R «3: -
‘fi
‘

,,‘

‘ée .‘.'A.~y‘

r.‘
I

_...:x.-

W»

t

. n»

1. ud-
, ". Jam.- ,
.«.. .. 3” ”w...- ‘

u.

Ii

. (mm—,— v.

 

 

\

' .-. 7V0 ..~u..— w

 

-.- - -»-.4»u...o

... VA»; n .. a... .
.

9097/

SQLLQDSI

This is to certify that the
dissertation entitled

THE EFFECT OF BEARING SOLO STATUS WITHIN AN
APPLICANT POOL ON NEWLY HIRED FEMALES'
COGNITIONS, AFFECT, AND BEHAVIOR

presented by
JENNIFER ZOPHY GILLESPIE

has been accepted towards fulfillment
of the requirements for the

Doctoral degree in Psychology

 

 

/%/flé,

Major Professor’sfififiature
SfflK/o 7

Date

MSU is an Affinnative Action/Equal Opportunity Institution

---.-n-.-- --u-o---0------—--o---a-c-n-n-n-n-o—o-n-p-n-u-.-.-.-.-.-

 

4. o ~o--—o——+- —

«aw—o ~ 9 ~0r~

o -.. —-fi-—-—.-—-. 0‘

_.‘o -o~.

--o O —o— O———ov-.———-o -"~

4.’-"""-9

 

 

UBRARY

Michigan State
University

 

 

 

PLACE IN RETURN Box to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c:/CIRC/DaIeDue.p65-p.15

THE EFFECT OF BEARING SOLO STATUS WITHIN AN APPLICANT POOL
ON NEWLY HIRED FEMALES’ COGNITIONS, AFFECT, AND BEHAVIOR

By

Jennifer Zophy Gillespie

A DISSERTATION

Submitted to
Michigan State University
In partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Psychology

2004

ABSTRACT

THE EFFECT OF BEARING SOLO STATUS WITHIN AN APPLICANT POOL
ON NEWLY HIRED FEMALES’ COGNITIONS, AFFECT, AND BEHAVIOR

By

Jennifer Zophy Gillespie

The process by which bearing solo status within one’s applicant pool affects
newly hired females’ cognitions, affect, and behavior (including the presence of
individual difference variables that may moderate the relationship) has not yet been
specified. However, existing bodies of work (on solo status and stereotype threat)
collectively yield points of theoretical convergence from which a framework can be
developed. Such a framework was developed and then used (along with Heilman’s
research on preferential selection) to propose a model of the relationship between bearing
solos status within an applicant pool and the cognitions, affect, and behavior of females
recently hired for a traditionally masculine job. Study results are used to provide

recommendations for future research and practice.

TABLE OF CONTENTS

LIST OF TABLES .............................................................................................................. v
LIST OF FIGURES ............................................................................................................ vi
INTRODUCTION ............................................................................................................... I
CONSTRAINTS AND CAVEATS .................................................................................... 4
RELEVANT RESEARCH .................................................................................................. 5
Solo Status ............................................................................................................... 5
Consequences .............................................................................................. 8
Potential Moderators ................................................................................... 9
Potential Mediators ................................................................................... 11
Summary ................................................................................................... 15
Stereotype Threat .................................................................................................. 18
The Present Review: A Cautionary Note .................................................. 22
Antecedents ............................................................................................... 24
Consequences ............................................................................................ 27
Potential Moderators ................................................................................. 30
Potential Mediators ................................................................................... 32
Summary ................................................................................................... 34
Solo Status within a Selection Context ................................................................. 36
Heilman (I980) ......................................................................................... 39
Heilman and Blader (2001) ....................................................................... 4O
Heilman and Alcott (2001) ........................................................................ 42
Summary ................................................................................................... 46

POINTS OF THEREOTICAL CONVERGENCE BETWEEN THE SOLO STATUS
AND STEREOTYPE THREAT LITERATURES ........................................................... 50
PROPOSED MODEL ....................................................................................................... 55
METHOD .......................................................................................................................... 64
Overview ............................................................................................................... 64
Participant Recruitment and Sign-Up Process .......................................... 65
Key Features of the Cover Story ............................................................... 66
Procedure ............................................................................................................... 68
Study Design and Participants ............................................................................... 69
Solo Status Manipulation ...................................................................................... 71
Measures ................................................................................................................ 74
Self-Efficacy .............................................................................................. 74
Selection Test ............................................................................................ 74
Perception of Others’ Confidence in One’s Own Ability ......................... 75
Work Sample Performance ....................................................................... 76

iii

. .- fl-l

Affect ......................................................................................................... 78

Task Difficulty Choice .............................................................................. 80
Feedback-Seeking ..................................................................................... 81

Stereotype Threat ...................................................................................... 81
Meta-Perception of Preferential Selection ................................................ 83

RESULTS .......................................................................................................................... 85
Manipulation Check .............................................................................................. 85

Tests of Hypotheses and Model ............................................................................ 85
Individual Hypothesis Tests ...................................................................... 88

Model Test ................................................................................................. 91
Respecifications Decision Points Common to All Models ................................... 95
Unique Respecifications by Dependent Variable .................................................. 98

Task difficulty choice .................................................................... 98

F eedback-seeking .......................................................................... 98

Work sample performance .......................................................... 102

Positive aroused affect ................................................................ 104

Negative aroused affect ............................................................... 104

F ollow-up Analyses ............................................................................................. 106
DISCUSSION ................................................................................................................. 108
Key Findings ....................................................................................................... 108
Unexpected Findings ........................................................................................... 110

