and insightful feedback. Our constant non-academic chit-chats were
what kept me going throughout my PhD journey. I am forever grateful for the times you reminded me to
take care of my mental health.
I would also like to thank my committee members: Dr. Richard John for providing mentorship
even before I started grad school, Dr. Lynn Miller for constantly helping me level up my research craft,
and Dr. Albert "Skip" Rizzo for his constant reminder that my research should always serve those in need
in the real world.
In addition, I want to give a huge shoutout to my older brother Dr. Haifeng “Johnson” Qiao for
his sympathetic ear, my parents for their unconditional support, my best friend Emily Wang for keeping
me sane, my senior lab mate Dr. Steffie Kim for riding the struggle bus with me every week, my gamemaker Omar for his infinite patience, and my research assistants – Danuisca, Yazmin, and Anthony – for
firmly believing in my vision.
Finally, to my husband Harman: I couldn’t have gotten into grad school nor could I have
completed this dissertation without you. We did it!!!

iv
Table of Contents
Dedication..................................................................................................................................................... ii
Acknowledgements...................................................................................................................................... iii
List of Tables .................................................................................................................................................v
List of Figures...............................................................................................................................................vi
Abstract...................................................................................................................................................... viii
Chapter 1 – A Review of Past Theories and Research Methodologies in Social Nonacceptance.................1
Chapter 2 – Examining Nonacceptance Contexts’ Effect on Affect and Cognition: A SelfReport Vignette Study .................................................................................................................................33
Chapter 3 – Using a Faux Online Multiplayer Game to Simulate Different Nonacceptance
Contexts: Paradigm Development and Pilot Study .....................................................................................39
Chapter 4 – Comparing Findings between Spider Apocalypse and Existing Paradigms: A
Conceptual Validation .................................................................................................................................53
Chapter 5 – Comparing Nonacceptance Contexts on Mood, Cognition, and Behavior: An
Exploration ..................................................................................................................................................65
Chapter 6 – Testing the Hypothesized Nonacceptance-Response Process Model: A Path
Analysis .......................................................................................................................................................71
Chapter 7 – Conclusion and Discussion ......................................................................................................92
References....................................................................................................................................................98
Appendices ................................................................................................................................................103

v
List of Tables
Table 1: Experimental Paradigms in the Field of Social Nonacceptance ................................................ 18
Table 2: Vignette Setup and Manipulation.............................................................................................. 34
Table 3: Three-Way Mixed MANOVA Results of Type of Rejection (between subjects) ×
Emotional Closeness (between subjects) × Number of Rejecters (within subjects) ............................... 35
Table 4: Means, Standard Deviations, and Analyses of Variance in Likelihood, Anger,
Sadness, and Intention Interpretation for Type of Rejection ................................................................... 36
Table 5: Means, Standard Deviations, and Analyses of Variance in Likelihood, Anger,
Sadness, and Intention Interpretation for Emotional Closeness.............................................................. 36
Table 6: Means, Standard Deviations, and Analyses of Variance in Likelihood, Anger,
Sadness, and Intention Interpretation for Number of Rejecters .............................................................. 37
Table 7: Two-Way MANOVA Results of Number of Rejecters × Type of Rejection on
Change in Anger, Sadness, and Happiness.............................................................................................. 47
Table 8: Means, Standard Deviations, and Analyses of Variance in Change in Anger,
Sadness, and Happiness for Rejection Types.......................................................................................... 47
Table 9: Pairwise Comparison of Means Within Rejection Types on Change in Anger,
Sadness, and Happiness........................................................................................................................... 48
Table 10: Means, Standard Deviations, and Analyses of Variance in Change in Anger,
Sadness, and Happiness for Number of Rejecters................................................................................... 48
Table 11: Two-Way MANOVA Results of Number of Rejecters × Type of Rejection on
Partner Evaluation, Compensation Adjustment, and Rating Adjustment................................................ 49
Table 12: Means, Standard Deviations, and Analyses of Variance in Partner Evaluation,
Compensation Adjustment, and Rating Adjustment for Rejection Types............................................... 50
Table 13: Pairwise Comparison of Means Within Rejection Types on Partner Evaluation,
Compensation Adjustment, and Rating Adjustment ............................................................................... 50
Table 14: Means, Standard Deviations, and Analyses of Variance in Partner Evaluation,
Compensation Adjustment, and Rating Adjustment for Number of Rejecters ....................................... 51
Table 15: Estimated Regression Weight Parameters and Standard Errors for Predicting
Aggression in Two Steps......................................................................................................................... 60
Table 16: Estimated Regression Weight Parameters and Standard Errors for Predicting Social
Approach-Avoid in Two Steps................................................................................................................ 62
Table 17: Two-Way MANOVA Results of Number of Rejecters × Type of Rejection on
Change in Anger, Sadness, and Happiness.............................................................................................. 66
Table 18: Means, Standard Deviations, and Analyses of Variance in Change in Anger,
Sadness, and Happiness........................................................................................................................... 67
Table 19: Pairwise Comparison of Means Within Rejection Types on Change in Anger,
Sadness, and Happiness........................................................................................................................... 67
Table 20: Two-Way MANOVA Results of Number of Rejecters × Type of Rejection on
Partner Evaluation and Post-Game Behavior .......................................................................................... 68
Table 21: Means, Standard Deviations, and Analyses of Variance in Partner Evaluation and
Post-Game Behavior for Rejection Types............................................................................................... 69
Table 22: Pairwise Comparison of Means Within Rejection Types on Partner Evaluation and
Post-Game Behavior................................................................................................................................ 69
Table 23: Means, Standard Deviations, and Analyses of Variance in Partner Evaluation and
Post-Game Behavior for Number of Rejecters........................................................................................ 69
Table 24: Motivations Items Correlation................................................................................................. 72
Table 25: Means of Dependent Variables Within Each Condition ......................................................... 78

vi
List of Figures
Figure 1: The Need to Belong Model (1995) ............................................................................................ 3
Figure 2: The Self-Esteem as Sociometer Model (1995) .......................................................................... 4
Figure 3: Repeated Social Acceptance in the Sociometer Model (1995).................................................. 4
Figure 4: Repeated Social Rejection in the Sociometer Model (1995) ..................................................... 5
Figure 5: The Need-Threat Model of Ostracism (1997)............................................................................ 6
Figure 6: The Social Reconnection Model (2007)..................................................................................... 7
Figure 7: The Temporal Need-Threat Model (2009)................................................................................. 8
Figure 8: Taxonomy of Social Exclusion by Wesselmann et al. (2016).................................................... 9
Figure 9: Taxonomy of Social Exclusion by Freedman et al. (2016)...................................................... 10
Figure 10: Proposed Taxonomy of Social Nonacceptance ...................................................................... 26
Figure 11: Proposed Process Model of Social Nonacceptance Response ............................................... 28
Figure 12: Parallel Mediation Model to be Tested .................................................................................. 30
Figure 13: Games that Inspired Spider Apocalypse ................................................................................ 40
Figure 14: Introduction Screens of Spider Apocalypse ........................................................................... 41
Figure 15: Example Gameplay Screen of Spider Apocalypse................................................................. 42
Figure 16: Spider Apocalypse Experimental Conditions......................................................................... 43
Figure 17: Mediation Model of Implicit vs. Control Conditions on Aggression Mediated by
Anger ....................................................................................................................................................... 60
Figure 18: Mediation Model of Unknown vs. Control Conditions on Social Approach vs.
Avoid Mediated by Negative Mood......................................................................................................... 63
Figure 19: Stepwise Partial Least Squares Models for Predicting Partner Evaluation............................ 74
Figure 20: Stepwise Partial Least Squares Models for Predicting Subsequent Social Approach
vs. Avoidance Behavior ........................................................................................................................... 76
Figure 21: PLS Model of Motivations to Partner Evaluation.................................................................. 79
Figure 22: PLS Model of Emotions to Motivations to Partner Evaluation ............................................. 79
Figure 23: PLS Model of Rejection vs. non-Rejection to Partner Evaluation, Mediated by
Emotion and Motivation.......................................................................................................................... 80
Figure 24: PLS Model of Control vs. Irrelevant Rejection to Partner Evaluation, Mediated by
Emotion and Motivation.......................................................................................................................... 81
Figure 25: PLS Model of Personal vs. Unknown Rejection to Partner Evaluation, Mediated
by Emotion and Motivation..................................................................................................................... 82
Figure 26: PLS Model of Personal vs. Implicit Rejection to Partner Evaluation, Mediated by
Emotion and Motivation.......................................................................................................................... 83
Figure 27: PLS Model of Implicit vs. Unknown Rejection to Partner Evaluation, Mediated by
Emotion and Motivation.......................................................................................................................... 83
Figure 28: PLS Model of Personal Rejection vs. Control Condition to Partner Evaluation,
Mediated by Emotion and Motivation..................................................................................................... 84
Figure 29: PLS Model of Motivations to Subsequent Social Approach vs. Avoidance
Behavior................................................................................................................................................... 85
Figure 30: PLS Model of Emotions to Motivations to Subsequent Social Approach vs.
Avoidance Behavior................................................................................................................................. 86
Figure 31: PLS Model of Rejection vs. non-Rejection to Subsequent Social Approach vs.
Avoidance Behavior, Mediated by Emotion and Motivation .................................................................. 86
Figure 32: PLS Model of Control vs. Irrelevant Rejection to Subsequent Social Approach vs.
Avoidance Behavior, Mediated by Emotion and Motivation .................................................................. 87
Figure 33: PLS Model of Personal vs. Unknown Rejection to Subsequent Social Approach vs.
Avoidance Behavior, Mediated by Emotion and Motivation .................................................................. 88
Figure 34: PLS Model of Personal vs. Implicit Rejection to Subsequent Social Approach vs.
Avoidance Behavior, Mediated by Emotion and Motivation .................................................................. 89

vii
Figure 35: PLS Model of Implicit vs. Unknown Rejection to Subsequent Social Approach vs.
Avoidance Behavior, Mediated by Emotion and Motivation .................................................................. 89
Figure 36: PLS Model of Personal Rejection vs. Control Condition to Subsequent Social
Approach vs. Avoidance Behavior, Mediated by Emotion and Motivation ............................................ 90

viii
Abstract
Social nonacceptance, a common yet impactful human experience, has been extensively studied in the
field of social psychology. However, the field still lacks a comprehensive framework for the
conceptualization and experimentation of social rejection. This dissertation addresses this gap by
proposing a taxonomy and an integrative framework for understanding social rejection. A novel
experimental paradigm, Spider Apocalypse, is introduced to test these concepts in a controlled,
ecologically valid environment. Pilot studies using this paradigm reveal significant effects of rejection
contexts on mood, cognition, and behavior, supporting the proposed taxonomy and framework. This work
advances the theoretical understanding of social rejection and provides a robust tool for future research.

1
Chapter 1 – A Review of Past Theories and Research Methodologies in Social Nonacceptance
As has been said many times before – humans are inherently social creatures. Yet, we constantly
experience social rejection, ostracism, and exclusion. It is not uncommon to see people cast out of their
religious groups, rejected by their romantic interests, or ignored by their close friends. Perhaps due to the
phenomenon’s prevalence, many researchers have examined the effects of social rejection on human
cognition, affect, and behavior. Since the 1990s, several theories, models, paradigms, and individual
difference measures have been proposed to investigate the impact of social rejection.
However, despite the many theories and experimental methodologies, there is no inclusive,
systematic framework for conceptualizing and testing the effects of social rejection. This lack of a
thorough conceptualization has led to a paradoxical pattern of empirical findings in how individuals
respond to social rejection. For example, some research shows that humans exhibit antisocial behaviors
post-rejection (see: Twenge et al., 2001; Smith & Williams, 2004; Wesselmann et al., 2010; Chester &
DeWall, 2017), while other evidence shows that humans exhibit prosocial behaviors post-rejection (see:
Maner et al., 2007; DeWall, 2010; Walasek et al., 2019).
The lack of an overall conceptual structure is often hinted at when researchers attempt to decode
this pattern. A 2005 review by Williams and Gerber suggested two possible explanations: 1) participants
respond prosocially when their need to belong is threatened, but they respond antisocially when their need
for control is threatened, or 2) participants are more likely to be antisocial if they can respond in an
implicit way, but they respond more prosocially if they respond in an explicit way. A follow-up review by
Williams (2007) and a more recent review by Wesselmann & and Williams (2017) both pointed out that
the contradictory findings could also be due to differences in experimental paradigms. Williams (2007)
argued: “Undoubtedly, these paradigms themselves may account for some of the discrepant outcomes
(i.e., pro- versus antisocial responses), so it is wise to consider each and to note which paradigms are
associated with which outcomes.” Along the same line of argument, Richman (2013) proposed that the
ways in which we respond to rejection would depend on how we construe the context of the experience –
i.e., how we interpret the rejection experience would affect how we respond to it.

2
Additionally, a review by Wesselmann et al. (2015) noted that the behavioral measures available
to participants could be the reason for the conflicting findings: “Often, researchers only give participants
one behavioral option. Because of this, it is hard to rule out the possibility that ostracized participants are
simply responding more extremely than included participants using whatever option they are given
because it is the only option they have to fortify some of their basic needs.” These varied explanations
offered in previous reviews serve as evidence that the field still lacks a coherent framework for and
systematic testing of the impact of social rejection.
The Current Chapter
The main aims of this paper are: 1) to outline a thorough taxonomy of social rejection, and 2) to propose
a new, integrative framework of social rejection. To do so, I shall start by offering a systematic review of
experimental methodologies, where I examine past theories, paradigms, measurements, and moderating
variables. The articles selected for review were identified by searching keywords including “social
rejection”, “social exclusion”, and “ostracism”. Related articles not captured by these keywords were
identified using connectedpapers.com. A thorough search resulted in approximately 300 papers. During
the review, I will draw from existing research to argue that 1) there does not yet exist a thorough
taxonomy of social rejection types, 2) past experimental paradigms in the field of social rejection do not
offer systematic manipulations of the different types of rejections, 3) measurements of behaviors and
behavioral intentions are constantly used interchangeably, 4) behavioral measurements have mostly only
offered one behavior option, and 5) moderating variables have been tested in isolation and rarely
replicated.
Though this review is broad and comprehensive, it is not without constraints. The current review does not
look at cultural differences, identity-based rejection (gender, ethnicity, political orientation, etc.),
rejection in children/adolescents, neural/physio responses to rejection, or long-term persistent rejection.
This review focuses solely on experimentally induced rejection manipulations, and their effects on affect,
cognition, and behavior.

3
Theories
In this section, I outline existing theories of social rejection in the order of publication. I will
attempt to demonstrate that even though the theories may be supported by research they often only
capture a small segment of the social rejection process.
The Need to Belong Model
Baumeister and Leary’s (1995) need to belong model arguably kickstarted the field of research in
social rejection. The central idea of this model claims that all humans have a fundamental need to belong
– i.e., an inherent motivation to form and maintain social bonds. Baumeister and Leary further claim that
the need to belong should a) produce effects under many different contexts, b) produce cognitive and
emotional responses, c) lead to ill effects if this need is thwarted, d) elicit goal-oriented behavior designed
to satisfy it, e) apply to all individuals, f) not be derivative of other motives, g) affect a wide variety of
behaviors, and h) have implications that go beyond immediate psychological functioning.
In the context of social rejection, the need to belong model functions as follows:
Figure 1
The Need to Belong Model (1995)
The model suggests that a rejection experience would thwart one’s need to belong, thus creating negative
cognitive and affective responses, which lead to one’s motivation to satisfy the need to form and maintain
social bonds. The need to belong model acts as the foundation of many subsequent models of social
rejection. Those models also employ the process of rejection → affected internal state → motivation to
recuperate, but with variations depending upon their unique conceptualizations.
The Sociometer Model
The self-esteem as sociometer model (Leary et al., 1995), proposed around the same time as the
need to belong model, incorporates a similar process structure. The model proposes that self-esteem is an

4
indicator (aka. Sociometer) of one’s social relationship quality. In other words, the model suggests that
how individuals feel about themselves in each moment is a subjective index of how they perceive
themselves to be included/excluded by others. This subjective index informs and motivates the individual
to avoid further social rejection/exclusion. Thus, the self-esteem as sociometer model functions as
follows:
Figure 2
The Self-Esteem as Sociometer Model (1995)
The model suggests that a rejection experience would decrease one’s self-esteem, thus motivating
one to avoid any possible future rejection. The sociometer model and the need to belong share many
similarities. Both models suggest that a rejection experience would affect some internal state, and the
affected state would motivate outward behavior. In the need-to-belong model, the proposed internal state
is a thwarted need to belong, and the behavior is to satisfy the need; on the other hand, in the sociometer
model, the proposed internal state is the decreased self-esteem, and the behavior is to avoid future
rejection. Furthermore, the sociometer model offers an explanation of how repeated rejection (or
acceptance) can affect one’s chronic self-esteem and social expectations:
Figure 3
Repeated Social Acceptance in the Sociometer Model

5
Or alternatively:
Figure 4
Repeated Social Rejection in the Sociometer Model
The sociometer model thus concludes that repeated rejection leads to chronic low self-esteem,
which results in one’s expectation of being rejected in the future.
The Need-Threat Model
The need-threat model of ostracism (Williams & Sommer, 1997) posits that ostracism threatens
four basic needs of humans: 1) ostracism deprives people of a sense of belongingness to others, 2)
ostracism threatens its victims’ abilities to maintain high self-esteem, 3) ostracism robs individuals of a
sense of control over their interactions with others, and 4) individuals repeatedly exposed to ostracism
may question whether their existence is meaningful or important. From this outline, one can see that it
incorporates the need to belong model (see point (1) above) as well as the self-esteem as sociometer
model (see point (2) above). The model also offers an explanation of how repeated ostracism affects one’s
internal state, suggesting that repeated ostracism leads to individuals questioning the importance of their
own existence. The authors of this model made two clarifying statements: first, the model only applies to
ostracism, conceptualized by the authors as “the perception of being ignored by others in one’s presence”
(Williams & Sommer, 1997), and the authors do not claim that this model works with rejections (e.g.,
being told explicitly that one is not welcomed); second, the model assumes that depending on which
need(s) have been threatened, the individuals will behave accordingly to regain or strengthen the deprived
need.

6
Figure 5
The Need-Threat Model of Ostracism (1997)
The Social Reconnection Model
Building upon previous models, Maner et al. (2007) proposed the social reconnection model. The
model’s structure is much like that of the need to belong model and the need-threat model, suggesting that
when a social need is thwarted, humans seek alternative means of satisfying the drive. Though previous
models do not address exactly what humans do to resatisfy their thwarted needs, the social reconnection
model specifically claims that after experiencing rejection/exclusion, humans will wish to form bonds
with other people in order to recuperate our need for social connection. Furthermore, the authors of the
social reconnection model explained that affiliative needs do not happen under all rejection/exclusion
circumstances. They claim that after experiencing rejection/exclusion, humans will only seek to satisfy
their affiliative motives if we feel like others can provide realistic sources of social reconnection. Such
perception is dependent on: 1) Whether the other person is the one who just rejected the agent – because
humans do not wish to affiliate with individuals who were the original rejecters; 2) Whether the agent
thinks they will interact with this other person – because if there will not be real interactions, the agent
will not have affiliative motives; and 3) Individual differences – e.g., fear of rejection – because those

7
who are high on fear of rejection will have negative expectations about novel social interactions, and they
would tend to avoid novel social encounters for fear that they will be rejected again.
Figure 6
The Social Reconnection Model (2007)
The Temporal Need-Threat Model
Approximately a decade after the development of the need-threat model (Williams & Sommer,
1997), Williams (2009) expanded it with a temporal component. This new temporal need-threat model
suggests that after detecting cues of ostracism, humans respond in three stages: 1) during the immediate,
reflexive stage, we experience emotional pain, negative affect, and our needs are threatened (belonging,
self-esteem, control, and meaningful existence); then, 2) during the reflective stage, we assess the
meaning and motivation of the ostracism, and try to come up with ways to resatisfy those threatened
needs; finally, 3) if ostracism is experienced persistently over time, we enter the resignation stage, where
we feel helpless and depressed.

