Copyright

by
Susana Lucia Kugeares
2002

The Dissertation Committee for Susana Lucia Kugeares

certifies that this is the approved version of the following dissertation:

Social Anxiety in Dating Initiation: An Experimental

Investigation of an Evolved Mating-Specific Anxiety Mechanism

Committee:
____________________________________
Michael J. Telch, Supervisor
____________________________________
David M. Buss, Co-supervisor
____________________________________
Patrick K. Randall
____________________________________
Cindy Meston
____________________________________
Devendra Singh
____________________________________
Frank Wicker

Social Anxiety in Dating Initiation: An Experimental

Investigation of an Evolved Mating-Specific Anxiety Mechanism

by
Susana Lucia Kugeares, B.A., M.A.

Dissertation
Presented to the Faculty of the Graduate School of
the University of Texas at Austin
in Partial Fulfillment
of the Requirements
for the Degree of

Doctor of Philosophy

The University of Texas at Austin

December, 2002

Dedication
To my family

Acknowledgements
Without question, the undertaking of this dissertation project has been the
most grueling but most rewarding task of my life thus far. Even years ago when I
was only entering graduate school (before the turn of this century), I recognized
the enormous fortitude and courage this project would require of me and the
commensurate pride that I would feel if I accomplished it. My premonitions
turned out to be true. But what I had not predicted was how great the contributions
from others would be in order for this project to move from birth to completion,
like a child that took a village to raise it.
Part of that village was in Austin, when I started this project. Thanks goes
to Dr. Judy Langlois who generously let me use faces from her research on
attractiveness to pilot my study, and to Jasper Smits, whose help on little things
was actually quite big. Thanks also to Dr. Jan Kamphuis, one of the most gifted
writers I have met, for his invaluable editorial comments on earlier drafts (and for
using the word elegant to describe my study). I play back that tape and have not
forgotten your help. On this project and in the development of my professional
confidence in general, I will always owe you at least a little thanks for the seeds
you planted.
The UT Psychology Department has been blessed with some of the most
capable and skilled computer support staff that I have come across in any setting
v

and I am pleased to have known them. Bruce Turbeville is a treasure in our

department and deserves thanks for more than just his extensive computer
knowledge. He was an advocate for us students and fought our battles even
though he was not asked. How lucky for me to have had an office just around the
corner from you in old Mezes Hall! I also credit both Bruce and Gary Zucker for
the final stimuli in this project. It was their initial idea to beef up the stimuli by
creating a logo and more for my fictitious dating service. Their simple five
minutes of off-the-cuff suggestions may very well be the reason for my significant
results. Thanks also to Susanna Douglas for her help with so many graphics and
printing questions, but also for her wisdom about my angst in leaving Austin, her
sweetness, and for her friendship. She is one of the most balanced and actualized
individuals I know, and I hope we find ourselves still in touch long from now.
My research assistants for this project also deserve thanks. They were not
only competent, creative, and bright, they made data collection the most fun phase
of this entire project. What a summer. My special thanks goes to Sean Maguire,
Rebecca Block, Dina Segovia, and Bernardo de la Fuente, in whom I entrusted the
most critical tasks of the data phase and who exceeded my expectations. I know
each of you are, and will continue to, excel in the directions you have chosen in
life and I would always be happy to help if I can.

vi

My dissertation committee was extraordinarily helpful and truly something

to brag about. All are experts who have made significant impacts in their
respective fields of research. Thanks to Dr. Dev Singh for broadening my
understanding of the biological side of this work and for his delicious cooking; to
Dr. Cindy Meston who balanced sweetness and fun with a sharp minded
suggestions for the study; and to Dr. Frank Wicker who complimented me by
joining the committee out of pure interest for the topic even though he was high in
demand. Finally, Dr. Pat Randall was a late addition to the committee but was
invaluable to me through the analyses and final defense. He went above and
beyond in making time for our meetings and in translating statistics into English
until I began to learn the language myself. Thank you for your patience and for
your certainty about my MVD results. I did not get to pick your brain nearly as
much as I wish. You are an invaluable asset to this department and a pleasure to
work to boot.
Dr. David Buss has been a huge influence in my scholarly growth. Hearing
Davids first lecture at UT, when he was still faculty at the University of
Michigan, lit a fire inside of me. How exciting for me to discover that a branch of
study actually existed that combined my two greatest intellectual loves! the study
of human behavior and emotion with the principles of evolution. David has never
failed to impress me with the precision of his thinking, his determination in

vii

reaching goals, his passion for his work, and his ability to transmit that excitement
to others. I hope that our paths continue to cross in the future, and if your lecture
invitations in the future ever lead you to Greece (or Tampa), let me know.
To Dr. Mike Telch, my committee chair and now a friend, I owe enormous
thanks. This dissertation simply would not have happened without his mentorship.
He took me on despite how novice I was about research and never flagged in his
support. His guidance with research design, statistical knowledge, and
conceptualization was superior. His intellectual creativity made it fun and his
enthusiasm reassured my at my deepest doubts. I cannot overstate how grateful I
am for his expertise, patience, and hospitality. Mike, as you said, it is a shame we
began working together so late in the game, but I hope that collaboration and
contact continues on.
The last of my Austin village were Penny Frohlich, Amanda Gregory,
Tracy Sloan, and Kevin Larson, who provided fun during the wee hours of the
morning still working in the computer lab, an ear for venting frustrations and
laments, and rationality during the inevitable crises that arose in this process and
innumerable other things that good friends do that help keep one motivated and
sane. Thanks for making my graduate school years the most fun and stimulating
ones of my life.

viii

As the dissertation process extended, for me, beyond state lines, I owe
certain thanks to folks in Florida as well. Thanks to Dr. Art Rosenblatt, the Chief
of Psychology at the Tampa VA hospital, to whom I owe being gainfully
employed without yet being dissertated. Drs. John Schinka and Glenn Curtiss
helped me walk my first statistical steps and, as true believers in science, directly
pointed out the flaws in my study. Both astute researchers and statisticians, they
were also great fun in meetings and I hope future research opportunities arise for
more of the same. Drs. Karen Nicholson, Sid Davis, and Ron Gironda have earned
my respect as well-rounded psychologists and my gratitude for their support of me,
each in his/her own unique way. My co-intern and now colleague at the VA, Dr.
Jennifer Duchnick, has been a support from day one, feeling joyful when I met yet
another goal and angry too when I met another obstacle. Her quiet determination
is an example to be followed. Thanks for accompanying me on this trek.
Help and support can come in many forms and my family provided most of
them. However, I want to especially thank my mother for being my library
connection. Her love of library research compensated for my lack of it and was
invaluable to me. My sister Zoe was a support from afar, and thanks to my father
who provided a cushion to fall on if I needed it and was always there to remind me
that I had already eaten the cow and only the tail remained. Sometimes, there is
no wisdom like that of the Greeks. I owe my sister, Christina, enormous thanks

ix

for so many things. Through the most chaotic times, she was a constant with
support as well as practical help, keeping my refrigerator stocked, my kitchen ratfree, and my clocks ticking. She was proud of me even during my discouraged
moments and was purely elated at the high points. I could not have asked for a
better cheerleader in my corner. Thank you all for believing in me.

Social Anxiety In Dating Initiation: An Experimental

Investigation of an Evolved Mating-Specific Anxiety Mechanism

Publication No.______________

Susana Lucia Kugeares, Ph.D.

The University of Texas at Austin, 2002
Supervisors: Michael J. Telch and David M. Buss
A theoretical debate is currently ongoing about whether human emotional
states were molded by evolutionary pressures to serve particular functions. Social
anxiety has entered this debate. Non-evolutionary and evolutionarily adaptive
functional accounts of social anxiety have suggested that it serves as a ready-made
excuse for poor social performance, as a signal of submissiveness to prevent
hostility from dominant social partners, or as a cognitive interruption of a serious
social blunder that would lead to group expulsion. The evolutionary accounts are
domain-general, in that they attempt to explain the adaptive value of social anxiety
with a single function for all social contexts. The aim of the present study was to
test a domain-specific account of social anxiety, limiting investigation to the
mating domain. This study proposed that mating anxiety helps solve the adaptive
problem of the costliness of being rejected by a potential mate. To accomplish
xi

this, the mating anxiety mechanism was hypothesized to estimate the likelihood of
rejection by a potential mate by calculating the discrepancy between their
respective levels of desirability (Mate Value Discrepancy). Gender differences in
mating anxiety were also expected based on documented sex differences in
preference for physical attractiveness (PA) and social status/financial resources
(S/R) in a mate. A fictitious dating service was created in which participants
viewed mate profiles of four types: high PA-high S/R, high PA-low S/R, low PAhigh S/R, and low PA-low S/R. A 2 (Profile PA) x 2 (Profile S/R) x 3 (Replicates
for each mate type) x 2 (Participant Gender) design, with 2 covariates (self-rated
PA and self-rated S/R) was used to test the effects of profile and participant
characteristics on Anxiety, Likelihood of Rejection (LR), and Likelihood of
Asking the individual for a date (LA). Hierarchical Linear Modeling was used to
test the predictions about mate value discrepancy (MVD) and LR on Anxiety.
MVD had a highly significant effect on Anxiety and on LR. LR had a significant
effect on Anxiety as predicted, but did not mediate the effect of MVD on Anxiety.
A gender differences in anxiety were found in the effect of profile status/resources
on anxiety, but not for other profile or participant characteristics as expected. The
implications of an unexpected sex difference (men having greater anxiety than
women) are discussed. The results suggest strong support a domain-specific
approach and the position that human emotions serve adaptive functions.

xii

Table of Contents
List of Tables................................................................. xvii
List of Figures.................................................................... xviii
List of Appendices..................................................................... xix
Chapter 1: Introduction
1.1 Examining Social Anxiety Through an Evolutionary Lens..... 1
1.2 The Adaptationist Perspective of the Human Mind..... 3
1.3 An Evolutionary View of Emotional Disorders....... 6
1.3.1 Proper Classification of Dysfunction.... 6
1.3.2 Misconstrual as Dysfunction......................... 7
1.4 The Function of Social Anxiety....... 8
1.4.1 A Proximal Theory: Self-Handicapping Strategy...... 9
1.4.2 An Evolutionary Account: Submissiveness Signal Hypothesis. 11
1.4.3 An Evolutionary Account: Social Exclusion Model.. 15
1.4.4 A Review of the Proposed Functions of Social Anxiety....... 20
1.5 A Review of Relevant Dating Anxiety Research......... 21
1.5.1 Physical Attractiveness and Social Anxiety......23
1.5.2 Social Skill Deficits and Social Anxiety.................. . 24
1.5.3 Cognitive Interference and Social Anxiety .. 26
1.5.4 Dysfunctional Self-Evaluations and Social Anxiety.... 27
1.5.5 The Self-Presentational Model..............28
1.5.6 A Review of the Features of Dating Anxiety 31
Chapter 2: Present Study
2.1 Mate Value Hypothesis: A Context-Specific Theory of
xiii

Social Anxiety. 34
2.1.1 Mate Value Discrepancy and Likelihood of Rejection
(Hypothesis 1).... 36
2.1.2 Gender Differences in Mating Anxiety (Hypothesis 2).... 37
2.2 Specific Contributions..................................................38
Chapter 3: Method
3.1 Overview of Experimental Design....................................... 39
3.2 Participants....................................................................... 39
3.2.1 Exclusionary Criteria.... 40
3.2.2 Participant Characteristics. 40
3.3 Stimuli: Date Profiles.. 44
3.3.1 Physical Attractiveness in Profiles 44
3.3.2 Status/Resources in Profiles.. 45
3.4 Measures...................................................... 47
3.4.1 Participant Measures: Mating-Relevant Variables...47
3.4.2 Participant Measures: Anxiety and Personality Variables... 48
3.4.3 Profile Measures... 51
3.5 Procedure......................................................................... 53
3.6 Steps to Enhance Manipulation Integrity. 54
3.7 Experimental Conditions......................................................55
Chapter 4: Statistical Analyses
4.1 Sample Size and Power........................................................... 56
4.2 Analytic Strategy................................................. 56
4.2.1 Hierarchical Linear Modeling.. 57
4.3 Specific Analyses................................................. 58
xiv

4.3.1 Manipulation Checks.... 58

4.3.2 Replicate Effects....... 59
4.3.3 Effect of Profile and Participant Characteristics....... 59
4.3.4 Effects of MVD and Likelihood of Rejection on Anxiety.... 59
4.3.5 Mediation Effect....... 60
Chapter 5: Results
5.1 Manipulation Checks... 62
5.2 Effect of Replicate... 64
5.3 Effect of Profile Characteristics... 66
5.3.1 Effect of Profile PA and S/R on Anxiety...... 66
5.3.2 Effect of Profile PA and S/R on Likelihood of Rejection.....70
5.3.3 Effect of Profile PA and S/R on Likelihood of Approach.... 73
5.4 Effect of Participant Characteristics.... 75
5.4.1 Effect of PA and S/R on Anxiety...... 75
5.4.2 Effect of Participant PA and S/R on Likelihood of Rejection.. 76
5.4.3 Effect of Participant PA and S/R on Likelihood of Approach.. 77
5.5 Mate Value Discrepancy and Likelihood of Rejection Hypotheses.... 77
5.5.1 MVD and Anxiety.... 77
5.5.2 MVD and Likelihood of Rejection... 84
5.5.3 Likelihood of Rejection and Anxiety... 86
5.5.4 Mediation Hypothesis of Likelihood of Rejection 87
Chapter 6: Discussion
6.1 Credibility of the Deception......... 89
6.2 Integrity of the Experimental Manipulations... 90
6.3 Summary of the Main Findings ...... 93
xv

6.4 Factors That Influence Anxiety During Dating Initiation.... 94

6.4.1 The Influence of Prospective Mate Characteristics.. 94
6.4.2 The Influence of Gender of Individuals Evaluating
Potential Mates. 96
6.4.3 The Influence of Attractiveness and Status/Resources of
Individuals Evaluating Potential Mates.... 100
6.4.4 The Importance of Mate Value Discrepancy 102
6.4.5 The Role of Likelihood of Rejection.... 105
6.5 Implications of Exploratory Findings...... 106
6.6 Implications for Competing Accounts of Social Anxiety.... 108
6.7 Limitations of the Present Study...... 110
6.8 Directions for Future Research.... 113
6.9 Clinical Implications.... 116
Appendices.............................................................................. 119
References... 127
Vita.................................................................................. 137

xvi

List of Tables

Table 1. Demographics and Dating Status of Study Sample..............................41

Table 2. Means, Standard Deviations, and t-test Comparisons for Gender on
Dating history, Desirability, and Anxiety Variables........................ 43
Table 3. Mean Attractiveness and Sexiness Ratings from Pilot Study for the
Faces Selected for the Profile Stimuli.. 46
Table 4. Means and Standard Errors on Manipulation Check Variables for
Men and Women.......................... 63
Table 5. Means of Anxiety, Likelihood of Rejection, and Likelihood of
Approach scores by Profile Type and Participant Gender....................... 67
Table 6. Effect of Level 2 Participant Variables on MVD-Anxiety
Relationship Above and Below the Breakpoint............................... 82
Table 7. Effect of Male and Female Level 2 Variables on MVD-Anxiety
Relationship Above and Below Breakpoint......................... 83
Table 8. Moderating Effects of Level 2 Variables on the Relationship of
MVD on Likelihood of Rejection.............................85
Table 9. Moderating Effects of Level 2 Variables on the Likelihood of
Rejection-Anxiety Relationship............................... 87

xvii

List of Figures

Figure 1. Means for Men and Women on Perceived Attractiveness, Financial

Resources, and Social Status for Replicates within Mate Type.. 65
Figure 2. Significant 3-way Interaction Effect of Gender x PA x S/R on
Anxiety........................ 69
Figure 3. Significant Effect of Gender x PA x S/R on Likelihood of Rejection
scores... 72
Figure 4. Effect of Gender x PA x S/R on Likelihood of Approach.......... 74
Figure 5. Effect of Self-rated Physical Attractiveness and Status/Resources
on Anxiety for Men and Women.................... 75
Figure 6. Linear and Piecewise Relationships of Mate Value Discrepancy on
Anxiety............ 78
Figure 7. Piecewise Regressions of Mate Value Discrepancy on Anxiety with
Different Optimal Breakpoints for Women and Men..............79
Figure 8. Effects of Profile Desirability and Participant Desirability on
Anxiety........ 80
Figure 9. Effect of Profile Desirability on Anxiety as a Function of
Participant Self-rated Desirability................... 81
Figure 10. Effect of Mate Value Discrepancy on Likelihood of Rejection
Estimates.................. 85
Figure 11. Effect of Likelihood of Rejection on Level of Anxiety............ 86
Figure 12. P-values of Likelihood of Rejection as a Mediator of Mate Value
Discrepancy-Anxiety relationship............ 88

xviii

List of Appendices

Appendix A. Deceptive Introduction to the Testing Session.... 119

Appendix B. WHR Instruction Page (Female Version).... 120
Appendix C. Welcome to Love Letters Instruction Page.. 121
Appendix D. LoveLetters - Market Research Survey... 122
Appendix E. Example Stimulus Profile.... 123
Appendix F. Profile Assessment Survey... 124
Appendix G. Interest Letter... 125
Appendix H. Assessment of Deception Survey.... 126

xix

CHAPTER 1: INTRODUCTION

1.1 Examining Social Anxiety Through An Evolutionary Lens

The question of what constitutes a mental disorder is currently being hotly
debated (Lilienfeld & Marino, 1995; Cosmides & Tooby, 1999; Wakefield, 1992;
Wakefield, 1999). On one side of this debate, Darwinian psychiatrists and
evolutionary psychologists contend that emotions are the purposeful output of
evolved psychological mechanisms, designed by natural selection to solve certain
specific and recurrent adaptive challenges faced by ancestral humans. Therefore,
understanding the evolved function of an emotion is considered necessary in order
to identify emotional dysfunction. For instance, proponents of this hypothesis
view panic attacks as the functional product of an evolved defense system that has
cognitive, physiological, motivational, and behavioral aspects designed to protect
against imminent threats to survival (Nesse, 1987; Nesse & Williams, 1994).
Knowing that the evolved function of panic attacks is to signal alarm in the
presence of immediate danger enables clinicians to distinguish true alarm panic
attacks from false alarm panic attacks, which occur in the absence of any real
environmental threat. This distinction is essential for identifying dysfunctions in
the panic mechanism and making judgments about the presence of panic disorder
(American Psychiatric Association, 1994).

Increasingly during this decade, an evolutionary lens has been used to

reexamine the nature of disorder, the function of emotions (e.g., Gilbert, 1989),
and psychopathology. Theories have arisen about the potential adaptive value of
depression (McGuire, & Troisi, 1998; Sloman & Price, 1987), psychopathy
(Mealey, 1995), eating disorders (Abed, 1998), and mood changes (Price, 1998)
to name a few. Social anxiety has also been reconceptualized as an adaptation
(e.g., Trower & Gilbert, 1989; Trower, Gilbert, & Sherling, 1990). Two main
theories have been proposed: the Submissiveness Signal Hypothesis (Ohman,
1987) and the Social Exclusion Model (Baumeister & Tice, 1990). Each suggests
that social anxiety is a defense against some threat to survival posed by social
interactions. However, both theories err in one important way: They attempt to
explain the adaptive value of social anxiety in all social contexts with a single
function. Different contexts, however, likely pose different challenges to survival
and reproduction, which, in turn, demand different solutions (Tooby & Cosmides,
1990). The current study will take a more context-specific approach to social
anxiety and will focus on anxiety aroused in mating contexts, from here on
referred to as mating anxiety.
First, the adaptationist perspective of psychological mechanisms, which is
the theoretical foundation of Darwinian psychiatry, will be briefly reviewed and
distinctions between true mechanism dysfunction and misconstrual as dysfunction

will be outlined. Next, two evolutionary theories of social anxiety, the

Submissiveness Signal Hypothesis (SSH) and the Social Exclusion Model (SEM),
will be presented in some detail and their limitations discussed. In addition,
relevant clinical literature on dating anxiety will be reviewed. Finally, a model of
a mating anxiety mechanism, of which at least one design feature evolved to solve
a mating-specific adaptive problem, will be proposed and tested. Gender
differences in mate preferences will be briefly reviewed as they pertain to certain
expected design features of the mating anxiety mechanism.