Study Limitations ................................................................................................ 111
Directions for Future Research ........................................................................... 114
Concluding Comment ......................................................................................... 116
APPENDICES ................................................................................................................. 1 17
Appendix A ......................................................................................................... 118
Appendix B ......................................................................................................... l 19
Appendix C ......................................................................................................... 120
REFERENCES ................................................................................................................ 127

iv

LIST OF TABLES

Table l - Sample Definitions ofStereotype Threat ........................................................... 19
Table 2 - Sample Definitions of a Stereotype (from Greenwald & Banaji, I995) ............ 20
Table 3 - Sample Operationalizations ofStereotype Threat ............................................. 25

Table 4 - Potential Moderators of the Effect of Stereotype Threat on Test Performance
(from Steele et al.. pp. 391-392, 394-397) ................................................................ 31

Table 5 - Study Measures and Sample Sizes ..................................................................... 70

Table 6 - Exploratory Factor Analysis Results for POCAb Measure Using Principal Axis
Factoring .................................................................................................................... 76

Table 7 - Results of Principal Axis Factor Analysis with Oblique Rotation for Affect
Measure ..................................................................................................................... 79

Table 8 - lntercorrelations and lntemal Consistency Reliabilities for Affect Factors ...... 80

Table 9 - Exploratory Factor Analysis Results for Final Stereotype Threat Measure Using
Principal Axis Factoring ............................................................................................ 82

Table 10 - Means And Standard Deviations of Major Study Variables ............................ 86
Table 11 - Reliability Coefficients and lntercorrelations of Major Study Variables ........ 87

Table 12a — The Relationship Between Self-Efficacy and Feedback-Seeking. Controlling

for MPPS ................................................................................................................... 90
Table 12b - The Relationship Between POCA and Feedback-Seeking, Controlling for

MPPS ......................................................................................................................... 90
Table 13 - The Interactive Effect of MPPSa and Initial Self-Efficacy on POCAb ............ 91
Table 14 - Initial Fit Statistics by Dependent Variable ..................................................... 95
Table 15 - Fit Statistics For Second Iteration Of Models By Dependent Variable ........... 96
Table 16 - Fit Statistics For Third Iteration Of Models By Dependent Variable .............. 97

Table 17 - Personal Awareness of and Perception of Managers’ Awareness of Stereotype
................................................................................................................................. 1 l2

LIST OF FIGURES
Figure 1 - The 5010 status phenomenon from the perspective of the solo status literature I 7

Figure 2 - The solo status phenomenon from the perspective of the stereotype threat
literature .................................................................................................................... 35

Figure 3 - The solo status phenomenon from the perspective Heilman and her colleagues’
research ...................................................................................................................... 47

Figure 4 - Points oftheoretical convergence between the solo status and stereotype threat

literatures ................................................................................................................... 5 I
Figure 5 - Proposed model ................................................................................................ 56
Figure 6 - Content of e-mail message sent to participants in the non-solo condition ....... 73
Figure 7 - Sample structural model ................................................................................... 93

Figure 8 - Final path model for task difficulty choice (with standardized regression

weights) ................................................................................................................... 100
Figure 9 - Final path model for feedback-seeking (with standardized regression weights)
................................................................................................................................. 101
Figure 10 - Final path model for job performance (with standardized regression weights)
................................................................................................................................. 103
Figure l 1 - F inal path model for positive aroused affect (with standardized regression
weights) ................................................................................................................... 105
Figure 12 - Overall path model (with standardized regression weights) ........................ 107

vi

THE EFFECT OF BEARING SOLO STATUS WITHIN AN APPLICANT POOL ON
NEWLY HIRED FEMALES’ COGNITIONS. AFFECT. AND BEHAVIOR

An increasing number of women are gaining employment in jobs historically
occupied by men, such as managerial and professional positions (Amott & Matthaei.
1996; Burke, 1996; Federal Glass Ceiling Commission. 1995; Powell. Butterfield, &
Parent, 2002). In such cases, these women may be the only female in an otherwise male
applicant pool. This condition — being the only member of one’s social category in an
otherwise homogenous group — is referred to as bearing solo status (Lord & Saenz,
1985). The effect of bearing solo status within an applicant pool on solos’ thoughts,
feelings, and actions of newly hired females remains largely unexplored. Thus, the
purpose of this study was to develop and test a model of the relationship between bearing
solos status within an applicant pool and the cognitions, affect. and behavior of females
recently hired for a traditionally masculine job.

Three streams of research address this issue. First, early qualitative research by
sociologist Rosabeth Moss Kanter (1977a; 1977b) suggests women working in
predominately masculine environments experience greater visibility and scrutiny of their
work as well as confinement to stereotypic conceptions, ultimately resulting in lower job
performance. Empirical research since then has found general support for Kanter’s
contention that being a solo is a negative experience (e.g., Lord & Saenz, 1985; Niemann
& Dovidio, 1998; Sekaquaptewa & Thompson, 2002; cf, Fuegen & Biemat, 2002).
Unfortunately, the solo status literature remains plagued with what Lawrence (1997)

termed “black box” studies, which do not measure intervening process variables, and thus

there is currently no clear model of the solo status phenomenon (e.g.. Riordan. 2000:
Sekaquaptewa & Thompson, 2003; Young & James. 2001 ).

Second, Claude Steele and his colleagues recently introduced the concept of
stereotype threat — loosely defined as the pressure an individual faces when at risk of
confirming a negative, self-relevant group stereotype (Steele & Aronson. 1995). These
researchers have suggested that bearing solo status may prompt stereotype threat.
resulting in “underperformance” (e.g., Steele, Spencer. & Aronson. 2002). Steele and
Davies (2003) claim the basic tenant of this line of research —- stereotype threat
undermines test performance — “has been shown in multiple groups, on many tests. in
laboratory and real-life testing situations, in many countries, by many different
investigators, in published experiments that now approach 100 in number” (p. 312).
Unfortunately, the knowledge of the effects of stereotype threat may have been acquired
at understanding the phenomenon itself, as there are currently an abundance of features
suggested to characterize the threat (Wheeler & Petty, 2001) and a conspicuous absence
of any model that integrates them.