8
Figure 7
The Temporal Need-Threat Model (2009)
Additionally, Williams added that prosocial responses are present when inclusionary needs
(belonging and self-esteem) are threatened, whereas antisocial responses are present when the power
needs (control and meaningful existence) are threatened.
The Multi-Motive Model
In a more recent take on the taxonomy of rejection, Richman (2013) proposed a multi-motive
model of responses to rejection. It is the first model to formally suggest that humans respond to rejection
in three ways – to aggress, to appease, or to withdraw. Unlike the social reconnection model, where
Maner et al. (2007) theorized that our response depends on our interpretation of the subsequent social
situational context, the multi-motive model claims that our response depends on our current interpretation
of the rejection context. The perceived context is composed of six construals: 1) the perception of fairness
of the rejection, 2) the expectations of relationship repair, 3) the pervasiveness or chronicity of rejection,
4) the value of the damaged relationship, 5) the perceived cost of the rejection, and 6) the possibility of
relational alternatives. Richman posits that each of these construals would motivate our behaviors
differently. E.g., a sense of unfairness of rejection would make humans respond more aggressively, a high

9
value of relationship would prompt humans to act prosocially in hopes of repairing the relationship, and a
perceived high probability of chronicity would motivate humans to withdraw and avoid future rejections,
etc. When the combination of the construals is taken together, one motive for behavior may dominate,
which leads to a behavioral response.
Terminology and Taxonomy
As this field of research progressed over the years, so did the terminology and taxonomy.
However, researchers were not always careful in distinguishing the different types of social exclusion,
ostracism, and rejection. This often results in researchers falling prey to the jingle-jangle fallacy (Block,
1995), where two different rejection phenomena bear the same name (jingle fallacy), or the same rejection
phenomena are given multiple names (jangle fallacy). Here, I will provide a brief overview of the current
terminology and taxonomy in social exclusion, ostracism, and rejection research.
In a 2016 review, Wesselmann et al. offered a taxonomy, visually represented in the figure below:
Figure 8
Taxonomy of Social Exclusion by Wesselmann et al. (2016)
Wesselmann et al. (2016) considered social exclusion the umbrella term that encapsulates all
experiences of being kept apart from others physically or emotionally. Within social exclusion, the
authors grouped the exclusion experiences as either social rejection (receiving direct negative attention)
or ostracism (being ignored by others).
A review published that same year (Freedman et al., 2016) offered a similar taxonomy but broke
things down a step further, as shown in the figure below:

10
Figure 9
Taxonomy of Social Exclusion by Freedman et al. (2016)
Freeman et al. (2016) categorized exclusion experiences based on the type of communication between the
actor and the target of exclusion. Similar to the taxonomy proposed by Wesselmann et al. (2016),
Freedman et al. theorized that if there is no communication between the excluder and the excluded, then it
is an ostracism, and if there is direct communication, then it is a rejection. In Freedman et al.’s
categorization, they further defined that if the communication explicitly indicated that social exclusion
has occurred, then it is an explicit rejection, but if the communication did not explicitly indicate social
exclusion, then it is an ambiguous rejection.
Though not formally developed from a theoretical perspective, the third level of distinction
emerged from experimental paradigms. Both the Video Exchange paradigm (Vorauer et al., 2003; Maner
et al., 2007) and the Get Acquainted paradigm (Leary et al., 1995) manipulated explicit rejection into two
separate conditions – personal rejection (where the rejecter explicitly said they do not want to work with
the participant), and irrelevant rejection (where the rejecter did not work with the participants due to
external reasons). By adding this distinction, it was found that an explicit personal rejection affects
participants more so than an explicit irrelevant rejection does, though in both antisocial and prosocial
directions (see: DeWall et al., 2008; Romero-Canyas et al., 2010; Rajchert et al., 2019).

11
Moving forward in this current review, I will use the term “nonacceptance” as an umbrella term,
where the defining features of the experience (e.g., perceived delivery method of nonacceptance,
perceived reason for nonacceptance, assessment of emotional closeness to the source of nonacceptance,
etc.) are adjectives tagged before the phrase. For example, when a person purposefully ignores their
romantic partner, it is an “indirect personal nonacceptance by a romantic partner”. I chose
“nonacceptance” as an umbrella term because it can capture a wider range than existing terms.
“Rejection” often entails a level of deliberativeness, but does not cover seemingly unintended acts – for
example, not being invited to a social gathering cannot be categorized as a rejection. Both “exclusion”
and “ostracism” imply that the source is often a group either detaching or barring the entry of the
target(s), but these terms do not cover acts by individuals – for example, being turned down by a romantic
interest cannot be categorized as an exclusion or ostracism. Rejection, exclusion, and ostracism are lowerorder terms that describe specific types of nonacceptance, and I shall still use these terms when referring
to certain existing experimental paradigms (e.g., Cyberball creates an ostracism experience, whereas the
Video Exchange paradigm creates a rejection experience; discussed in the section below). However,
nonacceptance can be used to describe all forms of rejection, exclusion, and ostracism; whether direct,
indirect, group1
, or individual.
Paradigms
To test the theories proposed, researchers have devised and implemented numerous experimental
paradigms. In this section, I will 1) review some of the more popular paradigms used in this field of
research, and 2) distinguish/compare the multiple features of each paradigm. I will attempt to demonstrate
that existing experimental paradigms do not offer systematic manipulations of the different types of
nonacceptance and I will use this review as part of the basis of the theoretical framework I propose.
1 Group, in social psychology, is defined as a collection of individuals who come together and interact with
each other. However, one can experience a group rejection as either a “group consensus to reject” or a
“collection of individuals who reject independently”. All paradigms in this field adhere to the latter
definition – the more “individualistic” type of group rejection. Therefore, when I say “group” rejection in
this manuscript, I’m referring to the experience of being in a group where each individual has made their
own decision to reject the subject.

12
A potential issue arises within the field when researchers compare findings from different
paradigms. Some paradigms utilize nonacceptance via experience, where participants were deceived into
believing that they are currently undergoing social nonacceptance. These paradigms include Cyberball,
Video Exchange, Get Acquainted, Chat Room, and so on. Other paradigms utilize nonacceptance via
mental simulation, where participants were explicitly asked to imagine/enact/recall a scenario of social
nonacceptance. One popular paradigm that does not fall into either of these two categories is the Future
Life Alone paradigm, which I will discuss in detail in a later section. Each of these paradigms offers
fundamentally different nonacceptance experiences, and drawing conclusions from cross-paradigm
findings is problematic. Failure to distinguish the multiple features of each paradigm can have a central
impact on how we comprehend empirical findings. For each of the paradigms we review, we will
explicitly identify what we see as key features of the paradigm that may distinguish it and its impact from
other paradigms that have been used.
Nonacceptance via Experience
Cyberball. Cyberball (Williams et al., 2000) is arguably the most popular paradigm in this field
of research. Its popularity is established not only because of its simplicity in implementation but also in
its robust findings (Hartgerink et al., 2015).
In this paradigm, the participant plays a ball-tossing computer game with two or more other
players. The participant is told that the other players are online participants located elsewhere, when in
fact, the other players are characters pre-programmed to behave as the experimenter designed them.
During the ball-tossing game, the group has one ball that the members toss to one another. When a player
receives the ball, they choose the player they wish to throw the ball to.
For most of the experiments that use Cyberball, a sense of nonacceptance is achieved when the
other players gradually stop tossing the ball to the participant, and more so to one another. Eventually, the
player no longer receives the ball from the other two players. In this condition, the participant is meant to
feel gradually excluded from a group they initially felt included in. In this paradigm, the experienced
nonacceptance is often indirect (i.e., the participant is not explicitly informed that the nonacceptance is

13
occurring), from a group of strangers, and the reasons for nonacceptance are commonly not made clear
to the participant.
Chat Room Paradigm. The chatroom paradigm (first used by Gardner et al. (2000), later
formalized by Williams et al. (2002)) uses a similar setup to that of Cyberball. In this paradigm, the
participant joins an online chatroom with two or more other participants. The participant is told that the
other members in the chatroom are online participants located elsewhere, when in fact, the other members
are trained confederates. The participants are prompted to have light-hearted conversations in the
chatroom.
To achieve a sense of nonacceptance, the other members of the group chat seem to find a
common interest (e.g., liking the same fictional band, or having gone to the same fictional high school)
and start to chat with one another about the shared interest without including the participant in the
conversation. Much like Cyberball’s manipulation, the participant is meant to feel gradually excluded
from a group that they initially felt included in.
In this paradigm, the experienced nonacceptance is often indirect, from a group of strangers, and
the reasons for nonacceptance are not made clear to the participant.
O-Cam. Another paradigm that uses a similar setup is the O-cam (Goodacre & Zadro, 2010).
Instead of a virtual ball toss or chat room, the paradigm uses a video conference platform. The participant
joins a conference call with two or more other participants. To achieve a sense of nonacceptance, the
other members (confederates or pre-recordings) of the conference call start to chat with one another
instead of chatting with the participant. In this paradigm, the experienced nonacceptance is also indirect,
from a group of strangers, and the reasons for nonacceptance are not made clear to the participant.
Video Exchange Paradigm and its variations. The video exchange paradigm originated from
the Signal Bias Model by Vorauer et al. (2003) but was later formalized by Maner et al. (2007). In this
paradigm, the participant is told that they and another participant will each make a self-introduction
videotape. The videotapes will be exchanged so they can view each other’s self-introduction, before being
paired up to do a task together. The participant goes into a private room to make the video introduction,

14
and the experimenter enters to take their videotape in exchange for a videotape supposedly made by
another participant who is currently in another private room. Then the participant watches the selfintroductory video produced by the other participant (a confederate).
For the personal reason manipulation, the experimenter comes into the room to tell the participant
that “the other person left the experiment because they did not wish to meet you and do the task with
you”. In this condition, the participant is meant to feel that they had been purposefully rejected by the
other participant.
For the irrelevant reason manipulation, the experimenter comes into the room to tell the
participant that “the other person left the experimenter due to a personal emergency unrelated to you –
they didn’t get to view your self-introduction video”. In this condition, the participant is still left to do an
individual task instead of a group task, but they are told that the other participant did not purposefully
mean to reject them.
In this paradigm, the experienced nonacceptance is direct (i.e., the participant is explicitly
informed that they had been rejected), by an individual stranger, and the reasons for nonacceptance are
made clear to the participant (personal vs. irrelevant reasons).
Variations. There are a handful of variations in the field that existed prior to the video exchange
paradigm, which don’t necessarily use videotapes. For example, some let participants write lengthy selfintroductory paragraphs (Downey & Feldman, 1996) instead of making a videotape, and others have
participants interact face-to-face first with the confederate before a rejection manipulation is implemented
(Ayduk et al., 1999), etc. The central idea of all these variations remains the same: the participant extends
a self-introduction and expects to be paired in a partnership, only to find that the other person has rejected
them, whether due to personal or irrelevant reasons.
Get Acquainted Paradigm. One paradigm that stemmed from the sociometer model is the get
acquainted paradigm (Leary et al., 1995). In this paradigm, the participant arrives at the lab and starts
chatting with other participants. After a while, the experimenter tells all participants that there is one
group task and a few individual tasks. Each participant then enters a separate room and fills out a form to

15
indicate with whom they would like to work on the group task. The setup concludes with the
experimenter collecting all forms from participants. This framework is designed such that the more
socially accepted a participant is, the more likely they will be chosen by other participants to do the group
task with.
For the personal reason manipulation, the experimenter comes into the room to tell the participant
that “you are selected to do the individual task because none of the other participants indicated they
would like to work with you on the group task”. In this condition, the participant is meant to feel that they
had been purposefully rejected and excluded by others in the group.
For the irrelevant reason manipulation, the participant is also assigned to do the individual task,
but the reason given is “because you have been randomly assigned to do so; your/their rankings of each
other were not factored in”. In this condition, the participant is still meant to feel excluded from others,
while knowing that others did not mean to do so purposefully.
In this paradigm, the experienced nonacceptance is direct (i.e., the participant is explicitly
informed that they had been rejected/excluded), from a group of strangers, and the reasons for
nonacceptance are made clear to the participant (personal vs. irrelevant reasons).
Ostracism Online. In an attempt to create a more ecologically valid paradigm, Wolf et al. (2014)
introduced Ostracism Online. In this paradigm, the participant creates an online profile of themself, and
enters a virtual space with other users (who are pre-programmed bots). The participant is instructed to
read everyone else’s profiles, and to leave a “like” on the profile if the participant finds them interesting.
At the end of the procedure, the participant sees how many “likes” they received from other users. A
sense of nonacceptance is achieved by giving the participant fewer “likes” than all the other profiles.
In this paradigm, the experienced nonacceptance is indirect (i.e., the participant is not explicitly
informed that they had been rejected/excluded), from a group of strangers, and the reasons for
nonacceptance are not made clear to the participant.

16
Rejection via Mental Simulation
Imagine. In studies that use the Imagine manipulation, participants are often asked to imagine
themselves in a scenario where they are not accepted. For example, an experiment by Poon & Chen
(2014) asked participants to imagine being ostracized at work, or an experiment by Wirth et al. (2017)
asked participants to imagine not being picked to play trivia.
These manipulations are easy to implement, but do not offer great internal or external validity.
The scenarios brought to the participants are often drastically different from one another depending on the
experiment, and there is no experimental control over the features of the vignettes presented to
participants.
Recall. Another mental simulation nonacceptance manipulation asks participants to recall a time
they felt rejected / excluded / ostracized / ignored (see: Bernstein et al., 2008; Chen et al., 2012; Klages &
Wirth, 2014). Though the instructions are simple and kept constant, the actual memories conjured by
participants are not specified. For example, one participant might recall being shunned by a group of close
friends, whereas another participant might recall being stood-up by a date. Researchers that used the
recall manipulation did not seek to control the type of rejection memory participants retrieved, which is a
cause for concern in understanding participants’ subsequent measured affect, cognition, and behavior.
O-Train. The O-Train paradigm developed by Zadro et al. (2005) asks participants to act out a
nonacceptance scenario. A group of participants is each given a script to follow. The script outlines a
scene taking place on a train, where three participants are chatting initially, but two of the three
participants begin to have a conversation excluding the third participant. The scenario itself is very similar
to the Chat Room paradigm and O-Cam, except participants in O-Train are not deceived to believe that
the rejection scenario is real.
In this paradigm, the simulated nonacceptance is indirect (i.e., the participant is not explicitly
informed that they had been rejected/excluded), from a group of strangers, and the intentions are not
made clear to the participant. However, all the participants are aware that the interaction is scripted, such
that the act of nonacceptance is unintended by the source(s).

17
Future Life Alone
The Future Life Alone paradigm (Twenge et al., 2001) does not belong in either of the two
categories listed above, because it incorporates a bit of both. It uses mental simulation as well as
deception. In this paradigm, participants fill out numerous personality measures. Then, in the future-lifealone condition, the participant is told that their personality test results show they’re more likely to end up
alone in life. This test result is made-up, and it is intended to induce a feeling of exclusion.
However, one may speculate that this manipulation induces a sense of loneliness instead of
exclusion. Though the manipulation tells participants that their personal relationships won’t last, it never
really tells the participants why their relationships won’t last. The participants can interpret this future
outcome as “it’ll get harder for me to make new friends” or “I will outgrow my friends”, instead of “my
friends will exclude me”. The scenario described to the participants is one of solitude and isolation, and
less so about being rejected or excluded.
Paradigms by Dimension
The lack of a systematic framework has allowed these distinct manipulations to exist within the
same field of research, where the paradigm can elicit fundamentally different nonacceptance experiences
(Romero-Canyas et al., 2010; Ren et al., 2016). Comparing empirical findings from these paradigms is
fundamentally problematic, because the impact of each of the different factors has yet to be thoroughly
examined. A group vs. an individual nonacceptance may have different effects on one’s response, because
an individual nonacceptance signals one’s relationship with another person, whereas a group
nonacceptance signals one’s own standing within a social circle. A direct vs. an indirect nonacceptance
may have different effects on one’s response, because a direct nonacceptance leaves no room for doubt
that one is not accepted, whereas an indirect nonacceptance causes one to wonder whether an act of
nonacceptance has happened.
Thus, some key dimensions I have identified in nonacceptance manipulations are: 1) how many
people are involved in the scenario, 2) is the nonacceptance directly communicated or indirect, and 3) are

18
the reasons for nonacceptance made clear to the target, and if so, what are they. See Table 1 below for the
different dimensions of rejection manipulation, and how each paradigm fits into those dimensions.
Table 1
Experimental Paradigms in the Field of Social Nonacceptance.
Measurements
After a feeling of nonacceptance is elicited among participants, various measures are utilized to
investigate the effect of nonacceptance. In this section, I will review the affective, cognitive, and
behavioral measurements. I will attempt to demonstrate that most existing behavioral measures within this
field have mostly only offered one behavior option, and that withdraw was not used as a behavior option
in measurement until recently.
Affect
The most common measurement after putting participants through a nonacceptance condition is
mood. The vast majority of these studies that measure mood as a dependent variable uses a single
generalized/aggregated mood index, instead of separating the different emotions and examining the
impact of nonacceptance. Most studies found that participants reported worsened mood after experiencing
nonacceptance – i.e., decrease in positive affect and/or increase in negative affect – than control condition
participants (Nezlek et al., 1997; Zadro et al., 2004; Gonsalkorale & Williams, 2007; Jones et al., 2009;
Hawkley et al., 2011; Klages & Wirth, 2014; Wirth et al., 2017). However, there are exceptions with a
few studies, where mood was found to be non-significant (e.g., Maner et al., 2007; DeWall, 2010).

19
Even though worsened mood has been consistently found, it is unclear whether mood is a
mediator between nonacceptance and cognition/behavior. When looking at mood as a single index,
Twenge et al. (2001) found that mood did not mediate the effect of personal rejection on subsequent
aggressive behavior. Similarly, Hales & Williams (2018) also concluded that mood did not mediate the
effect of group ostracism on participants’ subsequent interest in joining extreme groups. When looking at
different affects separately, Chow et al. (2008) found that anger did mediate the effect of group ostracism
on subsequent social behavior, but sadness did not – angry participants were more likely to assign an
unappealing snack than an appealing snack to other participants. However, this effect did not replicate in
a later study done by Rajchert et al. (2017), where they found that anger did not mediate the effect of
personal rejection on hot sauce allocation to other participants.
Cognition
While reviewing the cognitive measures, I found that oftentimes researchers would claim
participants to be more “prosocial” or “antisocial” when the measurements are only cognitive, not
behavioral. For example, Ayduk et al. (1999; Study 2) concluded that women were more hostile after
experiencing a personal rejection, but hostility was operationalized as self-reported liking of their
communication partner. Similarly, but in the opposite direction, Romero-Canyas et al. (2010) concluded
that humans were more ingratiating after experiencing nonacceptance, but ingratiating behavior was
operationalized as an indicated willingness-to-pay for other group members.
Although one might argue that cognition directed towards others is a behavior, vast amounts of
empirical evidence in the attitudes literature tells us that there’s a considerable gap between a person’s
intention and their behavior (for a meta-analytic review, see: Armitage & Conner, 2001). Therefore, this
current review treats all self-report measures that don’t have external effects as “cognitive measures”,
whereas self-report or behavioral measures that are framed as having effects on other people are grouped
as “behavioral measures”. For example, a participant’s evaluation of others is treated as a cognitive
measurement; however, if the participant’s evaluation is framed as influencing others’ monetary

20
compensation, then the evaluation is treated as a behavioral measurement. Behavioral measures will be
discussed in a later section.
Self-Reflecting Cognition. A common self-reflecting cognition that is measured is participants’
subjective feeling of rejection (”how rejected / ostracized / excluded do you feel right now?”). This is
normally used as a manipulation check, and the responses are compared between participants in the
nonacceptance condition vs. participants in the control condition.
Another common self-reflecting cognition measured is the need-threat or need-satisfaction. Both
need-threat and need-satisfaction measure the degree to which participants’ four fundamental needs are
threatened (Williams et al., 2000; Zadro et al., 2004; ). The four needs are derived from the need-threat
model (Williams & Sommer, 1997), which are: Belongingness (e.g., “how much do you feel you
belonged to the group?”), Meaningful Existence (e.g., “how true is the statement: ‘Life is
meaningless’?”), Self-Esteem (e.g., “to what extent do you think the other participants value you as a
person?”), and Control (e.g., “how true is the statement: ‘I am in control of my life’?”). The scores of
each item are then aggregated to compose one overall need-threat/need-satisfaction score.
Directed Cognition. One of the directed cognitive measurements utilized is having participants
evaluate other people. Some studies measured participants’ evaluation of other people’s traits – e.g.,
smart, likable, intelligent, popular, friendly, kind, etc. (Leary et al., 1995; Ayduk et al., 2003), while other
studies measured participants’ evaluation of their own attitude towards other people – e.g., personal liking
and trust of group members (Jones et al., 2009), judgment of importance and closeness they feel towards
other participants (Böckler et al., 2021), etc.
Another set of directed cognitive measurements is of participants indicating intention for certain
behaviors. For example, an intention to retaliate against those who rejected the participant (van Beest &
Williams, 2006; Hales et al., 2015; Chester & DeWall, 2017; Rudert et al., 2017; Rajchert et al., 2019), an
intention to reaffiliate with those who rejected the participant (Wirth et al., 2010), an intention to engage
in other prosocial (Gamian-Wilk et al., 2020) or antisocial acts (Klages & Wirth, 2014; Tuscherer et al.,
2016), or an intention to withdraw (Pfundmair et al., 2015).