1.2 The Adaptationist Perspective of the Human Mind

As in biology, the adaptationist program in psychology is to understand
features of the human mind as elements of special problem-solving machinery
(Williams, 1985 cited in Symons, 1992, p. 140). Why take this approach to
understanding human psychology? The human mind, like the vertebrate eye,
shows evidence of complex design, the coordinated assembly of many novel,
functionally integrated features that accumulate slowly over hundreds of
thousands of generations (Tooby & Cosmides, 1992, p. 5). The chance that
random events created as complex a structure as the human mind is near zero. In
addition, the presence of complex design in any adaptation belies the pressures
that natural selection exerted during the mechanisms evolution. An adaptation is

a structure or mechanism designed by natural selection to serve a particular

function. That function is the solution to an adaptive problem (i.e., one whose
solution affected reproduction) that recurrently challenged the fitness of ancestral
humans. An adaptation may co-opt mechanisms or parts of mechanisms that
previously existed but also possesses at least one aspect (e.g., a particular input,
output, decision rule) that is unique to it and that dovetails with the corresponding
adaptive problems. The result is a close functional mesh between an adaptive
problem from ancestral, deep time and a mechanisms features that evolve to
solve it (Tooby & Cosmides, 1992). In terms of the human mind, evolved
psychological adaptations are information-processing mechanisms designed to
take in and process particular environmental information and produce particular
output. The features of such a mechanism were selected (by natural and sexual
selection pressures) because they helped to solve the particular adaptive problem
over the course of the mechanisms evolutionary history. An example of such a
mechanism is sexual jealousy. It occurs in predictable contexts, with predictable
triggers, and has predictable emotional, behavioral, and physiological outcomes
that, for men, solve the problem of paternity uncertainty (Buss, Larsen, Westen, &
Semmelroth, 1992).
Two important points must be reiterated with respect to studying
psychological adaptations. First, psychological mechanisms evolved to solve

adaptive problems that existed recurrently over long spans of prehistoric time.
This is assumed because of the long span of time necessary to accumulate, bit by
bit, each additional modification and feature that results in complexly designed
adaptations. As a result, a given psychological mechanism may not have adaptive
consequences in the modern environments, if the environmental context relevant
to the mechanism has significantly changed. This requires that adaptations be
distinguished from adaptiveness (Symons, 1992). Second, all observed structures
and behavioral phenomena need not be adaptations. A given structure may be a
byproduct of an adaptation, that is a non-purposeful consequence of an adaptation
much like heat is an incidental byproduct of a light bulbs original lightproducing design. However, a byproduct can only be understood in relation to the
functionally designed adaptation that produces it. It must be remembered that any
emotion, including mating anxiety, may be merely a byproduct of some other
adaptation.
Since features of psychological mechanisms evolved to solve particular
adaptive problems, and those problems were specific to context, a psychological
mechanisms function, if one exists, can only be understood as it relates to the
ancestral human context of the adaptive problem it was designed to solve and in
which it is currently activated. By identifying adaptations one carves the
phenotype at its natural, functional joints (Symons, 1992, p. 140).

1.3 An Evolutionary View of Emotional Disorders

Using an evolutionary perspective, emotional disorders can be classified
as 1) the output of truly defective mechanisms (i.e. true disorder), 2) the
maladaptive output of properly functioning mechanisms, or 3) the misconstrual
of a mechanisms proper output as dysfunction.

1.3.1 Proper Classification of Dysfunction

True disorder is defined by a failure of the mechanism to perform as it was
designed by natural selection to perform within those specific contexts in which it
evolved (Buss, Haselton, Shackelford, & Bleske, 1996). In theory, a
psychological mechanism can be faulty at any of several stages of information
processing. It could fail to become activated in the proper context; fail to respond
to proper cues (inputs); fail to produce its proper output; or fail to interact with
related mechanisms as it was designed to (Buss, et al., 1996). Additional
information processing steps that can go awry and lead to mechanism dysfunction
are outlined by Buss and colleagues (1996).
The latter two classes of emotional disorders, not true mechanism
dysfunction, may still lead to problematic consequences and even warrant
treatment (Cosmides & Tooby, 1999). They are discussed below.

1.3.2 Misconstrual as Dysfunction

Some emotional disorders may be the product of a properly functioning
mechanism which nevertheless produces maladaptive output. The output may be
maladaptive due to a mismatch between modern environments and the past
evolutionary environment which posed the adaptive problems for which the
mechanism is geared to solve. For example, evolutionarily adaptive taste
preferences for fat and sugar from the past result in high rates of heart disease and
obesity in todays agricultural societies. Maladaptive outputs can also result from
a mechanisms built-in likelihood of making false positives errors, an adaptive
design feature in contexts where the cost of a false negative error is higher than a
false positive error. Anxiety disorders (i.e., clinically treatable levels of anxiety)
may be caused by this kind of built-in over-responsiveness (Nesse, 1990). In
addition, since many psychological mechanisms are designed to counter the
strategic interference from others, the correct and usually adaptive output of a
mechanism might still be outdone by the strategies of another individual and
ultimately fail to solve the adaptive problem (Buss, et al., 1996).
Other disorders may be the adaptive output of properly functioning
mechanisms, yet be classified as disorders because they cause pain or discomfort.
Evidence has recently been found of the adaptive value of several physiological
mechanisms that were traditionally considered dysfunctions. Fever (Kluger,

1997), pregnancy sickness (Profet, 1987), and allergies (Profet, 1991) are among
the physiological defenses for which adaptive benefits are suggested. If anxiety
similarly serves an adaptive defensive function, there may be detrimental
consequences to blocking it with medications (Nesse, 1990).
Last, socially nonconforming behavior can sometimes be labeled as
dysfunction when it might be the functional and proper output of an evolved
psychological mechanism. Psychopathy is one example. Despite its harmful
effects on social groups and even, in some ways, on the individual himself (e.g.,
incarceration), recent evolutionary theories view psychopathy as a frequencydependent adaptation a coordinated system of emotions, behaviors, motivations,
cognitions, and physiological reactions that promote a short-term sexual strategy
and a cheating social lifestyle (Harpending & Sobus, 1987; Mealey, 1995). We
will now turn our attention to social anxiety in particular.

1.4 The Function of Social Anxiety

Since before the birth of Darwinian medicine or evolutionary psychology,
anxiety has been thought of as a functional product of natural selection, with
evolutionarily-based features (e.g., Seligman, 1971; McNally, 1987). With
respect to social anxiety, three theories of its function have been proposed: the
Self-Handicapping Strategy (Snyder, Smith, Augelli, & Ingram, 1985), the

Submissiveness Signal Hypothesis (Ohman, 1987), and the Social Exclusion

Model (Baumeister & Tice, 1990), the latter two being evolutionary theories. The
first also warrants review since it stands out as the only functional account of
social anxiety that is not based in evolutionary theory.
Three general functions have been proposed for social anxiety (Nesse,
1990): communication, motivation, and cognition. The Self-Handicapping
Strategy (SHS) and the Submissiveness Signal Hypothesis (SSH) propose
communicative functions of social anxiety. The Social Exclusion Model (SEM)
proposes a primarily cognitive function of social anxiety, with motivational
aspects.

1.4.1 A Proximal Theory: The Self-Handicapping Strategy

The Self-Handicapping Strategy (SHS) is a theory about the function of
social anxiety in shy or socially anxious individuals. Snyder and colleagues
(1985) proposed that in these individuals, social anxiety and shyness serve to
control attributions made by oneself or by others about social performance. They
suggest that a self-handicapping presentation channels attributions of poor social
performance away from the individuals general worth or skill-level, and onto
internal qualities that are less relevant to self-esteem (i.e., to shyness or
situational anxiety). According to the SHS, socially anxious individuals doubt

their ability to create the positive social impressions they desire, and, therefore,
seek alternate strategies of self-presentation that, at least, minimize the negative
implications on themselves of failure (Snyder, et al., 1985). Snyder and
colleagues proposed that when an individual expects a high likelihood of failure
in a social situation yet avoidance is not possible, an anxious self-presentation is
employed to deflect failure from being attributed to a true lack of ability or to
other shortcomings. By impeding successful social performance, the social
anxiety serves as a ready excuse in the event that a social failure occurs
(Snyder, et al., 1985, p. 971). In addition, if the opposite occurs (that is social
success), the implications of evaluation by others of the individuals abilities are
even more positive than without the impediment at all. The authors hypothesize
that only individuals who feel a threat to an important dimension of their selfesteem (either intelligence, ability, or attractiveness) will use a self-handicapping
strategy in this way.
Indeed, individuals reported more social anxiety only in situations where
anxiety could be used to explain poor performance but not in situations where it
was precluded as an explanation (Snyder, et al., 1985). Other research found that
when anxiety was precluded as an explanation for failure, subjects reported other
explanations for poor performance such as lack of effort (Smith, Snyder, &
Handelsman, 1982). They interpreted this result as further evidence of the

10

tendency of shy individuals to deflect negative evaluation away from enduring

internal qualities to other less important characteristics or to situational variables.
This pattern, however, was found exclusively in chronically shy men. To
explain this, the authors suggested that when women are faced with negative
social evaluation, they are more likely to use a strategy referred to as passive
accommodation, a self-presentational stance of silence, passivity, and
agreeableness, rather than using the strategies of social withdrawal and avoidance
as men tend to (Pilkonis, 1977; Snyder & Smith, 1982). However, they do not
offer a specific explanation for why men but not women use social anxiety as a
self-invoked handicap.

1.4.2 An Evolutionary Account: The Submissiveness Signal Hypothesis

Another theory of the communicative function of social anxiety is the
Submissiveness Signal Hypothesis (Ohman, 1986; Ohman, 1987). It proposes
that social anxiety is a system of physiological and emotional responses whose
function is to display submissiveness to a dominant other in a social group. The
SSH suggests that, in humans, social anxiety is aroused in situations that involve
social dominance or submissiveness in some way. These could be instances of
jockeying for social dominance or merely interacting with a more dominant and
potentially hostile other. The theory is based on the premise that, in many

11

animals, natural selection pressures are stronger for appearing able to win a
physical confrontation than for actual success in confrontations. The theory does
not state explicitly but implies that the appearance is associated with the actual
ability to win a physical confrontation. Hence, selection pressures led to the
evolution of signals of dominance and submissiveness because they were reliable
indicators of the likelihood of success in a physical confrontation (Ohman, 1987).
Ohman offered several pieces of evidence to support his claim that
socially aroused fear is distinct from fear evoked by other stimuli. He reported
that facial and behavioral expressions of fear differ when triggered in different
contexts. For instance, anxiety induced by the presence of a predator is more
automatic and is geared more for physical escape than socially related anxiety.
He also noted similarities between the behaviors associated with social anxiety in
humans and indicators of submissiveness in other animals, such as crouching,
freezing, and making submissive facial expressions. In non-human primates,
dominance conflicts frequently arise during mating and feeding interactions and
when meeting strangers. He also suggested that, in non-human primates,
dominance hierarchies are equally important for both males and females in
determining which individuals have access to limited resources (such as food and
mates). If these situations present opportunities for dominance conflicts in
humans as well, and if social anxiety serves the function proposed by the SSH,

12

then humans should feel social anxiety in the same contexts that involve
competition for social dominance as other animals: mating, feeding, and meeting
strangers.
Unfortunately, Ohmans (1987) SSH stops short of proposing specific
benefits that a display of submissiveness confers on the signal sender, nor does it
posit what danger(s) from the signal receiver submissiveness displays stave off.
The SSH might suggest that one possible benefit to the sender of a
submissiveness signal might be eliciting sympathetic reactions rather than
hostility from dominant others when the individual might not otherwise possess
the requisite qualities or skills to be dominant himself (or herself). Ohmans
theory implies that certain kinds of individuals (or any individual in a particular
type of social encounter) become motivated to exhibit a signal of submissiveness.
But, it fails to specify what could be thought of as design features of the
mechanism that governs submissiveness displays. For example, are the primary
triggers for displaying social anxiety located within the individual (e.g., small
physical size)? within the interactional partner (e.g., higher social status or
aggressive demeanor)? or in some other feature of the surrounding environment
(e.g., scarcity of a particular resource, thereby intensifying competition)?
Ohman predicts that social anxiety in humans should begin to manifest
itself during the developmental periods when individuals start to become sensitive

13

to dominance hierarchies, namely, puberty and early adulthood. Therefore,

individuals who experience social anxiety before puberty should show evidence
of sensitization to dominance hierarchies. Ohman predicts that adolescents will
experience greater social anxiety than adults because they are not yet mature
enough to assume a dominant position over adults, and therefore have more
reason to display submissiveness. Hence, adolescents should feel more anxious
during interactions with adults than with other adolescents. Given the ubiquity of
interactions where dominance-submissiveness conflicts can arise, almost all
people should feel some amount of social anxiety. The exception to this would be
those individuals who are extremely high in social status.
Other predictions can be inferred from the SSH. The SSH should predict
that social anxiety would be highest when the social dominance of the interaction
partner is much higher than the individual. Because facial expressions and
fixations of gaze are central features of dominance displays, one would expect
anxiety to increase in response to dominant facial expressions and eye contact (of
course, these would have to be defined). Individuals might also be predicted to be
more perceptive to facial expressions and eye contact during those types of
interactions that involve dominance and submissiveness than during other social
interactions (which also need to be specified). Based on Ohmans observation
that dominance hierarchies are equally important for male and female non-human

14

primates, one would infer that the SSH holds dominance hierarchies to be equally
important to human males and females. It follows that the SSH would predict no
gender differences in social anxiety that could not be accounted for by the
proposed social dominance effect.

1.4.3 An Evolutionary Account: The Social Exclusion Model

Exclusion theory attempts to explain anxiety in a number of contexts:
specific situational anxiety, such as anxiety about losing ones job, chronic
existential anxiety, anxiety about death, agoraphobia, and anxiety over sexual
urges and sexual dysfunctions (Baumeister & Tice, 1990). It also provides
alternative accounts for a broad range of previous theories, such as the
development of the self and culture (as defenses against anxiety), attachment
theory and separation anxiety, self esteem, guilt, and the Oedipal complex. The
Social Exclusion Model proposes that the root of all these anxieties is fear of
group exclusion, whether that group is a family, a social unit, a dyadic
relationship, or culture on the whole. Therefore, all these anxieties are considered
forms of social anxiety.
The model argues the following. Human have evolved a strong preference
for living in social groups because of the survival benefits groups confer, such as
better access to resources, safety from predators, and cooperation at difficult

15

tasks. As a result, any event that separates an individual from an important

social group can put an individual at risk of expulsion from that group, remind the
individual of past social exclusions, and may even represent the threat of chronic
aloneness (e.g., darkness is proposed as a representation of this threat). These
events which threaten the loss of group benefits will arouse affective distress
(Baumeister & Tice, 1990, p. 167). Next, Baumeister and Tice name three kinds
of behavior or traits that increase the risk of being ostracized or rejected by a
group: 1) failing to contribute adequately to a groups welfare or survival, 2)
violating group norms about appropriate interpersonal behavior, and 3) being
unattractive or possessing dislikable personality traits.
The proposed function of anxiety in the Social Exclusion Model seems to
revolve around one particular threat to group membership -- making a serious
social blunder. The SEM proposes that when an individual is about to commit a
social blunder, which would lead to negative evaluation by others, social anxiety
becomes activated to interrupt the current course of action (first function), shift
attention to the source of the threat to group inclusion, which is negative
evaluation by others (second function), and prompt cognitive reassessment of the
course of action before proceeding (third function).
Based on the three primary reasons listed for group exclusion, the SEM
would suggest that being judged by other group members as unattractive,

16

abnormal, or incompetent should arouse the most social anxiety. It is not made
clear how being unattractive risks group expulsion. The SEM only states that
unattractiveness makes a person less desirable as an interactional partner
(Baumeister & Tice, 1990, p.169). Furthermore, the importance of attractiveness
does not follow from the SEMs premise that group living evolved in response to
the benefits to survival, since attractiveness is clearly more related to reproduction
than to survival.
How does incompetence lead to group exclusion? The SEM leaves this
vague, but it seems plausible that the model might view it as related to the first
threat to social exclusion, failures to contribute adequately to a groups
functioning.
How likely is social deviance to result in social exclusion? Evidence from
chimpanzee behavior suggests that deviant behavior can provoke a combination
of fear and hostility from other group members and does, indeed, lead to
exclusion from social interactions (Goodall, 1986). Observations of chimpanzees
suggests that excluding individuals who display abnormal behavior may be
adaptive for survival since abnormal behavior (e.g., limping or paralysis) is often
the result of disease. Therefore, excluding abnormal individuals from social
interactions could protect against contagion. It is possible that similar decision

17

rules apply in humans. It is also possible that rejection of abnormal or socially

deviant humans is more closely related to social costs than threats of contagion.
Baumeister and Tices proposal is consistent with some of the current
clinical knowledge about social anxiety. The attention of socially anxious
individuals often does turn to problems more readily than that of non-anxious
individuals. Socially anxious people have also been observed to behave in ways
that will avoid arousing the anger or ill will of the other person. For instance,
they avoid controversial topics and topics about which they feel they know little;
they smile and nod more (especially women); they behave in ways that are
agreeable, non-conflictual and non-threatening to their interaction partner
(Baumeister & Tice, 1990). In general, their interaction style is passive.
However, predictions of the SEM contradict other social anxiety research
findings. The proposed shift in attention onto the interaction partner is not found
in socially anxious individuals. Although attentional focus does change when
social anxiety is aroused, it usually becomes focused inward on the self, rather
than on the interaction partner. Research has consistently shown that socially
anxious individuals remember less about an interaction partner than do nonanxious counterparts (Hope & Heimberg, 1990).
An oversight of the SEM is its emphasis of survival benefits of group
living, disregarding the valuable reproductive benefits that groups provide. Buss

18

(1990) articulates three such reproductively beneficial characteristics of group

living: a greater number of potential mates, a greater number of potential allies
with whom to exchange reproductively valuable resources, and cohabitation with
kin with whom acts of reciprocal altruism (which enhance inclusive fitness) can
be performed. For example, an ostracized individual will not only lose group help
to obtain food, but loses access to potential mates, an arguably more costly loss
(Buss, 1990). Selection pressures for group living were probably influenced by
its benefits to reproduction as much as the benefits to survival.
The SEM also fails to consider the potential effects on social anxiety of
differences in the desirability of different social groups. Individuals desire
membership in certain groups (e.g., those that are high in status and power or that
contain attractive members) more than other groups (Buss, 1990). When social
anxiety is activated, its trigger may have been a cue that a member of a desirable
group is evaluating the individual as inadequate in some important quality. As
Buss (1990) points out, the threat may not be group expulsion, per se, but rather
exclusion from reproductively valuable resources.
Another weakness of the SEM is its lack of explanation about why
competence, attractiveness, and normality would be central to fears of group
rejection over other qualities. With respect to avoidance as a response closely
related to social anxiety, the SEM is mostly silent. It does acknowledge that shy

19

individuals avoid those situations and behaviors that they feel anxious about,
those in which they anticipate being rejected or evaluated negatively. But, its
authors make no attempt to explain any possible adaptive function of avoidance.
They state that avoidance keeps the individual safe from acute anxiety, but
acknowledge the costs of possible chronic loneliness and lost opportunities to
build social bonds.

1.4.4 A Review of the Proposed Functions of Social Anxiety

Three theories of the function of social anxiety were presented. First,
anxiety may provide men with a ready excuse for attributions about social
failures, to protect against threats to self-esteem (Snyder, et al., 1985). Second,
anxiety may signal submissiveness to more dominant individuals, perhaps to
avoid hostility and to negotiate positions of social dominance in relevant contexts
(Ohman, 1987). Third, anxiety may serve a cognitive interrupt function, to
prevent the commission of social blunders that could lead to expulsion from an
important social group. It may simultaneously function to shift attention to signs
of negative evaluation by interaction partners, again to protect against social
exclusion (Baumeister & Tice, 1990). As previously mentioned, a major
theoretical shortcoming of both evolutionary theories of social anxiety is their

20

attempt to capture with a single mechanism the function of social anxiety in all
social contexts.

1.5 A Review of the Relevant Dating Anxiety Research

Like other psychological adaptations such as sexual jealousy (Buss, et al.,
1992) or panic (Nesse, 1987), an evolved social anxiety mechanism should show
evidence of the coordination of cognitive, motivational, behavioral, attentional,
and physiological outputs. The body of research on dating anxiety provides
information about some of these aspects of social anxiety, but without suggesting
a potential function. Most models focus on explaining the immediate antecedents
of the experience of social evaluative anxiety and on describing its micro-level
triggers and concomitant behaviors. The explanations for social anxiety presented
in the Social Skill Deficits, Dysfunctional Self-Evaluations, and Cognitive
Interference models will be reviewed, as well as physiological correlates and the
possible role played by physical attractiveness. In these models, social anxiety is
implicitly (and in one case explicitly) viewed as an error in normal healthy
functioning. Then, the Self-Presentational Model, which integrates cognitive,
motivational, and behavioral aspects of social anxiety, will be outlined. Though
implied throughout the model, the SPM stops just short of hypothesizing a
possible function of social anxiety.

21

Dating is defined as interactions with potential romantic partners prior to

the development of a full-fledged relationship (Hope & Heimberg, 1990, p. 220).
Within the clinical literature on dating anxiety, a variety of terms -- heterosocial
anxiety, low frequency dating, communication apprehension -- have been used to
capture the construct and to identify dating anxious individuals, those who are
characterized by both high anxiety during interactions with members of the
opposite sex and by low success with creating a desirable image during these
interactions (e.g., Jaremko, Myers, Daner, Moore, & Allin, 1982). In addition to
the research focused specifically on dating anxiety, a significant portion of
general social anxiety research has operationally defined social anxiety as that
aroused in opposite sex interactions. This suggests that researchers who, on one
hand, consider social anxiety to be a response to a broad range of social contexts,
may, on the other hand, have intuited something especially potent, or perhaps
universal, about dating interactions for arousing anxiety. Therefore, a brief
review of the dating anxiety literature has relevance to the understanding of a
mating anxiety mechanism. For simplicity, this review will use the term dating
anxiety, except when specifying the subject selection criteria used for a particular
study (e.g., selection on the basis of date frequency).

