THE EFFECTS OF BIAS CRIMES ON SEXUAL MINORITY INDIVIDUALS: A STUDY OF
MINORITY IDENTITY, DISCRIMINATION, AND FEAR RESPONSE
By
Brittany Kay Lannert

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Psychology – Doctor of Philosophy
2014

ABSTRACT
THE EFFECTS OF BIAS CRIMES ON SEXUAL MINORITY INDIVIDUALS: A STUDY OF
MINORITY IDENTITY, DISCRIMINATION, AND FEAR RESPONSE
By
Brittany Kay Lannert
Vicarious traumatization of non-victim members of a community targeted by bias crimes
has been suggested by previous qualitative research (Noelle, 2002). The present study extended
this research in a quantitative examination of a model of proximal physiological, emotional, and
cognitive responses among lesbian, gay, bisexual (LGB) and heterosexual participants during
and immediately following exposure to bias crime information, using a novel quasiexperimental, mixed-factorial design with ecologically-valid stimuli. Participants listened to
news information about neutral, general threat, and bias threat events while skin conductance
level (SCL) and startle eyeblink magnitude were recorded, providing reports of affect following
each clip. Participants reported pre-task negative beliefs about the world and completed post-task
measures of negative cognitions, identity centrality, public regard, and additional variables of
interest (e.g., past victimization).
I hypothesized that responses would differ between LGB and heterosexual participants,
such that LGB participants would show exaggerated defensive mobilization to bias threats. I
further hypothesized that collective identity would magnify defensive responses. Group
differences were observed in startle, affect, and SCL. LGB participants’ affective response and
startle magnitude across conditions were consistent with defensive mobilization, while their
pattern of SCL was inconsistent. Both groups demonstrated declines in negative cognitions,
opposite of prediction. No relations were observed between collective identity and these
responses. Results are discussed in light of defensive mobilization to threat, nonlinear relations

between trauma and physiological responses to threat, social identity, and threat-related
cognition. These results suggest that LGB participants experience bias crime events differently
from heterosexual participants, and indicates that they respond to them as threats, with subjective
and physiological markers of defensive mobilization. Changes in worldview are not consistent
with known responses to acute trauma exposure. Long-term consequences of exposure to bias
crime information for mental health and well-being among LGB young adults are unclear.
Potential pathways to mental health outcomes, such as stress-related dysregulation of the stress
response, are discussed. Future research further explicating this phenomena is strongly
recommended. The present study contributes to the growing literature on minority mental health
and may have implications for clinical practice by informing clinicians of stress-related
experiences of LGB individuals.

TABLE OF CONTENTS

LIST OF TABLES…………………………………………………………………..……..…... vii
LIST OF FIGURES………………………………………………………………..………….... xii
CHAPTER ONE
Introduction………………………………………………………………………..…………….. 1
CHAPTER TWO
The Physiology and Cognition of Stress and Trauma……………………………….….……….. 8
Stress and the Threat Response…………………………………………….….………… 8
Trauma and the Defense Cascade……………………………………………….……… 11
The Defense Cascade and LGB Responses to Bias Crimes…………………….……… 13
Methods in Physiological Assessment of Threat Response……………………….…… 15
Trauma and Cognition…………………………………………………………….……. 20
CHAPTER THREE
Minority Stress………………………………………………………….……...………….....…. 24
Risk of Psychopathology in LGB Populations……………………………………….… 26
LGB Bias Victimization across the Lifespan................................................................... 29
Vicarious Victimization and Discrimination as Trauma: Examples from
Other Literatures................................................................................................... 32
The Impact of Bias Crimes for Victims and Non-Victims............................................... 34
CHAPTER FOUR
Collective Identity........................................................................................................................ 37
Collective Identity, Discrimination, and Psychological Well-Being............................... 41
Collective Identity and Exposure to Bias Crimes............................................................ 42
CHAPTER FIVE
Research with LGB Individuals: History and Methodological Considerations........................... 44
Classification of Participants as LGB.............................................................................. 45
Sampling Methods with LGB Populations...................................................................... 50
Assessment and Analysis of Data with LGB Populations............................................... 52
CHAPTER SIX
Rationale and Hypotheses........................................................................................................... 54
Specific Hypotheses........................................................................................................ 56
CHAPTER SEVEN
Methods....................................................................................................................................... 58
Design.............................................................................................................................. 58
Participants...................................................................................................................... 58
iv

Materials.......................................................................................................................... 63
Auditory stimuli................................................................................................... 63
Psychophysiological data recording.................................................................... 66
Questionnaire Measures.................................................................................................. 66
Demographics...................................................................................................... 67
Sexual orientation................................................................................................ 67
Sexual orientation collective identity.................................................................. 68
Subjective ratings of affect during task presentation.......................................... 68
Assumptive worldview........................................................................................ 70
Negative cognitions about the world................................................................... 71
Past experiences of bias victimization................................................................. 71
Tendency toward negative affect/neuroticism..................................................... 72
Life Events Checklist........................................................................................... 72
Other post-traumatic cognitions (full scale)......................................................... 73
Attention to task................................................................................................... 74
Procedures........................................................................................................................ 74
Identification and enrollment of participants....................................................... 74
Lab task procedures............................................................................................. 75
CHAPTER EIGHT
Results......................................................................................................................................... 80
Preliminary Data Analyses.............................................................................................. 80
Missing values..................................................................................................... 80
Psychophysiological data reduction.................................................................... 81
Questionnaire data preparation........................................................................... 81
Evaluation of statistical and theoretical assumptions......................................... 82
Confirmation of the structure of the World Assumptions Questionnaire........... 82
Individual differences in neuroticism as a possible covariate............................. 83
Primary Analyses............................................................................................................ 84
Descriptive data.................................................................................................. 85
Hypothesis testing............................................................................................... 95
Additional Analyses...................................................................................................... 105
CHAPTER NINE
Discussion.................................................................................................................................. 107
Clinical and Research Implications............................................................................... 117
Strengths, Limitations, and Future Directions............................................................... 119
Conclusions.................................................................................................................... 125
APPENDICES
Appendix A – Participant Consent Form...................................................................... 128
Appendix B – Copies of Measures................................................................................ 132
Appendix C – Figures for Confirmatory Analysis of the Structure of the WAQ.......... 159
Appendix D – Supplemental Tables for Main Variables of Interest............................. 163
Appendix E – Supplementary Variables of Interest...................................................... 183

v

REFERENCES........................................................................................................................... 194

vi

LIST OF TABLES

Table 1.

Sample Demographic Characteristics, Full Sample and in Comparison
by Kinsey Group.................................................................................................. 60

Table 2.

Sample Sexual Orientation Characteristics, Full Sample and by
Kinsey Group (N = 60)........................................................................................ 62

Table 3.

Correlations between Average Skin Conductance Level (SCL) across Clip
Conditions and Dimensions of Collective Identity within the LGB Subsample (in
Plain Text; N = 28) and Heterosexual Subsample (in Bold; N = 32).................. 92

Table 4.

Correlations between Startle Reflex EMG Magnitude across Clip Conditions and
Dimensions of Collective Identity within the LGB Subsample (in Plain Text; N =
27) and Heterosexual Subsample (in Bold; N = 30)............................................ 92

Table 5.

Correlations between Subjective Ratings of Arousal during Neutral, General
Threat, and Bias Threat Clips and Dimensions of Collective Identity within the
LGB Subsample (in Plain Text; N = 28) and Heterosexual Subsample (in Bold; N
= 32)..................................................................................................................... 93

Table 6.

Correlations between Subjective Ratings of Affect Valence and Dimensions of
Collective Identity within the LGB Subsample (in Plain Text; N = 28) and
Heterosexual Subsample (in Bold; N = 32)......................................................... 93

Table 7.

Correlations between Subjective Ratings of Dominance and Dimensions of
Collective Identity within the LGB Subsample (in Plain Text; N = 28) and
Heterosexual Subsample (in Bold; N = 32)......................................................... 94

Table 8.

Correlations between Subjective Ratings of Interpersonal Warmth and
Dimensions of Collective Identity within the LGB Subsample (in Plain Text; N =
28) and Heterosexual Subsample (in Bold; N = 32)........................................... 94

Table 9.

Correlations between Pre- and Post-Task Negative Cognitions about the World
and Dimensions of Collective Identity within the LGB Subsample (in Plain Text;
N = 28) and Heterosexual Subsample (in Bold; N = 32).................................... 95

Table 10.

Results of a 2x3 Mixed Model Repeated Measures Analysis of Variance
Comparing Average Skin Conductance Level During Three Clip Conditions
(Neutral, General Threat, and Bias Threat), Across Heterosexual and LGB
Participants (N = 60)........................................................................................... 96

Table 11.

Results of a 2x4 Mixed Model Repeated Measures Analysis of Variance
Comparing Average Startle Reflex EMG Magnitude During Four Clip Conditions
vii

(Neutral, General Threat, Bias Threat, and Intertrial Interval) Between
Heterosexual and LGB Participants (N = 60)..................................................... 96
Table 12.

Results of a Series of 2x3 Mixed Model Repeated Measures Analyses of Variance
Comparing Four Subjective Ratings of Affect (Arousal, Valence, Dominance, and
Warmth) During Three Clip Conditions (Neutral, General Threat, and Bias
Threat) Between Heterosexual and LGB Participants (N = 60).......................... 98

Table 13.

Results of a Mixed Model Repeated Measures Analysis of Variance Comparing
Negative Cognitions about the World Pre- and Post-Laboratory Task, Between
Heterosexual and LGB Participants (N = 60)................................................... 100

Table 14.

Correlations between Public Regard and Difference Scores between Average Skin
Conductance Level (SCL) across Stimulus Conditions, within the LGB Subsample
(in Plain Text; N = 28) and Heterosexual Subsample (in Bold; N = 32)........... 103

Table 15.

Correlations between Public Regard and Difference Scores between SAM
Arousal Ratings across Stimulus Conditions, within the LGB Subsample (in
Plain Text; N = 28) and Heterosexual Subsample (in Bold; N = 32)................ 104

Table 16.

Means, Standard Deviations, and Ranges of Values for Skin Conductance
Level (SCL), by Group (N = 60)........................................................................ 163

Table 17.

Means, Standard Deviations, and Ranges of Values for Startle Reflex
EMG, by Group (N = 57).................................................................................. 163

Table 18.

Means, Standard Deviations, and Observed Ranges of Values for
Subjective Affective Responses, by Group (N = 60)........................................... 164

Table 19.

Means, Standard Deviations, and Observed Ranges of Values for Preand Post-Task PTCI Negative Cognitions about the World, by Group
(N = 60).............................................................................................................. 164

Table 20.

Means, Standard Deviations, and Observed Ranges of Values for Sexual
Orientation Centrality, Private Regard, and Public Regard (N = 60)............... 165

Table 21.

Correlations between Average Skin Conductance Level (SCL) and Startle Reflex
EMG across Clip Conditions in the LGB Subsample (in Plain Text; N = 27) and
Heterosexual Subsample (in Bold; N = 30)........................................................ 165

Table 22.

Correlations between Average Skin Conductance Level (SCL) and Subjective
Ratings of Arousal across Clip Conditions in the LGB Subsample (in Plain Text;
N = 28) and Heterosexual Subsample (in Bold; N = 32)................................... 166

Table 23.

Correlations between Average Skin Conductance Level (SCL) and Subjective

viii

Ratings of Affect Valence across Clip Conditions in the LGB Subsample (in Plain
Text; N = 28) and Heterosexual Subsample (in Bold; N = 32).......................... 166
Table 24.

Correlations between Average Skin Conductance Level (SCL) and Subjective
Ratings of Dominance across Clip Conditions in the LGB Subsample (in Plain
Text; N = 28) and Heterosexual Subsample (in Bold; N = 32).......................... 167

Table 25.

Correlations between Average Skin Conductance Level (SCL) and Subjective
Ratings of Interpersonal Warmth across Clip Conditions in the LGB Subsample
(in Plain Text; N = 28) and Heterosexual Subsample (in Bold; N = 32)........... 167

Table 26.

Correlations between Average Skin Conductance Level (SCL) across Clip
Conditions and Pre- to Post-task PTCI Negative Cognitions about the World
Subscale Scores in the LGB Subsample (in Plain Text; N = 28) and Heterosexual
Subsample (in Bold; N = 32).............................................................................. 168

Table 27.

Correlations between Startle Reflex EMG and Subjective Ratings of
Arousal across Clip Conditions in the LGB Subsample (in Plain Text; N = 27)
and Heterosexual Subsample (in Bold; N = 30)................................................. 168

Table 28.

Correlations between Startle Reflex EMG and Subjective Ratings of Affect
Valence across Clip Conditions in the LGB Subsample (in Plain Text; N = 27)
and Heterosexual Subsample (in Bold; N = 30)................................................. 169

Table 29.

Correlations between Startle Reflex EMG and Subjective Ratings of Dominance
across Clip Conditions in the LGB Subsample (in Plain Text; N = 27) and
Heterosexual Subsample (in Bold; N = 30)....................................................... 169

Table 30.

Correlations between Startle Reflex EMG and Subjective Ratings of Interpersonal
Warmth across Clip Conditions in the LGB Subsample (in Plain Text; N = 27)
and Heterosexual Subsample (in Bold; N = 30)................................................ 170

Table 31.

Correlations between Startle Reflex EMG across Clip Conditions and Pre- to
Post-task PTCI Negative Cognitions about the World Subscale Scores in the LGB
Subsample (in Plain Text; N = 27) and Heterosexual Subsample (in Bold; N =
30)...................................................................................................................... 170

Table 32.

Correlations between Subjective Ratings of Arousal across Clip Conditions and
Pre- to Post-task PTCI Negative Cognitions about the World Subscale Scores in
the LGB Subsample (in Plain Text; N = 28) and Heterosexual Subsample (in
Bold; N = 32)..................................................................................................... 171

Table 33.

Correlations between Subjective Ratings of Affect Valence across Clip Conditions
and Pre- to Post-task PTCI Negative Cognitions about the World Subscale Scores
in the LGB Subsample (in Plain Text; N = 28) and Heterosexual Subsample (in
Bold; N = 32)..................................................................................................... 171
ix

Table 34.

Correlations between Subjective Ratings of Dominance across Clip Conditions
and Pre- to Post-task PTCI Negative Cognitions about the World Subscale Scores
in the LGB Subsample (in Plain Text; N = 28) and Heterosexual Subsample (in
Bold; N = 32)..................................................................................................... 172

Table 35.

Correlations between Subjective Ratings of Interpersonal Warmth across Clip
Conditions and Pre- to Post-task PTCI Negative Cognitions about the World
Subscale Scores in the LGB Subsample (in Plain Text; N = 28) and Heterosexual
Subsample (in Bold; N = 32)............................................................................. 172

Table 36.

Results of a 2x3 Mixed Model Repeated Measures Analysis of Covariance
Comparing the Interactive Effects of Sexual Orientation Group, Clip
Condition, and Collective Identity Variables on Average Skin Conductance
Level (N = 60).................................................................................................... 173

Table 37.

Results of a 2x4 Mixed Model Repeated Measures Analysis of Covariance
Comparing the Interactive Effects of Sexual Orientation Group, Clip
Condition, and Collective Identity Variables on Startle Eyeblink Reflex
Magnitude (N = 57)........................................................................................... 174

Table 38.

Results of a 2x3 Mixed Model Repeated Measures Analysis of Covariance
Comparing the Interactive Effects of Sexual Orientation Group, Clip
Condition, and Collective Identity Variables on Subjective Ratings of
Arousal (N = 60)................................................................................................ 176

Table 39.

Results of a 2x3 Mixed Model Repeated Measures Analysis of Covariance
Comparing the Interactive Effects of Sexual Orientation Group, Clip Condition,
and Collective Identity Variables on Subjective Affect Valence (N = 60).......... 177

Table 40.

Results of a 2x3 Mixed Model Repeated Measures Analysis of Covariance
Comparing the Interactive Effects of Sexual Orientation Group, Clip Condition,
and Collective Identity Variables on Subjective Dominance (N = 60)............... 178

Table 41.

Results of a 2x3 Mixed Model Repeated Measures Analysis of Covariance
Comparing the Interactive Effects of Sexual Orientation Group, Clip Condition,
and Collective Identity Variables on Interpersonal Warmth (N = 60)............... 179

Table 42.

Results of a Mixed Model Repeated Measures Analysis of Covariance Comparing
the Interactive Effects of Sexual Orientation Group, Clip Condition, and
Collective Identity Variables on Pre-Task to Post-Task PTCI Negative Cognitions
about the World Subscale Scores (N = 60)......................................................... 181

Table 43.

Descriptive Data and Group Comparisons for Supplementary Questionnaire
Variables (N = 60).............................................................................................. 183
x

Table 44.

Bivariate Correlations between Secondary Variables of Interest and Average
Skin Conductance Level across Clip Conditions, within the LGB Subsample
(in Plain Text; N = 27) and Heterosexual Subsample (in Bold; N = 30)........... 187

Table 45.

Bivariate Correlations between Secondary Variables of Interest and Average
Startle Reflex Magnitude across Clip Conditions, within the LGB Subsample
(in Plain Text; N = 28) and Heterosexual Subsample (in Bold; N = 32)........... 188

Table 46.

Bivariate Correlations between Secondary Variables of Interest and Subjective
Ratings of Arousal across Clip Conditions, within the LGB Subsample (in
Plain Text; N = 28) and Heterosexual Subsample (in Bold; N = 32)................ 189

Table 47.

Bivariate Correlations between Secondary Variables of Interest and Subjective
Ratings of Affect Valence across Clip Conditions, within the LGB Subsample
(in Plain Text; N = 28) and Heterosexual Subsample (in Bold; N = 32)........... 190

Table 48.

Bivariate Correlations between Secondary Variables of Interest and Subjective
Ratings of Dominance across Clip Conditions, within the LGB Subsample
(in Plain Text; N = 28) and Heterosexual Subsample (in Bold; N = 32)......... 191

Table 49.

Bivariate Correlations between Secondary Variables of Interest and Subjective
Ratings of Interpersonal Warmth across Clip Conditions, within the LGB
Subsample (in Plain Text; N = 28) and Heterosexual Subsample (in Bold;
N = 32).............................................................................................................. 192

Table 50.

Bivariate Correlations between Secondary Variables of Interest and Pre-task
Negative Cognitions about the World (PTCI), Post-Task PTCI, and Change in
PTCI (Post – Pre), within the LGB Subsample (in Plain Text; N = 28) and
Heterosexual Subsample (in Bold; N = 32)...................................................... 193

xi

LIST OF FIGURES

Figure 1.

Average Skin Conductance Level (SCL) across Stimulus Conditions,
by Group (N = 60)............................................................................................... 86

Figure 2.

Average Startle Reflex EMG Magnitude across Stimulus Conditions,
by Group (N = 57)............................................................................................... 86

Figure 3.

SAM Arousal and Valence Ratings across Stimulus Conditions, by
Group (N = 60).................................................................................................... 88

Figure 4.

SAM Dominance and Warmth Ratings across Stimulus Conditions, by
Group (N = 60).................................................................................................... 88

Figure 5.

Average Pre- and Post-Task PTCI Negative Cognitions about the
World, by Group (N = 60).................................................................................... 89

Figure 6.

Average Centrality, Private Regard, and Public Regard Scores, by
Group (N = 60)..................................................................................................... 90

Figure 7.

Self-Assessment Manikin (SAM) Visual Figures and Rating Scales for Self-Report
Ratings of Affective Arousal, Valence, Dominance, and Interpersonal
Warmth................................................................................................................ 140

Figure 8.

CFA model for the World Assumptions Questionnaire....................................... 153

Figure 9.

Standardized Estimates for the World Assumptions Questionnaire within the
Full Sample (N = 60).......................................................................................... 154

Figure 10.

Standardized Estimates for the World Assumptions Questionnaire within the
LGB Sample (N = 28)......................................................................................... 155

Figure 11.

Standardized Estimates for the World Assumptions Questionnaire within the
Heterosexual Sample (N = 32)........................................................................... 156

Figure 12.

Estimated Marginal Means of Average Skin Conductance Level across
Task Conditions and Kinsey Groups, Controlling for Centrality, Private
Regard, and Public Regard (N = 60)................................................................. 174

Figure 13.

Estimated Marginal Means of Startle Reflex Magnitude across Task
Conditions and Kinsey Groups, Controlling for Centrality, Private Regard,
and Public Regard (N = 57)............................................................................... 175

Figure 14.

Estimated Marginal Means of Subjective Arousal and Affect Valence
xii

Ratings across Task Conditions and Kinsey Groups, Controlling for
Centrality, Private Regard, and Public Regard (N = 60).................................. 178
Figure 15.

Estimated Marginal Means of Subjective Dominance and Interpersonal
Warmth Ratings across Task Conditions and Kinsey Groups, Controlling for
Centrality, Private Regard, and Public Regard (N = 60).................................. 180

Figure 16.

Estimated Marginal Means of Pre- and Post-Task PTCI Negative Cognitions
about the World Subscale Scores by Kinsey Group, Controlling for Centrality,
Private Regard, and Public Regard (N = 60)..................................................... 182

xiii

CHAPTER ONE
Introduction
The October 1998 murder of openly-gay college student Matthew Shepard in Laramie,
Wyoming, sparked a contentious national debate on the subject of bias-motivated crimes against
lesbian, gay, and bisexual (LGB) individuals and prompted calls for bias crime legislation
protecting sexual minorities at the federal level (Torkelson, 2008). The news – and details – of
his victimization were broadcast across the nation and sparked strong emotions and bitter debates
among gay rights and anti-gay rights activists alike (Brooke, 1998). One prominent gay rights
activist, Andrew Sullivan, remarked on the effects of this crime on the gay community: “I think a
lot of gay people, when they first heard of that horrifying event, felt sort of punched in the
stomach. I mean it kind of encapsulated all of our fears of being victimized” (“New Details
Emerge in Matthew Shepard Murder,” November 26, 2004, n.p.). Shepard became a symbol of
anti-gay victimization, and his case prompted calls for hate crime legislation protecting sexual
minorities at the federal level (Torkelson, 2008). Following numerous unsuccessful lobbying
attempts by gay rights groups, on October 28, 2009, President Barack Obama signed into law the
Matthew Shepard Act, extending the coverage of federal hate crime laws to crimes targeted
against individuals on the basis of sexual orientation, gender, and gender identity (Geen, 2009).
The Hate Crime and Statistics Act passed in April of 1990 and required the U.S.
Department of Justice to maintain statistics on bias-motivated crimes, including crimes against
sexual minorities (Otis & Skinner, 1996). In the year of Shepard’s death, the Federal Bureau of
Investigation reported 1,260 crimes motivated by sexual orientation bias, approximately 16% of
all bias crimes reported that year (Federal Bureau of Investigation [FBI], 1998). In the most
recent summary report released by the FBI (2012), 1,318 sexual orientation bias crimes were

1

reported, constituting approximately 19.6% of bias crimes. True base rates of bias-motivated
victimization of sexual minorities are uncertain, however, as bias crimes are underreported by
victims and agencies do not consistently release information on bias crimes (Berrill & Herek,
1990; Dunbar, 2006; Herek, Gillis, & Cogan, 1999; Otis, 2007; Otis & Skinner, 1996).
The Bureau of Justice Statistics (BJS) National Crime Victimization Survey (NCVS)
interviewed a nationally-drawn sample of 143,120 individuals age 12 or older in 2011 regarding
their experiences of crime victimization (Sandholtz, Langton, & Planty, 2013). Between 2007
and 2011, respondents reported more than 46,000 victimization events they perceived to be
motivated by offender bias against sexual orientation or gender expression, nearly ten times more
(annually) than recorded in the FBI statistics. Of these events, less than one-fourth were reported
to authorities by the victim, and only one in ten resulted in a formal complaint. Notably, in this
survey the percentage of respondents (reporting all types of bias victimization) who said they did
not report the crime to authorities because they believed the police could not or would not help
increased from 14% in 2003-2006 to 24% in 2007-2011.
Once a crime is identified by prosecutors as bias-motivated, individuals convicted of
perpetration are subject to greater legal penalties and longer sentences (Berrill & Herek, 1990;
Dunbar, 2006; Geen, 2009; Sullaway, 2004). Increased penalties for bias crime offenders reflect,
in part, the severe nature of these crimes. Research on LGB bias crime victimization in the
United States suggests that LGB bias crimes typically involve multiple offenders, are on average
more severe than ethnic or religious bias crimes, are likely to involve assault, sexual assault,
sexual harassment, and stalking, and involve greater victim distress than non-bias crimes
(Dunbar, 2006; Herek et al., 1999; McDevitt, Balboni, Garcia, & Gu, 2001; Rose & Mechanic,

2

2002). In the NCVS, 92% of bias crime events reported between 2007-2011 were classified as
violent offenses, as compared to less than 25% of non-bias crimes (Sandholtz et al., 2013).
In addition, advocates of social policies against bias victimization have argued that
crimes motivated by bias against an identity affect members of a community for whom this
identity is salient by creating feelings of fear, helplessness, and shock, as well as anger and a
desire for retaliation, as they may lead community members to interpret these events as potential
personal threats (Craig, 1999; Smith & Silva, 2011; Sullaway, 2004). However, this purported
phenomenon has been subjected to little empirical scrutiny. Studies on bias victimization of
sexual minorities typically examine only direct experiences of victimization and exclude
awareness of victimization of others, although researchers note that this may be a relevant factor
in explaining perceptions of victimization and fear of crime among LGB individuals (Otis,
2007). The present study will attempt to remedy this gap in the psychological literature by
examining non-victim responses to bias crimes among individuals who identify as lesbian, gay,
or bisexual, using sound empirical methodology and best research practices. Given that legal,
social, and political arguments about bias crimes frequently rely on the untested assumption of
vicarious effects, and that bias crimes are increasingly the subject of national debate, such a
study is critical to the advancement of our understanding of these putative phenomena.
Given that bias crimes are so frequently defined and prosecuted in reference to the
presumed increased harm inflicted upon targeted communities, it is striking that so little
empirical research examining this purported effect. To date, only three extant studies have
examined the psychological effects of exposure to information about bias crimes on non-victim
members of a targeted population (Craig, 1999; Noelle, 2002; Perry & Alvi, 2012). Craig (1999)
compared post-task subjective reports of emotional responses to the presentation of videotaped,

3

scripted scenes of racially-motivated bias crimes across a sample of White and Black male
college students. He found that the Black participants, as compared to the White participants,
reported greater anger, believing that the crime was more likely to occur to them and was typical
of an experience that they would expect to encounter, and feeling that they would be more likely
to return to the scene later with friends.
Noelle (2002) conducted qualitative analyses of interviews with 9 individuals who selfidentified as lesbian, gay, or bisexual and who reported that they were significantly affected by
Matthew Shepard’s murder. She questioned whether non-victim individuals with LGB identities
might exhibit strong personal reactions to salient, publicized bias crime events and posited that
sexual minorities might respond to such events with shifts in their assumptive worldviews (i.e.,
belief in the benevolence, meaningfulness, and controllability of the world) that are consistent
with cognitive changes following experiences of trauma and crime victimization. These selfselected individuals broadly reported identifying with Shepard and feeling personally threatened,
with a significant proportion reporting increased vigilance to threat, increased vulnerability, and
decreased belief in the justness and benevolence of the world. A subsample reported defensive
changes in behavior such as removing identifying insignia from clothing, while a second group
reported greater identification with and involvement in the LGB community. One participant
reported sub-clinical levels of post-traumatic stress symptoms, including intrusive aversive
thoughts of the event and increased avoidance of event cues.
Most recently, Perry and Alvi (2012) interviewed members of Canadian minority groups
(n = 39) in an effort to understand bias crime vicarious victimization experiences of individuals
within a shared identity group. Because the authors were interested in both shared and specific
responses to bias crimes across minority groups, they interviewed members of seven

4

communities: Aboriginal peoples; African Canadians; Asians; Jews; Muslims; lesbian, gay,
bisexual, and transgender people; and South Asians. A majority (75%) of participants in the full
sample reported that, when learning about victimization of a member of their community, they
feared a similar incident could happen to them or another member of their community. Asked to
rate the impact of bias incidents on the community, 100% indicated that it made them feel
fearful, vulnerable, and suspicious of other groups, and 91.7% reported feeling powerless.
Among the consistent narrative themes identified by the authors were feelings of shock, surprise,
and anger, persistent feelings of fear and vulnerability, behavioral changes aimed at increasing
safety such as social withdrawal, and a belief that such crimes represent purposeful messages of
inferiority and social exclusion against members of the targeted group. The authors argued that
bias crimes serve as symbolic acts aimed at increasing fear and vulnerability among a targeted
community, and that they exert “profound and negative impact[s] on affected communities” (69)
marked by fears of victimization and defensive efforts to increase safety.
The Noelle (2002), Craig (1999), and Perry and Alvi (2012) studies suggest that members
of a targeted community may experience emotional, cognitive, and behavioral changes to the
threat of bias-motivated victimization. However, further research is needed in order to expand
upon these foundations, more comprehensively examining these phenomena while addressing
limitations of existing work. While these studies are novel and advance our understanding of
vicarious responding to bias events, the qualitative interview, correlational and retrospective
design of these studies necessarily limit interpretation of and extrapolation from their findings.
Although all three studies argue for a collective or shared identity status as the putative
mechanism of vicarious threat response, none directly assess this proposed relationship. Finally,
these studies do not examine potential links between exposure to bias crimes and mental health

5

outcomes, which are of particular concern to clinical scientists. Given evidence that LGB
individuals seek mental health services at proportionally higher rates than the rest of the
population (Cochran, Sullivan, & Mays, 2003; Meyer, 2003), it is essential that psychologists
understand the effects of discrimination and other forms of stress on this population.
The present study is designed to assess vicarious responses to bias crimes among lesbian,
gay, and bisexual young adults using quantitative methodology within a quasi-experimental
design. It quantitatively examines group and individual differences in physiological, cognitive,
and affective responses to identity-threat (bias crime) stimuli based on a proposed model
developed from a literature review integrating the biological, psychological, and sociological
fields. The study employs real-time recording of physiological data, subjective self-report
measures of affective response, and pre-post measures of world cognitions in a multi-method
approach designed to address both the research questions at hand and the limitations of prior
research.
This dissertation is organized into 9 chapters. Chapter 2 reviews the biological and
cognitive processes involved in stress and trauma, and proposes peripheral psychophysiological
measurement and assessment of beliefs about the world as methods assess these processes.
Chapter 3 reviews the minority stress model (Meyer, 2003), provides an overview of the
empirical literature on the prevalence and correlates of LGB victimization and mental health
outcomes, and briefly reviews analogous literature in the fields of racial and sexual experiences
of discrimination. A moderator of vicarious threat response, collective identity, is examined in
Chapter 4. Chapter 5 provides a brief overview of the relevant historical context and
methodological issues of conducting research with lesbian, gay, and bisexual individuals.
Chapter 6 integrates the preceding review and provides a rationale and hypotheses for the present

6

study. Chapter 7 outlines in detail the methodology of the study, including study design,
participant identification and recruitment, and data collection. Chapter 8 presents the results of
the study. Chapter 9 provides a discussion of the results and their implications for future research
and clinical practice. Finally, a series of appendices provide study materials and supplemental
information on methodology and statistical analyses.

7

CHAPTER TWO
The Physiology and Cognition of Stress and Trauma
Stress and the Threat Response
The question of whether or not LGB individuals have vicarious responses to bias crime
events is fundamentally a question of stress and the threat response. Stressors are environmental
challenges that are perceived to overwhelm the individual’s immediate capacity to cope,
prompting the individual to react via physiological, cognitive, and/or behavioral changes aimed
at adapting to the stressor (Thoits, 2010). A wide range of stimuli can serve as stressors,
including physical illness, social disruptions, impediments to goals, acute physical danger, and
perception of threats. Stress reactions may occur in response to both objectively threatening
circumstances (e.g., being chased by a lion) and from subjectively threatening circumstances
(e.g., perceptions of discrimination or social rejection). Individuals may exhibit the physiological
changes that mark the threat response in reaction to both immediate and future (anticipated)
events (Berntson & Cacioppo, 2007; Hamm & Weike, 2005).
Physiological responses to environmental stress are characterized by interactions between
central and peripheral nervous system structures (Berntson & Cacioppo, 2007; Bradley & Lang,
2007; Hamm & Weike, 2005; Lang, Davis, & Öhman, 2000). The central nervous system
consists of the brain and spinal cord, whereas peripheral nervous system structures include the
somatic and autonomic nervous systems and extend throughout the musculature and organs. The
somatic nervous system is primarily involved in the reception of external stimuli (e.g., touch)
and coordination of motor movements, and the autonomic nervous system controls visceral
responses and affects cardiovascular, electrodermal, endocrine, and gastrointestinal functioning
(Berntson, Quigley, & Lozano, 2007; Dawson, Schell, & Filion, 2007). The two divisions of the

8

autonomic nervous system, the sympathetic (SNS) and parasympathetic (PNS), are responsible
for mobilizing and conserving the body’s energy resources through the activation and
suppression of physiological arousal, respectively.
When confronted with a stressor, the sympathetic nervous system mobilizes an individual
to confront or avoid the perceived threat (Berntson & Cacioppo, 2007; Cohen et al., 1998; Keary,
Hughes, & Palmieri, 2009; Lang, 1995). Central and peripheral nervous system structures
interact to excite or inhibit autonomic arousal, and central nervous system functions (such as
primary and secondary appraisals; Smith & Ellsworth, 1985) can elicit or override immediate
autonomic responses (Lang et al., 2000; Paterson & Neufeld, 1987). The individual’s subjective
emotional experience is likewise organized in accordance with two motivational systems
(pleasant-appetitive and unpleasant-defensive) which motivate the individual to approach or
withdraw from situations according to physical survival needs (Bradley, Codispoti, Cuthbert, &
Lang, 2001). Emotional responses in the face of survival threats are typically perceived along
two distinct but interactive primary dimensions, arousal (intensity of ANS activation) and
valence (pleasure/displeasure), with positive emotions supporting appetitive motivation and
negative emotions supporting defensive motivation (Lang, 1995).
When confronted with a threat, individuals conduct a series of rapid appraisals of both
the positive or negative valence of the stressor (primary appraisals) and the individual’s capacity
to cope (secondary appraisals; Bovin & Marx, 2011; Carter, 2007; Smith & Ellsworth, 1985;
Taylor, 1999). Perception of threat activates what is known as the defense cascade, a sequence of
physiological response patterns organized according to the imminence of threat and intensity of
fear (Bradley & Lang, 2007; Bradley et al., 2001; Fanselow, 1994; Lang, 1995). Moderate fear
(typical of initial detection of a threat) is characterized by a period of initial freezing marked by a

9

reduction in heart rate, behavioral immobility, moderate increases in autonomic arousal, and
modulation of the startle reflex. At lower levels of fear, the startle reflex may be inhibited rather
than potentiated (Globisch, Hamm, Esteves, & Öhman, 1999). As fear increases further (such as
when contact with a threat becomes imminent), the individual mobilizes physiological resources
in order to act defensively and prompted to action by sharply increased heart rate, autonomic
arousal, and startle potentiation (Fanselow, 1994; Hamm & Weike, 2005; Lang et al., 2000). At
the same time, central nervous system structures within the limbic system (including the
amygdala and hippocampus) interact in a process of memory acquisition and the development of
fear-related memories (Lanius, Bluhm, & Frewen, 2011; Peri, Ben-Shakhar, Orr, & Shalev,
2000; Pole, 2007). The encoding of threat-related memories is a form of aversive conditioning
that may account for the intrusive and reexperiencing symptoms of posttraumatic stress disorder
(PTSD; Bovin & Marx, 2011; Hamm & Weike, 2005; Peri et al., 2000; Pole, 2007). Early
research on the physiology of threat response and defensive mobilization was primarily
conducted using animal models (e.g., Globisch et al., 1999). As technology and methodological
sensitivity have improved, researchers have increasingly examined these responses in humans
and found strikingly similar processes (Bradley & Lang, 2007; Lang et al., 2000).
Of concern for mental health practitioners, repeated activations of this system may have
deleterious effects over time (Berntson & Cacioppo, 2007; Lanius et al., 2011; Thoits, 2010).
Initially, the effects of chronic and inescapable stress manifest in heightened physiological
reactivity (Edgar, Keller, Heller, & Miller, 2007; Peri et al., 2000). Over time, the responsiveness
of this system may be attenuated as the stress response system downregulates (Edgar et al., 2007;
Thoits, 2010). As Bernstein and Cacioppo (2007) note, this initial heightening of response
followed by attenuation under conditions of prolonged stress may provide adaptive value insofar

10

as it allows the organism to keep up with environmental demands by conserving physiological
resources in high stress environments.
However, these adaptive benefits may be exceeded by maladaptive consequences, as seen
in cases of prolonged fear conditioning (Gros, Hawk, & Moscovitch, 2009; Lang et al., 2000;
Peri et al., 2000. The development of phobias (social and specific), in which individuals respond
to normally non-aversive or only mildly aversive stimuli with intense activation of the threat
response system, are believed to represent the extremes of fear conditioning (Bradley et al.,
2001; Gros et al., 2009; Hamm, Cuthbert, Globisch, & Vaitl, 1997; Klorman, Weissbert, &
Wiessenfeld, 1977). The majority of studies examining patterns of physiological responding to
threatening stimuli have used picture-viewing designs in which electrodermal and startle reflex
potentiation responses are recorded concurrently with picture presentation (e.g., Bradley et al,
2001). In these studies, the final stage of the defensive cascade (overt activation, with potentiated
startle and heightened ANS activation) is rarely observed, save in the cases of individuals with
phobias exposed to phobic content or some individuals with PTSD exposed to trauma content
(Gros et al., 2009; Peri et al., 2000). The present study hypothesizes that exposure to bias crime
information is analogous to these cases insofar as it is proposed to lead to activation of similar
mechanisms of fear conditioning. However, this study proposes that these responses are more
analogous to those of traumatic stress, which involve both a process of fear conditioning and
changes in physiological and cognitive.
Trauma and the Defense Cascade
Popular discussions of vicarious responses to bias crimes typically frame these events as
a form of traumatic stress (Sullaway, 2004). The study of traumatic stress has historically
constituted a separate area of study in the psychological literature from general life stress, as

11

traumatic events have been believed to differ from general stressors in both the severity of the
stress event and in patterns of individual outcomes (Carter, 2007; Elhai et al., 2009; Hamm &
Weike, 2005; Peri et al., 2000; Thoits, 2010). Traumatic stressors are defined as those which are
threatening to the life and/or bodily integrity of a person, either directly experienced, witnessed,
or occurring to someone close to an individual (American Psychiatric Association [APA], 2013).
Physiological reactions to extreme stressors that are commonly defined as traumatic may have
unique components (for a review, see Peri et al., 2001).
Most research in traumatic stress assesses posttraumatic stress disorder (PTSD), a
psychiatric disorder with high rates of comorbidity with depression, anxiety, and substance use
disorders (Edgar et al., 2007). Post-traumatic stress symptoms, as defined in the fifth edition of
the Diagnostic and Statistical Manual (DSM 5; APA, 2013), include symptoms of intrusive
recollection, avoidance and/or emotional numbing, changes in cognitive and emotional
functioning (e.g., negative beliefs about the world, self-blame), and symptoms of hyperarousal
such as hypervigilance and heightened physiological reactivity. Findings as to whether PTSD is
marked by chronic or acute manifestations of hyperarousal are mixed (Cohen et al., 1998;
Casada, Amdur, Larsen, & Liberzon, 1998; Edgar et al., 2007; Gerardi, Keane, Cahoon, &
Klauminzer, 1994; Keary et al., 2009; Lanius et al., 2011; Pole, 2007). A meta-analysis of resting
baseline studies of individuals with PTSD diagnoses suggests that PTSD patients have larger
resting heart rates and diastolic blood pressure, suggesting long-term change in cardiovascular
functioning following repeated and prolonged hyperactivity (Pole, 2007). Similarly, the same
meta-analysis found that individuals with PTSD diagnoses generally exhibit larger startle
responses, greater heart rate, and greater skin conductance to aversive stimuli. However, this
meta-analysis drew upon studies with very limited variety in forms of traumatic stress (90%

12

combat trauma) and participant characteristics (primarily male, mid-thirties). Analyses also
showed significant heterogeneity in effect sizes for both resting and stimuli-driven physiological
responses across studies, suggesting that the substantial differences in sample characteristics and
study design likely affect results.
It is important to note that individuals with PTSD do not always exhibit hyperarousal at
resting baseline or in response to trauma cues (Lanius et al., 2011; McTeague et al., 2010;
Pitman, Orr, Forgue, de Jong, & Claiborn, 1987; Pole, 2007). Rather, approximately 30-40% of
PTSD patients exhibit physiological non-responsivity upon exposure to trauma cues (Pitman et
al., 1987). Recent research has suggested that individuals who experience multiple traumatic
events in their lifetime may display blunted reactivity in both autonomic nervous system
activation and startle reflex potentiation (McTeague et al., 2010), potentially in a downregulation
of the stress response as observed in models of general life stress (Edgar et al., 2007; Thoits,
2010). These findings reflect an ongoing area of uncertainty in the traumatic stress literature;
namely, what intraindividual, interindividual, experiential, and situational factors differentiate
responses to traumatic stressors (Bradley & Lang, 2007).
The Defense Cascade and LGB Responses to Bias Crimes
There is evidence within the stress and trauma literature to suggest that LGB individuals
may evidence different responses to bias crime events than heterosexual individuals. For one,
they may make different appraisals of these events. Substantial research indicates that appraisal
moderates the relationship between exposure to a threat and defensive responding (Averrill,
1973; Foa, Zinbarg, & Rothbaum, 1992; Paterson & Neufeld, 1987; Peri et al., 2000). The
appraisal of threat and subsequent stress response depends on a number of factors, as
summarized by Paterson & Neufeld (1987): threats perceived to be more severe, imminent,

13

probable, salient, and threatening to goals produce greater anticipatory stress. The magnitude of
anticipatory stress is proportional to the perceived severity of the threat. Ambiguity about the
severity and imminence of threats may lead to elevated nervous system responses (Foa et al.,
1992), and a perceived lack of control over the situation is also associated with heightened
defensive responding (Averrill, 1973).
As has been demonstrated with studies of general life stress, traumatic stress responses
may likewise be affected by appraisals of the controllability and predictability of the stressor
(Hamm & Weike, 2005; Paterson & Neufeld, 1987; Peri et al., 2000). Social stressors (e.g.,
social exclusion) have also been shown to affect stress hormone levels and reactivity in a unique
manner that other stressors do not (Berntson & Cacioppo, 2007). This may relate to findings that
interpersonal violence is associated with the highest risk of later development of PTSD (Roberts,
Austin, Cortiss, Vander Morris, & Koenen, 2010). Further, cumulative experiences of
unpredictable and uncontrollable stressors may lead to both dysregulation of the stress response
system and changes in processes of cognition and appraisal (Foa et al., 1992; McTeague et al.,
2010). Subjective perceptions of control (e.g., situational dominance) are thus also likely to
moderate emotional and physiological responses in response to threats (Bradley et al., 2001).
The processes involved in the development of symptoms of post-traumatic stress, and the
individual and situational factors that discriminate between individuals at risk for PTSD, are not
particularly well understood at this time (Bovin & Marx, 2011; Elhai et al., 2009). Why a
majority of individuals may experience a traumatic event in their lifetime but only a small
proportion of the population develops symptoms of post-traumatic stress remains unclear,
although some factors (e.g., past trauma, post-trauma support) have been identified (Ozer, Best,
Lipsey, & Weiss, 2003). However, the literature reviewed above suggests that the nature of

14

threat stimuli and appraisal processes may allow for prediction of immediate emotional and
cognitive responses upon acute exposure to threat, if not necessarily allowing for accurate
prediction of long-term consequences (Hamm & Weike, 2005; Peri et al., 2000). This is
particularly relevant to the questions at hand: for some individuals, bias crimes may serve as a
source of uncontrollable, ambiguous, and probable interpersonal threat, resulting in anticipatory
stress responses to bias crime cues. The threat response and mobilization of defensive
functioning may be assessed in-vivo using peripheral physiological measures, and participants
may report on their subjective emotional experience directly. Measurement of both physiological
and affective responses during exposure to bias crime information would have benefits for the
assessment of group and individual differences in responses to perceived threat.
Methods in Physiological Assessment of Threat Response
Psychophysiological methods to study processes of stress, anxiety, and trauma have
historically focused on measurements of peripheral (autonomic) nervous system functioning,
although use of central nervous system measures (PET, EEG, etc.) has increased in recent years
(Edgar et al., 2007). For this reason, markers of central nervous system responding to traumatic
events are still undergoing examination, whereas peripheral markers of threat response are
somewhat more established. Generally, CNS measures are also more expensive to use and may
be prohibitively costly in early studies of psychological phenomena. Consequently, the present
study will focus on the use of peripheral measures of defensive mobilization to threat.
Psychophysiological studies frequently collect baseline data on peripheral nervous
system functioning, such as resting heart rate, and/or event-related data (e.g., exposing the
participant to aversive or trauma-relevant images and measuring short-term PNS responses;
Cohen et al., 1998; Gerardi et al., 1994). Studies of physiology and psychopathology typically

15

assume a categorical approach to mental illness (e.g., examining a PTSD group vs. no diagnosis
group), although dimensional models have also been considered (Edgar et al., 2007).
Physiological studies of psychopathology processes can benefit the study of psychological
processes by providing data that is temporally concurrent with stimulus exposure and allowing
for collection of information on implicit or unconscious responses that may be particularly
impacted by response biases in interviews or questionnaires. Notably, psychophysiological
measures may capture responses that are rapidly suppressed or aborted and that do not reach the
level of conscious (reportable) awareness (Tassinary, Cacioppo, & Vanman, 2007). Thus,
psychophysiological measures can complement subjective reports of phenomenological
experience.
However, physiological studies of psychopathology also have limitations. The
relationship between physiological responses and psychological processes is not always clear,
and discrete observed changes in peripheral nervous system functioning (e.g., heart rate, skin
conductance) may be associated with a range of central nervous system processes of emotion and
motivation (Berntson et al., 2007; Bradley et al., 2001; Bradley & Lang, 2007; Nikula, 1991;
Tassinary et al., 2007). Researchers should be careful to select physiological measures that have
been shown to be related to the construct of interest and to use multi-method approaches to
measuring psychological processes wherever possible (Edgar et al., 2007). For the purposes of
examining differences in affective responses to bias crimes, the present study employs the use of
a standardized stimuli paradigm with psychophysiological and questionnaire measures. This
design allows for examination of group differences in defensive responding measured in vivo,
during presentation of information about events that are standardized for all participants.