21
Behavior
Here, we will review some of the more popular behavioral measurements. The postnonacceptance behaviors are categorized as prosocial, antisocial, or withdrawal (Richman, 2013).
Existing behavioral measurements in this field are often unipolar and used in isolation (Wesselmann et
al., 2015) where researchers draw conclusions about either prosocial, or antisocial, or withdrawal
responses to nonacceptance. Rarely do past studies give participants a choice to exhibit competing
behaviors (prosocial and antisocial) using bipolar measures or a combination of unipolar measures.
Prosocial Measures. In social nonacceptance research, prosocial behaviors are often
operationalized as either charitable behaviors or conformity.
Charitable behaviors are those that benefit others over oneself. For example, participants can
choose to donate a portion of their study compensation for a good cause. Some conflicting findings arise
with this measurement: van Beest & Williams (2006) and Twenge et al., (2007) found that participants
donated less money after experiencing group nonacceptance, whereas Carter-Sowell et al. (2008) found
that participants donated more money after experiencing group nonacceptance. However, this contrast in
findings can be attributed to the different methods of measurement delivery. In van Beest & Williams
(2006) and Twenge et al., (2007), the act of charitable giving was framed as unmonitored and anonymous
– participants were under the impression that the amount they donate would be unknown to others. In
Carter-Sowell et al. (2008), a confederate explicitly asked the participants to make a pledge – the amount
they donated would be obvious to the person who had requested it. Given this difference, one may
speculate that individuals are less willing to give money to charity after experiencing nonacceptance, but
will do so for impression management.
Participants’ conformity to other group members is often measured as whether they would offer
up the wrong answer to a question just because other group members agreed that the wrong answer is
correct. Williams et al. (2000), while measuring conformity, found that participants were more likely to
conform after a nonacceptance experience. However, this effect was not replicated in a subsequent study
(Wolf et al., 2014).

22
Antisocial Measures. On the other hand, antisocial behaviors are often operationalized as how
severe is the punishment that the participants bestow upon others, either in the form of a noise blast or
some hot sauce.
The noise-blast task is normally framed as a reaction-time competition between the participant
and the source of nonacceptance (Twenge & Campbell, 2003; Chen et al., 2012; Chester & DeWall,
2017) or a new experiment partner who was not involved in the nonacceptance experience (Twenge et al.,
2001; Twenge et al., 2007; DeWall et al., 2009; DeWall et al., 2010; Poon & Chenn, 2016). In this task,
participants are asked to set the volume and duration of an aversive noise that their opponent would
supposedly hear in their headphones for every lost round. Then, the noise volume and duration set by the
participant would be combined into a single index of aggression.
Similarly, the hot sauce task lets participants allocate a certain amount of hot sauce to another
participant, while knowing that the other participant strongly dislikes spicy foods (see: e.g., Warburton et
al., 2006; van Beest et al., 2011; Reiter-Scheidl et al., 2018). The weight of the hot sauce allocated by the
participant would be used as an indication of aggression.
Both Prosocial and Antisocial in One Measurement. All the measurements reviewed above are
unipolar – i.e., each measurement estimates whether one is more/less prosocial or more/less antisocial.
Typically, studies have used only a single unipolar measurement to conclude how participants behave
after experiencing nonacceptance. Very few measures in this body of research are bipolar – i.e., allowing
participants to behave either prosocially or antisocially with one measurement. One of the more
commonly used bipolar measures is the job evaluation task. For this task, participants are asked to
evaluate a job candidate who is applying for a research assistant job at the lab. Participants are told that
their evaluation of this job candidate will directly affect the likelihood of hire, such that a good evaluation
would increase the chances, and vice versa. Thus, whether a participant is prosocial or antisocial can be
estimated from the evaluation they give to this job candidate.
However, the job evaluation task is not without flaws. First, despite the bipolar nature of the job
evaluation task, researchers have used it mainly as an operationalization of aggression (DeWall et al.,

23
2009; Chen et al., 2012; Poon & Chen, 2014; Rajchert & Winiewski, 2017). Secondly, it is unclear
whether the evaluation scale provided has a neutral anchor. Therefore, researchers were only able to
compare score means from groups to determine which group of participants gave higher/lower
evaluations, but were unable to compare the scores to a neutral anchor. For example, rejected participants
often gave a lower score on average than control participants, and researchers have concluded that
rejected participants are more aggressive. It is possible that while rejected participants gave lower scores,
the average score is still considered above-neutral (i.e., a positive evaluation). In this case, concluding that
rejected participants are “more aggressive” would be incorrect, as rejected participants are simply “less
prosocial” than control participants, as they are not exhibiting aggressive behaviors.
Withdraw. Only in more recent studies, were withdrawal behavior options added to the list of
measurements. After experiencing nonacceptance, participants were given the choice to do a task with
either a) the same group of people who did not accept the participant, b) a group of new people, or c) just
the participant alone. Though solitude-seeking behavior was found to be more prominent among
nonaccepted participants than control participants, nonaccepted participants did not prefer solitudeseeking over the connection with a new group (Ren et al., 2016; Ren et al., 2021). However, research with
withdraw as an option is scant, hence more evidence is needed.
Moderating Variables
Aside from a myriad of experimental paradigms and measurements, researchers also considered a
long list of potential moderating variables. Each of these moderators is theoretically motivated, but there
hasn’t been an attempt to examine these systematically; instead, most of these were tested in isolation,
and the effects were not replicated.
Within the past two decades, researchers have found that higher depressive individuals are more
negatively affected by personal rejection (Nezlek et al., 1997); high narcissistic individuals are more
angry and aggressive after a personal rejection (Twenge & Campbell, 2003); neurotic participants
expressed higher need-threat when experiencing nonacceptance (Nezlek et al., 2012). High anxious
attachment was associated with increased feeling that life was meaningless after experiencing indirect

24
group nonacceptance (Shaver & Mikulincer, 2013), whereas high avoidance attachment was associated
with less distress after an indirect group nonacceptance experience (Yaakobi & Williams, 2016). High
rejection sensitivity was associated with a heightened dislike towards the rejecter (Ayduk et al., 2003) and
aggression towards the rejecter (Ayduk et al., 2008); but it was also found to predict an increase of
ingratiating behavior towards the rejecter (Romero-Canyas et al., 2010). Consistently, trait self-esteem
was found not to be a moderator of mood (Williams et al., 2000) or aggression (Twenge & Campbell,
2003). A set of personal beliefs were also associated with the rejection-response process. For example,
Poon & Chen (2014; Studies 1 & 4) found that people with weak just-world beliefs were more likely to
aggress against an innocent stranger; Chen et al. (2012; Study 1) found that people with destiny belief
(“relationships are formed by destiny, instead of growth”) were more aggressive after experiencing an
indirect group nonacceptance.
In summary, this body of research underscores the complex relationship between individual
personality traits and responses to rejection, revealing how different characteristics can distinctly shape
emotional and behavioral reactions. However, the lack of systematic examination and replication of these
findings highlights the crucial gap in the literature, suggesting the need for more comprehensive and
integrated investigation into moderating variables.
Proposal for a New Taxonomy and Model
Throughout this review, I concluded that: 1) there does not yet exist a thorough taxonomy of
social nonacceptance types, 2) past experimental paradigms in the field do not offer systematic
manipulations of the different types of nonacceptance, 3) measurements of behaviors and behavioral
intentions are constantly used interchangeably, 4) behavioral measurements have mostly only offered one
behavior option among the three possibilities that have been studied, and 5) moderating variables have
been tested in isolation and rarely replicated.
However, the different types of nonacceptance experiences and measurements can greatly affect
the response process. In other words, I argue that our perception and interpretation of a nonacceptance

25
situation, as well as what behavioral options are available to us, can greatly affect how we determine our
subsequent responses.
Therefore, in this section, I offer 1) my own take on a taxonomy of the different types of social
nonacceptance contexts, and 2) an integrative framework of a nonacceptance-response process.
Taxonomy
My taxonomy is built upon previous empirical studies on nonacceptance contexts (DeWall et al.,
2010; Nezlek et al., 2012), existing taxonomies (Wesselmann et al., 2016; Freedman et al., 2016), and
how various experimental paradigms differ from one another (Bernstein & Claypool, 2012a & 2012b;
Godwin et al., 2014). In a daily diary study, Nezlek et al. (2012) found that 1) people were more
negatively affected when experiencing an act of nonacceptance from close friends than from strangers,
and 2) subjective interpretations of the reasons behind the nonacceptance drastically change how people
feel about the incident. These findings tell us that nonacceptance experiences can differ greatly depending
on the relationship we have with the source of the nonacceptance and how we interpret the reasons behind
the nonacceptance. Furthermore, the number of people involved in the nonacceptance experience and
whether the nonacceptance was explicit or inferred may affect one’s experience. For example, the video
exchange paradigm is structured to be a directly informed, one-on-one rejection experience, whereas
Cyberball is designed to be an inferred/indirect, group ostracism experience.
When considering the dimensions together, a nonacceptance context can be constructed from
these four factors: 1) the number of people involved, 2) target’s relationship to those individual(s), 3)
whether the nonacceptance occurrence is explicit or implicit, and 4) the reason given for the
nonacceptance.

26
Figure 10
Proposed Taxonomy of Social Nonacceptance
The number of people involved can be an important factor in a nonacceptance situation. A oneon-one nonacceptance is more direct and personal, and it immediately sends a signal about one’s
relationship with the nonaccepting person. On the other hand, a group nonacceptance is perhaps more
threatening, because it reflects one’s standing within a bigger social ecosystem.
One’s relationship to the nonaccepting individual(s) can also be crucial in how one experiences
the nonacceptance. Nonacceptance from a stranger would feel extremely different than nonacceptance
from someone we are emotionally close to. Additionally, the relationship between the target and source of
nonacceptance greatly determines what nonacceptance-response behavior the victim would employ. For
example, if we deem the relationship to be important, we might try harder to salvage and course-correct
from whatever caused the nonacceptance; on the other hand, if we deem the relationship to be
unimportant, we might not exert the time/energy in reconnection, and seek out new connections instead.
In terms of how a nonacceptance is communicated between its source and its target, previous
taxonomies each addressed a version of the distinction between an explicit nonacceptance vs. an implicit

27
nonacceptance (Freedman et al., 2016; Wesselmann et al., 2016). The central distinction boils down to
whether or not the target was directly informed of the nonacceptance occurrence.
Lastly, the reason given for the nonacceptance plays a big role in nonacceptance- response.
Previous studies have repeatedly found that certain reasons (personal / internal / stable) given for the
nonacceptance affected participants more negatively than other reasons (irrelevant / external / unstable)
(Nezlek et al., 1997; Ayduk et al., 1999; Ayduk et al., 2003; DeWall et al., 2008; DeWall et al., 2009;
Romero- Canyas et al., 2010; Rajchert et al., 2019; Yaakobi, 2021).
One factor not included in the current taxonomy, but still of great importance, is one’s social
goals related to the nonacceptance instance. For example, being denied entry to a volleyball club can have
drastically different effects on individuals depending on why they wanted to join in the first place. If
one’s goal was merely to make new friends at the club, then being denied entry may be upsetting, but
there are still other means to achieve said social goal. On the other hand, if one’s goal was to practice with
good volleyball players in hopes to make varsity, being denied entry may be devastating, as there are few,
if any, alternative options to achieve the goal elsewhere.
Even though one’s goals related to the nonacceptance instance is crucial, it is extremely difficult
to structure and incorporate into the taxonomy. Not only do individuals have goals that are different from
other people’s, each individual also has a variety of goals in mind at any given time. These goals could be
relational (e.g., making new friends), professional (e.g., networking in hopes to switch jobs), spiritual
(e.g., joining a church), recreational (e.g., finding a rock-climbing partner), and so on. To incorporate this
factor into the taxonomy of nonacceptance would require a clear taxonomy of the different types of
human motivations and goals, which (to my knowledge) is yet to be established within the field, and is
definitely beyond the scope of the current project. Though it warrants further research, an investigation on
how the different social goals can affect one’s response to nonacceptance is not included within the
current set of studies.

28
By combining the nonacceptance context, one’s perception of the nonacceptance situation,
internal processing of the experience, and behavioral output, I proposed a process model of social
nonacceptance response:
Figure 11
Proposed Process Model of Social Nonacceptance Response
At the first level, the various aspects of a nonacceptance context are present. These include the
categories mentioned above in the taxonomy: 1) number of people involved, 2) relationship to the source
of nonacceptance, 3) whether the nonacceptance occurrence is explicit or implicit, and 4) reason for
nonacceptance.
At the second level, a perception of the nonacceptance experience is formed based on the external
context. The number of people involved and the relationship with the source inform the agent of who is
involved in the nonacceptance, whereas the delivery method and reasons given allow the agent to
perceive why the nonacceptance happened.
At the third level, the perception is processed to produce the corresponding affect (anger and
sadness) and motivation (need to lash out, need for solitude, and need for affiliation).

29
Finally, at the fourth level, the more salient affect or motivation will determine which behavior
the agent chooses to employ: Whether it be to aggress, to approach/reconnect/ingratiate, or to withdraw.
A potential moderating factor between level three and level four is the context of the subsequent
social context. For example, a person might have a high desire to lash out, a moderate need to reaffiliate,
and a low fear of future nonacceptance. If the subsequent social context does not allow lashing out, the
motivation to lash out would be inhibited, and the need to reaffiliate would edge out, causing the person
to behave prosocially instead of aggressing or withdrawing. Another angle from which to understand how
subsequent context can determine one’s post-nonacceptance behavior is whether the subsequent social
situation is similar to the context where one experienced the nonacceptance in. A study by Hong & Sun
(2018) found that participants were more wary of new groups that resembled the original rejecters.
Drawing from previous nonacceptance experiences, the resemblance in social context may act as a signal
that a future nonacceptance by a new but similar group is likely.
Project Aims
For the current project, we simplify the nonacceptance-response process model into the following
parallel mediation model:

30
Figure 12
Parallel Mediation Model to be Tested
We hypothesize that the different types of nonacceptance would affect participants’ mood and
motivations, which in turn affect their behavioral response.
A few things not tested/measured in the current project are: 1) relationship to the source of
nonacceptance and 2) participants’ perception of the context. These are omitted solely for the reason of
keeping the project within a designated scope and timeline. First, to fully test how the relationship to the
nonaccepting individuals affects participants’ response, we would need to set up the experiments such
that participants come in pairs or groups of strangers, classmates, friends, romantic partners, etc. For the
sake of simplicity, this current project only tests nonacceptance instances with online strangers. Second,
to test how participants’ perception of the event mediates the effect of nonacceptance context on
participants’ response, we would need a deception that allows us to measure participants’ perception
without raising any suspicion. However, the approach we’ve chosen for the current project is one where
we deceive the participants by claiming that this study is a playtest for a video game. Within this context

31
of deception, asking participants questions about the nonacceptance experience could potentially reveal
the true intentions of the study, as video game playtests would not ask how players felt during a
nonacceptance instance if it were truly unexpected. Therefore, for the current project outlined below, we
will place our focus primarily on how the number of people, the communication method, and the reasons
given for nonacceptance would affect participants’ mood and motivation, and how that subsequently
affects participants’ behavior towards different people.
The central aims of this project are as follows:
1. Use a simple self-report study to examine whether nonacceptance contexts affect how participants
think/feel/act;
2. Develop and pilot test a new experimental paradigm for social nonacceptance that offers flexible
manipulations of the different dimensions of nonacceptance context;
3. Compare findings using the new experimental paradigm and compare against empirical findings
from previous studies that used different paradigms to validate the new paradigm on a conceptual
level;
4. Compare all the nonacceptance contexts against one another and examine the degree to which
each condition affects the dependent variables;
5. Test the proposed nonacceptance-response process model using the data collected from our new
paradigm.
Each of the aims listed above is addressed with a corresponding following chapter. In Chapter 2, I report
on a vignette study where participants read eight scenarios where they could potentially be rejected by
others. Participants rated how likely they believed they would be rejected, how angry/sad they would feel
if they were rejected, and assessed the intentions of the rejecters. We manipulated participants’ emotional
closeness to the rejecters, number of rejecters, and type of rejection. In Chapter 3, I outline the
development process of our new experimental paradigm – Spider Apocalypse – and I report on pilot
findings. Participants were randomly assigned to one of 10 conditions within Spider Apocalypse, and
subsequently reported on their mood, assessment of the game difficulty, and whether the game felt

32
realistic (both pre- and post-debrief probes were administered). In Chapter 4, I identify the rejection
conditions that best fit each existing experimental paradigms and compare participants’ affective,
cognitive, and behavioral findings against findings from previous research. In Chapter 5, I report on
multiple MANOVAs that compare all nonacceptance conditions on mood, cognition, and behavior. I also
report on a logit loglinear analysis that compares the nonacceptance conditions on binary approach/avoid
behavior. In Chapter 6, I report on partial least squares path modeling that tests the hypothesized
nonacceptance-response process model. Finally, Chapter 7 contains my concluding remarks.

33
Chapter 2 – Examining Nonacceptance Contexts’ Effect on Affect and Cognition: A Self-Report
Vignette Study
As stated in Chapter 1, the study reported here examines whether the social context of a
nonacceptance scenario can affect individuals’ mood and cognition. Drawing upon the foundation laid in
Chapter 1, this study contributes to the literature by offering a systematic manipulation of nonacceptance.
Method
Design Overview
Respondents were recruited from Prolific for a vignette study. Each respondent read eight short
vignettes about social situations where they could potentially be socially rejected. We manipulated
situational factors in a mixed design: 2 (within-subjects, number of rejecters: one-on-one vs group
nonacceptance) by 2 (between-subjects, delivery of nonacceptance: explicit vs. ambiguous
nonacceptance) by 2 (between-subjects, relational proximity: close friends vs. acquaintances). For each of
the vignettes, participants rated: a) how likely they believe they’d be rejected, b) how angry/sad they
would feel if they were rejected, and c) their assessment of the intentions of the rejecter(s). Data
collection was performed and completed in January 2022.
Participants
A total of 400 United States respondents were recruited on Prolific. Forty-nine respondents were
removed for not completing the study. Each respondent received $5 USD. The final sample size was 351
(58.7% females, 71.8% Non-Hispanic White, mean age = 36.3).
Procedure
After reading a brief information sheet about the study and obtaining informed consent,
respondents read eight social situation vignettes, each about a social interaction that could result in social
rejection.
For each of the eight vignettes, we manipulated a) the type of rejection, defined as explicit vs.
ambiguous, and b) the emotional closeness with the rejecters, defined as close friends/peers vs.
acquaintances. Of the eight vignettes, four are individual nonacceptance scenarios (being rejected by one

34
person), and the other four are group nonacceptance scenarios (being rejected by a group of people). The
vignette setup and manipulation are presented in Table 2.
Table 2
Vignette Setup and Manipulation
Vignettes. Respondents were asked to imagine themselves in eight different scenarios outlined in
table 2. See Appendix A for the complete vignettes. Each scenario offers a unique social context, and was
presented in random order. We chose these eight scenarios to provide a range of different settings as a
means to generalize across various realistic social contexts. In the one-on-one rejection condition, the
scenarios outlined interactions between the respondent and one other person; in the group rejection
condition, the scenarios outlined interactions between the respondent and another group of people. In the
explicit rejection condition, the scenarios outlined nonacceptances that were clearly communicated to the
respondent; in the ambiguous rejection condition, the scenarios outlined nonacceptances that were not
communicated to the respondent, but were implied by the lack of communication. In the close
friends/peers condition, the person(s) in the scenarios were described to be emotionally close with the
respondent; in the new acquaintances condition, the person(s) in the scenarios were described to be not
emotionally intimate with the respondent.