22

1.5.1 Physical Attractiveness and Social Anxiety

A connection exists between physical attractiveness and social anxiety.
Individuals who are anxious about dating situations are rated as less attractive
than non-anxious individuals (Glasgow & Arkowitz, 1975). Physical
attractiveness is consistently correlated with a high frequency of dating and it
strongly predicts good social skills (e.g., Mathes, 1975; Mitchell & Orr, 1976;
Zakahi, Duran, & Adkins, 1994). In turn, individuals who take social initiatives,
such as asking questions, and in doing so demonstrate social confidence and calm
are evaluated as more attractive than are individuals who appear anxious. Others
have found that physically attractive individuals are higher in intelligence, social
status, and other positive attributes in addition to being superior in social skill
(Udry & Eckland, 1984).
What can explain these observations and the interplay between
attractiveness, social anxiety, and social skill? Dating anxiety may have its origin
in actual rejection experiences that were brought on by lack of physical
attractiveness. In addition to being advantageous in dating situations, physical
attractiveness positively affects other variables relevant to reproductive success,
such as occupational status (Garcia, Stinson, Ickes, Bissonnette, & Briggs, 1991).
Furthermore, attractive individuals may have more opportunities for social
interaction over the course of their development and, thus, develop better social

23

skills. Alternatively, good social skills may be attributed to attractive individuals

whether they possess those skills or not (Hope & Heimberg, 1990). It may also
be that good social skills may have a strong enough benefit on reproductive
success that it makes one appear more attractive. Finally, the presence of anxiety
itself may detract from ones appearance as attractive. One must keep in mind
that qualities that increase ones reproductive success are the qualities that are
likely to be considered desirable.

1.5.2 Social Skill Deficits and Social Anxiety

As mentioned above, anxiety is related to social skill. The Social Skills
Deficit theory explains social anxiety as the consequence of inadequate skills
necessary for smooth social interactions (Hope & Heimberg, 1990). These skills
include amount and directness of eye contact, head nodding, smiling, the timing
and length of pauses during conversation, and the conversational topics chosen.
The model suggests that, when social interaction skills are poor, high levels of
social anxiety are experienced (e.g., Larsen & Shackelford, 1996). Consistent
pairing of inadequate social skills with high anxiety creates a conditioned
association that can lead to anticipatory anxiety and avoidance of future social
interactions.

24

Social skill deficits have been consistently found among dating anxious
individuals. Relative to men with high frequencies of dates, low-dating men more
frequently leave pauses of longer than 10 seconds in their conversations, talk for
less total time, and introduce more negative topics into conversations. Their
female conversational partners are also less likely to respond positively to their
verbalizations than to those of men with low anxiety. Women with low dating
frequency are generally similar to low dating frequency men on this micro-level
of social skills. They make less eye contact, talk for less total time, and leave
more pauses of four seconds or longer than high frequency dating women do.
Another social skill relevant to dating is how to ask for a date, specifically the
directness of ones approach. Women who ask for a date directly are judged less
positively by others than women who ask indirectly (Muehlenhard, Koralewski,
Andrews, & Burdick, 1986). This suggests that demonstration of subtle social
skills can influence how attractive one appears to others.
Deficits in men and women on such subtle social behaviors as these can
lead to dating failures and to an increased likelihood of developing anxiety about
dating interactions. With respect to understanding what causes social anxiety,
these studies suggest that a lack of certain qualities or skills that would make an
individual desirable might activate social anxiety.

25

1.5.3 Cognitive Interference and Social Anxiety

The cognitive interference model of social anxiety is related to the social
skill deficit explanation. It attributes the social skills deficits observed in dating
anxious individuals to cognitive interference effects caused by anxiety. That is, it
claims that anxiety causes cognitive interference of the appropriate demonstration
of good social skills, which in turn leads to the operant conditioning of more
anxiety and avoidance. The Cognitive Interference model assumes that socially
anxious individuals do possess adequate social skills but are limited in their
ability to use them because anxiety interferes with attention and accurate
appraisals of the situation.
There is evidence that anxiety interferes with good social performance by
shifting attention onto the anxious individual himself and away from the
interaction partner. Normally, attending to an interaction partner permits the
discovery of important social information. For instance, most men pay more
attention to women who share their interests and attitudes, and prefer them as
interaction partners. Dating anxious men, however, do not show this preference
(Clark & Arkowitz, 1975), suggesting a lack of attention to important social
information from the partner. More support comes from the finding that high
anxious subjects focus attention on their own appearance during interactions,
rather than focusing on their partner (OBanion & Arkowitz, 1977).

26

Other research suggests that social anxiety does cause an attentional shift
onto the interaction partner but in a distorted and skewed way. Hope and
Heimberg (1990) reported that high anxious individuals have better recall for
negative evaluative comments made by interaction partners than positive ones.
High anxious individuals also anticipate more negative evaluations by others and
interpret neutral feedback more negatively (Hope & Heimberg, 1990). These
findings suggest the attention of a socially anxious individual becomes restricted
to the interaction partners evaluation of them. Whether social anxiety leads to
focusing attention onto the self or onto the partner in a skewed manner, valuable
information about the partners characteristics on which social decisions can be
base is missed.

1.5.4 Dysfunctional Self-Evaluations and Social Anxiety

This line of research considers social anxiety to be the consequence of
dysfunctional self-evaluations (Schlenker & Leary, 1985). These selfevaluations include negative self-appraisals and expectations of social
performance failures. Often underlying the latter are unattainably high standards
for social performance that are thought to originate from an excessive motivation
to impress others and gain their approval (Hope & Heimberg, 1990). Selfevaluations made by dating anxious individuals are skewed and negative, and

27

influenced by their level of social anxiety. There is evidence that an individuals

social anxiety is more closely related to his evaluation of his social performance
than to his actual social performance (Meichenbaum, 1977).
Dating anxious individuals make other self-appraisal errors as well. They
underestimate their level of social skill (Twentyman & McFall, 1975),
underestimate their social performance relative to ratings made by objective
judges (Clark & Arkowitz, 1975), and underestimate social performance
evaluations made by others (Smith & Sarason, 1975). Interestingly, low-anxious
men actually overestimate their social performance (Clark & Arkowitz, 1975).
Summarized in terms of evolved psychological mechanisms, the Dysfunctional
Self-Evaluations Model suggests that social anxiety is produced by faulty
information-processing algorithms in the mechanism, whether those algorithms
guide self-evaluation of social performance or social skill level, or generate
inferences about the evaluations made by others.

1.5.5 The Self-Presentational Model

The Self-Presentational Model (SPM; Schlenker & Leary, 1982; 1985)
integrates negative self-evaluations, physical attractiveness, and social skill
inadequacies into a coherent model of antecedents and triggers of social anxiety.
The SPM identifies two necessary conditions for the social anxiety mechanism to

28

be activated: a wish to create an impression of desirability to another person

combined with doubts about ones ability to successfully create that desired
impression. First, an individual forms an identification of himself, that is, an
image of himself that he wants to portray to the other individual. Next, this image
of himself is combined with his assessment of his ability to create the impression
he desires, or self-presentational outcome expectancies (Schlenker & Leary,
1982). This key component is akin to low self-efficacy in the social arena a low
expectation of ones ability to succeed in creating a positive impression of
oneself. Although no explicit attempt is made within the model to explain of the
function of social anxiety, one of its authors has linked the Self-Presentational
Model with Baumeister and Tices (1990) Social Exclusion theory, suggesting
that the underlying motivation to make a favorable social impression may stem
from an evolved motivation to ensure maximal group inclusion (Leary, 1990).
Schlenker and Learys (1982) SPM outlines which social contexts and
variables they believe arouse social anxiety and which do not but arouse other
emotions instead. The first two contexts do not arouse anxiety. Context 1: An
individual has low motivation to impress an interaction partner (or a potential
interaction partner) because that partner is unattractive or unimportant leads to
indifference. Unfortunately, no criteria are suggested for what makes a partner
important and or unimportant (a significant decision rule, from an

29

information-processing perspective). Context 2: An individual has little desire to

impress because he is certain that the other person holds him in high regard,
which leads to a feeling of complacency. Schlenker and Leary (1982) propose
that, in general, a lack of motivation to impress precludes social anxiety.
The last three contexts involve a high motivation to impress. Context 3:
A strong desire to create a positive impression combined with high selfpresentational outcome expectancies (that is, anticipation of a successful selfpresentation) results in a feeling of confidence. Context 4: When the same strong
desire to impress is paired with only moderate expectations of successful selfpresentation, an individual is left feeling challenged. Context 5: Finally,
feeling this strong desire in the context of low self-presentational expectancies
leads to anxiety.
The degree to which one wishes to create a positive impression is the first
half of Schlenker and Learys proposal (1982). Therefore, the SPM outlines
which contexts should arouse a desire to create a favorable impression. First, the
authors (1982) state that the desire to impress occurs when the other individual is
attractive, powerful or expert. Power presumably refers to social power and
status. Second, important performances, such as first dates or tests of some
kind, trigger the motivation to impress others. As a sidenote, important
performances may be those that involved attractive, powerful, or expert others,

30

but this is not implied by the SPM. Third, novel situations where others
expectations are ambiguous or where general rules of behavior are unclear should
arouse the motivation to impress. All of these situations that trigger a desire to
impress others should also elicit social anxiety.
Schlenker and Learys (1982) second requisite component for arousing an
anxious response is doubt about ones ability to create the desired impression.
The SPM identifies low self-esteem and poor social and communication skills as
the root of low performance outcome expectancies.

1.5.6 A Review of the Features of Dating Anxiety

To summarize, dating anxiety has cognitive, behavioral, attentional,
physical, and motivational correlates. Social anxiety may be the result of low
physical attractiveness. Dating anxious individuals are judged as less attractive
than others are and attractiveness is correlated with a high frequency of dating and
with other variables relevant to reproductive success (e.g., job status).
Social anxiety may be the result of inadequate social skills. Deficits in
social microskills such as amount and directness of eye contact, length of
conversational pauses, and conversational topics chosen have been found among
dating anxious individuals. In addition, better social skills leads to higher

31

judgments of attractiveness (Udry & Eckland, 1984) which suggests that poorer
social skills may detract from attractiveness and, in turn, increase social anxiety.
Rather than being the result of social skill deficits, social anxiety may
cause apparent skills deficits, or at least poor social performance, through
cognitive interference. This interference occurs through a shift in attention onto
the self or onto cues of negative feedback from the interaction partner, resulting in
missing important positive information about the interaction and the partner.
Anxiety also interferes with the accurate appraisal of social interactions,
such as interpreting neutral feedback more negatively, remembering more
negative than positive feedback, and anticipating more negative evaluation from
others. These attentional shifts and distorted appraisals can lead to poor social
performance, lower evaluations by others, and more anxiety. Anxiety-producing
dysfunctional appraisals also include expecting social performance failures,
having unattainably high standards for social performance, and underestimating
ones social skill, social performance, and others positive evaluations.
Schlenker and Learys (1985) Self-Presentational Model conceptualizes
social anxiety as the product of high motivation to appear desirable mixed with
doubt about ones ability to successfully create that impression. The motivation
to impress may be driven by an evolved desire to be maximally included in social
groups (Leary, 1990), and is proposed to be highest when the interaction partner

32

is attractive, powerful, or expert. Low self-esteem and poor social and

communication skills influence the development of low self-presentational
performance expectancies, which contribute to anxiety when in a context of high
motivation to impress.

33

CHAPTER 2: PRESENT STUDY

2.1 Mate Value Hypothesis: A Context-Specific Theory of Social Anxiety

Individuals have finite time, energy, and resources to invest in attaining a
mate. If rejected by a potential mate, an individual loses those resources
expended in the mating attempt. They lose the mating opportunity itself, the
opportunity to have spent those resources on a different mate, and other mating
opportunities if a negative evaluation of them is shared with others (Graziano,
Jensen-Campbell, Shebilske, & Lundgren, 1993). A mating attempt can also risk
retaliation from an already established mate who is high in power and status, has
strong kin ties, or who is jealous (Buss, 2000). Therefore, determining where to
best invest mating time and effort was likely a significant adaptive challenge over
the course of the evolution of ancestral humans. It is proposed that social anxiety
is context-specific rather than domain-general and different variables account for
social anxiety in different social contexts. One such context, mating, is the focus
of this study and a specific mating anxiety mechanism is proposed, in which at
least one design feature (e.g., a decision rule) specifically evolved to solve the
adaptive problem suggested. If one or more design features of such a
psychological mechanism were molded by selection pressures to defend against
the costs associated with approaching but being rejected by potential mates,

34

anxiety in mating situations should accomplish the following functions: (a)

mating anxiety should serve as a signal, or estimate, of the likelihood of rejection
by a potential mate; (b) the mechanism should lead to avoidance of mating
attempts where the likelihood of rejection is high; and (c) it should trigger
behaviors that minimize the likelihood of rejection during actual mating attempts.
The current study will test the first proposed function, limiting investigation to the
initial phase of mating, that is, evaluation of and by a potential mate. Since
presumably all individuals face this adaptive challenge, the proposed functions
are expected to be universal, in both men and women.
However, a psychological mechanism designed to solve problems related
to mating should be consistent with gender differences in evolved mate
preferences. Men and women differ in the degree to which they prefer certain
characteristics in a mate (Buss, 1989; Buss & Barnes, 1986). For instance,
women place more importance than men on a potential mates degree of
relationship commitment, social status and financial resources, and willingness to
invest resources in her and her offspring (Buss & Schmitt, 1993). Women also
have stronger preferences than men for physical strength, prowess, and social
confidence in a mate. In contrast, men place greater importance than women on
the physical attractiveness of a mate (Buss & Schmitt, 1993; Simpson &
Gangestad, 1992). The mate preferences of one sex are the primary determinants

35

of desirability in the other sex. Stated differently, the degree to which an

individual possesses those qualities desired by members of the opposite sex
determines his or her desirability, or mate value. Since these gender-related mate
preferences dictate desirability, it is hypothesized that differences in these
characteristics should lead to gender differences in mating anxiety. Based on the
documented sex differences in preference for Physical Attractiveness (PA) and
Social Status and Financial Resources (S/R) in a mate (Buss, 1989), this study
will test the effect of these two characteristics on mating anxiety in men and
women.

2.1.1 Mate Value Discrepancy and Likelihood of Rejection (Hypothesis 1)

The first proposed function of the mating anxiety mechanism is that
anxiety in mating situations serves as a signal of the likelihood of rejection.
Prediction 1: Anxiety was expected to increase as Likelihood of Rejection (LR)
increased, in both men and women.
It was further hypothesized that the mechanism bases its estimates of the
likelihood of rejection to some degree on the individuals desirability (mate value)
relative to that of the prospective mate (Mate Value Discrepancy). This
discrepancy was calculated as the difference between the potential mates
desirability minus the participants self-rated desirability. Other factors, such as

36

signals of availability and receptivity or interest, may also influence the likelihood
of rejection but are beyond the scope of this investigation. The following
predictions were made about Mate Value Discrepancy (MVD).
Prediction 2: For both men and women, Likelihood of Rejection (LR) was
expected to be higher for negative MVD scores (i.e., participant desirability was
lower than potential mates desirability); LR was expected to be lower when
MVD score was positive (i.e., when participant desirability exceeded potential
mates desirability).
Prediction 3: Based on the hypothesized association between Likelihood of
Rejection and Anxiety, a relationship similar to Prediction 2 was expected
between MVD and Anxiety.
Prediction 4: Finally, it was speculated that Likelihood of Rejection mediates the
relationship between MVD and Anxiety.

2.1.2 Gender Differences in Mating Anxiety (Hypothesis 2)

If mating anxiety is, indeed, the output of a context-specific evolved
mechanism, it should be consistent with other findings on human mating
behaviors. As such, mating anxiety was hypothesized to be influenced by the
differential emphasis that men and women place on the physical attractiveness
(PA) and status and resources (S/R) of a potential mate. It was also hypothesized

37

that different levels of participants own PA and S/R would lead to gender
differences in anxiety. The following specific predictions were made about the
effect of profile characteristics and participant characteristics on anxiety.
Prediction 5: Men were expected to show greater anxiety than women when
Target Physical Attractiveness level was high than low.
Prediction 6: Women were expected to show greater anxiety than men when
Targets Status/Resources level was high than low.
Prediction 7: Men were expected to show a steeper increase in anxiety as selfrated Status/Resources decreased.
Prediction 8: Women were expected to show a steeper increase in anxiety as selfrated Physical Attractiveness decreased.

2.2 Specific Contributions

The current model extends previous work in two important ways. First,
this study tests a context-specific model of social anxiety, as recommended by
Tooby and Cosmides (1992), hypothesizing design features relevant to only one
social context: mating. If these context-specific hypotheses of social anxiety are
supported by the findings, the domain-general previous theories are undermined.
Second, the present study is the first empirical test of an evolutionary explanation
of social anxiety.

38

CHAPTER 3: METHOD

3.1 Overview of Experimental Design

A 2 (profile PA) x 2 (profile S/R) x 3 (Replicate) x 2 (participant Gender)
mixed model was employed. Gender and participants self-ratings of their
physical attractiveness and status/resources were entered separately as covariates.
Dependent variables were Anticipated Anxiety, Likelihood of Rejection, and
Likelihood of Avoidance.

3.2 Participants
Participants were recruited from a large pool of introductory psychology
students at the University of Texas at Austin, for whom participation fulfilled part
of the course requirements. Since this population is active in mate-seeking, it was
considered an appropriate population on which to test hypotheses about mating
anxiety at the stage of attaining a mate (see Ohman, 1987 for similar
assumptions). Individuals were eligible to participate if they were not involved in
an exclusive dating relationship at the time of the study, were interested in
meeting new people, and spoke English as their first language.

39

3.2.1 Exclusionary Criteria

Because this study was designed to test hypotheses related to mate-seeking
and because only opposite sex profiles were viewed by participants due to
constraints of the testing procedure, participants who indicated any degree of
sexual interest in same sex mates or who indicated being married or engaged at
the time of the study (or who left these items blank) were excluded from the
study. These exclusions resulted in a total sample size to 248, 126 men and 122
women.

3.2.2 Participant Characteristics

As shown in Table 1, the total sample was made up of roughly equal
numbers of male and female undergraduate students at the University of Texas,
with a mean age of 18.1 years old. Most participants were available (90%),
interested (77%), and motivated (97%) to some degree to meet a new dating
partner. A greater proportion of men than women were not dating anyone at the
time of the study (71.4% of men, 46.7% of women). Of those participants were
reported no interest in a new dating partner, a large proportion were already
involved a monogamous relationship, but did not differ significantly from
interested participants on other important variables. (Participants who were not
interested in a new dating partner, or who were monogamous, were not

40

Table 1. Demographics and Dating Status of Study Sample

Total Sample
Variable
n
%
Gender
--Age (range = 17 24)
17 years
33
13.3%
18 years
183
73.8%
19 years
19
7.7%
20 + years
13
5.2%
Ethnicity
Caucasian
148
59.7%
Asian
41
16.5%
Hispanic
32
12.9%
African-American
11
4.4%
Other
14
5.6%
Year in college
230
92.7%
1st
2nd
6
2.4%
3rd 5th
9
3.6%
Current relationship status
Never dated
10
4.0%
None current, dated in past
137
55.2%
Casually dating
33
13.3%
Dating one & seeing others
18
7.3%
Dating one & open to others
25
10.1%
Monogamous, committed
24
9.7%
Number past sex partners
0
104 41.9 %
1-2
87
35.0 %
3-6
43
17.3 %
7+
10
4.0 %
Type of Mate typically want:
LTM
198 79.8 %
STM
50
20.2 %
Type of Mate typically get:
LTM
164 66.1 %
STM
76
30.6 %
Interest in new dating partner
No interest
8
3.2 %
Friends only
50
20.2 %
Mild
79
31.9 %
Moderate
79
31.9 %
High
32
12.9 %
Motivation
Not at all
7
2.8 %
Mild
62
25.0 %
Moderate
133 53.6 %
High
46
18.5 %
Note. LTM = Long-term mate; STM = Short-term mate.

41

Women
n
%
122
49.2%

n
126

Men
%
50.8%

23
89
4
6

18.9%
73.0%
3.3%
4.9%

10
94
15
7

7.9%
74.6%
11.9%
5.6%

84
14
13
5
5

68.9%
11.5%
10.7%
4.1%
4.0%

64
27
19
6
9

50.8%
21.4%
15.1%
4.8%
7.2%

115
0
7

94.3%
0%
5.8%

115
6
2

91.3%
4.8%
1.6%

2
55
20
12
17
15

1.6%
45.1%
16.4%
9.8%
13.9%
12.3%

8
82
13
6
8
9

6.3%
65.1%
10.3%
4.8%
6.3%
7.1%

47
45
24
4

38.5 %
36.9 %
19.7 %
3.2 %

57
42
19
6

45.2 %
33.4 %
15.1 %
4.8 %

111
11

91.0 %
9.0 %

87
39

69.0 %
31.0 %

82
36

67.2 %
29.5 %

82
40

65.1 %
31.7 %

4
29
45
34
10

3.3 %
23.8 %
36.9 %
27.9 %
8.2 %

4
21
34
45
22

3.2 %
16.7 %
27.0 %
35.7 %
17.5 %

5
31
63
23

4.1 %
25.4 %
51.6 %
18.9 %

2
31
70
23

1.6 %
24.6 %
55.6 %
18.3 %

significantly different from more interested or available participants on anxiety.)