16

One reliable method for examining defensive mobilization is to measure startle eyeblink
with electromyography (EMG). Electromyography (EMG) allows researchers to assess
potentiation (electrical activity) of the skeletomotor system that occurs with the experience of
emotions as they are communicated to others through the facial musculature (Blumenthal et al.,
2005; Tassinary et al., 2007; Vrana, Spence, & Lang, 1988). These reactions can be assessed
even when activity is too subtle to be viewed by the naked eye, and even in cases where
individuals make efforts to suppress or avoid the emotional reaction. Startle eyeblink designs
place electrodes over muscles near the outside of the eye (orbicularis oculi) and present sudden,
loud sounds at random intervals through participants’ headphones, then measure the magnitude
of the startle response as an index of fear potentiation (Blumenthal et al., 2005). These designs
may measure startle reflex potentiation during rest (without stimulus presentation) or during
presentation of stimuli of varying valence and/or personal relevance (Hamm et al., 1997;
Globisch et al., 1999; Gros et al., 2009; Pole, 2007). The startle eyeblink design has been widely
employed in psychophysiological assessments of trauma and posttraumatic stress disorder, in
particular as a reliable and sensitive measure of fear acquisition and learning (Edgar et al., 2007;
Hamm & Weike, 2005).
Measurements of the startle eyeblink reflex inform researchers of potentiated fear states,
as the startle reflex is a function of defensive behavioral cascades mediated by neural structures
associated with fear such as the amygdala (Bradley & Lang, 2007; Peri et al., 2000). As
previously noted, when confronted with aversive stimuli, generally humans and animals display
initially attenuated (decreased) stimuli when orienting to a sudden threat, then potentiated
(increased) startle as threat imminence increases (Bradley et al., 2001). This paradigm has been
shown to reliably differentiate responses to aversive or threatening stimuli, as compared to

17

neutral or positive stimuli, and the valence of the startle response has been shown to be
associated with ratings of stimuli aversiveness (Bradley et al., 2001; Bradley & Lang, 2007;
Hamm et al., 2007). However, as with any method there are limitations to consider. The
measurement of electrical activity from the facial musculature is subject to error due to noise
such as movement artifacts or spontaneous blinks, and so care should be taken to differentiate
startle responses from measurement error (Blumenthal et al., 2005). Typically, this
differentiation is conducted via post-task data processing by examining only a very small
window of responding (i.e., 200ms post-startle probe) and using computerized algorithms to
identify noise elements and smooth the data curves.
In addition to fear potentiation, researchers have identified autonomic nervous system
arousal as a marker of defensive mobilization to threat. One method of measuring autonomic
arousal is to examine skin conductance level, or SCL. In SCL studies, two electrodes are placed
at different points on the skin and an electrical current is passed between them, and the
magnitude of the response potential generated reflects the conductance of the skin is related to
activation of the sweat glands (Dawson et al., 2007; Fowles et al., 1981). Skin conductance
measures reflect general autonomic arousal and do not differentiate between different valences of
affect – that is, ANS arousal typically shows a linear increase with intensity of affect of any type
– but it does tend to show large changes in response to aversive stimuli and the anticipation of
threat (Bradley et al., 2001; Bradley & Lang, 2007; Nikula, 1991). The innervation of the
electrodermal system by the sympathetic nervous system alone allows for a straightforward
index of sympathetic arousal, in contrast to the dual innervation (and relative complexity) of the
cardiac system (Bernston et al., 2007).

18

As Foa, Zinbarg, and Rothbaum (1992) note, “repeated experience with uncontrollable,
unpredictable stress is necessary for formation of expectations that danger can occur at any time
and is inescapable” (p. 231). For LGB individuals who may perceive the risk of bias crime
victimization as a personally-relevant, uncontrollable, unpredictable, and probable event,
exposure to bias crime information may lead to anticipatory activation of the ANS and
engagement of the defensive cascade as marked by potentiated startle eyeblink reflexes. The use
of a standardized trauma cue paradigm with electrodermal measurement of startle eyeblink
responses and measurement of ANS arousal via skin conductance level would allow researchers
to examine this response in real time with high specificity of nervous system response.
Establishing a theoretical link between these responses and vulnerability to
psychopathology is complicated, however, because sexual minority individuals are exposed to
numerous social stressors as a function of their minority status, and these stressors have been
shown to relate to poorer mental health outcomes (Diaz, Ayala, Bein, Henne, & Marin, 2001;
Frost & Bastone, 2008; Meyer, 2003; Quinn & Chaudoir, 2009). Further, LGB individuals are
more likely to experience victimization and discrimination in adolescence and adulthood, which
likewise may increase the risk of subsequent psychopathology (Diaz et al., 2001; Dworkin & Yi,
2003; Herek et al., 1999; Levitt et al., 2009; Meyer, 2003; Savin-Williams, 2004; Quinn &
Chaudoir, 2009). Thus, it may be difficult to distinguish the dysregulating effects of responses to
bias crime information on mental health from the effects of related social stressors, and the
cumulative effects of these social stressors may affect responding on startle eyeblink tasks. Thus,
the present study includes measures of past victimization and stressful life events in order to
examine their cumulative effects on task responses.

19

Further, given the complex nature of PTSD (involving physiological, affective, and
cognitive changes over a period of time), a cross-sectional study such as the proposed
dissertation study is necessarily limited in its ability to assess the in-vivo effects of exposure to
bias crimes on the development of symptoms of trauma, as previously noted. At best, such a
study might assess whether immediate responses of LGB individuals to bias crimes are
consistent with elements of acute responses to threatening events believed to be predictive of
pathological outcomes – that is, peritraumatic responses associated with poorer outcomes
following exposure to stressful events (Bovin & Marx, 2011; Pole, 2007). There is sufficient
evidence to indicate that exaggerated startle responses and heightened ANS activation would
suggest a high level of defensive responding (Bradley et al., 2001; Edgar et al., 2007; Hamm &
Weike, 2005; Lang, 1995; Pole, 2007), but observing this pattern would not be sufficient in
suggesting a vicarious trauma response. To establish that such a response is traumatogenic would
require examination of responses that differentiate traumatic events from general stressors. To do
this, we may look to the field of traumatic cognitions research, which examines the development
of maladaptive beliefs about the world, others, and the self that may develop or be reinforced
after traumatic event exposure and which have been identified as symptoms of traumatic stress
(APA, 2013; Dagleish, 2004; Foa et al., 1999; Hagenaars, van Minnen, & Hoogduin, 2007;
Halligan, Michael, Clark, & Ehlers, 2003).
Trauma and Cognition
In addition to further explicating the role of physiological responses to trauma,
researchers have increasingly focused on the role that peritraumatic and posttraumatic cognitive
constructs and processes play (for a review, see Dagleish, 2004). Despite the limited presence of
cognitive symptoms in DSM-III and DSM-IV PTSD criteria, existing therapies for PTSD prior to

20

the release of DSM-5 frequently included components addressing post-traumatic cognitive
processes and beliefs (such as Cognitive Processing Therapy for PTSD; Resick & Schnicke,
1993). Studies of cognitive responses to traumatic events have looked at the role of dysfunctional
cognitions in predicting the development of symptoms of traumatic stress (Hagenaars et al.,
2007; Halligan et al., 2003), and researchers interested in post-traumatic growth processes have
also examined schemas and beliefs as explanatory factors in positive outcomes following trauma
exposure (Linley & Joseph, 2004).
Among the revisions to the fifth edition of the American Psychiatric Association
Diagnostic and Statistics Manual (APA, 2013) were the addition of specific cognitive symptoms
of post-traumatic stress disorder related to negative beliefs about the self, others, and the world
such as “The world is completely dangerous” and “No one can be trusted.” These revisions
emerged from a growing literature examining the role of cognitive processing in posttraumatic
outcomes, both with regard to cognitive beliefs about the event itself (Bovin & Marx, 2011; Foa
et al., 1999) and changes in broader cognitive schemas about the world, self, and others (e.g.,
assumptive worldview, Janoff-Bulman, 1989). As previously noted, neurobiological theories of
the development of PTSD propose that fear learning takes place through the development of fear
association networks that are encoded in a heightened autonomic state and subsequently
repeatedly activated and strengthened in response to real or perceived environmental threat cues
(Pole, 2007). There is evidence to suggest that peritraumatic cognitive responses (e.g., appraisal
of the severity and consequences of the event) and maladaptive posttraumatic cognitive coping
styles (e.g., negative cognitive beliefs about the self and world) predict increased severity of
posttraumatic symptomatology such as intrusive memories and hyperarousal (Agar, Kennedy, &
King, 2006; Ehlers, Mayou, & Bryant, 2003; Ehring, Ehlers, & Glucksman, 2008; Halligan et al.,

21

2003; Mueller, Moergeli, & Maercker, 2008; Nixon & Bryant, 2005; Regambal & Alden, 2009).
Within the context of the present study, examination of traumatogenic cognitions may be
appropriate for a number of reasons. Both Noelle (2002) and Perry and Alvi (2012) interpreted
their qualitative results in relation to changes in cognitive beliefs about the self, others, and
world. In addition, recent findings on the role of cognition in traumatic stress are consistent with
minority stress models that postulate that minority stress occurs in part through broad processes
of appraisal related to beliefs about the self, others, and the world (Meyer, 2003).
Research on cognitive models of posttraumatic stress have generally found relations
between negative beliefs/assumptions about the world and PTSD symptoms (Dekel, Solomon,
Elklit, & Ginzburg, 2004; Foa et al., 1999; Solomon, Iancu, & Tyano, 1997; Ullman, 1997), in
particular the sub-domains of benevolence (of the world and of others), self-worth, and
controllability (of the world and of the self). There is some evidence as well that relations
between these beliefs and the development of posttraumatic stress may be mediated by
differential coping responses (Goldenberg & Matheson, 2005). Diminished beliefs about the
world, self, and others have also been found to mediate the relationship between interpersonal
trauma exposure (e.g., IPV, sexual harassment) and other mental health outcomes such as
depression, grief, and general psychological distress (Collinsworth, Fitzgerald, & Drasgow,
2009; Lilly, Valdez, & Graham-Bermann, 2010; Wickie & Marwit, 2000).
Of interest, a link between vicarious exposure to trauma and change in beliefs about the
world has also been considered by researchers. One study (Feldman & Kaal, 2007)
retrospectively examined cognitive beliefs about the world in individuals vicariously exposed to
trauma (i.e., individuals acquainted with persons who had experienced a significant traumatic
stressor) and found relations between secondary exposure and more negative worldviews. Noting

22

that this is the only known study examining vicarious trauma and beliefs, the authors suggest that
vicarious exposure to traumatic events may similarly affect beliefs, although this study was
necessarily limited by the use of a retrospective design. Although prospective studies are few, a
handful of studies have prospectively measures domains of world beliefs and found that
increased psychological distress is associated with reductions in perceptions of controllability (of
self and world; Ranchor et al., 2010), perceptions of justice (Kaiser, Vick, & Major, 2004), and
self-worth and benevolence of others (Dekel et al., 2004).
To the extent that individuals demonstrate changes in cognition that have been observed
in trauma-exposed samples (e.g., more negative beliefs about the benevolence, controllability,
and predictability of the world, self, and others), there may be tentative (but by no means
definitive) evidence to suggest that similar processes of traumatogenic cognitive change could
occur in impacted individuals. The literature on defensive mobilization reviewed above,
however, suggests that the likelihood of activation of the defense cascade is dependent on
individual perceptions and appraisals of threat. LGB individuals at this time are frequently
exposed to social stressors that are likely to shape these perceptions of threat, including social
exclusion, discrimination, and witnessed or directly experienced violence (Meyer, 2003). These
group-specific experiences are collectively known as minority stress, and are explored in the next
chapter.

23

CHAPTER THREE
Minority Stress
Drawing from a rich theoretical and empirical literature on the relationship between
sociological stress and health spanning more than forty years (for a recent review, see Thoits,
2010), Meyer (1995) proposed the minority stress model to explain observed disparities in health
between LGB and heterosexual individuals. Minority stress refers to the group-specific and
cumulative stressful experiences faced by minority individuals as a result of social
stigmatization, which may include experiences of discrimination, social exclusion, inequality,
lack of opportunity, and threatened or real violence and victimization. As a consequence of
social processes, customs, and institutions, these stressors are inescapable and chronic (Gee, Ro,
Shariff-Marco, & Chae, 2009; Levitt et al., 2009). As stressors that are social in nature, they have
particular implications for the development of mental disorders (Cochran, 2001).
There is growing evidence to suggest that the cumulative effects of these social stressors
may impact both physical and mental health for minority individuals. A recent meta-analysis of
134 studies (N = 71,020) on perceived discrimination and physical and mental health outcomes
found that perceived discrimination was associated with larger physiological stress responses,
more engagement in harmful behaviors (e.g., alcohol use), and increased rates of mental illnesses
such as depression among a range of minority groups (Pascoe & Smart Richman, 2009).
Although the non-experimental nature of the studies analyzed necessarily limits inferences about
causality, the authors noted that those studies including quasi-experimental manipulations
generally showed greater magnitude of physiological and psychological stress responses
following discrimination manipulations.

24

For LGB individuals, sources of minority stress include experiences of individual
discrimination, systematic discrimination and exclusion from social institutions such as marriage,
internalized homophobia/homonegativity, experiences of victimization, and repeated threats of
victimization and marginalization (Meyer, 2003). In addition, LGB individuals face barriers to
health care, including systemic structural disparities such as access to insurance coverage, and
institutional biases that may deter individuals from seeking help such as screening questions
about marital status and heteronormative or highly gendered language on forms and
questionnaires (Dean et al., 2000). LGB individuals may also experience increased stress as a
result of efforts to conceal or hide their identity, which has been associated with increased risk of
negative mental health outcomes including depression (Frost & Bastone, 2008). Among the
myriad of other reported stressors affecting LGB populations are experiences of police
harassment, losing social support when revealing their sexual orientation identity, and having to
move away from friends and family to find social support (Diaz et al., 2001).
The minority stress model explains how distal stressors such as instances of bias
victimization against members of the LGB community may affect proximal individual processes
such as perception and appraisal (Meyer, 2003). It notes, importantly, that minority stress is a
social process that influences individuals, and individuals may respond to these stressors with
dysphoria, anxiety, and dysfunctional behavioral changes, or may respond with resilient coping
methods such as increasing social support and group solidarity to resist negative mental health
outcomes (Branscombe, Schmitt, & Harvey, 1999; Dean et al., 2000; Leonardelli & Tormala,
2003; Smith & Silva, 2011). However, the nature of LGB minority stressors may interfere with
utilization of active coping. For example, there is evidence to suggest that concealment,
internalized homophobia, and victimization are related to less perceived social support and, in

25

turn, more mental health problems and substance use among LGB individuals (Frost & Bastone,
2008; Leharot & Simon, 2011; Meyer, 2003; Quinn & Chandoir, 2009). For this reason, research
in minority stress among LGB individuals has generally focused on negative physical and mental
health outcomes (Meyer, 2003; Pascoe & Smart Richman, 2009).
While a growing body of research exists examining the effects of minority stress – in
particular, discrimination and traumatic response – in racial minority samples (Bryant-Davis,
2007; Carter, 2007; Gee et al., 2009; Pieterse, Carter, Evans, & Walter, 2010), research utilizing
the minority stress model in LGB samples is relatively underrepresented in the literature but
represents a rapidly growing area of research, which has consistently found links between
sources of minority stress and poorer mental health and well-being (Diaz et al., 2001; Dworkin &
Yi, 2003; Leharot & Simon, 2011; Levitt et al., 2009; Lewis, Derlega, Griffin, & Krowinski,
2003; Meyer, 2003).
Risk of Psychopathology in LGB Populations
Mood disorders, anxiety disorders, and substance use disorders have been linked to social
stressors, including stigma, experiences of discrimination, and victimization (Blanchard,
McKittrick, & Blanchard, 2001; Cochran, 2001; Connor-Smith & Compas, 2002; Rhodes &
Jason, 2002; Taylor & Turner, 2002). Thus, to the extent that LGB individuals experience greater
rates of social stress than non-LGB populations, we should expect to see elevated rates of these
disorders. To test this argument, Meyer (2003) performed a meta-analysis of 10 studies (Full N =
876 lesbian, gay, bisexual; 27,562 heterosexual) conducted prior to 2001 and assessing rates of
psychiatric diagnoses between LGB and heterosexual participants. His analysis found that LGB
men and women were approximately 2.5 times more likely to be diagnosed with a mental illness
(mood disorder, anxiety disorder, substance use disorder) over the course of their lifetime, with

26

significantly greater odds of mood disorders, anxiety disorders, and substance use. This analysis
was limited by the nature of the included studies, which used varying definitions of sexual
orientation, frequently assessed different disorders, and often relied on nonrandom sampling
techniques. In light of the limitations of nonrandom sampling with LGB participants, Meyer
stratified the data into studies using randomized or nonrandom sampling techniques. Notably, he
found that differences in rates of mood disorders, anxiety disorders, and substance use disorders
were found only in studies that randomized their sampling techniques. This may reflect a
limitation of nonrandom sampling methods such as snowball sampling, in that individuals with
less distress and greater overall functioning may be more likely to refer others (who may
likewise be less distressed and higher functioning) to participate in these studies (Moradi, Mohr,
Worthington, & Fassinger, 2009).
Early studies comparing heterosexual and LGB mental health have been criticized for the
use of brief screening scales rather than diagnostic interviews and questionnaires, volunteer or
convenience samples, and inconsistent or inappropriate methods of classifying individuals on the
basis of sexual orientation and identity (Gilman et al., 2001; Otis & Skinner, 1996). However,
recent studies in LGB mental health have been subject to greater methodological rigor in
epidemiological research, including the use of more advanced methods of sampling and
standardized, well-supported diagnostic instruments (Cochran, 2001). In general, more recent
studies using more sophisticated population-based methods, diagnostic interviews or
questionnaires, and multiple methods of sexual orientation classification have suggested that
rates of mental illness, particularly rates of anxiety disorders, mood disorders, and substance use
disorders, are higher among LGB populations than typically found in the general population,
although the magnitude of the difference (and which disorder is found to be elevated) varies

27

across studies, with observed effect sizes ranging widely from small to large (Cochran, 2001;
Cochran et al., 2003; Gilman et al., 2001; Green & Feinstein, 2012; McCabe, Hughes, Bostwick,
West, & Boyd, 2009).
It is also important to note, however, that not all studies have demonstrated significant
differences in rates of mental illness between LGB and heterosexual individuals (e.g., Alessi,
Meyer, & Martin, 2011). Further, many population-based representative studies, particularly
prior to the turn of the 21st century, were not designed to study sexual minority mental health a
priori and so frequently were not designed in consideration of the sampling and assessment
issues in research with LGB populations and relied on small subsamples (Meyer, 2003).
Generally, studies using sophisticated (e.g., multi-method) assessment of sexual orientation
increasingly show an elevation of risk for mood, anxiety, and/or substance disorders, with some
variation across subgroups and by gender, and this effect has been demonstrated in a range of
both nonrepresentative targeted sampling studies (e.g., Hellman, Sudderth, & Avery, 2002;
Herrell et al., 1999; Sheets & Mohr, 2009) and population-based representative studies (e.g.,
Burgard, Cochran, & Mays, 2005; Cochran & Mays, 2000; Cochran et al., 2003; Sandfort,
Bakker, Schellevis, & Vanwesenbeeck, 2006). LGB individuals are also more likely to utilize
mental health care services, which may reflect the increased need among this population
(Cochran et al., 2003; Meyer, 2003; Tjepkema, 2008).
Prior to the removal of homosexuality as a mental illness from the diagnostic cannon,
researchers generally ascribed increased rates of mental illness among LGB populations to the
(presumed) disorder of homosexuality itself (Cochran, 2001). However, in recent decades
researchers working with LGB populations have increasingly argued that the elevated risk of
psychopathology found among LGB adults is due to increased negative experiences and stress

28

incurred as members of a stigmatized minority group, including higher rates of victimization and
fears of victimization (Diaz et al., 2001; Dworkin & Yi, 2003; Herek et al., 1999; Leharot &
Simon, 2011; Levitt et al., 2009; Meyer, 2003; Savin-Williams, 2004). Cramer and colleagues
(2012), assessing LGB and heterosexual victims of crime upon presentation at a trauma center,
found that LGB orientation was associated with stronger acute stress responses and more
pervasive general anxiety. The authors argue that sexual minority status leads to the
internalization of social-identity threats, such that stigmatized individuals become hypervigilant
to cues of identity threat and develop biases to threat that increase vulnerability to the
development and maintenance of acute stress and anxiety responses.
LGB Bias Victimization across the Lifespan
One particularly damaging stressor disproportionately experienced by LGB individuals is
interpersonal violence in childhood and adulthood, including physical and sexual assault,
destruction of property and vandalism, and threats of violence (D’Augelli, Grossman, & Starks,
2006; House, Van Horn, Coppeans, & Stepleman, 2011; Roberts et al., 2010). Anti-gay violence
has been described as a central stressor affecting LGB populations, and past experiences of
victimization may increase fears of further victimization (Meyer, 2003; Otis, 2007).
Across multiple studies, a majority of LGB participants have reported childhood and
adolescent experiences of bias victimization – that is to say, victimization perceived to be due to
their sexual orientation - on the part of peers and family members, including bullying, social
isolation, and verbal, physical, and sexual abuse, and these experiences are typically elevated in
frequency, severity, and persistence as compared to non-LGB controls (D’Augelli et al., 2006;
Savin-Williams, 1994; Roberts et al., 2010) and even as compared to non-LGB siblings reared
within the same home (Balsam, Rothbaum, & Beauchaine, 2005). These data have been

29

proposed as one explanation for why risk of psychiatric disorders and suicidality for LGB
individuals is particularly high in adolescence and young adulthood (Cochran, 2001).
Victimization by peers and family members in youth has been associated with increased risk of
poorer school performance, truancy, dropping out of school, running away, conflict with the law,
substance abuse, and suicide (D’Augelli et al., 2006; Savin-Williams, 1994). Dragowski and
colleagues (2011) found that verbal and physical sexual orientation bias-motivated victimization
explained a significant proportion of the variance in posttraumatic stress symptoms among LGB
youths, above and beyond the effects of non-bias life stressors.
LGB adults also report disproportionate experiences of victimization, including physical
assault, sexual assault, threats of violence, and vandalism (Cochran, 2001; Dworkin & Yi, 2003;
Herek et al., 1999; House et al., 2011; Otis, 2007; Otis & Skinner, 1996; Rose & Mechanic,
2002). In recruiting participants for a study of vicarious traumatization in bias crimes, Noelle
(2002) noted that initial plans to exclude individuals with past experiences of victimization were
changed due to high rates of past bias victimization, as approximately 79% of the initial sample
reported at least one such experience. In one study, 73% of LGB participants recruited at a gay
pride rally reported having experienced at least one such incident of bias victimization, which
were correlated with traumatic stress symptoms and helpseeking behaviors (Rose & Mechanic,
2002). Otis and Skinner (1996) found that 50% of subjects (recruited via mailing lists, chain
referrals, and convenience sampling at a gay pride event) endorsed experiencing at least 1 act of
victimization in the previous two years that they attributed to sexual orientation bias (noting that
this is at the low end of most prevalence estimates). A study by Herek and colleagues (1999)
sampled over 2,000 gay and bisexual men and women through snowball and chain referral
sampling and found that one-fourth of the men and one-fifth of the women reported that they had

30

experienced at least one act of criminal victimization since age 16 that they attributed to sexual
orientation bias. Individuals who had experienced a bias crime within the past five years reported
elevated rates of traumatic stress, mood, and anxiety symptoms.
These three studies (Herek et al., 1999; Otis & Skinner, 1996; Rose & Mechanic, 2002)
reflect the limitations of early research on bias victimization, in particular the use of small, nonrepresentative samples recruited via convenience sampling, which tend to result in higherfunctioning samples (Moradi et al., 2009). These studies relied on retrospective reporting of
events and did not verify these incidents’ occurrence; consequently, the accuracy of the reported
rates is unclear. It is possible that higher rates of retrospective reporting of bias victimization are
associated with response biases associated with individual differences in recall and reporting
(e.g., neuroticism). There may also be an additional confounding effect of sampling; multiple
studies (typically with racial and gender minorities) suggest that greater perceived discrimination
is associated with greater identification with and involvement with one’s stigmatized group,
although the direction of this effect is unclear (Branscombe et al., 1999; Leonardelli & Tormala,
2003; Sellers & Shelton, 2003; Smith & Silva, 2011). Finally, these studies did not consistently
examine links to mental health outcomes.
More recent studies of LGB victimization have attempted to focus less on gauging the
rates of LGB victimization and more on assessing the relations (direct and indirect) between
experiences of violence and mental health outcomes (D’Augelli et al., 2006; House et al., 2011;
Leharot & Simon, 2011; Roberts et al., 2010). Much of this research has been conducted with
youth and adolescents, particularly through examination of bias victimization in schools (e.g.,
Birkett, Espelage, & Koenig, 2009). These studies have also substantiated a link between
experiences of bias victimization and poor mental health outcomes, although they have (as in

31

earlier studies) generally relied on nonrepresentative convenience, snowball, or chain sampling
techniques (D’Augelli et al., 2006; Dragowski, Halkitis, Grossman, & D’Augelli, 2011; House et
al., 2011; Leharot & Simon, 2011).
Studies examining the relationship between sexual orientation, crime victimization, and
mental health have increasingly employed more advanced sampling and assessment
methodology. Roberts and colleagues (2010) employed a nationally-representative sample (n =
33,046 self-identified heterosexual, 1022 self-identified gay, lesbian, or bisexual) and found that
LGB individuals reported more experiences of interpersonal trauma (childhood maltreatment,
assault, etc.), early exposure to traumatic events, and higher lifetime risk of PTSD. Although this
study did not specifically ask about bias motivation in defining traumatic events, the authors note
that bias victimization likely account for some of the difference in rates of exposure to traumatic
events.
Vicarious Victimization and Discrimination as Trauma: Examples from Other Literatures
In recent years, there has been a movement by some researchers working with
marginalized groups to categorize experiences of discrimination as traumatic, given the
observation that emotional reactions to acts of discrimination often include feelings of shock,
helplessness, and fear (Avina & O’Donahue, 2002; Bryant-Davis, 2007; Carter, 2007; Dunbar,
Liu, & Horvath, 1995: Harrell, 2000), and the cumulative effects of exposure to discrimination
include patterns of avoidance, hypervigilance, identity confusion, feelings of guilt and shame
(lowered self-worth), and interpersonal difficulties, as well as pervasive perceptions of the
environment as hostile (Pieterse et al., 2010; Sanchez-Hucles, 1999). Although the majority of
these studies have focused on racial and gender discrimination (and the intersectionality thereof),
they have generally adopted a broad view on the basis of social psychological theories that

32

“societal traumas include not only racism but also sexism, poverty, heterosexism, and religious
intolerance” (Bryant-Davis, 2007, p. 140). In addition to examining experiences of
discrimination as potential sources of traumatic stress, researchers in these fields have also
increasingly considered physiological processes as both markers of physical health and potential
mechanisms of stress-related changes in mental health (Carter, 2007).
There is evidence of changes in autonomic nervous system functioning during and as a
consequence of experiences of racial- and gender-based discrimination (Burrow & Ong, 2010;
Harrell, Halls, & Taliaferro, 2003; Kiang, Blumenthal, Carlson, Lawson, & Shell, 2009;
Richman, & Jonassaint, 2008). Racism has been identified as a stressor that may increase
vulnerability to negative mental health outcomes (Clark, Anderson, Clark, & Williams, 1999;
Williams & Williams-Morris, 2000), and racial discrimination has been linked to dysregulation
of the threat response system (DeSantis et al, 2007; Kiang et al., 2009; Lepore et al., 2006). For
example, exposure to a highly-publicized racial stressor (the Duke Lacrosse scandal) was
associated with sustained elevated cortisol levels in racial minority college students postexposure (as compared to pre-exposure levels; Richman & Jonassaint, 2008).
Similarly, research in the field of sexual harassment suggests that individuals may
experience negative outcomes in response to sexual harassment experienced both directly and
indirectly, and these negative outcomes are consistent with traumatic exposure (e.g., MinerRubino & Cortina, 2007); as such, multiple researchers have questioned whether individuals
exposed to environments in which sexual harassment occurs (experienced directly or indirectly)
may meet criteria for diagnoses of trauma-related disorders (Avina & O’Donahue, 2002;
Stockdale, Logan, & Weston, 2009; McDermut, Haaga, & Kirk, 2000; Wolfe, Crooks, Chiodo,
& Jaffe, 2009; Woods, Buchanan, & Settles, 2009). It is possible that similar patterns of response

33

may be observed in other minority groups exposed to social identity-based discrimination and
violence, such as sexual minorities exposed to information about bias-motivated violence.
The Impact of Bias Crimes for Victims and Non-Victims
The data on LGB experiences of violent victimization and mental health outcomes
highlight the high prevalence and harmful effects of direct trauma exposure among this
population (D’Augelli et al., 2006; Otis & Skinner, 1996; Savin-Williams, 1994; Roberts et al.,
2010). However, discrimination represents a “continuum of violence” with forms ranging from
stigmatizing and emotionally painful exclusion from and devaluation within social contexts, to
social and cultural institutions and practices that support systemic inequalities, to the threatened
or actual use of force and power in maintenance of these inequalities (Sanders-Phillips, 2009, p.
174). One recent study of LGB individuals exposed to non-life-threatening (non-Criterion A1)
bias-motivated events reported similar behavioral, cognitive, and emotional changes in response
as were observed in individuals exposed to life-threatening bias-motivated events (Alessi et al.,
2013), suggesting that responses to discrimination may likewise lie on a continuum of distress.
Most importantly for the purposes of the present study, there is some evidence that
individuals not directly exposed to victimization may also show symptoms of distress and
changes in cognition, mood, and behavior consistent with experiences of traumatic stress (e.g.,
Herek et al., 1999), although this research is sparse in comparison to the accumulated findings of
the racial discrimination literature (Bryant-Davis, 2007; Carter, 2007). In addition to the studies
explicitly examining vicarious responses to bias crime previously reviewed (Craig, 1999; Noelle,
2002; Perry & Alvi, 2012), other researchers have assessed whether simply living within a
broader social context in which victimization occurs is associated with psychological distress
(Duncan & Hatzenbuehler, 2014; Herek et al., 1999; Meyer & Grollman, 2014). Higher past-year

34

suicidality has been found among LGB adolescents living in neighborhoods with a higher
prevalence of bias crimes against sexual minorities (Duncan & Hatzenbuehler, 2014). LGB
participants across multiple studies have reported more fears of crime, fears of being victimized,
reduced belief in the benevolence of the world and of others, reduced perceptions of selfmastery, and more attributions to prejudice (Herek et al., 1999; Meyer & Grollman, 2014).
Witnessing hostility, incivility, and harassment against other LGB individuals on college
campuses has been shown to partially mediate the relations between sexual minority status and
anxiety (Woodford, Han, Craig, Lim, & Matney, 2014), substance use (Woodford, Krentzman, &
Gattis, 2012), and physical health problems (Woodford, Howell, & Silverschanz, 2012) among
college students. Living in states with discriminatory policies has been associated with higher
rates of psychiatric morbidity among LGB adults (Hatzenbuehler, McLaughlin, Keyes, & Hasin,
2010). In one study, LGB young adults who had been raised in highly stigmatizing environments
(based on level of structural/legal/organization stigma at the state level) demonstrated blunted
cortisol responses in a stress task (Hatzenbuehler & McLaughlin, 2014). The authors concluded
that “the stress of growing up in environments that target gays and lesbians for social exclusion
may exert biological effects that are similar to traumatic life experiences” (p. 39).
It is plausible to hypothesize that exposure to information about bias crime events would
serve as a form of minority stress with implications for physiological, emotional, and cognitive
functioning among LGB individuals. However, it is important to note that not all LGB
individuals who experience victimization and discrimination develop symptoms of
psychopathology or report lasting distress (Dean et al., 2000; Levitt et al., 2011; Otis, 2007). It is
likely that individual factors may affect responsiveness to actual and threatened victimization,
even if group-level effects exist. As theorized by Meyer (2003), one critical interindividual factor

35

likely to moderate responses to minority stressors is the salience of and evaluations that an
individual makes about their social identity. An individual’s perception of their social group
membership and the value placed upon that membership is known as collective identity
(Ashmore, Deaux, & McLaughlin-Volpe, 2003).

36

CHAPTER FOUR
Collective Identity
Collective identity, previously referred to as social identity, refers to an individual’s
evaluation of membership in a group perceived to have some shared characteristics and fate
(Ashmore et al., 2003). An individual may develop a collective identity with reference to an
ascribed group, such as a gender or ethnic group, or with reference to an achieved group, such as
a professional or occupational group. Collective identity is subjective and self-defined. As it is
psychological in nature, it may influence an individual regardless of the level of physical contact
with the collective group. Individuals have multiple collective identities that become more or less
salient within different social contexts. Collective identity has been identified as a relevant
construct for minority stress and coping among LGB-identified individuals (Meyer, 2003).
Current theories of collective identity are based heavily in Tajfel’s (1978) social identity
theory and Turner’s (1987) self-categorization theory, which refer to, respectively, an
individual’s self-conceptualization as a member of a social group and the cognitive processes
and conditions that then influence social categorization and behavior. Tajfel’s (1978) social
identity theory was developed initially in order to explain patterns of interactions between
groups, including social prejudice, discrimination, and stereotyping. Noting that individuals
placed into arbitrary groups exhibited preferences for the in-group and prejudices against outgroups, he argued that individuals define themselves and organize their behavior according to
their perceived status in social categories and the emotional significance of their group
membership. Thus, social categorization shapes self-identity, attributions, and behavior, and this
categorization depends on social context. According to self-categorization theory, an extension
of social identity theory that explains cognitive mechanisms of individual behavior in social

37

contexts, individuals attune to environmental cues relevant to their own identity (Smith & Silva,
2011).
Reviewing the literature on collective identity and noting the existence of considerable
disparity across models of collective identity, Ashmore, Deaux, and McLaughlin-Volpe (2003)
proposed unifying these disparate models according to shared, consistent characteristics. They
noted seven relatively consistent elements across models, albeit with differing rates of inclusion
and levels of emphasis. As a precondition, there must be some form of self-categorization, or
division into social categories. Second, there is some form of evaluation, or positive or negative
attitude to the identity. These include favorability judgments that individuals hold about the
group as well as judgments of how one believes others perceive the group. Third, there is some
judgment of the degree of importance of group membership to the individual’s overall selfconcept. These include both evaluations of “centrality,” or conscious appraisal of importance,
and “salience,” or readiness to act on an identity (which may not explicitly be recognized by the
individual). Fourth, there is often an attachment or sense of interdependence, including both a
feeling of belonging (such as affective responses to group symbols) and a sense of shared fate.
Fifth, there may be some variation in social embeddedness, or the degree to which one’s identity
is involved in social relationships in day-to-day life – which may in turn reflect an individual’s
level of group commitment or the cost of abandoning the identification. Sixth, there may be
behavioral involvement in group events or causes.
Finally, individuals may derive content and meaning from their collective identification,
whether with regard to self-attributions of traits or toward the development of a particular
ideology or social stance. Thus, individuals are likely to vary widely in their expression and
interpretation of any given collective identity according to their standing on each of these

38

dimensions. Although existing models of collective identity vary in the choice of and relative
emphasis that they place upon each of the elements outlined above, a recent meta-analysis of
nearly two hundred studies using a range of ethnic identity models and well-being concluded that
results did not vary by measures, reflecting the shared constructs of these models (Smith & Silva,
2011). Models of collective identity have been applied primarily in the study of inter- and intragroup processes, and have been applied heavily in studies of minority mental health (Leonardelli
& Tormala, 2003; Sanders-Thompson, 1996; Sellers & Shelton, 2003; Settles, Navarrete,
Pagano, Abdou, & Sidanius, 2010). In general, self-identity and self-categorization theories have
shaped the development of individualized models of collective identification for a broad range of
minority identities (Benish-Weisman & Horenczyk, 2010; Bombay, Matheson, & Anisman,
2010; Houkamau & Sibley, 2010; Quinn & Chaudoir, 2009; Rughinj, 2011; Yip & Cross, 2004).
Building upon social theories of collective identity and historical research with ethnic
minorities, Sellers and colleagues (1998) proposed the Multidimensional Model of Racial
Identity (MMRI) in an attempt to posit a mechanism by which group identity affects an
individual’s interpretation of and response to the social environment via situational appraisals
and behaviors. Noting the constraining influence of context on behavior and the importance of
subjective appraisals of environmental events, the MMRI is primarily concerned with the
individual’s perception of their identity, the “significance and meaning that [individuals] place
on race in defining themselves” (p. 19). Similarly, because the MMRI is concerned primarily
with predictive validity rather than internal validity, it adopts a cross-sectional approach that
focuses on an individual’s status along relevant dimensions of collective identification and does
not require the categorization of individuals into different stages of identity development.