35
Response measures. Following each vignette, respondents were asked to assess a) how likely
they believe they’d be rejected, b) how angry and sad they would feel if they were rejected, and c) their
assessment of the intentions of the rejecter(s). See Appendix B for the exact wording of these
measurements.
Results
Analysis Overview
In the following sections, I report results on the three-way mixed MANOVA of Type of Rejection
(between) x Emotional Closeness (between) x Number of Rejecters (within) on the dependent variables
(likelihood of rejection, anger, sadness, and intention interpretation).
Multivariate Tests
There was a significant multivariate effect for Type of Rejection (Wilks’ Lambda = .913, F =
8.209, p < .001), indicating that the type of rejection significantly influenced the combined dependent
variables. Emotional Closeness also showed a significant multivariate effect (Wilks’ Lambda = .913, F =
8.144, p < .001). The within-subjects factor, Number of Rejecters, was also significant (Wilks’ Lambda
= .505, F = 84.244, p < .001). See Table 3 for MANOVA multivariate test results.
Table 3
Three-Way Mixed MANOVA Results of Type of Rejection (between subjects) × Emotional Closeness
(between subjects) × Number of Rejecters (within subjects)
Wilk’s
Lambda F Sig. Partial η2
Between Subjects Effect
Intercept .053 1531.715 <.001 .947
Closeness: Close vs. New .913 8.144 <.001 .087
Type: Explicit vs. Ambiguous .913 8.209 <.001 .087
Closeness × Type .995 0.418 .796 .005
Within Subjects Effect
Number of Rejecters: One vs. Group .505 84.244 <.001 .495
Number of Rejecters × Closeness .982 1.578 .180 .018
Number of Rejecters × Type .974 2.307 .058 .026
Number of Rejecters × Closeness × Type .994 .483 .748 .006

36
Between-Subjects Effects
For Type of Rejection, significant effects were observed for Anger (p < .001), Sadness (p < .001),
and Intention Interpretation (p = .036); but not for Likelihood assessment (p = .777). Such that
respondents were more angry and sad in the ambiguous nonacceptance condition, and were more likely to
attribute the reason for nonacceptance to themselves rather than to an external reason. See Table 4 for
means, SDs, and analyses of variance in likelihood, anger, sadness, and intention interpretation.
Table 4
Means, Standard Deviations, and Analyses of Variance in Likelihood, Anger, Sadness, and Intention
Interpretation for Type of Rejection
Measure Explicit Ambiguous Sig. Partial η2
M SD M SD
Likelihood 3.247 .072 3.217 .079 .777 .000
Anger 2.228 .074 2.826 .081 <.001 .079
Sadness 2.917 .081 3.446 .088 <.001 .053
Intention Interpretation 3.106 .088 3.379 .096 .036 .013
For Emotional Closeness, significant effects were observed for Likelihood (p < .001), Anger (p
= .006), and Sadness (p < .001); but not for Intention Interpretation (p = .713). Such that respondents
perceived a higher likelihood of nonacceptance when interacting with strangers and new acquaintances,
but experienced more anger and sadness when facing a nonacceptance from close friends. See Table 5 for
means, SDs, and analyses of variance in likelihood, anger, sadness, and intention interpretation.
Table 5
Means, Standard Deviations, and Analyses of Variance in Likelihood, Anger, Sadness, and Intention
Interpretation for Emotional Closeness
Measure Close New Sig. Partial η2
M SD M SD
Likelihood 3.027 .075 3.437 .076 <.001 .041
Anger 2.678 .077 2.376 .078 .006 .021
Sadness 3.393 .084 2.971 .086 <.001 .034
Intention Interpretation 3.266 .091 3.218 .093 .713 .000
Within-Subjects Effects
For Number of Rejecters, there was a significant effect on Anger (p < .001), Sadness (p < .001),
and Intention Interpretation (p < .001); but not Likelihood (p = .407). Such that respondents were more

37
angry and sad in the group nonacceptance condition compared to one-on-one nonacceptance condition,
and were more likely to attribute the reason for nonacceptance to themselves rather than to an external
reason. See Table 6 for means, SDs, and analyses of variance in likelihood, anger, sadness, and intention
interpretation.
Table 6
Means, Standard Deviations, and Analyses of Variance in Likelihood, Anger, Sadness, and Intention
Interpretation for Number of Rejecters
Measure Individual Group Sig. Partial η2
M SD M SD
Likelihood 3.213 .054 3.252 .063 .407 .002
Anger 2.089 .054 2.965 .067 <.001 .433
Sadness 2.722 .063 3.642 .069 <.001 .445
Intention Interpretation 2.915 .073 3.570 .068 <.001 .292
Interaction Effects
No significant interactions were observed (ps > .05).
Conclusion
The current chapter reports on how nonacceptance contexts affect mood and cognition. We
manipulated a) the type of rejection (between-subjects) and b) the emotional closeness with the rejecters
(between-subjects) as well as c) the number of rejecters (within-subjects) in a 2 x 2 x 2 mixed design. The
respondents reported on a) how likely they believe they’d be rejected, b) how angry/sad they would feel if
they were rejected, and c) their assessment of the intentions of the rejecter(s).
Respondents were more angry and sad in the ambiguous nonacceptance condition, the close
friends/peers condition, and the group nonacceptance condition. Respondents more readily perceived
rejection happening when it's with new acquaintances than with close friends. Respondents were likely to
take the rejection personally in the ambiguous nonacceptance condition and group nonacceptance
condition. Some interpretations of these findings are as follows – a) We are less familiar with those who
are new acquaintance than those who we deem close to us, hence we perceive the likelihood of being
rejected to be higher with new acquaintances; therefore, when close friends and peers reject us, we are
more distressed over it because it is unexpected. b) When a rejection is ambiguous, it creates room for

38
rumination, which leads to a stronger negative emotional response as well as making us wonder whether
the reason for rejection is one attributed to ourselves. c) Finally, we are more emotionally distressed and
more ready to attribute the reason for rejection to ourselves when there are more rejecters than when there
is only one.
The results may act as preliminary evidence that nonacceptance contexts do in fact affect
individuals’ mood and cognition. However, this study is not without flaws. Vignettes are simulated
situations that do not capture the “in-the-moment” experiences, and many studies have suggested that
self-report hypothetical intentions don’t always match their actual responses to real situations (see: Exum
et al., 2010; Jerolmack & Khan 2014; FeldmanHall et al., 2012; Kormos et al., 2014; Exum & Layana,
2016). Furthermore, these vignettes contained 4 scenarios that are set up as one-on-one and 4 scenarios
that are set up as group, instead of randomizing the number of rejecters with the events scenarios; this
could be a potential confound, as we cannot be certain whether the group condition effects were found
due to the number of rejecters or the nature of the events themselves. Additionally, through self-report,
we were only able to measure respondents’ mood and cognition, but not their behavior. Thus, we are
unable to draw any conclusions on whether the change in mood and cognition would affect the way they
behave post-rejection. We address these caveats in the following chapters.

39
Chapter 3 – Using a Faux Online Multiplayer Game to Simulate Different Nonacceptance Contexts:
Paradigm Development and Pilot Study
To address the caveats presented in Chapter 2, and to further examine whether social contexts
affect individuals’ nonacceptance experience and response, the current chapter reports the development of
a new experimental paradigm – Spider Apocalypse. This new paradigm systematically manipulates the
different dimensions of nonacceptance scenarios in a faux multiplayer online game environment, which
allows respondents to make “in-the-moment” assessment and decisions.
Method
Design Overview
Respondents were recruited from Prolific to test a video game concept. Respondents were
randomly assigned to one of 10 conditions in a between-subjects design: 2 (group vs. one-on-one) by 5
(personal, unknown, implicit, irrelevant, and control). These 10 conditions are outlined below in the
Procedure section, and these conditions are derived from the taxonomy outlined in Chapter 1.
Respondents reported on both pre-game and post-game moods, and their assessment of the game
elements. Data collection was performed and completed in May 2023.
Participants
A total of 200 United States respondents, ages between 18-40, were recruited on Prolific. This
age range was pre-determined to ensure that respondents are familiar with PC gaming. Twenty-one
respondents were removed for not completing the study. By completing this study, each respondent
received $4 USD. The final sample size was 179 (50% females, 67% Non-Hispanic White, mean age =
28.6).
Spider Apocalypse
Spider Apocalypse is an arcade-style action-defense game made with Unity Engine that takes
place online on itch.io. Spider Apocalypse is inspired by classic arcade defense games like Space Invader,
Tank90, and Journey of the Prairie King from Stardew Valley.

40
Figure 13
Games that Inspired Spider Apocalypse
Note. The games are: (a) Space Invader, (b) Tank90, and (c) Journey of the Prairie King from Stardew
Valley.
Although Spider Apocalypse appears to be multi-player, the other “players” are, in fact, nonplayer characters (NPCs; or more commonly known as “bots”). In Spider Apocalypse, each player selects
a gender-neutral avatar and inputs an in-game name. Then, the players are instructed to cooperate and
protect their shared house (placed in the center of the screen) from incoming spiders (generated from the
edges of the screen, moving towards the center). Players can use the arrow keys on their keyboards to
move their avatars around, and the space bar on their keyboards to shoot arrows at the spiders. The arrows
only cause damage to the spiders, and do not harm the other players. If a spider successfully makes it past
the players into the house, the shared health of players drops by one heart. The game does not end even if

41
all health has been depleted, and the health bar is merely there to motivate players to protect the house. If
played till completion, the game lasts for a total of 180 seconds. This time duration is typical for a single
level of action-defense arcade game.
Procedure
After reading a brief information sheet about the study and obtaining informed consent,
respondents filled out a brief mood questionnaire derived from the PANAS (Watson et al., 1988). We
selected two moods each to represent Anger (angry, irritable), Sadness (sad, lonely), and Happiness
(happy, excited). Then, respondents clicked on a link to launch the itch.io web page for Spider
Apocalypse. First, respondents inputted a desired in-game name and selected a gender-neutral avatar to
represent themselves in-game. Then, a tutorial page was displayed, explaining the controls of the game:
arrow keys to move the avatar, spacebar to shoot arrows. The goal of the game was also displayed to the
participant: “Work together with other players to protect your house from spiders!”. Once the participant
had read the tutorial and goal, they entered a “queue” to be matched with other online “players”.
Figure 14
Introduction Screens of Spider Apocalypse

42
The above figure shows vertical slice (playable rough draft) of Spider Apocalypse introduction screens.
(a) depicts the avatar-selection and name-input screen. (b) depicts an example of chosen avatar and
player’s input name. (c) depicts the displayed game goal and tutorial. (d) depicts the “matching” screen
participants will see. These screen captures come from a version that is primarily used to test out UI and
logic flow, thus the actual design and text are subject to change before formal testing begins.
Figure 15
Example Gameplay Screen of Spider Apocalypse
The figure above shows the gameplay screen of Spider Apocalypse. Players’ shared house is placed in the
center of the map. The participant “dragonqueen” and the NPC “Gray” are free to move around and shoot
arrows at incoming spiders. Top left shows the players’ shared health in hearts; players would lose a heart
for each spider that reaches the house. The health indication is arbitrary, placed only to motivate
participants to play the game. Score shows the total number of spiders defeated. Time counts down on the
remaining time left in this game. Bottom left shows the chat box where participants will receive NPC
messages.

43
Many elements of this game can be fine-tuned, including the rate that the spiders are generated,
NPC skill and accuracy in shooting the spiders, total time for the game, number of hearts between the
players, the NPC messages sent to the participant, etc.
The respondents were randomly assigned to one of 10 conditions in a 2 (one-on-one vs. group) by
5 (nonacceptance type: personal, irrelevant, unknown, implicit, and control) between-subjects design.
Figure 16
Spider Apocalypse Experimental Conditions
In the personal nonacceptance condition, the respondent was loaded into the game with other
NPC(s). The game started, and players were free to move around / shoot arrows. At around the 90
seconds mark, the NPC(s) typed in the chat box: “you suck at this game, we are losing because of [you /
“participant in-game name”]. After the message was sent, the NPC(s) exited the game, leaving the
participant by themself to defend the house. A few seconds later, the game was paused and a system
message was displayed on-screen: “[NPC in-game name] has left the game. This game has been
terminated due to an insufficient number of players.” This condition is designed to provide a personal
nonacceptance experience – the participant is rejected by the NPC(s) where the reason is clearly attributed
to the participant themself.

44
In the unknown (explicit but without reason) nonacceptance condition, the respondent was loaded
into the game with other NPC(s). The game started, and players were free to move around / shoot arrows.
At around the 90 seconds mark, the NPC typed in the chat box: “bye”; then the NPC(s) exited from the
game, leaving the participant by themself to defend the house. A few seconds later, the game was paused
and a system message was displayed on-screen: “[NPC in-game name] has left the game. This game has
been terminated due to an insufficient number of players.” This condition is designed to communicate a
clear nonacceptance (“the other player left”) but without providing any reason. The ambiguity allows the
respondent to freely interpret why the nonacceptance occurred.
In the implicit nonacceptance condition, the participant was loaded into the game with other
NPC(s). The game started, and players were free to move around / shoot arrows. At around the 90
seconds mark, the NPC(s) exited from the game, leaving the participant by themself to defend the house.
A few seconds later, the game was paused and a system message was displayed on-screen: “This game
has been terminated due to an insufficient number of players.” This condition is designed to provide an
implicit nonacceptance experience – the participant is unsure whether a nonacceptance has taken place, or
if something else had happened. It adds an additional layer of ambiguity to the experience – not only are
the respondents unsure as to why the other players have left, the respondents aren’t even certain that the
nonacceptance has actually taken place.
In the irrelevant nonacceptance condition, the respondent was loaded into the game with other
NPC(s). The game started, and players were free to move around / shoot arrows. At around the 90
seconds mark, the NPC(s) suddenly exited the game without an explanation, leaving the participant by
themself to defend the house. A few seconds later, the game was paused and a system message was
displayed on-screen: “[NPC in-game name] has left the game due to a network issue. This game has been
terminated due to an insufficient number of players.” This condition is designed to provide an irrelevant
nonacceptance experience – the participant is separated from the NPC(s) due to an external reason
completely unrelated to the participant themself.

45
In the control condition, the respondent played the game with their NPC(s) till the timer ran out,
completing the game without any disruption. This condition was used as control as it is an inclusion
experience – the respondent cooperates with other player(s) as expected.
For all conditions listed above, a one-on-one version and a group-version were offered. In the
one-on-one version, the above conditions occurred with only one NPC, such that the nonacceptance only
came from one other “player”. In the group version, the above conditions occurred with three NPCs, such
that the nonacceptance came from several other “players” at once. This entire study provided a total of 10
conditions, consisting of eight experimental conditions and two control conditions.
After the game ended, participants were asked to assess their post-game moods, gameplay
experience, and evaluation of other players. They were asked a few game-design questions (Was the
game too easy/difficult? Did you find the controls intuitive? Etc.) to ensure that the game controls and
difficulty balance are adequate.
Then, participants underwent the behavioral measures. Respondents first completed a partner
rating task, where they provided recommendations on any adjustment to their game partner’s study
compensation (on a scale of -$4 to +$4 in $1 increments, anchored with a neutral “no change” option in
the middle) and platform reputation score (on a scale of 1-5 from decrease score to increase score,
anchored with a neutral “no change” option in the middle). Next, participants decided between playing
Spider Apocalypse again with new players, or playing the game again but by themselves.
Results
Analyses Overview
First, I report results on the gameplay assessment descriptives as a means to demonstrate the
playability of Spider Apocalypse. Second, I report the thematic analysis results on the deception probe
free responses. Third, I report results on the two-way MANOVA of Type of Rejection x Number of
Rejecters on the dependent variables.

46
Gameplay Assessment
On a scale of 1 (Very difficult) to 5 (Very easy), respondents generally felt that Spider
Apocalypse was slightly difficult (M = 2.84, SD = 1.077). On a scale of 1 (Strongly disagree) to 5
(Strongly agree), respondents were generally neutral on assessing how fun the gameplay experience was
(M = 2.97, SD = 1.178). On a scale of 1 (Strongly disagree) to 5 (Strongly agree), respondents reported
that the game was somewhat stressful (M = 3.2, SD = 1.224). Finally, on a scale of 1 (Far too little) to 5
(Far too much), respondents felt that the amount of spiders was trending too much (M = 3.31, SD = .721).
Overall, all of these metrics aligned with our design intentions, and we decided to not make changes to
the gameplay experience for subsequent studies.
Deception Probe
Generally, respondents reported no suspicion towards the NPCs. When explicitly asked about
whether the NPCs felt realistic, most respondents reported that they felt the multiplayer interaction
experience was realistic. However, a few respondents reported suspicion during pre-probe that the other
players are bots, though these respondents were in the minority. To address any concerns regarding
participants being suspicious of the bot players, we look toward Zadro et al.’s findings (2004) that even
when participants were explicitly told they’re socially rejected by computerized characters, they still
experienced the same amount of emotional distress as the participants that were told they were rejected by
real people.
Mood
We conducted a 2 (one-on-one vs. group conditions) by 5 (personal, unknown, implicit,
irrelevant, and control conditions) MANOVA on the mood measures – anger, sadness, and happiness.
The two-way MANOVA revealed a significant main effect for type of nonacceptance (Wilks’
Lambda = .831, p = .002) and number of rejecters (Wilks’ Lambda = .937, p = .012). No interaction effect
was observed (Wilks’ Lambda = .915, p = .238). See Table 7 for MANOVA multivariate test results.

47
Table 7
Two-Way MANOVA Results of Number of Rejecters × Type of Rejection on Change in Anger, Sadness,
and Happiness
Effect
Wilk’s
Lambda F Sig. Partial η2
Intercept .691 24.876 <.001 .309
Number of Rejecters: Individual vs. Group .937 3.754 .012 .063
Type of Rejection .831 2.633 .002 .060
Number of Rejecters × Type of Rejection .915 1.262 .238 .029
For Type of Rejection, significant effects were observed for Anger (p = .010) and Happiness (p
< .001), but not Sadness (p = .373). Personal Rejection condition produced a significantly higher increase
in anger than Irrelevant condition (p = .017) and Control condition (p = .016). Personal Rejection
condition also produced a significant decrease in happiness than Unknown condition (p < .001), Implicit
condition (p = .034), Irrelevant condition (p = .004), and Control condition (p = .001). See Table 8 for
means, SDs, and analyses of variance in change in anger, sadness, and happiness; see Table 9 for pairwise
comparison within the rejection types.
Table 8
Means, Standard Deviations, and Analyses of Variance in Change in Anger, Sadness, and Happiness for
Rejection Types
Measure Personal Unknown Implicit Irrelevant Control Partial η2
M SD M SD M SD M SD M SD
Anger 1.518 .243 .635 .199 .724 .277 .450 .230 .432 .235 .076*
Sadness -.016 .184 -.421 .151 -.260 .210 -.206 .210 -.472 .178 .025
Happiness -1.249 .261 .336 .214 -.075 .297 .044 .247 .178 .252 .129***

48
Table 9
Pairwise Comparison of Means Within Rejection Types on Change in Anger, Sadness, and Happiness
ANGER Personal Unknown Implicit Irrelevant Control
Personal
Unknown .883
Implicit .794 -.089
Irrelevant 1.068* .185 .274
Control 1.085* .202 .292 .018
SADNESS Personal Unknown Implicit Irrelevant Control
Personal
Unknown .406
Implicit .244 -.162
Irrelevant .191 -.215 -.053
Control .457 .051 .213 .266
HAPPINESS Personal Unknown Implicit Irrelevant Control
Personal
Unknown -1.585***
Implicit -1.174* .410
Irrelevant -1.293** .292 -.118
Control -1.427** .157 -.253 -.135
Note. Mean difference calculated using Row – Column. For example, the mean difference in Anger
of .883 between Personal Rejection and Unknown Reason is calculated using Personal – Unknown.
For Number of Rejecters, significant effects were observed for Anger (p = .035) and Happiness (p
= .002), but not Sadness (p = .941). One-on-one nonacceptance resulted to a higher increase in anger (p
= .035), and a higher decrease in happiness (p = .002). See Table 10 for means, SDs, and analyses of
variance in change in anger, sadness, and happiness.
Table 10
Means, Standard Deviations, and Analyses of Variance in Change in Anger, Sadness, and Happiness for
Number of Rejecters
Measure Individual Group Sig. Partial η2
M SD M SD
Anger .978 .150 .525 .151 .035 .026
Sadness -.269 .114 -.281 .114 .941 .000
Happiness -.514 .161 .208 .162 .002 .056
No significant interactions were observed (ps > .05).

49
Evaluation and Behavioral Reaction
We conducted a 2 (one-on-one vs. group conditions) by 5 (personal, unknown, implicit,
irrelevant, and control conditions) MANOVA on the evaluation and behavior reaction measures – partner
evaluation, compensation adjust, and rating adjustment.
The two-way MANOVA revealed a significant main effect for type of nonacceptance (Wilks’
Lambda = .839, p = .003) and number of rejecters (Wilks’ Lambda = .914, p = .002). No interaction effect
was observed (Wilks’ Lambda = .911, p = .203). See Table 11 for MANOVA multivariate test results.
Table 11
Two-Way MANOVA Results of Number of Rejecters × Type of Rejection on Partner Evaluation,
Compensation Adjustment, and Rating Adjustment
Effect
Wilk’s
Lambda F Sig. Partial η2
Intercept .044 1198.157 <.001 .956
Number of Rejecters: Individual vs. Group .914 5.258 .002 .086
Type of Rejection .839 2.520 .003 .057
Number of Rejecters × Type of Rejection .911 1.322 .203 .031
For Type of Rejection, significant effects were observed for partner evaluation (p = .002),
compensation adjustment (p < .001), and rating adjustment (p < .001). Personal Rejection condition
produced a significantly more negative partner evaluation than Irrelevant condition (p = .005) and Control
condition (p = .007), a significantly more punishing compensation recommendation than Irrelevant
condition (p = .010) and Control condition (p = .003), and a significantly lower rating recommendation
than Irrelevant condition (p = .007) and Control condition (p = .002). Furthermore, Unknown Reason
condition also produced a significantly more punishing compensation recommendation than Irrelevant
condition (p = .024) and Control condition (p = .007). See Table 12 for means, SDs, and analyses of
variance in change in anger, sadness, and happiness; see Table 13 for pairwise comparison within the
rejection types.

50
Table 12
Means, Standard Deviations, and Analyses of Variance in Partner Evaluation, Compensation Adjustment,
and Rating Adjustment
Measure Personal Unknown Implicit Irrelevant Control Partial η2
M SD M SD M SD M SD M SD
P Eval 2.369 .163 2.743 .134 2.909 .186 3.169 .155 3.152 1.58 .093**
Comp Adj -.551 .344 -.290 .282 .799 .392 1.037 .326 1.214 .332 .127***
Rate Adj 2.637 .188 2.997 .154 3.448 .214 3.531 .178 3.635 .182 .111***
Table 13
Pairwise Comparison of Means Within Rejection Types on Partner Evaluation, Compensation
Adjustment, and Rating Adjustment
P EVAL Personal Unknown Implicit Irrelevant Control
Personal
Unknown -.374
Implicit -.540 -.166
Irrelevant -.800** -.426 -.260
Control -.784** -.409 -.243 .016
COMP ADJ Personal Unknown Implicit Irrelevant Control
Personal
Unknown -.261
Implicit -1.350 -1.089
Irrelevant -1.588* -1.328* -.239
Control -1.756** -1.504** -.415 -.176
RATE ADJ Personal Unknown Implicit Irrelevant Control
Personal
Unknown -.359
Implicit -.811 -.452
Irrelevant -.894** -.535 -.083
Control -.998** -.639 -.187 -.104
For Number of Rejecters, significant effects were observed for partner evaluation (p = .003),
compensation adjustment (p = .001), and rating adjustment (p < .001). One-on-one rejection condition
resulted to significantly more negative partner evaluation, more punishing compensation
recommendation, and lower rating recommendation than the group rejection condition. See Table 14 for
means, SDs, and analyses of variance in change in Partner Evaluation, Compensation Adjustment, and
Rating Adjustment.