Finally, Madeline Heilman and her colleagues’ research (Heilman, 1980; Heilman
& Alcott, 2001; Heilman & Blader, 2001; see also Heilman & Haynes, in press)
addresses solo status in a selection context and thus is a helpful supplement to the
preceding two literatures despite the fact that it primarily focuses on others’ perception of
the solo (rather than on the solos themselves). Heilman and Blader’s (2001) results (see
also Heilman, 1980) suggest that females bearing solo status within a cohort of newly
hired individuals are perceived by co-workers to be beneficiaries of gender-based

preferential selection as well as less competent and less likely to succeed (relative to non-

solo females). Related, the results of Heilman and Alcott (2001) suggest that females
who are aware that their co-workers’ assume they were preferentially selected tend to
make more “timid, performance-limiting task decisions” (relative to those who assumed
they were selected based on merit), particularly if they had low initial self-confidence
about their ability to perform the task at hand (p. 574). In short. Heilman and her
colleagues’ research suggests the previously untested possibility that bearing solo status
within an applicant pool may prompt newly hired females to believe that others will view
them as preferentially selected, which may in turn negatively affect achievement.

Thus, to summarize, none of these three streams of research individually explains
the process by which bearing solo status within an applicant pool affects newly hired
females (including the presence of individual difference variables that may moderate the
relationship). However, the solo status and stereotype threat literatures collectively yield
points of theoretical convergence from which a framework can be developed. Thus, after
providing a review of both literatures, I develop such a framework and then use it (along
with Heilman’s research) to develop a specific and testable model of the relationship
between bearing solo status within an applicant pool and the cognitions, affect, and
behavior of females recently hired for a traditionally masculine job. In this way, I use the
general tentative framework to derive theoretically defensible predictions for the present
research question and to provide a greater understanding of the literatures from which I

drew in its development.

CONSTRAINTS AND CAVEATS

The purpose of the present study was to examine the effects of bearing solo status
within an applicant pool on the cognitions, affect, and behavior of females hired for a job
traditionally considered “masculine.” Thus, although I cite that fall outside of these
boundary conditions, the model that is eventually proposed applies only to situations that
fall within them.

References to the effects of “bearing solo status” are references to the effects of
“knowing that one bears solo status” (rather than the condition of solo status itself). This
caveat is not unique to the solo status literature, as it applies to most (if not all) studies of
solo status, but it is important to the present study because the demographic make-up of
one’s applicant pool is not necessarily apparent to the applicants themselves. This caveat
and the aforementioned boundary conditions collectively imply that the model that is
eventually proposed applies only to situations in which the relevant applicants are aware
that they bore solo status within their applicant pool.

Solo applicants may discover their 5010 status through a variety of methods.
Empirical research on the topic is lacking, but there is anecdotal evidence to support this
claim. For example, Ms. Sunalini Menon, a coffee quality control expert recently named
by her industry’s trade journal as one of the “People of the Year,” realized she bore solo
status within her applicant pool when she arrived for an interview with the Coffee Board
of India (Suryanarayanan, 2002). Similarly, Dr. Karen Vierow was aware that she bore
solo status within her applicant pool for an assistant professorship at Purdue University’s
School of Nuclear Engineering when this “distinction” made her the subject of a local

newspaper article (Jester, 2000).

References to females as “bearing lower social status” do not reflect the belief
that women are inferior but rather that stereotyping and preconceptions about women’s
abilities and suitability for “masculine” jobs are among the barriers to advancement for
professional and managerial women (Mattis, 1994; Morrison, 1992). This caveat is
consistent with the idea that professional and managerial women face a glass ceiling (a
subtle yet strong barrier that prevents women from moving up to senior levels of
management) that limits their advancement to top management in large organizations
(Burke, 1996, p. 205; Morrison & von Glinow, 1990; Morrison, White, & Van Velsor,
1987)

RELEVANT RESEARCH
Solo Status

As stated previously, Kanter’s (1977a; 1977b) seminal work laid the foundation
for current research on the experience of being a “token” or bearing “solo status” within
organizational contexts. Kanter defined tokenism in terms of numbers, with any social
group that represents less then 15% of the total group called “tokens.” Solo status is a
more specific case of tokenism and is defined as being the only member of one’s social
category in an otherwise homogenous group (Lord & Saenz, 1985). Whether a token or a
solo, Kanter suggests that salient minorities experience a greater sense of visibility and
thus these terms are sometimes used interchangeably. For example, Saenz and Lord
(1989) refer to the condition of solo status as a case of tokenism (e.g., he was a “token by
virtue of being the only man,” p. 699), while Sackett, DuBois, and Wiggens Noe (1991)

refer to both conditions under the broader term of tokenism without mention of “solo

status.” In the present paper, I am focusing on solo (versus token) status, but, given the
conceptual and terminological overlap, I also cite research on tokenism when relevant.
Solo status can occur within a variety of contexts and for a variety of people and
thus should not be understood as a chronic state or stigma (Crocker, Major, & Steele,
1998; Goffman, 1963). Kanter (1977a) examined the effects of being a salient minority
within an organization, a condition that is becoming increasingly rare as organizations
become larger and more diverse (academia being somewhat of an exception; see Fontaine
& Greenlee, 1993; Niemann & Dovidio, 1998). The majority of research on solo status
examines the effect of bearing solo status within a workgroup (i.e., a group of two or
more individuals who must interact cooperatively and adaptively in pursuit of shared,
valued objectives; Salas, Cannon-Bowers, & Blickensderfer, 1997, p. 305). This research
generally involves laboratory manipulations of the demographic composition of the
workgroup (e.g., Hyers & Swim, 1998; Saenz & Lord, 1989; Sekaquaptewa &
Thompson, 2002; 2003; Taps & Martin, 1990), although there are some studies that have
been conducted in a field setting (e.g., Sackett et al. 1991; Niemann & Dovidio, 1998).
Solo status can also occur within a selection context, defined here as the events
that occur during and after a selection decision is made (Gilliland, 1993). Bearing solo
status makes in a selection context makes one’s group membership (e.g., gender) more
salient and thus provides an external attribution for any outcomes received (Crocker &
Major, 1989; Weiner, Russell, & Lennan, 1979), including that of a selection decision
(Gilliland, 1993). Stated differently, bearing solo status in a selection context creates
uncertainty about the reason for hire (i.e. gender versus merit). I return to this point later

in a separate section when I review the small body of research on solo status within a

selection context (Heilman, 1980; Heilman & Blader, 2001; see also Craig & Feasel,
1998).