39

The model posits four major domains of racial identity: centrality, salience, regard, and
ideology. Centrality and salience refer, respectively, to the extent to which an individual
identifies him or herself as belonging a racial group and the relative importance of that
association within specific contexts, reflecting the two aspects of importance described by
Ashmore and colleagues (2003). Regard refers to the positive or negative affective judgments an
individual makes of the group, and reflects two aspects of evaluation: private regard, or the
individual’s subjective evaluation of the extent to which he or she feels positive or negative
feelings about the group and their membership within the group; and public regard, or the
individual’s evaluation of how he or she believes others evaluate the group (Sellers, Smith,
Shelton, Rowley, & Chavous, 1998). Finally, ideology refers to beliefs and attitudes about how
the individual feels members of the group should react (e.g., through distancing or assimilation),
and reflects the culturally-specific content and meaning that individuals derive from membership
in a racial group (Ashmore et al., 2003). The salience of public and private regard evaluations
may vary across social contexts, but centrality reflects the chronic accessibility of identity
constructs in appraisals and behavior (Sellers et al., 1998).
This model of identity, particularly with regard to the domains of centrality, salience, and
regard, has been successfully adapted for use with a wide variety of racial, ethnic, religious, and
sexual orientation groups, including Chinese Americans, Eastern European Roma adults, First
Nations peoples, individuals with stigmatized medical conditions, Maori peoples, Israeli
immigrants, and U.S. sexual minorities (Benish-Weisman & Horenczyk, 2010; Bombay,
Matheron, & Anisman, 2010; Houkamau & Sibley, 2010; Quinn & Chaudoir, 2009; Rughinj,
2011; Yip & Cross, 2004). The MMRI has also been used as a predictive model in many studies
of discrimination, perceptions of bias, well-being, and coping (e.g., Settles et al., 2010).

40

Literature on collective identity suggests that identity salience, centrality, and regard may reflect
general cognitive processes of social identification and may translate easily across social groups,
although ideology appears to reflect the unique experiences and values of particular social
groups (Sellers et al., 1998).
To date, no large-scale studies have compared the psychometric properties of the
modified MMRI across ethnic, gender, or sexual orientation minority groups, and so caution
must be taken in interpreting these constructs across identity groups. However, given that no
comparable, independent cross-sectional model of collective identity yet exists for LGB persons,
the constructs examined within the MMRI model have thus far successfully been examined
across other groups repeatedly, and meta-analysis suggests that the core constructs of collective
identity maintain relatively consistent relations with stress and well-being variables across
measures (Smith & Silva, 2011), applying this model to LGB participants seems appropriate for
the current study.
Collective Identity, Discrimination, and Psychological Well-Being
Discrimination against one’s group has been associated in some studies with less private
regard and lower self-esteem (e.g., Leonardelli & Tormala, 2003); however, under some
circumstances collective identity can serve a protective purpose (e.g., Branscombe et al., 1999).
There is some evidence to suggest that stigma may lead to increased group coping efforts and
group cohesion, whereas other research suggests that immediate and long-term effects may
include negative outcomes (Leonardelli & Tormala, 2003; Meyer, 2003). High public regard has
been shown to be either unrelated to individual psychological functioning or to relate to lower
levels of distress in samples (e.g., Sellers & Shelton, 2003), whereas high private regard has
consistently been shown to insulate members of stigmatized groups from negative mental health

41

outcomes (Settles et al., 2010). Highly-identified individuals may also be more likely to perceive
actions as discriminatory and report experiencing more psychological distress (Leonardelli &
Tormala, 2003; Sellers & Shelton, 2003).
A recent meta-analysis (Smith & Silva, 2011) of 184 studies of ethnic identity and one or
more aspects of well-being (n = 41,676) found varied effect sizes of identity on well-being across
studies, with greater effect sizes in younger populations. The authors speculated that these results
may follow from a relatively greater influence of identity on self-perception and well-being in
adolescence and young adulthood (Berman, Schwartz, Kurtines, & Berman, 2001), but caution
that relatively fewer studies have been conducted with middle-aged and older adults and so
further replication is needed. As also noted by Smith and Silva (2011), studies of the relationship
between collective identify, discrimination, and coping use cross-sectional data and by nature are
unable to employ experimental designs, which limits interpretation of causality. For the purposes
of the present study, examining identity in young adults may be particularly likely to reveal
relations with discrimination, given these findings.
Collective Identity and Exposure to Bias Crimes
Research on the effects of discrimination on minority groups has relied heavily of models
of minority stress and collective identity to explain both the processes by which individuals are
affected by stigma and victimization and to provide an explanation for the diversity of responses
to marginalization (Bryant-Davis, 2007; Carter, 2007; Sellers & Shelton, 2003; Settles et al.,
2010). These responses include both positive cognitive and behavioral coping strategies and
changes in self-worth, perceptions of group status, increased feelings of social isolation and
rejection, avoidance behaviors, and other risk factors for the development of psychopathology.

42

These models provide a useful framework within which to make predictions about the
differential responses of LGB individuals upon exposure to bias crimes.
As noted by Dunbar (2006), reactions to and effects from bias crimes are likely to vary
depending on the individual’s social identity. Individuals who identify highly with the in-group
may interpret events such as bias crimes as more salient to their group membership, consistent
with social identity and self-categorization theories (Smith & Zarate, 1992). Thus, I would
predict that individuals who perceive their LGB identity as more central would exhibit stronger
defensive responding to bias crime information. Similarly, lower public regard (perception that
out-group members value or devalue the in-group) seems likely to elicit stronger defensive
responding insofar as it relates to appraisals of the likelihood and imminence of threat. I would
predict that individuals who perceive public regard as low will demonstrate greater activation of
the threat response system. Finally, I would predict that the combination of high centrality and
low public regard would predict even greater defensive activation, as it would positively predict
immediate appraisals about threat salience, imminence, and likelihood.

43

CHAPTER FIVE
Research with LGB Individuals: History and Methodological Considerations
Before designing any research study with sexual minority individuals, it is important to
review the historical context and methodological considerations specific to research with this
population in order to provide a context within which to interpret existing research and a basis
upon which to determine best methodological practices in the present study. The history of
psychological research with LGB individuals has been colored by social prejudices that shaped
the conflation of sexual orientation with mental wellness for decades (Meyer, 2003). It was only
in the 1950s that psychological research with LGB-identified populations began to use nonpsychiatric samples. Prior to this time, research with sexual minorities often used evidence of
mental illness among institutionalized gay men to justify the assertion that homosexuality was
equated with mental illness (Cochran, 2001). As evidence mounted in support of the removal of
homosexuality as a diagnosable mental illness in 1974, studies assessing mental illness among
homosexual populations shifted to looking at non-psychiatric populations and generally failed to
find significant group differences in rates of psychopathology, leading to waning interest in such
research (Cochran, 2001; Meyer, 2003). However, prior to the onset of the HIV/AIDS epidemic
in the 1980’s, large, population-based national studies did not typically ask about sexual
orientation, and these early non-psychiatric studies often lacked the methodological rigor of
modern epidemiological studies, used small samples, and lacked a theoretical frame (Cochran,
2001; Otis & Skinner, 1996).
Prior to the 1990s, a small number of studies assessed mental illness among LGB
populations and found inconsistent results; in most cases, studies that demonstrated slightly
elevated (but not outside normal range) levels of mental illness among LGB samples were taken

44

as evidence of no difference (Meyer, 2003). Further, in sharp contrast to pre-1950’s studies,
studies of LGB mental health conducted from the 1970’s to the 1990’s generally excluded
participants with significant histories of psychosis, mania, or hospitalization (Hellman et al.,
2002). As discussed previously in this review, more recent research from the 1990s and 2000s
indicates that LGB individuals exhibit higher rates of mental disorders as compared to
heterosexual controls (Meyer, 2003). More recent epidemiological studies are more
methodologically rigorous than earlier studies, although they are not exempt from limitations and
some methodological issues can confound interpretation of specific results (Cochran, 2001).
Surveying the state of research on sexual orientation and psychopathology, Cochran
(2001) proposed a variety of explanations for why research on mental health and wellbeing with
LGB populations is scarce. First, the vestiges of historical debates continue to color modern
disagreements on the importance and interpretation of LGB data, which may lead researchers to
be discouraged from taking professional risks by publishing in a field wrought with ongoing
controversy and potential political implications. Second, population-based research surveys have
only recently begun to incorporate data on sexual orientation in their demographic
questionnaires; prior to the onset of the HIV/AIDS epidemic, these surveys generally did not ask
about sexual orientation or behavior. Finally, research with LGB populations is subject to
significant methodological limitations that may create barriers to conducting reliable and valid
research.
Classification of Participants as LGB
The base rate of homosexual and bisexual behavior among American men and women
has been popularly estimated to be approximately 10% of the population; however, depending on
the methodology of the study prevalence rates may vary as widely as 1-15% (Savin-Williams &

45

Ream, 2007). The wide variation in prevalence rates across studies is due in part to differences in
methods of identifying individuals as homosexual or bisexual. Studies that identify individuals
on the basis of self-reported sexual attraction generally show higher prevalence rates than those
classifying individuals based on reported sexual behavior or self-identification, and this pattern
has been replicated within studies that have collected data on both attraction and behavior
(Bagley & Tremblay, 1998; Cochran, 2001; Savin-Williams & Ream, 2007; Sell, Wells, &
Wypij, 1995).
Although studies assessing same-sex sexual attraction may show lifetime rates as high as
13-17% among both American men and women, studies of lifetime same-sex sexual behavior
have suggested somewhat lower rates of behavior ranging from 2-6% of men reporting same-sex
behavior exclusively, 1-4% of women reporting same-sex behavior exclusively, and 1-2% of
men and women reporting sexual behavior with both same-sex and opposite-sex partners
(Dickson, Paul, & Herbison, 2003; Laumann, Gagnon, Michael, & Michaels, 1994; Sell, Wells,
& Wypij, 1995). Prevalence rates of reported same-sex attractions and behavior may also vary
according to social setting and, in particular, in relation to the wider social acceptability of samesex relationships (e.g., Copas et al., 2002; Fergusson, Horwood, Ridder, & Beautrais, 2005;
Kuyper & Vanwesenbeeck, 2009). As well as affecting prevalence rates, differences in methods
of classifying individuals by sexual orientation across studies can make comparison of results
across studies difficult, or even produce different results (Midanik, Drabble, Trocki, & Sell,
2007; Moradi et al., 2009; Sell, 1997; Sell & Petrulio, 1996; Worthington & Reynolds, 2009). In
addition, each method of classification is subject to methodological strengths and limitations.
The most common method of assessing sexual orientation in the psychological literature
is to ask participants to self-identify by choosing a label of gay, lesbian, or bisexual (Cochran et

46

al., 2003; Hellman et al., 2002; Otis, 2007; Worthington & Reynolds, 2009). Another method of
identifying individuals as gay, lesbian, or bisexual that is common in national population-based
research relies on assessing reported sexual behavior histories, such that individuals reporting
only same-sex sexual partners are classified as gay or lesbian, individuals reporting both samesex and opposite-sex partners are classified as bisexual, and individuals reporting only oppositesex partners are classified as heterosexual (Cochran & Mays, 2000; Gilman et al., 2001). In an
effort to assess both behavior and attraction, some studies use the Kinsey scale, a 7-point scale
that attempts to assess a range of possible patterns of sexual behavior and attraction, ranging
from 0 (exclusively homosexual) to 6 (exclusively heterosexual) with intermediate levels
representing a gradient of bisexuality (Hellman et al., 2002).
Studies that ask individuals to self-identify as heterosexual, homosexual, lesbian, or
bisexual may suffer from non-response and response biases confounded with willingness to
disclose, although the natures of the potential confounding effects are currently unknown
(Cochran et al., 2003). For example, willingness to disclose sexual orientation may correlate with
willingness to disclose psychological distress, artificially inflating rates of distress; it may reflect
social context factors such as regional or local acceptability of disclosure of emotional states
and/or personal information, including sexuality (Cochran et al., 2003; Fassinger & Miller,
1996). Conversely, reluctance to disclose sexual orientation may correlate positively with
distress, such that more highly distressed individuals are less likely to disclose their sexual
orientation status. However, the importance of response biases (and non-response biases) in
sexual orientation remains unclear (Cochran, 2001).
Another concern with the use of self-identification methods alone is that sexual
orientation labels often fail to capture the complexity of sexual desires, fantasies, attitudes,

47

beliefs, and behaviors (Moradi et al., 2009).For example, one study of 2,372 adults recruited
online via email lists and in a large Midwestern university applied cluster analysis to determine
sexual orientation groups on the basis of sexual arousal, attraction, fantasies, behavior, and
romantic relationships, and then compared these groups on measures of group commitment and
negative attitudes toward homosexuality (Worthington & Reynolds, 2009). Using Ward’s (1963;
cited in Worthington & Reynolds, 2009) method of cluster analysis, the authors found 12 distinct
sexual orientation clusters, with differences in profiles of sexual attraction, behavior, and
relationships, as well as variation in group commitment and homonegativity. These results
suggest that the exclusive use of forced-choice self-identification labels may obscure significant
interindividual variation in attraction, behavior, group identification, and attitudes.
Studies classifying individuals as gay, lesbian, or bisexual on the basis of behavioral
histories alone may misclassify individuals whose sexual behavior does not reflect constructs
relevant to the hypothesis (Cochran et al., 2003; Cochran, 2001). A study examining
discrimination on the basis of sexual orientation identity, for example, may suffer from
categorization errors using a behavioral classification method. An additional limitation of relying
on behavioral reports of sexual interest is that these reports are often time-limited to the previous
year or previous two years, which may exclude participants without partners during those time
frames who would otherwise identify as LGB. One study of major (chronic, recurrent) mental
illness in an LGB sample found that 48.8% of the women and 57.9% of the men reported no
partners in the previous year – about ten times higher than typically found in a nonpsychiatric
LGB sample (Hellman et al., 2002). These results would suggest that studies assessing sexual
orientation on the basis of past year sexual behavior may be particularly limited by the exclusion

48

of a chronically ill individuals who have had limited sexual experiences in recent years and thus
were misclassified as (presumably) heterosexual.
Some studies attempt to compensate for this possibility by asking participants to report
all sexual partners across the lifespan, but this may result in overclassifying individuals as
homosexual or bisexual, given that studies have suggested that approximately half of individuals
who report past same-sex partners self-identify as heterosexual (in contrast to the 95% of those
self-identifying as homosexual or bisexual who report same-sex past partners; Cochran & Mays,
2000). These studies may likewise exclude individuals who have not yet engaged in same-sex
sexual experiences. One study of sexual orientation identity among LGB youths and young
adults (N = 72, mean age = 20.88, SD = 2.94) found that 10% of their sample had not yet had a
same-sex sexual experience, and this was significantly related to both the age of first disclosure
of orientation and levels of immersion in LGB social networks (Floyd & Stein, 2002). Thus, the
sole use of behavioral methods for classification of LGB individuals may introduce confounds
and produce biased samples that are less psychologically distressed, more comfortable with
disclosure of sexual orientation identity, and more immersed in LGB social networks.
To overcome these inherent classification limitations, the most carefully developed
studies use a combination of methods to classify individuals as gay, lesbian, or bisexual, asking
participants to self-identify an orientation as well as collecting information on sexual behavior,
fantasy, and desire (Floyd & Stein, 2002; Hellman et al., 2002; Roberts et al., 2010). The use of
multiple methods of classification has been recommended both to control for the potential effects
of social stigma that may lead to response biases and to provide a more richly detailed and
complete picture of sexual orientation (Moradi et al., 2009).

49

When time, sample, and monetary constraints do not preclude the use of multiple
methods of identification, obtaining information on attraction, fantasy, and behavior in addition
to forced-choice self-identification items is likely to provide the most detailed picture of sexual
orientation and sexual orientation identity. Given limitations in analysis, however, it may not
always be possible to compare a large number of orientation groups or control for all potential
confounding factors. Ultimately, the choice of classification system used in data analysis seems
to depend upon the nature of the question asked as well as a consideration of what potential
confounding factors an experimenter is willing to hazard. Given evidence of rich diversity in
sexual orientation beyond existing forced-choice categories, the potential impact of social stigma
on participant responses, and the known benefits and limitations of the use of single measures of
orientation, the choice of classification system employed should be theoretically driven and
consistent with sound experimental methodology.
Sampling Methods with LGB Populations
Given the low prevalence rates of homosexuality and bisexuality, the social
stigmatization of sexual minorities, and a history of discrimination against sexual minorities on
the part of scientific and health organizations, obtaining a testable sample of LGB participants
for psychological research often necessitates the use of non-representative or non-random
sampling methods (Savin-Williams, 1994). For example, it is common for studies with LGB
individuals to use convenience sampling, recruiting participants from gay pride events, LGB
organizations, or similar environments in which researchers expect to find a high proportion of
individuals identifying as LGB (Otis & Skinner, 1996: Hellman et al., 2002). Studies examining
LGB adolescents may recruit participants via community programs or youth centers aimed
specifically at LGB youth (D’Augelli et al., 2006).

50

To maximize recruitment, researchers may rely on snowball sampling, also known as
referral sampling, a non-random sampling technique in which enrolled participants refer
acquaintances for participation in a study (Hellman et al., 2002). Given the risk of bias inherent
in non-random, non-probability survey methods, and the possibility of regional differences when
sampling from a single city or area, researchers have increasingly attempted to assess the
prevalence of mental illness in LBG populations using nationally representative surveys
(Cochran & Mays, 2000; Gilman et al., 2001). However, only recently have national surveys of
mental health included information on sexual orientation identity and same-sex behavior, some
still do not include information on sexuality, and many of these surveys are focused primarily on
youth (Dean et al., 2000)
Choices about the location of convenience sampling, or the “anchors” of chain referral
sampling methods, may limit the generalizability of results. As Otis and Skinner (1996) note, the
use of convenience sampling at gay pride events, via organizational mailing lists, and through
subsequent chain referrals tends to result in a sample that is urban, relatively well-educated, and
typically young or middle-aged. Researchers examining LGB youth by recruiting participants at
community centers and youth organizations tend to access a population that is more willing to
disclose sexual orientation and more highly distressed than LGB youth not involved in such
organizations (Savin-Williams, 1994). Sampling at bars or other social venues frequented by
LGB adults may suffer from similar sampling effects, as well (Diaz et al., 2001). Snowball
sampling may produce samples with similar characteristics as individuals recruit friends and
acquaintances that may share environments and behaviors. The use of subsamples from national
data sets is limited by inconsistency across national studies in how sexual orientation is defined
and assessed, and researchers may be limited to only one aspect of sexual orientation such as

51

past sexual behavior (e.g., Cochran & Mays, 2000). Decisions about sampling methods should be
made in consideration of the research question, practical limitations, and potential confounding
effects of the recruitment method chosen (Moradi et al., 2009).
Assessment and Analysis of Data with LGB Populations
As with all psychological research studies conducted in laboratory settings, failing to
replicate real-world experiences may limit the generalizability of results (Moradi et al., 2009).
The method of administration of measures may affect responses. For example, Villarroel et al.
(2006) found that individuals were more willing to reveal same-sex attraction and behaviors
when interviews were conducted by computer rather than in-person, which may reflect concerns
related to stigma and social exclusion (Fassinger & Miller, 1996; Moradi et al., 2009). Given that
willingness to disclose sexual orientation may be correlated with other psychological variables,
this research indicates that providing participants with the perceived anonymity of responding
via computer may widen the pool of participants who identify as non-heterosexual and produce
more diverse and accurate samples (Villarroel et al., 2006). Researchers should be careful to
avoid using biased language when writing items so as to solicit the most accurate information
possible (Moradi et al., 2009). Presuming sample and data accuracy, however, considerations
around analysis of collected data remain.
Sexual orientation subgroups are often combined into broader categories (e.g., gay men
and lesbian women combined into a single ‘homosexual’ group), which may increase statistical
power but likewise obscure intragroup variation (Moradi et al., 2009; Worthington & Reynolds,
2009). It may be the case that some experiences are shared across sexual minority subgroups, but
it is unclear which experiences are likely to vary and there is evidence that subgroup differences
exist, including subgroup differences in patterns of psychopathology comorbidity (Cochran,

52

2001). As with decisions regarding sampling and classification, reviews of methodological issues
with LGB populations recommend weighing the risks and benefits of different methods of task
administration and data analysis according to the research questions and limitations (Moradi et
al., 2009). Given the significant and varied methodological issues involved in conducting
research with sexual minorities, it is important to choose methods of measurement that minimize
the risk of confounding factors while providing adequate means to test the research questions at
hand.
The present study collected a random sample of young adults from an undergraduate
subject pool, oversampling for LGB participants. Data was collected on sexual orientation selfcategorization and behavior, as well as on a continuous measure of attraction, fantasy, and
behavior (the Kinsey scale). For the purposes of statistical analyses, the Kinsey scale was chosen
as a primary grouping method due to its dimensional nature, established conventions for
grouping, and opportunity for participants to identify their sexual orientation based on their life
history of attraction, fantasy, and behavior (rather than self-identification or behavior alone).
Sexual orientation information was collected by computer and participant data was collected
anonymously in order to maximize participant disclosure. Finally, study staff were blind to group
status in order to minimize expectancy effects.

53

CHAPTER SIX
Rationale and Hypothesis
The current study is designed to examine three research questions. Do bias crimes serve
as a unique source of minority stress for LGB individuals that are experienced differently from
heterosexual individuals? Is exposure to bias crime information a potential source of
traumatogenic stress for LGB individuals, insofar as it evokes defensive threat responses and
leads to changes in cognitive worldview? Finally, do individuals vary in their physiological,
cognitive, and emotional responses to bias crimes as a function of their collective identity?
To date, two qualitative studies (Noelle, 2002; Perry & Alvi, 2012) and one quantitative
study (Craig, 1999) have examined responses of non-victim members of a targeted group to
racial (Craig, 1999), ethnic (Perry & Alvi, 2012), and sexual orientation bias crimes (Noelle,
2002), and have suggested that non-victim members of a targeted group may have threat-related
emotional, cognitive, and behavioral responses to bias crimes. However, these studies have relied
primarily on the use of qualitative and correlational designs, small sample sizes, nonrandom
sampling, and retrospective reporting of responses. The aim of the present dissertation study was
to build upon this work by assessing the questions posed above within a strong theoretical
framework and using sound quantitative and quasi-experimental methodology.
Minority stress models describe how distal events such as sexual minority bias crimes
may influence individual outcomes by gaining meaning in proximal processes such as perception
and appraisal (Meyer, 2003). Minority stress further defines the threat of violence against
minority individuals as a source of chronic and inescapable social stress, the effects of which are
moderated by individual factors such as identity (Cochran, 2001). The sexual harassment and
racial discrimination literature suggests that minority individuals are exposed to daily and life

54

stressors as a function of discrimination that may reach the level of traumatic threat (Avina &
O’Donahue, 2002; Bryant-Davis, 2007; Stockton et al., 2009; McDermut et al., 2000; Pieterse et
al., 2010; Wolfe et al., 2009; Woods et al., 2009), and that stress responses can occur through
indirect exposure such as working or living in an environment in which sexual harassment and/or
discrimination occur (DeSantis et al, 2007; Lim & Cortina, 2005; Miner-Rubino & Cortina,
2004; Miner-Rubino & Cortina, 2007; Raver & Gelfand, 2005).
Models of collective identity are based in social identity and self-categorization theories,
which support the hypothesis that threat to the community, interpreted as threat to self, may
affect immediate responses that would be marked by changes in physiological activity (Smith &
Silva, 2011; Tajfel, 1978; Turner, 1986). Individual aspects of identity may function to
alternately buffer or potentiate these reactions, and may have proximal (immediate physical
response) and distal (later development of pathology) effects via mechanisms of perception and
appraisal (Sellers, 1998). To the extent that individuals perceive identity as central and believe
public regard is low, they are theoretically more likely to both perceive a threat to a group
member on the basis of identity as a threat to the self and to appraise the likelihood of this threat
as high. The nature of bias crimes (as suggested by the subjects of the Noelle [2002] study) is
unpredictable and uncontrollable, which may exacerbate the stress response (Averill, 1973; Foa
et al., 1992). This is likely to activate autonomic defense cascade responses in preparation for
coping with perceived threat, which increases fear learning and fear acquisition via a variety of
nervous system and hormonal mechanisms (Hamm & Weike, 2005; Pole, 2007).
The study assesses proximal physiological, emotional, and cognitive responses during
and immediately following exposure to bias crime information, using a quasi-experimental,
mixed-factorial design with ecologically-valid stimuli. This study recruited heterosexual and

55

LGB individuals - identified using a continuous scale of same-sex attraction, fantasy, and
behavior - from a young-adult population of college students. Participants completed pre-task
questionnaires online assessing sexual orientation and assumptive worldview, then completed an
in-lab quasi-experimental task in which they listened to randomized bias-crime, non-bias crime,
and neutral stimuli in the form of auditory news reports while skin conductance level and startle
reflex potentiation were simultaneously and continuously recorded. Following the task, provided
reports of their cognitive and emotional responses to the clips, collective identity, post-task
measures of traumatogenic cognition, and supplemental variables of interest (e.g., past personal
bias victimization).
Specific Hypotheses
The following hypotheses are guided by three primary research questions:
1. Do LGB-identified individuals respond to exposure to bias crimes differently from
heterosexual-identified individuals?
a. LGB individuals show a greater magnitude of ANS activation and greater startle
response to bias-crime exposure than heterosexual individuals.
b. LGB individuals will report (subjectively) more negative affective responses and
greater arousal during exposure to bias-crime clips than will heterosexual
individuals.
c. LGB individuals will report more negative cognitions about the safety,
predictability, benevolence, and controllability of the world and of others
following exposure to news information about bias crimes following exposure to
bias crimes than will heterosexual individuals.

56

2. Do LGB-identified individuals respond to exposure to bias crimes as a source of potential
threat to the self?
a. LGB individuals will show a greater magnitude of ANS activation and greater
startle response to bias-crime clips than to general threat or neutral clips.
b. LGB individuals report more negative affective responses and greater arousal
during exposure to bias-crime clips than during neutral or general threat clips.
c. LGB individuals report more negative cognitions about the safety, predictability,
benevolence, and controllability of the world and of others following exposure to
news information about bias crimes than they reported prior to exposure.
3. Are responses to bias crimes related to domains of collective identity for LGB individuals
and heterosexual individuals?
a. Greater reported identity centrality predicts greater ANS activation and startle
responses in LGB individuals during exposure to bias crimes.
b. Lower reported public regard predicts greater ANS activation and startle
responses in LGB individuals during exposure to bias crimes.
c.

The interaction of high identity centrality and low public regard predicts greater
ANS activation and startle magnitude in LGB individuals during exposure to bias
crimes.

d.

Sexual orientation identity does not predict ANS activation or startle magnitude
for heterosexual individuals upon exposure to bias crimes.

57

CHAPTER SEVEN
Methods
Design
The study employed a mixed model repeated measures design. Two groups of
participants (LGB and heterosexual) completed pre-task questionnaires and then completed a lab
task between one and two weeks later. In the lab task, participants listened to audio only news
clips belonging to three conditions (neutral, general threat, and bias threat) while physiological
responses were continuously recorded. In the neutral (non-threat) condition, news content
consisted of non-threatening information such as reports of a manufacturing fair or a local art
festival. In the general threat condition, news content consisted of information regarding a
violent crime (an assault, murder, or arson) that was not identified by the newscaster as a biasmotivated crime. In the bias threat condition, news content consisted of information regarding a
violent crime (an assault, murder, or arson) that was identified by the newscaster as potentially
bias-motivated. All participants listened to all clips and completed questionnaire measures
following completion of the listening task.
Participants
Participants were recruited in the fall of 2012 and spring, summer, and fall of 2013 using
the Internet-based Human Participation in Research (HPR) system at a large, Midwestern
university. Individuals enrolled in the HPR system were primarily students enrolled in lowerlevel courses in the Psychology Department and received course credit for their participation.
Participants were also entered into a raffle in which an iPod Shuffle (cash value: $49.99) was
granted to a randomly selected participant each semester using a randomized computer
algorithm. The following inclusion and exclusion criteria were used to identify participants: all

58

participants were at least 18 years of age, had no significant impairments in hearing or vision,
and were reasonably fluent in reading and writing English. Participants meeting these criteria
were eligible to complete a global screening questionnaire via the HPR system, which was used
to determine enrollment for the lab study. A total of 2,352 students (31.2% male, 68.6% female,
<1% other) completed the global screening questionnaire.
The global screening questionnaire assessed sexual orientation on the basis of three
items. The first asked individuals to choose a sexual orientation label from a set of presented
categories (gay, lesbian, bisexual, heterosexual, queer, asexual, or other). The second item asked
individuals to describe the pattern of gender of their past sexual partners (“All men,” “Mostly
men, some women,” “Equal parts men and women,” “Mostly women, some men,” or “All
women”). The third item asked individuals to describe their sexual orientation on a modified
Kinsey scale, which asked individuals to rate their sexual orientation on a scale of 0
(“Exclusively heterosexual”) to 6 (“Exclusively homosexual”) on the basis of their total lifetime
experiences of sexual attraction, fantasy, and behavior. The scale was modified to add an
additional item, “Nonsexual/Asexual“, to allow individuals to identify as asexual. Individuals
who rated their sexual orientation at or above a 2 on the modified Kinsey scale (“More than
incidentally homosexual”) were classified as LGB. Individuals who rated their sexual orientation
as a 0 (“Exclusively heterosexual”) or 1 (“Only incidentally homosexual”) were classified as
heterosexual. Any individual identifying as LGB on the global screener, and a randomly-selected
subset of individuals identifying as heterosexual, was made eligible for participation in the full
study. Of the 2,352 individuals who completed the global screener, 8.3% (n = 204) were
classified as LGB, 91.1% (n = 2,133) were classified as heterosexual, and 0.6% (n = 15)
described themselves as “Nonsexual/Asexual.”

59

A total of 65 participants (33 heterosexual, 32 LGB; 61.5% female) attended the
laboratory session. Of those who attended, 4 (6.2%) experienced technical difficulties that led to
loss of more than 10% of data collected and so were excluded from the final sample. One
participant failed a validity check of attention to auditory clips and was also excluded from the
sample. An additional 5 participants (7.7%) experienced technical difficulties that resulted in <
10% questionnaire data loss, specifically the loss of a final demographic questionnaire which
provided information on family income, ethnicity, region of upbringing, and other sample
characteristics. In addition, 3 participants experienced technical errors that led to the recording of
startle reflex EMG data that was not analyzable, and corruption of a data file led to the loss of
SAM ratings for one participant. As participants with < 10% data loss retained the majority of
data on other variables of interest, they were retained within the final sample. The final sample
consisted of 60 participants (32 heterosexual, 28 LGB) primarily raised in the Midwest area of
the United States and from a diverse range of ethnic, racial, and socioeconomic backgrounds. A
complete summary of available sample demographic characteristics can be found in Table 1. A
summary of sample sexual orientation characteristics can be found in Table 2.
Table 1.
Sample Demographic Characteristics, Full Sample and in Comparison by Kinsey Group
Full Sample

Heterosexual

LGB

N = 60

N = 32

N = 28

N(%)

N(%)

N(%)

χ2 or t (p)

d/ϕ/V

20.03 (1.87)

20.06 (1.80)

20.00 (1.93)

0.13 (.90)

.03

Male

22 (36.7%)

14 (43.8%)

8 (28.6%)

Female

38 (63.3%)

18 (56.3%)

20 (71.4%)

10.38 (.24)

.42

Age, M(SD)
Sex

60

Table 1 (cont’d)
Race
Asian/AAPI

9 (16.0%)

8 (25.0%)

1 (3.6%)

Black/Af. Am.

6 (10.0%)

3 (9.4%)

3 (10.7%)

Latino/Chicano

4 (6.7%)

0 (0.0%)

4 (14.3%)

33 (55.0%)

16 (50.0%)

17 (60.7%)

3 (5.0%)

2 (6.3%)

1 (3.6%)

Hispanic

6 (10.0%)

1 (3.1%)

5 (17.9%)

Not Hispanic

49 (81.7%)

28 (87.5%)

21 (75.0%)

5 (8.3%)

2 (6.3%)

3 (10.7%)

Small town

12 (20.0%)

6 (18.8%)

6 (21.4%)

Suburban

21 (33.3%)

10 (34.5%)

10 (38.5%)

Small city

7 (11.7%)

3 (10.3%)

4 (14.3%)

City

10 (16.7%)

7 (24.1%)

3 (10.7%)

Other

1 (1.7%)

1 (3.4%)

0 (0.0%)

New England

2 (3.3%)

2 (6.3%)

0 (0.0%)

Mid-Atlantic

2 (3.3%)

1 (3.1%)

1 (3.6%)

44 (73.3%)

22 (68.8%)

22 (78.5%)

Pacific West

1 (1.7%)

0 (0.0%)

1 (3.6%)

Outside the U.S.

6 (10.0%)

4 (12.5%)

2 (7.1%)

Less than 20k

3 (5.0%)

2 (6.3%)

1 (3.6%)

20k-39.9k

6 (10.0%)

1 (3.1%)

5 (17.9%)

40k-59.9k

7 (11.7%)

4 (12.5%)

3 (10.7%)

60k-79.9k

13 (21.7%)

7 (21.9%)

6 (21.4%)

80k-99.9k

9 (15.0%)

5 (15.6%)

4 (14.3%)

100k-119.9k

6 (10.0%)

3 (9.4%)

3 (10.7%)

More than 120k

11 (18.3%)

7 (21.9%)

4 (14.3%)

White/Caucasian
Biracial/Multiracial

9.67 (.05)

.42

3.51 (.06)

.25

2.79 (.73)

.23

5.62 (.35)

.32

3.99 (.68)

.27

Ethnicity

Area of Upbringing
Rural

Region of Upbringing

Midwest

Family Income

61

Table 1 (cont’d)
Education
Diploma or GED

52 (86.7%)

28 (87.5%)

24 (85.7%)

Associate’s

1 (1.7%)

0 (0.0%)

1 (3.6%)

Bachelor’s

2 (3.3%)

1 (3.1%)

1 (3.6%)

5 (8.3%)

4 (12.5%)

1 (3.6%)

17 (28.3%)

9 (28.1%)

8 (28.6%)

Evangelical

2 (3.3%)

2 (6.3%)

0 (0.0%)

Jewish

1 (1.7%)

0 (0.0%)

1 (3.6%)

Muslim

1 (1.7%)

1 (3.4%)

0 (0.0%)

Atheist/Agnostic

16 (26.7%)

5 (15.6%)

11 (39.3%)

Other

13 (21.7%)

8 (25.0%)

3.51 (2.08)

4.03 (2.06)

1.15 (.56)

.14

5 (17.9%)

8.66 (.19)

.40

2.92 (1.98)

2.04 (.05)

.56

Religious Identity
Protestant
Roman Catholic

Religiosity, M(SD)

Note: Percentages do not add to 100 for some variables due to missing data (N = 5).
Table 2.
Sample Sexual Orientation Characteristics, Full Sample and by Kinsey Group (n = 60)
Full Sample
(n = 60)

Heterosexual
(n = 32)

LGB
(n = 28)

N(%)

N(%)

N(%)

0

29 (48.3%)

29 (90.6%)

0 (0.0%)

1

3 (5.0%)

3 (9.4%)

0 (0.0%)

2

14 (23.3%)

0 (0.0%)

14 (50.0%)

3

4 (6.7%)

0 (0.0%)

4 (14.3%)

4

3 (5.0%)

0 (0.0%)

3 (10.7%)

5

4 (6.7%)

0 (0.0%)

4 (14.3%)

6

3 (5.0%)

0 (0.0%)

3 (10.7%)

Lesbian

2 (3.3%)

0 (0.0%)

2 (7.1%)

Kinsey Scale

Self-Identification

62

Table 2 (cont’d)
Gay

6 (10.0%)

0 (0.0%)

6 (21.4%)

Bisexual

8 (13.3%)

1 (3.1%)

7 (25.0%)

Heterosexual

41 (68.3%)

31 (96.9%)

10 (35.7%)

3 (5.0%)

0 (0.0%)

3 (10.7%)

All men

18 (30.0%)

15 (46.9%)

3 (10.7%)

Mostly men, some women

18 (30.0%)

2 (6.3%)

16 (57.1%)

Equal numbers men and women

2 (3.3%)

0 (0.0%)

2 (7.1%)

Mostly women, some men

3 (5.0%)

0 (0.0%)

3 (10.7%)

All women

14 (23.4%)

14 (43.8%)

0 (0.0%)

No partners

5 (8.3%)

1 (3.1%)

4 (14.3%)

Other
Sexual Behavior

Materials
Auditory stimuli.
Auditory stimuli were 11 news clips gathered from the Internet and consisting of news
reports presented by local news stations in a variety of municipalities across the United States
and Canada. All news clips were presented in English and between 1m27s and 1m33s in length.
All clips in all conditions followed the following format: first, a presenter announced the topic
(e.g., “A crime was reported in the X district last night, and police are investigating whether or
not the event was a hate crime.”). Then, a reporter on scene conducted interviews with relevant
individuals (e.g., an identified victim, a witness, an attendee at a reported event). Finally, the
presenter summarized the news story and provided any additional information (e.g., “Police are
on the lookout for an individual believed to be wearing…”). All clips were publicly available via
local news websites and video indexing sites such as YouTube. During presentation of each clip
participants viewed a black screen, so only auditory information was provided.

63

News clips were identified for inclusion in the study via the following process: first, a
search was conducted to identify local news reports using two popular Internet index search
engines (Google and Bing) and a popular video indexing website (YouTube) using keywords
such as “bias crime,” “hate crime,” and “LGBT crime.” Approximately 30 potential clips were
initially identified based on meeting the topic (possible LGBT-related event) and length criteria
(approximately 1m30s ± 5s). Clips were then analyzed for presentation and content. Clips were
removed from the list if they a) did not follow the presentation format described above; b) did
not state explicitly that the event was potentially a bias-motivated crime; c) did not provide a
description of the event itself; d) contained excessive (defined as >5s duration) music; e)
contained music that was not presented in the initial 10s of the clip (as a “lead-in” to the news
report); f) explicitly stated perpetrator or victim race; or g) presented any difficulties re:
intelligibility (e.g., in one clip subtitles were used due to speaking impairments as a result of an
injury; this clip was excluded on this basis). Using these criteria, the list was narrowed to 7
potential clips. The final three clips were then selected to provide a range of events (assault,
murder, and arson), victim characteristics (male, female couple, LGBT-owned business), and
presenter and reporter gender (multiple presenters, at least one male and one female in each).
The selection of general threat clips was then conducted with the aim of matching the
salient event characteristics of the bias crime clips while also satisfying the same inclusion
criteria. An identical search process was conducted, using event type (“assault,” “murder,” or
“arson”) as the primary search terms. Clips were identified on the basis of matching
characteristics and length. Clips were rejected for inclusion if they a) did not follow the
presentation format given above; b) stated a potential bias motivation for the event (i.e., the
events described must be presented as not occurring due to any bias based on victim

64

characteristics); c) did not provide a description of the event itself; d) contained excessive
(defined as >5s duration) music; e) contained music that was not presented in the initial 10s of
the clip (as a “lead-in” to the news report); f) explicitly stated perpetrator or victim race; or g)
presented any difficulties re: intelligibility. The final three clips were selected to provide a
matched range of events (assault, murder, and arson), victim characteristics (male, male/female
couple, non-LGBT business), and presenter/reporter gender.
Finally, neutral clips were selected in similar fashion, with the aim of providing neutral
(non-emotive) information to compare the two threat conditions against a non-threatening
condition. Potential clips were identified based on length (1m30s ± 5s) and were rejected for
inclusion if they a) did not follow the presentation format given above; b) described any
potentially threatening or emotionally-evocative event (e.g., a crime, a high-intensity event such
as skydiving or rollercoasters, a bereavement or memorial); c) contained excessive (defined as
>5s duration) music; e) contained music that was not presented in the initial 10s of the clip (as a
“lead-in” to the news report); f) described any specific characteristics of involved individuals
(e.g., race, gender); g) presented any difficulties re: intelligibility. Five clips were selected for
inclusion as neutral stimuli and consisted of a range of neutral events (a manufacturing job fair,
changes in fall leaf colors, a local art festival, school board changes to class schedules, and a
drought’s effects on corn crops) and presenter/reporter gender.
Stimuli in all conditions were modified to remove visual content and normalize audio
content volume using widely-distributed freeware programs, AviSynth ver. 2.5 and Avidemux
ver. 2.6. All visual content was replaced by a blank (black) screen. Clips were further run
through volume analysis algorithms within these programs in order to assess the decibel range of
presented information, and volume was normalized such that all auditory data was presented

65

within the 55dB-65dB range when played using volume settings that were standardized across all
participants. A decibel meter was used to determine matching volume settings on the local task
computer such that clip volume was consistently within this range.
Auditory stimuli were presented in one of six potential orderings, determined based on
the generation of approximately 50 random orderings of the 11 clips using a computerized
random list generator. Six orderings were selected from this larger list based on meeting the
constraint that no bias threat clip should be presented immediately following a general threat
clip, and vice-versa (in order to further protect against any potential carry-over effects of
physiological arousal).
Psychophysiological data recording.
Continuous psychophysiological data were recorded using the Biosemi Active2 system
(Biosemi, The Netherlands). EMG recordings were taken from two Ag-AgCl electrodes
positioned approximately 25mm apart on the orbicularis oculi muscle below the left eye. Skin
conductance activity was recorded from two Ag-AgCl electrodes affixed to the distal phalanges
of the index and middle fingers of the left hand. Skin conductance level (SCL) sensors applied a
1V, 50Hz alternating voltage and measuring the returning current. During data acquisition, the
Common Mode Sense (CMS) active electrode and Driven Right Leg (DRL) passive electrode
formed a ground loop, consistent with Biosemi's design specifications. The function of the CMSDRL loop, in addition to serving as a reference, is to constrain the common mode voltage (i.e.
the average voltage of the participant), limiting the amount of current that can possibly return to
the participant. SCL signals were converted to an analog voltage of .2V/µS. All signals were
digitized at 24-bit resolution and a sampling rate of 1024 Hz.
Questionnaire Measures

66

Demographics. A demographic questionnaire was administered post-lab task to assess
age, sex, gender identification, city/suburban/rural upbringing, region of upbringing, education,
socioeconomic status, religious identification, religiosity and ethnicity. Appendix A contains
copies of all measures.
Sexual orientation. Sexual orientation information was collected in the Global Screening
Questionnaire using a multi-method approach. One item asked participants to self-select a sexual
orientation from a list of categories, including “lesbian,” “gay,” “bisexual,” “heterosexual,”
“queer,” “asexual,” and “other.” Participants selecting “other” were provided a free-response box
to describe their preferred sexual identity descriptor. A second item asked participants to report
on their sexual behavior history (kissing, fondling, intercourse) with regard to the gender of their
partners on a five point scale including “All men,” “Mostly men, some women,” “Equal numbers
of men and women,” “Mostly women, some men,” “All women,” and “I have not had any sexual
experiences.” Finally, a third item asked participants to report on their sexual orientation and
history of attraction, fantasy, and behavior using a version of the Kinsey scale. Participants were
asked, “Considering your history of attraction, fantasy, and behavior across your entire life, how
would you describe your sexual orientation?” Participants chose a response on a 7 point Likerttype scale, from 0 (Exclusively heterosexual) to 6 (Exclusively homosexual). An additional
response option not included in the original Kinsey scale, 7 (Non-sexual/Asexual) was also
provided to allow participants to identify as asexual. For the purposes of the present study, the
Kinsey scale variable was used to determine group categorization, with individuals reporting
values of 2 (Mostly heterosexual, but more than incidentally homosexual) and above classified as
LGB.