51
Table 14
Means, Standard Deviations, and Analyses of Variance in Partner Evaluation, Compensation Adjustment,
and Rating Adjustment
Measure Individual Group Sig. Partial η2
M SD M SD
Partner Evaluation 2.654 .101 3.083 .102 .003 .051
Compensation Adjust -0.045 .212 0.929 .214 .001 .058
Rating Adjust 2.931 .116 3.568 .117 <.001 .081
Social Approach and Avoid
A hierarchical loglinear model was used to analyze the three-way interaction between Type of
Rejection, Number of Rejecters, and Subsequent Social Approach/Avoidance Behavior. The analysis
began with a model including all main effects, which was systematically built up to include two-way and
three-way interactions.
The inclusion of the three-way interaction did not significantly improve model fit (χ²(4) = 2.354,
p = .671). Neither did the two-way effects of Type of Rejection x Approach/Avoid (χ²(4) = 5.185, p
= .269) and Type of Rejection x Number of Rejecters (χ²(4) = 5.230, p = .264). The final model (χ²(1) =
4.493, p = .034) included only a significant two-way interaction between Number of Rejecters and
Subsequent Social Approach/Avoidance Behavior.
The two-way interaction effect of Number of Rejecters x Subsequent Social Approach/Avoidance
Behavior was significant (β = -.176, p = .025), indicating that respondents were more likely to choose to
play with other online players if they were previously in the group condition.
Conclusion
The current chapter introduces a novel experimental paradigm – Spider Apocalypse – to study the
impact of various social contexts on individuals’ experiences of nonacceptance and their subsequent
responses. The design of the study, which involved creating an arcade-style action-defense game, allowed
for controlled manipulation of different aspects of nonacceptance scenarios within a multiplayer online
game setting. By utilizing 10 distinct conditions that varied in terms of group size (one-on-one vs. group)

52
and nonacceptance type (personal, unknown, implicit, irrelevant, and control), this study aimed to provide
a comprehensive understanding of how social context influences respondents’ reactions.
Initial results from the gameplay assessment suggest that Spider Apocalypse is a suitable tool for
this field of research. Respondents found the game slightly difficult but still enjoyable, slightly stressful,
and with an appropriate number of spiders as a challenge. Additionally, the thematic analysis of the
deception probe responses indicated that most participants did not express suspicion regarding the NPCs’
authenticity, supporting the game’s ability to create a realistic social environment.
The two-way MANOVA analysis revealed significant main effects of both the number of
rejecters and the type of nonacceptance on various dependent variables. Further analyses revealed that
Personal Rejection was the most robust rejection manipulation for mood and behavioral reactions
compared to other conditions; and one-on-one rejection condition resulted in more negative moods, more
punishing behavioral reactions, and more social avoidance tendencies. These findings provide initial
evidence that nonacceptance manipulations had a significant impact on respondents’ emotional responses,
evaluations of their game partners, and behavioral changes within the game.
Overall, the paradigm proves to work as intended, and successfully elicits changes in
mood/cognition/behavior in respondents. To further validate our paradigm, we set out to compare
findings from this study using Spider Apocalypse with findings from previous research that use existing
paradigms. We compare these findings in the following chapter.

53
Chapter 4 – Comparing Findings between Spider Apocalypse and Existing Paradigms: A
Conceptual Validation
The current chapter reports on empirical findings using Spider Apocalypse, and the comparison
of these findings with findings from past research. This is in part to further validate our paradigm – that
our conceptualization of the different rejection contexts maps onto past conceptualization.
Select Past Findings
In the process of selecting existing findings, I combed through the 78 publications I could find in
which social nonacceptance was experimentally manipulated. Of the list of publications, I selected 19
papers to compare Spider Apocalypse to. These 19 papers span across three different types of
nonacceptance – 1) individual personal nonacceptance, 2) implicit group nonacceptance (ambiguous as to
whether nonacceptance has happened), and 3) unknown group nonacceptance (the reason for the
nonacceptance is unknown to the participant). The reason why only 19 papers were selected out of 78, is
because many of the unselected papers had measurements that are not comparable to the ones used in the
current chapter’s study. Some of these measures include obedience (operationalized as the quality of
photos taken of campus; Riva et al., 2014), memory of social events (operationalized as the number of
social events recalled from reading a diary entry; Gardner et al., 2000), behavioral mimicry (Lakin et al.,
2008), identifying pictures of real vs. fake smiles (Bernstein et al., 2008), self-regulation (operationalized
as consumption of unhealthy foods; Oaten et al., 2008), dishonest intentions and behaviors
(operationalized as lying to get more money; Poon et al., 2013), etc. Therefore, only 19 papers were
deemed to be comparable to Spider Apocalypse and the measurements we use for this chapter’s
experiment.
Individual Personal Nonacceptance
Four selected papers used individual personal nonacceptance as their experimental manipulation.
These studies used the Video Exchange paradigm or its variants. In these studies, participants were
explicitly told that their study-partner chose not to work with them on a task. Within this nonacceptance
context, it was found that participants gave their study-partner more negative evaluations (Ayduk et al.,

54
1999; Maner et al., 2007), less monetary rewards for a task (Maner et al., 2007), more hot sauce in a
subsequent task (Ayduk et al., 2008), and more negative job rating (Rajchert & Winiewski, 2017). In
most of these selected studies, mood was not a variable of interest. Even when it was, post-rejection mood
was not found to be significantly different from the control condition (Maner et al., 2007).
Implicit Group Nonacceptance
Ten selected papers used implicit group nonacceptance as their experimental manipulation. These
studies used Cyberball, Euroball, O-cam, O-train, and the chatroom paradigm. In these studies,
participants experienced a very subtle nonacceptance where they’re uncertain whether the nonacceptance
has occurred. With this nonacceptance context, it was found that participants exhibited increased anger
(Smith & Williams, 2004; Chow et al., 2008; Chester & DeWall, 2017), increased sadness (Chow et al.,
2008), decreased positive mood and/or increased negative mood (Smith & Williams, 2004; van Beest &
Williams, 2006; Gonsalkorale & Williams, 2007; Carter-Sowell et al., 2008; Rajchert et al., 2017;
Gamian-Wilk et al., 2020), gave their rejecters worse ratings (Gonsalkorale & Williams, 2007), sent more
angry texts to the original rejecters (Smith & Williams, 2004), assigned them unappealing snacks for a
subsequent task (Chow et al., 2008), jabbed more pins into voodoo dolls that represented the rejecters
(Chester & DeWall, 2017). When measuring their behavior and cognition towards new participants that
were not their original rejecters, participants exhibited prosocial motivations like giving wrong answers to
appease the new group (Williams et al., 2000) and giving more money to charity (Carter-Sowell et al.,
2008), as well as antisocial motivations like giving more hot sauce to new participants in a subsequent
task (Rajchert et al., 2017). When measuring whether participants choose to approach new people or
withdraw in a subsequent social context, Ren et al. (2016) first found that participants indicated a higher
desire to join a new group than they want to withdraw; though this effect was not replicated in a
subsequent study published in the same paper. Interestingly, this effect was found in a later study
(Gamian-Wilk et al., 2020) where participants exhibited higher willingness to volunteer for an
organization with a group of new folks rather than volunteering alone.

55
Among these studies, it was found that needs mediated mood (van Beest & Williams, 2006), and
anger mediated aggressive behavior towards original rejecters (Chow et al., 2008). However, in a more
recent study, anger was not found to mediate the rejection-aggression link using the same type of
nonacceptance paradigm (Rajchert et al., 2017).
Unknown Group Nonacceptance
Five selected papers used unknown group nonacceptance as their experimental manipulation.
These studies used the Get Acquainted paradigm, where the nonacceptances were clear and explicit, but
the reason as to why the nonacceptance occurred was unknown to the participants. Due to the nature of
the paradigm (“nobody chose to work with you, so you will be doing another task”), most of these studies
examined participants’ response towards new individuals. However, the findings of these studies mostly
do not agree with one another. Namely, it was found that participants gave worse noise blasts to new
participants (Twenge et al., 2001; Twenge & Campbell, 2003), volunteered less in subsequent tasks
(Twenge et al., 2007), yet viewed new participants as more attractive and sociable (Maner et al., 2007).
Among these studies, it was found that mood did not mediate subsequent behavior (Twenge et al.,
2001; Twenge et al., 2007), and that post-nonacceptance mood was not significantly different from
control condition (Twenge et al., 2001; Maner et al., 2007).
Method
Design Overview
Respondents were recruited from Prolific to test a video game concept and randomly assigned to
either one of the experimental conditions or one of the control conditions. Respondents then reported on
both pre-game and post-game moods, their evaluation of their game partner(s), their behavior towards
their game partner(s), and whether they chose to approach or withdraw from a subsequent social situation.
Data collection was performed and completed in July 2023.
Participants
A total of 400 United States respondents ages between 18-40 were recruited from Prolific. This
age range was pre-determined to ensure that respondents were familiar with PC gaming. Twenty-three

56
respondents were removed for not completing this study. By completing the study, each respondent
received $4 USD. The final sample size was 377 (51.1% females, 65.7% Non-Hispanic White, mean age
= 29.43). For this chapter, we selected participants that were assigned to one of the conditions that
correspond to previous literature, so to compare the findings with previous findings. We selected the
individual personal nonacceptance condition (N = 39) and its corresponding “control” irrelevant personal
nonacceptance condition (N = 35). We chose the irrelevant personal nonacceptance condition (where the
game was interrupted due to a network error) because the Video Exchange paradigm used an irrelevant
condition as their control (where the study-partner had to leave due to an external reason). We also
selected the implicit group nonacceptance condition (N = 36), the unknown group nonacceptance
condition (N = 25), and the inclusion control condition (N = 36) where respondents were able to complete
the game with other players. The final sample size that we’ll be examining in this chapter is 171.
Procedure
The same procedure used in Chapter 3 was used for this current study. Respondents read a brief
information sheet and gave consent, filled out a brief mood questionnaire derived from the PANAS that
included angry, irritable, sad, lonely, happy, and excited. Then, respondents played Spider Apocalypse.
The respondents were randomly assigned to one of 10 conditions in a 2 (one-on-one vs. group) by 5
(nonacceptance type: personal, irrelevant, unknown, implicit, and control) between-subjects design. For
the purpose of comparison, we will only be focusing on the selected conditions and their analyses.
After the game ended, participants were asked to assess their post-game moods, gameplay
experience, and evaluation of other players. Then, participants underwent the behavioral measures.
Respondents first completed a partner rating task, where they provided recommendations on any
adjustment to their game partner’s study compensation (on a scale of -$4 to +$4 in $1 increments,
anchored with a neutral “no change” option in the middle) and platform reputation score (on a scale of 1-5
from decrease score to increase score, anchored with a neutral “no change” option in the middle). Next,
participants decided between playing Spider Apocalypse again with new players, or playing the game
again but by themselves.

57
Before being debriefed on the purpose of the study, a deception probe question was administered
in the form of an open text box. After debriefing, participants were asked to explicitly assess how realistic
the NPCs felt to them during the game.
Comparison Results
Personal One-on-One Nonacceptance vs. Irrelevant Personal Nonacceptance
We ran a series of t-tests to examine whether there was a mean difference between individual
personal nonacceptance vs. irrelevant personal nonacceptance. As stated in the section above, we chose
the irrelevant personal nonacceptance condition (where the game was interrupted due to a network error)
as the control comparison because the Video Exchange paradigm studies all used an irrelevant condition
as their control (where the study-partner had to leave due to an external reason).
Previous research has found that personal one-on-one nonacceptance caused respondents to give a
more negative evaluation of their rejecter (Ayduk et al., 1999; Maner et al., 2007). We found a similar
effect, that our respondents gave a more negative evaluation of their play partners (t(72) = -4.240, p
< .001).
Previous research has also found that personal one-on-one nonacceptance caused antisocial
behaviors among participants towards their original rejecters (less monetary rewards for a task (Maner et
al., 2007), more hot sauce in a subsequent task (Ayduk et al., 2008), and more negative job rating
(Rajchert & Winiewski, 2017)). We found similar effects, that our respondents were more antisocial
towards their original rejecters by suggesting a decrease in their reputation scores (t(72) = -4.368, p
< .001) and reducing their study compensation (t(72) = -4.300, p < .001).
Lastly, Maner et al. (2007) found that post-rejection mood was not significantly different from
control condition. Here, we observed a different finding – respondents that underwent the personal oneon-one nonacceptance condition reported significantly increased anger (t(72) = 2.511, p = .014), sadness
(t(72) = 2.629, p = .010), and decreased happiness (t(72) = -2.008, p = .048). If we aggregate all mood
measures into one single mood index (reverse-coding positive affect), respondents in the personal one-on-

58
one nonacceptance reported significantly higher negative mood than respondents in the irrelevant
nonacceptance condition (t(72) = 2.768, p = .007).
Implicit Group Nonacceptance vs. Inclusion Control
We ran a series of t-tests and a chi-square test to examine whether there was a mean difference
between implicit group nonacceptance vs. control condition on mood, cognition, and behavior. We then
conducted a mediation analysis to examine if anger mediates the rejection-aggression link, as previously
found by Chow et al. (2008) but not replicated by Rajchert et al. (2017).
Previous research has found that implicit group nonacceptance caused respondents to give a more
negative evaluation of their rejecters (Gonsalkorale & Williams, 2007). However, we did not replicate
this effect – respondents in our study did not give significantly different partner evaluation between the
two conditions (t(70) = -1.262, p = .221).
Previous research has also found that participants exhibited increased anger (Smith & Williams,
2004; Chow et al., 2008; Chester & DeWall, 2017) and sadness (Chow et al., 2008), and increased
general negative mood (Smith & Williams, 2004; van Beest & Williams, 2006; Gonsalkorale & Williams,
2007; Carter-Sowell et al., 2008; Rajchert et al., 2017; Gamian-Wilk et al., 2020). Our study partially
replicated these findings – a) implicit group nonacceptance did increase respondents’ anger (t(70) =
2.386, p = .020), b) but implicit group nonacceptance did not increase respondents’ sadness (t(70) = .644,
p = .521), c) though general negative mood was found to significantly increase among respondents in the
implicit group nonacceptance condition (t(70) = 3.034, p = .003).
Additionally, previous research has found that participants exhibited antisocial behaviors towards
their original rejecters (sent more angry texts to the original rejecters (Smith & Williams, 2004), assigned
them unappealing snacks for a subsequent task (Chow et al., 2008), jabbed more pins into voodoo dolls
that represent the rejecters (Chester & DeWall, 2017)). We found similar effects, but only in respondents’
recommended study compensation adjustment, where respondents in the implicit group nonacceptance
condition suggested a decrease in their play partners’ study compensation (t(70) = -2.178, p = .033).

59
Rating adjustment was not found to be significantly different between nonacceptance and control
conditions (t(70) = -.796, p = .429).
When it comes to post-rejection response towards new individuals, previous research has found
conflicting findings between whether participants were more prosocial vs antisocial (Williams et al.,
2000; Carter-Sowell et al., 2008; Rajchert et al., 2017), as well as whether they were more likely to
approach or avoid a subsequent social situation (Ren et al., 2016; Gamian-Wilk et al., 2020). In our study,
we were unable to test whether respondents would be prosocial vs. antisocial towards new individuals, as
it was not included in our experimental design; but we were able to test if respondents were more likely to
socially approach or withdraw. To do so, we conducted a chi-square test of independence to examine the
relation between nonacceptance-context and subsequent social approach/avoid behavior. The relation
between these variables was found to be insignificant (X^2(1, N = 72) = .229, p = .633) – whether
respondents chose to approach or withdraw did not differ between implicit group nonacceptance
condition vs. control condition. This aligns with the findings from Ren et al. (2016) but not with the
findings from Gamian-Wilk et al. (2020).
Lastly, Chow et al. (2008) found that anger mediated the rejection-aggression link, whereas
Rajchert et al. (2017) did not find this effect. We tested this mediation hypothesis using Conditional
Process analysis (Hayes, 2022) to estimate the direct and indirect effects of implicit group nonacceptance
on aggressive behavior. Table 15 presents the estimated raw parameters for predicting anger from the
nonacceptance vs. control conditions (Step 1), and for predicting subsequent aggressive behavior (Step 2).
Parameter estimates for predictors, standard errors, t-statics, and p-values are based on 10,000 bootstrap
samples.

60
Table 15
Estimated Regression Weight Parameters and Standard Errors for Predicting Aggression in Two Steps
STEP 1: Predicting Anger
Predictors B SE(B)
Constant -.083 .154
Implicit vs. Control .528* .220
Adjusted R2 .062
STEP 2: Predicting Aggression
Model 1 Model 2
Predictors B SE(B) B SE(B)
Constant .631*** .134 .613*** .134
Implicit vs. Control -.347 -.344 -.245 .222
Anger -.198 .220
Adjusted R2 .031 .043
The results of the mediation analysis are summarized in the figure below. Respondents that were
in the implicit group nonacceptance condition reported increased anger (a = .528, p = .016). However,
angered respondents did not respond any more or less aggressively (b = -0.193, p = .123). A bootstrap
confidence interval for the indirect effect (ab = -.101, p = .213) was also not significant. Finally, there
was no indication that the nonacceptance manipulation influenced aggressive behavior (c’ = -.245, p
= .269). Overall, our findings align with that of Rajchert et al. (2017) – anger did not mediate the effect
between nonacceptance and aggressive behavior.
Figure 17
Mediation Model of Implicit vs. Control Conditions on Aggression Mediated by Anger
Note. Aggression operationalized as the combination of Compensation Adjustment and Rating
Adjustment towards play partners, as those measures allow respondents to punish (aggress) towards the
NPCs.

61
Unknown Group Nonacceptance vs. Inclusion Control
We ran a chi-square test to examine whether respondents from unknown group nonacceptance vs.
control condition chose differently in a subsequent social approach/avoid decision. We then conducted a
series of t-tests to examine whether post-rejection mood differed between groups, and conducted a
mediation analysis to examine whether post-rejection mood mediated the effect that rejection has on
social approach/avoidance.
Previous research has found that participants gave worse noise blasts to new participants (Twenge
et al., 2001; Twenge & Campbell, 2003), volunteered less in subsequent tasks (Twenge et al., 2007), yet
viewed new participants as more attractive and sociable (Maner et al., 2007). In our study, we were
unable to test whether respondents would be prosocial vs. antisocial towards new individuals, as it was
not included in our experimental design; but we were able to test if respondents were more likely to
socially approach or withdraw. To do so, we conducted a chi-square test of independence to examine the
relation between nonacceptance-context and subsequent social approach/avoid behavior. The relation
between these variables was found to be insignificant (X^2(1, N = 61) = .159, p = .690) – whether
respondents chose to approach or withdraw did not differ between the unknown group nonacceptance
condition vs. control condition. This lack of significance is perhaps reflected in the inconsistent findings
from past research, such that participants can react in many different ways, potentially depending on a)
what subsequent behavior was offered as options, and b) individual difference variables. This warrants a
deeper dive in the future, but will not be further explored within the current set of works.
Previous research has also failed to find significance in post-rejection mood (Twenge et al., 2001;
Maner et al., 2007). To our surprise, our findings replicated this lack of significance. Respondents in the
unknown group nonacceptance condition did not report significantly different levels of anger (t(37.348) =
1.928, p = .061), sadness (t(59) = -.304, p = .762), or happiness (t(59) = -1.656, p = .103). Thus, when
mood was aggregated into a single negative mood index, it also did not exhibit a significant effect
between the groups (t(59) = 1.780, p = .080).

62
Lastly, previous research did not find a mediation effect of aggregated mood between
nonacceptance and subsequent social behavior (Twenge et al., 2001; Twenge et al., 2007). We tested this
hypothesis using Conditional Process analysis (Hayes, 2022) to estimate direct and indirect effects of
unknown group nonacceptance on subsequent social approach/avoid behavior. Table 16 presents the
estimated raw parameters for predicting aggregated mood index (Step 1), and for predicting likelihood of
approach/avoid (Step 2). Parameter estimates for predictors, standard errors, t-statics, and p-values are
based on 10,000 bootstrap samples.
Table 16
Estimated Regression Weight Parameters and Standard Errors for Predicting Social Approach-Avoid in
Two Steps
STEP 1: Predicting Negative Mood
Predictors B SE(B)
Constant -0.994** .364
Unknown vs. Control 1.144 .671
Adjusted R2 .035
STEP 2: Predicting Social Approach-Avoid
Model 1 Model 2
Predictors B SE(B) B SE(B)
Constant .611*** .081 .606*** .085
Unknown vs. Control -.051 .128 -.045 .133
Negative Mood -.006 .026
Adjusted R2 -.014 -.031
The results of the mediation analysis is summarized in the figure below. All paths were
insignificant. In some way, our findings align with that of Twenge et al. (2001) and Twenge et al. (2007)
where they did not find a significant mediation effect.