The consequences of solo status on the solos themselves have been investigated in
both fieldl and lab2 settings. In a prototypical field study, Niemann and Dovidio (1998)
examined how racial/ethnic solo status related to the job satisfaction of 425 White and
racial/ethnic-minority members of the American Psychological Association working in
psychology departments in academia. Regression analyses suggest that minority faculty
members bearing solo minority status within their departments felt less satisfied with
their job. Specifically, solo minorities had lower levels of job satisfaction than did non-
solo minorities (B = .19, t (156) = 2.48, p < .014). Moreover, a priori tests revealed the
same pattern of significant differences within all ethnic sub—groups (i.e., African
Americans, Asians, and Hispanics). Although recognizing the methodological limits of a
survey for establishing causality, the authors’ results suggest that perceived
distinctiveness associated with solo minority status (which I define and discuss in a
separate section) mediates the link between $010 status and job (dis)satisfaction.

In a prototypical lab study, Lord and Saenz (1985) led 48 male and female
undergraduates to believe that they were sharing their views on everyday topics with
three other participants (actually videotaped confederates), who were either all of the
participants’ own sex or all of the opposite sex. Some of the study participants were

asked to contribute to the dialogue, while others were asked to observe. The authors’

 

' e.g., Alexander & Thoits, I985; Garland & Price, 1977; Hamner, Kim, Baird, & Bigoness, 1974; Heikes,
1991; Kanter, 1977; Niemann & Dovidio, I998; Ott, I989; Sackett et al., 1991; Yoder & Aniakudo, 1997;
Yoder & Sinnett, 1985

2 Craig & Rand, I998; Crocker & McGraw, 1984; Cohen & Swim, 1995; Karakowsky & Siegel, 1999;
Lord & Saenz, 1985; Pollak & Niemann, I998; Saenz & Lord, 1989; Taylor, Fiske, Etcoff, & Ruderrnan,
1978

(using ANOVA-based analyses) found that solo participants who were asked to
contribute (as opposed to observe) had lower performance than non-solos (regardless of
sex) — i.e., they remembered fewer of the opinions expressed (both by themselves and by
others). In contrast, participants who were asked to act as observers remembered more of
what the solo (versus non—solo) participants said. The authors interpreted these results to
be a function of solos’ perceived need to be more aware of the potential of being
evaluated by others.
Consequences

As stated previously, existing research suggests that bearing solo status is a
negative experience (e.g., Crocker & McGraw, 1984; Floge & Merrill, 1986; Kanter,
1977a; 1977b; Lord & Saenz. 1985; Niemann & Dovidio, 1998; Sekaquaptewa &
Thompson, 2002; c.f., Fuegen & Biemat, 2002).

First, existing research suggests that solo status has a negative effect on solos’
cognitions. For example, it has been found to negatively affect male and female solos’
learning (Saenz & Lord, 1989), to make racial minority solos feel more representative of
their race (Pollak & Niemann, 1998), and to prompt male and females to have increased
expectations of being stereotyped (Cohen & Swim, 1995). Second, solo status has a
negative effect on solos’ affect. For example, solos report lower levels of job satisfaction.
This is true for racial minorities (Niemann & Dovidio, 1998) and for women (Kanter,
1977a). Additionally, both male and females solos express discomfort by reporting a
greater desire to change the composition of their group and to change to a different group
(Cohen & Swim, 1995). Finally, solo status has a negative effect on solos’ behavior. For

example, female solos exhibit lower levels of performance (relative to male counterparts,

Sackett, et al. 1991; Sekaquaptewa & Thompson, 2003; see also Alexander & Thoits,
1985; Lord & Saenz. 1985). Additionally. there is evidence to suggest that both male and
female solos exhibit lower levels of emergent leadership behavior when their gender is
incongruent with the perceived gender orientation of the group's task (Karakowsky &
Siegel, 1999).

Potential Moderators

There are three constructs that have been proposed as potential moderators of the
relationship between solo status and cognitions, affect, and behavior. Here, I describe
each construct along with the relevant empirical evidence. Later, I describe the theoretical
rationale that may underlie their suggested moderated effect.

The first potential moderator is social status, such that the negative effects of solo
status are stronger for those with lower social status (i.e., women, racial-ethnic
minorities, and those with low socio-economic status, Amott & Matthaei, I996). The
abundance of evidence demonstrating the absence of any negative effect for male solos
(e.g., Cohen & Swim, 1995; Sackett et al., 1991; Sekaquaptewa & Thompson, 2002;
Yoder, 1991; 1994), even in situations where they are the workplace minority (e.g., Floge
& Merrill, 1986; Heikes, 1991; Ott, 1989; c.f., Young & James, 2001) provides partial
support for this point. However, it is worth emphasizing that social status arises from the
context and is relative to the status of other group members, which may explain why
Sekaquaptewa and Thompson (2002) found that white female solos exhibited
performance decrements in the presence of the relatively higher-status white males but

not in the presence of the relatively lower-African American females.

The second potential moderator is initial .sel/Lcon/idem'c. such that solos status’
negative effects are more pronounced for individuals with low (versus high) initial self-
confidence for the task. To investigate this possibility. Cohen and Swim (1995)
manipulated the solo status (within a work group) and task confidence of participants and
found that many of the negative effect of solo status (e.g.. preference for a different
group) were stronger for women with low (versus high) confidence. Thus, this study
suggests that initial self-confidence may moderate the relationship between solo status
and cognitions, affect, and behavior.