67

Sexual orientation collective identity. A modified version of the Multidimensional
Inventory of Black Identity (MIBI) was administered post-lab task to assess dimensions of
sexual orientation collective identity. The MIBI has been modified for use in a wide range of
minority populations and has demonstrated good psychometric properties across populations (for
a full review, see Chapter Four, above). The modified version of the MIBI used in this study
consisted of 20 statements assessing collective identity centrality, public regard, and private
regard. An example statement is, “I have a strong sense of belonging to other people who share
my sexual orientation.” Participants indicated their agreement with each statement using a 7
point Likert-type scale ranging from 1 (Strongly Disagree) to 7 (Strongly Agree). Scale scores
were produced for identity centrality, public regard, and private regard by averaging the items
within each scale and ranged from 1 to 5. In the present study, reliability of the identity centrality
subscale (Cronbach’s α) was good for the full sample (α = 0.71), acceptable for the heterosexual
subsample (α = 0.63), and good for the LGB subsample (α = 0.79). The reliability of the public
regard subscale was excellent for the full sample (α = 0.91), good for the heterosexual subsample
(α = 0.79), and good for the LGB subsample (α = 0.89). The reliability of the private regard
subscale was acceptable for the full sample (α = 0.67), good for the heterosexual subsample (α =
.70), and acceptable for the LGB subsample (α = 0.65).
Subjective ratings of affect during task presentation. Following presentation of each
trial in the lab task (stimulus plus intertrial interval), participants completed three computer items
rating the magnitude of arousal (high vs. low), valence of affect (positive/pleasant vs.
negative/unpleasant), and subjective perceptions of controllability/dominance (in
control/dominant vs. not in control/non-dominant) and warmth (interpersonal closeness) using a
modified version of the Self-Assessment Manikin (SAM) system developed by Peter Lang

68

(1980; Lang, Bradley, & Cuthbert, 1997). The standard SAM system uses sets of 9 graphic
figures arranged along a 9-point scale representing a continuum of arousal, valence, and
dominance from low to high. An additional scale using the same 9-point scale of reference but
without graphic representations was added to assess interpersonal warmth (the warm/cold scale),
and was presented immediately following presentation of the three SAM figures. Figure 7 in
Appendix B provides examples of these items. Participants were instructed to select the number
in each scale that best represented their responses while listening to the auditory clips. The SAM
system has been widely used as a quick and reliable method of obtaining subjective ratings of
affective responses during exposure to images, sounds, film clips, and painful stimuli (Bradley &
Lang, 1994). Internal (split-half) reliability coefficients for the arousal and valence scales have
been as high as .93 and .94, respectively, in previous studies (Lang et al., 1997).
Psychophysiological studies have demonstrated correlations between SAM valence ratings and
startle magnitude of r = -.53 (Bradley & Lang, 2000), and the correlation between SAM arousal
ratings and electrodermal activity has also been shown to be high (r = .81; Lang, 1995).
Subjective ratings of arousal, valence, dominance, and warmth were computed for each
category of stimuli by averaging SAM responses across all clips within each category (e.g.,
average of all ratings of arousal for bias crime clips). In the present sample, reliabilities of
arousal ratings were good for the full sample (α = 0.84), heterosexual subsample (α = 0.87), and
LGB subsample (α = 0.77). Reliabilities of valence ratings were also good for the full sample (α
= 0.80), heterosexual subsample (α = 0.81), and LGB subsample (α = 0.75). Reliabilities of
dominance ratings were good for the full sample (α = 0.88), excellent for the heterosexual
subsample (α = 0.91), and good for the LGB subsample (α = 0.74). Reliabilities of warmth

69

ratings were good for the full sample (α = .71), acceptable for the heterosexual subsample (α =
0.67), and good for the LGB subsample (α = 0.73).
Assumptive worldview. The World Assumptions Questionnaire (WAQ; Janoff-Bulman,
1989; Kaler, 2009) was administered pre- and post-lab task to assess pre- and post-task beliefs
about the world. The WAQ is an alternative measure of assumptive worldview from the World
Assumptions Scale (WAS; Janoff-Bulman, 1992) based on a psychometric analysis of the
limitations of the WAS (Kaler, 2009). The WAQ consists of 22 statements rated on a 6-point
Likert-type scale ranging from 1 (Strongly Disagree) to 6 (Strongly Agree). Factor analysis in
the Kaler (2009) study indicated four subscales: Controllability of Events, Comprehensibility and
Predictability of People, Truthworthiness and Goodness of People, and Safety and Vulnerability.
The WAQ subscales showed improved internal reliability (α = .74 to .82) over WAS subscales.
The WAQ is also moderately related to similar trauma measures such as the Posttraumatic
Cognitions Inventory (PTCI; Foa et al., 1999) Negative Cognitions subscale (r = .46) and
measures of general distress (but not related to measures of social desirability).
In the present study, subscale scores were computed by averaging items within each
subscale to produce four scale scores according to the factors identified by Kaler (2009). Scales
were scored such that higher scores reflected more positive world beliefs (e.g., greater belief in
the controllability of the world). Scale scores were computed separately for pre-task and posttask responses, for a total of 8 scales. In the current study, reliabilities (Cronbach’s alpha) of the
pre- and post-task subscales ranged from unacceptable (post-task Safety, heterosexual sample α
= .32) to good (pre-task Controllability and Predictability of People, LGB sample α = .84).
Unacceptable reliability estimates were observed within the heterosexual sample on two pre-task
scales (Controllability α = .57; Safety α = .49) and one post-task scale (Safety α = .32). Given

70

this, closer examination of this scale’s psychometric properties was conducted to assess whether
the factor structure reported by Kaler was replicated within the present sample, and is
summarized under Results, Preliminary Analyses, below.
Negative cognitions about the world. The Negative Cognitions about the World
subscale of the PTCI was administered pre- and post-task to provide an additional measure of
changes in cognitions related to exposure to information about bias-motivated crimes. The
Negative Beliefs about the World subscale consists of 7 statements rated on a Likert-type scale
ranging from 1 (Totally Disagree) to 7 (Totally Agree), such as, “People can’t be trusted.” Two
overall scale scores were computed by averaging responses across the 7 items at pre-task and
post-task. The reliability of the pre-task scale was good for the full sample (α = 0.73), acceptable
for the heterosexual subsample (α = 0.68), and good for the LGB subsample (α = 0.75). The
reliability of the post-task scale was good for the full sample (α = 0.80), the heterosexual
subsample (α = 0.79), and the LGB subsample (α = 0.80).
Past experiences of bias victimization. Past experiences of bias victimization were
assessed post-lab task using a slightly-modified version of a brief, 7-item inventory developed by
D’Augelli and colleagues (1992; with author permission) that has been used in studies with both
LGB youth (D’Augelli, Pilkington, & Hershberger, 2002) and LGB adults (Herek et al., 1997).
Seven types of victimization were assessed, including verbal victimization (verbal insults, threats
of violence, and threats to disclose the individual’s orientation) and physical victimization
(objects being thrown, physical assault, threat of assault with a weapon, and sexual assault).
Participants were asked to estimate the numbers of times they had experienced such events in
their lifetime. A sample item is, “How many times have you been threatened with physical
violence because you were, or were thought to be, lesbian, gay, or bisexual?” Mean frequency

71

scores may be computed for verbal victimization experiences, physical victimization
experiences, and total victimization experiences. An additional item (“How many times has this
happened within the past month?”) was added to the original questionnaire to assess past-month
frequency of victimization. Total estimated past victimization events ranged from 0 to 101. As a
continuous sum score representing number of reported events was found to be highly skewed
(skewness = 3.04, SE = .32) due to a small subsample of participants reporting very high rates of
victimization, a binary variable was created to represent the presence (“1”) or absence (“0”) of
reported past victimization for the present analyses.
Tendency toward negative affect/neuroticism. The neuroticism scale of the
International Personality Item Pool (IPIP-NEO; Goldberg, 1999; Goldberg, 2006) was
administered post-lab task to assess participant tendency toward negative affect/neuroticism. The
IPIP Neuroticism scale consists of 20 statements (5 reverse-scored) such as, “I get stressed out
easily.” Participants rated the extent to which each statement described them on a 5-point Likert
type scale ranging from 1 (Very Inaccurate) to 5 (Very Accurate). The IPIP Neuroticism scale
has demonstrated both high internal reliability (α = .91) and strong convergent validity (r = .86)
with the Revised NEO Personality Inventory (NEO-PI-R) Neuroticism scale. A single scale
score for neuroticism was computed by averaging all 20 items after reverse-scoring 10 items. In
the present sample, the internal reliability (Cronbach’s α) of the IPIP Neuroticism scale was good
for the full sample (α = 0.89), the heterosexual subsample (α = 0.89), and the LGB subsample (α
= 0.89).
Life Events Checklist. The Life Events Checklist (LEC) was administered post-lab task
to participants to assess for exposure to other potentially traumatic prior life events. The LEC
was developed at the National Center for PTSD as a method of assessing for potentially

72

traumatic life events, and consists of 17 items assessing a range of events (example: “Natural
disaster (for example, flood, hurricane, tornado, earthquake”). Participants indicate whether each
event happened to them, was witnessed by them, occurred to someone close to them, did not
apply to them, or if they were unsure/did not know whether the event applied to them. The LEC
has been shown to be correlated with both general psychological distress and PTSD symptoms in
clinical samples (Gray, Litz, Hsu, & Lombardo, 2004). In the present study, events were coded
dichotomously for participant exposure such that events directly experienced, witnessed, or
learned about after occurring to someone close were coded as “1” and events coded as “not sure”
or “doesn’t apply” were coded as “0”. Events were then summed to produce an overall life
events score (range: 0 – 17).
Other post-traumatic cognitions (full scale). The full Post-Traumatic Cognitions
Inventory (PTCI; Foa et al., 1999) was administered post-lab task to assess for the presence of
post-traumatic cognitions related to any prior stressful life experience. Participants were
instructed to complete the scale in reference to any stressful life experience endorsed on the LEC
that continued to bother them at the time of the lab assessment. The PTCI consists of 33
statements rated on a Likert-type scale ranging from 1 (Totally Disagree) to 7 (Totally Agree),
such as, “My reactions since the event mean that I am going crazy.” Principal-components
analyses of the PTCI have yielded three primary factors: Negative Cognitions about the Self,
Negative Cognitions about the World, and Self Blame. In initial studies of PTCI reliability and
validity, internal consistencies for the three scales ranged from 0.88 to 0.97. The PTCI has also
demonstrated adequate convergent, discriminant, and predictive validity in follow-up studies
(Beck et al., 2004). Participants were instructed to complete this inventory in reference to any
event that occurred in the past and continues to bother them. Scale scores were computed for the

73

three subscales by averaging items within each scale. In the present sample, the internal
reliability of the Negative Cognitions about the Self subscale was excellent in the full sample (α
= 0.94), good in the heterosexual subsample (α = 0.89), and excellent in the LGB subsample (α =
0.96). The internal reliability of the Negative Cognitions about the World subscale was good in
the full sample (α = 0.86), heterosexual subsample (α = 0.81), and LGB subsample (α = 0.89).
The internal reliability of the Self Blame subscale was good in the full sample (α = 0.78),
heterosexual subsample (α = 0.73), and LGB subsample (0.82).
Attention to task. A series of questions regarding the factual content of the auditory
news clips was created in order to provide a rough measure of participant attention to task. Two
items were created for each clip, for a total of 22 accuracy items. An example item for a general
threat clip (“Couple Shot”) is, “Where did this event occur?” with possible responses of “a park,”
“a street,” “a home” (correct response), or “a store.” Response frequencies were examined to
determine the level of item difficulty across the sample. Six items were answered incorrectly by
more than 25% of the sample and so were judged to be poorly discriminating. The remaining 16
items were coded dichotomously for accuracy of response (1 = correct, 0 = incorrect) and
summed to produce an overall accuracy score. Scores of 12 or above were judged to represent
acceptable attention to clip content. Only one participant among those who completed the lab
task with less than 10% data loss due to technical error failed this validity check (Score = 7) and
was not included in the final sample.
Procedures
Identification and enrollment of participants.
Participants meeting basic inclusion criteria were eligible to complete a global screening
questionnaire through the MSU Human Participation in Research (HPR) system consisting of

74

three items assessing sexual orientation, a pre-task World Assumptions Questionnaire, and the
Negative Beliefs about the World subscale of the PTCI. Prior to completing this questionnaire,
participants were provided with information about the purpose of the questionnaire, possible
risks and benefits, and the possibility that they may be invited to complete a lab task following
completion of the screening questionnaire. Participants indicated their consent by clicking a
button labeled “I Consent” and were then presented with screening items. Participants who
identified as LGB on the basis of the modified Kinsey scale (2 and above = LGB), and a
randomly selected subsample of participants who identified as heterosexual (0, 1), were invited
to participate in an in-lab task. Individuals who elected to enroll in the study were invited to
complete the in-lab task no less than one week but no more than two weeks following
completion of the global screening questionnaire and identification of eligibility. To ensure that
participation in the task was completed within this time frame, the global screening questionnaire
was made available to individuals enrolled in the HPR system at varying time and dates within
each week until a maximum of 100 individuals completed the questionnaire, at which point
access to the global questionnaire was closed. At this time, eligible participants were identified
based on their responses on the modified Kinsey scale and invited to enroll in the study within
the specified time frame. This identification and enrollment procedure was completed on a
weekly basis. Participant confidentiality and participation tracking was maintained through the
use of identification numbers assigned by the HPR system such that researchers were not
exposed to identifying information such as names or email addresses.
Lab task procedures.
All procedures were approved by the Institutional Review Board (IRB) of Michigan State
University prior to beginning the study. Eligible participants were invited to attend a 75-minute

75

lab-based session conducted at the Michigan State University Department of Psychology. All
sessions were conducted by a team of trained graduate and undergraduate research assistants
who were blind to the sexual orientation status of participants. Upon arrival at the laboratory,
participants were asked to read a written consent form detailing their rights as research
participants, the risks and benefits of participation in the study, and a basic outline of what
participation in the study entailed.
Given that the potential risks of inadvertent disclosure of sexual orientation status were
judged by the researcher and IRB members to outweigh the need to maintain a written record of
consent, and the study design did not necessitate retention of identifying information, a waiver of
written consent was obtained. Participants were informed verbally and in writing that the study
was being conducted anonymously and so no identifying information would be collected.
Participants were afforded the opportunity to ask questions about their rights as a research
participant at this time and researchers fully answered any questions regarding the consent
procedure to ensure that participants understand their rights. Participants were informed that they
may choose not to answer any of the questions or withdraw from the study at any time, and that
their decision to do so would not affect their relationship with MSU. In lieu of written consent,
verbal consent was obtained by explaining the consent procedure and asking participants to state
their consent to participate explicitly (e.g., “I consent to participating in this study”). Participants
who elected not to participate in the study at the time of consent were thanked for reviewing the
consent information but did not receive HPR credit for their attendance. Participants who
consented to participate in the study and began the task were awarded HPR credit for their
participation regardless of their decision to not answer questions or terminate their participation
early.

76

Participants were seated approximately 60cm from a 19-inch computer monitor (for more
detailed information, see Materials, above). Researchers provided a verbal overview of study
procedures and introduced study materials to the participants in order to familiarize them with
the task and mitigate potential anxiety over study procedures (e.g., physiological recording).
Researchers explained the function and safety of the equipment and answered any additional
participant questions at this time. After answering any additional questions, electrodes were
attached to the middle and index fingers of the left hand, below the left eye, and on the forehead,
using double-sided adhesive, medical tape, and a small amount of conductive gel. Once
electrodes were attached, researchers examined physiological recording data to ensure that
equipment was functioning within normal parameters and made adjustments if needed.
Researchers retained a written record of any abnormalities in task administration, physiological
recording, and/or participant behavior to protect study validity.
Once researchers verified that all equipment was functioning within normal parameters,
they introduced the computer task. Researchers described the structure of the task and played an
example of an acoustic startle probe. Participants were asked to listen carefully to the
information provided in each auditory clip and to do their best to imagine the events described in
as much detail as possible. Researchers also explained the SAM subjective emotion response
system and showed example scales. After explaining the task, researchers answered any
questions. Once participants indicated that they understood the task, participants put on
headphones and researchers left the room. Participants were instructed to press any button on the
computer keyboard once the researcher had left to begin the task.
The computer task was entirely automated and was similar to that developed by Quevado
and colleagues (2009), which has been shown to reliably elicit affectively-modulated startle

77

responses across a wide range of age groups. Modifications to the paradigm consisted of the use
of auditory stimuli in place of visual stimuli to reduce potential confounding effects of imagery,
the constraint that no bias threat and general threat clip should be played immediately following
one another, and the increase of the interval between startle probes from 5 to 10 seconds in order
to allow concomitant skin conductance measures to return to baseline between clips.
Participants completed 11 trials (3 bias threat, 3 general threat, and 5 neutral), which
were presented in one of six randomized orderings created via a publically-available
computerized random list generator (www.random.org/lists/). Which ordering was presented was
pre-determined by participant number via random list generator to ensure that no group had any
greater or lesser chance of being presented with any particular ordering. For each trial,
participants were first instructed to listen to the auditory clip and effortfully imagine the
described events while listening. To ensure attention to task, participants were informed that they
would be asked questions regarding the content of the clip following clip presentation.
For each trial, the following procedures were followed. During presentation of the
stimuli, three binaurally-presented acoustic startle probes consisting of white noise played at
95dB were randomly presented to elicit a startle reflex response with the constraint that no two
startle probes were presented within a single 10 second interval (to enable psychophysiological
measures to return to baseline between probes). Following presentation of the auditory stimulus,
an intertrial interval (ITI) of 10 seconds was presented during which no auditory stimulus was
given and a single acoustic startle probe of 95dB was presented at a random point within the
interval. Following the ITI, participants were be presented with four questions about their
subjective emotional response to the clip (for detailed information, see Measures, below).
Participants were also be presented with two questions related to the semantic content of the clip

78

(e.g., if the incident occurred in a stated location such as a park, participants were asked to select
the location from a list of possible locations) to ensure continued attention to task. Upon
completion of these questions, participants proceeded to the next trial until all eleven trials were
completed.
Following completion of the trials, participants were disconnected from the
psychophysiological equipment and completed post-task questionnaires on their demographics,
victimization history, collective identification with LGB or heterosexual individuals as a group,
world assumptions, tendency toward negative affect/neuroticism, and post-traumatic cognitions
related to any significant stressful life event. All questionnaires were presented on the computer
and participants used the number pad to enter their responses. Once participants had completed
the questionnaires, they were informed that they had completed the study and were invited to ask
any additional questions about the study aims, their participation, or any other related matter.
Researchers ensured that participants were provided with a copy of the consent form, including
contact information for the researchers and IRB, prior to leaving the laboratory.

79

CHAPTER EIGHT
Results
Preliminary Data Analyses
Missing values.
A total of 2,318 data points were included in analyses of the present study, across 32
primary and 6 supplemental variables comprised of a combination of physiological and
questionnaire data. A total of 58 (2.5%) data points were missing in the final data set. As
previously mentioned, of those participants who completed the lab task and were retained within
the final sample (N = 60), 8 (13.3%) experienced technical errors leading to <10% loss of
questionnaire data. Five participants (8.3%) experienced the loss of a final demographic
questionnaire and the final two items of a past bias victimization questionnaire (assessing
frequency of sexual assaults prior to age 18 and within the past year). Three participants (5.0%)
produced non-analyzable startle reflex EMG data. Consideration of the method of data loss (via
technical errors) does not suggest any theoretical mechanism of nonrandom missingness. Little’s
MCAR test failed to reject the null hypothesis that these data points were missing completely at
random, χ2(56, N = 60) = 24.95, p = 1.00.
Options for possible imputation of missing data for variables of interest in the primary
analyses (e.g., startle reflex EMG) were explored. The most powerful method of imputation,
multiple imputation, was not feasible for the purposes of the current study as the statistical
software in use (SPSS) was unable to provide pooled estimates for mixed model ANOVA
analyses, and an agreed-upon set of guidelines for averaging test values across iterations was not
available at the time of analysis. Therefore, given the small proportion of overall missingness
within the full data set, a pairwise deletion method was employed as the most parsimonious

80

solution. Because demographic information was not necessary for the purposes of primary
hypothesis testing and imputation of a large number of categorical values was not feasible,
missing values for these six participants were also excluded pairwise from frequency analyses of
sample characteristics. (As information on age and sex were collected on the pre-task
questionnaire, these demographic variables were retained for all participants.)
Psychophysiological data reduction.
Offline analyses were performed using BrainVision Analyzer 2 software (BrainProducts,
Gilching, Germany). EMG activity was bandpass filtered with cutoffs of 28 and 256 Hz (24
dB/Oct rolloff and 48 dB/Oct rolloff, respectively). Individual trials were inspected and rejected
if the startle reflex began less than 20ms following probe onset or if an unrelated blink occurred
within 50ms of the startle probe. EMG startle reflex activity was rectified in the post-startle
window and smoothed using a 6-point running average. Startle amplitude was then quantified as
the maximum response in the 20-200 ms post-probe window relative to the average activity in
the 200-ms pre-probe baseline period. Skin conductance activity was defined as the average
value of the SCL curve during presentation of the audio clip relative to the 1000ms pre-stimulus
baseline. Finally, startle amplitude and average SCL for each subject were converted to t-scores
to reduce between-subject variability unrelated to variables of interest in the current study.
Questionnaire data preparation.
Raw questionnaire data were preserved in EPrime output files, which were converted to
Microsoft Excel files for the purposes of efficient data entry. Audio response questionnaire files
and post-task questionnaire files were run through separate macros (repeatable automated task
programs within Excel) to organize data into single-row sets of values for each participant,
which were then entered into the statistical analysis program (SPSS). A random subsample of

81

data points within each macro-produced single-row set of values was compared to the
accompanying raw data value to assess the accuracy of this process. No discrepancies were
observed between raw data files and data entered into the analysis file. Data were also checked to
ensure that entered values were consistent with the theoretical range of possible responses to
each questionnaire item. A small number of participants appeared to have double-entered
responses on a fraction of categorical items assessing demographic characteristics in which the
participant was asked to enter a number and then press the enter key (e.g., “Would you describe
yourself as Hispanic or Latino/a? 1. Yes; 2. No,” participant response: “22{ENTER}”). None of
these responses appeared ambiguous upon examination and so they were recoded. No other
discrepancies were observed.
Evaluation of statistical and theoretical assumptions.
Remaining variables were evaluated for assumptions of the general linear model once the
data had been checked for accuracy of entry and coding. Standardized residuals were retained
after each analysis and examined for the presence of outliers. Cook’s distance and single fitted
values did not reveal any influential cases.
Confirmation of the structure of the World Assumptions Questionnaire.
Due to unacceptable reliability of some subscales of the World Assumptions
Questionnaire (see Methods, above), a confirmatory factor analysis was employed in order to
examine whether the factor structure observed by Kaler (2009) in scale development was
adequately replicated in the present study. Factor analysis of the questionnaire data within the
final sample (N = 62) alone was not feasible due to failure to meet common recommended
guidelines of both minimum necessary sample size and minimum ratio of sample size to model
variables (for a review, see MacCallum, Widaman, Zhang and Hong, 1999). However, as the

82

WAQ was administered to all individuals who completed the pre-task screening questionnaire,
sufficient sample size was achieved by including questionnaire data from this larger screening
sample. In total, WAQ questionnaire data was available for 2,352 individuals, of whom 204
identified as LGB based on Kinsey scale cutoffs. An equal number of heterosexual individuals
were selected from the larger sample at random using an automated computer algorithm, for a
total sample of 408 individuals.
Following an initial process of item selection, content rating, exploratory factor analysis
(EFA), item reduction, and confirmatory factor analysis (CFA), Kaler’s (2009) WAQ was
comprised of 22 items loading on four subscales: Controllability and Predictability of the World,
Controllability and Predictability of People, Trustworthiness and Goodness of People, and Safety
and Vulnerability. To assess the fit of the questionnaire to this proposed structure within the
present sample, a CFA was conducted using IBM SPSS AMOS v.22.0.0. Appendix C presents
the models of these analyses. Determination of model fit was based on suggested cutoffs
frequently cited in the literature for modifying indices (Steiger & Lind, 1980; Bentler, 1990).
Fit of the specified model was assessed with maximum likelihood estimation. Factors
were allowed to correlate with one another, error covariances were constrained to zero, and
factor variances were constrained to one. The model for this analysis was recursive and can be
seen in Figure 8. Fit statistics of the model consistently did not support fit in the current sample,
χ2 (203) = 648.38, p < .001, RMSEA = .07 (90% CI = .067-.080), CFI = .83, SRMR = .11.
Standardized estimates for the full sample and LGB and heterosexual subsamples are displayed
in Figures 9 – 11. The factor structure of the WAQ could not be replicated in the present sample
and therefore was not included in analyses for this study.
Individual differences in neuroticism as a possible covariate.

83

Given the possibility that group differences could be a function of differences in tendency
toward negative affect across groups, this supplementary measure was included in data collection
and explicitly examined a priori as a possible covariate in primary analyses. An independent
samples t-test was conducted to assess whether neuroticism scores varied across groups (see
Appendix E). No significant difference was observed. Given this observation and limited sample
size, neuroticism was not included as a covariate in primary hypothesis testing.
Primary Analyses
A series of mixed model repeated measures ANOVA and ANCOVA models were
employed to test the hypotheses of the study, with outcome variables (SCL, startle reflex EMG
magnitude, SAM ratings, PTCI) entered in separate models. Identity variables (centrality, private
regard, and public regard) were entered as continuous covariate in models examining their
moderating effects. In analyses examining startle reflex magnitude, intertrial interval startle
magnitude scores were also entered into the model.
Mauchly’s (1940) test of sphericity was used to assess for violations of sphericity in
repeated measures analyses of variance in which more than two levels of outcome variable were
examined. Where violations were observed, the Greenhouse-Geisser correction (Greenhouse &
Geisser, 1959) was applied. Effect sizes (Cohen’s d and ηp2) were computed for all analyses. For
within-subjects comparisons, effect size estimates were corrected for dependence between means
according to the guidelines provided by Morris and DeShon (2002). As final sample size was
less than initially planned, observed power estimates (α = .05) based on observed effect sizes
were also computed for all analyses. Due to smaller than expected sample size, descriptive
information is first examined for the pattern and magnitude of observed relations between
variables in conjunction with null-hypothesis significance testing.

84

Descriptive data.
Figures 1 through 4 graphically display the means and confidence intervals for
participants’ average skin conductance level (SCL) during clip presentation, average startle
reflex EMG magnitude during the three clip conditions and intertrial interval, subjective ratings
of affect (arousal, valence, dominance, and warmth) following clip presentation, and pre- and
post-task PTCI Negative Cognitions about the World scores, stratified by Kinsey group.
Numerical means, standard deviations, and ranges are presented in Tables 16 – 19 in Appendix
D and interrcorrelations between outcome variables of interest are presented in Tables 21 – 35 in
Appendix D.
Visual examination of mean trends suggests that heterosexual and LGB participants
displayed different patterns of physiological responding across clip conditions (Figures 1 and 2,
below). LGB participants’ mean autonomic arousal was highest in the general threat condition,
somewhat lower in the bias threat condition, and lowest in the neutral condition. Heterosexual
participants’ mean autonomic arousal was lowest in the general threat condition, with similar
higher levels of mean arousal during neutral and bias threat clips. With regard to startle reflex
potentiation, LGB participants displayed larger startle reflex during threat condition clips (both
general threat and bias threat) as compared to neutral clips. Heterosexual participants’ greatest
mean startle reflex potentiation magnitude was observed during presentation of general threat
clips, with moderate potentiation during presentation of neutral clips, and lowest potentiation
during presentation of bias threat clips.

85

Figure 1.
Average Skin Conductance Level (SCL) across Stimulus Conditions, by Group (N = 60)

Average SCL Scores (T)

60
55
50
45
40
35
30
Heterosexual
Neutral

LGB
General Threat

Bias

Note: Error bars represent the 95% confidence interval of the mean.
Figure 2.

Average Startle EMG Magnitude Tscore

Average Startle Reflex EMG Magnitude across Stimulus Conditions, by Group (N = 57)
60
55
50
45
40

35
30
Heterosexual
Neutral

LGB

General Threat

Bias

Note: Error bars represent the 95% confidence interval of the mean.
86

Examination of the mean self-reported levels of arousal and affect valence during clip
presentation (Figures 3 and 4) similarly suggests that the pattern of affective response to bias
crime material differed across groups. Heterosexual participants rated neutral clips as least
arousing, with similar ratings of arousal for general threat and bias threat clips. LGB participants
rated neutral clips as least arousing, with general threat clips evoking more arousal and bias
threat clips evoking the most arousal. With regard to affect valence, the pattern of response is
highly similar to that of arousal. Heterosexual participants rating neutral clips as most positive,
with general threat and bias threat clips more negative (but similarly so). Similarly, LGB
participants rated neutral clips as the most positive, with general threat clips rated more
negatively. However, visual inspection suggests that LGB participants’ ratings of affect valence
during bias threat clips was lower than affect valence reported during general threat clips.
Similar patterns were observed with respect to dominance and interpersonal warmth, such that
LGB individuals generally appeared to report lower dominance and warmth during bias crime
clips, as compared to heterosexual participants and as compared to ratings of general threat clips.

87

Figure 3.

Average SAM Score

SAM Arousal and Valence Ratings across Stimulus Conditions, by Group (N = 60)
9
8
7
6
5
4
3
2
1
0
Neutral

General
Threat

Bias Threat

Neutral

Arousal

General
Threat

Bias Threat

Valence
Heterosexual

LGB

Note: Error bars represent the 95% confidence interval of the mean.
Figure 4.

Average SAM Score

SAM Dominance and Warmth Ratings across Stimulus Conditions, by Group (N = 60)
9
8
7
6
5
4
3
2
1
0
Neutral

General
Threat

Bias Threat

Neutral

Dominance

General
Threat
Warmth

Heterosexual
88

LGB

Bias Threat

Note: Error bars represent the 95% confidence interval of the mean.
The observed pattern of pre- to post-task change in PTCI Negative Cognitions about the
World scores (Figure 5) was contrary to hypotheses. At pre-task, LGB participants appeared to
report slightly higher levels of negative cognitions than did heterosexual participants. However,
both groups reported lower levels of negative cognitions post-task, contrary to predictions that
reported negative cognitions would increase following exposure to threat information.
Figure 5.
Average Pre- and Post-Task PTCI Negative Cognitions about the World, by Group (N = 60)
5
4.5

PTCI Score

4
3.5
3
2.5
2
1.5
1
Heterosexual

LGB

Pre-Task

Post-Task

Note: Error bars represent the 95% confidence interval of the mean, adjusted for between-subject
variability using Cousineau’s (2005) method.
Figure 6 graphically displays the means and 95% confidence intervals of the three
dimensions of collective identity assessed within the present study (centrality, private regard, and
public regard), stratified by Kinsey group. The precise values of the means, standard deviations,

89

and observed ranges of these variables are presented in Table 20 in Appendix D. Of note, LGB
participants reported lower public regard as compared to heterosexual participants.
Figure 6.
Average Centrality, Private Regard, and Public Regard Scores, by Group (N = 60)
7
6

Scale Score

5
4
3
2
1

0
Heterosexual
Centrality

LGB
Private Regard

Public Regard

Note: Error bars represent the 95% confidence interval of the mean.
Tables 3 – 9 below display the bivariate correlations between collective identity variables
and outcome variables of interest, by Kinsey group. Correlations between identity variables and
outcome variables of interest were not statistically significant, but some trends were observed
within the present sample.
For the LGB group, identity centrality was marginally negatively related to private regard
and strongly negatively correlated with public regard. Private and public regard were strongly
positively related. Centrality showed a weak positive relationship with autonomic arousal during
bias crime clips and startle reflex magnitude during neutral clips, a moderate positive
relationship with subjective ratings of affect during neutral clips and with pre-task PTCI scores,
90

and a weak positive relationship with post-task PTCI scores. Private regard was weakly
positively correlated with SCL, affect valence, and dominance during neutral clips, weakly
correlated with self-reported arousal during general threat clips, and moderately correlated with
reported warmth during general and bias threat clips. Public regard had a weak negative
correlation with startle reflex magnitude, and a weak positive correlation with self-reported
arousal, during general threat clips. No correlations were observed between identity variables of
interest and change in PTCI scores.
For the heterosexual group, identity centrality, private regard, and public regard showed
small to moderate positive interrelations. Identity centrality was weakly negatively correlated
with SCL and moderately negatively correlated with startle reflex magnitude during general
threat clips. Identity centrality also showed weak, moderate, and strong negative correlations to
self-reported arousal during bias, neutral, and general threat clips, respectively. A weak positive
correlation between centrality and self-reported dominance during bias crime clips was also
observed. Finally, centrality was weakly negatively correlated with both pre- and post-task PTCI
Negative Cognitions about the World scale scores. Private regard was strongly negative
correlated with startle reflex magnitude during general threat clips, weakly positively related to
self-reported arousal during bias crime clips, and moderately negatively correlated with reported
affect valence during neutral clips. Public regard was moderately positively correlated with SCL
during neutral clips and strongly negatively correlated with SCL during viewing of bias crime
clips. Weak negative correlations were also observed between public regard, self-reported
arousal during neutral and general threat clips, and self-reported interpersonal warmth during
general threat clips. A weak positive correlation was observed between public regard and selfreported dominance during general threat clips. As within the LGB sample, no correlations were

91

observed between identity variables of interest and change in PTCI scores for heterosexual
participants.
Table 3.
Correlations between Average Skin Conductance Level (SCL) across Clip Conditions and
Dimensions of Collective Identity within the LGB Subsample (in Plain Text; N = 28) and
Heterosexual Subsample (in Bold; N = 32)
1

2

3

4

5

6

-

-.17

-.40*

-.16

-.07

.24

2. Private Regard

.33

-

.40*

.23

-.05

-.19

3. Public Regard

.20

.25

-

.03

.03

-.06

4. Neutral SCL

.14

.04

.38*

-

-.56*

-.47*

5. GT SCL

-.21

.05

.14

-.44*

-

-.47*

6. Bias SCL

.07

-.09

-.49*

-.49*

-.56*

-

1. Centrality

*p < .05
Table 4.
Correlations between Startle Reflex EMG Magnitude across Clip Conditions and Dimensions of
Collective Identity within the LGB Subsample (in Plain Text; N = 27) and Heterosexual
Subsample (in Bold; N = 30)
1

2

3

4

5

6

-

-.17

-.40*

-.20

.15

-.07

2. Private Regard

.33

-

.40*

.17

-.02

.10

3. Public Regard

.20

.25

-

-.06

-.25

.16

4. Neutral Startle

.19

.05

.13

-

-.02

-.23

5. GT Startle

-.35

-.40*

.03

-.06

-

-.55*

6. Bias Startle

.30

.17

-.03

-.43*

-.35

-

1. Centrality

*p < .05

92

Table 5.
Correlations between Subjective Ratings of Arousal during Neutral, General Threat, and Bias
Threat Clips and Dimensions of Collective Identity within the LGB Subsample (in Plain Text; N
= 28) and Heterosexual Subsample (in Bold; N = 32)
1

2

3

4

5

6

-

-.17

-.40*

.35

.08

.10

2. Private Regard

.33

-

.40*

.05

.24

.10

3. Public Regard

.20

.25

-

-.07

.06

.01

4. Neutral Arousal

-.30

-.03

-.25

-

.33

.18

5. GT Arousal

-.43*

-.04

-.24

.49*

-

.41*

6. Bias Arousal

-.24

.20

.17

.32

.63*

-

1. Centrality

*p < .05
Table 6.
Correlations between Subjective Ratings of Affect Valence and Dimensions of Collective Identity
within the LGB Subsample (in Plain Text; N = 28) and Heterosexual Subsample (in Bold; N =
32)
1

2

3

4

5

6

-

-.17

-.40*

.03

-.04

-.11

2. Private Regard

.33

-

.40*

.29

.05

.00

3. Public Regard

.20

.25

-

.08

-.03

-.02

4. Neutral Valence

-.08

-.36*

-.01

-

.47*

.17

5. GT Valence

.09

-.10

-.10

.01

-

.58*

6. Bias Valence

.18

-.15

-.16

-.02

.82*

-

1. Centrality

*p < .05

93

Table 7.
Correlations between Subjective Ratings of Dominance and Dimensions of Collective Identity
within the LGB Subsample (in Plain Text; N = 28) and Heterosexual Subsample (in Bold; N =
32)
1

2

3

4

5

6

-

-.17

-.40*

.15

-.13

-.17

2. Private Regard

.33

-

.40*

.26

.09

.05

3. Public Regard

.20

.25

-

.02

.08

.08

4. Neutral Dominance

-.13

-.08

.04

-

.35

-.27

5. GT Dominance

.15

-.01

.21

.45*

-

.37*

6. Bias Dominance

.21

-.12

.06

.39*

.84*

-

1. Centrality

*p < .05
Table 8.
Correlations between Subjective Ratings of Interpersonal Warmth and Dimensions of Collective
Identity within the LGB Subsample (in Plain Text; N = 28) and Heterosexual Subsample (in
Bold; N = 32)
1

2

3

4

5

6

-

-.17

-.40*

.05

.06

.10

2. Private Regard

.33

-

.40*

.00

.31

.31

3. Public Regard

.20

.25

-

-.07

-.03

-.01

4. Neutral Warmth

.07

.14

.02

-

.04

-.23

5. GT Warmth

.08

-.09

-.24

-.09

-

.78*

6. Bias Warmth

.18

.01

-.10

-.05

.53*

-

1. Centrality

*p < .05

94

Table 9.
Correlations between Pre- and Post-Task Negative Cognitions about the World and Dimensions
of Collective Identity within the LGB Subsample (in Plain Text; N = 28) and Heterosexual
Subsample (in Bold; N = 32)
1

2

3

4

5

6

-

-.17

-.40*

.05

.20

-.17

2. Private Regard

.33

-

.40*

.00

.07

.11

3. Public Regard

.20

.25

-

-.07

-.01

-.00

4. Pre-Task PTCI

-.20

-.03

-.06

-

.75*

-.33

5. Post-Task PTCI

-.28

-.03

-.13

-.09

-

.38*

6. Change in PTCI

-.17

-.01

-.12

-.05

.62*

-

1. Centrality

*p < .05
Hypothesis testing.
The first research question asked whether LGB and heterosexual participants would show
differing patterns of response to bias crime information across physiological, affective, and
cognitive domains. First, I predicted that LGB individuals would show greater magnitude of
ANS activation and startle reflex in response to bias crime exposure than would heterosexual
individuals. Tables 10 and 11 present the results of two mixed model repeated measures
ANOVAs testing these hypotheses.
Statistical comparison of patterns of autonomic nervous system responding using the
general linear model indicated a significant interaction between sexual orientation group and clip
condition. Post-hoc analyses with Bonferroni correction indicated that the average SCL varied
between groups in the neutral (F[1,58] = 5.98, p = .02, ηp2 = .09, observed power = .67) and
general threat (F[1,58] = 10.45, p = .002, ηp2 = .15, observed power = .89) conditions, such that
the heterosexual participants demonstrated greater autonomic nervous system activation during
95

neutral clips relative to LGB participants, and LGB participants demonstrated greater autonomic
nervous system activation during general threat clips as compared to heterosexual participants.
Table 10.
Results of a 2x3 Mixed Model Repeated Measures Analysis of Variance Comparing Average Skin
Conductance Level During Three Clip Conditions (Neutral, General Threat, and Bias Threat),
Across Heterosexual and LGB Participants (N = 60)
Power
df

F

p

ηp2

(obs.)

Group

1

2.92

.09

.05

.39

Condition

2

1.09

.34

.02

.24

2

5.75

.004

.09

.86

Main effects

Two-way interactions
Group*Condition

With regard to startle reflex magnitude, statistical analysis did not reveal a significant
interaction between group and condition or any significant main effect of group or condition.
Small to medium effect sizes were observed for the main effects of group and the interaction of
group and condition.
Table 11.
Results of a 2x4 Mixed Model Repeated Measures Analysis of Variance Comparing Average
Startle Reflex EMG Magnitude During Four Clip Conditions (Neutral, General Threat, Bias
Threat, and Intertrial Interval) Between Heterosexual and LGB Participants (N = 60)
Power
df

F
96

p

ηp2

(obs.)