Overall, more participants in this sample reported typically wanting to find a
long-term (serious, committed) relationship than a short-term (casual sex)
relationship. Men and women reported seeking long-term relationships to about
the same degree, but when forced to choose which they wanted, the preference for
long-term relationships was overwhelmingly large for women (91% vs. 9%)
compared to men (69% versus 31%). Men in this study were seeking a short-term
relationship significantly more so than were women.
Men and women in this sample rated themselves as above average in
Overall Desirability (Mate Value) and in specific components of desirability, such
as physical attractiveness, listed in Table 2. Few participants rated themselves on
the lower half of the rating scale for Mate Value (9.5% of men and 4.1% of
women).
The sample had a high mean level of global self-esteem (SEQ) and was in
the normal range on all anxiety measures. Statistically significant differences
were found between men and women (women > men) for state anxiety (STAI-S),
trait anxiety (STAI-T), and fear of social situations (LSAS-F). However, these
differences were small and not meaningful. Men and women did not show
significant differences in their levels of dating anxiety (DAS) or social avoidance
(LSAS-A).

42

43

6.70
1.96
16.13
12.84
4.19
3.03
2.31
1.88

240
244
141
245
248
248
248
248
248
248
248
--

Desirability variables
Self-rated Physical Attractiveness
Self-rated Status/Resources
Self-rated Overall Desirability
Waist:Hip Ratio
1.38
1.29
1.31
--

7.31
5.11
1.31
12.01
1.78
1.94
1.04
0.73

SD

122
122
122
122

117
120
73
122
122
122
122
122

6.08
6.28
6.59
0.78

7.80
1.76
15.90
15.66
4.14
2.20
2.14
1.85

Women

1.32
1.16
1.26
.009

8.95
2.52
1.19
13.81
1.75
1.52
0.98
0.77

SD

Anxiety measures
DAS (range = 23-161)
247 82.38 24.82
122
84.03 23.48
LSAS-F (range = 0-72)
248 22.55 11.35
122
24.49 11.61
LSAS-A (range = 0-72)
248 21.06 11.68
122
22.40 11.58
STAI-S (range = 20-80)
248 35.21 11.21
122
37.53 11.97
STAI-T (range = 20-80)
248 39.28 10.47
122
42.01 10.82
SEQ (range = 10-40)
248 33.10
5.57
122
32.12
5.77
Note. LTM = Long-term mate; STM = Short-term mate; DAS = Dating Anxiety Survey; LSAS =
Liebowitz Social Anxiety Scale (Fear & Avoidance Subscales); STAI = State-Trait Anxiety
Inventory (State & Trait subscales); SEQ = Self-Esteem Questionnaire; BIDR = Balanced
Inventory of Desirable Responding (SDE = Self-Deceptive Enhancement, IM = Impression
Management).
***p .001, **p .01, *p .05.

5.92
6.16
6.49
--

Measure
Dating history variables
# past dating partners
# past sex partners
Age 1st intercourse (yrs.)
Longest relationship (mos.)
Degree seeking LTM
Degree seeking STM
Level of interest
Level of motivation

Total sample

Table 2. Means, Standard Deviations, and t-test Comparisons for Gender on Dating history,
Desirability, and Anxiety Variables

126
126
126
126
126
126

126
126
126
123

123
124
68
123
126
126
126
126

80.80
20.66
19.77
32.96
36.63
34.06

5.76
6.04
6.40
0.87

5.60
2.15
16.40
10.04
4.25
3.84
2.48
1.90

Men

26.04
10.80
11.68
9.96
9.43
5.23

1.42
1.41
1.36
.006

5.12
6.74
1.39
9.12
1.82
1.96
1.06
0.70

SD

1.02
2.69**
1.78
3.27***
4.18***
-2.77**

1.86
1.45
1.16
-9.55**

2.37*
-0.60
-1.94
3.76***
-0.47
-7.37**
-2.59**
-0.56

3.3 Stimuli: Date Profiles

Four Mate Types were created from two levels each of the independent
variables, PA and S/R. Three replicates were created for each Mate Type,
resulting in 12 profile stimuli. Each profile consisted of a face and descriptive
information. An example Profile is included as Appendix E. PA level was
conveyed by the photo and by written information about basic physical
appearance (e.g., height, weight). S/R was represented by information about
current educational level and college major (if applicable), future career goal,
current automobile, and level of ambition. In order to minimize the potential
effect of mating strategy (i.e., long- or short-term mating preference) on the
dependent variables, information about both current and future financial resources
was included in the profiles. Filler information, such as favorite food, leisure
activities, dating activities, and a fictitious codename, was also included to
increase the believability of the profiles being part of a dating service. The
profiles were counterbalanced to avoid order effects.

3.3.1 Physical Attractiveness in Profiles

Height and weight figures were generated to be within one SD of mean
height and weight for male and female undergraduate college students at this
university.

44

Faces were selected from a pilot study in which 21 male and 21 female
undergraduate introductory psychology students rated 36 male and 37 female
faces acquired from various hairstyle magazines. Each face was rated by
opposite-sex and same-sex participants on estimated age, attractiveness, and
sexiness. Ratings of the latter two characteristics were made on a scale from1,
not at all attractive or sexy, to 9, extremely attractive or sexy. Based on
means of opposite sex raters only (listed in Table 3), the three male and three
female faces with the highest PA ratings were selected for the high PA profiles,
and the three males and females with the lowest Attractiveness means were used
for the low PA profiles. This resulted in 24 final profiles, 12 profiles of women
and 12 of men.

3.3.2 Status/Resources in Profiles

The occupations selected to be career goals for the fictitious dates were
based on previous research on occupational prestige (Chartrand, Dohm, Dawis, &
Lofquist, 1987). A high degree of stability in judgments of occupational prestige
and status has been shown for the last 70 years (Deeg & Paterson, 1946). The six
high status occupations used were: cardiologist, radiologist, clinical psychologist,
lawyer, stockbroker, and banker/financial investor. The low status occupations
were: courier/delivery driver, electronics and computer repair, hairstylist, phone

45

46

1
2
3
1
2
3
1
2
3
1
2
3

High PA Low S/R

Low PA High S/R

Low PA Low S/R

Replicate

High PA High S/R

Mate Type

1.71 (1.27)
1.29 (0.94)
2.86 (1.33)

2.33 (1.63)
2.05 (1.25)
1.33 (0.94)

7.48 (1.63)
7.67 (1.39)
7.14 (1.64)

7.38 (1.54)
6.86 (1.71)
8.00 (1.47)

1.48 (0.83)
2.24 (1.25)
3.05 (1.49)

2.05 (1.12)
2.29 (1.27)
2.00 (1.10)

7.14 (1.21)
7.29 (1.06)
7.76 (1.08)

7.14 (1.32)
7.38 (1.23)
7.48 (1.06)

Attractiveness
Women
Men
(n = 21)
(n = 21)
M (SD)
M (SD)

1.24 (0.77)
1.10 (0.79)
2.14 (1.51)

1.95 (1.34)
1.33 (0.96)
1.10 (0.85)

6.48 (1.92)
7.00 (1.49)
7.19 (1.71)

Men
(n = 21)
M (SD)

1.43 (0.60)
2.05 (1.10)
2.71 (1.64)

1.71 (0.83)
1.86 (0.94)
1.67 (0.80)

6.91 (1.46)
7.52 (1.25)
8.10 (0.95)

7.52 (1.31)
7.29 (1.37)
7.43 (1.37)

Sexiness

7.24 (1.93)
7.14 (1.85)
8.10 (1.54)

Women
(n = 21)
M (SD)

Face Characteristic

Table 3. Mean Attractiveness and Sexiness Ratings from Pilot Study for the Faces Selected for the Profile Stimuli

sales representative, security guard, and postal clerk. Current educational level
and area of study (if applicable) were generated to logically correspond to the
occupations. The same 12 sets of S/R information (current educational level,
career goal, level of ambition, college major if applicable, and current
automobile) were used for male and female profiles. They were randomly
assigned to the faces.

3.4 Measures
3.4.1 Participant Measures: Mating-Relevant Variables
Demographics measure. This measure, entitled Love Letters Market
Research Survey, was created for this study and is included as Appendix D. It
measured participant age, ethnicity, year in college, past dating history, current
relationship status, level of interest and motivation to meet new people, and
sexual orientation. Filler items were also included to promote believability of the
fictitious dating service.
Dating Preferences and Beliefs. This questionnaire consisted of 34
items frequently used in human mating research, which assess various aspects of
desirability (or mate value) such as physical attractiveness, social status, financial
resources, intelligence, and desirability as a short-term mate (STM) and a longterm mate (LTM). It also assessed participants degree of seeking a LTM or STM

47

and which type of mate they typically find. Items were rated on 7-point or 9-point
Likert scales; two items were forced choice. Although these items have been used
in numerous other studies to measure mate value, no psychometric data is
available.
Waist-to-Hip Ratio (WHR). WHR was calculated by dividing the
circumference of the narrowest point of the waist by the circumference of the
widest point of the buttocks or hips. This measurement is highly correlated with
health and physical attractiveness in men (Singh, 1994) and women (Singh,
1993), and also fertility in women. Each participant measured his or her WHR
using a standard flexible tape measure and a mirror, privately in a nearby
restroom. Participants were given both written and verbal instructions on how to
measure their WHRs, as shown by Appendix B.

3.4.2 Participant Measures: Anxiety and Personality Variables

Dating Anxiety Survey. (DAS; Calvert, Moore, & Jensen, 1987). This 23item self-report measure, entitled Opposite Sex Interactions for this study,
assesses level of anxiety aroused in situations associated with getting, or being on,
a date. Anxiety is rated on a 7-point Likert scale from 1, not at all anxious, to 7,
extremely anxious. DAS scores are correlated with dating history and with
social anxiety, and measure three reliable factors of dating anxiety: Passive

48

Contact, Active Intentions for Dating, and Dating Interactions. Support for
construct and concurrent validity have also been shown (Calvert, Moore, &
Jensen, 1987).
Liebowitz Social Anxiety Scale-Self Report Version (LSAS; Liebowitz,
1987). This self-report measure of fear and avoidance of social situations is based
on the original LSAS clinician-administered interview. The LSAS is commonly
used in pharmacological treatment studies for social phobia (Liebowitz, 1993) and
has been shown to have good internal consistency and test-retest reliability
(Heimberg, Horner, Juster, Safren, Brown, Schneier, & Liebowitz, 1999). The
self-report version has shown high correlations with the LSAS interview version
(rs = .78 to .85) and high internal consistency (rs = ..79 to .95; Baker, Heinrichs,
Kim, & Hofmann, 2002). The questionnaire asks participants to rate their fear
and avoidance of 24 social interaction and performance-oriented situations, thus
created two subscales of 24 items each. Items are rated on a 4-point scale, from 0,
none for Fear or never for Avoidance, to 3, severe for Fear or Usually for
Avoidance. The title Social Anxiety and Avoidance Survey was used in this
study.
State-Trait Anxiety Inventory (STAI; Spielberger, Gorsuch, & Lushene,
1970). This is a widely used, self-report inventory for the assessment of state and
trait anxiety. It contains 2 subscales of 20 items each rated on a 4-point Likert

49

scale. Items depict positive and negative feelings. The state anxiety subscale
(STAI-S) measures the intensity of various current feelings, from not at all to
very much so. The trait anxiety subscale (STAI-T) measures how often
participants generally experience various feelings, thoughts, and attitudes, from
almost never to almost always. For this study, the subscales were entitled
How You Feel (Present) or (General). Scores can range from 20 to 80. Both
subscales have demonstrated good psychometric properties, including a high
internal consistency of about .90 (Spielberger, Gorsuch, & Lushene, 1970).
Participants completed the STAI-S twice, first to establish a baseline anxiety level
and, second, after completing the fictitious dating service, to assess postexperiment anxiety. At the post-assessment, the measure was entitled Reactions
Survey.
Self-Esteem Questionnaire (SEQ; Rosenberg, 1965). This 10-item scale
assesses global self-esteem. Items are rated on a scale of 1-4 for a range of 10-40.
The SEQ has been shown to have high internal consistency (.92; Rosenberg,
1965) as well as good test-retest reliability (r = .85).
Balanced Inventory of Desirable Responding (BIDR; Paulhus, 1994).
This 40-item inventory, renamed the Personal Decisions & Responses Scale for
this study, measures the tendency to endorse socially desirable responses in selfreport. It consists of two subscales: Self-Deception scale (SDS) and Impression

50

Management scale (IMS). The SDS reflects an individuals tendency to attribute

positive characteristics to the self, or self-deceptive positivity (Paulhus, 1986).
This subscale has been shown to be closely associated with self-esteem and other
indicators of adjustment (Paulhus, 1991). The Impression Management scale
captures a more intentional or conscious distortion in self-report endorsement
aimed at being evaluated by a specific audience. The IMS scale is more closely
related with indices such as the Like scale of the Eysenck Personality Inventory
(Eysenck & Eysenck, 1964 in Paulhus, 1991). Items were rated on a 7-point scale
from 1 not very true of me to 7 very true of me, producing a range of 20 to
140.

3.4.3 Profile Measures

Profile Assessment Survey. This 15-item measure was created for this
study and was completed 12 times by each participant (one per profile). It is
included as Appendix F. The first 8 items asked participants to rate the profile on
aspects of desirability including PA, Resources, Social Status, and Overall
Desirability, using a 9-point Likert scale from not at all desirable to extremely
desirable. For the last 7 items, participants were asked to imagine asking the
profile individual to meet or go on a date and then rate their likelihood of asking,
Anticipated Anxiety, and the Likelihood of Rejection, using a 100-point

51

probability/severity scale, where 0 represented not at all likely or no anxiety at

all and 100 was completely certain or extreme anxiety. Additional filler
items were also included.
Date Choice. This was a single item asking participants to record the
name of the profile individual they chose to ask for a date and their reasons,
briefly, for their choice. To promote the believability of the dating service,
written instructions to participants explained that this information was requested
in order to keep track of which types of clients receive the greatest response.
Interest Letter. This pre-formatted letter, included as Appendix G, was
designed to support credibility of the dating service and to arouse more real-life
anxiety about the possibility of being rejected by a desired, prospective date. It
did not assess variables relevant to the study. Participants filled in their responses
to closed-ended or multiple choice questions about their interests, preferences,
basic personal information, and suggested activities for a first date with the profile
individual. They were also given the option to write additional comments on the
back of the Letter. Gender-specific versions were created for male and female
participants.
Assessment of Deception Survey. This 10-item measure was designed for
this study to measure participant beliefs, attitudes, and behaviors that reflected
their degree of belief in the procedures deceptive nature. For instance,

52

participants were asked if they already [knew] any of the Love Letters members
and to what extent [they] believed this dating service was in Austin.
Participants rated the items on a 0%-100% probability scale. This measure was
entitled Opinion Survey and is included as Appendix H.

3.5 Procedure
Participants were run in same sex groups of 2-8. They were seated at
every other seat, which decreased likelihood of participants using social support
to dispelling their anxiety. The test session had two sections. Participants were
led to believe that only the first section of the session was an experiment and that
the second part was not part of the study. The explanation given to participants is
included as Appendix A.
Section 1 lasted about one hour and was led by one same-sex
experimenter. After the session was explained, participants completed the Dating
Preferences and Beliefs measure, SEQ, DAS, BIDR, LSAS, STAI, and then took
their WHR measurements individually in a nearby restroom.
Section 2 of the session, the dating service, lasted 1-2 hours and was led
by two opposite-sex experimenters, who presented themselves as representatives
of the fictitious Love Letters dating service. Participants were each given a
notebook of 12 profiles, Interest Letter, and envelope with fictitious return

53

address and logo, and were given a folder containing the Welcome to Love
Letters page, Market Research Survey, 12 Profile Assessment Surveys, Date
Choice measure, instruction page for completing the Interest Letter, and STAI
post-test (Reactions Survey). After completing these measures and the Interest
Letter, each participant had his or her photo taken, deposited the sealed Interest
Letter in a designated box, and completed the Opinion Survey. The photo was
included to arouse more real-life anxiety about being evaluated by a potential
date.
Once all participants were finished, the group was debriefed and informed
about the deceptive aspects of the study. They were also asked to refrain from
discussing the deception with other potential study participants. After participants
departed, experimenters together completed an Anxiety and Belief Evaluation
Form on each participant to document behavioral indicators of anxiety and degree
of belief.

3.6 Steps to Enhance Manipulation Integrity

In order to enhance the believability of the deception, the following
materials and procedures were included: 1) different experimenters were used for
each section of testing; 2) after WHR measurements were taken, the first
experimenter told participants the study was over, signed all credit sheets, and left

54

the testing room with all materials; 2) the latter two experimenters introduced
themselves as Love Letters representatives and wore shirts with a Love Letters
company logo on them; 3) the two representatives behaved in a professionally
detached manner, as if unfamiliar, with the first section experimenter; 4)
participants were told that staying for the Dating Service part of the session was
voluntary and would not affect their credit; 5) all Section 2 data collection sheets
included the Love Letters logo; 6) the Welcome to Love Letters instruction page
(see Appendix C) was the first page in the data collection folders.

3.7 Experimental Conditions

Participants rated their Anticipated Anxiety, Likelihood of Rejection, and
Likelihood of Avoidance to three Replicates each of four Mate Types: 1) high
PA-high S/R, 2) high PA-low S/R, 3) low PA-high S/R, and 4) low PA-low S/R.

55

CHAPTER 4: STATISTICAL ANALYSES

4.1 Sample Size And Power

Cohens (1977) norm for a medium effect size when using ANOVA is .25.
Given a desired level of power of .8 and alpha of .05 for a two-tailed test, a cell
size of 52 is adequate to detect this size effect. A cell size of 41 is adequate to
achieve a level of power of .70. With a cell size of 100, a level of power of .98 is
achieved to detect a medium effect size (Cohen, 1977; Table 8.3.13). Therefore, a
total sample size of 100 will be adequate to detect moderate size effects of
participant characteristics and profile characteristics on the dependent variables.
To test the predicted interactions of PA and S/R with gender, a total sample size
of 200 is adequate to attain a level of power of .80.

4.2 Analytic Strategy

First, repeated measures MANOVAs were used to confirm that profiles
were perceived by participants as intended. Next, the effect of Replicate was
tested using MANOVAs with Replicate nested within Profile Types. To test the
effects of profile characteristics on the dependent variables, 2 (profile PA) x 2
(profile S/R) x 2 (participant Gender) repeated measures analyses of variance
(ANOVAs), with Replicate nested within cells, were performed. Analyses of

56

covariance (ANCOVAs) were used to test the effects of participant

characteristics. The predictions involving MVD, Likelihood of Rejection, and
Anxiety used hierarchical linear modeling (HLM) Version 5 (Bryk &
Raudenbush, 1992; Bryk, Raudenbush, & Congdon, 1996), which is described
below.

4.2.1 Hierarchical Linear Modeling

HLM is a multi-level, random regression procedure designed for analyzing
data that have a nested structure, where repeated observations are made of each
individual and where measurement occasions are not identical for all individuals.
Here, profile ratings were nested within participant and participants were not
sampled in all the same MVD conditions. This is because, for example, if a
participant rated his own desirability as very high (9 on the 1-9 scale), he would
not have any positive MVD scores (in which profile desirability exceeded his
own), and therefore would not have been exposed to all MVD conditions.
Within HLM, each level of the nested data is represented by its own submodel in the analyses. Variables at the level of repeated observations within
individuals are considered Level 1, while variables at the participant level (i.e.,
characteristics between individuals) are considered Level 2. At Level 1, a
regression equation is computed for each individual and a slope coefficient

57

(unstandardized ) is derived. These slopes are then considered a sample drawn

from a population of slopes, for which the mean is computed and tested against
zero using a t ratio. Then, significant variation among individual slopes can be
tested for with a chi-square statistic.
In Level 2, the degree to which the individual Level 1 slopes were
influenced by various participant characteristics (e.g., gender) are evaluated and
are represented by a new regression equation. The slope coefficient (,
unstandardized) of these regressions indicates the strength of association of the
participant characteristic (e.g., DAS score) with the Level 1 relationship (e.g.,
MVD-anxiety relationship) and a t-ratio is used to determine whether that slope
() differs significantly from zero.

4.3 Specific Analyses

4.3.1 Manipulation Checks
To confirm that the profiles were perceived by participants as intended, a
2 (profile PA) x 2 (profile S/R) x 2 (participant Gender) repeated measures
analysis of variance (ANOVA), with Replicate nested within cells, was conducted
on three manipulation check items: perceived physical attractiveness, perceived
financial resources, and perceived social status.

58

4.3.2 Replicate Effects

Since three replicates were used to represent each mate type, the effect of
Replicate was examined although it was not a main variable of interest. To
examine this, MANOVAs with Replicate nested within Profile Types were
conducted on the manipulation check items and the dependent variables.

4.3.3 Effect of Profile and Participant Characteristics

Repeated measures ANOVAs were used to test the effect of profile PA,
profile S/R, and participant gender on Anxiety. Although no specific predictions
had been made about gender differences in LR and LA, similar analyses were
conducted on these variables. Next, to test the effects of the participant
characteristics (self-rated PA and self-rated S/R) on the dependent variables, each
was added to the model as a covariate. Significant two-way and three-way
interactions were followed up with simple effects analyses.