63
Figure 18
Mediation Model of Unknown vs. Control Conditions on Social Approach vs. Avoid Mediated by Negative
Mood
Conclusion
The current chapter reports on empirical findings using Spider Apocalypse as manipulation, and
compares these findings with previous research. We selected a range of studies from existing literature,
and categorized them into three distinct types of nonacceptance: personal one-on-one nonacceptance,
implicit group nonacceptance, and unknown group nonacceptance. Our findings generally aligned with
previous findings, but had trouble replicating certain findings when previous literature had conflicting
evidence.
For individual personal nonacceptance, we observed parallels with previous research.
Respondents in the Spider Apocalypse’s one-on-one personal nonacceptance condition gave more
negative evaluations of their game partners, mirroring prior findings. Respondents also exhibited
increased antisocial behaviors, successfully capturing the dynamics found by previous research. However,
we found a noteworthy departure from previous research in terms of respondents’ post-rejection mood
change, as our respondents reported significantly higher levels of anger and sadness, and decreased
happiness; whereas prior literature did not.
For implicit group nonacceptance, our findings revealed a mixed bag of results. Unlike previous
research, we did not observe significant differences in partner evaluations between the implicit group
nonacceptance condition and the control condition. However, similar to past studies, participants in our

64
implicit group nonacceptance condition did exhibit heightened anger, and general negative mood,
suggesting that Spider Apocalypse’s implicit group nonacceptance condition successfully triggers these
emotional responses. Additionally, our participants displayed some antisocial behavior in the form of
suggested reductions in study compensation, albeit without any significant differences in rating
adjustments. We found no substantial differences in respondents’ choices to socially approach or
withdraw in subsequent situations, which may indicate variability in the impact of implicit group
nonacceptance on social behavior.
In the case of unknown group nonacceptance, the results were complex. While prior studies
revealed differing responses to new participants, our research design did not explore this aspect.
Additionally, we replicated the lack of significance in post-rejection mood observed in past research, with
no substantial differences in anger, sadness, or happiness between the unknown group nonacceptance and
control conditions. The mediation analysis suggested that post-rejection mood did not mediate the effect
of rejection on subsequent social approach or avoidance, aligning with previous findings in this context.
Our goal for this chapter was to validate the Spider Apocalypse paradigm by comparing its
outcomes with established findings; the comparisons outlined in the above thus act as a conceptual
validation that Spider Apocalypse operates similarly to previous existing paradigms. As a next step, we
use data from the same study in this chapter to compare mood, cognition, and behavior across all tested
nonacceptance contexts. We report the findings in the following chapter.

65
Chapter 5 – Comparing Nonacceptance Contexts on Mood, Cognition, and Behavior: An
Exploration
In this chapter, we dive deeper into how nonacceptance contexts can affect individuals’ mood,
cognition, and behavior. We utilize the same data used in Chapter 4, and report on several exploratory
analyses.
Method
Design Overview
As outlined in Chapter 4, respondents were recruited from Prolific to test a video game concept
and randomly assigned to either one of the experimental conditions or one of the control conditions.
Respondents then reported on both pre-game and post-game moods, their evaluation of their game
partner(s), their behavior towards their game partner(s), and whether they choose to approach or withdraw
from a subsequent social situation. Data collection was performed and completed in July 2023.
Participants
A total of 400 United States respondents ages between 18-40 were recruited from Prolific. This
age range was pre-determined to ensure that respondents are familiar with PC gaming. Twenty-three
respondents were removed for not completing this study. All respondents received $4 USD. The final
sample size was 377 (51.1% females, 65.7% Non-Hispanic White, mean age = 29.43). For this chapter,
we will be using all the data collected to compare findings across all experimental conditions.
Procedure
The same procedure used in Chapter 3 was used for this current study. Respondents read an
information sheet, provided consent, and completed the mood questions. Then, they played Spider
Apocalypse. Participants were randomly assigned to one of 10 conditions based on one-on-one or group
gameplay within various nonacceptance contexts. After the game, respondents rated their post-game
moods, gameplay experience, and other players. They also completed the partner rating task and
suggested compensation adjustments for their game partner(s). Then, respondents chose whether to play

66
Spider Apocalypse again with new people, or by themselves. Deception probes were administered both
before and after the debrief.
Results
Analyses Overview
First, I report results on the two-way MANOVA of Type of Nonacceptance x Number of
Rejecters on mood variables (anger, sadness, and happiness). Second, I report results on the two-way
MANOVA of Types of Nonacceptance x Number of Rejecters on post-game evaluation of game partners
and the aggregated behaviors towards the game partners (compensation adjustment and rating
adjustment). Third, I report on the logit loglinear analysis of Types of Nonacceptance x Number of
Rejecters on the binary outcome of social approach vs. social avoidance.
Mood
We conducted a 2 (one-on-one vs. group conditions) by 5 (personal, unknown, implicit,
irrelevant, and control conditions) MANOVA on the mood measures – anger, sadness, and happiness.
The two-way MANOVA revealed a significant main effect for types of nonacceptance (Wilks’
Lambda = .899, F(12,966) = 3.299, p < .001). No main effects were observed with number of rejecters
(Wilks’ Lambda = .991, F(3,365) = 1.093, p = .352) or the interaction term (Wilks’ Lambda = .959,
F(3,365) = 1.268, p = .232).
Table 17
Two-Way MANOVA Results of Number of Rejecters × Type of Rejection on Change in Anger, Sadness,
and Happiness
Effect
Wilk’s
Lambda F Sig. Partial η2
Intercept .758 38.830 <.001 .242
Number of Rejecters: Individual vs. Group .991 1.093 .352 .009
Type of Rejection .899 3.299 <.001 .035
Number of Rejecters × Type of Rejection .959 1.268 .232 .014
Tukey’s HSD post-hoc analysis was conducted to explore the main effect of types of
nonacceptance further. All three mood measures (all ps < .001) were observed to be significant in a test of
between-subjects effects. See Table 18 for means, SDs, and analyses of variance in change in anger,
sadness, and happiness; see Table 19 for pairwise comparison within the rejection types.

67
Table 18
Means, Standard Deviations, and Analyses of Variance in Change in Anger, Sadness, and Happiness
Measure Personal Unknown Implicit Irrelevant Control Partial η2
M SD M SD M SD M SD M SD
Anger 1.072 .169 .558 .177 .672 .168 .393 .182 .056 .162 .052***
Sadness .315 .124 -.276 .130 -.149 .123 -.362 .133 -.344 .119 .052***
Happiness -1.178 .186 -.355 .194 -.519 .184 -.336 .200 .152 .179 .070***
Table 19
Pairwise Comparison of Means Within Rejection Types on Change in Anger, Sadness, and Happiness
ANGER Personal Unknown Implicit Irrelevant Control
Personal
Unknown .514
Implicit .399 -.114
Irrelevant .679 .165 .279
Control 1.015*** .502 .616 .337
SADNESS Personal Unknown Implicit Irrelevant Control
Personal
Unknown .591*
Implicit .464 -.127
Irrelevant .677** .086 .213
Control .659** .068 .195 -.018
HAPPINESS Personal Unknown Implicit Irrelevant Control
Personal
Unknown -.824*
Implicit -.659 .165
Irrelevant -.843* -.019 -.184
Control -1.331*** -.507 -.672 -.488
Respondents in the personal nonacceptance condition were significantly more angry than
respondents in the control condition (p < .001). Respondents in the personal nonacceptance condition
were significantly more sad than respondents in the unknown condition (i.e., rejected but not sure why; p
= .011), irrelevant condition (i.e., game terminated due to network error; p = .002), and control condition
(p = .002). Respondents in the personal nonacceptance condition were significantly less happy than
respondents in the unknown condition (p = .023), irrelevant condition (p = .022), and control condition (p
< .001).
Evaluation and Behavioral Reaction
We conducted a 2 (one-on-one vs. group conditions) by 5 (personal, unknown, implicit,
irrelevant, and control conditions) MANOVA on the respondents’ post-game evaluation of their game

68
partners and their post-game behavior (aggregating rating adjustment and compensation adjustment)
towards their game partners.
The two-way MANOVA revealed significant main effects of both types of nonacceptance
(Wilks’ Lambda = .838, F(8,732) = 8.425, p < .001) and number of rejecters (Wilks’ Lambda = .962,
F(2,366) = 7.132, p < .001). The interaction term between the two was not observed to be significant
(Wilks’ Lambda = .963, F(8,732) = 1.737, p = .087).
Table 20
Two-Way MANOVA Results of Number of Rejecters × Type of Rejection on Partner Evaluation and PostGame Behavior
Effect
Wilk’s
Lambda F Sig. Partial η2
Intercept .070 2443.790 <.001 .930
Number of Rejecters: Individual vs. Group .962 7.132 <.001 .038
Type of Rejection .838 8.425 <.001 .084
Number of Rejecters × Type of Rejection .963 1.737 .087 .019
Type of Rejection. For partner evaluation, respondents in the personal nonacceptance condition
gave worse partner evaluations than respondents in the implicit nonacceptance condition (i.e., ambiguous
as to whether the nonacceptance has occurred; p < .001), irrelevant condition (p < .001), and control
condition (p < .001). Those in the unknown nonacceptance condition also gave worse partner evaluations
than respondents in the irrelevant condition (p = .027) and control condition (p = .004). For post-game
aggressive behavior, respondents in the personal nonacceptance condition were significantly more
aggressive towards their original rejecters (recommended lowered reputation rating and decreased study
compensation) than respondents in the unknown nonacceptance condition (p = 0.023), implicit
nonacceptance condition (p < .001), irrelevant nonacceptance condition (p < .001), and control condition
(p < .001). Additionally, respondents in the unknown nonacceptance condition were more aggressive
towards their original rejecters than respondents in the irrelevant nonacceptance condition (p = 0.024).
See Table 21 for means, SDs, and analyses of variance in change Partner Evaluation and Post-Game
Behavior; see Table 22 for pairwise comparison within the rejection types.

69
Table 21
Means, Standard Deviations, and Analyses of Variance in Partner Evaluation and Post-Game Behavior
Measure Personal Unknown Implicit Irrelevant Control Partial η2
M SD M SD M SD M SD M SD
P Eval 2.326 .097 2.644 .101 2.954 .096 3.083 .104 3.137 .093 .118***
PG Behavior -.425 .109 .062 .115 .289 .109 .565 .118 .484 .105 .123***
Table 22
Pairwise Comparison of Means Within Rejection Types on Partner Evaluation and Post-Game Behavior
P Eval Personal Unknown Implicit Irrelevant Control
Personal
Unknown -.318
Implicit -.628*** -.310
Irrelevant -.757*** -.439* -.129
Control -.811*** -.493** -.183 -.054
PG Behavior Personal Unknown Implicit Irrelevant Control
Personal
Unknown -.487*
Implicit -.715*** -.227
Irrelevant -.990*** -.503* -.276
Control -.910*** -.423 -.195 .080
Number of Rejecters. For partner evaluation, respondents in the individual rejection condition
gave worse partner evaluations than respondents in the group condition (p = .009). For post-game
aggressive behavior, respondents in the individual rejection condition were more punishing than
respondents in the group condition (p < .001). See Table 23 for means, SDs, and analyses of variance in
Partner Evaluation and Post-Game Behavior.
Table 23
Means, Standard Deviations, and Analyses of Variance in Partner Evaluation and Post-Game Behavior
Measure Individual Group Sig. Partial η2
M SD M SD
Partner Evaluation 2.713 .058 2.934 .066 .009 .019
Post Game Behavior .016 .065 .374 .075 <.001 .034
Social Approach and Avoid
A hierarchical loglinear model was used to analyze the three-way interaction between Type of
Rejection, Number of Rejecters, and Subsequent Social Approach/Avoidance Behavior. The analysis
began with a model including all main effects, which was systematically built up to include two-way and
three-way interactions.

70
The two-way interactions between Type of Rejection x Social Approach/Avoidance, Number of
Rejecters x Social Approach/Avoidance, and Type of Rejection x Number of Rejecters yielded p-values
of .240, .245, and .205 respectively.The three-way interaction between Type of Rejection, Number of
Rejecters, and Subsequent Social Approach/Avoidance Behavior resulted in a p-value of .716. None of
the p-values above were below .05, indicating that there was no statistically significant interaction
between these variables in this model.
Conclusion
The current chapter reports on findings of all the tested nonacceptance contexts on respondents’
mood, cognition, and behavior. The same dataset used in Chapter 4 was used here. Through a variety of
analyses, we shed light on the interplay between types of nonacceptance and number of rejecters and how
these factors affect individuals.
In regards to mood, evaluation, and behavior towards the original rejecters, respondents in the
personal nonacceptance condition reported significantly higher levels of anger and sadness, lowered level
of happiness, worse partner evaluation, and were more aggressive towards their original rejecters.
Additionally, respondents in the individual rejection conditions gave more negative partner evaluations
and were more punishing towards their game partners. However, social approach/avoid was not
significantly predicted by either of the manipulation variables.
Overall, Spider Apocalypse’s flexible manipulations extended our understanding of how
nonacceptance contexts can influence participants’ post-rejection mood, evaluation, behavior, and social
approach/avoid. These findings further emphasize the importance of considering the impact of social
contexts when examining the rejection-response link.

71
Chapter 6 – Testing the Hypothesized Nonacceptance-Response Process Model: A Path Analysis
Through chapters 4 and 5, we established the validity of Spider Apocalypse. This current chapter
collects a new set of data, then tests the hypothesized Nonacceptance-Response Process Model using
Spider Apocalypse. I report the partial least squares (PLS) models that test the relationship among
rejection contexts, emotions, motivations, and behavioral responses.
Method
Design Overview
A similar study methodology is used here as in the previous two chapters. Respondents were
recruited from Prolific to test a video game concept, and were randomly assigned to either one of the
experimental conditions or one of the control conditions. Respondents then reported on both pre-game
and post-game moods, their evaluation of their game partner(s), and whether they choose to approach or
withdraw from a subsequent social situation. The only addition for this experiment is that respondents
also reported on pre-game and post-game motivations (need to lash out, need for solitude, and need for
affiliation; see Procedure on the actual items used). Data collection was performed and completed in
November 2023.
Participants
A total of 400 United States respondents ages between 18-40 were recruited from Prolific. This
age range was pre-determined to ensure that respondents are familiar with PC gaming. Twenty-four
respondents were removed from data analyses for not completing the study. Each respondent received $4
USD. The final sample size was 376 (50.5% females, 65.0% Non-Hispanic White, mean age = 29.31).
Procedure
A procedure similar to that used in Chapters 4 and 5 was used for this current study. Respondents
read an information sheet, provided consent, and completed the pre-game mood and motivation questions.
Then, they played Spider Apocalypse. Participants were randomly assigned to one of 10 conditions based
on one-on-one or group gameplay within various nonacceptance contexts. After the game, respondents
rated their post-game moods and motivations, gameplay experience, and the other players. The newly

72
added motivation questions included three items for each of the three motivations, resulting in a total of
nine items. For each of the items, respondents reported from 1 (Very slightly or not at all) to 5
(Extremely) on whether they have the desire to:
1. Express frustration (Need to Lash Out 1)
2. Lash out at other people (Need to Lash Out 2)
3. Hurt other people (Need to Lash Out 3)
4. Be alone (Need for Solitude 1)
5. Withdraw from other people (Need for Solitude 2)
6. Avoid social contact (Need for Solitude 3)
7. Seek social relationships (Need for Affiliation 1)
8. Feel close to other people (Need for Affiliation 2)
9. Feel connected to other people (Need for Affiliation 3)
To evaluate these new items, we conducted a reliability analysis. Cronbach’s alpha was used to
assess the internal consistency of each of the motivations. The analysis revealed a Cronbach’s alpha
of .560 for the Need to Lash Out, an alpha of .711 for the Need for Solitude, and an alpha of .675 for the
Need for Affiliation. The alpha value for the Need to Lash Out is lower than the standard acceptable
range; this could be because we only have three items per motivation measure. Readers should bare this in
mind in interpreting results below. See Table 24 for correlation matrix of the motivation items.
Table 24
Motivations Items Correlation
LO1 LO2 LO3 S1 S2 S3 AFF1 AFF2 AFF3
LO1
LO2 .369**
LO3 .269** .437**
S1 .210** .107* .149**
S2 .350** .285** .240** .378**
S3 .307** .255** .200** .461** .516**
AFF1 -.110* .037 .057 -.109* -.049 -.106*
AFF2 -.076 -.008 .02 -.160** -.135** -.121* .413**
AFF3 -.145** -.025 -.054 -.111* -.138** -.136** .411** .404**
Note. LO = Lash Out; S = Solitude; AFF = Affiliation.

73
After the emotion and motivation questions, respondents completed the partner rating task and
suggested compensation adjustments for their game partner(s). Then, respondents chose whether to play
Spider Apocalypse again with new people, or by themselves. Deception probes were administered both
before and after the debrief.
Results
Analyses Overview
To investigate how motivations affect partner evaluation and how emotions affect motivations, I
first report results on the partial least squares (PLS) model testing with paths from motivations to Partner
Evaluation (Figure 19a). I then report results on PLS with paths from emotions to motivations to Partner
Evaluation (Figure 19b). To investigate how rejection contexts influence this process, I report results on
PLS with paths from Rejection vs. Non-Rejection conditions contrast, the two Non-Rejection conditions
contrast (control vs. irrelevant), and the three Rejection conditions contrasts (personal vs. unknown;
personal vs. implicit; unknown vs. implicit) on Partner Evaluation with significant emotions and
motivations as mediators (Figure 19c). This stepwise approach is chosen to compensate for the modest
sample size (N = 376), such that non predictive paths can be omitted in the rejection contexts contrast
models. See figures below for stepwise PLS models.

74
Figure 19
Stepwise Partial Least Squares Models for Predicting Partner Evaluation
(a)
(b)

75
(c)
We ran a separate set of models to investigate how motivations affect subsequent social behavior
and how emotions affect motivations. I first report results on PLS model testing with paths from
motivations to Social Approach Avoid (Figure 20a). I then report results on PLS with paths from
emotions to motivations to Social Approach Avoid (Figure 20b). To investigate how rejection contexts
influence this process, I report results on PLS with paths from Rejection vs. Non-Rejection conditions
contrast, the two Non-Rejection conditions contrast (control vs. irrelevant), and the three Rejection
conditions contrasts (personal vs. unknown; personal vs. implicit; unknown vs. implicit) on Social
Approach Avoid with significant emotions and motivations as mediators (Figure 20c). See figures below
for stepwise PLS models.

76
Figure 20
Stepwise Partial Least Squares Models for Predicting Subsequent Social Approach vs. Avoidance
Behavior
(a)
(b)

77
(c)
Group vs. individual contrasts are not reported in the analyses below, as the contrast was not
significant in any of the models tested. Therefore, the group and individual conditions are collapsed
within the experimental and control conditions.
All analyses were conducted using SmartPLS 4.0 (Ringle et al., 2022), with 5000 bootstrap
resamples.2
Table 25 reports the means of each measure within each condition. Of the measures reported in
the table, the change in need to lash out had relatively low deviation from zero compared to the items
from the other two motivations. This could be because aggression is generally hard to measure with selfreport, as respondents can be reluctant to explicitly admit their desire to harm other people, especially in a
psychology study. Future research may need to re-write the need to lash out items in order to capture
more variance.
2 Emotion (anger and sadness) and motivation (need to lash out, need for solitude, and need for affiliation)
variables were the change scores of their corresponding items. Another set of identical models were run
using residualized change scores instead, and all significant paths were identical to those using regular
change scores. To see outputs using the residualized change scores, see Appendix C.

78
Table 25
Means of Dependent Variables Within Each Condition
Personal Unknown Implicit Irrelevant Control
Anger 0.47 0.11 0.14 0.21 0.15
Angry 0.42 0.08 0.20 0.26 0.15
Irritable 0.51 0.14 0.07 0.15 0.14
Sadness 0.03 -0.17 -0.02 0.15 -0.25
Sad 0.13 -0.12 0.10 0.14 -0.24
Lonely -0.07 -0.21 -0.13 0.15 -0.25
Lash Out 0.19 0.08 0.04 0.06 0.05
Lash out 1 0.34 0.10 0.08 0.17 0.01
Lash out 2 0.20 0.10 0.04 -0.02 0.13
Lash out 3 0.04 0.03 0.01 0.03 0.02
Solitude 0.03 -0.32 -0.27 -0.18 -0.15
Solitude 1 -0.06 -0.39 -0.34 -0.18 -0.14
Solitude 2 0.10 -0.21 -0.15 -0.18 -0.16
Solitude 3 0.06 -0.36 -0.31 -0.18 -0.15
Affiliation -0.45 -0.24 -0.23 -0.15 -0.17
Affiliation 1 -0.41 -0.06 -0.10 -0.05 -0.13
Affiliation 2 -0.49 -0.45 -0.30 -0.35 -0.20
Affiliation 3 -0.46 -0.21 -0.30 -0.06 -0.18
PG Behavior 1.05 1.52 2.10 2.18 2.44
Compensation Adj -0.49 0.08 0.77 0.92 1.22
Ratings Adj 2.59 2.96 3.42 3.43 3.66
Approach/Avoid 0.58 0.53 0.49 0.58 0.55
Partner Evaluation
Motivations to Partner Evaluation.
The need to lash out (β = -.211, p < .001) and the need for affiliation (β = .145, p = .005)
significantly predicted how respondents punish/reward their game partners. The higher the need to lash
out, the more punishing the respondents are; and the higher the need for affiliation, the less punishing the
respondents are. The need for solitude did not significantly predict partner evaluation (β = -.092, p
= .162), and thus will be omitted from further analyses.