Table 11 (cont’d)
Main effects
Group

1

1.12

.29

.02

.18

Condition

2

.09

.91

.00

.06

2

2.62

.08

.05

.51

Two-way interactions
Group*Condition

I also predicted that LGB participants would report higher levels of arousal and lower
affect valence during exposure to bias crime clips than did heterosexuals. Results of a series of
mixed-model repeated measures ANOVAs testing these hypotheses are presented in Table 12.
Also presented for exploratory purposes are the dimensions of dominance and interpersonal
warmth.
Statistical tests of significance were largely consistent with the hypothesized pattern of
results. An interaction between condition and group in predicting ratings of arousal was not
significant, although main effects of both condition and group on arousal ratings were observed.
For both groups, arousal ratings for the general threat clips (t[59] = 8.01, p < .001, d = 1.04, 95%
CI of the difference = 1.14 – 1.90) and bias threat clips (t[59] = 9.19, p < .001, d = 1.20, 95% CI
of the difference = 1.78 – 2.78) were higher than ratings for the neutral clips. Arousal ratings for
the bias crime clips were greater than those for the general threat clips for the full sample, t(59) =
3.83, p < .001, d = .51, 95% CI of the difference = .36 – 1.16. A between-group difference was
observed in ratings of arousal to bias crime clips; specifically, LGB individuals reported greater
arousal in response to bias crime clips than did heterosexual participants, t(55.56) = 3.03, p <
.01, d = .81, 95% CI of the difference = 0.45 – 2.19.

97

Similarly, the interaction between condition and group in predicting ratings of affect
valence was not significant, but main effects of condition and group were observed. Both groups
rated their affective response to general threat (t[59] = 11.46, p < .001, d = , 95% CI of the
difference = -1.48) and bias threat clips (t[59] = 14.44, p < .001, d = -1.89, 95% CI of the mean =
-3.22 – -2.44) as more negative than their response to neutral clips. Participants also rated affect
in response to the bias crime clips as more negative than the general threat clips t(59) = -7.60, p
< .001, d = -1.01, 95% CI = -1.07 – -.62. Comparing ratings across the groups, LGB individuals
reported less positive affect in response to bias crime clips than did heterosexual individuals,
t(49.99) = 3.27, p < .01, d = .92, 95% CI = .35 – 1.57. LGB participants also reported feeling less
dominant during bias crime clips than did heterosexual participants, t(58) = 2.27, p = .03, d =
.58, 95% CI = .09 – 1.68. No significant difference was observed between groups in ratings of
interpersonal warmth during bias crime clips, t(58) = 1.49, p = .14, d = .38, 95% CI = -.24 –
1.62.
Table 12.
Results of a Series of 2x3 Mixed Model Repeated Measures Analyses of Variance Comparing
Four Subjective Ratings of Affect (Arousal, Valence, Dominance, and Warmth) During Three
Clip Conditions (Neutral, General Threat, and Bias Threat) Between Heterosexual and LGB
Participants (N = 60)
Power
df

F

p

ηp2

(obs.)

1

5.95

<.01

.09

.67

1.78

61.70

<.001

.52

1.00

Arousal
Main effects
Group
Condition

98

Table 12 (cont’d)
Two-way interactions
Group*Condition

1.78

2.86

.07

.05

.52

1

5.95

<.01

.09

.39

1.43

43.26

<.01

.43

1.00

1.43

3.36

.06

.06

.62

1

5.59

.34

.02

.16

1.44

8.21

<.01

.12

.96

1.44

3.31

.06

.06

.52

1

1.83

.18

.03

.16

1.53

28.18

<.001

.33

1.00

1.53

0.96

.39

.02

.21

Valence
Main effects
Group
Condition
Two-way interactions
Group*Condition
Dominance
Main effects
Group
Condition
Two-way interactions
Group*Condition
Warmth
Main effects
Group
Condition
Two-way interactions
Group*Condition

With regard to negative cognitions about the world, I predicted that LGB participants
would show greater increases in negative cognitions about the world than would heterosexual
participants. Results of a mixed-model repeated measures ANOVA testing this hypothesis are
99

presented in Table 13. Contrary to prediction, no interaction between group and task was
observed. A main effect of task was observed. Specifically, both groups displayed a decline in
PTCI scores from pre-task to post-task, t(59) = 4.15, p < .001, d = .54, 95% CI of the mean = .16
– .47. No significant main effect of group was observed.
Table 13.
Results of a Mixed Model Repeated Measures Analysis of Variance Comparing Negative
Cognitions about the World Pre- and Post-Laboratory Task, Between Heterosexual and LGB
Participants (N = 60)
Power
df

F

p

ηp2

(obs.)

Group

1

2.28

.15

.04

.30

Task

1

17.52

<.001

.23

.98

1

0.64

.43

.01

.12

Main effects

Two-way interactions
Group*Task

The second research question asked whether LGB participants would show a pattern of
responding to bias crime information that is consistent with defensive mobilization against
threat. I hypothesized that LGB individuals would show a greater magnitude of ANS activation
and greater startle response to bias-crime clips than to general threat or neutral clips. Contrary to
prediction, LGB participants’ average autonomic nervous system arousal during bias crime clips
did not significantly differ from arousal during general threat (t[27] = -1.17, p = .25, d = -.22,
95% CI of the difference = -8.01 – 2.19) or neutral clips (t[27] = 1.67, p = .11, d = .32, 95% CI
of the difference = -.94 – 9.30). LGB participants demonstrated greater autonomic nervous
100

system activation during viewing of general threat clips than during viewing of neutral clips,
t(27) = 2.68, p = .01, d = .51, 95% CI of the difference = 1.66 – 12.51.
As previously noted, visual examination of group means suggests that LGB participants’
greatest mean magnitude of startle reflex potentiation was observed during presentation of both
bias threat and general threat clips, with somewhat lower potentiation observed during
presentation of neutral clips. This is not consistent with expectations that startle reflex
potentiation during bias crime clips would exceed potentiation during general threat and neutral
clip presentation. Statistical analysis of these trends within the general linear model did not
reveal any significant differences between mean autonomic arousal during bias crime clip
presentation and mean autonomic arousal during presentation of general threat or neutral threat
clips. Similarly, no statistically-significant differences were observed between clip conditions in
mean startle reflex magnitude for LGB participants.
I further hypothesized that LGB individuals would self-report greater arousal and less
positive affect during exposure to bias crime clips than during general threat or neutral clips. As
predicted, LGB participants rated bias crime clips as more arousing (t[27] = 8.25, p < .001, d =
.92, 95% CI = 2.03 – 3.37) and less positive (t[27] = -17.46, p < .001, d = -3.34, 95% CI = -3.55
– -2.80) than neutral clips. LGB participants also reported feeling more arousal (t[27] = 4.83, p <
.001, d = .20, 95% CI = .73 – 1.80) and less positive affect (t[27] = -9.40, p < .001, d = -1.79,
95% CI = -1.54 – -.99) during bias crime clips than during general threat clips.
Finally, I predicted that LGB participants would report more negative cognitions about
the world following exposure to bias crime clips. The observed pattern of pre- to post-task
change in PTCI Negative Cognitions about the World scores was contrary to hypotheses. As

101

previously noted, LGB participants reported fewer negative cognitions about the world following
the lab task than they had prior to the lab task.
The third research question asked whether domains of collective identity (specifically,
centrality and public regard) moderated relations between group membership and responses to
bias crime information for LGB and heterosexual individuals. A series of 2x3 mixed-model
repeated measures analyses of variance conducted, with the three dimensions of collective
identity entered as covariates, to examine whether identity centrality, private regard, and public
regard moderated relations between sexual orientation group, clip condition, and outcome
variables of interest. Tables 36-42 in Appendix D display the results of these analyses. Figures
12 - 16 in Appendix D display the marginal means of outcome variables under consideration, at
grand mean values of identity centrality, private regard, and public regard.
I hypothesized that higher identity centrality would positively predict autonomic nervous
system activation, startle reflex magnitude, negative affect (i.e., greater ratings of arousal and
less positive reported valence), and more negative beliefs about the world in LGB participants
during exposure to bias crimes, but not heterosexual participants. No significant three-way
interactions were observed between centrality, clip condition, and group for any of the specified
outcome variables. Similarly, no significant interactions were observed between centrality and
Kinsey group, or between centrality and clip condition. Notably, examination of marginal means
suggest that the trends observed across groups in outcome variables were maintained in
physiological measures and SAM ratings of arousal and affect valence when controlling for
identity variables. Group differences in SAM ratings of dominance and warmth were attenuated
when controlling for identity variables. In the case of change in PTCI over time, controlling for
identity variables suggested divergent pre- to post-PTCI slopes between LGB and heterosexual

102

participants (though this overall interaction of group and condition remained non-significant in
statistical testing).
With respect to public regard, I hypothesized that lower reported public regard would
predict heightened ANS activation, greater startle reflex potentiation, more negative affect (i.e.,
greater arousal and less positive valence), and more negative cognitions about the world in LGB
individuals during exposure to bias crimes, but not heterosexual participants. An interaction was
observed between public regard, clip condition, and sexual orientation on average SCL. To
clarify the nature of the interaction between public regard, clip condition, and sexual orientation
on SCL, difference scores were calculated between average SCL scores during each clip
condition (bias – general threat, bias – neutral, general threat - neutral), and bivariate correlations
between public regard and these difference scores were computed separately for the LGB and
heterosexual subsamples. These matrices are presented below in Table 14. Examination of these
correlation matrices suggests that the nature of this interaction is as follows: for heterosexual
participants, greater perceived public regard was associated with smaller differences in
autonomic nervous system activation during bias crime clips as compared to during general
threat or neutral clips. No significant relations were observed between public regard and
differences in autonomic nervous system activation among LGB participants, contrary to
predictions.
Table 14.
Correlations between Public Regard and Difference Scores between Average Skin Conductance
Level (SCL) across Stimulus Conditions, within the LGB Subsample (in Plain Text; N = 28) and
Heterosexual Subsample (in Bold; N = 32)
LGB

Het

103

1

2

3

4

Table 14 (cont’d)
M (SD)

M (SD)

r

r

r

r

1. Public Regard

3.64 (1.26)

5.64 (1.01)

-

-.05

-.05

.00

2. Bias – GT SCL

-2.91 (13.16)

5.48 (14.32)

-.36*

-

.44*

-.53*

3. Bias – Neutral SCL

4.18 (13.21)

1.02 (13.65)

-.51*

.55*

-

.53*

4. GT – Neutral SCL

7.09 (13.98)

-4.47 (13.24)

-.14

-.51*

.43*

-

*p < .05
An interaction was also observed between public regard, clip condition, and sexual
orientation on subjective ratings of arousal. To clarify the nature of this interaction, difference
scores were calculated between SAM arousal ratings during each clip condition (bias – general
threat, bias – neutral, general threat - neutral), and bivariate correlations between public regard
and these difference scores were computed separately for the LGB and heterosexual subsamples.
Means, standard deviations, and correlation matrices of difference scores are presented below in
Table 15. Examination of these correlation matrices suggests that the nature of this interaction is
as follows: for heterosexual participants, greater reported public regard for heterosexuals was
positively associated with greater (positive) differences in reported arousal during bias crime
clips as compared to reported arousal during general threat clips. There was no such association
observed within the LGB subsample, contrary to predictions.
Table 15.
Correlations between Public Regard and Difference Scores between SAM Arousal Ratings
across Stimulus Conditions, within the LGB Subsample (in Plain Text; N = 28) and Heterosexual
Subsample (in Bold; N = 32)
LGB

Het

104

1

2

3

4

Table 15 (cont’d)
M (SD)

M (SD)

r

r

r

r

1. Public Regard

3.64 (1.26)

5.64 (1.01)

-

-.05

.06

.12

2. Bias – GT Arousal

1.26 (1.38)

0.32 (1.56)

.46*

-

.60*

-.24

3. Bias – Neutral Arousal

2.70 (1.73)

1.91 (2.03)

.34

.67*

-

.63*

4. GT – Neutral Arousal

1.44 (1.42)

1.59 (1.53)

-.02

-.14

.65*

-

*p < .05
Finally, I hypothesized that the interaction of high identity centrality and low public
regard would positively predicted greater defensive mobilization on the part of LGB participants
(but not heterosexual participants), marked by greater ANS activation, greater startle reflex
potentiation, more reported negative affect, and greater increases in negative cognitions about the
world from pre-task to post-task assessment. Contrary to this hypothesis, no significant
interactions were observed between identity centrality, public regard, group, and condition in the
present analyses.
Additional Analyses
Descriptive statistics, group comparisons, and bivariate correlations for additional
variables of interest (e.g., neuroticism) can be found in Appendix E, Tables 43 - 50. In brief,
notable observations include the following: for both groups, neuroticism was positively
correlated (weakly to moderately) with negative cognitions about the world and weakly
negatively correlated with autonomic arousal during general threat clips. Stressful life
experiences were weakly negatively associated with autonomic arousal during general threat
clips for the LGB subsample. Within the LGB subsample, past direct experiences of bias
victimization were weakly positively associated with self-reported arousal, post-task negative

105

cognitions about the world, and change in negative cognitions, and weakly negatively associated
with reported dominance and warmth during bias crime clips.

106

CHAPTER NINE
Discussion
The present study examined non-victim responses to auditory information about biasmotivated crimes against lesbian, gay, and bisexual (LGB) individuals in order to examine
whether LGB individuals evidence physiological, affective, and cognitive responses consistent
with vicarious threat. Based on a review of relevant literature in biological, social, and clinical
psychology, a model of threat responding, cumulative physiological consequences of stress, and
long-term cognitive changes in cognitions about the world was proposed to explain proximal and
distal pathways to negative mental health consequences of exposure to bias crime information
for LGB populations (Lannert, 2014). This model proposes that individuals who are members of
a targeted group are more likely to appraise bias-motivated violence against another member of
that group as a source of uncontrollable, unpredictable, and personally-relevant threat, and to
mobilize physiological and cognitive defensive systems in response. Over time, the combined
consequences of repeated activation of the threat response system, cognitive changes in beliefs
about the self, others, and the world, and behavioral changes aimed at increasing safety are
proposed to place LGB individuals at higher risk for stress-related psychiatric disorders such as
depression, anxiety, PTSD, and substance use disorders.
This dissertation study examined in-vivo responses to bias crime information in a sample
of heterosexual and LGB young adults as a preliminary examination of this model. Four forms of
response were considered: subjective affective responses as reported by participants, a
physiological marker of generalized autonomic arousal (skin conductance level), a physiological
marker of defensive responding to threat (startle reflex EMG magnitude), and changes in
negative beliefs about the world consistent with traumatogenic cognitions. In addition, the role of

107

collective (social) identity was examined as a potential moderator of these responses within the
LGB subsample. I hypothesized that LGB and heterosexual participants would evidence different
patterns of responses to bias crime information on physiological, affective, and cognitive
measures. I predicted that LGB individuals would display a pattern of proximal responses to bias
crime information consistent with defensive mobilization to threat. Specifically, I hypothesized
that LGB participants would display heightened autonomic arousal (SCL), potentiated startle
reflex magnitude, greater subjective arousal, and more negative affect in response to bias crime
clips, as compared to heterosexual participants and as compared to their responses to general
threat clips. I further hypothesized that LGB participants would show increases in negative
cognitions about the world following exposure. I proposed that domains of collective identity (in
particular, centrality and public regard) would moderate these responses such that high centrality
and high public regard would magnify defensive mobilization, both individually and in
interaction with one another. Results of the present study provide partial support for these
hypotheses. In general, participants’ subjective reports of affective response are consistent with
hypotheses, physiological measures were equivocal or inconsistent, and observed changes in
negative cognitions about the world were opposite of predictions.
Firstly, examination of group means indicated that the patterns of response across clip
conditions for all outcome variables, except for negative beliefs about the world, were different
for LGB and heterosexual participants. In the case of subjective ratings of affect, these
differences were statistically significant and effect sizes in the present sample were generally
medium to large. LGB participants rated bias crime clips as more aversive (more arousing, less
positive) than did heterosexual participants, and endorsed feeling less dominant during exposure
to bias crime clips. A non-significant but consistent trend was observed with ratings of

108

interpersonal warmth, with LGB participants reporting feeling less warm during bias crime clips
than heterosexual participants. With regard to autonomic arousal, heterosexual and LGB
participants produced different patterns of activation, and these group differences in pattern were
also statistically significant, with generally small to medium effect sizes observed. LGB
participants showed similar potentiation of startle reflex during bias threat and general threat
clips, greater than that exhibited during neutral clips, whereas heterosexual participants’ startle
reflex magnitude during bias threat clips was comparable to their response to neutral clips. Both
LGB and heterosexual participants displayed similar reductions in negative beliefs about the
world from pre-task to post-task assessments, contrary to the prediction that negative beliefs
would increase for LGB participants following the task and with no clear differences in the
degree of change by group.
Given past qualitative (Noelle, 2002; Perry & Alvi, 2012) and quantitative (Herek et al.,
1999) work suggesting that sexual minority individuals are more likely to respond to these events
as threatening, it is not surprising that patterns of response varied between LGB and heterosexual
participants. However, while the patterns of responses observed within the LGB sample were
generally suggestive of defensive responding, there was no support for traumatogenic response
(that is, both defensive responding to threat and alterations in cognition) as described in the
rationale and hypotheses. Participants’ reported affective responses (i.e., high arousal, low
valence, and low dominance) and observed patterns of startle reflex potentiation in response to
bias crime information are consistent with defensive mobilization to threat (Bradley et al, 2001;
Hamm & Weike, 2005; Lang, 2010). Notably, heterosexual participants evidenced potentiated
startle during general threat clips only, suggesting that LGB participants uniquely responded to
bias crimes as threatening. LGB participants evidenced similar levels of startle potentiation

109

during bias and general threat clips, and the greatest autonomic arousal during general threat
clips (followed closely by bias crime clips, then neutral clips). These patterns are not consistent
with the hypothesis that LGB individuals experience bias crime events as exceptionally
threatening as compared to non-targeted events. Finally, both LGB and heterosexual participants
reported fewer negative appraisals of the world following the lab task – entirely opposite of
hypotheses, and inconsistent with acute responses to traumatic events.
That LGB participants’ autonomic nervous system activation and startle reflex magnitude
observed during bias crime clips were not greater than that observed during general threat clips
(as was hypothesized) could be due to any of three factors. First, with regard to startle reflex
potentiation, results may reflect a ceiling effect in which startle reflex magnitude was maximized
during threat exposure for LGB participants, regardless of content. Second, the prediction that
LGB participants would exhibit greater startle potentiation and autonomic arousal in the bias
crime condition as compared to the general threat condition was based upon the assumption that
responses to bias crime events within that subgroup would be most consistent with those
observed in extreme cases of fear learning (such as social and specific phobias; Gros et al., 2009;
Hamm et al., 1997; Klorman et al., 1977). Though the present data suggest that LGB participants
respond to bias crime clips as a comparable source of threat to other, non-bias-motivated violent
crimes, they may not necessarily respond to them as an imminent threat in the way that an
individual with a phobia may respond to the feared object or situation. Imminence of threat is an
important dimension of threat attribution that demonstrates positive relations with the degree of
observed defensive activation (Fanselow, 1994). In the present study, I hypothesized that LGB
individuals would perceive bias crime events as imminent threats given the frequency of these
events and their uncontrollable, unpredictable, and (perceived) likely nature, and so demonstrate

110

high levels of defensive activation analogous to individuals exposed to phobic stimuli (Hamm et
al., 1997). However, it may be that LGB individuals experience learning about these events as
threatening but not necessarily imminent – that is, with some psychological distance.
Finally, the assumption that the pattern of response observed in a phobic response and
one observed in a traumatic response are identical may itself be faulty. Although phobic
individuals consistently show sharply potentiated startle reflex and autonomic arousal in
response to phobia cues (Gerlach, Wilhelm, Gruber, & Roth, 2001; Gros et al., 2009; McTeague
et al., 2009; Weerts & Lang, 1978), this pattern is not necessarily always observed in cases of
trauma and other anxiety disorders. For example, trauma researchers have observed that
individuals with PTSD often show blunted physiological responses to trauma cues (McTeague et
al., 2010), and studies of physiological reactivity in PTSD frequently find only one peripheral
measure (e.g., heart race or SCL) elevated during trauma cue exposures (e.g., Gerardi et al.,
1994; McDonagh-Coyle et al., 2001; Orr et al., 1998; Pitman et al., 1987). Recent work suggests
that individuals with a history of a single traumatic event typically show magnified responses to
trauma cues, whereas individuals with a history of multiple traumatic events often show a
blunting response, with a non-linear dose-response effect (Lanius et al., 2011; McTeague et al.,
2010). In fact, this finding has been generalized across the whole spectrum of anxiety disorders,
such that individuals with disorders marked by central fear responses to imminent threats (e.g.,
specific phobia) tended to show pronounced patterns of defensive responding across
physiological measures, while individuals with conditions characterized by more long-term,
chronic apprehension and avoidance (e.g., chronic PTSD, generalized anxiety disorder) showed
paradoxical blunting of responses consistent with the long-term effects of dysregulation of the
functioning of central nervous system structures such as the amygdala (McTeague & Lang,

111

2012). The authors noted that this pattern was observed within disorders as well, with selfreported level of distress inversely correlated with the magnitude of defensive physiological
mobilization observed at the highest levels of distress.
Thus, it may be that the lack of consistency between subjective ratings of affective
response and observed patterns within the physiological data is at least partly due to the presence
of subgroups of LGB individuals who evidence physiological blunting in response to prolonged
experiences of stress and emotional dysregulation. In the present sample, summed past stressful
life events was weakly negatively correlated with startle magnitude during bias crime clips for
LGB participants. Past bias victimization (dichotomously coded), for LGB participants, was
positively correlated with self-reported arousal but not with physiological arousal during bias
crime clips. It is possible that the relationship between past bias victimization and physiological
responses to bias crime information is a non-linear one obscured by the use of linear models, as
suggested by the previously-mentioned research (McTeague & Lang, 2012). As LGB individuals
as a group experience higher rates of victimization during childhood, adolescence, and adulthood
(D’Augelli et al., 2006; House et al., 2011; Roberts et al., 2010), this possibility is feasible.
Although the overall number of stressful life events reported by participants in the present study
did not differ, a significant proportion (nearly 80%) of the LGB sample reported at least one
episode of bias-motivated victimization in their lifetime (as compared to 25% of the heterosexual
sample). The likelihood of this scenario is further suggested by recent findings that youth raised
in environments with high rates of stigma against LGB individuals show blunted cortisol
reactivity in stress-induction tasks (Hatzenbuehler & McLaughlin, 2014). Given the small
sample size of the present study, however, it was not possible to model these relations with any
degree of confidence at this time.

112

Changes in pre- to post-task cognitions about the world were not consistent with
traumatogenic responses. Rather than demonstrating increases in negative beliefs about the
world, LGB and heterosexual individuals both showed decreases in negative beliefs from pretask to post-task assessment. Positive change in worldview (as opposed to no change) demands
explanation, which may be provided by the social-cognitive literature on positive illusions
(Taylor & Brown, 1988; Taylor & Brown, 1994). Positive illusions refer to the tendency for
humans to maintain exaggerated evaluations of performance, perceptions of mastery and control,
and unrealistic optimism about future outcomes, which have been positively linked to
psychological health. As Taylor and Brown (1994) note: “the mentally healthy person appears to
have the enviable capacity to distort reality in a direction that enhances self-esteem, maintains
beliefs in personal efficacy, and promotes an optimistic view of the future” (p. 204). These
positive illusions have been found to be generally adaptive, insofar as they are correlated with
contentment, are positively associated with elements of social bonding such as increased social
connections, and enhance motivation, persistence, and performance across tasks (Janoff-Bulman,
1989; Taylor & Brown, 1994). Belief in a just world for the self, for example, has been linked to
decreases in depression and stress, and increases in life satisfaction (Lipkus, Dalbert, & Siegler,
1996).
Notably, positive biases have been shown to become more prominent as threats to the self
are introduced (Greenwald, 1981), as seen in studies of the just-world hypothesis wherein
individuals respond to information about crimes by inflating their perceptions of control and
frequently assigning blame (believed to be due to an inflated sense of control and self-efficacy)
to the victim (Furnham, 2003; Hafer & Bègue, 2005). Of particular relevance to the present
question, a series of studies have suggested that individuals who share a social identity with a

113

crime victim are even more likely than those who do not share an identity to make efforts to
psychologically distance themselves from the threat of the situation by either derogating the
victim or distancing themselves from the targeted group (Correia et al., 2012). The same
researchers recently found that not endorsing just-world beliefs or distancing from the victim led
to low self-esteem when individuals were highly identified with the group (Ramos, Correia, &
Alves, 2014). Finally, this pattern of response has also been observed in responses to
discriminatory events, specifically (Stroebe, Dovidio, Barreto, Ellemers, & John, 2011). That
LGB individuals demonstrated a decline in negative cognitions is consistent with this process
and suggests that they responded to the news information provided as a threat. Presented with
potentially threatening information, the individuals in this sample may have engaged in an
immediate process of cognitive restructuring, inflating their positive perceptions of the safety
and benevolence of the world and others in order to defend against the negative mood states and
cognitions evoked by exposure to aversive information.
It may then be that these individuals are showing a form of coping in response to threat,
though this pattern of response is not consistent with the changes in cognitions about the world
observed in trauma survivors (Foa et al., 1999; Janoff-Bulman, 1989). Cross-sectional and
retrospective studies of cognitive appraisal in trauma have typically focused on appraisals related
to the self, such as self-blame (e.g., Foa et al., 1999) or appraisals about the severity and meaning
of symptoms of acute stress such as intrusive memories (e.g., Steil & Ehlers, 2000). Longitudinal
studies have also found links between negative appraisals of the self and symptoms and later
symptoms of posttraumatic stress (Ehlers et al., 2003; Ehring et al., 2008; Nixon & Bryant, 2005;
Mueller et al., 2008; O’Donnell, Elliott, Wolfgang, & Creamer, 2007; Regambal & Alden,

114

2009), and have also suggested that pre-trauma negative cognitions about the self may predict
PTSD symptoms following trauma exposure (Bryant & Guthrie, 2005).
Negative appraisals of the world and PTSD symptoms have also been found to be related
(albeit, typically more weakly) in correlational and cross-sectional studies (Agar et al., 2006;
Kolts, Robinson, & Tracy, 2004; Laposa & Alden, 2003), although other studies have found that
this effect was not significant when controlling for negative self-appraisals (Beck et al., 2004),
maladaptive control strategies (e.g., worry and self-punishment; Bennett, Beck, & Clap, 2009)
and other factors such as gender and depressed mood (Karl, Rabe, Zöllner, Maercker, & Stopa,
2009; Moser, Hajcak, Simons, & Foa, 2007). Longitudinal studies have also suggested that the
path between acute (shortly post-exposure) negative cognitions about the world and PTSD
symptom severity may be mediated by negative self-appraisals and have suggested that internal
perceptions of threat (e.g., the individual’s perceived lack of ability to cope) may be more
influential in the development of PTSD than are perceptions of external threat (O’Donnell et al.,
2007). The present study did not assess negative beliefs about the self in relation to vicarious
exposure to bias crime information, which limits comparisons between types of negative
cognitions and which would likely have provided richer information about the LGB individuals’
cognitive responses to such events. Regardless, although observed changes in negative
cognitions about the world suggest individuals in the present study were responding to a
threatening event, there is no evidence that these changes are consistent with changes observed
following exposure to trauma.
The final research question under consideration was whether dimensions of collective
identity moderated LGB participants’ responses to bias crime events. Small to moderate positive
correlations were observed between identity centrality and average SCL during bias crime clips,

115

pre-task PTCI, and post-task PTCI for LGB participants. Public regard did not correlate with any
outcome variable of interest. These relations between identity centrality and outcome variables
of interest were not significant in general linear models, however. Notably, significant
interactions were observed between clip condition and identity measures for heterosexual
participants, but not for LGB participants. This is contrary to expectation, given literature
indicating that more central social identity has found to moderate the severity of anxiety and
trauma symptoms in individuals who have undergone discriminatory events such as exposure to
racism (Khaylis, Waelde, & Bruce, 2007) and exposure to sexism (Eliezer, Major, & Mendes,
2010). Collective identity has been little studied in sexual minority samples, however, and the
possibility that it does not exhibit the same pattern of relations cannot be denied.
Conversely, these results may be primarily a function of methodological and sample
constraints. Of greatest concern, given how group assignment was conducted (using a continuous
scale of sexual orientation rather than self-identification), the meaning of participant responses
on this scale is unclear. The wording of the items (e.g., “In general, my sexual orientation is an
important part of my self-image”) suggests that responses to these items are most likely to reflect
a general centrality of sexual orientation identity, as intended for comparison across groups.
However, although group differences in outcome measures were observed using this
classification system, this manner of group assignment may not as clearly map onto the cognitive
processes of self-identification via explicit endorsement of group membership, as presumed by
the collective identity model employed (Sellers et al., 1998). In addition, the structure and
functioning of this scale has not been psychometrically examined with Kinsey-derived LGB
subsamples and so interpretation of subscale scores in the present study is tenuous at best. (For

116

additional discussion of sexual orientation categorization in the present study, see Strengths,
Limitations, and Future Directions, below.)
The types of discriminatory events presented as stimuli in the present study (arson,
assault with injury, murder) were selected because they were judged to be the most threatening
form of bias-motivated events and therefore believed to be most likely to evoke defensive
responding on the part of like-identified individuals. In general, these clips appeared to function
as predicted, as responses to threat condition clips (general and bias) across groups were
typically more aversive than were responses to neutral clips. Consistent with predictions, there is
evidence that LGB participants respond to bias crime events as more aversive than heterosexual
participants in both self-report and physiological measures, and LGB participants described bias
crimes as more aversive than general crime events. However, there is no support in the present
findings for the hypothesis that exposure to these events provokes a response consistent with
defensive mobilization to imminent threat consistent with the extreme fear learning of a phobia,
as assessed by physiological measures, or with trauma exposure, as assessed by changes in
traumatogenic cognitions. It is plausible that subgroup differences exist such that individuals are
likely to vary in their responses according to their personal experiences of life stress and bias
victimization. Further, the grouping method used in the present study may not be the most
sensitive for the assessment of responses to an identity threat, and none of the predicted relations
between domains of collective identity and responses to bias crime information were observed in
significance testing within the LGB sample.
Clinical and Research Implications
As this study was an early quasi-experimental examination of a model of a vicariouslyexperienced source of minority stress for LGB individuals, direct clinical implications of these

117

findings are somewhat limited. The present findings provide support for the hypothesis that LGB
individuals experience bias crime events as a source of stress unique from heterosexual
participants, and partial support for the hypothesis that this experience is consistent with
responses to threats such as are observed in responses to threatening stimuli. According to the
full theoretical model (Lannert, 2014), repeated exposure to these events may lead to
strengthening of fear associations and greater likelihood of defensive mobilization in the
presence of ambiguous and clear threats, as well as being experienced as a general stressor, with
implications for behavior by increasing the risks of avoidance and social withdrawal (Foa et al.,
1992; Peri et al., 2000). Repeated physiological activation, combined with psychological shifts in
worldview and self-evaluation, may then lead to alterations in physiological and psychological
functioning that increase the risk of trauma-related mental illnesses such as posttraumatic stress
and other anxiety disorders, mood disorders, and substance use disorders as allostatic load is
increased incrementally (Berntson & Cacioppo, 2007; McTeague et al., 2010).
This study examines the immediate proximal responses to exposure to bias crime
information in light of these potential downstream effects, although it does not purport to assess
mental health outcomes directly in light of the suggested cumulative nature of these stress
effects. Future studies may incorporate estimates of lifetime vicarious exposure to such events as
well as direct measures of psychological well-being in order to evaluate the links between these
in-vivo responses and mental health outcomes. Overall, further examination of these processes
may have implications for understanding the development of a vulnerability to anxiety symptoms
in response to social identity threat, which may inform further research and clinical practice.
Before considering how future research may build upon these findings, it is necessary to address

118

the strengths and limitations of the current study design.
Strengths, Limitations, and Future Directions
The present study has a number of strengths. First, it is the first study to examine realtime responses to bias crime information in non-victim members of a targeted group, building
and expanding upon past retrospective and qualitative studies (Craig, 1999; Noelle, 2002; Perry
& Alvi, 2012) within a quantitative approach. Across disciplines, it contributes to the
sociological, legal, and psychological literature on crime victimization by examining an oftenargued but little studied assumption of community intimidation in bias crimes (Sullaway, 2004).
It expands upon the sparse literature on LGB mental health by testing existing qualitative
findings with empirically-sound quantitative methodology. Methodologically, its design allows
for examination of complex processes using systematic data collection and minimizing
additional sources of variance. For example, by ensuring that study facilitators remained blind to
sexual orientation status, the study is able to minimize potential observer-expectancy effects. To
maximize ecological validity, stimuli consisted of news reports of bias crimes, analogous to
likely sources of exposure to bias crime information for most LGB individuals (Craig, 1999).
Clips were carefully selected to minimize possible influences of other intersecting identities.
Physiological data were collected and analyzed in accordance with the best available practices
and recommendations of the field (Blumenthal et al., 2005; Fowles et al., 1981). Finally, the
study design allowed for longitudinal assessment of variables of interest (i.e., PTCI) where
included in the pre-study questionnaire.
This study was not without limitations, however. Primary among these was a relatively
small sample size. Examination of effect size and observed power statistics suggests that the
present study was underpowered for the analysis of some of the proposed hypotheses. Namely,

119

analyses failed to achieve sufficient power (commonly identified as .80 or above for a Type I
error rate of alpha = .05; Cohen, 1992) to detect significant group differences for the majority of
statistical comparisons made, save for some observed relations (e.g., self-reported arousal and
affect valence) in which observed effect sizes were large or very large. One significant
consequence of low power is that these results cannot reasonably be generalized to a broader
population, given the increased risk that findings are a sample-specific function of the particular
composition of LGB and heterosexual participants in this study.
However, the presumption of replicable effect sizes cannot be made, given that prior
research indicates that effect sizes tend to become attenuated as sample size increases (Button et
al., 2013; Ferguson, 2009; Wilkinson, 1999). Therefore, an additional consequence of the small
sample size in the present study is that effect sizes observed in the present sample likely
overestimate those within the broader population from which this sample was derived. Similarly,
smaller samples sizes are more prone to sampling bias and may be less likely to accurately
represent the characteristics of the broader population (Button et al., 2013; Wunsch, 1986). There
is some evidence that this may be the case in the present study. For example, the present sample
did not display the same distribution of demographic characteristics (e.g., race, ethnicity) as that
of the university population or of the surrounding municipality. In fact, a larger proportion of
racial and ethnic minority individuals were observed within the LGB subsample than within the
heterosexual subsample. It is possible that the pattern of relations between variables observed
within the present sample are at least partially a function of sampling error.
Finally, small sample size was problematic in the present study insofar as it precluded the
use of some analyses (e.g., covariate models) that may further elucidate relations between the
constructs under consideration. For example, prior research has suggested that different

120

classification strategies for sexual orientation are likely to yield different, if overlapping, samples
of individuals, often with different characteristics and results (Midanik et al., 2007; Moradi et al.,
2009; Sell, 1997; Sell & Petrulio, 1996; Worthington & Reynolds, 2009). The present study
collected data on three forms of sexual orientation classification (self-identification, sexual
behavior, and the Kinsey scale). The Kinsey scale (Kinsey, 1948) was chosen as the grouping
method in this study because it allowed participants to rate themselves on a continuum based on
their perception of their life histories of attraction, fantasy, and behavior rather than solely on
self-categorization or behavior. Conceptualizing sexual orientation as a spectrum rather than as
discrete categories is more consistent with recent research on sexuality and sexual orientation
identity suggesting a continuum of orientation (Vrangalova & Savin-Williams, 2012).
The nature of the quasi-experimental design employed necessitated initial categorization
into two group for comparison, and so the original Kinsey scale cutoffs (in which individuals
reporting “more than incidental” same-sex attraction, fantasy, and behavior are categorized as
LGB) were used in accordance with convention (Kinsey, 1948; Sell, 1997) and with the
presumption that this would be strongly associated with differences in the appraisal of bias crime
events with regard to self-relevance, likelihood, prevalence, and imminence (Paterson &
Neufeld, 1987). No direct evaluation (e.g., through questioning) of individuals’ appraisals of
these events was conducted, however, and although many of the group differences observed
(e.g., in self-ratings of affect) suggest that appraisals likely differed by group, the precise nature
of the relation between the grouping method and threat appraisals is not clear. Future studies
should directly ask participants about their appraisals of these events. Similarly, the meaning of
collective identity variables assessed is unclear, given that these scales were worded to assess
general dimensions of identity (e.g., centrality of sexual orientation) to allow for comparison

121

across groups, rather than in specific reference to the individual’s own self-categorization (e.g.,
centrality of lesbian identity), which may evoke more meaning for participants. It is possible that
when asked about their specific identity, as in prior modifications of the Multidimensional
Inventory of Black Identity (Benish-Weisman & Horenczyk, 2010; Bombay et al., 2010;
Houkamau & Sibley, 2010; Quinn & Chaudoir, 2009; Rughinj, 2011; Yip & Cross, 2004),
responses may have differed.
At the time of study design, it was presumed that comparisons of different methods of
identification (and, potentially, a continuous measure of sexual orientation) might be employed;
however, given sample size constraints this was not feasible. It is very possible that different
results may have been achieved by comparing groups on the basis of another classification
system such as self-identification. Further, larger sample size may have allowed for comparison
between different combinations of methods of categorization, such as comparing the responses
of individuals who report same-sex sexual behavior but do not identify as LGB to those of
individuals who both report same-sex sexual behavior and self-identify as LGB. As concealment
of sexual identity has been linked to both stigma consciousness and negative mental health
outcomes (Frost & Bastone, 2008; Quinn & Chandoir, 2009), such an analysis would likely have
yielded rich information about the complexities of social identity, sexuality, and discrimination
that was not available in the present sample.
Similarly, an additional weakness of the present study is that it administered only some
measures prior to the laboratory exposure. This study design is therefore limited in its ability to
draw linkages between exposure to bias crime information and additional variables of interest.
For example, the collective identity measure was administered only post-exposure. Thus, the
direction of the observed correlations between identity dimensions and some outcome measures

122

for LGB participants cannot be indicated with any confidence. It is possible that participants may
demonstrate pre-task to post-task changes in collective identity domains, as observed in studies
of responses to observed discrimination in which minority individuals were more likely to
distance themselves from the targeted in-group in response to perceived threat than were
majority individuals (Correia et al., 2012; Ramos, Correia, & Alves, 2014).
The failure to administer the collective identity questionnaire pre-exposure also precluded
a closer analysis of its psychometric properties. Although the internal reliability statistics of the
sub scales of the modified MIBI/MISOI ranged from acceptable to excellent across both groups,
and scale means differed between groups in theoretically expected directions (i.e., lower public
regard was observed within the LGB subsample), it was not possible to confirm the underlying
structure of this scale in the present sample. Of particular interest would be an assessment of
invariance across groups, including construct invariance. Given that the structure of one
questionnaire proposed for use in this study (the World Assumptions Questionnaire) was not
replicated within a larger pre-task screening sample, as well as exhibiting measurement variance
between the LGB and heterosexual groups, the risk of scale structure failing to conform to
theoretical expectations among LGB individuals cannot be ignored. It is also unclear what group
participants who endorsed more-than-incidental same-sex feelings, attractions, and behavior, but
who identified as heterosexual, used as a reference point in answering items on this scale.
Further evaluation of the psychometric properties of the modified MIBI/MISOI is necessary.
Additional sources of variance that fell beyond the scope of the present study but which
deserve further consideration in future research include specific characteristics of the crime
events, including characteristics of perpetrators and victims. In the interest of limiting these
potential sources of variance to strictly test differences between the presentation of bias and non-

123

bias crime information, auditory clips were carefully selected based on stringent criteria which
attempted to exclude many of these sources of variance (e.g., deliberately selecting clips that did
not explicitly describe the race of perpetrators or victims). To that same end, news clips were
selected to address a range of potential events (assault, murder, and arson), with a range of
potential victim genders, and with a range of presenter and reporter genders. Thus, selected clips
in the bias and general threat conditions were matched according to a broad range of criteria and
are presumed to be only minimally affected by these sources of variance. However, despite
careful selection of stimuli it was not possible in the present study to eliminate all such sources
of variance, and participants may have responded somewhat differently to clips according to this
information in ways that could not be measured in the current design (e.g., by drawing
assumptions about victim or perpetrator characteristics based on information given about names
or locations of crimes). In addition, despite efforts to vary specific characteristics explicitly
stated in the clips such as gender, the limited sample size of the present study prevented in-depth
analysis of these differences at the level of individual clips. Given the complexity of social
identity and the likelihood that individual differences in characteristics not assessed here-in play
a role in individual responses to bias crime events, future studies designed to carefully parse
these differences are advised.
Of particular interest would be: examining the intersectionality of identities (e.g., sexual
orientation, racial, economic, and gender) in vicarious responses to bias crimes, especially given
evidence that perceptions of event severity for personally-experienced LGBT bias-motivated
events may vary according to race and socioeconomic status (Meyer, 2010). Individual
differences in addition to social identity (e.g., neuroticism), should also be evaluated, given
observed small to large correlations between neuroticism and many of the outcome variables

124

examined. Similarly, given the high rates of past bias victimization endorsed by the LGB
participants in this sample, additional studies assessing a range of types and severity of bias
victimization within this context would be critical for parsing apart responses due to cued
activation of existing fear structures in individuals with significant trauma histories (Lanius et
al., 2011; McTeague et al., 2010; Pitman et al., 1987; Pole, 2007), versus vicarious activation of
threat response in individuals without significant trauma history. Finally, although distal mental
health outcomes were not examined in the present study, they are proposed to be associated with
the effects of repeated exposure to information about bias crime events (Lannert, 2014). As such,
future studies could examine correlations between cumulative exposure to such information (e.g.,
through participants’ reports of how often they have heard of these events) and symptoms of
psychological distress in order to begin disentangling those relations.
Conclusions
In conclusion, this thesis presents a novel and carefully-planned study designed to
examine how heterosexual and LGB individuals respond to information about bias crimes, and
whether these responses constitute a traumatogenic response. This study was partially successful
in achieving its aims. Present results suggest that LGB individuals likely experience bias crimes
differently from heterosexual individuals, and that their experience involves some degree of
defensive mobilization against threat. There is no evidence that these responses are consistent
with a traumatogenic response based on the present findings, however, and links between the
responses observed in the present study and long-term mental health outcomes are unclear.
Limitations of the present study, including small sample size, methods of group categorization,
and the absence of some measures at the pre-task assessment, precluded more in-depth
assessment of the data such as comparison using different grouping methods. These results

125

tentatively suggest, however, that some response process unique to individuals who endorse
more than incidental same-sex fantasy, attraction, and behavior is occurring when exposed to
bias crime information, and that this process involves both central and peripheral nervous system
stress responses believed to be linked to distal mental health outcomes over time. Future research
is strongly suggested in order to more fully explicate these findings and to more closely examine
the clinical implications of bias-motivated violence and discrimination against sexual minorities.