4.3.4 Effects of MVD and Likelihood of Rejection on Anxiety

The three predicted relationships among MVD, Likelihood of Rejection,
and Anxiety were tested in Level 1, unconditional (i.e., no Level 2 variables)
models. For each significant relationship, an exploratory visual examination of
the data plots was performed to detect any non-linearity of the relationship. When

59

this was the case, a piecewise regression (a 2-piece linear function that has a
breakpoint at which the slope changes significantly) was tested for its goodness of
fit against the linear model, using a chi square statistic. No specific predictions
had been laid out about the influence of participant characteristics (i.e., Level 2
variables) on the Level 1 relationships of interest. However, when significant
variation was found among individual slopes in the Level 1 models, additional
exploratory analyses were conducted at Level 2. Participant variables tested at
Level 2 included gender, self-esteem (SEQ), socially desirable response style
(BIDR), various anxiety measures (DAS, LSAS, STAI), and mating-relevant
factors (SOI, WHR, participant self-rated MV, and degree seeking a LTM or
STM).

4.3.5 Mediation Effect

Baron and Kenny (1986) have outlined that a variable (LR) can be
considered a mediator only if (a) variations in level of the independent variable
(MVD) significantly account for variations in the presumed mediator (LR); (b)
variations in the presumed mediator (LR) significantly account for variations in
the dependent variable (anxiety); and that (c) when these relationships (a and
b) are controlled, the significant relationship between the independent and
dependent variables (MVD and anxiety) becomes non-significant or significantly

60

reduced. Provided the preceding analyses of LR and Anxiety, MVD and LR, and
MVD and Anxiety were found to be significant, LR was added to the Level-1
model of MVD and anxiety to test the mediation hypothesis.

61

CHAPTER 5: RESULTS

5.1 Manipulation Checks

The results of the Assessment of Deception measure suggested that most
participants had believed that the dating service was real. Almost 50% of
participants (47.1%) rated their degree of belief as 75% or higher, with about 28%
rating their belief as 100%. About one-fifth of participants (21.3%) rated a low
degree of belief (< 25%) in the dating service being real. Men and women did not
significantly differ in their scores of degree of belief. With respect to outcome
measures, participants ratings of how strongly they believed the dating service
was real showed no significant relationship with anxiety (p = .67), Likelihood of
Rejection estimates (p = .25), or Likelihood of Approach scores (p = .58).
Results of the repeated measures ANOVAs for the manipulation check
variables suggested that the stimuli were perceived as intended by both men and
women. Means are listed in Table 4. Profiles designated as high in Physical
Attractiveness were rated significantly more attractive than profiles designated as
low in Physical Attractiveness F (1, 236) = 2133.54, p < .001. Likewise, profiles
designated as high in Status and Resources (S/R) were rated significantly higher
in financial resources, F (1, 227) = 1655.58, p < .001 and significantly higher in
social status, F (1, 236) = 437.00, p < .001, than profiles designated as low in S/R.

62

63
3.22
3.68

Perceived Status
Women
Men
5.13
5.20

Note. PA = Physical Attractiveness; S/R = Status/Resources.

0.12
0.12

0.11
0.10

6.90
6.76

2.74
3.25

Perceived Resources
Women
Men
0.15
0.15

0.13
0.13

0.11
0.11

2.01
2.74

2.15
2.48

Measure
Perceived PA
Women
Men
0.10
0.10

High S/R
M
SD

Low S/R
M
SD

Low PA

5.25
5.69

3.68
4.19

6.15
7.15

0.13
0.13

0.12
0.11

0.13
0.13

7.45
7.33

7.93
7.78

0.10
0.10

0.10
0.10

0.12
0.12

High S/R
M
SD
7.12
7.46

High PA
Low S/R
M
SD

Profile Type

Table 4. Means and Standard Errors on Manipulation Check Variables for Men and Women

The profiles of women were rated as slightly more attractive than the
profiles of men, F (1, 236) = 27.06, p < .001 (M = 4.96, SE = 0.08 and M = 4.36,
SE = 0.80, respectively). Male and female profiles were rated similarly with
respect to financial resources and social status.

5.2 Effect of Replicate

Since three replicates were used to represent each Mate Type, the effect of
Replicate on the manipulation check items and on the dependent variables was
examined. Male and female replicates nested within Mate Types were not
perceived by participants as equivalent in physical attractiveness, (8, n = 238) =
20.25, p < .001, financial resources, (8, n = 229) = 23.76, p < .001, or in social
status, (8, n = 238) = 14.58, p < .001, but were roughly similar as illustrated in
Figure 1. With respect to the dependent variables, Replicates within the Mate
Types evoked significantly different levels of anxiety in participants, (35, n =
225) = 295.48, p < .001, significantly different estimates of the likelihood of
being rejected by the profile individual, (35, n = 235) = 94.98, p < .001, and
significantly different estimates of the likelihood of asking the Profile individual
to meet for a date, (35, n = 237) = 318.72, p < .001. Furthermore, female
replicates (i.e., profiles of women) evoked significantly greater anxiety ratings, F
(8, 216) = 6.75, p < .001, likelihood of rejection estimates, F (8, 226) = 2.03, p =

64

Perceived PA

Figure 1. Means for Men and Women on Perceived Attractiveness, Financial

Resources, and Social Status for each Replicate within Mate Type
Women
Men

9
8
7
6
5
4
3
2
1
LL1 LL2 LL3

LH1 LH2 LH3

HL1 HL2 HL3

HH1 HH2 HH3

Fin. Resources

Replicates

9
8
7
6
5
4
3
2
1
LL1

LL2 LL3

LH1 LH2 LH3

HL1 HL2 HL3

HH1 HH2 HH3

Social Status

Replicates

9
8
7
6
5
4
3
2
1
LL1

LL2 LL3

LH1 LH2 LH3

HL1 HL2 HL3

Replicates

65

HH1 HH2 HH3

.04, and Likelihood of Approach ratings, F (8, 228) = 6.28, p < .001, than male
replicates. These Replicate findings support the inclusion of replicates in the
design to represent each Mate Type since participants anxiety, estimated
likelihood of being rejected, and estimated likelihood of asking for a date are not
equivalent across prospective mates. Since Replicate was not a key variable of
interest, the effect of Replicate was not examined in further analyses.

5.3 Effect of Profile Characteristics

5.3.1 Effect of Profile PA and S/R on Anxiety
Means and SDs for the dependent measures, Anxiety, Likelihood of
Rejection, and Likelihood of Approach, are listed by Mate Type in Table 5.
Significant main effects for PA, F (1, 223) = 589.38, p < .001, S/R, F (1, 223) =
99.03, p < .001, and Gender, F (1, 223) = 4.85, p < .03 were found for anxiety.
Men reported higher Anxiety than women, by about 5 points on the 0-100 point
scale (M = 29.59, SD = 1.55 and M = 24.67, SD = 1.60, respectively). As
expected, high attractive profiles evoked higher levels of anxiety than low
attractive profiles. Similarly, profiles high in status and resources evoked higher
levels of anxiety than profiles low in status and resources. These main effects
were qualified by two significant 2-way interactions: S/R x Gender, F (1, 223) =

66

67

32.60
31.52
4.35
9.21

Likelihood of Rejection
Women
Men

Likelihood of Approach
Women
Men
0.99
0.98

1.88
1.85

1.40
1.36

4.67
12.67

34.37
34.72

6.90
12.68

1.11
1.10

1.92
1.90

1.48
1.43

High S/R
M
SD

Note. PA = Physical Attractiveness; S/R = Status/Resources.

6.78
10.58

Measure
Anticipated Anxiety
Women
Men

Low S/R
M
SD

Low PA

24.19
43.98

44.20
49.84

34.18
44.15

2.18
2.15

1.78
1.76

2.38
2.30

37.59
51.23

47.71
50.05

50.83
50.93

2.49
2.46

1.76
1.74

2.57
2.49

High S/R
M
SD

High PA
Low S/R
M
SD

Profile Type

Table 5. Means of Anxiety, Likelihood of Rejection, and Likelihood of Approach

scores by Profile Type and Participant Gender

9.36, p = .002, and PA x S/R, F (1, 223) = 71.21, p < .001. The interaction of PA
x Gender was not significant, contrary to prediction.
With respect to the conditional effect of S/R, men and women did not
differ significantly in their reported anxiety to high S/R profiles; but in response
to low S/R profiles, women had significantly lower anxiety than men, F (1, 237) =
11.14, p = .001. The significant PA x S/R interaction indicated that when profile
PA was high, participants anxiety was significantly greater for high S/R profiles
relative to low S/R profiles, F (1, 232) = 101.96, p < .001. However, under
conditions of low PA, the effect of profile S/R level on anxiety did not reach the
level of statistical significance (p = .068). That is, when participants were
examined as a whole, S/R level did not significantly affect anxiety when profile
PA was low.
The significant 3-way interaction of PA x S/R x Gender, F (1, 223) =
22.20, p < .001 helped further clarify the effect of S/R level on anxiety to low PA
profiles. The 3-way interaction is illustrated in Figure 2. Under conditions of low
profile PA, anxiety in women was not significantly influenced by the S/R level of
the profile (p = .824), whereas anxiety in men was significantly higher when
profile S/R level was high, F (1, 121) = 5.72, p = .018. In the high profile PA
condition, both women (F (1, 114) = 78.73, p < .001) and men (F (1, 118) =
24.39, p < .001) showed significantly higher anxiety to high S/R profiles than low

68

S/R profiles. However, as evident in Figure 2, S/R level led to a bigger difference
in anxiety within the high PA condition for women than men. That is, when PA
was high, womens anxiety to low S/R profiles was much lower than to high S/R
profiles.

Figure 2. Significant 3-way Interaction Effect of Gender x PA x S/R on Anxiety

Anxiety (0 - 100)

Women (n = 109)
60
50
40

High PA
Low PA

30
20
10
0
Low S/R

High S/R

Anxiety (0 - 100)

Men (n = 116)
60
50
40

High PA
Low PA

30
20
10
0
Low S/R

High S/R

One further test was performed to examine the effect of self-presentational

bias on anxiety. First, the assumption of the homogeneity of slopes was tested, by

69

adding each BIDR subscale to the repeated measures ANOVA analysis. No

significant interactions were found between each subscale (Self-Deceptive
Enhancement or Impression Management) and the other between-subject variable
(participant gender). Therefore, an ANCOVA was justified. Self-Deceptive
Enhancement (SDE) had a significant effect on anxiety, F (1, 221) = 14.71, p <
.001, while Impression Management (IM) subscale did not. Both men and
women showed a significant effect of SDE on anxiety scores. As self-deception
decreased, anxiety scores increased. When SDE was covaried, statistically
significant effects remained for Profile PA (F (1, 222) = 42.38, p < .001), Gender
(F (1, 222) = 8.96, p = .003), S/R x Gender (F (1, 222) = 9.30, p < .001), and PA
x S/R x Gender (F (1, 222) = 20.68, p < .001). However, the main effect of S/R
on anxiety and the interaction of PA x S/R became non-significant (p = .23 and p
= .13, respectively).

5.3.2 Effect of Profile PA and S/R on Likelihood of Rejection

Significant main effects were found for PA, F (1, 233) = 151.16, p < .001,
and S/R, F (1, 233) = 21.58, p < .001. That is, participants made higher estimates
of the probability of being rejected by profile individuals who were high in
physical attractiveness or who had high status and resources than who were low in
those characteristics. Men and women did not differ significantly in their

70

estimates of Likelihood of Rejection. These main effects were qualified by a

significant 3-way interaction of Gender x PA x S/R, F (1, 233) = 4.55, p = .03.
Figure 3 illustrates this interaction. Women predicted significantly higher
likelihood of being rejected by profiles with high S/R than profiles with low S/R
only when profile PA was high, F (1, 116) = 14.46, p < .001, and not when profile
PA was low. In other words, women expected a higher likelihood of being
rejected by a highly attractive man who also had high status and resources than by
a highly attractive man who had low status and resources, while status and
resource level made no significant difference in how likely women felt they were
of being rejected by a low attractive man. The reverse was true for men. That is,
men made significantly higher estimates of being rejected by a prospective mate
with high S/R than low S/R if her PA level was low, F (1, 121) = 13.31, p < .001,
whereas profile S/R level made no significant difference in mens estimates of
their likelihood of being rejected by a high PA woman.

71

Figure 3. Significant Effect of Gender x PA x S/R on Likelihood of Rejection

scores

Lik.of Rejection

Women (n = 116)
55
45

High PA
Low PA

35
25

Low S/R High S/R

Lik.of Rejection

Men (n = 119)
55

High PA
Low PA

45
35
25

Low S/R High S/R

A follow-up ANCOVA to examine the effects of Self-Deceptive

Enhancement (SDE) and Impression Management (IM) showed a significant
effect for SDE, F (1, 232) = 12.98, p = .001, but not for Impression Management.
The previously significant effects of PA, S/R, and PA x S/R x Gender on
Likelihood of Rejection remained significant when SDE was covaried. In

72

addition, the interaction of PA x Gender became significant, F (1, 232) = 4.39, p

= .04.

5.3.3 Effect of Profile PA and S/R on Likelihood of Approach

Participants reported higher Likelihood of Approach ratings for profiles
with high PA than low PA, F (1, 235) = 426.64, p < .001, and for profiles with
high S/R than low S/R, F (1,235) = 119.69, p < .001. For all mate types, men
reported a significantly higher likelihood of asking the prospective dates to meet
than did women (M = 29.27, SD = 1.31, and M = 17.70, SD = 1.33, respectively).
Although men were significantly more likely than women to approach both high
and low PA profiles, the influence of high PA on likelihood of approach was
significantly greater for men, as seen in the significant PA x Gender interaction, F
(1, 235) = 11.35, p = .001. In contrast, gender did not moderate the effects of S/R
on likelihood of approach.
The interaction of PA x S/R was significant, F (1, 235) = 47.41, p < .001,
indicating that participants had higher Likelihood of Approach ratings to high S/R
profiles than low S/R profiles at both levels of profile PA, but under conditions of
high PA, S/R level exerted a significantly greater effect on LA ratings. The
preceding 2-way interactions were qualified by a significant 3-way interaction of
PA x S/R x Gender, F (1, 235) = 14.40, p < .001. For women, profile S/R level

73

had no effect on Likelihood of Approach ratings when profile PA was low,

whereas, when profile PA was high, women reported higher Likelihood of
Approach ratings for high S/R profiles, F (1, 119) = 61.51, p < .001. For men, the
effects of S/R level on LA were greater under the high PA condition (Figure 4).

Figure 4. Effect of Gender x PA x S/R on Likelihood of Approach

Lik.of Approach

Women (n = 117)
40

High PA
Low PA

30
20
10
0

Low S/R High S/R

Lik.of Approach

Men (n = 120)
60
50
40
30
20
10
0

High PA
Low PA

Low S/R High S/R

74

5.4 Effect of Participant Characteristics

5.4.1 Effect of Participant PA and S/R on Anxiety
As expected, participants with low self-rated physical attractiveness
reported significantly higher anxiety than participants with high self-assessed
physical attractiveness, F (1, 221) = 9.81, p = .002. Likewise, participants with
low self-assessed status and resources reported higher anxiety than participants
with high status and resources, F (1, 221) = 5.89, p = .016. Contrary to
prediction, the 2-way interactions of Gender x Participant PA and Gender x
Participant S/R were not significant (see Figure 5). Men and women did not
differ in the degree to which their level of physical attractiveness or their level of
status and resources influenced their anxiety. Likewise, no significant higher
order interactions of participant PA and participant S/R on Anxiety were found.

Figure 5. Effect of Self-Rated Physical Attractiveness and Status/Resources on

Anxiety for Men and Women
Anxiety (0 - 100)

50
40

Men
Women

30
20
10
Lo PA

Hi PA

Lo S/R

Hi S/R

Participant Characteristic

75

Exploratory analyses about the effect of participant Waist-to-Hip Ratio

(WHR) were also examined. WHR was significantly associated with self-rated
desirability for women, F (1, 120) = 4.35, p = .04, and for men F (1, 121) = 4.69,
p = .03. However, WHR did not have a significant effect on anxiety for women
(p = .92) or for men (p = .90). In addition, no significant relationship was found
between WHR and number of past sex partners for women (p= .42) or men (p=
.82) or number of past dating partners for women (p= .54) or men (p= .82).

5.4.2 Effect of Participant PA and S/R on Likelihood of Rejection

Participant level of physical attractiveness and status/resources had a
similar effect on Likelihood of Rejection as on Anxiety. Participants estimated a
much higher likelihood of being rejected if they rated themselves as low in
physical attractiveness, F (1, 231) = 46.22, p < .001, or low in status and
resources, F (1, 221) = 5.89, p = .016. The effect of participants PA level and
S/R level on LR did not differ significantly for men and women. There were no
significant higher order interactions of participant PA and participant S/R on LR.
Exploratory analyses about the effect of participant Waist-to-Hip Ratio (WHR) on
Likelihood of Rejection showed no significant relationship for women (p = .72) or
for men (p = .10).

76

5.4.3 Effect of Participant PA and S/R on Likelihood of Approach

In contrast to Anxiety and Likelihood of Rejection, participant self-rated
physical attractiveness and status/resources had no significant effect on
Likelihood of Approach ratings. This was the case for both men and women.
There were also no significant higher order interactions with Self-PA or Self-S/R
and Likelihood of Approach. No significant relationship was found between
participant WHR and Likelihood of Approach scores for either men (p = .52) or
women (p = .58).

5.5 Mate Value Discrepancy (MVD) and Likelihood of Rejection Hypotheses

5.5.1 MVD and Anxiety
The influence of MVD on Anxiety, tested at Level 1, was significant and
in the direction predicted, ( = 7.933, t = 30.96, p < .001). Participant anxiety was
greater at positive values of MVD (i.e., when Profile-MV exceeded Self-MV).
This relationship is represented by the dotted line in Figure 6. A follow-up
exploratory examination of the data, however, suggested that the relationship
between MVD and Anxiety was nonlinear. Depicted by the solid line in Figure 6,
a piecewise regression with a breakpoint at 3 fit the data significantly better than
the linear model, (2 = 212.07, 3 df, p < .001). The first piece of the line showed
that when Profile-MV was much lower than participant MV (below a MVD score

77

of 3), participants reported modestly, but significantly, greater anxiety scores at

each unit increase of MVD ( = 3.52, t = 8.22, p < .001). However, above a
discrepancy value of 3 (where Profile MV was slightly less than, equal to, or
exceeded participant MV), anxiety was greater for each unit increase of MVD by
a rate almost three times that of the first part of the curve, ( = 9.85, t = 26.24, p <
.001). In other words, the growth trajectory for anxiety when Profile MV fell far
below participant MV followed a different pattern than when Profile MV was
close to or exceeded participant MV.

Figure 6. Linear and Piecewise Relationships of Mate Value Discrepancy on

Anxiety
125

Anxiety

100
75

Linear

50

Piecewise

25
0
-25
-8 -6 -4 -2

Mate Value Discrepancy

Unexpectedly, men and women showed different optimal breakpoints

when they were analyzed separately. As shown in Figure 7, the slope of womens
anxiety as a function of MVD had a breakpoint of 3, similar to that of the total
sample ( = 3.71, t = 5.75, p < .001 before breakpoint; = 10.11, t = 20.15, p <
78

.001 after breakpoint). For men, however, the change in slope of anxiety as a
function of MVD occurred at a larger negative MVD value (a breakpoint of 5),
where Profile desirability was much lower than the participants desirability ( = 1.38, t = -1.35, p = .177 before breakpoint; = 8.514, t = 20.07, p < .001 after
breakpoint). Below the breakpoint, the effect of MVD on mens anxiety was nonsignificant.

Figure 7. Piecewise Regressions of Mate Value Discrepancy on Anxiety with

Different Optimal Breakpoints for Women and Men
125

Anxiety

100
75

Women
Men

50
25
0
-25
-8

-6

-4

-2

Mate Value Discrepancy

Since a direct test of the effect of MVD on Anxiety above and beyond the
effects of Profile-MV or Participant-MV could not be conducted in HLM, two
additional analyses were performed. First, the effects of Profile-MV and
Participant-MV were tested in independent models. Both variables had a
significant effect on anxiety, shown in Figure 8 below. As one would expect,
greater anxiety resulted from higher levels of Profile-MV ( = 8.12, t = 31.15, p <
79

.001) and from lower levels of Participant-MV ( = -3.16, t = -3.62, p = .001).

Neither variable showed a distinct breakpoint in its linear relationship with
anxiety. They failed to capture the change in slope in the distribution of anxiety
scores that MVD did capture.