The third potential moderator is what I will refer to as evaluative pressure. That
is, it is possible that solo status’ negative effects only exist in situations where one is
called on to demonstrate his/her ability and skills under the scrutiny of others
(Sekaquaptewa & Thompson. 2002, p. 705). To investigate this possibility,
Sekaquaptewa and Thompson (2002) conducted two lab studies similar to the one
conducted by Lord and Saenz (1985), but with separate learning and testing phases
(which represented low and high evaluative pressure, respectively). These two studies
collectively suggest that white female solos experience performance decrements in the
presence of a group comprised of white males but not in the presence of a group
comprised of African American females. Sekaquaptewa and Thompson (2002) interpret
their results to suggest “the impact of being different from the rest of the group is greatest
when it counts the most:” when one is under evaluative pressure (p. 705). .

To summarize, existing research suggests that the effects of solo status may be
ameliorated by high social status, high self-confidence, and low evaluative pressure, and,

conversely, exacerbated by low social status, low self-confidence, and high evaluative

10

pressure. Notably, these three constructs are likely related insofar as women and
minorities are more often called upon to demonstrate their skills and abilities under the
scrutiny of others due to unfavorable stereotypes about their capabilities. which
Sekaquaptewa and Thompson (2002, p. 705) suggest is the case. In this way. theoretical
rationale underlying each of these proposed moderators might be drawn from the
previously introduced construct of stereotype threat (e.g.. Steele & Aronson, 1995). That
is, solos with low social status are often the targets of negative stereotypes and thus can
find themselves in situations marked by evaluative pressure where they are at risk of
confirming these stereotypes (which is exacerbated in situations where they do not feel
confident about their own ability to perform). 1 return to this idea later in the paper.
Potential Mediators

As stated previously, the solo status literature (and, more broadly. the
demographic composition literature) is plagued with what Lawrence (1997) called “black
box” studies, which do not measure intervening process variables. The need to identify
and test mediating mechanisms has been echoed by others (e.g., Riordan, 2000;
Sekaquaptewa & Thompson, 2003; Young & James, 2001). To this end, I now review the
four potential mediators of the relationship between solo status and cognitions, affect. and
behavior that have been identified by existing research.

First, performance expectancies — defined as beliefs about how well one can
perform in a given situation3 — have been has been found to mediate the relationship
between solo status and behavior for females. Specifically, Sekaquaptewa and Thompson

(2003) suggest that female solos have lower performance expectancies (compared to

 

3 In the applied psychological literature, perceptions of one’s capability to complete a task are referred to as
one’s self-eflicacy (see Gist & Mitchell, 1992)

11

males and to non-solos) and that these lower expectancies translate into lower
performance. The authors point out that this idea is not new in that previous research has
demonstrated that female solos have lower performance expectancies than their non-solo
counterparts (e.g., Stangor, Carr, & Kiang, 1998) and that expectancies predict
performance (e.g., Eccles, 1994). However, Sekaquaptewa and Thompson (2003) were
the first to test this hypothesis in a mediational model. And, consistent with predictions.
the authors found that the effect of solo status on performance was partially mediated by
performance expectancies, suggesting that women who were expecting to bear solo status
developed lower expectancies about their upcoming performance compared to their male
counterparts, and that this led to lower levels of actual performance than that of non-solos
(p. 73). The authors point out that the fact that the indirect effect of performance
expectancies suggest that it may “therefore be but one of numerous factors that account
for the influence of solo status on men and women’s performance outcomes”
(Sekaquaptewa & Thompson, 2003, p. 73).

The next two potential mediators each represent a concern with how one is
perceived by others — i.e., solos’ evaluative concerns and stereotypic concerns.
Specifically, Saenz and Lord (1989, p. 698) suggest that solos “spend too much time
worrying about being evaluated” and thus suffer cognitive deficits, while Sekaquaptewa
and Thompson (2002, p. 705) suggest that solos’ “negative stereotypic expectations” —
i.e., the expectation that one will be negatively stereotyped by others — play a role in their
underperforrnance. These two constructs, which are different from performance

expectancies in that they reflect a concern for what others think (rather than a concern for

12

one’s own ability), have been suggested to mediate the relationship between solo status
and cognitions and behavior.

Saenz and Lord (1989; see also Lord & Saenz, 1985; Saenz, 1994) infer the
presence of evaluative concerns from solo participants’ differential performance in
“evaluator” and “evaluatee” conditions. That is, the authors suggest and find support for
the idea that, if solos spend too much time worrying about being evaluated, this problem
will be alleviated in situations where they are asked to evaluate others. Unfortunately, no
examination of solo status (of which I am aware) directly measures evaluative and test for
their presence in a mediational model.

Two studies (within the solo status literature") suggest that stereotypic concerns
may mediate the relationship between solo status and affect and behavior. First, Cohen
and Swim (1995) found that male and female solos expect to be stereotyped to a greater
degree than non-solos and that they expressed discomfort such that they report desire to
change the composition of the group. However, the authors did not test for stereotypic
concerns’ mediating presence. Moreover, the authors did not specify the content of the
stereotype hypothesized to drive the stereotypic concerns, although it likely involves the
idea that one’s social identity is considered less competent and/or less valued within a
given context. Second, Sekaquaptewa and Thompson (2002), in an examination of the
performance outcomes of African American solos, tested for the mediating presence of
stereotypic concerns, but the authors measured the construct by assessing the activation

of general (as opposed to context specific) racial stereotypes (e.g., African Americans are

 

’ There are other examinations of the mediating role of stereotypic concerns in the relationship between
solo status and achievement-related outcomes (e.g.. Stangor, et al., 1998). However, the authors of these
studies place themselves within the stereotype threat (as opposed to the solo status) literature and,
accordingly, will be reviewed later in the paper.