126

APPENDICES

127

APPENDIX A
Participant Consent Form
PARTICIPANT INFORMATION AND CONSENT FORM
Michigan State University
Department of Psychology
1.

What is the purpose of the research? You are being asked to participate in a research

study of individual differences in responses to different types of news stories. We hope to learn
how individuals respond to different types of news information, including physical, emotional,
and cognitive reactions. We are also interested in how individual differences such as identity,
personality, sexual orientation, and past life experiences affect responses to different types of
news information. You must be 18 years of age to participate in this study. If you choose to
participate in the study, your participation will take about 1 hour to 1.5 hours.
2.

What will I do as a participant? If you consent to participate in the study, we will

attach sensors to the fingers of your left hand, under your left eye, and your forehead using a
mild adhesive and a small amount of water-based conductive gel. These sensors are not
dangerous and should not be painful or uncomfortable in any way. Once the sensors are attached,
we will ask you to complete a computer task during which you will wear headphones and be
presented with different auditory clips of news stories. We will ask you to focus on the stories
and to imagine the events of the stories as they are described. As you listen, we will record your
physical responses (muscle activity and nervous system arousal) through the sensors placed on
your skin. At different times during the news clips you will also hear brief and random bursts of
white noise. After you listen to each news clip, you will answer some questions about your
feelings and thoughts during the clip before listening to the next news clip. After you have heard

128

all of the news clips, we will remove the sensors and you will complete questionnaires on
demographic information, identity, personality, beliefs about the world, stressful past life
experiences, and negative thoughts about stressful experiences.
3.

What are the possible benefits of participating? You will not directly benefit from

participation in this study. However, your participation in this study may contribute to the
understanding of social group and individual differences in responses to different types of news
information, and so may have benefits for society as a whole.
4.

What are the possible risks of participating? The risks of participating in this study are

minimal and include some possible temporary anxiety or discomfort from exposure to news
information that you may find difficult or emotional, such as news information about violent
crimes. You may also experience some possible mild discomfort answering questionnaires about
sensitive topics such as stressful life experiences. If you experience more than mild or transient
discomfort during this study, we can provide you with information about counseling resources
that are available to you as an MSU student.
5.

How will you protect my privacy and confidentiality? Your responses in this study

will remain entirely anonymous. We will assign you an identification number and will not collect
or store personal information such as your name. The study staff will not directly enter
attendance in HPR or in any other way be exposed to any of your identifying information. All
data will be electronic and will be stored on a secure server within the Psychology Department,
marked by an ID number, and password-protected. Only trained study staff will have access to
the data. The results of this study may be published or presented at professional meetings, but the
identities of all research participants will remain anonymous. No member of the study staff will
be able to link your responses to identifying information.

129

6.

Can I say no, choose not to answer questions, or stop participating if I want?

Participation in all or part of this study is entirely voluntary. You can choose not to participate in
the study, can participate in all parts of the study, or participate in just some parts of the study.
You have the right to withdraw from this study at any point if you change your mind, with no
penalty or negative consequences. Your decision about whether to participate or not will not
affect your relationship with Michigan State University. If you choose to withdraw from the
study early, you may receive only partial HPR credit for your time. Whether you choose to
participate or not will have no effect on your grade.
7.

How will I be compensated for my participation? For completing the study, you will

receive 3 HPR credits. These credits may be used to fulfill course requirements or as extra credit,
depending on the requirements set by your course instructor. If you would rather complete an
alternative assignment than participate in research studies, please consult your instructor.
In addition, upon completion of this study your HPR Id will be entered into a raffle for the
opportunity to win an iPod shuffle (value: $49). At the end of the semester, the HPR Ids of all
participants who completed this study will be entered into an Excel spreadsheet. Random
numbers will be generated for each HPR ID and compared to a single, separate, randomlygenerated number. The participant whose randomly-generated ID number is closest to the
randomly-generated second number will be contacted by Leslie Baldwin (HPR Administrator)
and asked to pick up their iPod shuffle. At no time during the raffle process will study staff see
your identifying information; only HPR Ids will be used during this process.
Only 1 iPod shuffle will be awarded by raffle. The raffle will be conducted following the closure
of the HPR system at the end of the semester. Each participant is estimated to have a 1 in 120
chance of winning the raffle.

130

8.

What if I have further questions or concerns? If you have concerns or questions about

this study, such as scientific issues, how to do any part of it, or to report an injury, please contact
the researchers:
Brittany Lannert, M.A.

Dr. Alytia Levendosky

316 Physics Rd., Room 44
East Lansing, MI

316 Physics Rd., Room 107C

48824

East Lansing, MI

Phone: (918) 857-7024

48824

Phone: (517) 353-6396

Email: lannertb@msu.edu

Email: levendo1@msu.edu

If you have questions or concerns about your role and rights as a research participant,
would like to obtain information or offer input, or would like to register a complain about this
study, you may contact, anonymously if you wish, the Michigan State University’s Human
Research Protection Program at (517) 355-2180, Fax (517) 432-4503, or email irb@msu.edu or
regular mail at 207 Olds Hall, MSU, East Lansing, MI 48824.
Your participation today does not obligate you to participate in any future studies. You
will be given a copy of this form to keep.
If you choose not to consent to participate in this study, you may leave at this time. If you
choose to consent to participation in this study, your decision to begin the study indicates your
consent. You may choose to withdraw from the study at any time without penalty, but may not
receive full HPR credit for your participation and may not be entered into the raffle.

131

APPENDIX B
Copies of Measures
Global Screening Sexual Orientation Items
Lab Protocol
SAM Subjective Task Ratings
Demographics Questionnaire
Post-Traumatic Cognitions Inventory – Negative Beliefs about the World Subscale
Sexual Orientation Identity Questionnaire
World Assumptions Questionnaire
Multidimensional Inventory of Sexual Orientation Identity
IP-IP Neuroticism Scale
Post-Traumatic Cognitions Inventory
Past Victimization Experiences Scale

132

Global Screening Questionnaire (HPR)
1. Are you age 18 or older?
_____ Yes
_____ No
2. What is your gender?
_____ Male
_____ Female
_____ Other
3. Based on your reactions and experiences over your entire life, how would you describe
your sexual orientation?
_____ Lesbian
_____ Gay
_____ Bisexual
_____ Heterosexual
_____ Queer
_____ Asexual
_____ Other
4. Based on your history of sexual experiences (kissing, fondling, intercourse), how would
you describe the gender of your sexual partners?
_____ All men
_____ Mostly men, some women
_____ Equal numbers of men and women
_____ Mostly women, some men

133

_____ Equal numbers of men and women
_____ All women
_____ I have not had any sexual experiences
5. Considering your history of attraction and fantasy across your entire life, how would you
describe your sexual orientation?
_____ Exclusively heterosexual
_____ Mostly heterosexual, only incidentally homosexual
_____ Mostly heterosexual, but more than incidentally homosexual
_____ Equally heterosexual and homosexual
_____ Mostly homosexual, but more than incidentally heterosexual
_____ Mostly homosexual, only incidentally heterosexual
_____ Exclusively homosexual
_____ Non-sexual/Asexual

134

Lab Protocol
(Portions in quotes “” are to be read to participants. Italicized segments are instructions for
research staff.)
Greet participants and review information about the study provided in the consent form. Answer
any participant questions and ensure that they understand the information contained in the
consent form. If they choose to participate in the study, ask them to sign and date the consent
form. Provide the participant with a blank copy of the consent form to take with them.
“Thank you for coming today and for agreeing to participate in this study. Today, we will
ask you to complete two tasks. First, we will ask you to complete a computer task where you will
listen to clips of news reports while you imagine the events described by the clips in as much
detail as possible. At random times during each clip, you will hear a short burst of white noise.
After each news report, you will rate how you felt while listening to the clip using the computer.
The listening task should take approximately 35-40 minutes. After you have listened to all of the
audio clips, you will be asked to complete a series of questionnaires on the computer.”
“During the listening task, we will be recording your muscle activity and nervous system
functioning using six small sensors that we will attach to your skin with a mild adhesive.” Show
the participant the electrodes and the adhesives. “We will also apply a small amount of this
water-based gel to help conduct the signal.” Show the participant the conductive gel. “You
should not feel any discomfort from the sensors, and most people report that they do not notice
them after a short time. The gel that we use is like a thick saltwater and should not cause any
irritation to your skin. We will attach one sensor to the index finger of your left hand, one sensor
to the middle finger of your left hand, two sensors below your left eye, and two sensors on your
forehead. Do you have any questions about this?”

135

Answer any questions about the physiological recording. When the participant indicates
that they are comfortable with the equipment, attach the electrodes in the specified locations,
making sure to perform a gentle debriding of the electrode sites with a cotton swab prior to
application. As you attach the electrodes, explain to the participant what you are doing (“Okay,
now I’m going to put a sensor on your left index finger”) to help them remain comfortable
during set-up. When the electrodes are attached, check for signal quality and make any
adjustments as necessary. When recording is of sufficient quality, start the computer program.
Show the participant the headphones.
“During the listening task, you will be wearing these headphones. For now, we will keep
the headphones off while we explain your task. You will be asked to listen to nine audio clips of
news reports that are similar to the sort of news story that you might hear on your local evening
news. While listening to each clip, we would like you to try and imagine the events described in
as much detail as possible. Please listen carefully, because after each clip you will be asked to
answer a few questions about the information provided in the clip. You will also be asked to rate
how you felt as you listened after each clip.”
Display the SAM scales.
“Look at these set of figures. Each of these figures is arranged along a continuum. We
call this set of figures SAM, and you will be using these figures to rate how you felt while
listening to each clip. You will use all three scales – making three ratings – for each auditory
clip. SAM shows three different kinds of feelings: Happy vs. Unhappy, Excited vs. Calm, and
Controlled vs. In Control.”
Display the first SAM figure. Point to the appropriate parts of the scale and the computer
keyboard as you read the following description.

136

“You can see that each SAM figure varies along each scale, and each SAM figure has a
number underneath it ranging from 1 to 9. The first SAM scale is the happy-unhappy scale,
which ranges from a smile to a frown. At one extreme of the happy vs. unhappy scale, you felt
happy, pleased, satisfied, contented, hopeful. If you felt completely happy while listening to the
clip, you can indicate this by pressing a ‘9’ on the keyboard when you see this scale. The other
end of the scale is when you felt complete unhappy, annoyed, unsatisfied, melancholic,
despaired, bored. You can indicate feeling completely unhappy by pressing a ‘1’ on the keyboard
when you see this scale. If you felt completely neutral, neither happy nor sad, select the figure in
the middle by pressing a ‘5’ on the keyboard. This figure also allows you to describe
intermediate feelings of pleasure, by choosing any of the other numbers. This allows you to
make more finely graded ratings of how you felt while listening to the clips.”
Display the second SAM figure. Point to the appropriate parts of the scale and the
computer keyboard as you read the following description.
“The excited vs. calm dimension is the second type of feeling displayed here. At one
extreme of the scale you felt stimulated, excited, frenzied, jittery, wide-awake, aroused. If you
felt completely aroused while listening to the clip, press a ‘9’ on the keyboard when you see this
scale. The other end of the scale is when you felt completely relaxed, calm, sluggish, dull,
sleepy, unaroused. You can indicate you felt completely calm by choosing a ‘1’ on the keyboard
when you see this scale. As with the happy-unhappy scale, you can represent intermediate levels
by selecting any of the other figures. If you are not at all excited nor at all calm, press ‘5’ on the
keyboard to select the figure in the middle of the row.”
Display the third SAM figure. Point to the appropriate parts of the scale and the
computer keyboard as you read the following description.

137

The last SAM scale that you will rate is the dimension of controlled vs. in control. At one
end of the scale you have feelings characterized as completely controlled, influenced, cared-for,
awed, submissive, guided. Please indicate feeling controlled by pressing a ‘1’ on the keyboard
when you see this figure. At the other extreme of this scale, you felt completely controlling,
influential, in control, important, dominant, autonomous. You can indicate you felt dominant by
pressing a ‘9’ on the keyboard. Note that when the figure is large, you feel important and
influential, and that it will be very small when you feel controlled and guided. If you feel neither
in control nor controlled you should press ‘5’ to select the middle figure. Remember, you can
also represent your feelings between these endpoints by pressing any other number on the scale.”
Click the space bar to progress to the WARMTH scale.
“Finally, you will be asked to rate how you feel about other people using this number
scale. We call this scale the ‘warmth’ scale. At one end of the scale you have feelings
characterized as extremely cold, uninterested in others, disengaged, withdrawn, and flat. Please
indicate feeling extremely cold toward others by pressing the “1” on the scale. At the other end
of the scale you have feelings characterized as extremely warm, interested in others, engaged,
connected, and affiliative. When you feel extremely warm toward others, please press the “9” on
the scale. As with the SAM scales, if you feel neither warm nor cold toward others press the “5”
in the middle of the scale. You can also represent your feelings between these points by pressing
the other numbers on the scale.”
“Some of the news clips may prompt emotional experiences; others may seem relatively
neutral. Your rating of each clip should reflect your immediate personal experience – that is, how
you felt in the moment - and no more. Please rate each one as you actually felt while listening to
the clip.”

138

Cue the computer program to the start screen.
“In a minute, we will ask you to put on the headphones and we will leave the room while
you complete the task. Remember, while listening to the clips please do your best to imagine the
events described and pay close attention to the information that you are given by each clip. If at
any time you wish to stop the study, please knock on the mirror on this wall and we will
discontinue your participation. If we need you to stop momentarily so that we can check the
equipment, we will knock on the mirror. When you have listened to all of the clips and rated
your responses for each, you will see a screen telling you that you have completed this portion of
the experiment. At that time, we will get you started on the questionnaires. Are there any
questions before we begin?”
Answer participant questions, if any.
“Okay, once we have left the room, please put on the headphones. You may begin the
task at any time by pressing the space bar.”
Leave the room and monitor participant performance and physiological recordings from
the adjoining room. When the participant completes the listening task, re-enter the testing room
and proceed to the questionnaires.
“Please complete the questionnaires as accurately and completely as possible. If you have
any questions about any of the items, please let us know.”
When they have completed the questionnaires, thank the participant for their time and
provide them with a copy of the consent form (with contact information) to take with them.

139

SAM Ratings (Lab, Computer Task)
Figure 7.
Self-Assessment Manikin (SAM) Visual Figures and Rating Scales for Self-Report Ratings of
Affective Arousal, Valence, Dominance, and Interpersonal Warmth.

1

2

3

4

5

6

7

8

Less Happy

1

More Happy

2

3

4

5

6

7

8

Less Aroused/Excited

1

9

2

9

More Aroused/Excited

3

4

5

6

7

Less Dominant/In Control

8

9

More Dominant/In Control

WRM
1

2

3

4

5

Less Warm/Connected

6

7

8

9

More Warm/Connected

140

Demographics Questionnaire (Lab)
1) Age: ___________ years
2) Sex: ____M

____F

3) Gender identification: _____ Male
_____ Female
_____ Transgender
_____ Other
4) Current year at MSU: _____ Freshman
_____ Sophomore
_____ Junior
_____ Senior
_____ Graduate Student
5) Prior to enrolling at MSU, what is the highest level of education you had completed?
_____ Some high school
_____ High school/GED
_____ Associate’s degree
_____ Bachelor’s degree
_____ Master’s degree
_____ Doctorate
_____ Other professional degree
Please specify: ______________________
6) What is your family’s average annual income? Please estimate:
_____ Less than $20,000

141

_____ $20,000 - $39,999
_____ $40,000 - $59,999
_____ $60,000 - $79,999
_____ $80,000 - $99,999
_____ $100,000 - $119,999
_____ More than $120,000
7) Would you describe yourself as Hispanic or Latino/a?
_____ Yes, I am Hispanic or Latino/a
_____ No, I am not Hispanic or Latino/a
8) How would you describe yourself?
_____ American Indian/Native American
_____ Asian/Asian American/Pacific Islander
_____ Black/African American
_____ Latino/Hispanic/Chicano
_____ Biracial/Multiracial
Please specify: _____________________
_____ White/Caucasian
_____ Other
Please specify: _____________________
9) How would you describe the place where you grew up (spent the greatest amount of time
during your childhood and adolescence)?
_____ A rural area
_____ A small town

142

_____ A suburban area
_____ A small city
_____ A City
_____ Other:
_____________________________
10) If you grew up within the United States, how would you describe the region of the
country where you grew up (spent the greatest amount of time during your childhood and
adolescence)?
_____ New England (Maine, New Hampshire, Vermont,
Massachusetts, Rhode Island, Connecticut)
_____ Mid-Atlantic (New York, Pennsylvania, New Jersey)
_____ East North Central Midwest (Wisconsin, Michigan, Illinois,
Indiana, Ohio)
_____ West North Central Midwest (Missouri, North Dakota,
South Dakota, Nebraska, Kansas, Minnesota, Iowa)
_____ South Atlantic (Delaware, Maryland, District of Columbia,
Virginia, West Virginia, North Carolina, South Carolina,
Georgia, Florida)
_____ East South Central (Kentucky, Tennessee, Mississippi,
Alabama)
_____ West South Central (Oklahoma, Texas, Arkansas,
Louisiana)

143

_____ West Mountain Region (Idaho, Montana, Wyoming,
Nevada, Utah, Colorado, Arizona, New Mexico)
_____ Pacific West (Alaska, Washington, Oregon, California,
Hawaii)
_____ I did not grow up within the United States
11) What is your religious affiliation?
_____ Protestant Christian
_____ Roman Catholic
_____ Evangelical Christian
_____ Jewish
_____ Muslim
______ Atheist/Agnostic
______ Other
12) How important would you say your religion is in your life?
1

2

3

4

Not at all important

5

6

7

Extremely Important

144

WAQ
Please rate the following statements on how much you agree or disagree with them using the
following scale:
1 = Strongly Agree
2 = Agree
3 = Slightly Agree
4 = Slightly Disagree
5 = Disagree
6 = Strongly Disagree
_____ 1. Most people can be trusted.
_____ 2. I don’t feel in control of the events that happen to me.
_____ 3. You usually can know what is going to happen in your life.
_____ 4. It is difficult for me to take most of what people say at face-value.
_____ 5. It is very difficult to know what others are thinking.
_____ 6. Anyone can experience a very bad event.
_____ 7. People often behave in unpredictable ways.
_____ 8. People are less safe than they usually realize
_____ 9. For the most part, I believe people are good.
_____ 10. I have a great deal of control over what will happen in my life.
_____ 11. You never know what’s going to happen tomorrow.
_____ 12. Other people are usually trustworthy.
_____ 13. People’s lives are very fragile.
_____ 14. It is hard to know exactly what motivates another person.

145

_____ 15. Most people cannot be trusted.
_____ 16. People fool themselves into feeling safe.
_____ 17. It is hard to understand why people do what they do.
_____ 18. Most of what happens to me happens because I choose it.
_____ 19. Terrible things might happen to me.
_____ 20. It is ultimately up to me to determine how events in my life will happen.
_____ 21. It can be very difficult to predict other people’s behavior.
_____ 22. What people say and what they do are often very different things.

146

PTCI-WBS
Please read each statement carefully and tell us how much you AGREE or DISAGREE with
each statement, using the scale below. There are no right or wrong answers to these statements.
Just tell us how you feel right now.
1

Totally Disagree

2

Disagree Very Much

3

Disagree Slightly

4

Neutral

5

Agree Slightly

6

Agree Very Much

7

Totally Agree

_____ 1. People can’t be trusted.
_____ 2. I have to be on guard all the time.
_____ 3. You can never know who will harm you.
_____ 4. I have to be especially careful because you never know what can happen next.
_____ 5. The world is a dangerous place.
_____ 6. I can’t rely on other people.
_____ 7. People are not what they seem.

147

MISOI (Lab)
The following are a series of statements about possible thoughts, feelings, and beliefs that you
might have about being a heterosexual, gay, lesbian, or bisexual person, and about people who
have the same sexual orientation as you do. Please indicate your agreement or disagreement with
these statements using the scale on the right.
Strongly Disagree

Neutral

Strongly Agree

1. Overall, my sexual orientation
has very little to do with how I

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

feel about myself.
2. I feel good about other people
who share my sexual orientation.
3. Overall, people of my sexual
orientation are considered good by
others.
4. In general, my sexual
orientation is an important part of
my self-image.
5. I am happy that I am of my
sexual orientation.
6. I feel that people who share my
sexual orientation have made

148

major accomplishments and
advancements.
7. My destiny is tied to the destiny

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

of other people who share my
sexual orientation.
8. My sexual orientation is
unimportant to my sense of what
kind of person I am.
9. In general, others respect
people of my sexual orientation.
10. Most people consider people
of my sexual orientation, on
average, to be more ineffective
than other people.
11. I have a strong sense of
belonging to people who share my
sexual orientation.
12. I often regret my sexual
orientation.
13. I have a strong attachment to
other people of my sexual
orientation.

149

14. My sexual orientation is an

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

important reflection of who I am.
15. My sexual orientation is not a
major factor in my social
relationships.
16. People of my sexual
orientation are not respected by
the broader society.
17. In general, other groups view
people of my sexual orientation in
a positive manner.
18. I am proud of my sexual
orientation.
19. I feel that the community of
people with my sexual orientation
has made valuable contributions
to this society.
20. Society views people of my
sexual orientation as an asset.

150

IP-IP-N (Lab)
Please rate how well each of the following statements describes you using the scale below:
1 = Very Inaccurate
2 = Moderately Inaccurate
3 = Neither Inaccurate nor Accurate
4 = Moderately Accurate
5 = Very Accurate
_____ 1. I often feel blue.
_____ 2. I dislike myself.
_____ 3. I am often down in the dumps.
_____ 4. I seldom feel blue.
_____ 5. I have frequent mood swings.
_____ 6. I feel comfortable with myself.
_____ 7. I rarely get irritated.
_____ 8. I am not easily bothered by things.
_____ 9. I panic easily.
_____10. I am filled with doubts about things.
_____ 11. I am very pleased with myself.
_____ 12. I feel threatened easily.

151

_____ 13. I am relaxed most of the time.
_____ 14. I seldom get mad.
_____ 15. I am not easily frustrated.
_____ 16. I get stressed out easily.
_____ 17. I fear for the worst.
_____ 18. I remain calm under pressure.
_____ 19. I rarely lose my composure.
_____ 20. I worry about things.

152

PTCI (Lab)
We are interested in the kind of thoughts which you may have had after a stressful or
traumatic life experience. Stressful or traumatic life experiences may include events in which
you were at risk of serious bodily injury or death, events in which you witnessed someone at risk
of serious bodily injury or death, the sudden and unexpected loss of someone close to you, times
when you felt threatened by a person or situation, sudden or dramatic life changes, or other
experiences that you found highly stressful.
Please select the life experience which you found most stressful and which still bothers
you, and complete the survey in reference to that experience. When you are done with the
survey, please turn the last page over and write a brief description (2-3 sentences is okay) of the
experience.
Please read each statement carefully and tell us how much you AGREE or DISAGREE
with each statement. People react to these events in many different ways. There are no right or
wrong answers to these statements.
1

Totally Disagree

2

Disagree Very Much

3

Disagree Slightly

4

Neutral

5

Agree Slightly

6

Agree Very Much

7

Totally Agree

_____ 1. The event happened because of the way I acted.
_____ 2. I can’t trust that I will do the right thing.
153

_____ 3. I am a weak person.
_____ 4. I will not be able to control my anger and will do something terrible.
_____ 5. I can’t deal with even the slightest upset.
_____ 6. I used to be a happy person but now I am always miserable.
_____ 7. People can’t be trusted.
_____ 8. I have to be on guard all the time.
_____ 9. I feel dead inside.
_____ 10. You can never know who will harm you.
_____ 11. I have to be especially careful because you never know what can happen next.
_____ 12. I am inadequate.
_____ 13. I will not be able to control my emotions, and something terrible will happen.
_____ 14. If I think about the event, I will not be able to handle it.
_____ 15. The event happened to me because of the sort of person I am.
_____ 16. My reactions since the event mean that I am going crazy.
_____ 17. I will never be able to feel normal emotions again.
_____ 18. The world is a dangerous place.
_____ 19. Somebody else would have stopped the event from happening.

154

_____ 20. I have permanently changed for the worse.
_____ 21. I feel like an object, not a person.
_____ 22. Somebody else would not have gotten into this situation.
_____ 23. I can’t rely on other people.
_____ 24. I feel isolated and set apart from others.
_____ 25. I have no future.
_____ 26. I can’t stop bad things from happening to me.
_____ 27. People are not what they seem.
_____ 28. My life has been destroyed by the trauma.
_____ 29. There is something wrong with me as a person.
_____ 30. My reactions since the event show that I am a lousy coper.
_____ 31. There is something about me that made the event happen.
_____ 32. I will not be able to tolerate my thoughts about the event, and I will fall apart.
_____ 33. I feel like I don’t know myself anymore.
_____ 34. You never know when something terrible will happen.
_____ 35. I can’t rely on myself.
_____ 36. Nothing good can happen to me anymore.

155

PVQ (Lab)
For the questions below, please consider how many times you have had the following
experiences in your life. If you are not sure, please give your best estimate. Please answer using
numbers rather than phrases; that is to say, please estimate the number (0, 3, 4, 10, etc.) rather
than saying “a few,” “many,” etc.
1. How many times have you been verbally insulted (yelled at, mocked, criticized) because
you were, or were thought to be, lesbian, gay, or bisexual?
Number: ___________
How many times has this occurred in the past year?
Number: ___________
How many times has this occurred in the past month?
Number: ___________
2. How many times have you been threatened with physical violence because you were, or
were thought to be, lesbian, gay, or bisexual?
Number: ___________
How many times has this occurred in the past year?
Number: ___________
How many times has this occurred in the past month?
Number: ___________
3. How many times have you had an object thrown at you because you were, or were
thought to be, lesbian, gay, or bisexual?
Number: ___________
How many times has this occurred in the past year?

156

Number: ___________
How many times has this occurred in the past month?
Number: ___________
4. How many times have you been punched, kicked, or beaten because you were, or were
thought to be, lesbian, gay, or bisexual?
Number: ___________
How many times has this occurred in the past year?
Number: ___________
How many times has this occurred in the past month?
Number: ___________
5. How many times have you been threatened with a knife, gun, or other weapon
because you are, or were thought to be, lesbian, gay, or bisexual?
Number: ___________
How many times has this occurred in the past year?
Number: ___________
How many times has this occurred in the past month?
Number: ___________
6. How many times have you been attacked sexually (forced to have a sexual experience,
raped) because you are, or were thought to be, lesbian, gay, or bisexual?
Number: ___________
How many times has this occurred in the past year?
Number: ___________
How many times has this occurred in the past month?

157

Number: ___________

158

APPENDIX C
Figures for Confirmatory Analysis of the Structure of the WAQ
Figure 8.
CFA model for the World Assumptions Questionnaire

159

Figure 9.
Standardized Estimates for the World Assumptions Questionnaire within the Full Sample (N =
60)

160

Figure 10.
Standardized Estimates for the World Assumptions Questionnaire within the LGB Sample (N =
28)

161

Figure 11.
Standardized Estimates for the World Assumptions Questionnaire within the Heterosexual
Sample (N = 32)

162

APPENDIX D
Supplemental Tables for Main Variables of Interest
Descriptive Statistics
Table 16.
Means, Standard Deviations, and Ranges of Values for Skin Conductance Level (SCL), by Group
(N = 60)
Heterosexual
Variable

M (SD)

LGB

Range

M (SD)

Range

Neutral 51.15 (7.59)

38.73-61.55

46.24 (7.94)

38.46-61.43

General Threat 46.68 (7.98)

38.48-61.48

53.33 (7.91)

38.60-61.49

Bias Threat 52.17 (8.22)

38.47-61.52

50.42 (7.45)

38.86-61.49

Average SCL

Table 17.
Means, Standard Deviations, and Ranges of Values for Startle Reflex EMG, by Group (N = 57)

Variable

Heterosexual

LGB

(N = 30)

(N = 27)

M (SD)

Range

M (SD)

Range

Inter-trial Interval 53.90 (9.28)

37.33-64.77

51.26 (9.50)

37.29-64.88

Neutral 48.97 (7.10)

38.53-63.85

47.78 (6.56)

37.18-63.11

General Threat 50.95 (8.79)

36.60-63.59

50.60 (8.09)

38.24-64.96

Bias Threat 46.18 (7.97)

35.32-62.63

50.35 (10.26)

35.48-64.98

Startle EMG

163

Table 18.
Means, Standard Deviations, and Observed Ranges of Values for Subjective Affective Responses,
by Group (N = 60)

Variable

Heterosexual

LGB

(N = 32)

(N = 28)

M (SD)

Range

M (SD)

Range

Arousal

3.06 (1.44)

1.00-6.40

3.56 (1.32)

1.40-6.00

Valence

5.79 (1.13)

2.20-8.60

5.47 (.68)

3.00-7.00

Dominance

4.96 (1.97)

1.20-9.00

5.09 (1.11)

1.20-7.40

Warmth

5.83 (1.15)

2.00-7.40

5.61 (.87)

3.40-7.40

Arousal

4.65 (1.58)

1.33-6.67

5.02 (1.13)

2.33-7.33

Valence

3.74 (1.30)

1.33-7.00

3.56 (.75)

2.00-5.00

Dominance

4.65 (1.46)

1.67-9.00

4.56 (.87)

2.33-6.00

Warmth

4.42 (1.22)

1.67-6.33

4.32 (1.27)

2.00-7.33

Arousal

4.97 (1.96)

1.00-8.00

6.29 (1.38)

3.33-8.33

Valence

3.26 (1.43)

1.00-7.00

2.30 (.80)

1.00-4.00

Dominance

4.63 (1.66)

1.33-9.00

3.74 (1.37)

1.67-6.33

Warmth

4.49 (1.60)

1.00-8.00

3.80 (1.99)

1.33-9.00

Neutral

General Threat

Bias Threat

Table 19.
Means, Standard Deviations, and Observed Ranges of Values for Pre- and Post-Task PTCI
Negative Cognitions about the World, by Group (N = 60)
Heterosexual
Variable

M (SD)

Range
164

LGB
M (SD)

Range

Table 19 (cont’d)
Neg. Cognitions
Pre-Task

4.29 (0.70)

3.00-5.29

4.63 (0.77)

3.43-6.71

Post-Task

4.03 (0.89)

1.86-5.71

4.25 (0.78)

2.29-5.43

Table 20.
Means, Standard Deviations, and Observed Ranges of Values for Sexual Orientation Centrality,
Private Regard, and Public Regard (N = 60)
Heterosexual
Variable

LGB

M (SD)

Range

M (SD)

Range

t(p)

d

Centrality

3.81 (0.96)

2.00-5.38

3.50 (1.08)

1.00-5.50

1.18 (.24)

.31

Private Regard

5.56 (1.01)

3.00-7.00

5.15 (0.97)

3.17-7.00

1.59 (.12)

.41

Public Regard

5.64 (1.01)

2.67-7.00

3.64 (1.26)

2.00-6.50

6.83 (<.01)

1.77

Table 21.
Correlations between Average Skin Conductance Level (SCL) and Startle Reflex EMG across
Clip Conditions in the LGB Subsample (in Plain Text; N = 27) and Heterosexual Subsample (in
Bold; N = 30)
1

2

3

4

5

6

-

-.56*

-.47*

.25

.04

-.08

2. GT SCL

-.44*

-

-.47*

.02

.15

.04

3. Bias SCL

-.49*

-.56*

-

-.28

-.21

.04

4. Neutral Startle

.04

.00

-.04

-

-.02

-.23

5. GT Startle

.08

-.29

.20

-.06

-

-.55*

6. Bias Startle

-.14

.14

-.01

-.43*

-.35

-

1. Neutral SCL

*p < .05
165

Table 22.
Correlations between Average Skin Conductance Level (SCL) and Subjective Ratings of Arousal
across Clip Conditions in the LGB Subsample (in Plain Text; N = 28) and Heterosexual
Subsample (in Bold; N = 32)
1

2

3

4

5

6

-

-.56*

-.47*

.33

-.25

-.46*

2. GT SCL

-.44*

-

-.47*

-.36

.08

.15

3. Bias SCL

-.49*

-.56*

-

.03

.17

.33

4. Neutral Arousal

-.24

-.06

.27

-

.33

.18

5. GT Arousal

-.03

-.15

.18

.49*

-

.41*

6. Bias Arousal

.18

.01

-.17

.32

.63*

-

1. Neutral SCL

*p < .05
Table 23.
Correlations between Average Skin Conductance Level (SCL) and Subjective Ratings of Affect
Valence across Clip Conditions in the LGB Subsample (in Plain Text; N = 28) and Heterosexual
Subsample (in Bold; N = 32)
1

2

3

4

5

6

-

-.56*

-.47*

.22

.44*

.28

2. GT SCL

-.44*

-

-.47*

-.08

-.22

.07

3. Bias SCL

-.49*

-.56*

-

-.15

-.24

-.37

4. Neutral Valence

.02

.04

-.06

-

.47*

.17

5. GT Valence

-.25

.17

.07

.01

-

.58*

6. Bias Valence

-.18

.04

.12

-.02

.82*

-

1. Neutral SCL

*p < .05

166

Table 24.
Correlations between Average Skin Conductance Level (SCL) and Subjective Ratings of
Dominance across Clip Conditions in the LGB Subsample (in Plain Text; N = 28) and
Heterosexual Subsample (in Bold; N = 32)
1

2

3

4

5

6

-

-.56*

-.47*

.13

.13

.07

2. GT SCL

-.44*

-

-.47*

-.20

-.05

.35

3. Bias SCL

-.49*

-.56*

-

.07

-.09

-.44*

4. Neutral Dominance

.10

.20

-.29

-

.35

-.27

5. GT Dominance

.14

.01

-.14

.45*

-

.37

6. Bias Dominance

.20

-.12

-.08

.39*

.84*

-

1. Neutral SCL

*p < .05
Table 25.
Correlations between Average Skin Conductance Level (SCL) and Subjective Ratings of
Interpersonal Warmth across Clip Conditions in the LGB Subsample (in Plain Text; N = 28) and
Heterosexual Subsample (in Bold; N = 32)
1

2

3

4

5

6

-

-.56*

-.47*

.28

.17

.13

2. GT SCL

-.44*

-

-.47*

-.05

-.08

.05

3. Bias SCL

-.49*

-.56*

-

-.25

-.10

-.19

4. Neutral Warmth

.08

-.01

-.06

-

.04

-.23

5. GT Warmth

-.23

.15

.07

-.09

-

.78*

6. Bias Warmth

.02

-.13

.11

-.05

.53*

-

1. Neutral SCL

*p < .05

167

Table 26.
Correlations between Average Skin Conductance Level (SCL) across Clip Conditions and Preto Post-task PTCI Negative Cognitions about the World Subscale Scores in the LGB Subsample
(in Plain Text; N = 28) and Heterosexual Subsample (in Bold; N = 32)
1

2

3

4

5

6

-

-.56*

-.47*

-.11

.21

.46*

2. GT SCL

-.44*

-

-.47*

-.21

-.41*

-.29

3. Bias SCL

-.49*

-.56*

-

.34

.21

-.18

4. Pre-Task PTCI

-.13

.14

-.01

-

.75*

-.33

5. Post-Task PTCI

.12

.12

-.23

.72*

-

-.38*

6. Change in PTCI

.31

.02

-.31

-.11

.62*

-

1. Neutral SCL

*p < .05
Table 27.
Correlations between Startle Reflex EMG and Subjective Ratings of Arousal across Clip
Conditions in the LGB Subsample (in Plain Text; N = 27) and Heterosexual Subsample (in Bold;
N = 30)
1

2

3

4

5

6

-

-.02

-.23

.03

-.22

-.19

2. GT Startle

-.06

-

-.55*

.09

.25

.13

3. Bias Startle

-.43*

-.35

-

-.10

.18

-.13

4. Neutral Arousal

-.05

.01

-.35

-

.33

.18

5. GT Arousal

-.16

.26

-.27

.49*

-

.41*

6. Bias Arousal

.07

-.05

-.13

.32

.63*

-

1. Neutral Startle

*p < .05

168

Table 28.
Correlations between Startle Reflex EMG and Subjective Ratings of Affect Valence across Clip
Conditions in the LGB Subsample (in Plain Text; N = 27) and Heterosexual Subsample (in Bold;
N = 30)
1

2

3

4

5

6

-

-.02

-.23

-.13

.16

-.16

2. GT Startle

-.06

-

-.55*

.05

.20

.33

3. Bias Startle

-.43*

-.35

-

.06

-.11

.22

.09

.12

.18

-

.47*

.17

5. GT Valence

-.38*

.04

.18

.01

-

.58*

6. Bias Valence

-.32

.05

.12

-.02

.82*

-

1. Neutral Startle

4. Neutral Valence

*p < .05
Table 29.
Correlations between Startle Reflex EMG and Subjective Ratings of Dominance across Clip
Conditions in the LGB Subsample (in Plain Text; N = 27) and Heterosexual Subsample (in Bold;
N = 30)
1

2

3

4

5

6

-

-.02

-.23

.18

-.27

-.09

2. GT Startle

-.06

-

-.55*

-.14

-.07

.03

3. Bias Startle

-.43*

-.35

-

-.13

.10

.16

4. Neutral Dominance

-.40*

-.13

.30

-

.35

-.27

5. GT Dominance

-.02

-.29

.04

.45*

-

.37

6. Bias Dominance

.05

-.27

-.10

.39*

.84*

-

1. Neutral Startle

*p < .05

169

Table 30.
Correlations between Startle Reflex EMG and Subjective Ratings of Interpersonal Warmth
across Clip Conditions in the LGB Subsample (in Plain Text; N = 27) and Heterosexual
Subsample (in Bold; N = 30)
1

2

3

4

5

6

-

-.02

-.23

-.14

-.20

-.20

2. GT Startle

-.06

-

-.55*

.29

.14

-.06

3. Bias Startle

-.43*

-.35

-

-.07

-.36

-.18

4. Neutral Warmth

.28

-.03

.10

-

.04

-.23

5. GT Warmth

-.05

.01

.10

-.09

-

.78*

6. Bias Warmth

.17

.02

-.19

-.05

.53*

-

1. Neutral Startle

*p < .05
Table 31.
Correlations between Startle Reflex EMG across Clip Conditions and Pre- to Post-task PTCI
Negative Cognitions about the World Subscale Scores in the LGB Subsample (in Plain Text; N =
27) and Heterosexual Subsample (in Bold; N = 30)
1

2

3

4

5

6

-

-.02

-.23

-.19

.15

.47*

2. GT Startle

-.06

-

-.55*

-.29

-.22

.08

3. Bias Startle

-.43*

-.35

-

.23

-.07

-.41*

4. Pre-Task PTCI

.11

.06

-.21

-

.75*

-.33

5. Post-Task PTCI

-.07

.00

-.22

.72*

-

-.38*

6. Change in PTCI

-.23

-.06

-.08

-.11

.62*

-

1. Neutral Startle

*p < .05

170

Table 32.
Correlations between Subjective Ratings of Arousal across Clip Conditions and Pre- to Post-task
PTCI Negative Cognitions about the World Subscale Scores in the LGB Subsample (in Plain
Text; N = 28) and Heterosexual Subsample (in Bold; N = 32)
1

2

3

4

5

6

-

.33

.18

.04

.07

.04

2. GT Arousal

.49*

-

.41*

-.09

-.28

-.27

3. Bias Arousal

.32

.63*

-

.07

-.11

-.25

4. Pre-Task PTCI

-.11

.12

-.14

-

.75*

-.33

5. Post-Task PTCI

.06

.10

.02

.72*

-

-.38*

6. Change in PTCI

.22

.00

.18

-.11

.62*

-

1. Neutral Arousal

*p < .05
Table 33.
Correlations between Subjective Ratings of Affect Valence across Clip Conditions and Pre- to
Post-task PTCI Negative Cognitions about the World Subscale Scores in the LGB Subsample (in
Plain Text; N = 28) and Heterosexual Subsample (in Bold; N = 32)
1

2

3

4

5

6

-

.47*

.17

.18

.25

.11

2. GT Valence

.01

-

.58*

.05

.29

.34

3. Bias Valence

-.02

.82*

-

-.26

-.26

-.01

4. Pre-Task PTCI

.06

-.08

.01

-

.75*

-.33

5. Post-Task PTCI

-.05

-.01

.05

.72*

-

-.38*

6. Change in PTCI

-.13

.07

.06

-.11

.62*

-

1. Neutral Valence

*p < .05

171

Table 34.
Correlations between Subjective Ratings of Dominance across Clip Conditions and Pre- to Posttask PTCI Negative Cognitions about the World Subscale Scores in the LGB Subsample (in Plain
Text; N = 28) and Heterosexual Subsample (in Bold; N = 32)
1

2

3

4

5

6

-

.35

-.27

.14

.32

.26

2. GT Dominance

.45*

-

.37

.43*

.40*

-.03

3. Bias Dominance

.39*

.84*

-

.09

-.11

-.28

4. Pre-Task PTCI

-.18

-.09

-.30

-

.75*

-.33

5. Post-Task PTCI

.07

-.20

-.34

.72*

-

-.38*

6. Change in PTCI

.30

-.19

-.14

-.11

.62*

-

1. Neutral Dominance

*p < .05
Table 35.
Correlations between Subjective Ratings of Interpersonal Warmth across Clip Conditions and
Pre- to Post-task PTCI Negative Cognitions about the World Subscale Scores in the LGB
Subsample (in Plain Text; N = 28) and Heterosexual Subsample (in Bold; N = 32)
1

2

3

4

5

6

-

.04

-.23

.03

.17

.19

2. GT Dominance

-.09

-

.78*

.20

.28

.11

3. Bias Dominance

-.05

.53*

-

-.02

.00

.02

4. Pre-Task PTCI

-.05

.08

.08

-

.75*

-.33

5. Post-Task PTCI

-.06

-.05

-.15

.72*

-

-.38*

6. Change in PTCI

-.03

.16

-.30

-.11

.62*

-

1. Neutral Dominance

*p < .05

172

Collective Identity Analyses
Table 36.
Results of a 2x3 Mixed Model Repeated Measures Analysis of Covariance Comparing the
Interactive Effects of Sexual Orientation Group, Clip Condition, and Collective Identity
Variables on Average Skin Conductance Level (N = 60)
Power
df

F

p

ηp2

(obs.)