Figure 8. Effects of Profile Desirability and Participant Desirability on Anxiety

100

Anxiety

75
Profile MV

50

Participant MV

25
0
-25
1 2 3 4 5 6 7 8 9

Next, a Level 2 model was used to test whether the effect of Profile-MV
on Anxiety differed depending on the level of Participant-MV. This was
conceptually similar, but not identical, to the interaction of the two variables in
MVD. The result was marginal but non-significant ( = -0.38, t = -1.71, p = .09).
However, when men and women were analyzed separately, a significant effect
was found for women ( = - 0.92, t = -3.03, p = .003), shown in Figure 9. For
women, higher Profile MV led to greater anxiety when self-rated MV was lower.
Participant-MV did not significantly influence the effect of Profile-MV on
Anxiety for men ( = 0.05, t = 0.20, p = .841).
80

Profile-MV on Anxiety

Figure 9. Effect of Profile Desirability on Anxiety as a Function of Participant

Self-rated Desirability
20
15

Women
Men

10
5
0
2

Participant Self-Rated MV
Since significant individual variation was found above and below the
breakpoint, exploratory analyses of several participant characteristics (Level 2
variables) were performed to examine their potential moderating influence. Slope
coefficients () and t-ratios with Bonferroni-corrected significance levels are
listed in Table 6. Higher scores on the DAS and LSAS-Fear scale heightened the
effect of MVD on Anxiety. No other Level 2 variables showed significant
moderating effects. Interestingly, Level 2 effects were found only for the slope
above the breakpoint for the sample and not below.
Although gender itself did not moderate the MVD-Anxiety relationship,
gender differences were found in the effects of other Level 2 variables. Table 7
includes slope coefficients () and t-ratios for these Level 2 effects by gender. As

81

82

1
-0.536
0.006
0.009
0.004
0.052
0.033
-0.008
-0.011
0.001
0.194
-0.121
0.184
-0.250
4.926

se
0.842
0.018
0.041
0.036
0.040
0.046
0.091
0.032
0.033
0.349
0.348
0.256
0.218
11.686

t-ratio
-0.636
0.359
0.232
0.107
1.288
0.711
-0.083
-0.330
0.025
0.555
-0.346
0.716
-1.148
-0.422

2
-0.375
0.058
0.091
0.073
0.066
0.090
-0.140
-0.060
-0.047
0.287
0.257
-0.263
0.310
-20.30

se
0.748
0.016
0.030
0.032
0.037
0.032
0.067
0.027
0.029
0.279
0.283
0.222
0.190
9.963

t-ratio
-0.501
3.628*
3.045*
2.278
1.814
2.810
-2.074
-2.198
-1.655
1.029
0.905
-1.184
1.632
-2.038

Above breakpoint

Note. DAS = Dating Anxiety Survey; LSAS = Liebowitz Social Anxiety Scale,
Fear and Avoidance subscales; STAI = State-Trait Anxiety Inventory, State and
Trait subscales; SEQ = Self-Esteem Questionnaire; BIDR = Balanced Inventory
of Desirable Responding, SDE = Self-Deceptive Enhancement, IM = Impression
Management; PA = Physical Attractiveness; S/R = Status/Resources; LTM =
Long-term Mate; STM = Short-term Mate; WHR = Waist-to-Hip Ratio.
*p .05.

Fixed Effect
Gender
DAS
LSAS-F
LSAS-A
STAI-S
STAI-T
SEQ
BIDR-SDE
BIDR-IM
Participant PA
Participant S/R
Seek LTM
Seek STM
Gender x WHR

Below breakpoint

Table 6. Effect of Level 2 Participant Variables on MVD-Anxiety Relationship

Above and Below the Breakpoint

83
0.008
0.115
0.124
0.086
0.133
-0.369
-0.050
0.060
0.610
0.030
34.09

DAS
LSAS-F
LSAS-A
STAI-S
STAI-T
SEQ
BIDR-SDE
BIDR-IM
Participant PA
Participant S/R
WHR

** p .01; * p .05.

-0.025
-0.095
-0.110
0.004
-0.072
0.206
0.016
0.009
0.618
0.285
-4.564

DAS
LSAS-F
LSAS-A
STAI-S
STAI-T
SEQ
BIDR-SDE
BIDR-IM
Participant PA
Participant S/R
WHR

Fixed Effect

0.058
0.090
0.082
0.101
0.114
0.317
0.078
0.077
0.652
0.664
26.089

0.026
0.061
0.066
0.058
0.065
0.118
0.045
0.053
0.591
0.624
6.298

0.134
1.285
1.523
0.855
1.165
-1.165
-0.644
0.778
0.936
0.046
1.307

-0.933
-1.550
-1.658
0.074
-1.112
1.743
0.362
0.169
1.045
0.456
-0.725
0.053
0.088
0.060
0.064
0.061
-0.034
-0.057
-0.046
0.491
0.081
-20.20

Men (breakpoint -5)

0.064
0.110
0.097
0.086
0.154
-0.308
-0.068
-0.033
-0.707
-0.061
-7.527
0.019
0.039
0.040
0.053
0.045
0.086
0.035
0.030
0.280
0.291
6.338

0.020
0.036
0.041
0.041
0.038
0.080
0.037
0.043
0.384
0.427
6.146

2.810*
2.260
1.497
1.202
1.366
-0.396
-1.642
-1.537
1.754
0.278
-3.186*

3.235*
3.041*
2.344
2.083
4.046**
-3.84**
-1.850
-0.775
-1.842
-0.143
-1.225

Above breakpoint
se
t ratio

Women (breakpoint -3)

Below breakpoint
se
t ratio

Table 7. Effect of Male and Female Level 2 Variables on MVD-Anxiety

Relationship Above and Below Breakpoint

had been observed for the total sample, no Level 2 variables had a significant
effect on the MVD-Anxiety relationship below the breakpoint for men or women.
For women, higher trait anxiety (STAI-T), higher dating anxiety (DAS), higher
fear of social situations (LSAS-F), and lower self-esteem scores (on SEQ)
increased the degree to which positive MVD scores led to higher anxiety. For
men, dating anxiety (DAS) was a significant moderator as it was for women. In
addition, lower WHR in men was associated with a steeper slope after the
breakpoint in the effect of MVD on Anxiety.

5.5.2 MVD and Likelihood of Rejection

MVD had a significant influence on Likelihood of Rejection ( = 2.80, t =
13.02, p < .001), indicating that positive MVD values led to greater estimates of
the probability of being rejected. Figure 10 illustrates the relationship.
Examination of scatterplots and comparison of deviance scores for several
possible breakpoints showed no distinct breakpoint and suggested the relationship
between MVD and LR was adequately captured by a single-slope linear function.
Significant individual variation was found (2 = 501.98, 195 df, p< .001),
but follow-up exploratory analyses of the effect of Level 2 variables found no
significant effects of any Level 2 variables, as listed in Table 8.

84

Likelihood of Rejection

Figure 10. Effect of Mate Value Discrepancy on Likelihood of Rejection

Estimates
75

50

25

0
-8 -6 -4 -2 0

Mate Value Discrepancy

Table 8. Moderating Effects of Level 2 Variables on the Relationship of MVD on

Likelihood of Rejection
Fixed Effect

Coefficient

se

t ratio

Gender
DAS
LSAS-F
LSAS-A
STAI-S
STAI-T
SEQ
BIDR-SDE
BIDR-IM
Participant PA
Participant S/R
G x STAI-T
G x WHR

0.707
0.006
0.036
0.025
0.032
0.052
-0.069
-0.018
-0.014
-0.172
0.041
-0.0002
-12.97

0.428
0.008
0.020
0.017
0.019
0.021
0.040
0.015
0.012
0.147
0.172
0.044
6.547

1.651
0.698
1.790
11.491
1.673
2.512
-1.734
-1.204
-1.137
-1.169
0.241
-0.005
-1.982

*p .05.

85

5.5.3 Likelihood of Rejection and Anxiety

To test Prediction 1 of Hypothesis 2, Likelihood of Rejection and Anxiety
were entered at Level 1 (n for this analysis was 244). As predicted, greater
estimated likelihood of being rejected was associated with higher anxiety scores,
by a factor of .61 ( = 0.609, t = 16.893, p < .001). This relationship is shown in
Figure 11.

Figure 11. Effect of Likelihood of Rejection on Level of Anxiety

Anxiety

100
75
50
25
0
0

25

50

75

100

Likelihood of Rejection
Significant individual variation existed (2 = 501.98, 195 df, p < .001) and
was explored with additional Level 2 analyses. Table 9 shows coefficients and
significance levels for these analyses. Significant moderating effects on the LRAnxiety relationship were found for fear of social situations (LSAS-Fear; =
0.009, t = 3.01, p = .003) and for trait anxiety (STAI-Trait; = 0.009, t = 2.97, p =
.003). At higher scores on each of these scales, higher LR estimates led to greater

86

anxiety. Gender had no significant effect in the degree to which Anxiety

increased at higher LR estimates.
Table 9. Moderating Effects of Level 2 Variables on the Likelihood of RejectionAnxiety Relationship
Fixed Effect

Coefficient

se

t ratio

Gender
DAS
LSAS-F
LSAS-A
STAI-S
STAI-T
SEQ
BIDR-SDE
BIDR-IM
Participant PA
Participant S/R
G x DAS
G x LSAS-F
G x LSAS-A
G x STAI-T
G x WHR

-0.026
0.003
0.009
0.007
0.006
0.009
-0.008
-0.004
-0.001
0.025
0.033
0.001
0.005
0.003
-0.006
-1.055

0.072
0.001
0.003
0.003
0.003
0.003
0.006
0.003
0.002
0.027
0.026
0.003
0.006
0.006
0.007
1.019

-0.361
2.323
3.006*
2.495
1.651
2.970*
-1.296
-1.357
-0.243
0.908
1.294
0.192
0.811
0.536
-0.885
-1.036

*p .05.

5.5.4 Mediation Hypothesis of Likelihood of Rejection

The results of the preceding tests satisfied the first three steps of Baron
and Kennys (1986) guidelines for identifying a mediator variable. For the fourth
step, LR was added to the piecewise Level 1 model of MVD and Anxiety. The
previously significant relationship between MVD and Anxiety remained
87

significant both above and below the breakpoint ( = 2.49, t = 5.73, p < .001
below breakpoint; = 9.39, t = 25.84, p < .001 above breakpoint). To test for
partial mediation, a comparison was made of the 95% confidence intervals for the
MVD-Anxiety slopes and the slopes after LR was added. Considerable overlap
was found. These results suggest that LR did not function as a full or partial
mediator of the MVD-Anxiety relationship.

Figure 12. P-values of Likelihood of Rejection as a Mediator of Mate Value

Discrepancy-Anxiety relationship

LR
p < .001

p < .001

MVD

Anxiety
p < .001

88

CHAPTER 6: DISCUSSION

The aim of the present study was to test specific design features of a
mating context-specific model of social anxiety. The design features included
characteristics of potential mates, characteristics of participants, and cognitive
appraisals made by participants (e.g., estimates of the likelihood of rejection and
mate value discrepancy) that may underlie differences in anxiety. Before
reviewing the major findings, the degree to which participants believed the
deceptive aspects of the procedure and the extent to which the profile stimuli
successfully represented the mate types deserve comment.

6.1 Credibility of the Deception

The success of the deceptive aspects of the testing procedure (i.e.,
convincing participants that the dating service was real) was crucial. Almost half
of participants rated themselves as having believed the dating service to a high
degree (75% or greater), with more than one-quarter rating their degree of belief
as 100%. Furthermore, two-thirds (68.1%) of participants reported that some
aspect of the testing procedure raised their anxiety. About 20% of participants
reported a low level of belief (less than 25%) that the dating service was real, but
were not significantly different on any outcome measures from participants who

89

reported a higher degree of belief. Furthermore, anecdotal reports and

observations of participants gave a similar impression. Participants often
displayed signs of nervousness during testing or relief once debriefed.
Approximately one-third of participants took the opportunity provided to primp
before having their photo taken or provided their own photos to the
experimenters, to be viewed by the prospective mate, which suggested they
believed their letter and photo would be evaluated by the profile individual they
selected. Many participants showed interest in the fictitious dating service
company and asked (either out of interest or out of anxiety) whether a similar
profile would be created on them to be viewed by other dating service members.
These data and observations suggest that participants were successfully deceived
and believed they were taking part in a real dating service.

6.2 Integrity of the Experimental Manipulation

As part of the experimental manipulation, participants were instructed to
imagine asking the profile individual to meet for a date. Participants showed a
high degree of adherence to the instructional set and engagement in the imaginal
task. Over of the sample (77.6%) reported having imagined asking for a date to
a high degree (75% or greater).

90

The stimuli created for this study were designed to represent four types of
mates, which varied on the characteristics of physical attractiveness (low and
high) and social status and resources (low and high). Participants perceptions of
the profiles on the manipulated characteristics showed very high consistency with
the mate type manipulations. The physical attractiveness manipulation accounted
for 90% of the variance in perceptions of profile attractiveness and the
status/resources manipulation accounted for 88% of the variance in perceived
financial resources. The information used to represent high and low financial
resources, which had not been tested before, can be considered valid and useful
for future studies. Similarly, the faces used in the profiles had greater ecological
validity (believability) for a dating service methodology than other faces used in
attractiveness research (e.g., Langlois, Ritter, Roggman, & Vaughn, 1991;
Langlois, Roggman, Casey, & Ritter, 1987). Perceived social status showed a
more unexpected and interesting result.
In designing the profiles, social status was considered conceptually closely
related to financial resources, as is obvious by the use of a single manipulated
variable called status/resources. Therefore, it was thought that information
representing high or low financial resources (e.g., current automobile, future
occupation such as physician or delivery truck driver) would also convey the
impression of high or low social status. The results of the manipulation check

91

analyses unexpectedly showed that profile physical attractiveness accounted for a

slightly larger amount of variance (using partial Eta2) in perceived social status
(71%) than did the S/R manipulation (65%). Certain cues of social status and
financial resources may have been conveyed by the photos (e.g., hairstyle, photo
background). In addition, physical attractiveness in the profile may have simply
been more a salient aspect of the profiles than status/resources information.
The use of replicates in this study also deserves mention. Participants
perceived the replicates of each mate type as significantly different on all
manipulation check variables (although still roughly similar as discussed earlier).
The replicates also elicited significantly different levels of anxiety, different
estimates of rejection likelihood, and different likelihood of being asked for a
date. The presence of significant differences in the perceptions of and responses
to the replicates supports their inclusion in the study and underscores their
importance for future studies of this kind. Despite the differences among
replicates, the effects of the experimentally manipulated profile characteristics
were robust. Together, these data support the integrity of the manipulation and
the validity of these stimuli for similar future research. It is also worthy of note
that no conscious effort at deception (as measured by the BIDR-Impression
Management scale) came into play in anxiety or likelihood of approach ratings.

92

Discussion of the effect of self-deception is discussed in a later section. The lack

of significance for impression management supports the integrity of these data.

6.3 Summary of the Main Findings

The main findings of the present study can be summarized as follows.
Anxiety about asking for a date was influenced by characteristics of the
prospective mates. Higher anxiety was triggered by high physical attractiveness
and by high status/resources in the profiles. Men had higher anxiety than women.
Anxiety was also greater when participants were lower in self-rated physical
attractiveness and self-rated status/resources. Furthermore, the relative
desirability of participants (i.e., the discrepancy between participants judgments
of their own desirability and their subjective impressions of the profiles
desirability) significantly influenced anxiety. Participants had the least anxiety
when they perceived a potential mate as much less desirable than themselves, and
had the most anxiety when a potential mates desirability exceeded their own.
The effect of relative desirability on anxiety showed a critical threshold, or
breakpoint, beyond which there was a threefold increase in the amount of anxiety
triggered by increasingly large discrepancies in mate value. On the continuum of
mate value discrepancies, the location of this breakpoint was lower for men than
for women. Finally, it was found that participants estimated likelihood of being

93

rejected by a potential mate was significantly associated with level of anxiety and
with discrepancy in mate value. However, likelihood of rejection estimates did
not mediate the effect of mate value discrepancy on anxiety. The implications of
these findings and additional unexpected findings are discussed below.

6.4 Factors That Influence Anxiety during Dating Initiation

6.4.1 The Influence of Prospective Mate Characteristics
Higher physical attractiveness in a potential mate led to higher anxiety
when both objective measures of physical attractiveness (i.e., determined by pilot
testing) and the subjective perceptions of participants were used. The same was
found for higher status/resources in the profiles. Level of physical attractiveness
in the potential mates accounted for a much larger amount of variance in anxiety
than Status/Resources. Physical attractiveness accounted for 63.1% of withinsubject variance in anxiety compared to only 1.4% accounted for by the status and
resources of the profiles (with the interaction of the two accounting for another
4% of the within-subject variance in anxiety). This large difference may be partly
a result of the design of the stimuli. The photos in the profile stimuli, which
represented physical attractiveness, were much more prominent or salient than the
brief written information used to convey status/resources information. This
imbalance in the potency of the manipulations of each variable may account for

94

the large difference in the amount of variance in anxiety attributed to each

variable. However, it is also possible that physical attractiveness plays a larger
role than status/resources during brief first impressions of a potential mate, similar
to the present study.
Status/Resource level appeared to potentiate the effect of high profile
physical attractiveness on participants anxiety. When a profile was highly
attractive, having high status and resources as well elicited greater anxiety in
participants. This effect had not been expected. It contradicts Ohmans (1987)
hypothesis that higher social status of the interaction partner is the primary factor
underlying social anxiety.
To our knowledge, this study is the first to examine the influence of the
attractiveness of the interaction partner on anxiety. Clinical research on dating
anxiety, or on social anxiety more broadly, has tended to focus on the
attractiveness of the anxious individual and its relationship to anxiety (e.g.,
Baumeister and Tice, 1990; Glasgow & Arkowitz, 1975, Thornton & Moore,
1993). Ohman (1987) proposed the effect of an interaction partner characteristic
(social status) on anxiety but did not test this. The physical attractiveness of
interaction partners (i.e., potential mates), along with other characteristics, have
been examined in human mating research as they relate to desire and to partner
selection (Buss, 1994; Buss & Schmitt, 1993; Townsend & Levy, 1990), but have

95

not been examined as they relate to anxiety. The present finding is the first
empirical evidence that (non-pathological) social anxiety is more strongly driven
by characteristics of the interaction partner than characteristics of the subject.

6.4.2 The Influence of Gender of Individuals Evaluating Potential Mates

Men reported higher anxiety than women across all mate types. This
general gender effect had not been expected. Several follow-up questions were
tested to clarify this finding. Were the men in this study simply more anxious
than the women? Anxiety measures indicated they were not higher in dating
anxiety (DAS) or in social anxiety than female participants. In fact, men showed
significantly lower scores than women on the LSAS-Fear and the STAI, both state
and trait.
Men in this study rated the (female) profiles as higher in physical
attractiveness than women did (rating male profiles). Were men in this study
presented with profiles that were simply more attractive (and, therefore, more
anxiety-provoking) than those viewed by women? Analysis of the pilot data, in
which the attractiveness of profile faces were rated by both men and women,
suggested this was not the case. The female faces incorporated into the profile
stimuli were not objectively more attractive than the male faces. This incidental
finding suggests that perceptions of the attractiveness of opposite sex individuals

96

may change in situations where mating potential exists, at least for men. This is
consistent with the point made by others that certain environmental criteria must
be present for a properly functioning psychological mechanism to be activated
(Tooby & Cosmides, 1990; Nesse, 1990) as is the case with physiological
mechanisms (e.g., calluses) and supports the context-specificity of social anxiety.
In light of the tests described above, the gender difference found in overall
anxiety to the stimuli likely represents a true sex difference in proneness to
experiencing anxiety in mating situations. Sex differences are precedented in
other domains of anxiety (Shear, Feske, & Greeno, 2000; Yonkers & Kidner,
2002), but by and large women show higher anxiety than men across a number of
anxiety domains: agoraphobic anxiety and panic (Bekker, 1996), generalized
anxiety, post-trauma anxiety and avoidance (Fullerton, Ursano, Epstein, Crowley,
Vance, Kao, Dougall & Baum, 2001), fear to blood/injury (Kleinknecht, 1988)
and to small animals, and even general social anxiety (e.g., Dell'Osso, Saettoni,
Papasogli, Rucci, Ciapparelli, Di Poggio, Ducci, Hardoy, & Cassano, 2002).
Although not representative of a disorder, the present sex difference in anxiety in
mating initiation contexts is the first known to this author to show a higher level
of anxiety in men than women. This finding, unique in the anxiety literature, also
underscores the value of a context-specific approach to conceptualizing and
investigating psychological mechanisms (Tooby & Cosmides, 1990; Tooby &

97

Cosmides, 1992). This sex difference is analogous that found by Silverman,

which had also been guided by context-specific evolutionary hypotheses, in which
women showed superior spatial location memory abilities compared to men the
reverse of the typical finding of mens superiority in this cognitive domain.
Previous domain-general approaches to social anxiety did not uncover this sex
difference in mating initiation anxiety.
Greater pressure or responsibility on men to initiate mating attempts over
ancestral evolutionary time may underlie this observed sex difference. It may
also reflect stronger interest or motivation for mating interactions with novel
mates (i.e., the profile stimuli). Alternatively, as we know, men generally show
greater receptivity to sexual opportunities than women (e.g., Clark & Hatfield,
1989), a tendency to overinterpret womens behavior as indicating sexual interest
(e.g., Haselton & Buss, 2000; Koukounas & Letch, 2001), and greater willingness
to engage in a broader variety of sexual behaviors (e.g., Schmitt, Shackelford,
Duntley, Tooke, & Buss, 2001). Based on their self-reports, men were more
likely to ask the prospective dates to meet than were women in this study. This
contrasts with evidence of greater avoidance and social withdrawal in socially
anxious men than women (Pilkonis, 1977; Snyder & Smith, 1982). When mating
anxiety occurs in normal individuals (i.e., without dysfunctional levels of social
anxiety), it may be more accurately thought of as a positive sort of anticipation

98

rather than an aversive anxiety that would lead to avoidance. This has
implications on the understanding of the function of mating anxiety and is
discussed in a later section.
Women and men were different in how profile status and resources
affected their anxiety. Both showed greater anxiety when a profile was highly
attractive and high in status and resources. However, women also showed the
reverse. That is, a highly attractive potential mate who had low status/resources
elicited a much lower level of anxiety in women than in men. These results
provided support for the expectation, in Prediction 6, that womens anxiety would
be more sensitive to differences in level of status/resources of potential mates.
However, the specific prediction that high S/R in the potential mate would lead to
greater anxiety in women than in men was not borne out. Rather, low profile S/R
led to significantly less anxiety in women than in men. Studies of the effect of
status and resources are limited to their influence on mate preferences and
desirability, and have consistently shown that higher status/resources in a mate is
more important to women than to men (e.g., Buss & Schmitt, 1993). Since
anxiety in the current context appears to track desirability, this finding may help
fine-tune our understanding of womens desire for status/resources. Women may
have a relatively stronger dislike for low status/resources in a mate than a desire
for high status and resources.