13

lazy). The authors note that assessing stereotype activation is difficult in achievement
setting and suggest, “negative stereotypic expectations may play a role, but more
sensitive assessments may be required to better investigate this role” (p. 705). I argue that
perhaps a more construct valid measure is required — i.e., one that actually assess the
stereotypic concerns relevant to the situation rather than the activation of general
stereotype. This idea is theoretically consistent with the stereotype threat literature,
which, as I will demonstrate later, emphasizes the perceived pressure of stereotype
confirmation over the actual activation of the stereotype itself.

Finally, as previously discussed, the results of Niemann and Dovidio (1998)
suggest that perceived distinctiveness mediates the role between solo status (within a
department) and affect (i.e., job satisfaction). The authors define perceived distinctiveness
as “the subjective experience of visibility, contrast, and role encapsulation and of being
viewed primarily in terms of category membership” and state that the construct “refers to
feelings of difference derived from visible differences between people” (p. 55). Niemann
and Dovidio (1998) argue that bearing solo status may produce negative consequences
for members of traditionally underrepresented and stigmatized groups by increasing
feelings of distinctiveness based on group membership, which can “increase the salience
of negative stereotypic expectancies (Goffman, 1963; Milliken & Martins, 1996; Steele,
1997; Steele & Aronson, 1995)” (p. 56).

This line of reasoning suggests that “stereotypic expectations” and “perceived
distinctiveness” are overlapping constructs. Niemann and Dovidio (1998) themselves
point out that the items used to measure perceived distinctiveness represent both “feelings

of distinctiveness” and “perceptions of stigmatization and inequitable treatment of

14

members of minority groups” (p. 67). Respective sample items include "I would feel

more comfortable at work if more of my colleagues were members of my ethnic group"

and "To be hired for faculty positions in my department, minority candidates must be

better qualified than White candidates". The authors even go so far as to say that

“although these dimensions are conceptually separable in the laboratory, they may be

closely linked in the experience of minorities in more naturalistic settings” (p. 67).
Summary

The solo status literature suggests that solo status negatively relates to favorable
(i.e., performance-facilitating) cognitions, behavior, and affect. It also suggests some
potential moderators and mediators of this relationship. Here, I review these ideas —
which have been presented in a fragmented fashion — and suggest how they may be
combined to form what I’m terrning “The solo status phenomenon from the perspective
of the solo status literature” (shown in Figure 1). This framework is admittedly imprecise
in it has never been explicitly suggested by any one solo status researcher (including
Kanter) but it bears the advantage of efficiently communicating the cumulative findings
of the solo status literature to date.

As shown in Figure 1, the solo status literature collectively suggests that solo
status’ negative effects occur through one or more of the following mechanisms:
performance expectancies, evaluative concerns, stereotypic concerns, and perceived
distinctiveness. As stated previously, evaluative concerns and stereotypic concerns have
not received empirical support in a mediational model (which is footnoted in the Figure
1). Moreover, there has been no discussion of the relationships amongst these four

constructs, making it difficult to determine how they may interplay or overlap in any

15

given context. As stated previously, the exact content of solos’ stereotypic concerns
likely varies by person and situation. However, I believe that females bearing solo status
within a selection context are concerned about the stereotype that they are less competent
and less likely to succeed as well as perhaps the idea that, if they are hired, others will
think they were preferentially selected on the basis of sex (versus merit). I return to both
of these suggestions later in the paper.

The solo status literature has yielded three potential moderators of the relationship
between solo status and cognitions, affect, and behavior: social status, initial self-
confidence, and evaluative pressure. Two of these three constructs — social status and
evaluative pressure — are effectively irrelevant in a traditional selection context such that
it is inherently a high evaluative pressure situation and solos are likely to be of lower
social status. However, the idea that female solos’ initial self-confidence also moderates
this relationship (with initially high self-confidence potentially acting to buffer against

solo status’ negative effects) is relevant across contexts.

16

.BBOE 32253:. a E tomasm :33qu 3280.. so: we; 623:8 £5. _

 

 

 

Eoom 033335 REE

 

 

 

 

 

 

_
_
2 mafia 2336 oocovccoofiom
_
_

 

 

 

rllu't IIIIIII l |||||||| ll IIIIIIIIIII Ola lllllllllllllllllllllllllll L

 

$538556

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3208...:—
._o_>m;op 25250 L 38% Bow
oaboflgm 4
.202:on
3:30:00
o>:m3m>m

 

 

 

moficfiooaxm

_
_
_
_
_
_
_
_
_
new quota . _ _
_
_
_
_
_
_
_
oozes—cote; "

 

 

 

.2338: 38% 28 2: Co 35333 2: 80¢ coaquconq 98% 28 oat—t . N 333%

17

Stereotype Threat

Claude Steele and his colleagues (Steele. 1997: Steele & Aronson, 1995) recently
introduced the construct of stereotype threat. loosely defined as the pressure an individual
faces when at risk of confirming a negative. self-relevant group stereotype (Steele &
Aronson, 1995). I say “loosely" because there is an absence of consensus as to the exact
definition of "stereotype threat” (even within Steele’s own work; See Table l). A precise
definition of a construct is critical, facilitating its operationalization and the development
of its nomological net and hence future research (Cronbach & Meehl, 1955). Thus, it
should be no surprise that there has been a proliferation of research proposing a variety of
potential mediators and moderators of the stereotype threat phenomenon with little to no
discussion of how they relate to one another (e.g., Steele et al., 2002). However, even
with its difficulties, the stereotype threat literature provides several theoretical insights
that are directly relevant to the current paper. Thus. further clarification of the construct is
necessary.