Group

1

2.92

.09

.05

.39

Condition

2

.31

.74

.01

.09

Group*Condition

2

.49

.61

.01

.13

Centrality*Condition

2

.47

.63

.01

.12

Private Reg.* Con.

2

.20

.82

.00

.08

Public Reg.*Con.

2

.29

.75

.01

.07

Group*Centr*Con

2

.15

.87

.01

.07

Group*PubReg*Con

2

.75

.47

.02

.18

4

1.19

.32

.05

.36

Main effects

Two-way interactions

Three-way interactions

Four-way interaction
Gp*Cent*PubR*Con

173

Figure 12.
Estimated Marginal Means of Average Skin Conductance Level across Task Conditions and
Kinsey Groups, Controlling for Centrality, Private Regard, and Public Regard (N = 60)

Average SCL Scores (T)

60
55
50
45
40
35
30
Heterosexual

LGB
Stimulus Condition

Neutral

General Threat

Bias

Table 37.
Results of a 2x4 Mixed Model Repeated Measures Analysis of Covariance Comparing the
Interactive Effects of Sexual Orientation Group, Clip Condition, and Collective Identity
Variables on Startle Eyeblink Reflex Magnitude (N = 57)
Power
df

F

p

ηp2

(obs.)

Group

1

.11

.74

.00

.06

Condition

3

.99

.32

.00

.06

Main effects

Two-way interactions

174

Table 37 (cont’d)
Group*Condition

3

2.62

.08

.05

.51

Centrality*Condition

3

2.03

.12

.12

.49

Private Reg.* Con.

3

.56

.65

.04

.16

Public Reg.*Con.

3

2.16

.11

.12

.52

Group*Centr*Con

3

.97

.42

.06

.24

Group*PubReg*Con

3

1.27

.30

.08

.32

6

1.16

.33

.07

.44

Three-way interactions

Four-way interaction
Gp*Cent*PubR*Con

Figure 13.
Estimated Marginal Means of Startle Reflex Magnitude across Task Conditions and Kinsey
Groups, Controlling for Centrality, Private Regard, and Public Regard (N = 57)

Startle Reflex Magnitude (T)

60
55
50
45
40
35
30
Heterosexual
Neutral

LGB
General Threat

175

Bias

Table 38.
Results of a 2x3 Mixed Model Repeated Measures Analysis of Covariance Comparing the
Interactive Effects of Sexual Orientation Group, Clip Condition, and Collective Identity
Variables on Subjective Ratings of Arousal (N = 60)
Power
df

F

p

ηp2

(obs.)

1

.03

.87

.00

.05

1.66

1.20

.30

.02

.23

Group*Condition

1.66

2.27

.11

.04

.41

Centrality*Condition

1.66

2.57

.19

.03

.32

Private Reg.* Con.

1.66

1.37

.39

.02

.19

Public Reg.*Con.

1.66

1.00

.49

.01

.15

Group*Centr*Con

1.66

5.55

.06

.07

.67

Group*PubReg*Con

1.66

4.37

.07

.05

.56

3.31

4.12

.06

.10

.74

Main effects
Group
Condition
Two-way interactions

Three-way interactions

Four-way interaction
Gp*Cent*PubR*Con

176

Table 39.
Results of a 2x3 Mixed Model Repeated Measures Analysis of Covariance Comparing the
Interactive Effects of Sexual Orientation Group, Clip Condition, and Collective Identity
Variables on Subjective Affect Valence (N = 60)
Power
df

F

p

ηp2

(obs.)

1

.82

.37

.02

.15

1.41

2.97

.08

.06

.39

Group*Condition

1.41

2.94

.08

.06

.47

Centrality*Condition

1.41

1.99

.16

.04

.34

Private Reg.* Con.

1.41

.43

.58

.01

.11

Public Reg.*Con.

1.41

1.02

.34

.02

.19

Group*Centr*Con

1.41

3.69

.04

.07

.56

Group*PubReg*Con

1.41

2.68

.09

.05

.43

2.83

1.90

.14

.07

.46

Main effects
Group
Condition
Two-way interactions

Three-way interactions

Four-way interaction
Gp*Cent*PubR*Con

177

Figure 14.
Estimated Marginal Means of Subjective Arousal and Affect Valence Ratings across Task
Conditions and Kinsey Groups, Controlling for Centrality, Private Regard, and Public Regard
(N = 60)
9

Average SAM Score

8
7
6
5
4
3
2
1
0
Neutral

General
Threat

Bias Threat

Neutral

General
Threat

Arousal

Bias Threat

Valence
Heterosexual

LGB

Table 40.
Results of a 2x3 Mixed Model Repeated Measures Analysis of Covariance Comparing the
Interactive Effects of Sexual Orientation Group, Clip Condition, and Collective Identity
Variables on Subjective Dominance (N = 60)
Power
df

F

p

ηp2

(obs.)

1

.29

.59

.01

.08

Main effects
Group

178

Table 40 (cont’d)
Condition

1.42

1.80

.18

.03

.31

Group*Condition

1.42

.10

.84

.00

.06

Centrality*Condition

1.42

1.60

.21

.03

.28

Private Reg.* Con.

1.42

1.31

.27

.03

.24

Public Reg.*Con.

1.42

2.24

.13

.04

.37

Group*Centr*Con

1.42

.22

.73

.00

.08

Group*PubReg*Con

1.42

.26

.70

.01

.08

2.83

2.48

.07

.09

.58

Two-way interactions

Three-way interactions

Four-way interaction
Gp*Cent*PubR*Con

Table 41.
Results of a 2x3 Mixed Model Repeated Measures Analysis of Covariance Comparing the
Interactive Effects of Sexual Orientation Group, Clip Condition, and Collective Identity
Variables on Interpersonal Warmth (N = 60)
Power
df

F

p

ηp2

(obs.)

1

3.26

.08

.06

.43

1.51

.27

.70

.01

.09

1.51

.02

.95

.00

.05

Main effects
Group
Condition
Two-way interactions
Group*Condition

179

Table 41 (cont’d)
Centrality*Condition

1.51

.51

.55

.01

.12

Private Reg.* Con.

1.51

.06

.90

.00

.06

Public Reg.*Con.

1.51

.92

.38

.02

.18

Group*Centr*Con

1.51

.12

.83

.00

.07

Group*PubReg*Con

1.51

.14

.82

.00

.07

3.02

1.46

.23

.05

.37

Three-way interactions

Four-way interaction
Gp*Cent*PubR*Con

Figure 15.
Estimated Marginal Means of Subjective Dominance and Interpersonal Warmth Ratings across
Task Conditions and Kinsey Groups, Controlling for Centrality, Private Regard, and Public
Regard (N = 60)
9

Average SAM Score

8
7
6
5
4
3
2
1
0
Neutral

General
Threat

Bias Threat

Neutral

Dominance

General
Threat
Warmth

Heterosexual
180

LGB

Bias Threat

Table 42.
Results of a Mixed Model Repeated Measures Analysis of Covariance Comparing the Interactive
Effects of Sexual Orientation Group, Clip Condition, and Collective Identity Variables on PreTask to Post-Task PTCI Negative Cognitions about the World Subscale Scores (N = 60)
Power
df

F

p

ηp2

(obs.)

Group

1

.50

.48

.01

.11

Time

1

.10

.75

.00

.06

Group*Time

1

.03

.88

.00

.05

Centrality*Time

1

.02

.89

.00

.05

Private Reg.* Time

1

.61

.44

.01

.12

Public Reg.*Time

1

.06

.82

.00

.06

Group*Centr*Time

1

.01

.92

.00

.05

Group*PubReg*Time

1

.03

.86

.00

.05

2

.27

.76

.01

.09

Main effects

Two-way interactions

.

Three-way interactions

Four-way interaction
Gp*Cent*PubR*Time

181

Figure 16.
Estimated Marginal Means of Pre- and Post-Task PTCI Negative Cognitions about the World
Subscale Scores by Kinsey Group, Controlling for Centrality, Private Regard, and Public
Regard (N = 60)
5
4.5

PTCI Score

4
3.5
3
2.5
2
1.5
1
Heterosexual

LGB

Pre-Task

182

Post-Task

APPENDIX E
Supplementary Variables of Interest
Secondary Analyses
Supplementary variables - descriptive data and comparisons.
Descriptive data for supplementary variables are presented in Table 43. T-tests were used
to assess whether continuous variable means differed across groups. A chi-square test was used
to assess whether past bias victimization (yes/no) frequencies varied across groups. Presence of
bias victimization varied across group, with a greater proportion of LGB individuals reporting
that had experienced any type of past victimization that they believed to be motivated by bias.
Table 43.
Descriptive Data and Group Comparisons for Supplementary Questionnaire Variables (N = 60)

Heterosexual

LGB

M (SD)

M (SD)

Neuroticism

2.63 (.63)

2.93 (.72)

Range

1.40-3.80

1.75-4.35

Sum Life Events

3.16 (2.41)

2.93 (2.11)

1-17

4-17

Other PTCI Self

2.20 (.77)

2.40 (1.27)

Range

1.05-3.90

1.00-5.62

Other PTCI World

4.22 (1.06)

4.07 (1.29)

Range

2.00-6.00

1.14-6.29

Other PTCI Self-Blame

2.63 (1.14)

2.44 (1.39)

Range

1.00-4.80

1.00-6.40

Range

183

t (df)

p

d

-1.76 (58)

.08

.46

.39 (58)

.70

.10

-.74 (57)

.47

.20

.49 (57)

.63

.13

.56 (57)

.58

.15

Table 43 (cont’d)

Bias Victimization
History (Yes/No)

N (%)

N (%)

Χ2

p

ϕ

8 (25)

22 (79)

17.14

<.001

.54

Bivariate correlations between secondary variables of interest and average skin
conductance level across clip conditions, by sexual orientation group, are presented in Table 44.
Some notable interrelations between secondary variables of interest were observed. Within both
groups, PTCI subscales were moderately to strongly interrelated, and neuroticism was strongly
correlated with negative beliefs about the self and moderately correlated with negative world
beliefs. Other stressful life events were only negligibly related to other variables of interest.
Groups differed with regard to self-blame following a self-identified past stressful life event,
negative beliefs about the self, and bias victimization. Within the LGB sample, self-blame was
strongly correlated with neuroticism, whereas no such relation was observed in the heterosexual
sample. Having personally experienced bias victimization was weakly correlated with negative
beliefs about the self and self-blame in the LGB sample. In the heterosexual sample, bias
victimization was moderately positively correlated with neuroticism and negative beliefs about
the world, and weakly related to negative beliefs about the self.
Secondary variables of interest were correlated with average SCL (Table 44) as follows:
For both groups, neuroticism and PTCI Negative Cognitions about the Self were weakly
negatively correlated with SCL during general threat clips. Within the heterosexual sample, no
other notable relations were observed. However, within the LGB sample additional relations
were observed between average SCL across clip conditions and other-event PTCI subscales. The
Negative Cognitions about the Self subscale was weakly positively correlated with average SCL
184

during bias crime clips. The Negative Cognitions about the World subscale showed a weak
positive correlation with SCL during neutral clips, a strong negative correlation with SCL during
general threat clips, and a moderate positive correlation with SCL during bias crime clips.
Finally, the Self Blame subscale was weakly positively correlated with neutral SCL and weakly
negatively correlated with general threat SCL for LGB individuals.
Table 45 presents bivariate correlations between secondary variables of interest and
startle reflex magnitude. No shared correlation trends were observed between sexual orientation
groups. For the LGB subsample, negative beliefs about the world were moderately negatively
correlated with startle reflex magnitude during general threat clips. A weak positive correlation
was also observed between the Self-Blame subscale of the PTCI and startle reflex magnitude
during neutral clips. For the heterosexual sample, neuroticism and PTCI Negative Cognitions
about the Self were weakly negative correlated with startle reflex magnitude during general
threat clips. Negative beliefs about the world were weakly negatively correlated with bias startle.
Finally, the PTCI Self-Blame subscale was moderately positively correlated with neutral startle
and moderately negatively correlated with bias startle.
Tables 46-49 present bivariate correlations between secondary variables of interest and
self-rated dimensions of affective response during each clip condition. A number of differences
were observed across groups. Some select observations: For heterosexual participants, selfreported arousal was associated negatively with life events (in the bias condition) and self-blame
(in the general threat and bias conditions). For LGB participants, arousal in the bias condition
was negatively associated with neuroticism and positively associated with bias victimization. For
heterosexual participants, stressful life events were positively associated with ratings of affect
valence in both threat conditions, while this relationship was negative for the LGB participants.

185

For LGB participants, negative beliefs about the world were negatively associated with reported
valence in the bias condition. Reported dominance in the general threat condition was positively
associated with past stressful life events and negatively associated with negative self-cognitions,
negative world beliefs, and past bias victimization for the heterosexual group. It was positively
associated with negative world beliefs and past bias victimization for the LGB group.
Dominance in the bias threat condition was negatively associated with bias victimization for the
heterosexual participants, and negatively associated with negative self-cognitions and self-blame
for the LGB group. Reported interpersonal warmth in the bias condition was positively
associated with stressful life events, negative self-cognitions, and self-blame for heterosexual
participants, and negatively associated with past bias victimization for LGB participants.
Finally, Table 50 presents bivariate correlations between additional variables of interest
and task-linked PTCI Negative Cognitions about the World subscale scores. For both groups,
neuroticism, negative self-cognitions, and self-blame were weakly positively associated with pretask PTCI scores. Neuroticism was moderately associated with post-task scores for the
heterosexual group and strongly associated with post-task scores for the LGB group. Bias
victimization was weakly positively associated with post-task PTCI scores for both groups.
Within the LGB group, change in PTCI scores was weakly positively associated with
neuroticism and self-blame, and moderately positively associated with bias victimization.

186

Table 44.
Bivariate Correlations between Secondary Variables of Interest and Average Skin Conductance Level across Clip Conditions, within
the LGB Subsample (in Plain Text; N = 27) and Heterosexual Subsample (in Bold; N = 30)
Sum Life

PTCI

Bias

SCL

SCL Gen.

SCL Bias

Neuroticism

Events

PTCI Self

PTCI World

Blame

Victim

Neutral

Threat

Threat

1

2

3

4

5

6

7

8

9

r

(p)

r

(p)

r

1. -

-

.03

(.88) .75

2. .03

(.89)

-

-

3. .53

(<.01) -.05

(.81) -

4. .32

(.08)

-.13

5. .10

(.59)

6. .41

(p)

r

(p)

r

(p)

(p)

r

(p)

r

(.06)

.49

(<.01) .20

(.32)

.08

(.70)

.08

(.71)

.12

(.54)

.14

(.47)

.17

-

.54

(<.01) .72

(<.01) .30

(.12)

(.47) .45

(.01)

-

-

.44

(.02)

.23

-.18

(.34) .51

(<.01) .41

(.02)

-

-

(.01)

.12

(.52) .35

(.05)

.33

(.07)

.12

7. .11

(.56)

-.18

(.33) .10

(.58)

.07

(.70)

.03

8. -.20

(.28)

.15

(.41) -.20 (.28)

-.12

9. .09

(.62)

.02

(.91) .10

.05

(<.01) .35

-.13 (.52)

(.60)

r

(p)

-.20 (.32)

.13

(.52)

(.38)

-.29 (.14)

.12

(.52)

.04

(.85)

-.23 (.25)

.20

(.30)

(.25)

.23

(.24)

-.54 (<.01) .33

(.09)

.26

(.18)

.22

(.27)

-.24 (.22)

.02

(.91)

(.53)

-

-

-.00 (.99)

.09

-.10 (.62)

(.87)

-.07 (.69)

-

-.56 (<.01) -.47 (.01)

(.53)

-.04 (.85)

-.07 (.71)

-.44 (.01)

-

(.80)

.01

.13

-.49 (<.01)

-.56 (<.01)

187

(.98)

R

(.47)

-

(p)

(.63)

-

-.47 (.01)
-

-

Table 45.
Bivariate Correlations between Secondary Variables of Interest and Average Startle Reflex Magnitude across Clip Conditions, within
the LGB Subsample (in Plain Text; N = 28) and Heterosexual Subsample (in Bold; N = 32)
Startle
Sum Life

PTCI

Bias

Startle

General

Startle Bias

Neuroticism

Events

PTCI Self

PTCI World

Blame

Victim

Neutral

Threat

Threat

1

2

3

4

5

6

7

8

9

r

(p)

r

(p)

r

1. -

-

.03

(.88) .75

2. .03

(.89)

-

-

(p)

r

(p)

(p)

r

(p)

r

(p)

r

(p)

(<.01) .35

(.06)

.49

(<.01) .20

(.32)

-.10

(.62)

.08

(.68)

-.11

(.60)

(.52)

.08

(.71)

.12

(.54)

.14

(.47)

.00

(.99)

.15

(.45)

-.19

(.34)

3. .53

(<.01)

-.05

(.81) -

-

.54

(<.01) .72

(<.01) .30

(.12)

.00

(.99)

-.03

(.88)

-.17

(.34)

4. .32

(.08)

-.13

(.47) .45

(.01)

-

-

.44

(.02)

.23

(.25)

-.07

(.74)

-.32

(.10)

-.01

(.97)

5. .10

(.59)

-.18

(.34) .51

(<.01) .41

(.02)

-

-

.26

(.18)

.22

(.26)

.06

(.77)

-.11

(.60)

6. .41

(.01)

.12

(.52) .35

(.05)

.33

(.07)

.12

(.53)

-

-

-.07

(.73)

.19

(.34)

-.09

(.64)

7. -.12

(.53)

-.33

(.08) .10

(.62)

.15

(.43)

.34

(.07)

-.03

(.89)

-

-

-.02

(.93)

-.23

(.25)

8. .33

(.07)

.14

(.45) .18

(.36)

.01

(.96)

-.17

(.37)

.10

(.59)

-.06

(.77)

-

-

-.55

(<.01)

9. -.25

(.19)

.15

(.43) -.26

(.17)

-.22

(.26)

-.31

(.10)

.10

(.59)

-.43

(.02)

-.35

(.06)

-

-

-.13

(p)

r

188

r

Table 46.
Bivariate Correlations between Secondary Variables of Interest and Subjective Ratings of Arousal across Clip Conditions, within the
LGB Subsample (in Plain Text; N = 28) and Heterosexual Subsample (in Bold; N = 32)
Sum Life

PTCI

PTCI

Bias

Neutral

Bias

Neuroticism

Events

PTCI Self

World

Blame

Victim

Arousal

GT Arousal

Arousal

1

2

3

4

5

6

7

8

9

r

(p)

r

(p)

r

1. -

-

.03

(.88) .75

2. .03

(.89)

-

-

(p)

r

(p)

(p)

r

(p)

r

(p)

r

(p)

(<.01) .35

(.06)

.49

(<.01) .20

(.32)

-.07

(.74)

-.31

(.11)

-.29

(.13)

(.52)

.08

(.71)

.12

(.54)

.14

(.47)

.12

(.56)

.15

(.45)

-.11

(.57)

3. .53

(<.01)

-.05

(.81) -

-

.54

(<.01) .72

(<.01) .30

(.12)

.17

(.39)

-.39

(.04)

-.14

(.48)

4. .32

(.08)

-.13

(.47) .45

(.01)

-

-

.44

(.02)

.23

(.25)

.25

(.21)

-.27

(.16)

-.01

(.95)

5. .10

(.59)

-.18

(.34) .51

(<.01) .41

(.02)

-

-

.26

(.18)

.16

(.42)

-.15

(.44)

-.08

(.68)

6. .41

(.01)

.12

(.52) .35

(.05)

.33

(.07)

.12

(.53)

-

-

-.07

(.71)

-.02

(.94)

.20

(.32)

7. .18

(.32)

-.05

(.81) .01

(.96)

.13

(.48)

.18

(.33)

.06

(.75)

-

-

.33

(.09)

.18

(.37)

8. .14

(.44)

-.10

(.58) .06

(.76)

.15

(.42)

-.21

(.26)

-.04

(.83)

.49

(<.01)

-

-

.41

(.03)

9. -.07

(.71)

-.26

(.15) -.12

(.54)

.15

(.43)

-.35

(.06)

-.15

(.41)

.32

(.07)

.63

(<.01)

-

-

-.13

(p)

r

189

r

Table 47.
Bivariate Correlations between Secondary Variables of Interest and Subjective Ratings of Affect Valence across Clip Conditions,
within the LGB Subsample (in Plain Text; N = 28) and Heterosexual Subsample (in Bold; N = 32)
Sum Life

PTCI

Bias

Neutral

Neuroticism

Events

PTCI Self

PTCI World

Blame

Victim

Valence

GT Valence

Bias Valence

1

2

3

4

5

6

7

8

9

r

(p)

r

(p)

r

(p)

1. -

-

.03

(.88)

.75

2. .03

(.89)

-

-

3. .53

(<.01)

-.05

4. .32

(.08)

5. .10

r

(p)

r

(p)

(<.01) .35

(.06)

.49

-.13

(.52)

.08

(.71)

.12

(.81)

-

-

.54

-.13

(.47)

.45

(.01)

-

(.59)

-.18

(.34)

.51

6. .41

(.01)

.12

(.52)

(p)

r

(p)

r

(p)

r

(p)

(<.01) .20

(.32)

.06

(.76)

.37

(.05)

.15

(.45)

(.54)

.14

(.47)

-.40

(.04)

-.12

(.53)

-.16

(.43)

(<.01) .72

(<.01) .30

(.12)

.08

(.68)

.34

(.07)

-.11

(.57)

-

.44

(.02)

.23

(.25)

.25

(.21)

.19

(.34)

-.36

(.06)

(<.01) .41

(.02)

-

-

.26

(.18)

.12

(.55)

.31

(.11)

-.09

(.64)

.35

(.05)

(.07)

.12

(.53)

-

-

.16

(.42)

.12

(.54)

-.02

(.91)

7. -.26

(.15)

.13

(.49)

-.58

(<.01) -.09

(.63)

-.34

(.07)

-.27

(.14)

-

-

.47

(.01)

.17

(.38)

8. .02

(.93)

.55

(<.01) -.11

(.55)

-.10

(.61)

.12

(.53)

.10

(.59)

.01

(.97)

-

-

.58

(<.01)

9. -.04

(.82)

.51

(<.01) -.13

(.47)

-.18

(.34)

.12

(.53)

.08

(.66)

-.02

(.90)

.82

(<.01)

-

-

.33

190

r

Table 48.
Bivariate Correlations between Secondary Variables of Interest and Subjective Ratings of Dominance across Clip Conditions, within
the LGB Subsample (in Plain Text; N = 28) and Heterosexual Subsample (in Bold; N = 32)
Sum Life

PTCI

Bias

Neutral

GT

Bias

Neuroticism

Events

PTCI Self

PTCI World

Blame

Victim

Dominance

Dominance

Dominance

1

2

3

4

5

6

7

8

9

r

(p)

r

(p)

r

(p)

1. -

-

.03

(.88)

.75

2. .03

(.89)

-

-

3. .53

(<.01)

-.05

4. .32

(.08)

5. .10

r

(p)

r

(p)

(<.01) .35

(.06)

.49

-.13

(.52)

.08

(.71)

.12

(.81)

-

-

.54

-.13

(.47)

.45

(.01)

-

(.59)

-.18

(.34)

.51

6. .41

(.01)

.12

(.52)

7. -.15

(.41)

-.08

8. -.19

(.29)

9. -.27

(.13)

(p)

r

(p)

r

(p)

r

(p)

(<.01) .20

(.32)

.04

(.86)

.20

(.30)

-.05

(.81)

(.54)

.14

(.47)

-.30

(.13)

.12

(.56)

.19

(.34)

(<.01) .72

(<.01) .30

(.12)

.19

(.33)

.02

(.92)

-.43

(.02)

-

.44

(.02)

.23

(.25)

.34

(.08)

.40

(.04)

-.24

(.22)

(<.01) .41

(.02)

-

-

.26

(.18)

.11

(.59)

-.05

(.82)

-.24

(.21)

.35

(.05)

.33

(.07)

.12

(.53)

-

-

-.09

(.66)

.21

(.29)

-.12

(.53)

(.67)

-.41

(.02)

-.09

(.63)

-.19

(.31)

-.26

(.15)

-

-

.35

(.07)

-.27

(.17)

.20

(.28)

-.31

(.10)

-.09

(.65)

.03

(.89)

-.26

(.15)

.45

(<.01)

-

-

.37

(.05)

.08

(.66)

-.24

(.19)

-.20

(.29)

.11

(.55)

-.31

(.09)

.39

(.03)

.84

(<.01)

-

-

191

r

Table 49.
Bivariate Correlations between Secondary Variables of Interest and Subjective Ratings of Interpersonal Warmth across Clip
Conditions, within the LGB Subsample (in Plain Text; N = 28) and Heterosexual Subsample (in Bold; N = 32)
Sum Life

PTCI

Bias

Neutral

Neuroticism

Events

PTCI Self

PTCI World

Blame

Victim

Warmth

GT Warmth

Bias Warmth

1

2

3

4

5

6

7

8

9

r

(p)

r

(p)

r

(p)

1. -

-

.03

(.88)

.75

2. .03

(.89)

-

-

3. .53

(<.01)

-.05

4. .32

(.08)

5. .10

(.59)

(p)

r

(p)

(p)

r

(p)

r

(p)

r

(p)

(<.01) .35

(.06)

.49

(<.01) .20

(.32)

-.02

(.91)

.20

(.32)

.13

(.50)

-.13

(.52)

.08

(.71)

.12

(.54)

.14

(.47)

-.11

(.57)

.12

(.55)

.01

(.97)

(.81)

-

-

.54

(<.01) .72

(<.01) .30

(.12)

-.01

(.95)

.01

(.98)

-.15

(.45)

-.13

(.47)

.45

(.01)

-

-

.44

(.02)

.23

(.25)

.12

(.54)

.29

(.13)

-.01

(.97)

-.18

(.34)

.51

(<.01) .41

(.02)

-

-

.26

(.18)

-.02

(.94)

.01

(.94)

-.14

(.47)

6. .41 ( (.01)

.12

(.52)

.35

(.05)

.33

(.07)

.12

(.53)

-

-

.32

(.10)

.02

(.93)

-.23

(.24)

7. -.36

(.04)

-.06

(.75)

-.42

(.02)

-.03

(.87)

-.23

(.22)

-.50

(<.01)

-

-

.04

(.85)

-.23

(.24)

8. .11

(.56)

.50

(<.01) -.06

(.77)

-.01

(.96)

-.05

(.81)

.24

(.19)

-.09

(.61)

-

-

.78

(<.01)

9. .03

(.89)

.26

(.15)

(.28)

-.03

(.90)

.22

(.25)

-.06

(.76)

-.05

(.80)

.53

(<.01)

-

-

.20

r

192

r

Table 50.
Bivariate Correlations between Secondary Variables of Interest and Pre-task Negative Cognitions about the World (PTCI), Post-Task
PTCI, and Change in PTCI (Post – Pre), within the LGB Subsample (in Plain Text; N = 28) and Heterosexual Subsample (in Bold; N
= 32)
Sum Life

PTCI

Bias

Pre-World

Post-World

Change in

Neuroticism

Events

PTCI Self

PTCI World

Blame

Victim

Cog.

Cog.

World Cog.

1

2

3

4

5

6

7

8

9

r

(p)

r

(p)

r

(p)

1. -

-

.03

(.88)

.75

2. .03

(.89)

-

-

3. .53

(<.01)

-.05

4. .32

(.08)

5. .10

(.59)

r

(p)

r

(p)

(<.01) .35

(.06)

.49

-.13

(.52)

.08

(.71)

(.81)

-

-

.54

-.13

(.47)

.45

(.01)

-

-.18

(.34)

.51

6. .41 ( (.01)

.12

(.52)

7. .24

(.18)

-.02

8. .32

(.08)

9. .18

(.33)

(p)

r

(p)

r

(p)

r

(p)

(<.01) .20

(.32)

.26

(.19)

.41

(.03)

.23

(.25)

.12

(.54)

.14

(.47)

.11

(.58)

.12

(.54)

.02

(.91)

(<.01)

.72

(<.01) .30

(.12)

.25

(.20)

.46

(.01)

.31

(.11)

-

.44

(.02)

.23

(.25)

.70

(<.01)

.87

(<.01)

.27

(.17)

(<.01) .41

(.02)

-

-

.26

(.18)

.24

(.21)

.44

(.02)

.28

(.14)

.35

(.05)

.33

(.07)

.12

(.53)

-

-

.04

(.84)

.25

(.20)

.30

(.12)

(.90)

.24

(.20)

.74

(<.01)

.20

(.29)

.18

(.34)

-

-

.75

(<.01)

-.33

(.08)

-.07

(.70)

.28

(.12)

.84

(<.01)

.23

(.22)

.20

(.26)

.72

<.01)

-

-

.38

(.05)

-.08

(.69)

.11

(.55)

.28

(.13)

.08

(.66)

.09

(.61)

-.11

(.56)

.62

(<.01)

-

-

193

r

REFERENCES

194

REFERENCES

Agar, E., Kennedy, P., & King, N.S. (2006). The role of negative cognitive appraisals in PTSD
symptoms following spinal cord injuries. Behavioral and Cognitive Psychotherapy, 34,
437-452.
Alessi, E.J., Meyer, I.H., & Martin, J.I. (2011). PTSD and sexual orientation: An examination of
criterion A1 and non-criterion A1 events. Psychological Trauma: Research, Practice,
and Policy,n.p specified.
Allport, G.W. (1954). The Nature of Prejudice. Cambridge, MA: Perseus Books.
American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders
(5th ed.). Arlington, VA: American Psychiatric Publishing.
Ashmore, R.D., Deaux, K., & McLaughlin-Volpe, T. (2004). An organizing framework for
collective identity: Articulation and significance of multidimensionality. Psychological
Bulletin, 130, 80-114.
Averill, J.R. (1973). Personal control over aversive stimuli and its relationship to stress.
Psychological Bulletin, 80(4), 286-303.
Avina, C., & O’Donohue, W. (2002). Sexual harassment and PTSD: Is sexual harassment
diagnosable trauma? Journal of Traumatic Stress, 15(1), 69-75.
Bagley, C., & Tremblay, P. (1998). On the prevalence of homosexuality and bisexuality, in a
random community survey of 750 men ages 18 to 27. Journal of Homosexuality, 36(2),
1-18.
Balsam, K.F., Rothblum, E.D, & Beauchaine, T.P. (2005). Victimization over the life span: A
comparison of lesbian, gay, bisexual, and heterosexual siblings. Journal of Consulting
and Clinical Psychology, 77(3), 477-487.
Beck, J.G., Coffey, S.F., Palyo, S.A., Gudmundsdottir, B., Miller, L.M., & Colder, C.R. (2004).
Psychometric properties of the Posttraumatic Cognitions Inventory (PTCI): A replication
with motor vehicle accident survivors. Psychological Assessment, 16(3), 289-298.
Becker, J.D. (1973). A model for the encoding of experiential information. In R. Schank & K.
Colby (Eds.), Computer Models of Thought and Language. San Francisco, CA: W.H.
Freeman & Co.
Benish-Weisman, M., & Horenczyk, G. (2010). Cultural identity and perceived success among
Israeli immigrants: An emic approach. International Journal of Intercultural Relations,
34, 516-526.

195

Bennett, S. A., Beck, J. G., & Clapp, J. D. (2009). Understanding the relationship between
posttraumatic stress disorder and trauma cognitions: The impact of thought control
strategies. Behavior Research and Therapy, 47(12), 1018-1023.
Bentler, P. M. (1990). Comparative fit indexes in structural models. Psychological Bulletin,
107(2), 238-246.
Berman, A.M., Schwartz, S.J., Kurtines, W.M., & Berman, S.L. (2001). The process of
exploration in identity formation: The role of style and competence. Journal of
Adolescence, 24, 513-528.
Berntson, G.G., & Cacioppo, J.T. (2007). From homeostasis to allodynamic regulation. In
Cacioppo, John T., Tassinary, L. G., & Berntson, G. G. (Eds.). Handbook of
psychophysiology (3rd ed.; 182-207). New York, NY: Cambridge University Press.
Berntson, G.G., Quigley, K.S., & Lozano, D. (2007). Cardiovascular psychophysiology. In
Cacioppo, John T., Tassinary, L. G., & Berntson, G. G. (Eds.). Handbook of
psychophysiology (3rd ed.; 182-207). New York, NY: Cambridge University Press.
Berrill, K.T., & Herek, G.M. (1990). Primary and secondary victimization in anti-gay hate
crimes: Official response and public policy. Journal of Interpersonal Violence, 5(3), 401413.
Birkett, M., Espelage, D.L., & Koenig, B. (2009). LGB and questioning students in schools: The
moderating effects of homophobic bullying and school climate on negative outcomes.
Journal of Youth and Adolescence, 38, 989-1000.
Blanchard, R. J., McKittrick, C. R., & Blanchard, D. C. (2001). Animal models of social stress:
effects on behavior and brain neurochemical systems. Physiology & behavior, 73(3), 261271.
Blumenthal, T.D., Cuthbert, B.N., Filion, D.L., Hackley, S., Lipp, O.V., & Boxtel, A.V. (2005).
Committee report: Guidelines for human startle eyeblink electromyographic studies.
Psychophysiology, 42, 1-15.
Bombay, A., Matheson, K., & Anisman, H. (2010). Decomposing identity: Differential
relationships between several aspects of ethnic identity and the negative effects of
perceived discrimination among First Nations adults in Canada. Cultural Diversity and
Ethnic Minority Psychology, 16(4), 507-516.
Bowling, N.A., & Beehr, T.A. (2006). Workplace harassment from the victim’s perspective: A
theoretical model and meta-analysis. Journal of Applied Psychology, 91, 998-1012.
Bovin, M.J., & Marx, B.P. (2011). The importance of the peritraumatic experience in defining
traumatic stress. Psychological Bulletin, 137(1), 47-67.

196

Boyle, C. (May 20, 2012). Ex-roommate of gay Rutgers teen faces sentencing. The New York
Daily News. Retrieved from http://www.nydailynews.com/news/ex-roommate-gayrutgers-teen-faces-sentencing-article-1.1081299?localLinksEnabled=false.
Bradley, M. M., Codispoti, M., Cuthbert, B. N., & Lang, P. J. (2001). Emotion and motivation I:
Defensive and appetitive reactions in picture processing. Emotion, 1(3), 276.
Bradley, M.M., & Lang, P.J. (2000). Affective reactions to acoustic stimuli. Psychophysiology,
37, 204-215.
Bradley, M.M., & Lang, P.J. (2007). Emotion and motivation. In Cacioppo, John T., Tassinary,
L. G., & Berntson, G. G. (Eds.). Handbook of psychophysiology (3rd ed.; 581-605). New
York, NY: Cambridge University Press.
Bradley, M. M., & Lang, P. J. (1994). Measuring emotion: the self-assessment manikin and the
semantic differential. Journal of Behavior Therapy and Experimental Psychiatry, 25(1),
49-59.
Branscombe, N.R., Schmitt, M.T., & Harvey, R.D. (1999). Perceiving pervasive discrimination
among African-Americans: Implications for group identification and well-being. Journal
of Personality and Social Psychology, 77, 135-149.
Brooke, J. (October 13, 1998). Gay man dies from attack, fanning outrage and debate. The New
York Times. Retrieved from http://www.nytimes.com/1998/10/13/us/gay-man-dies-fromattack-fanning-outrage-and-debate.html?pagewanted=1
Bryant, R. A., & Guthrie, R. M. (2007). Maladaptive self-appraisals before trauma exposure
predict posttraumatic stress disorder. Journal of Consulting and Clinical Psychology,
75(5), 812-815.
Bryant-Davis, T. (2007). Healing requires recognition: The case for race-based traumatic stress.
The Counseling Psychologist, 35(1), 135-143.
Burgard, S.A., Cochran, S., & Mays, V. (2005). Alcohol and tobacco use patterns among
heterosexually and homosexually experienced California women. Drug and Alcohol
Dependence, 77(1), 61-70.
Burrow, A.L., & Ong, A.D. (2010). The moderating role of racial identity in exposure and
reactivity to daily racial microaggressions. Self and Identity, 9, 383-402.
Button, K.S., Ioannidis, J.P.A., Mokrysz, C., Nosek, B.A., Flint, J., Robinson, E.S.J., & Munafὸ,
M.R. (2013). Power failure: Why small sample size undermines the reliability of
neuroscience. Nature Reviews Neuroscience, 14, 365-376.

197

Cacioppo, J.T., Petty, R.E., Losch, M.E., & Kim, H.S. (1986). Electromyographic activity over
facial muscles regions can differentiate the valence and intensity of affective reactions.
Journal of Personality and Social Psychology, 50, 260-268.
Carter, R.T. (2007). Racism and psychological and emotional injury: Recognizing and assessing
race-based traumatic stress. The Counseling Psychologist, 35(1), 13-105.
Casada, J.H., Amdur, R., Larsen, R., & Liberzon, I. (1998). Psychophysiologic responsivity in
posttraumatic stress disorder: Generalized hyperresponsiveness versus trauma specificity.
Biological Psychiatry, 44, 1037-1044.
Clark, R., Anderson, N.B., Clark, V.R., & Williams, D.R. (1999). Racism as a stressor for
African Americans: A biopsychosocial model. American Psychologist, 54(10), 805-16.
Cochran, S. D. (2001). Emerging issues in research on lesbians and gay men's mental health:
Does sexual orientation really matter. American Psychologist, 56, 931-947.
Cochran, S.D., & Mays, V.M. (2000). Lifetime prevalence of suicide symptoms and affective
disorders among men reporting same-sex sexual partners: Results from NHANES III.
American Journal of Public Health, 90(4), 573-578.
Cochran, S.D., Sullivan, J.G., & Mays, V.M. (2003). Prevalence of mental disorders,
psychological distress, and mental health services use among lesbian, gay, and bisexual
adults in the United States. Journal of Consulting and Clinical Psychology, 71, 53-61.
Cohen, H., Zohar, J., & Matar, M. (2003). The relevance of differential response to trauma in an
animal model of posttraumatic stress disorder. Biological Psychiatry, 53(6), 463-473.
Cohen, H., Kotler, M., Matar, M. A., Kaplan, Z., Loewenthal, U., Midodownik, H., & Cassuto,
Y. (1998). Analysis of heart rate variability in posttraumatic stress disorder patients in
response to a trauma-related reminder. Biological Psychiatry, 44, 1054-1059.
Cohen, J. (1992). A power primer. Psychological Bulletin, 112(1), 155-159.
Collinsworth, L.L., Fitzgerald, L.F., & Drasgow, F. (2009). In harm’s way: Factors related to
psychological distress following sexual harassment. Psychology of Women Quarterly,
33(4), 475-490.
Connor-Smith, J. K., & Compas, B. E. (2002). Vulnerability to social stress: Coping as a
mediator or moderator of sociotropy and symptoms of anxiety and depression. Cognitive
Therapy and Research, 26(1), 39-55.
Copas, A.J., Welling, K., Erens, B., Mercer, C.H., McManus, S., Fenton, K.A., et al. (2002). The
accuracy of reported sensitive sexual behavior in Britain: Exploring the extent of change
1990-2000. Journal of Sexually Transmitted Infections, 78, 26-30.