99

In Prediction 5, mens anxiety was expected to have been more sensitive

than womens anxiety to higher levels of physical attractiveness in the potential
mates. This prediction was not supported. Instead, women showed roughly the
same degree of difference as men between their anxiety to low attractive profiles
as to highly attractive profiles. Womens anxiety during mating initiation,
therefore, was influenced just as much as mens by the physical attractiveness of a
potential mate. Although, as mentioned, previous research has not examined the
influence of partner characteristics on anxiety, this finding is consistent with other
studies that failed to find a gender difference in the effect of subject attractiveness
on anxiety about social interaction, or communication apprehension (Zakahi,
Duran, & Adkins, 1994).

6.4.3 The Influence of Attractiveness and Status/Resources of Individuals

Evaluating Potential Mates
Participants with lower self-rated physical attractiveness and lower selfrated status/resources had higher anxiety. These findings are consistent with
clinical research findings showing that subjects high in dating anxiety are judged
to be low in physical attractiveness by objective raters (Glasgow & Arkowitz,
1975) and by the participants themselves (Mitchell & Orr, 1976). As one would
expect, lower (self-rated) overall desirability led to greater anxiety. Previous to

100

the present study, the relationship of desirability as a mate and anxiety had not
been examined. This could be related to the significant relationship of selfdeception and anxiety, and the issue of distorted yet healthy positive selfperceptions. Participants in this study rated themselves as higher than the average
in their level of desirability as a mate. In addition, higher self-deception on the
BIDR-SDS scale was associated with lower anxiety in men and in women. As
higher self-deception has been linked with positive self-esteem (Paulhus, 1991),
the present sample can be thought of normal individuals who likely maintain a
healthy degree of positive distortion, or illusory self-efficacy, about their assets,
abilities, and their mating desirability.
No support was found for the predicted gender differences for self-judged
physical attractiveness and status/resources. It was expected (Prediction 7) that
men would show higher anxiety than women when self-rated status/resources
were low. We also expected women to show higher anxiety than men when selfrated attractiveness was low (Prediction 8). The results indicated that men and
women were influenced to the same degree by their self-rated level on these
characteristics. Previous work with clinical populations found no gender
difference in the influence of negative self-appraisals on anxiety in subjects with
high dating anxiety (Clark & Arkowitz, 1975; Schlenker & Leary, 1985). The
present finding extends this observation to non-pathological levels of anxiety.

101

Overall, negative self-appraisals were associated with higher anxiety.

This is consistent with the clinical literature that attributes social anxiety to
dysfunctional self-evaluations (e.g., Schlenker & Leary, 1985). However, the
small size of the influence of self-appraisals on anxiety in this study calls into
question the importance of negative self-evaluations as a primary factor
underlying dating anxiety in non-pathologically anxious individuals. Instead, the
role of negative self-appraisals may become important in the context of
comparisons to appraisals of others as suggested by the significant finding for
Mate Value Discrepancy, discussed below.

6.4.4 The Importance of Mate Value Discrepancy

One of the major hypotheses, that anxiety would track the size and
direction of the discrepancy between participants and potential mates respective
levels of desirability (Prediction 3), was strongly supported. Mate Value
Discrepancy accounted for 75% of the variance in anxiety within individuals.
This MVD effect was present in men and women as expected in Hypothesis 1.
An interesting and unexpected finding was the gender difference in the
location of the MVD breakpoint or threshold. Men apparently began to
experience higher levels of anxiety at larger negative Mate Value Discrepancy
scores than women. The mating anxiety mechanism in men became activated at a

102

lower point, i.e., in response to prospective mates who were much lower in
desirability than did women. Womens mating anxiety, on the other hand,
appeared to become activated when the desirability of the prospective mate was
closer to her own.
Mens lower MVD breakpoint fits conceptually with the more general
finding of higher mating initiation anxiety in men than women. That is, when
opportunities for mating exist, men not only experience higher anxiety than
women but do so for a broader range of potential mates with respect to level of
desirability. In other words, men cast a wider net in the mating pool and are,
therefore, anxious to a broader range of potential mates than women. This greater
anxiety could help alert men to the resulting greater (or more frequent) possibility
of rejection.
The lower breakpoint for men may also be related to a relaxation of
mating standards in certain contexts, which leads to a broader range of potential
mates being acceptable to men (Buss, 1994). The fact that this loosening of
standards occurs in short-term mating for men makes one wonder whether the
dating service methodology of this study (or real dating services) pulled for shortterm mating psychology more than long-term mating strategies. Only a fraction
of participants (25%) rated themselves high in currently seeking a short-term mate

103

(5% of women; 25% of men). However, participants mating intentions toward

the profile individuals were not measured after viewing and rating the profiles.
No direct statistical comparison was possible to test the effect of MVD on
anxiety above and beyond the effect of profile desirability or self-rated
desirability on anxiety. An alternative, but not equivalent, way to examining the
relative mate value concept was to examine the moderating influence of
participant desirability (self-rated) on the relationship of profile desirability to
anxiety. Consistent with the MVD finding, womens level of self-rated
desirability moderated the effect of profile desirability on anxiety. However, men
did not show this moderating effect. This finding is most likely the result of a
restriction in range at the lower end of the distribution of MVD scores for men, a
floor effect, and does not likely suggest that self-rated desirability is irrelevant to
the effect of profile desirability on mens anxiety. Only 13.1% of MVD scores
for men fell below the male breakpoint compared to 42.9% of womens MVD
scores below the female breakpoint. If provided with profiles that were even less
desirable than those included in the current study, men might produce more MVD
scores below the low male breakpoint and thereby increase the ability to detect
any conditional effect of profile desirability on anxiety as in women.

104

6.4.5 The Role of Likelihood of Rejection

Higher Likelihood of Rejection estimates led to higher anxiety, confirming
Prediction 1. Likelihood of Rejection accounted for a significant amount of the
variance in anxiety within individuals (28.2%), but a smaller degree than
accounted for by MVD. MVD significant predicted Likelihood of Rejection
scores, supporting Prediction 2. Participants reported higher chances of being
rejected by potential mates who were higher in relative desirability and lower
chances of being rejected by mates with lower relative desirability. However,
Likelihood of Rejection estimates did not mediate the relationship of Mate Value
Discrepancy and anxiety. Therefore, Prediction 4 was not supported.
Although the present mediation hypothesis had been exploratory in nature,
past clinical research suggested that high anxious individuals have higher
expectations of being negatively evaluated than others (Hope & Heimberg, 1990),
which would increase their predicted likelihood of being rejected. The absence of
a mediation effect for likelihood of rejection may be due to a lack of
pathologically anxious individuals in this study. Among normal individuals, the
likelihood of rejection may play a less important role in anxiety than among
anxious individuals. Research has shown that individuals low in dating anxiety
actually overestimate their performance in a social interaction (Clark & Arkowitz,
1975), and possibly underestimate the chances of rejection. To speculate, dating

105

anxiety in typically non-anxious individuals may set in motion a different strategy

for approaching a potential mate than in typically anxious individuals (i.e., those
with low self-confidence or low dating self-efficacy). That strategy in nonanxious individuals may focus attention, cognitions, and efforts on maximizing
the chances of success during a mating attempt (which may encourage them to
take a more active, confident, or risky approach) rather than playing it safe.
Research tells us that low self-confidence is related to avoidance of mating
attempts with attractive individuals for fear of rejection (Stroebe, 1977) and that
frequent rejection can reduce self-confidence (Nesse, 1990). Further research is
needed to clarify the possibility that estimates of the likelihood of rejection play a
more central role for individuals with pathological levels of dating anxiety.

6.5 Implications of Exploratory Findings

The Waist-to-Hip Ratio findings in this study were somewhat unexpected.
WHR was significantly associated with self-assessments of overall desirability as
a mate, consistent with what one would expect based on the general WHR
literature. Likewise, self-rated desirability was associated with number of past
dating partners for women and for men. However, WHR itself was not related to
number of past sexual partners or dating partners for women or for men. This is
inconsistent with previous research that has shown a link between desirable WHR

106

and greater number of sex partners (e.g., Singh, 1994). Self-rated desirability was
associated with number of past sex partners only for men and not for women. In
addition, WHR heightened the effect of MVD on anxiety in men but not women.
To speculate, other factors may govern whether or not and how frequently
individuals become involved in new mating relationships. For instance,
individuals with less desirable WHRs may find themselves seeking new mates
because they lose their mates more often whereas those with desirable WHRs may
find themselves in new relationships because more opportunities or offers by new
mates arise for them. If this were the case, the effect of WHR on dating and
sexual history could remain obscured without examining the reasons that a new
mate was found. Alternatively, self-rated desirability may developmentally
mediate the effect that WHR has on number of dating partners. Desirable WHR
may increase self-confidence, which is attractive to potential mates in its own
right (Buss, 1990), and may increase ones frequency of approaching mates or
ones receptivity to mating offers which then results in greater dating experience.
These possibilities remain speculative, however, until more research in this area
can be done. Nevertheless, this study indicated that WHR is significantly
associated with subjective perceptions of ones own desirability as a mate. In
order to expand on the present study, further research could also examine the
effect of WHR in the potential mates on dating anxiety in participants.

107

6.5 Implications for Competing Accounts of Social Anxiety

The results of this study support the context-specificity of social anxiety,
and underscore the importance of using a domain-specific approach to examining
psychological phenomena. The results undermine competing theories of social
anxiety and these will be briefly discussed.
The Submissiveness Signal Hypothesis (Ohman, 1987) proposed that
relative social status of an individual underlies social anxiety. Higher social
status in a profile (interaction partner) did lead to higher anxiety, consistent with
the SSH, but was of lesser influence than the physical attractiveness of the
prospective mate. For instance, even when a potential mates social status was
low, individuals experienced higher anxiety if the mate was physically attractive.
This finding undermines the SSH, at least in mating initiation contexts. However,
it does not preclude relative social status from playing a more prominent role in
anxiety in non-mating social intercourse. Further study of the SSH could include
testing the effects of a social status discrepancy score on anxiety a more direct
test of the SSH. In addition, it will be important to test the SSHs hypothesized
adaptive function of preventing hostility from others because, as shown in this
study, individuals may interpret high social status in others as indicating
something other than increased likelihood of hostility (e.g., higher desirability as a
mate, increased likelihood of being rejected for a date).

108

The present study did not address the issue of social incompetence
relevant to the Social Exclusion explanation of social anxiety (Baumeister & Tice,
1990), but has implications on the Social Exclusion Model that deserve comment.
As reviewed earlier, the SEM asserts that anxiety occurs after a social blunder has
been made that could lead to group exclusion. In the present study, individuals
reported (and demonstrated) anxiety before any blunder had been made. The
SEM gives no rationale to expect that individuals would be more likely to commit
a social blunder with highly attractive individuals. Yet, anxiety was greater for
imagined interactions with highly attractive individuals, suggesting that some
variable other than the likelihood of committing a serious social blunder was
driving anxiety. Perhaps, a serious social blunder is one that leads to the loss of
desirable mating opportunities, consistent with Busss (1990) argument? If so,
the SEM likely points in the correct general direction of the function of social
anxiety, but falls short by inadequate specificity and operationalization of its
concepts.
According to the Self-Handicapping Strategy, individuals should become
anxious when they feel a threat to an important dimension to their self-esteem
(Snyder, et al., 1985). It follows that an individual with lower self-esteem should
be more apt to feel threatened and, therefore, would feel socially anxious more
often or more intensely. The present data provide little support for this theory.

109

Scores on the Self-Esteem Questionnaire (Rosenberg, 1979) showed a statistically

significant albeit small effect on anxiety accounting for only .008% of the
variance in anxiety. Better tests of the Self-Handicapping Strategy are needed,
but this preliminary finding suggests that ones trait level of self-esteem has little
influence on mating anxiety.

6.6 Limitations of the Present Study

As mentioned earlier, the manipulation of status and resources in the
stimuli appeared to be much weaker than that of physical attractiveness. This
likely explains the finding that status/resources accounted for substantially less of
the variance in anxiety than did physical attractiveness and is consistent with the
changes in significance of the effect of status/resources when other participant
characteristics were covaried. In future studies, status and resources might be
more powerfully conveyed through the use of videotape representations of
potential mates where clothing, mannerisms of speech, and other non-verbal
information can also convey status and resource information in addition to written
descriptors of income, profession, and level of ambition.
Few participants had mate value discrepancy scores in the very low range,
which may account for the failure to demonstrate the relative desirability effect in
men when testing the effect of profile desirability conditional on level of

110

participant desirability. Another problem was idiosyncratic patterns of anxiety

scores from some individuals that did not fit the distribution for the population.
Presenting participants with more prospective mate stimuli (replicates) would
create a denser sampling and minimize the diluting influences of idiosyncratic
mating preferences or idiosyncratic response styles. In addition, maximizing the
range of sampling by ensuring that many or all participants are sampled at all
levels of mate value discrepancy would be a challenge to accomplish but would
increase the ability to detect the profile MV effect conditional on participant MV,
as well as further clarify the MVD effect.
The hypothetical nature of the assessment methodology may limit the
generalizability of the results. Asking participants to imagine the level of anxiety
they would experience if they asked the target individual for a date may not
correspond to participants actual anxiety when in the real-life situation. Although
others have used fantasy and imaginal tasks to investigate psychological
mechanisms (e.g., in homicidal ideation; Buss & Duntley, 1998; Kenrick &
Sheets, 1993), the relationship between imagined or anticipated anxiety with
actual anxiety in a given situation is tenuous.
Second, the dating service methodology poses limitations on the
generalizability of the findings. Participants were not strictly selected based on
their interest in using a dating service, although they were asked to volunteer for

111

participation only if they were not in a committed monogamous relationship and

were interested in meeting new people. It is possible that the sample differed
from individuals who spontaneously seek out and use dating services in aspects of
their mate value, mate preferences, or proneness to becoming anxious about
dating. Furthermore, the present findings may not generalize to other methods of
initiating contact with potential mates. It must also be remembered that this
method of meeting potential mates is a novelty of modern times and speculation
about the design features of an evolved psychological mechanism must be
relevant to the problems faced in an ancestral context (discussed further below).
Another limitation to the generalizability of the present findings is that
only moderate levels of anxiety were elicited in the present study. The triggers of
very high mating anxiety were, therefore, not examined in this study. The factors
that lead to high levels of mating anxiety could be qualitatively different than
those that elicit moderate levels. This possibility suggests the importance of
further research to more comprehensively understand the mating anxiety
mechanism.
Caution is also necessary in generalizing the present findings to different
populations, particularly to older adults. Differences between younger and older
adults have been found for preferences in short versus long-term mating (Mathes,
King, Miller, & Reed, 2002) and may exist in the factors that lead to anxiety

112

during mate initiation as well. For instance, the incidental finding that physical
attractiveness had a stronger influence on perceptions of social status than did
financial resources may be the case in young adults at the peak of their
reproductive value, whereas older adults may base their perceptions of social
status more heavily on characteristics linked to financial status. Whether older
age is associated with changes in how strongly men and women desire certain
traits in a mate (other than preferred age in a mate; Buunk, Dijkstra, Kenrick, &
Warntjes, 2001) have not been researched. However, it is reasonable to suspect
such differences, which could in turn influence which factors drive anxiety in
older adults when initiating mating.

6.7 Directions for Future Research

The immediate next steps in the present line of research are examination
of the second and third hypothesized functions of the mating anxiety mechanism:
the relationship of anxiety to behavioral avoidance and to performance during
mating attempts. It was proposed that mating anxiety functions to avoid the costs
associated with rejection by a potential mate by (a) avoiding mating attempts with
a high probability of rejection and (b) improving performance during actual
mating attempts (to minimize the probability of rejection). Preliminary analyses
of the present data suggest that anxiety was positively associated with the

113

likelihood of asking the prospective mate to meet, which may be evidence against
the hypothesis that mating anxiety will lead to avoidance. The data contained few
high scores for Likelihood of Rejection, which prevents this question from being
answered adequately without further study. Speculatively, a higher likelihood of
avoidance may be associated with anxiety only when the mating anxiety
mechanism is dysfunctional. Hence, the frequency of avoidance could be a key
discriminator between a functional and dysfunctional mating anxiety mechanism.
It is also possible that a critical threshold in anxiety exists beyond which
avoidance becomes the most likely behavioral output of the mechanism, but
below which a recruitment of resources (e.g., attentional, behavioral, effort) that
would boost the chances of success during the mating attempt is the most likely
output. Alternatively, the mating anxiety mechanisms solution to the costs of
rejection may most often be moderate anxiety which triggers increased interest
and effort, whereas avoidance is a low frequency output because the costs of not
trying to attain a mate are much higher than the costs of being rejected by one.
Further research is needed to address these questions by presenting participants
with a broader range of rejection likelihoods and by examining behavioral
responses during mating attempts that may indicate better preparation or
performance associated with various levels of anxiety. Other factors may also
influence the likelihood of rejection besides relative desirability or the

114

attractiveness and status of the potential mate. These might include signals of
availability, receptivity, or interest. Further research on these factors may help
elucidate the lesser importance of likelihood of rejection that was found in this
sample under these circumstances.
Similar research with older populations is necessary in order to determine
whether the present findings are unique to young adults at the peak of their
reproductive value. Do older men and women face different adaptive problems in
seeking a mate that the younger group studied here does not face? Might the
anxiety-provoking features of prospective mates carry different relative weights
for men and women seeking a mate at an older age? Besides loss of youth, older
adults are more likely to have children by a previous mate. How might this
influence mating anxiety? These questions, which have clinical implications as
well, can only be answered by further research using older populations. More
direct tests of the competing evolutionary accounts of social anxiety, suggested in
earlier sections, are also theoretically valuable avenues for further investigation.
Additional speculation is possible regarding other ancestrally-relevant
adaptive problems that mating anxiety may have been designed to solve. That is,
anxiety during mating initiation may address the costs associated with rejection
but may also have helped prevent or minimize other costs possibly associated
with approaching someone as a potential mate. For instance, the present study

115

found that more highly attractive potential mates aroused higher anxiety. Using
traditional societies today as an analog, ancestral females who were attractive
likely became mated at a young age. Therefore, the chances of an attractive
young woman being already mated was higher than for a less attractive young
woman. This could have meant that approaching a highly attractive woman
would risk retaliation from her high status or powerful mate. In the reverse,
women may have risked paying certain costs if they approached, or even were
approached by, an attractive potential mate. It has been shown that women place
more emphasis on physical attractiveness in a short-term mate (Buss, 1994).
Therefore, if a mated woman approached or was approached by an attractive man,
her mate (if aware of this interaction) might infer that she had short-term mating
interests in the competitor male and might inflict costs on her such as increased
mate-guarding, restriction of her behaviors, and even violence (Shackelford &
Buss, 2000). This adaptive problem for women could drive womens anxiety
during mating initiation, more so than mens.

6.8 Clinical Implications

The large effect of mate value discrepancy on anxiety during mating
initiation suggests an avenue for treatment of socially anxious individuals: to
decrease mate value discrepancies. Socially anxious individuals tend to have

116

significant distortions in their interpretation of external social cues, their detection

of negative responses from others, and their conclusions about how they appear to
others (Clark & McManus, 2002). Therefore, their appraisals of self and potential
mate, which make up mate value discrepancy, are also likely skewed. To the
extent that a socially anxious individuals mate value discrepancy was based on
realistic appraisals, efforts could be spent boosting the individuals perceived
desirability (e.g., improving appearance, improving social interaction skills),
helping the individual better gauge and interpret cues of receptivity from potential
mates, and focus mating attempts to those who are most likely to be receptive.
Unrealistic appraisals may be based in over- or under-estimation of the
importance of certain characteristics to an individuals desirability. Adjusting an
individuals assessment of his relative desirability would involve clarification of
misinformation or misperceptions of what potential mates actually desire and seek
in a mate. Recent research has shown that self-judgments of attractiveness
increase or decrease, respectively, when an individual is exposed to same-sex
individuals who are low or high in attractiveness (Thornton & Moore, 1993).
Individuals with pathologically high dating anxiety may be especially vulnerable
to this contrast effect. Therefore, interventions to minimize the distorting impact
of such social comparison would benefit those with dating anxiety who have
unrealistically low self-judgments of attractiveness and desirability.

117

Dating is not the only social domain in which individuals experience

pathologically high anxiety. It would be interesting to examine the extent to
which anxiety about non-mating social interactions would remain after treatment
of mate value discrepancy distortions, to assess any generalization of treatment
effects beyond dating situations. For instance, could fears of being judged as low
in mate value underlie the anxiety that occurs in other domains, such as in public
speaking phobias? Such clinical applications of the importance of relative mate
value would enhance the theoretical understanding of the mating anxiety
mechanism and help clarify what drives social anxiety context by context.

118

Appendix A. Deceptive Introduction to the Testing Session

There are two parts to todays session.