Stereotypes have been defined in a variety of ways (Greenwald & Banaji, 1995;
See Table 2). For the present paper, stereotypes are defined as a cognitive structure that
contains the perceiver's knowledge, beliefs, and expectancies about some human group
(Hamilton & Trolier, 1986, p. 133). Stereotype threat has been characterized as a
stereotype pertaining to one’s own group (i.e., self-stereotypes; Wheeler & Petty, 2001).
However, I argue that stereotype threat more accurately refers to a stereotype pertaining
to one’s own group that is believed to be held by others (e.g., Steele & Aronson, 1995)
and thus differs from a self-stereotype in that it does necessitate that the target internalize

the stereotype (e.g., Crocker, et al., 1998).

18

 

Sam .5 Emcee b30to>fim£ 2.83 85
@2668 ow £me 28 85 co 09008on 05 .6 £58 E v83: 8 Bowen“ on :3 use 88 8:8
was—38 of m_ 39:: 3.03%: .mEZE E 28 85135 an 8 8323 Mono .8 852985
3008 cemcoc<
as .80:on .03on cm 3 .23st JESSE émcofioa $8802 mo tam mm 25 65 macaw m Bone 3502on gunmen m. 555

$3. a: BBS sabesxea Ecanmeocm £5 :3 03 .83 E50 Maze 5 55 32ch

ES .2050 Co 8.3 2: E :ocmwcofifimzoiom m we @36sz 808 0&6»:on 2t 888 a 9 E828

comcoc<AWMWW2m 35 358$ Mono (8 En do Soc 28 wcfiica 65 8on 09302on a :03 We mosaics 2t . W
8% .3 0&8»:on gamma: :2: mo 3:8 5 @8on co 58 on 2:8 one 5.: magmas 5%
0:0 3:33 6st _ $55625 mm mat. .cocmzzm 3:5qu m E .2828 8 cinemas fl mwsofin
£00 3 293m use 5053 9 95cm x Sena 08002on “Stage a 35 moNEwoooc 0:0 5:3 £380 83% subspace”

Am as 39:: suggest. we 8 have 95 monomeomxo

88 c 5.30
as .20on £8:on ca dzocw mbco Sons 89308on 3568 no woman 82:35.“ 82¢ 28 80:3 mouse? a 5 mafia

 

229a :oEcmoQ

 

Becafi mmbomxmumKo €225wa mEEom

H 2an

33: 902%

$8: 32: a Na:

3%:

828% Q 556::

Aoom .3 bomeo 85 9 w:mw:o_os 85:53 2: =: 33:30: :85:

95:55 2: 65 Eovafi 2: 836 8 E2053 fl 35:02:88 :3 n3:83: Rotowouwo <
2M: 3 6:88 @3630 :5: “we game: So 80.: 33mm: :5: 4:823:

9 £588: a 6m: 05 8 2:: b3 $5880 :35» 625895 :86 m mm 3&8:va <
92 .3 anew 53:3: 08%

Son: 8683035 :5 $023 ,omwBBQE @5383: 2: 25:8 :2: 23835 Q>E:woo <

A? .3 533:: :m 3 womcmig on 8 33238 mm 52:3 £09533

 

C no C :8:ow :3: a $5538 daew 3:50 B: Son: 25:: :o:m~=80:ow m mm @9902on 3:50 :<
33:
«com ED av 80:39: 8: .3 2:8: mo 95% m mo 339:3: 6:85: 05 Son: @320: mo :8 <
Gm: C “.597: 22 .8 Eowofio m 53> 33688 (Home: @8903me :m mm 0:500:03 <
203.5 55:5va

 

«33 #855 a: 3:32:30 59% mmbomgmfi :X: m::2.E§mQ Susan

N 2an

20

Accordingly, to distinguish between the stereotypes that one holds about one's
self and the stereotypes that one believes that others hold about one’s self. 1 will refer to
the latter with the recently proposed construct of meta-stereotypes. Meta-stereotfinas' are
an extension of the literature on meta-perceptions (defined as beliefs about another's
impressions of one’s self; e.g., Kenny & DePaulo, 1993; Vorauer & Claude, 1998:
Vorauer & Miller, 1997) and refer to one’s beliefs regarding the stereotype that others
hold about one’s own group (Vorauer, Main, & O'Connell, 1998). Thus, the content of
meta-stereotypes reflect the assumptions that people hold about the stereotypes others
have about the social group(s) to which they belong. Vorauer, Hunter, Main, and Roy
(2000) empirically examined meta-stereotypes in the context of social interaction and
found that the potential for evaluation by (as opposed to mere exposure to) another in a

valued domain5

is a precondition for meta-stereotype activation. This point is an
important one, as it suggests that “stereotype threat” and “the activation of an unfavorable
meta-stereotype” are tapping a similar phenomenon.

A conventional definition of stereotype threat is “the realization that one’s
performance on a particular task might confirm a negative stereotype about ones’ group"
(Bosson, Haymovitz, & Pinel, 2003, p. 247; Steele & Aronson, 1995). However, because
the potential for (in this case, unfavorable) evaluation by someone else in a valued
domain is a precondition for meta-stereotype activation (Vorauer, et al., 2000), I argue
that a more precise and parsimonious definition of the construct is “the activation of an

unfavorable meta-stereotype.” The problem with this definition is that it is not well-

supported by the existing stereotype threat literature, which focuses on self- (vs. meta-)

 

5 Readers familiar with the stereotype threat literature might wonder how this relates to domain
identification, a point I return to later

21

stereotypes. Given that the purpose of this paper is not to develop new constructs (or new
labels for existing constructs), I will use the aforementioned, existing definition to guide
further discussion.

Necessary conditions for stereotype threat to occur are that a negative stereotype
must exist (Steele, 1997), the target must be aware of the stereotype, and the target must
believe that the relevant entity (i.e., the one making the evaluation) is aware of the
stereotype.