198

Correia, I., Alves, H., Sutton, R., Ramos, M., Gouveia-Pereira, M., & Vala, J. (2012). When do
people derogate or psychologically distance themselves from victims? Belief in a just
world and ingroup identification. Personality and Individual Differences, 53(6), 747-752.
Cortina, L.M., Fitzgerald, L.F., & Drasgow, F. (2002). Contextualizing Latina experiences of
sexual harassment: Preliminary tests of a structural model. Basic and Applied Social
Psychology, 24, 295-311.
Cortina, L.M., & Wasti, S.A. (2005). Profiles in coping: Responses to sexual harassment across
persons, organizations, and cultures. Journal of Applied Psychology, 90, 182-192.
Cousineau, D. (2005). Confidence intervals in within-subject designs: A simpler solution to
Loftus and Masson’s method. Tutorial in Quantitative Methods for Psychology, 1(1), 4–
45.
Craig, K. M. (1999). Retaliation, fear, or rage: An investigation of African American and White
reactions to racist hate crimes. Journal of Interpersonal Violence, 14, 2, 138-151.
Cramer, R.J., McNiel, D.E., Holley, S.R., Shumway, M., & Boccellari, A. (2012). Mental health
in violent crime victims: Does sexual orientation matter? Law and Human Behavior,
36(2), 87-95.
D’Augelli, A.R. (1992). Lesbian and gay male undergraduates’ experiences of harassment and
fear on campus. Journal of Interpersonal Violence, 7, 383-395.
D’Augelli, A.R., Grossman, A.H., & Starks, M.T. (2006). Childhood gender atypicality,
victimization, and posttraumatic stress disorder among lesbian, gay, and bisexual youth.
Journal of Interpersonal Violence, 21(11), 1462-1482.
D’Augelli, A.R., Pilkington, N.W., & Hershberger, S.L. (2002). Incidence and mental health
impact of sexual orientation victimization of lesbian, gay, and bisexual youths in high
school. School Psychology Quarterly, 17, 148-167.
Dagleish, T. (2004). Cognitive approaches to posttraumatic stress disorder: The evolution of
multirepresentational theorizing. Psychological Bulletin, 130(2), 228-260.
Dawson, M.E., Schell, A.M., & Filion, D.L. (2007). The electrodermal system. In Cacioppo,
John T., Tassinary, L. G., & Berntson, G. G. (Eds.). Handbook of psychophysiology (3rd
ed.; 159-181). New York, NY: Cambridge University Press.
Dean, L., Meyer, I.H., Robinson, K., Sell, R.L., Sember, R., Silenzio, V.M.B., et al., (2000).
Lesbian, gay, bisexual, and transgender health: Findings and concerns. Journal of the
Gay and Lesbian Medical Association, 4(3), 102-151.
Dekel, R., Solomon, Z., Elklit, A., & Ginzburg, K. (2004). World assumptions and combatrelated posttraumatic stress disorder. The Journal of social psychology, 144(4), 407-420.
199

DeSantis, A.S., Adam, E.K., Doane, L.D., Mineka, S., Zinbarg, R.E., & Craske, M.G. (2007).
Racial/ethnic differences in cortisol diurnal rhythms in a community sample of
adolescents. Journal of Adolescent Health, 41(1), 3-13.
Diaz, R.M., Ayala, G., Bein, E., Henne, J., & Marin, B.V. (2001). The impact of homophobia,
poverty, and racism on the mental health of gay and bisexual Latino men: Findings from
3 U.S. cities. American Journal of Public Health, 91(6).
Dickson, N., Paul, C., & Herbison, P. (2003). Same-sex attraction in a birth cohort: Prevalence
and persistence in early adulthood. Social Science & Medicine, 56, 1607-1615.
Dragowski, E.A., Halkitis, P.N., Grossman, A.H., & D’Augelli, A.R. (2011). Sexual orientation
victimization and posttraumatic stress symptoms among lesbian, gay, and bisexual youth.
Journal of Gay & Lesbian Social Services, 23(2), 226-249.
Dunbar, E. (2006). Race, gender, and sexual orientation in hate crime victimization: Identity
politics or identity risk? Violence and Victims, 21(3), 323-337.
Dunbar, E. W., Liu, J. F. & Horvath, A. M. (1995). Coping with culture-based conflict:
Implications for counseling research and practice. Cultural Diversity & Mental Health, 1,
2, 139-148.
Duncan, D. T., & Hatzenbuehler, M. L. (2014). Lesbian, gay, bisexual, and transgender hate
crimes and suicidality among a population-based sample of sexual-minority adolescents
in Boston. American Journal of Public Health, 104(2), 272-278.
Dworkin, S.H., & Yi, H. (2003). LGBT identity, violence, and social justice: The psychological
is political. International Journal for the Advancement of Counselling, 25(4), 269-279.
Edgar, J.C., Keller, J., Heller, W., & Miller, G.A. (2007). Psychophysiology in research on
psychopathology. In Cacioppo, John T., Tassinary, L. G., & Berntson, G. G. (Eds.).
Handbook of psychophysiology (3rd ed.; 665-683). New York, NY: Cambridge
University Press.
Ehlers, A., Mayou, R. A., & Bryant, B. (2003). Cognitive predictors of posttraumatic stress
disorder in children: Results of a prospective longitudinal study. Behaviour Research and
Therapy, 41(1), 1-10.
Ehring, T., Ehlers, A., & Glucksman, E. (2008). Do cognitive models help in predicting the
severity of posttraumatic stress disorder, phobia, and depression after motor vehicle
accidents? A prospective longitudinal study. Journal of Consulting and Clinical
Psychology, 76(2), 219.
Elhai, J. D., Engadahl, R. M., Palmieri, P. A., Nifeh, J. A., Schweinle, A., & Jacobs, G. (2009).
Assessing posttraumatic stress disorder with or without reference to a single, worst
200

traumatic event: Examining differences in factor structure. Psychological Assessment, 21,
629-634.
Eliezer, D., Major, B., & Mendes, W. B. (2010). The costs of caring: Gender identification
increases threat following exposure to sexism. Journal of Experimental Social
Psychology, 46(1), 159-165.
Elklit, A., Shevlin, M., Solomon, Z, & Dekel, R. (2007). Factor structure and concurrent validity
of the World Assumptions Scale. Journal of Traumatic Stress, 20(3), 291-301.
Fanselow, M. S. (1994). Neural organization of the defensive behavior system responsible for
fear. Psychonomic Bulletin & Review, 1(4), 429-438.
Fassinger, R.E., & Miller, B.A. (1996). Validation of an inclusive model of sexual minority
identity formation on a sample of gay men. Journal of Homosexuality, 32(2), 53-78.
Federal Bureau of Investigation (1998). Hate Crime Statistics, 1998. Retrieved from
http://www.fbi.gov/about-us/cjis/ucr/hate-crime/1998
Federal Bureau of Investigation (2012). Hate Crime Statistics, 2012. Retrieved from
http://www.fbi.gov/about-us/cjis/ucr/hate-crime/2012
Feldman, D.B., & Kaal, K.J. (2007). Vicarious trauma and assumptive worldview: Beliefs about
the world in acquaintances of trauma victims. Traumatology, 13(3), 21-31.
Ferguson, C.J. (2009). An effect size primer: A guide for clinicians and researchers. Professional
Psychology: Research and Practice, 40(5), 532-538.
Fergusson, D.M., Horwood, L.J., Riddler, E.M., & Beautrais, A.L. (2005). Sexual orientation
and mental health in a birth cohort of young adults. Psychological Medicine, 25, 971-981.
Floyd, F.J., & Stein, T.S. (2002). Sexual orientation identity formation among gay, lesbian, and
bisexual youths: Multiple patterns of milestone. Journal of Research on Adolescence,
12(2), 167-191.
Foa, E.B., Tolin, D.F., Ehlers, A., Clark, D.M., & Orsillo, S.M. (1999). The Post-Traumatic
Cognitions Inventory (PTCI): Development and validation. Psychological Assessment,
11(3), 303-314.
Foa, E.B., Zinbarg, R., & Rothbaum, B.O. (1992). Uncontrollability and unpredictability in PostTraumatic Stress Disorder: An animal model. Psychological Bulletin, 112(2), 218-238.
Fowles, D.C., Christie, M.J., Edelberg, R., Grings, W.W., Lykken, D.T., & Venables, P.H.
(1981). Publication recommendations for electrodermal measurements.
Psychophysiology, 18(3), 232-239.

201

Frost, D.M., & Bastone, L.M. (2008). The role of stigma concealment in the retrospective high
school experiences of gay, lesbian, and bisexual individuals. Journal of LGBT Youth,
5(1), 27-36.
Furnham, A. (2003). Belief in a just world: Research progress over the past decade. Personality
and Individual Differences, 34(5), 795-817.
Gee, G.C., Ro, A., Shariff-Marco, S., & Chae, D. (2009). Racial discrimination and health
among Asian Americans: Evidence, assessment, and directions for future research.
Epidemiologic Reviews, 31(1), 130-151.
Geen, J. (October 28, 2009). Mother of Matthew Shepard welcomes U.S. hate crimes bill. Pink
News. Retrieved from http://www.pinknews.co.uk/2009/10/28/mother-of-matthewshepard-welcomes-us-hate-crimes-bill
Gerardi, R.J., Keane, T.M., Cahoon, B.J., & Klauminzer, G.W. (1994). An in vivo assessment of
physiological arousal in posttraumatic stress disorder. Journal of Abnormal Psychology,
103(4), 825-827.
Gerlach, A. L., Wilhelm, F. H., Gruber, K., & Roth, W. T. (2001). Blushing and physiological
arousability in social phobia. Journal of Abnormal Psychology, 110(2), 247-258.
Gilman, S.E., Cochran, S.D., Mays, V.M., Hughes, M., Ostrow, D., & Kessler, R.C. (2001). Risk
of psychiatric disorders among individuals reporting same-sex sexual partners in the
National Comorbidity Survey. American Journal of Public Health, 91(6), 933-939.
Globisch, J., Hamm, A.O., Esteves, F., & Öhman, A. (1999). Fear appears fast: Temporal course
of startle reflex potentiation in animal fearful subjects. Psychophysiology, 36, 66-75.
Goldberg, L.R. (1999). A broad-bandwidth, public domain, personality inventory measuring the
lower-level facets of several five-factor models. In I. Mervielde, I. Deary, F. De Fruyt, &
F. Ostendorf (Eds.), Personality Psychology in Europe, Vol. 7 (pp. 7-28). Tilburg, The
Netherlands: Tilburg University Press.
Goldberg, L.R., Johnson, J.A., Eber, H.W., Hogan, R., Ashton, M.C., Cloninger, C.R., & Gough,
H.C. (2006). The International Personality Item Pool and the future of public-domain
personality measures. Journal of Research in Personality, 40, 84-96.
Goldenberg, I., & Matheson, K. (2005). Inner representations, coping, and posttraumatic stress
symptomatology in a community sample of trauma survivors. Basic and Applied Social
Psychology, 27, 361-369.
Gray, M.J., Litz, B.T., Hsu, J.L., & Lombardo, T.W. (2004). Psychometric properties of the Life
Events Checklist. Assessment, 11(4), 330-341.

202

Green, K.E., & Feinstein, B.A. (2012). Substance use in lesbian, gay, and bisexual populations:
An update on empirical research and implications for treatment. Psychology of Addictive
Behaviors, 26(2), 265-278.
Greenhouse, S.W., & Geisser, S. (1959). On Methods in the Analysis of Profile Data.
Psychometrika, 24, 95-112.
Greenwald, A.G. (1981). Self and memory. Psychology of Learning and Motivation, 15, 201236.
Gros, D.F., Hawk, L.W., & Moscovitch, D.A. (2009). The psychophysiology of social anxiety:
Emotional modulation of the startle reflex during socially-relevant and –irrelevant
pictures. International Journal of Psychophysiology, 73(3), 207-211.
Hafer, C. L., & Begue, L. (2005). Experimental research on just-world theory: problems,
developments, and future challenges. Psychological Bulletin, 131(1), 128-167.
Hagenaars, M.A., van Minnen, A., & Hoogduin, K.A.L. (2007). Peritraumatic psychological and
somatoform dissociation in predicting PTSD symptoms: A prospective study. The
Journal of Nervous and Mental Disease, 195(1), 952-954.
Halligan, S.L., Michael, T., Clark, D.M., & Ehlers, A. (2003). Posttraumatic stress disorder
following assault: The role of cognitive processing, trauma memory, and appraisals.
Journal of Consulting and Clinical Psychology, 71(5), 419-431.
Hamm, A. O., Cuthbert, B. N., Globisch, J., & Vaitl, D. (1997). Fear and the startle reflex: Blink
modulation and autonomic response patterns in animal and mutilation fearful subjects.
Psychophysiology, 34(1), 97-107.
Hamm, A.O., & Weike, A.I. (2005). The neuropsychology of fear learning and fear regulation.
International Journal of Psychophysiology, 57(1), 5-14.
Harrell, S.P. (2000). A multidimensional conceptualization of racism-related stress: Implications
for the well-being of people of color. American Journal of Orthopsychiatry, 70(1), 42-57.
Harrell, J. P., Halls, S. & Taliaferro, J. (2003). Physiological responses to racism and
discrimination: an assessment of the evidence. Am. J. Public Health, 93, 243-48.
Harris, H.N., & Valentiner, D.P. (2002). World assumptions, sexual assault, depression, and
fearful attitudes toward relationships. Journal of Interpersonal Violence, 17, 286-305.
Hatzenbuehler, M. L., & McLaughlin, K. A. (2014). Structural stigma and hypothalamic–
pituitary–adrenocortical axis reactivity in lesbian, gay, and bisexual young adults. Annals
of Behavioral Medicine, 47(1), 39-47.

203

Hatzenbuehler, M. L., McLaughlin, K. A., Keyes, K. M., & Hasin, D. S. (2010). The impact of
institutional discrimination on psychiatric disorders in lesbian, gay, and bisexual
populations: a prospective study. American Journal of Public Health, 100(3), 452-459.
Hellman, R.E., Sudderth, L., & Avery, A.M. (2002). Major mental illness in a sexual minority
psychiatric sample. Journal of the Gay and Lesbian Medical Association, 6(3-4), 97-106.
Herek, G. M., Gillis, J. R. & Cogan, J. C. (1999). Psychological sequelae of hate crime
victimization among lesbian, gay, and bisexual adults. Journal of Consulting & Clinical
Psychology, 67, 6, 945-951.
Herek, G.M, Gillis, J.R., Cogan, J.C., & Glunt, E.K. (1997). Hate crime victimization among
lesbian, gay, and bisexual adults: Prevalence, psychological correlates, and
methodological issues. Journal of Interpersonal Violence, 12, 195-215.
Herrell, R., Goldberg, J., True, W. R., Ramakrishnan, V., Lyons, M., Eisen, S., & Tsuang, M. T.
(1999). Sexual orientation and suicidality: A co-twin control study in adult men. Archives
of General Psychiatry, 56(10), 867-874.
Hitlan, R.T., Schneider, K.T., & Walsh, B. (2006). Upsetting behavior: Reactions to personal and
bystander sexual harassment experiences. Sex Roles: A Journal of Research, 55, 187-195.
Hogg, M.A., & Williams, K.D. (2000). From I to we: Social identity and the collective self.
Group Dynamics: Theory, Research, and Practice, 4, 81-97.
Houkamau, C.A., & Sibley, C.G. (2010). The Multidimensional Model of Maori Identity and
Cultural Engagement. New Zealand Journal of Psychology, 39(1), 8-28.
House, A.S., Van Horn, E., Coppeans, C., & Stepleman, L.M. (2011). Interpersonal trauma and
discriminatory events as predictors of suicidal and nonsuicidal self-injury in gay, lesbian,
bisexual, and transgender persons. Traumatology, 17(2), 75-85.
Janoff-Bulman, R. (1989). Assumptive worldviews and the stress of traumatic events:
Applications of the schema construct. Social Cognition, Special Issue: Social Cognition
and Stress, 7, 113-136.
Kaiser, C.B., Vick, S.B., & Major, B. (2004). A prospective investigation of the relationship
between just-world beliefs and the desire for revenge after September 11, 2001.
Psychological Science, 15(7), 503-506.
Kaler, M.E. (2009). The World Assumptions Questionnaire: Development of a measure of the
assumptive world (Doctoral dissertation). Retrieved from ProQuest Dissertations and
Theses Database. (AAT 3371864).

204

Kaler, M.E., Frazier, P.A., Anders, S.L., Tashiro, T., Tomich, P., Tennen, H., & Park, C. (2008).
Assessing the properties of the World Assumptions Scale. Journal of Traumatic Stress,
21(3), 326-332.
Karl, A., Rabe, S., Zöllner, T., Maercker, A., & Stopa, L. (2009). Negative self-appraisals in
treatment-seeking survivors of motor vehicle accidents. Journal of Anxiety Disorders,
23(6), 775-781.
Keary, T.A., Hughes, J.W., & Palmieri, P.A. (2009). Women with posttraumatic stress disorder
have larger decreases in heart rate variability during stress tasks. International Journal of
Psychophysiology, 73(3), 257-264.
Khaylis, A., Waelde, L., & Bruce, E. (2007). The role of ethnic identity in the relationship of
race-related stress to PTSD symptoms among young adults. Journal of Trauma &
Dissociation, 8(4), 91-105.
Kiang, L., Blumenthal, T., Carlson, E.N., Lawson, Y.N., & Shell, J.C. (2009). Physiological
responses to racial rejection images among African American youth. Journal of Youth
and Adolescence, 38, 164-174.
Kilpatrick, D.G., Saunders, B.E., Amick-McMullen, A., Best, C.L., Veronen, L.G., & Resnick,
H.S. (1989). Victims and crime factors associated with the development of crime-related
posttraumatic stress disorder. Behavior Therapy, 20, 199-214.
Kinsey, A.C., Pomeroy, W.B., & Martin, C.E. (1948). Sexual Behavior in the Human Male.
Philadelphia, PA: W.B. Saunders.
Klorman R., Weissbert R.P., Wiessenfeld A.R. (1977). Individual differences in fear and
autonomic reactions to affective stimulation. Psychophysiology, 14, 45–51.
Kolts, R.L., Robinson, A.M., & Tracy, J.J. (2004). The relationship of sociotropy and autonomy
to posttraumatic cognitions and PTSD symptomatology in trauma survivors. Journal of
Clinical Psychology, 60(1), 53-63.
Kuyper, L., & Vanwesenbeeck, I. (2009). High levels of same-sex experiences in the
Netherlands: Prevalences of same-sex experiences in historical and international
perspective. Journal of Homosexuality, 56(8), 993-1010.
Lanius, R.A., Bluhm, R.L., & Frewen, P.A. (2011). How understanding the neurobiology of
complex post-traumatic stress disorder can inform clinical practice: A social cognitive
and affective neuroscience approach. Acta Psychiatrica Scandinavica, 124(5), 331-348.
Lang, P. J. (1980). Behavioral treatment and bio-behavioral assessment: Computer applications.
In J. B. Sidowski, J. H. Johnson, & T. A. Williams (Eds.), Technology in mental health
care delivery (pp. 119-137). Norwood, NJ: Ablex.

205

Lang, P.J. (1995). The emotion probe: Studies of motivation and attention. American
Psychologist, 50(5), 372-385.
Lang, P.J., Bradley, M.M., & Cuthbert, B.N. (1997). International Affective Picture System
(IAPS): Technical Manual and Affective Ratings. NIMH Center for the Study of Emotion
and Attention, University of Florida, Gainesville, FL.
Lang, P.J., Davis, M., & Öhman, A. (2000). Fear and anxiety: Animal models and human
cognitive psychophysiology. Journal of Affective Disorders, 61(3), 137-159.
Lannert, B.K. (2014). Traumatogenic processes and pathways to mental health outcomes for
sexual minorities exposed to bias crime information. Trauma, Violence, and Abuse.
Advance online publication. doi: 10.1177/1524838014526067
Laposa, J. M., & Alden, L. E. (2003). Posttraumatic stress disorder in the emergency room:
exploration of a cognitive model. Behaviour Research and Therapy, 41(1), 49-65.
Laumann, E.O., Gagnon, J., Michael, R.T., & Michaels, S. (1994). The social organization of
sexuality: Sexual practices in the United States. Chicago: University of Chicago Press.
Leharot, K., Simon, J.M. (2011). The impact of minority stress on mental health and substance
use among sexual minority women. Journal of Consulting and Clinical Psychology, 79,
159-170.
Leonardelli, G.J., & Tormala, Z.L. (2003). The negative impact of perceiving discrimination on
collective well-being: The mediating role of perceived ingroup status. European Journal
of Social Psychology, 33(4), 507-514.
Lepore, S.J., Revenson, T.A., Weinberger, S.L., Weston, P., Frisina, P.G., Robertson, R., Mentor
Portillo, M., Jones, H., & Cross, W. (2006). Effects of social stressors on cardiovascular
reactivity in Black and White women. Annals of Behavioral Medicine, 31, 120-127.
Lewis, R.J., Derlega, V.J., Griffin, J.L., Krowinski, A.C. (2003). Stressors for gay men and
lesbians: Life stress, gay-related stress, stigma consciousness, and depressive symptoms.
Journal of Social Psychology, 22(6), 716-729.
Levitt, H.M., Ovrebo, E., Anderson-Cleveland, M.B., Leone, C., Jeong, J.Y., Arm, J.R. & Horne,
S.G. (2009). Balancing dangers: GLBT experience in a time of anti-GLBT legislation.
Journal of Counseling Psychology, 56(1), 67-81.
Leonardelli, G.J., & Tormala, Z.L. (2003). The negative impact of perceiving discrimination on
collective well-being: The mediating role of perceived ingroup status. European Journal
of Social Psychology, 33(4), 507-514.

206

Lilly, M. M., Valdez, C. E., & Graham-Bermann, S. A. (2011). The mediating effect of world
assumptions on the relationship between trauma exposure and depression. Journal of
Interpersonal Violence, 26(12), 2499.
Lim, S.G.P., & Cortina, L.M. (2005). Interpersonal mistreatment in the workplace: The interface
and impact of general incivility and sexual harassment. Journal of Applied Psychology,
90, 483-496.
Linley, P.A., & Joseph, S. (2004). Positive change following trauma and adversity: A review.
Journal of Traumatic Stress, 17(1), 11-21.
Lipkus, I. M., Dalbert, C., & Siegler, I. C. (1996). The importance of distinguishing the belief in
a just world for self versus for others: Implications for psychological well-being.
Personality and Social Psychology Bulletin, 22(7), 666-677.
Little, R.J.A., & Rubin, D.B. (1987). Statistical Analysis with Missing Data. New York, New
York: Wiley.
Long, M. E., & Elhai, J. D. (2009). Posttraumatic stress disorder's traumatic stressor criterion:
History, controversy, clinical and legal implications. Psychological Injury and Law, 2,
167-178.
Long, M. E., Elhai, J. D., Schweinle, A., Gray, M. J., Grubaugh, A. L., & Frueh, B. C. (2008).
Differences in posttraumatic stress disorder diagnostic rates and symptoms severity
between Criterion A1 and non-Criterion A1 stressors. Journal of Anxiety Disorders, 22,
1255-1263.
MacCallum, R.C., Widaman, K.F., Zhang, S., & Hong, S. (1999). Sample size in factor analysis.
Psychological Methods, 4(1), 84-99.
Mauchly, J. W. (1940). Significance test for sphericity of a normal n-variate distribution. The
Annals of Mathematical Statistics, 11, 204-209.
Mays, V.M., & Cochran, S.D. (2001). Mental health correlates of perceived discrimination
among lesbian, gay, and bisexual adults in the United States. American Journal of Public
Health, 91(11), 1869-1876.
McCabe, S., Hughes, T., Bostwick, W., West, B., & Boyd, C. (2009). Sexual orientation,
substance use behaviors and substance dependence in the United States. Addiction, 104,
1333-1345.
McDermut, J.F., Haaga, D.A.F., & Kirk, L. (2000). An evaluation of stress symptoms associated
with academic sexual harassment. Journal of Traumatic Stress, 13(3), 397-411.
McDevitt, J., Balboni, J., Garcia, L., & Gu, J. (2001). Consequences for victims: A comparison
of bias-and non-bias-motivated assaults. American Behavioral Scientist, 45(4), 697-713.
207

McDonagh-Coyle, A., McHugo, G. J., Friedman, M. J., Schnurr, P. P., Zayfert, C., & Descamps,
M. (2001). Psychophysiological reactivity in female sexual abuse survivors. Journal of
Traumatic Stress, 14(4), 667-683.
McTeague, L. M., & Lang, P. J. (2012). The anxiety spectrum and the reflex physiology of
defense: from circumscribed fear to broad distress. Depression and Anxiety, 29(4), 264281.
McTeague, L.M., Lang, P.J., Laplant, M.C., Cuthbert, B.N., Shumen, J.R., & Bradley, M.M.
(2010). Aversive imagery in Post-Traumatic Stress Disorder: Trauma recurrence,
comorbidity, and physiological reactivity. Biological Psychiatry, 67(4), 346-356.
McTeague, L. M., Lang, P.J., Laplante, M.C., Cuthbert, B.N., Strauss, C.C., & Bradley, M.M.
(2009). Fearful imagery in social phobia: generalization, comorbidity, and physiological
reactivity. Biological Psychiatry, 65(5), 374-382.
Meyer, I.H. (1995). Minority stress and mental health in gay men. Journal of Health and Social
Behavior, 36(1), 38-56.
Meyer, I. M. (2003). Prejudice, social stress, and mental health in lesbian, gay, and bisexual
populations: Conceptual issues and research evidence. Psychological Bulletin, 129, 674697.
Meyer, D., & Grollman, E. A. (2014). Sexual orientation and fear at night: Gender differences
among sexual minorities and heterosexuals. Journal of Homosexuality, 61(4), 453-470.
Midanik, L. T., Drabble, L., Trocki, K., & Sell, R. L. (2007). Sexual orientation and alcohol use:
Identity versus behavior measures. Journal of LBGT Health Research, 3(1), 25-35.
Miner-Rubino, K., & Cortina, L.M. (2004). Working in a context of hostility toward women:
Implications for employees’ wellbeing. Journal of Occupational Health Psychology, 9,
107-122.
Miner-Rubino, K., & Cortina, L.M. (2007). Beyond targets: Consequences of vicarious exposure
to misogyny at work. Journal of Applied Psychology, 92, 1254-1269.
Moradi, B., Mohr, J.J., Worthington, R.L., & Fassinger, R.E. (2009). Counseling psychology
research on sexual (orientation) minority issues: Conceptual and methodological
challenges and opportunities. Journal of Counseling Psychology, 56(1), 5-22.
Morris, S.B., & DeShon, R.P. (2002). Combining Effect Size Estimates in Meta-Analysis with
Repeated Measures and Independent-Groups Designs. Psychological Methods, 7(1), 105125.

208

Moser, J. S., Hajcak, G., Simons, R. F., & Foa, E. B. (2007). Posttraumatic stress disorder
symptoms in trauma-exposed college students: The role of trauma-related cognitions,
gender, and negative affect. Journal of Anxiety Disorders, 21(8), 1039-1049.
Mueller, J., Moergeli, H., & Maercker, A. (2008). Disclosure and social acknowledgement as
predictors of recovery from posttraumatic stress: A longitudinal study in crime victims.
Canadian Journal of Psychiatry, 53(3), 160.
New details emerge in Matthew Shepard murder. (November 26, 2004). ABC News 20/20.
Retrieved from http://abcnews.go.com/2020/story?id=277685&page=1.
Nikula, R. (1991). Psychological correlates of nonspecific skin conductance responses.
Psychophysiology, 28(1), 86-90.
Nixon, R. D., & Bryant, R. A. (2005). Are negative cognitions associated with severe acute
trauma responses? Behaviour Change, 22(1), 22-28.
Noelle, M. (2002). The ripple effect of the Matthew Shepard murder: Impact on the assumptive
worlds of members of the targeted group. American Behavioral Scientist, 46(1), 27-50.
O'Donnell, M. L., Elliott, P., Wolfgang, B. J., & Creamer, M. (2007). Posttraumatic appraisals in
the development and persistence of posttraumatic stress symptoms. Journal of Traumatic
Stress, 20(2), 173-182.
Orr, S. P., Lasko, N. B., Metzger, L. J., Berry, N. J., Ahern, C. E., & Pitman, R. K. (1998).
Psychophysiologic assessment of women with posttraumatic stress disorder resulting
from childhood sexual abuse. Journal of Consulting and Clinical Psychology, 66(6), 906913.
Otis, M.D. (2007). Perceptions of victimization risk and fear of crime among lesbians and gay
men. Journal of Interpersonal Violence, 22(2), 198-217.
Otis, M.D., & Skinner, W.F. (1996). The prevalence of victimization and its effect on mental
well-being among lesbians and gay people. Journal of Homosexuality, 30, 93-121.
Ozer, E. J., Best, S. R., Lipsey, T. L., & Weiss, D. S. (2003). Predictors of posttraumatic stress
disorder and symptoms in adults: A meta-analysis. Psychological Bulletin, 129(1), 52-73.
Pascoe, E.A., & Smart Richman, L. (2009). Perceived discrimination and health: A meta-analytic
review. Psychological Bulletin, 135(4), 531-554.
Paterson, R.J., & Neufeld, R.W. (1987). Clear danger: Situational determinants of the appraisal
of threat. Psychological Bulletin, 101(3), 404-416.

209

Peri, T., Ben-Shakhar, G., Orr, S.P., & Shalev, A.Y. (2000). Psychophysiologic assessment of
aversive conditioning in posttraumatic stress disorder. Biological Psychiatry, 47(6), 512519.
Perry, B., & Alvi, S. (2012). ‘We are all vulnerable’: The in terrorem effects of hate crimes.
International Review of Victimology, 18(1), 57-71.
Pieterse, A.L., Carter, R.T., Evans, S.A., & Walter, R.A. (2010). An exploratory examination of
the associations among racial and ethnic discrimination, racial climate, and traumarelated symptoms in a college student population. Journal of Counseling Psychology,
57(3), 255-263.
Pitman, R., Orr, S., Forgue, D., de Jong, J., & Claiborn, J. (1987). Psychophysiologic assessment
of posttraumatic stress disorder imagery in Vietnam combat veterans. Archives of
General Psychiatry, 44, 970-975.
Pole, N. (2007). The psychophysiology of posttraumatic stress disorder: A meta-analysis.
Psychological Bulletin, 133(5), 725-746.
Poteat, V.P., Mereish, E.H., DiGiovanni, C.D., & Koenig, B.W. (2011). The effects of general
and homophobic victimization on adolescents’ psychosocial and educational concerns:
The importance of intersecting identities and parent support. Journal of Counseling
Psychology, 58(4), 597-609.
Quevedo, K., Smith, T., Donzella, B., Schunk, E., & Gunnar, M. (2009). The startle response:
Developmental effects and a paradigm for children and adults. Developmental
Psychobiology, 52, 78-89.
Quinn, D.M., & Chaudoir, S.R. (2009). Living with a concealable stigmatized identity: The
impact of anticipated stigma, centrality, salience, and cultural stigma on psychological
distress and health. Journal of Personality and Social Psychology, 97(4), 634-651.
Ramos, M. R., Correia, I., & Alves, H. (2014). To believe or not to believe in a just world? The
psychological costs of threats to the belief in a just world and the role of attributions. Self
and Identity, 13(3), 257-273.
Ranchor, A.V., Wardle, J., Steptoe, A., Henselmans, I., Ormel, J., & Sanderman, R. (2010). The
adaptive role of perceived control before and after cancer diagnosis: A prospective study.
Social Science and Medicine, 70(11), 1825-1831.
Raver, J.L., & Gelfand, M.J. (2005). Beyond the individual victim: Linking sexual harassment,
team processes, and team performance. Academy of Management Journal, 48, 388-400.
Regambal, M.J., & Alden, L.E. (2009). Pathways to intrusive memories in a trauma analogue
paradigm: A structural equation model. Depression and Anxiety, 26, 155-166.

210

Rhodes, J. E., & Jason, L. A. (1990). A social stress model of substance abuse. Journal of
consulting and clinical psychology, 58(4), 395.
Richman, L. S., & Jonassaint, C. (2008). The effects of race-related stress on cortisol reactivity
in the laboratory: Implications of the Duke lacrosse scandal. Annals of Behavioral
Medicine, 35(1), 105-110.
Roberts, A.L., Austin, S.B., Cortiss, H.L., Vander Morris, A.K., & Koenen, K.C. (2010).
Pervasive trauma exposure among U.S. sexual orientation minority adults and risk of
posttraumatic stress disorder. American Journal of Public Health, 100(12), 2433-2441.
Rose, S.M., & Mechanic, M.B. (2002). Psychological distress, crime features, and help-seeking
behaviors related to homophobic bias incidents. American Behavioral Scientist, 46(1),
14-26.
Rughinj, C. (2011). Measuring and using Roma/Gypsy ethnicity. Ethnic and Racial Studies,
34(4), 594-619.
Sanchez-Hucles, J.V. (1999). Racism: Emotional abusiveness and psychological trauma for
ethnic minorities. Journal of Emotional Abuse, 1, 69-87.
Sanders-Phillips, K. (2009). Racial discrimination: a continuum of violence exposure for
children of color. Clinical Child and Family Psychology Review, 12(2), 174-195.
Sandfort, T.G.M., Bakker, F., Schellevis, F., & Vanwesenbeeck, I. (2009). Coping styles as
mediator of sexual orientation-related health differences. Archives of Sexual Behavior,
38(2), 253-263.
Sandholtz, N., Langton, L. & Planty, M. (2013). Hate crime victimization, 2003-2011. Special
report NCJ 241291: U.S. Department of Justice, Office of Justice Programs, Bureau of
Justice Statistics. Retrieved from http://www.bjs.gov/content/pub/pdf/hcv0311.pdf.
Sapolsky, R.M. (2004). Why Zebras Don’t Get Ulcers (3rd Edition). New York, New York: Holt.
Savin-Williams, R. C. (1994). Verbal and physical abuse as Stressors in the lives of lesbian, gay
male and bisexual youths: Associations with school problems, running away, substance
abuse, prostitution, and suicide. Journal of Consulting & Clinical Psychology, 62, 2, 261269.
Savin-Williams, R.C., & Ream, G.L. (2007). Prevalence and stability of sexual orientation
components during adolescence and young adulthood. Archives of Sexual Behavior, 26,
385-394.
Sell, R.L. (1997). Defining and measuring sexual orientation: A review. Archives of Sexual
Behavior, 26(6), 643-658.

211

Sell, R. L., & Petrulio, C. (1996). Sampling homosexuals, bisexuals, gays, and lesbians for
public health research: A review of the literature from 1990 to 1992. Journal of
Homosexuality, 30(4), 31-47.
Sell, R. L., Wells, J.A., & Wypij, D. (1995). The prevalence of homosexual behavior and
attraction in the United States, the United Kingdom and France: Results of national
population-based samples. Archives of Sexual Behavior, 24(3), 235-249.
Sellers, R. M., & Shelton, J. N. (2003). Racial identity, discrimination, and mental health among
African Americans. Journal of Personality and Social Psychology, 84(5), 1079-1092.
Sellers, R.M., Smith, M.A., Shelton, J.N., Rowley, S.A.J., & Chavous, T.M. (1998).
Multidimensional model of racial identity: A reconceptualization of African American
racial identity. Personality and Social Psychology Review, 2(1), 18-39.
Settles, I.H., Navarrete, C.D., Pagano, S.J., Abdou, C.M., & Sidanius, J. (2010). Racial identity
and depression among African American women. Cultural Diversity and Ethnic Minority
Psychology, 16(2), 248-255.
Sheets, R.L., & Mohr, J. (2009). Perceived social support from friends and family and
psychosocial functioning in bisexual young adult college students. Journal of Counseling
Psychology, 56(1), 152-163.
Shupe, E.I., Cortina, L.M., Ramos, A., Fitzgerald, L.F., & Salisbury, J. (2002). The incidence
and outcomes of sexual harassment among Hispanic and non-Hispanic white women: A
comparison across levels of cultural affiliation. Psychology of Women Quarterly, 26(4),
298-308.
Smith, C. A., & Ellsworth, P. C. (1985). Patterns of cognitive appraisal in emotion. Journal of
personality and social psychology, 48(4), 813.
Smith, T.B., & Silva, L. (2011). Ethnic identity and personal well-being of people of color: A
meta-analysis. Journal of Counseling Psychology, 58(1), 42-60.
Smith, E. R., & Zarate, M. A. (1992). Exemplar-based model of social judgment. Psychological
Review, 99(1), 3-21.
Solomon, Z., Iancu, I., & Tyano, S. (1997). World assumptions following disaster. Journal of
Applied Social Psychology, 27(20), 1785-1798.
Steiger, J. H., & Lind, J. C. (1980). Statistically based tests for the number of common factors.
Annual Meeting of the Psychometric Society, Iowa City, IA (758), 424-453.
Steil, R., & Ehlers, A. (2000). Dysfunctional meaning of posttraumatic intrusions in chronic
PTSD. Behaviour Research and Therapy, 38(6), 537-558.

212

Sterling, P., & Eyer, J. (1988). Allostasis: A new paradigm to explain arousal pathology. In: S.
Fisher and J. Reason (Eds.), Handbook of Life Stress, Cognition and Health. J Wiley &
Sons, New York.
Stockdale, M.S., Logan, T.K., & Weston, R. (2009). Sexual harassment and posttraumatic stress
disorder: Damages beyond prior abuse. Law and Human Behavior, 33, 405-418.
Stroebe, K., Dovidio, J. F., Barreto, M., Ellemers, N., & John, M. S. (2011). Is the world a just
place? Countering the negative consequences of pervasive discrimination by affirming
the world as just. British Journal of Social Psychology, 50(3), 484-500.
Sullaway, M. (2004). Psychological perspectives on hate crime laws. Psychology, Public Policy,
and Law, 10, 250-292.
Tajfel, H. (Ed.). (1978). Differentiation between social groups: Studies in the social psychology
of intergroup relations. London: Academic Press.
Tangri, S.S., Burt, M.R., & Johnson, L.B. (1982). Sexual harassment at work: Three explanatory
models. Journal of Social Issues, 38(4), 33-54.
Tassinary, L.G., Cacioppo, J.T., & Vanman, E.J. (2007). The skeletomotor system: Surface
electromyography. In Cacioppo, John T., Tassinary, L. G., & Berntson, G. G. (Eds.).
Handbook of psychophysiology (3rd ed.; 267-293). New York, NY: Cambridge
University Press.
Taylor, S. (1999). Health Psychology (4th Edition). Los Angeles: McGraw Hill.
Taylor, S. E., & Brown, J. D. (1988). Illusion and well-being: a social psychological perspective
on mental health. Psychological Bulletin, 103(2), 193-210.
Taylor, S. E., & Brown, J. D. (1994). Positive illusions and well-being revisited: separating fact
from fiction. Psychological Bulletin, 116(1), 21-27.
Taylor, J., & Turner, R. J. (2002). Perceived discrimination, social stress, and depression in the
transition to adulthood: Racial contrasts. Social Psychology Quarterly, 65(3), 213-225.
Thoits, P.A. (2010). Stress and health: Major findings and policy implications. Journal of Health
and Social Behavior, 51(1), S41-S53.
Tjepkema, M. (2008). Health care use among gay, lesbian, and bisexual Canadians. Health
Reports, 19(1), 53-64.
Torkelson, J. (October 3, 2008). Mother’s mission: Matthew Shepard’s death changes things.
Rocky Mountain News. Retrieved from http://m.rockymountainnews.com/news/2008/oct/
03/10-years-later-matthew-shepard-hasnt-been-forgotte/

213

Turner, J.C., Oakes, P.J., Haslam, S.A., & McGarty, C. (1994). Self and collective: Cognition
and social context. Personality and Social Psychology Bulletin, 20(5), 454-463.
Ullman, S. E. (1997). Attributions, world assumptions, and recovery from sexual assault. Journal
of Child Sexual Abuse, 6(1), 1-19.
Villarroel, M.A., Turner, C.F., Eggleston, E., Al-Tayyib, A., Rogers, S.M., Roman, A.M., et al.
(2006). Same-gender sex in the United States: Impact of T-ACASI on prevalence
estimates. Public Opinion Quarterly, 70(2), 166-169.
Vrana, S. R., Spence, E. L., & Lang, P. J. (1988). The startle probe response: A new measure of
emotion?. Journal of abnormal psychology, 97(4), 487.
Vrangalova, Z., & Savin-Williams, R. C. (2012). Mostly heterosexual and mostly gay/lesbian:
Evidence for new sexual orientation identities. Archives of Sexual Behavior, 41(1), 85101.
Weerts, T. C., & Lang, P. J. (1978). Psychophysiology of fear imagery: differences between
focal phobia and social performance anxiety. Journal of Consulting and Clinical
Psychology, 46(5), 1157-1159.
Wendt, J., Lotze, M., Weike, A.I., Hosten, N., & Hamm, A.O (2008). Brain activation and
defensive response mobilization during sustained exposure to phobia-related and other
affective pictures in spider phobia. Psychophysiology, 45, 205-215.
Wickie, S. K., & Marwit, S. J. (2001). Assumptive world views and the grief reactions of parents
of murdered children. Journal of Death and Dying, 42(2), 101-114.
Wilkinson, L. (1999). Statistical methods in psychology journals: Guidelines and explanations.
American Psychology, 54(8), 594-604.
Williams, D.R., & Williams-Morris, R. (2000). Racism and mental health: The African
American experience. Ethnicity and Health, 5, 243-268.
Willness, C.R., Steel, P., & Lee, K. (2007). A meta-analysis of the antecedents and consequences
of workplace sexual harassment. Personnel Psychology, 60, 127-160.
Wolfe, D.A., Crooks, C.C., Chiodo, D, & Jaffe, P. (2009). Child maltreatment, bullying, genderbased harassment, and adolescent dating violence: Making the connections. Psychology
of Women Quarterly, 33(1), 21-24.
Woodford, M. R., Han, Y., Craig, S., Lim, C., & Matney, M. M. (2014). Discrimination and
mental health among sexual minority college students: The type and form of
discrimination does matter. Journal of Gay & Lesbian Mental Health, 18(2), 142-163.

214

Woodford, M. R., Howell, M. L., Silverschanz, P., & Yu, L. (2012). “That's so gay!”: Examining
the covariates of hearing this expression among gay, lesbian, and bisexual college
students. Journal of American College Health, 60(6), 429-434.
Woodford, M. R., Krentzman, A. R., & Gattis, M. N. (2012). Alcohol and drug use among
sexual minority college students and their heterosexual counterparts: the effects of
experiencing and witnessing incivility and hostility on campus. Substance abuse and
rehabilitation, 3, 11.
Woods, K.C., Buchanan, N.T., & Settles, I.H. (2009). Sexual harassment across the color line:
Experiences and outcomes of cross- vs. intra-racial sexual harassment among Black
women. Cultural Diversity and Ethnic Minority Psychology, 15, 67-76.
Worthington, R.L., & Reynolds, A.I. (2009). Within-group differences in sexual orientation and
identity. Journal of Counseling Psychology, 56(1), 44-55.
Wunsch, D. (1986). Survey research: Determining sample size and representative response.
Business Education Forum, 40(5), 31-34.
Yip, T., & Cross, W.E. (2004). A daily diary study of mental health and community involvement
outcomes for three Chinese American social identities. Cultural Diversity and Ethnic
Minority Psychology, 10(4), 394-408.

215