The first part is the

experiment, which is on the connection between personality and dating

life in college students. You will get a packet of questionnaires that
ask about your attitudes and feelings about dating, your confidence
about dating, your preferences in dating, and a bit about your dating
experiences. When you finish the questionnaires, we will take some
additional physical measurements that have been associated with
dating in previous research students. I will explain more about these
when you reach that part of the packet. Any questions about todays
experiment?
Let me explain the second part of todays session. It is not
actually part of our research, but rather is intended to help a new
company in Austin with their market testing. The company is a new
Introductions/Dating Service, called Love Letters. They are interested
in finding out what appeal their type of service has for a younger
population than their usual clientele that is, young adults less than 25
years old. They contacted our Relationships Research lab, because of
our knowledge about peoples feelings during dating, and asked for
help designing their market research, including helping them connect
with adults of this age range. So, I arranged for a couple of their
Representatives to come here to campus around the time we should be
finishing, so they can show you their service and get your impressions.
Since youll be spending that extra time here today, we will give you
credit for it in addition to the time you spend on the experiment. Any
questions? When the Representatives get here, you can ask them any
questions that I may not have the answer to.
119

Appendix B.
INSTRUCTIONS FOR WAIST AND HIP MEASUREMENTS
Take this sheet, a pencil, and a tape measure to the nearest restroom (the session facilitators will
show you where it is).

To Measure WAIST

To Measure HIPS

Note: Clothes add inaccuracy to the measurements, so be sure the tape measure is pressed
only against your skin, and not your clothing, while measuring.
WAIST MEASUREMENT:
Step 1: Stand facing the mirror and find the narrowest part of your torso (your waist).
Step 2: Place the tape around your waist, flat against your skin with centimeters showing.
Step 3: Take the measurement at the end of a normal exhale, with the tape held snugly against the
skin, but NOT pressed into your flesh.
Step 4: Take your waist measurement three times, for maximal accuracy. Each time, measure the
narrowest part of your waist. Record the measurements in centimenters (to the nearest 0.5 cm) here:
WAIST: (1) _______ cm

(2) ________ cm

(3) ________ cm

HIP MEASUREMENT:
Step 1: Face away from the mirror. Look over your shoulder and locate the area of maximum width
across your buttocks. Be sure your clothes are low enough to not interfere with the measurements.
Step 2: Place the tape measure snugly around your hips at their widest point, as in the diagram
do not press into your flesh or leave it loose.
Step 3: Take three measurements of your hips. Each time, measure the widest point of your
buttocks. Record the measurements in the centimeters (to the nearest 0.5 cm) here:
HIPS: (1) _______ cm

(2) ________ cm
120

(3) ________ cm

Appendix C.

121

Appendix D.

122

Appendix E. Example Stimulus Profile

123

Appendix F. Profile Assessment Survey

124

Appendix G. Interest Letter

125

Appendix H. Assessment of Deception Survey

126

REFERENCES
Abed, R. T. (1998). The sexual competition hypothesis for eating disorders.
British Journal of Medical Psychology, 71, 525-547.
American Psychiatric Association. (1994). Diagnostic and statistical manual of
mental disorders (4th ed.; DSM-IV). Washington DC: Author.
Baker, S. L., Heinrichs, N., Kim, H. J., & Hofmann, S. G. (2002). The Liebowitz
Social Anxiety Scale as a self-report instrument: a preliminary psychometric
analysis. Behaviour Research and Therapy, 40, 701-715.
Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable
distinction in social psychological research: Conceptual, strategic, and
statistical considerations. Journal of Social and Personality Psychology, 51,
1173-1182.
Baumeister, R. F., & Tice, D. M. (1990). Anxiety and social exclusion. Journal
of Social and Clinical Psychology, 9, 165-195.
Bekker, M. H. J. (1996). Agoraphobia and gender: A review. Clinical
Psychology Review, 16, 129-146.
Bryk, A. S., & Raudenbush, S. W. (1992). Hierarchical lineal models:
Applications and data analysis methods. Newbury Park, CA: Sage.
Bryk, A. S., & Raudenbush, S. W., & Congdon, R. T., Jr. (1996). HLM:
Hierarchical linear and nonlinear modeling with the HLM/2L and HLM/3L
Programs. Chicago, IL: Scientific Software International.
Buss, D. M. (1989). Sex differences in human mate preferences: Evolutionary
hypotheses tested in 37 cultures. Behavioral and Brain Sciences, 12, 1-49.
Buss, D. M. (1990). The evolution of anxiety and social exclusion. Journal of
Social and Clinical Psychology, 9, 196-201.
127

Buss, D. M. (1994). The evolution of desire: strategies of human mating. New

York: Basic Books.
Buss, D. M. (2000). The dangerous passion: why jealousy is as necessary as
love and sex. New York: The Free Press.
Buss, D. M., & Barnes, M. (1986). Preferences in human mate selection.
Journal of Personality and Social Psychology, 50, 559-570.
Buss, D. M., & Duntley, J. (1998). Homicidal Ideation. Paper presented at the
annual meeting of the Human Behavior and Evolution Society, Davis, CA,
US.
Buss, D. M., Haselton, M. G., Shackelford, T., Bleske, A. (1996). The
evolutionary psychology of mental disorder: The proper function model.
Unpublished manuscript.
Buss, D. M., Larsen, R. J., Westen, D., & Semmelroth, J. (1992). Sex differences
in jealousy: Evolution, physiology, and psychology. Psychological Science,
3, 251-255.
Buss, D. M., & Schmitt, D. P. (1993). Sexual strategies theory: An evolutionary
perspective on human mating. Psychological Review, 100, 204-232.
Buunk, B. P., Dijkstra, P., Kenrick, D. T., & Warntjes, A. (2001). Age
preferences for mates as related to gender, own age, and involvement level.
Evolution and Human Behavior, 22, 241-250.
Calvert, J. D., Moore, D., & Jensen, B. J. (1987). Psychometric evaluation of the
Dating Anxiety Survey: A self-report questionnaire for the assessment of
dating anxiety in males and females. Journal of Psychopathology and
Behavioral Assessment, 9, 341-350.
Chartrand, J. M., Dohm, T. E., Dawis, R. V., & Lofquist, L. H. (1987).
Estimating occupational prestige. Journal of Vocational Behavior, 31, 14-25.

128

Clark, D. M., & McManus, F. (2002). Information processing in social phobia.

Biological Psychiatry, 51, 92-100.
Clark, J. V., & Arkowitz, H. (1975). Social anxiety and self-evaluation of
interpersonal performance. Psychological Reports, 36, 211-221.
Clark, R. D., & Hatfield, E. (1989). Gender differences in receptivity to sexual
offers. Journal of Psychology and Human Sexuality, 2, 39-55.
Cohen, J. (1977). Statistical power analysis for the behavioral sciences. Revised
Edition. New York: Academic Press.
Cosmides, L., & Tooby, J. (1999). Toward an evolutionary taxonomy of
treatable conditions. Journal of Abnormal Psychology, 108, 453-464.
Deeg, M. E., & Paterson, D. G. (1946). Changes in social status of occupations.
Occupations, 25, 205-208.
Dell'Osso, L., Saettoni, M. Papasogli, A., Rucci, P., Ciapparelli, A., Di Poggio,
A.B., Ducci, F., Hardoy, C., & Cassano, G. B. (2002). Social anxiety
spectrum: Gender differences in Italian high school students. Journal of
Nervous & Mental Disease, 190, 225-232.
Fullerton, C. S., Ursano, R. J., Epstein, R. S., Crowley, B., Vance, K., Kao, T.,
Dougall, A. & Baum, A. (2001). Gender differences in posttraumatic stress
disorder after motor vehicle accidents. American Journal of Psychiatry, 158,
1486-1491.
Garcia, S., Stinson, L., Ickes, W., Bissonnette, V., & Briggs, S. R. (1991).
Shyness and physical attractiveness in mixed-sex dyads. Journal of
Personality and Social Psychology, 61, 35-49.
Gilbert, P. (1989). Human Nature and Suffering. Hillsdale, New Jersey:
Lawrence Erlbaum.

129

Glasgow, R., & Arkowitz, H. (1975). The behavioral assessment of male and
female social competence in dyadic heterosexual interactions. Behavior
Therapy, 6, 488-498.
Goodall, J. (1986). Social rejection, exclusion, and shunning among the Gombe
chimpanzees. Ethology and Sociobiology, 7, 227-236.
Graziano, W. G., Jensen-Campbell, L. A., Shebilske, L. J., & Lundgren, S. R.
(1993). Social influence, sex differences, and judgments of beauty: Putting
the interpersonal back in interpersonal attraction. Journal of Personality &
Social Psychology, 65, 522-531.
Harpending, H. C., & Sobus, J. (1987). Sociopathy as an adaptation. Ethology
and Sociobiology, 8 Suppl, 63-72.
Haselton, M. G., & Buss, D. M. (2000). Error management theory: A new
perspective on biases in cross-sex mind reading. Journal of Personality and
Social Psychology, 78, 81-91.
Heimberg, R. G., Horner, K. J., Juster, H. R., Safren, S. A., Brown, E. J.,
Schneier, F. R., & Liebowitz, M. R. (1999). Psychometric properties of the
Liebowitz Social Anxiety Scale. Psychological Medicine, 29, 199-212.
Hope, D. A., & Heimberg, R. G. (1990). Dating anxiety. in Harold Leitenberg
(Ed.) Handbook of Social and Evaluative Anxiety (pp. 217-246). New York:
Plenum Press.
Jaremko, M. E., Myers, E. J., Daner, S., Moore, S., & Allin, J. (1982).
Differences in daters: Effects of sex, dating frequency, and dating frequency
of partner. Behavioral Assessment, 4, 307-316.
Kenrick, D. T., & Sheets, V. (1993). Homicidal fantasies. Ethology &
Sociobiology, 14, 231-246.
Kleinknecht, R. A. (1988). Specificity and psychosocial correlates of
blood/injury fear and fainting. Behaviour Research & Therapy, 26, 303-309.

130

Kluger, M. (1997). The adaptive value of fever. In P. MacKowiak (Ed.) Fever:

Basic measurement and management (2nd ed., pp. 255-266). Philadelphia:
Lippincott-Raven.
Koukounas, E., & Letch, N. M. (2001). Psychological correlates of perception of
sexual intent in women. Journal of Social Psychology, 141, 443-456.
Langlois, J. H., Ritter, J. M., Roggman, L. A., & Vaughn, L. S. (1991). Facial
diversity and infant preferences for attractive faces. Developmental
Psychology, 27, 79-84.
Langlois, J. H., Roggman, L. A., Casey, R. J., & Ritter, J. M. (1987). Infant
preferences for attractive faces: Rudiments of a stereotype? Developmental
Psychology, 23, 363-369.
Larsen, R. J., & Shackelford, T. K. (1996). Gaze avoidance: Personality and
social judgments of people who avoid direct fact-to-face contact. Personality
and Individual Differences, 21, 907-917.
Leary, M. R. (1990). Responses to social exclusion: Social anxiety, jealousy,
loneliness, depression, and low self-esteem. Journal of Social and Clinical
Psychology, 9, 221-229.
Liebowitz, M. R. (1987). Social phobia. Modern Problems in Pharmacopsychiatry, 22, 141-173.
Liebowitz, M. R. (1993). Pharmacotherapy of social phobia. Journal of Clinical
Psychiatry, 54 (suppl), 31-35.
Lilienfeld, S. O., & Marino, L. (1995). Mental disorder as a Roschian concept:
A critique of Wakefields Harmful Dysfunction analysis. Journal of
Abnormal Psychology, 104, 411-420.
Mathes, E. W. (1975). The effects of physical attractiveness and anxiety on
heterosexual attraction over a series of five encounters. Journal of Marriage
and the Family, 37, 769-773.

131

Mathes, E. W., King, C. A., Miller, J. K., & Reed, R. M. (2002). An

evolutionary perspective on the interaction of age and sex differences in shortterm sexual strategies. Psychological Reports, 90, 949-956.
McGuire, M. T., and Troisi, A. (1998). Prevalence differences in depression
among males and females: Are there evolutionary explanations? British
Journal of Medical Psychology, 71, 479-491.
McNally, R. J. (1987). Preparedness and phobias: A review. Psychological
Bulletin, 101, 283-303.
Mealey, L. (1995). The sociobiology of sociopathy: An integrated evolutionary
model. Behavioral and Brain Sciences, 18, 523-541.
Meichenbaum, D. H. (1977). Cognitive behavior modification. New York:
Plenum Press.
Mitchell, K. R., & Orr, F. E. (1976). Heterosexual social competence, anxiety,
avoidance, and self-judged physical attractiveness. Perceptual and Motor
Skills, 43, 553-554.
Muehlenhard, C. L., Koralewski, M. A., Andrews, A. L., & Burdick, C. A.
(1986). Verbal and nonverbal cues that convey interest in dating: Two
studies. Behavior Therapy, 17, 404-419.
Nesse, R. M. (1990). Evolutionary explanations of emotions. Human Nature, 1,
261-289.
Nesse, R. M. (1987). An evolutionary perspective on panic disorder and
agoraphobia. Ethology and Sociobiology, 8, 73s-83s.
Nesse, R. M., & Williams, G. C. (1994). Why we get sick: The new science of
Darwinian medicine. New York: Random House Times Books.
OBanion, K., & Arkowitz, H. (1977). Social anxiety and selective memory for
affective information about the self. Social Behavior and Personality, 5, 321328.
132

Ohman, A. (1986). Face the beast and fear the face: Animal and social fears as
prototypes for evolutionary analyses of emotion. Psychophysiology, 23, 123145.
Ohman, A. (1987). Evolution, learning, and phobias: An interactional analysis.
in D. Magnusson & A. Ohman (Eds.) Psychopathology: An interactional
perspective (pp. 143-158). Orlando: Academic Press.
Paulhus, D. L. (1986). Self-deception and impression management in test
responses. In A. Angleitner & J. S. Wiggins (Eds.), Personality assessment
via questionnaire (pp.142-165). New York: Springer.
Paulhus, D. L. (1991). Enhancement and denial in socially desirable responding.
Journal of Personality and Social Psychology, 60, 307-317.
Paulhus, D. L. (1994). Balanced inventory of desirable responding: Reference
manual for BIDR Version 6. Unpublished manuscript, University of British
Columbia.
Pilkonis, P. A. (1977). The behavioral consequences of shyness. Journal of
Personality, 45, 596-611.
Price, J. (1998). The adaptive function of mood change. British Journal of
Medical Psychology, 71, 465-477.
Profet, M. (1987). The evolution of pregnancy sickness as protection to the
embryo against Pleistocene teratogens. Evolutionary Theory, 8, 177-190.
Profet, M. (1991). The function of allergy: Immunological defense against
toxins. The Quarterly Review of Biology, 66, 23-62.
Rosemberg, F. K., Garcia. J. H., & Perez S., L. (1998). Staying well in an ailing
society. The Behavior Therapist, 21, 174-175.
Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ:
Princeton University Press.

133

Schlenker, B. R., & Leary, M. R. (1982). Social anxiety and self-presentation: A

conceptualization and model. Psychological Bulletin, 92, 641-649.
Schlenker, B. R., & Leary, M. R. (1985). Social anxiety and communication
about the self. Journal of Language and Social Psychology, 4, 171-192.
Schmitt, D. P., Shackelford, T. K., Duntley, J., Tooke, W., & Buss, D. M. (2001).
The desire for sexual variety as a key to understanding human mating
strategies. Personal Relationships, 8, 425-455.
Seligman, M. (1971). Phobias and preparedness. Behavior Therapy, 2, 307-320.
Shackelford, T. K., & Buss, D. M. (2000). Marital satisfaction and spousal cost
infliction. Personality and Individual Differences, 28, 917-928.
Shear, M. K., Feske, U., & Greeno, C. (2000). Gender differences in anxiety
disorders: Clinical implications. In E. Frank (Ed.) Gender and its effects on
psychopathology. American Psychopathological Association series. (pp. 151165). Washington, DC: American Psychiatric Publishing.
Simpson, J. A., & Gangestad, S. W. (1992). Sociosexuality and romantic partner
choice. Journal of Personality, 60, 31-51.
Singh, D. (1993). Adaptive significance of female physical attractiveness: Role
of waist-to-hip ratio. Journal of Personality and Social Psychology, 65, 293307.
Singh, D. (1994). Effect of waist-to-hip ratio on womens ratings of
attractiveness of male line drawings. Lecture to Department of Psychology,
University of Texas at Austin.
Sloman, L., & Price, J. S. (1987). Losing behaviour (yielding subroutine) and
human depression: Proximate and selective mechanisms. Ethology and
Sociobiology, 8, 99s-109s.

134

Smith, R. E., & Sarason, I. G. (1975). Social anxiety and the evaluation of
negative interpersonal feedback. Journal of Consulting and Clinical
Psychology, 43, 429.
Smith, T. W., Snyder, C. R., & Handelsman, M. (1982). On the self-serving
function on an academic wooden leg: Test anxiety as a self-handicapping
strategy. Journal of Personality and Social Psychology, 42, 314-321.
Snyder, C. R., & Smith, T. W. (1982). Symptoms as self-handicapping
strategies: The virtues of old wine in a new bottle. In G. Weary & H. L.
Mirels (Eds.), Integrations of clinical and social psychology (pp. 104-127).
New York: Oxford University Press.
Snyder, C. R, Smith, T. W., Augelli, R. W., & Ingram, R. E. (1985). On the selfserving function of social anxiety: Shyness as a self-handicapping strategy.
Journal of Personality and Social Psychology, 48, 970-980.
Spielberger, C. D., Gorsuch, R. L., & Lushene, R. E. (1970). Manual for the
State-Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists
Press.
Stroebe, W. (1977). Self-esteem and interpersonal attraction. In S. W. Duck
(Ed.), Theory and practice in interpersonal attraction (pp.79-104). London:
Academic Press.
Symons, D. (1992). On the use and misuse of Darwinism in the study of human
behavior. In J. H. Barkow, L. Cosmides, & J. Tooby (Eds.) The adapted
mind: Evolutionary psychology and the generation of culture (pp. 137-159).
New York: Oxford University Press.
Thornton, B., & Moore, S. (1993). Physical attractiveness contrast effect:
Implications for self-esteem and evaluations of the social self. Personality
and Social Psychology Bulletin, 19, 474-480.
Tooby, J. & Cosmides, L. (1990). The past explains the present: Emotional
adaptations and the structure of ancestral environments. Ethology and
Sociobiology, 11, 375-424.
135

Tooby, J. & Cosmides, L. (1992). The psychological foundations of culture. in

J. H. Barkow, L. Cosmides, & J. Tooby (Eds.) The adapted mind:
Evolutionary psychology and the generation of culture (pp. 19-136). New
York: Oxford University Press.
Townsend, J. M., & Levy, G. D. (1990). Effects of potential partners physical
attractiveness and socioeconomic status on sexuality and partner selection.
Archives of Sexual Behavior, 19, 149-164.
Trower, P., & Gilbert, P. (1989). New theoretical conceptions of social anxiety
and social phobia. Clinical Psychology Review, 9, 19-35.
Trower, P., Gilbert, P., & Sherling, G. (1990). Social anxiety, evolution, and
self-presentation: An interdisciplinary perspective. In Harold Leitenberg
(Ed.) Handbook of Social and Evaluative Anxiety. New York: Plenum Press,
pp. 11-45.
Twentyman, C. T., & McFall, R. M. (1975). Behavioral training of social skills
in shy males. Journal of Consulting and Clinical Psychology, 43, 384-395.
Udry, J. R., & Eckland, B. K. (1984). Benefits of being attractive: Differential
payoffs for men and women. Psychological Reports, 54, 47-56.
Wakefield, J. C. (1992). Disorder as harmful dysfunction: A conceptual critique
of DSM-III-Rs definition of mental disorder. Psychological Review, 99, 232247.
Wakefield, J. (1999). Evolutionary versus prototype analyses of the concept of
disorder. Journal of Abnormal Psychology, 108, 374-399.
Yonkers, K. A., & Kidner, C. L. (2002). Sex differences in anxiety disorders. In
F. Lewis-Hall, T. S. Williams, J. A. Panetta, & J. M. Herrera (Eds.),
Psychiatric illness in women: Emerging treatments and research (pp. 5-30).
Washington, DC: American Psychiatric Publishing, Inc.
Zakahi, W. R., Duran R. L., & Adkins, M. (1994). Social anxiety, only skin
deep? The relationship between ratings of physical attractiveness and social
anxiety. Communication Research Reports, 11, 23-31.
136

VITA

Susana Lucia Kugeares was born in Ann Arbor, Michigan, on November

30, 1970 to Sophia Manoulian Kugeares, of Dearborn, MI, and Sotirios George
Kugeares of Nemea, Korinthias, Greece. After completing high school in
Clearwater, Florida, in 1988, she entered Florida State University where she
earned her Bachelors of Science degree in Psychology. After a subsequent year
of work in research on psychopathy at Florida State University and the Florida
Correctional Institute in Tallahassee, she entered the University of Texas at
Austin in August 1993 to pursue her doctoral degree in Clinical Psychology. She
received her Master of Arts degree in May 1996 for research on gender
differences in behavioral manifestations of psychopathic personality. She began
studying evolutionary psychology in 1996. In August 2000, she moved to Tampa,
Florida to complete her internship in Clinical Psychology at the James A. Haley
Veterans Hospital, where she then remained as staff in Psychology. Her present
clinical work focuses on veterans coping with disease, treatment of sexual
dysfunctions, and psychological assessment of severe psychopathology and
malingering.

Permanent address:

2815 W. Sitios Street, Apt. #1

Tampa, FL 33629-6137

This dissertation was typed by the author.

137