79
Figure 21
PLS Model of Motivations to Partner Evaluation
Emotions to Motivations to Partner Evaluation.
Anger significantly predicted both the need to lash out (β = .428, p < .001) and the need for
affiliation (β = -.247, p < .001). However, sadness does not significantly predict either of the motivations
(need to lash out: β = .109, p = .151; need for affiliation: β = -.041, p = .604), and thus will be omitted
from further analyses.
Figure 22
PLS Model of Emotions to Motivations to Partner Evaluation

80
Rejection vs. Non-Rejection.
Independent variable Rejection vs. Non-Reject was contrasted coded as 0.4 and -0.6. Non-Reject
conditions included control condition (played game till completion) as well as irrelevant condition
(disconnected game due to “network error”). Rejection conditions included personal rejection (players
explicitly stating that they’re leaving the game because of the participant), unknown reason (players
stating they’re leaving but not giving a reason), and implicit rejection (players leaving the game without
saying anything).
The Rejection vs. Non-Reject contrast did not significantly predict anger (β = .086, p = .399), and
thus this contrast did not play a part in the rejection-response process.
Figure 23
PLS Model of Rejection vs. non-Rejection to Partner Evaluation, Mediated by Emotion and Motivation
Control vs. Irrelevant Rejection.
Independent variable Control vs. Irrelevant Rejection was contrasted coded as -0.5 and 0.5
respectively, with all other conditions coded as 0. In the control condition, respondents played the game
till completion; in the irrelevant condition, respondents were disconnected from the game due to a
“network error”.

81
Similar to the Rejection vs. Non-Reject contrast, the Control vs. Irrelevant Rejection contrast did
not significantly predict anger (β = .031, p = .499), and thus this contrast did not play a part in the
rejection-response process.
Figure 24
PLS Model of Control vs. Irrelevant Rejection to Partner Evaluation, Mediated by Emotion and
Motivation
Personal Rejection vs. Unknown Reason.
Independent variable Personal Rejection vs. Unknown Reason was contrasted coded as 0.5 and -
0.5 respectively, with all other conditions coded as 0. In the Personal Rejection condition, respondents
were told “you suck at this game, we are losing because of you” before their game partners leave the
game; in the Unknown Reason condition, respondents were told “bye” before their game partners leave
the game.
The Personal Rejection vs. Unknown Reason contrast significantly predicted change in anger (β
= .149, p = .012). Such that respondents experienced more post-game anger in the personal rejection
condition. The total indirect effect of this model was significant (β = -.022, p = .027), and the Personal
Rejection vs. Unknown Reason contrast had a significant specific indirect effect on partner evaluation
through anger and the need to lash out (β = -.016, p = .034), but not through anger and the need for
affiliation (β = -.006, p = .129).

82
Figure 25
PLS Model of Personal vs. Unknown Rejection to Partner Evaluation, Mediated by Emotion and
Motivation
Personal Rejection vs. Implicit Rejection.
Independent variable Personal Rejection vs. Implicit Rejection was contrasted coded as 0.5 and -
0.5 respectively, with all other conditions coded as 0. In the Personal Rejection condition, respondents
were told “you suck at this game, we are losing because of you” before their game partners leave the
game; in the Implicit Rejection condition, respondents did not receive any messages from their game
partners before their partners left and the game was terminated.
The Personal Rejection vs. Implicit Rejection contrast significantly predicted change in anger (β
= .141, p = .005). Such that respondents experienced more post-game anger in the personal rejection
condition. The total indirect effect of this model was significant (β = -.021, p = .033), and the Personal
Rejection vs. Unknown Reason contrast had a significant specific indirect effect on partner evaluation
through anger and the need to lash out (β = -.015, p = .041), but not through anger and the need for
affiliation (β = -.006, p = .139).

83
Figure 26
PLS Model of Personal vs. Implicit Rejection to Partner Evaluation, Mediated by Emotion and
Motivation
Implicit Rejection vs. Unknown Reason.
Independent variable Implicit Rejection vs. Unknown Reason was contrasted coded as 0.5 and -
0.5 respectively, with all other conditions coded as 0. In the Unknown Reason condition, respondents
were told “bye” before their game partners leave the game; in the Implicit Rejection condition,
respondents did not receive any messages from their game partners before their partners left and the game
was terminated.
The Implicit Rejection vs. Unknown Reason contrast did not significantly predict anger (β = .010,
p = .832), and thus this contrast did not play a part in the rejection-response process.
Figure 27
PLS Model of Implicit vs. Unknown Rejection to Partner Evaluation, Mediated by Emotion and
Motivation

84
Personal Rejection vs. Control.
Finally, from the item descriptives of the different conditions, we decided to examine the contrast
of Personal Rejection vs. Control condition on partner evaluation. Independent variable Personal
Rejection vs. Control was contrasted coded as 0.5 and -0.5 respectively, with all other conditions coded as
0. In the Personal Rejection condition, respondents were told “you suck at this game, we are losing
because of you” before their game partners leave the game; in the Control condition, respondents played
the game till completion.
The Personal Rejection vs. Control contrast significantly predicted change in anger (β = .133, p
= .013). Such that respondents experienced more post-game anger in the personal rejection condition.
Although the total indirect effect of this model was significant (β = -.020, p = .048), the Personal
Rejection vs. Control contrast did not have a significant specific indirect effect on partner evaluation
through either anger and the need to lash out (β = -.014, p = .054), or through anger and the need for
affiliation (β = -.006, p = .166).
Figure 28
PLS Model of Personal Rejection vs. Control Condition to Partner Evaluation, Mediated by Emotion and
Motivation

85
Subsequent Social Approach/Avoidance Behavior
Motivations to Social Approach/Avoid.
The need for solitude (β = -.061, p = .048) significantly predicted whether respondents choose to
play Spider Apocalypse with new online players (social approach behavior) or by themselves alone
(social avoidance behavior). A higher need for solitude resulted in a higher likelihood of social avoidance
behavior (OR = 1.063 more likely to choose to play alone with every one point increase in the need for
solitude). However, the need to lash out (β = -.006, p = .912) and the need for affiliation (β = -.001, p
= .743) did not significantly predict respondents’ subsequent social approach/avoidance behaviors, and
thus will be omitted from further analyses.
Figure 29
PLS Model of Motivations to Subsequent Social Approach vs. Avoidance Behavior
Emotions to Motivations to Social Approach/Avoid.
Both anger (β = .281, p < .001) and sadness (β = .205, p = .004) significantly predicted the need
for solitude. An increase in anger and/or an increase in sadness resulted inan increase in the need for
solitude.

86
Figure 30
PLS Model of Emotions to Motivations to Subsequent Social Approach vs. Avoidance Behavior
Rejection vs. Non-Rejection.
Independent variable Rejection vs. Non-Reject was contrasted coded as 0.4 and -0.6. Non-Reject
conditions included control condition (played game till completion) as well as irrelevant condition
(disconnected game due to “network error”). Rejection conditions included personal rejection (players
explicitly stating that they’re leaving the game because of the participant), unknown reason (players
stating they’re leaving but not giving a reason), and implicit rejection (players leaving the game without
saying anything).
The Rejection vs. Non-Reject contrast significantly predicted change in sadness (β = .210, p
= .037), but not a change in anger (β = .086, p = .402). Rejected respondents experienced a higher
increase in sadness than respondents in the non-rejection conditions. However, the total indirect effect of
this model was insignificant (β = -.011, p = .101).
Figure 31
PLS Model of Rejection vs. non-Rejection to Subsequent Social Approach vs. Avoidance Behavior,
Mediated by Emotion and Motivation

87
Control vs. Irrelevant Rejection.
Independent variable Control vs. Irrelevant Rejection was contrasted coded as -0.5 and 0.5
respectively, with all other conditions coded as 0. In the control condition, respondents played the game
till completion; in the irrelevant condition, respondents were disconnected from the game due to a
“network error”.
The Control vs. Irrelevant Rejection contrast did not significantly predict anger (β = .031, p
= .501) or sadness (β = .067, p = .151), and thus this contrast did not play a part in the rejection-response
process.
Figure 32
PLS Model of Control vs. Irrelevant Rejection to Subsequent Social Approach vs. Avoidance Behavior,
Mediated by Emotion and Motivation
Personal Rejection vs. Unknown Reason.
Independent variable Personal Rejection vs. Unknown Reason was contrasted coded as 0.5 and -
0.5 respectively, with all other conditions coded as 0. In the Personal Rejection condition, respondents
were told “you suck at this game, we are losing because of you” before their game partners leave the
game; in the Unknown Reason condition, respondents were told “bye” before their game partners leave
the game.
The Personal Rejection vs. Unknown Reason contrast significantly predicted change in anger (β
= .149, p = .012), such that respondents experienced a higher increase in anger in the personal rejection
condition than those in the Unknown Reason condition. However, the Personal Rejection vs. Unknown
Reason contrast did not significantly predict change in sadness (β = .099, p = .099). The total indirect

88
effect from the Personal Rejection vs. Unknown Reason to Social Approach/Avoidance Behavior was
insignificant (β = -.003, p = .163).
Figure 33
PLS Model of Personal vs. Unknown Rejection to Subsequent Social Approach vs. Avoidance Behavior,
Mediated by Emotion and Motivation
Personal Rejection vs. Implicit Rejection.
Independent variable Personal Rejection vs. Implicit Rejection was contrasted coded as 0.5 and -
0.5 respectively, with all other conditions coded as 0. In the Personal Rejection condition, respondents
were told “you suck at this game, we are losing because of you” before their game partners leave the
game; in the Implicit Rejection condition, respondents did not receive any messages from their game
partners before their partners left and the game was terminated.
The Personal Rejection vs. Implicit Rejection contrast significantly predicted change in anger (β
= .142, p = .005), such that respondents experienced a higher increase in anger in the personal rejection
condition than those in the Implicit Rejection condition. However, the Personal Rejection vs. Implicit
Rejection contrast did not significantly predict change in sadness (β = .078, p = .185). The total indirect
effect from the Personal Rejection vs. Implicit Rejection to Social Approach/Avoidance Behavior was
insignificant (β = -.003, p = .160).

89
Figure 34
PLS Model of Personal vs. Implicit Rejection to Subsequent Social Approach vs. Avoidance Behavior,
Mediated by Emotion and Motivation
Implicit Rejection vs. Unknown Reason.
Independent variable Implicit Rejection vs. Unknown Reason was contrasted coded as 0.5 and -
0.5 respectively, with all other conditions coded as 0. In the Unknown Reason condition, respondents
were told “bye” before their game partners leave the game; in the Implicit Rejection condition,
respondents did not receive any messages from their game partners before their partners left and the game
was terminated.
The Implicit Rejection vs. Unknown Reason contrast did not significantly predict anger (β = .010,
p = .834) or sadness (β = .024, p = .604), and thus this contrast did not play a part in the rejectionresponse process.
Figure 35
PLS Model of Implicit vs. Unknown Rejection to Subsequent Social Approach vs. Avoidance Behavior,
Mediated by Emotion and Motivation

90
Personal Rejection vs. Control.
Finally, from the item descriptives of the different conditions, we decided to examine the contrast
of Personal Rejection vs. Control condition on subsequent social approach/avoidance behavior.
Independent variable Personal Rejection vs. Control was contrasted coded as 0.5 and -0.5 respectively,
with all other conditions coded as 0. In the Personal Rejection condition, respondents were told “you suck
at this game, we are losing because of you” before their game partners leave the game; in the Control
condition, respondents played the game till completion.
The Personal Rejection vs. Control contrast significantly predicted change in anger (β = .134, p
= .013) and sadness (β = .154, p = .006). Respondents experienced a higher increase in both anger and
sadness in the personal rejection condition than those in the control condition. The total indirect effect
from the Personal Rejection vs. Control to Social Approach/Avoidance Behavior was insignificant (β =
-.004, p = .124).
Figure 36
PLS Model of Personal Rejection vs. Control Condition to Subsequent Social Approach vs. Avoidance
Behavior, Mediated by Emotion and Motivation
Conclusion
The current chapter focuses on testing the hypothesized Nonacceptance-Response Process Model
using PLS path analysis. We observed patterns in how emotions and motivations affect subsequent social
behaviors both towards their original game partners and towards new people. Furthermore, this chapter
plays a crucial role in our understanding of how experiences within Spider Apocalypse affect the
psychological dynamics of post-rejection response.

91
Anger was found to be a pivotal emotion in the nonacceptance-response process, significantly
affecting all three motivations (the need to lash out, the need for solitude, and the need for affiliation).
Notably, the need to lash out and the need for affiliation influenced respondents’ reward/punishment
behavior towards their game partners, whereas the need for solitude emerges as the critical predictor of
subsequent social approach/avoidance behavior. When anger is present in the model, sadness rarely had a
significant effect above and beyond the effect of anger. Sadness only had significant paths through the
need for solitude to predicting subsequent social approach/avoidance behavior, and that effect was only
seen under the reject vs. non-reject contrast and the personal vs. control contrast. This finding highlights
the importance of separating anger and sadness when examining one’s post-rejection emotional response,
to identify the different motivations elicited from these emotions, and to distinguish how the motivations
affect subsequent social behavior. Previous work in this area has typically ignored the differences
between different emotions and has only focused on the impact of positive versus negative emotions.
The Personal Rejection condition seemed to have produced the most robust effects in the
nonacceptance-response process, evident in its significant contrasts with the Unknown Reason condition,
the Implicit Rejection condition, and the Control condition when predicting both Partner Evaluation and
Subsequent Social Approach/Avoidance Behavior.

92
Chapter 7 – Conclusion and Discussion
This dissertation embarked on an in-depth exploration of social nonacceptance, a fundamental
aspect of human social interactions. Through chapters 1 through 6, we navigated the complex dynamics
of social rejection, utilizing innovative methodologies and theoretical frameworks to investigate the
nuanced ways in which individuals respond to nonacceptance.
Chapter 1 laid the groundwork by reviewing past theories, methodologies, and findings in the
field of social nonacceptance research. It highlighted the paradoxical empirical findings, and the absence
of a coherent framework to explain these contradictions.
Chapter 2 presented a vignette study examining the effects of social nonacceptance on mood and
cognition. This chapter uncovered how ambiguous and explicit rejections differently impact individuals’
emotional and cognitive states, contributing to broader understanding of social rejection contexts on
individuals.
Chapter 3 introduced and pilot tested a novel experimental paradigm Spider Apocalypse to
simulate various nonacceptance scenarios in a more dynamic and interactive online gaming environment.
This paradigm sought to address the limitations of traditional methods used in social nonacceptance
research.
Chapter 4 collected data that compared findings from Spider Apocalypse with findings from
existing paradigms, validating the game as an effective alternative to studying social nonacceptance. This
chapter bridged the gap between traditional methods and innovative approaches.
Chapter 5 dove deeper into the comprehensive analysis using the data collected for Chapter 4,
comparing different nonacceptance contexts regarding their impact on mood and behavior. This chapter
shed light on the intricate ways these context variables interplay in response to social rejection.
Chapter 6 collected a final set of data, which tested the hypothesized Nonacceptance-Response
Process Model using path analysis. It highlighted the pivotal role of anger and dissected the complex
relationship between emotions, motivations, and social behaviors both towards respondents’ game
partners as well as new individuals.

93
Contributions to the Field of Research
1. Taxonomy of Social Nonacceptance: This dissertation provides a detailed outline of a
taxonomy of social nonacceptance. The taxonomy distinguishes different types of contexts of
nonacceptance, which most prior literature overlooks. This taxonomy includes factors such as
the number of people involved, the relationship to the nonaccepting individuals, whether the
nonacceptance is explicit or implicit, and the reasons given for the nonacceptance.
2. Integrative framework: This dissertation proposes a new integrative framework for
understanding social nonacceptance. It incorporates various dimensions of nonacceptance
contexts and their effects on mood, motivation, and behavior. The model suggests that the
perception and interpretation of a nonacceptance context, as well as the subsequent available
behavioral options, can significantly influence an individual’s post-nonacceptance response.
3. Review of experimental paradigms: This dissertation reviews major existing experimental
paradigms in social rejection research, including Cyberball, chat room paradigms, and video
exchange paradigms. It highlights the lack of systematic manipulation and the issues arising
from comparing findings across different paradigms, which may lead to contradictory
findings.
4. New experimental paradigm: The dissertation introduces a new experimental paradigm,
Spider Apocalypse. This game-based paradigm allows for flexible manipulations of
nonacceptance contexts and aim to provide more ecologically valid and controlled conditions
for studying post-nonacceptance response.
5. Empirical findings using Spider Apocalypse: This dissertation provides empirical evidence
through various studies that demonstrate how different nonacceptance contexts affect
individuals’ mood, cognition, and behavior. The findings support the proposed taxonomy and
integrative framework, and offer insights into the complex dynamics of social nonacceptance.

94
Spider Apocalypse as a Novel Paradigm
The introduction of Spider Apocalypse marks a significant advancement in the experimental
paradigms of social nonacceptance research. Its video game format has demonstrated high engagement
and ecological validity. The validation of Spider Apocalypse through extensive data collection is evident
in the participant feedback regarding the game’s realism and challenge level (Chapter 3), the replication
of results comparing to previous findings in the field (Chapter 4), as well as the different rejection
contexts producing significant effects on anger, happiness, partner evaluation, compensation adjustment,
and rating adjustment (Chapter 5).
Furthermore, its flexibility in manipulating various dimensions of social nonacceptance – such as
NPC behaviors, number of NPCs, game difficulty, and player message content – makes it a versatile and
powerful tool for future research in this field. Unlike many of the existing paradigms in the field, Spider
Apocalypse requires no coding experience to set up. We have premade a GUI in Unity to allow
researchers to change almost all aspects of the game.
Interpretation of Findings through Spider Apocalypse
Our findings using Spider Apocalypse highlight the profound impact of rejection context on
individuals’ emotional, motivational, and behavioral responses. Notably, personal rejection emerged as
the most influential nonacceptance context – significantly increasing anger and sadness, increasing the
need to lash out and need for solitude, decreasing the need for affiliation, and resulted in more punishing
and withdrawal social behaviors when comparing with other contexts (Chapters 5 & 6). Although, this
finding was contradictory to our own vignettes study (Chapter 2) where we found ambiguous
nonacceptance to result in a higher self-report of anger and sadness. This further validates the notion that
mental simulations of social rejection elicited via vignettes don’t always accurately reflect what
respondents would do/feel when experiencing real rejection. Thus, it is essential to put participants
through realistic simulated rejection scenarios rather than just situational vignettes.
Across different contexts, anger was a key driver in the nonacceptance-response process,
influencing all three motivations (need to lash out, need for solitude, and need for affiliation). These

95
motivations in turn were related to different social behaviors. In respondents’ interactions with their
original game partners, the need to lash out and the need for affiliation were significantly predictive of
individuals’ behaviors; whereas in respondents’ decisions about engaging or avoiding new social
interactions, the need for solitude was the critical predictor (Chapter 6). The lack of significance for
sadness validates the notion that researchers need to tease out the different emotions in the rejectionresponse process, as opposed to aggregating across multiple emotions into a single mood index.
Finally, the findings in this dissertation using Spider Apocalypse emphasize the multifaceted
nature of social nonacceptance, and how the contexts of nonacceptance can greatly affect one’s
experience and behavior; this is supported by the nuanced differences we detected in responses to various
types of rejections (personal vs. unknown vs. implicit). This supports the notion that rejection is not a
monolithic experience, but rather varies significantly based on context and individual perception.
Relation to Previous Research
The use of Spider Apocalypse enabled the conceptual replication of certain findings from
previous research, particularly with past findings that used individual personal nonacceptance paradigms.
However, the results only partially supported findings from other contexts like implicit group and
unknown group nonacceptance paradigms (Chapter 4).
This project thus not only aligns with existing literature, but also extends it by providing a
comprehensive and systematic exploration into the nonacceptance-response process. It adds a nuanced
understanding of how individuals respond to nonacceptance by comparing the different rejection contexts,
an approach not utilized in prior studies.
Future Directions
1. Enhancing Spider Apocalypse: The Spider Apocalypse paradigm offers ample
opportunities for modification to induce more distinct or intensified effects. For instance,
varying the game's duration or creating more pronounced differences between rejection
contexts could enhance the ecological validity of the research. Unexplored dimensions,

96
such as emotional closeness to rejecters or subjective appraisals of nonacceptance, should
also be incorporated to examine their interplay with rejection contexts.
2. Cultural and political orientation variations: Investigating the response to social
nonacceptance in diverse cultural settings and among individuals with different political
orientations could offer valuable insights. Such studies would illuminate how societal
norms and personal beliefs shape reactions to rejection, potentially leading to culturally
sensitive interventions and policies.
3. Additional moderators and mediators: One significant area of exploration is how gender
and personality traits may influence emotional regulation and decision-making in the
rejection-response process. Additionally, researchers may incorporate other state-emotion
and -motivations of interest into the rejection-response process.
4. Physiological response to different types of rejection: Integrating physiological
measurements like heart rate variability, brain activity, or cortisol levels could provide a
deeper understanding of how rejection impacts the mind-body connection. This approach
could reveal the biological underpinnings of emotional and behavioral responses to
rejection.
5. Individual differences regarding respondents’ personal relationships with video games:
Measuring respondents’ perception of self in the context of video games could shed more
light on post-rejection response. For example, respondents who do not game often, or
self-perceive to be “bad at video games”, may perhaps experience less distress when they
are rejected, because they easily rationalize it as “my skills are lacking, of course others
would find it hard to play with me”. Whereas respondents who game often or selfperceive to be “good at video games”, may perhaps experience more anger when they are
rejected, because the rejection directly contradicts their perception of self – “I am good at
video games, why was I rejected?”

97
Concluding Remarks
The current research identifies a gap in research in the field of social nonacceptance and
introduces a novel paradigm to address such gap. Through six chapters, we demonstrated a significant
advance in understanding the complex dynamics of social rejection. The introduction of Spider
Apocalypse overcame the limitations of traditional research paradigms, and also provided insights into the
nuanced effects of different rejection contexts. This body of work has significant implications for both
theoretical advancement and practical application. Theoretically, it provides a clearer framework for
understanding the varied dimensions of social rejection. Practically, it offers a more comprehensive and
flexible tool to examine the various elements of the nonacceptance-response process. Therefore, this
research paves way for future studies to further unravel the complexities of social rejection.

98
References
Ayduk, O., Downey, G., Testa, A., Yen, Y., & Shoda, Y. (1999). Does rejection elicit hostility in rejection
sensitive women? Social Cognition, 17(2), 245-271. https://doi.org/10.1521/soco.1999.17.2.245
Ayduk, O., Gyurak, A., & Luerssen, A. (2008). Individual differences in the rejection-aggression link in
the hot sauce paradigm: The case of rejection sensitivity. Journal of Experimental Social
Psychology, 44(3), 775-782. https://doi.org/10.1016/j.jesp.2007.07.004
Ayduk, O., May, D., Downey, G., & Higgins, E. T. (2003). Tactical differences in coping with rejection
sensitivity: The role of prevention pride. Personality and Social Psychology Bulletin, 29(4), 435-
448. https://doi.org/10.1177/0146167202250911
Bernstein, M. F., & Claypool, H. M. (2012a). Not all social exclusions are created equal: Emotional
distress following social exclusion is moderated by exclusion paradigm. Social Influence, 7(2),
113-130. https://doi.org/10.1080/15534510.2012.664326
Bernstein, M. F., & Claypool, H. M. (2012b). Social exclusion and pain sensitivity: Why exclusion
sometimes hurts and sometimes numbs. Personality and Social Psychology Bulletin, 38(2), 185-
196. https://doi.org/10.1177/0146167211422449
Bernstein, M. J., Young, S. G., Brown, C. M., Sacco, D. F., & Claypool, H. M. (2008). Adaptive
responses to social exclusion: Social rejection improves detection of real and fake smiles.
Psychological Science, 19(10), 981–983. https://doi.org/10.1111/j.1467-9280.2008.02187.x
Carter-Sowell, A. R., Chen, Z., & Williams, K. D. (2008). Ostracism increases social susceptibility. Social
Influence, 3(3), 143–153. https://doi.org/10.1080/15534510802204868
Chester, D., & DeWall, C. N. (2017). Combating the sting of rejection with the pleasure of revenge: A
new look at how emotion shapes aggression. Interpersonal Relations and Group Processes,
112(3), 413-430. https://doi.org/10.1037/pspi0000080
Chow, R. M., Tiedens, L. Z., & Govan, C. L. (2008). Excluded emotions: The role of anger in antisocial
responses to ostracism. Journal of Experimental Social Psychology, 44(3), 896–903.
https://doi.org/10.1016/j.jesp.2007.09.004
DeWall, C. N. (2010). Forming a basis for acceptance: Excluded people form attitudes to agree with
potential affiliates. Social Influence, 5(4), 245-260. https://doi.org/10.1080/15534511003783536
DeWall, C. N., Baumeister, R. F., & Vohs, K. D. (2008). Satiated with belongingness? Effects of
acceptance, rejection, and task framing on self-regulatory performance. Journal of Personality and
Social Psychology, 95(6), 1367-1382. https://doi.org/10.1037/a0012632
DeWall, C. N., Twenge, J. M., Bushman, B. J., Im, C., & Williams, K. D. (2010). A little acceptance goes
a long way: Applying social impact theory to the rejection-aggression link. Social Psychological
and Personality Science, 1(2), 168-174. https://doi.org/10.1177/1948550610361387
DeWall, C. N., Twenge, J. M., Gitter, S. A., & Baumeister, R. F. (2009). It's the thought that counts: The
role of hostile cognition in shaping aggressive responses to social exclusion. Journal of
Personality and Social Psychology, 96(1), 45-59. https://doi.org/10.1037/a0013196

99
Exum, M.L., Bouffard, J.A. (2010). Testing theories of criminal decision making: Some empirical
questions about hypothetical scenarios. In: Piquero, A., Weisburd, D. (eds) Handbook of
Quantitative Criminology. Springer, New York, NY. https://doi.org/10.1007/978-0-387-77650-
7_28
Exum, M.L., & Layana, M.C. (2018). A test of the predictive validity of hypothetical intentions to offend.
Journal of Crime and Justice, 41(2), 136-154. https://doi.org/10.1080/0735648X.2016.1244486
FeldmanHall, O., Mobbs, D., Evans, D., Hiscox, L., Navrady, L., & Dalgleish, T. (2012). What we say
and what we do: The relationship between real and hypothetical moral choices. Cognition,
123(3), 434–441. https://doi.org/10.1016/j.cognition.2012.02.001
Freedman, G., Williams, K. D., & Beer, J. S. (2016). Softening the blow of social exclusion: The
responsive theory of social exclusion. Frontiers in Psychology, 7, 1570-1570.
https://doi.org/10.3389/fpsyg.2016.01570
Gamian-Wilk, M., Madeja-Bien, K., & Bjørkelo, B. (2020). To comply or not to comply: Social exclusion
increases compliance in cyberball but only with the non-rejecters. Forum Oświatowe, 32(2), 75-
85. https://doi.org/10.34862/fo.2020.2.3
Gardner, W. L., Pickett, C. L., & Brewer, M. B. (2000). Social exclusion and selective memory: How the
need to belong influences memory for social events. Personality and Social Psychology Bulletin,
26(4), 486-496. https://doi.org/10.1177/0146167200266007
Godwin, A., MacNevin, G., Zadro, L., Iannuzzelli, R., Weston, S., Gonsalkorale, K., & Devine, P. (2014).
Are all ostracism experiences equal? A comparison of the autobiographical recall, Cyberball, and
O-Cam paradigms. Behavioral Research Methods, 46(3), 660-667.
https://doi.org/10.3758/s13428-013-0408-0
Gonsalkorale, K., & Williams, K. D. (2007). The KKK won't let me play: Ostracism even by a despised
outgroup hurts. European Journal of Social Psychology, 37(6), 1176–1186.
https://doi.org/10.1002/ejsp.392
Goodacre, R., & Zadro, L. (2010). O-Cam: A new paradigm for investigating the effects of ostracism.
Behavioral Research Methods, 42(3), 768-774. https://doi.org/10.3758/brm.42.3.768
Hong, J., & Sun, C. (2018). To defend or to affiliate: The effects of categorical similarity cues after social
exclusion. The Journal of Social psychology, 158(6), 647-662.
https://doi.org/10.1080/00224545.2017.1373621
Jerolmack, C., & Khan, S. (2014). Toward an understanding of the relationship between accounts and
action. Sociological Methods & Research, 43(2), 236-247.
https://doi.org/10.1177/0049124114523397
Kormos, C., & Gifford, R. (2014). The validity of self-report measures of proenvironmental behavior: A
meta-analytic review. Journal of Environmental Psychology, 40, 359-371.
https://doi.org/10.1016/j.jenvp.2014.09.003
Lakin, J. L., Chartrand, T. L., & Arkin, R. M. (2008). I am too just like you: Nonconscious mimicry as an
automatic behavioral response to social exclusion. Psychological Science, 19(8), 816–822.
https://doi.org/10.1111/j.1467-9280.2008.02162.x

100
Leary, M. R., Tambor, E. S., Terdal, S. K., & Downs, D. L. (1995). Self-esteem as an interpersonal
monitor: The sociometer hypothesis. Journal of Personality and Social Psychology, 68(3), 518–
530. https://doi.org/10.1037/0022-3514.68.3.518
Maner, J. K., DeWall, C. N., Baumeister, R. F., & Schaller, M. (2007). Does social exclusion motivate
interpersonal reconnection? Resolving the “porcupine problem”. Journal of Personality and
Social Psychology, 92(1), 42-55. https://doi.org/10.1037/0022-3514.92.1.42
Nezlek, J. B., Kowalski, R. M., Leary, M. R., Blevins, T., & Holgate, S. (1997). Personality moderators of
reactions to interpersonal rejection: Depression and trait self-esteem. Personality and Social
Psychology Bulletin, 23(12), 1235-1244. https://doi.org/10.1177/01461672972312001
Nezlek, J. B., Wesselmann, E. D., Wheeler, L., & Williams, K. D. (2012). Ostracism in everyday life.
Group Dynamics: Theory, Research, and Practice, 16(2), 91–104.
https://doi.org/10.1037/a0028029
Oaten, M., Williams, K. D., Jones, A., & Zadro, L. (2008). The effects of ostracism on self-regulation in
the socially anxious. Journal of Social and Clinical Psychology, 27(5), 471–504.
https://doi.org/10.1521/jscp.2008.27.5.471
Poon, K. T., Chen, Z., & DeWall, C. N. (2013). Feeling entitled to more: Ostracism increases dishonest
behavior. Personality and Social Psychology Bulletin, 39(9), 1227–1239.
https://doi.org/10.1177/0146167213493187
Rajchert, J., & Winiewski, M. (2017). Strength of excitation is negatively associated with aggressive
behavior after interpersonal rejection. Frontiers in Psychology, 8, 296.
https://doi.org/10.3389/fpsyg.2017.00296
Rajchert, J., Żółtak, T., Szulawski, M., & Jasielska, D. (2019). Effects of rejection by a friend for
someone else on emotions and behavior. Frontiers in Psychology, 10, 746-746.
https://doi.org/10.3389/fpsyg.2019.00764
Richman, L. S. (2013). The multi-motive model of responses to rejection-related experiences. In The
Oxford Handbook of Social Exclusion.
https://doi.org/10.1093/oxfordhb/9780195398700.013.0005
Ren, D., Wesselmann, E., & Williams, K. D. (2016). Evidence for another response to ostracism: Solitude
seeking. Social Psychological and Personality Science, 7(3), 204–212.
https://doi.org/10.1177/1948550615616169
Ringle, C. M., Wende, S., & Becker, J. (2022). SmartPLS 4. Retrieved from https://www.smartpls.com
Riva, P., Williams, K. D., Torstrick, A. M., & Montali, L. (2014). Orders to shoot (a camera): Effects of
ostracism on obedience. The Journal of Social Psychology, 154(3), 208–216.
https://doi.org/10.1080/00224545.2014.883354
Romero-Canyas, R., Downey, G., Reddy, K. S., Rodriguez, S., Cavanaugh, T. J., & Pelayo, R. (2010).
Paying to belong: When does rejection trigger ingratiation? Journal of Personality and Social
Psychology, 99(5), 802-823. https://doi.org/10.1037/a0020013

101
Smith, A., & Williams, K. D. (2004). R U there? Ostracism by cell phone text messages. Group
Dynamics, 8(4), 291-301. https://doi.org/10.1037/1089-2699.8.4.291
Twenge, J. M., Baumeister, R. F., DeWall, C. N., Ciarocco, N. J., & Bartels, J. M. (2007). Social
exclusion decreases prosocial behavior. Journal of Personality and Social Psychology, 92(1), 56–
66. https://doi.org/10.1037/0022-3514.92.1.56
Twenge, J. M., Baumeister, R. F., Tice, D. M., & Stucke, T. S. (2001). If you can't join them, beat them:
Effects of social exclusion on aggressive behavior. Journal of Personality and Social Psychology,
81(6), 1058–1069. https://doi.org/10.1037/0022-3514.81.6.1058
Twenge, J. M., & Campbell, W. K. (2003). "Isn't it fun to get the respect that we're going to deserve?"
Narcissism, social rejection, and aggression. Personality & social psychology bulletin, 29(2),
261–272. https://doi.org/10.1177/0146167202239051
van Beest, I., & Williams, K. D. (2006). When inclusion costs and ostracism pays, ostracism still hurts.
Journal of Personality and Social Psychology, 91(5), 918–928. https://doi.org/10.1037/0022-
3514.91.5.918
Vorauer, J. D., Cameron, J. J., Holmes, J. G., & Pearce, D. G. (2003). Invisible overtures: Fears of
rejection and the signal amplification bias. Journal of Personality and Social Psychology, 84(4),
793–812. https://doi.org/10.1037/0022-3514.84.4.793
Walasek, L., Juanchich, M., & Sirota, M. (2019). Adaptive cooperation in the face of social exclusion.
Journal of Experimental Social Psychology, 82, 35-46. https://doi.org/10.1016/j.jesp.2018.11.005
Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive
and negative affect: the PANAS scales. Journal of Personality and Social Psychology, 54(6),
1063–1070. https://doi.org/10.1037//0022-3514.54.6.1063
Wesselmann, E. D., Butler, F. A., Williams, K. D., & Pickett, C. L. (2010). Adding injury to insult:
Unexpected rejection leads to more aggressive responses. Aggressive Behavior, 36(4), 232-237.
https://doi.org/10.1002/ab.20347
Wesselmann, E. D., Grzybowski, M. R., Steakley-Freeman, D. M., DeSouza, E. R., Nezlek, J. B., &
Williams, K. D. (2016). Social exclusion in everyday life. In Social Exclusion (pp. 3-23).
https://doi.org/10.1007/978-3-319-33033-4_1
Wesselmann, E. D., Ren, D., & Williams, K. D. (2015). Motivations for responses to ostracism. Frontiers
in Psychology, 6, 40-40. https://doi.org/10.3389/fpsyg.2015.00040
Wesselmann, E. D., Williams, K. D. (2017). Social life and social death: Inclusion, ostracism, and
rejection in groups. Group Processes and Intergroup Relations, 20(5), 693-706.
https://doi.org/10.1177/1368430217708861
Williams, K. D. (2007). Ostracism. Annual Review of Psychology, 58(1), 425-452.
https://doi.org/10.1146/annurev.psych.58.110405.085641
Williams, K. D., Cheung, C. K. T., & Choi, W. (2000). Cyberostracism: Effects of being ignored over the
Internet. Journal of Personality and Social Psychology, 79(5), 748–
762. https://doi.org/10.1037/0022-3514.79.5.748

102
Williams, K. D., & Gerber, J. (2005). Ostracism: The making of the ignored and excluded mind.
Interaction Studies, 6(3), 359-374. https://doi.org/10.1075/is.6.3.04wil
Williams, K. D., Govan, C. L., Croker, V., Tynan, D., Cruickshank, M., & Lam, A. (2002). Investigations
into differences between social- and cyberostracism. Group Dynamics: Theory, Research, and
Practice, 6(1), 65–77. https://doi.org/10.1037/1089-2699.6.1.65
Wolf, W., Levordashka, A., Ruff, J. R., Kraaijeveld, S., Lueckmann, J., & Williams, K. D. (2014).
Ostracism Online: A social media ostracism paradigm. Behavior Research Methods, 47(2), 361-
373. https://doi.org/10.3758/s13428-014-0475-x
Yaakobi, E. (2021). Immediate moderating effects of attribution on ostracism distress. Australian
Psychologist, 56(3), 193-203. https://doi.org/10.1080/00050067.2020.1829452

103
Appendices
Appendix A
Social Nonacceptance Situational Vignettes for Chapter 2
1. You are organizing a weekend gathering, and you invite a [close friend / new acquaintance] to
attend. They say they will let you know:
2. A [close friend / new acquaintance] is applying for a position at the company that you work at.
You offer to help them prepare for their interviews. They say they will let you know:
3. You finished a difficult project at work/school and want to celebrate by going out, so you ask a
[close / new] colleague/classmate if they wish to join you. They say they will let you know:
4. A [close / new] colleague/classmate is struggling with a complicated project, so you offer to help
them out on some tasks. They say they will let you know:
5. You are interested in joining your [close friends’ / new acquaintances’] hobby group (for
example: team sports, book club, board game group, etc.), so you reach out to them and request to
join. They say they will let you know:
6. You are planning on a night-out with a group of [close friends / new acquaintances]. The group is
trying to decide on a place for dinner, and you suggest a restaurant to everyone. They say they
will let you know:
7. You hear that a group of your [close / new] colleagues/classmates are planning to hang out at a
bar for happy hour, and you ask them if you could join them. They say they will let you know:
8. You are working on a team project with some [close / new] colleagues/classmates, and you
suggest a new idea for consideration. They say they will let you know:

104
Appendix B
Response Measures for Chapter 2
All items answered on a scale of 1 to 7:
1. How LIKELY are they to [turn you down / not respond]?
2. How ANGRY would you feel if they [turned you down / did not respond]?
3. How SAD would you feel if they [turned you down / did not respond]?
4. Imagine that they [turned down / did not respond to] your request – How likely do you think it is
BECAUSE OF SOMETHING ABOUT YOU PERSONALLY?

105
Appendix C

106

107

108

Abstract (if available)

Abstract Social nonacceptance, a common yet impactful human experience, has been extensively studied in the field of social psychology. However, the field still lacks a comprehensive framework for the conceptualization and experimentation of social rejection. This dissertation addresses this gap by proposing a taxonomy and an integrative framework for understanding social rejection. A novel experimental paradigm, Spider Apocalypse, is introduced to test these concepts in a controlled, ecologically valid environment. Pilot studies using this paradigm reveal significant effects of rejection contexts on mood, cognition, and behavior, supporting the proposed taxonomy and framework. This work advances the theoretical understanding of social rejection and provides a robust tool for future research.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Examining Pavlovian-to-instrumental transfer effect in human mate selection behavior

PDF

Generalization of social rejection in social networks

PDF

Beyond active and passive social media use: habit mechanisms are behind frequent posting and scrolling on Twitter/X

PDF

Branes, topology, and perturbation theory in two-dimensional quantum gravity

PDF

Displaced aggression as a function of target power

PDF

Classrooms are game settings: learning through and with play

PDF

Predictors and outcomes across the transition to fatherhood

PDF

Meta-analysis on the misattribution of arousal

PDF

Exploring exposures: provider perspectives of CBT with underserved, diverse youth in-person and via telehealth

PDF

Effects of racial exclusion from media

PDF

Trust in the 100% remote workplace in high growth technology consulting firms

PDF

Investigating the effects of Pavlovian cues on social behavior

PDF

The metaphysics of social groups: structure, location, and change

PDF

Montage of interaction: conceptual exploration and creative practice

PDF

Implicit theory ideology about human attributes and aggression

PDF

Pain and multi-morbidity among veterans: theory-guided, data-driven, and narrative approaches

PDF

Exploring the implementation of the Indian Child of Welfare Act of 1978

PDF

Leadership psychological safety: exploring its development and relationship with leader-member exchange theory

PDF

Personality and culture revisited: Integrating the social cognitive approach to personality and the dynamic constructivist approach to culture

PDF

Psychological distance in the public’s response to terrorism: An investigation of the 2016 Orlando nightclub shooting

Asset Metadata

Creator Qiao, Aili (author)

Core Title Jilted and tilted: an exploration of post-rejection response and introduction of a novel experimental paradigm

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Psychology

Degree Conferral Date 2024-08

Publication Date 06/19/2024

Defense Date 06/18/2024

Publisher Los Angeles, California (original), University of Southern California (original), University of Southern California. Libraries (digital)

Tag aggression,Anger,approach,avoidance,exclusion,nonacceptance,OAI-PMH Harvest,Sadness,Social exclusion,social rejection,video games,withdraw

Format theses (aat)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Read, Stephen (committee chair), John, Richard (committee member), Miller, Lynn (committee member), Rizzo, Albert (committee member)

Creator Email aqiao@usc.edu,qiaoalice@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC113996WQW

Unique identifier UC113996WQW

Identifier etd-QiaoAili-13119.pdf (filename)

Legacy Identifier etd-QiaoAili-13119

Document Type Dissertation

Format theses (aat)

Rights Qiao, Aili

Internet Media Type application/pdf

Type texts

Source 20240619-usctheses-batch-1171 (batch), University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright.

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Repository Email cisadmin@lib.usc.edu