The Present Review: A Cautionary Note

While the present paper focuses on females applying for jobs that have
historically been occupied by males, Steele and Aronson’s ( 1995) seminal research
focused on African Americans taking cognitively loaded measures of knowledge, skills.
and abilities. Such measures are among the most valid predictors of performance but bear
the undesirable consequence of yielding racial sub-group differences (with an average
majority-minority difference of .83 standard deviations; Schmitt, Clause, & Pulakos,
1996). These sub-group differences are meaningful because scores on such tests are
frequently the basis for high-stakes decisions in a variety of contexts (e.g., employment,
academic admission, licensure, and certification; see Sackett, Schmitt, Ellingson, &
Kabin, 2001). Thus, Steele and Aronson’s (1995) finding that the African American—
White difference in standardized test performance was larger when participants were in
the stereotype threat condition than when they were in the non-threat condition is an
important one, as it suggests that “the instructional set under which examinees approach a

test can affect test results” (Sackett, Schmitt, Ellingson, & Kabin, 2001, p. 310).

22

Unfortunately, many mainstream news stories (e. g., the Public Broadcast Service
[PBS] show Frontline; Chandler, 1999) and introductory psychology textbooks have
erroneously interpreted Steele and Aronson’s (1995) research to mean that eliminating
stereotype threat will eliminate the African American—White difference in standardized
test performance (see Sackett et al., 2001; Sackett, 2003; Sackett, Hardison, & Cullen,
2004). The source of the misunderstanding is Steele and Aronson’s (1995) decision to
statistically control for differences in students’ prior standardized test performance.
Therefore, the practical implications drawn from the Steele and Aronson (1995) data
have been called into question. Specifically, although there was some observed
stereotype threat effect on performance over-and-above than that predicted by
standardized cognitive ability, reducing the threat effect and ignoring ability will have no
impact on test performance. Since racial difference on cognitive tests is a well-
established finding and since ability tests predict the kinds of performance that Steele and
Aronson (1995) used in their study, one cannot infer that removing the treat will remove
difference in performance between the groups.

Although Sackett and his colleagues (Sackett, 2003; Sackett et al., 2001; 2004)
call into question implications drawn from the stereotype threat literature for reducing
race differences in performance, they do not invalidate the relevance of the stereotype
threat literature for issues regarding stereotype threat based on gender. This is because on
the kinds of jobs or tasks relevant to the present study there is no a priori reason to expect
ability differences between men and women. Or, put another way, a priori differences for
which Steele and Aronson (1995) controlled through statistical means are assumed not to

exist. Whether they do or do not is an empirical question in any one study and can be

23

investigated, but there is no reason to expect that difference. Assuming no difference
occurs, then the fact that Steele and Aronson observe a stereotype threat effect with
equated racial groups means that one too should expect an effect in gender groups with
equal ability.

Antecedents

What prompts one to experience stereotype threat? Unfortunately, there is not a
“definitive set of such cues” (Steele et al., 2002, p. 419) and thus researchers have
manipulated the construct in a variety of ways (see Table 3). The focus of the present
paper is solo status, and there is laboratory and field evidence to suggest that bearing solo
status can prompt one to experience stereotype threat. Stangor et al. (1998) prompted
stereotype threat for female (but not male) participants in a laboratory study by
manipulating anticipated solo status. Unfortunately, the authors did not directly measure
stereotype threat, but rather inferred its presence from its “effects” (e.g. decreased
performance expectancies). The content of the threatening stereotype was that the spatial-
abilities task (that they were about to perform) was one in which men were known to
perform better than women. That is, the stereotype suggested that females were
considered less competent and less likely to succeed (relative to males).

More recently, Roberson and her colleagues (2003) examined African American
professionals and found a significant correlation (r=.16, p<.05) between bearing solo
status within their department and reported experience of stereotype threat, as measured
with items such as “Other people expect me to perform poorly because of my race.”
Thus, the content of the threatening stereotype concerns the idea that Afi‘ican Americans

are less capable of performing the job (than Whites).

24

Table 3

Sample ()perationalizations ()fStereatype Threat

 

Operationalization

Study

 

Asked participants to indicate race

Presented inherently threatening subject
matter

Made participants aware of unfavorable
stereotype

Placed participants in situation where
they bore solo status

Told participants that members of their
group usually performed worse on
this task

Told participants that task was

diagnostic of their ability

Used male (vs. female or neutral)
character in math problem

Exposed participants to stereotypic
commercials

Measured participants’ perceived

stereotype threat

Steele & Aronson (1995, Study 4)

Osborne (2001), Schmader, (2002), Quinn &
Spencer (2001)

Aronson, et al. (1999), Leyens, Desert,
Croizet, & Darcis (2000)

Carr, Stangor, & Kiang, (1998), lnzlicht &
Ben-Zeev (2000)

Carr, et al. (1998, Study 1), O'Brien &
Crandall (2003), Spencer, Steele, & Quinn,
1999

Gonzales, Blanton, & Williams (2002),
McKay, Doverspike, Bowen-Hilton, &
Martin (2002), Steele & Aronson (1995,
Studies 1, 2, & 3)

Walsh, Hickey, & Duffy (1999, Study 1)

Davies, Spencer, Quinn, & Gerhardstein
(2002)
Roberson, Deitch, Brief, & Block (2003).

Steele & Aronson, (1995, Study 4)

25

For female solos applying for a traditionally masculine job (e.g., managerial and
professional positions. Amott & Matthaei, 1996; Burke. 1996; Federal Glass Ceiling
Commission. 1995: Powell. Butterfield. & Parent, 2002), I believe the relevant meta-
stereotype relates to their comparatively lower levels of competence and likelihood of
success in the workplace (versus men). While there is no study within the stereotype
threat literature that addresses stereotypic conceptions of women in the workplace, those
that address stereotypic conceptions of women in academic settings (e.g., Ambady, Paik,
J. Steele, Owen-Smith, & Mitchell, 2003; Brown & Pinel, 2003; Ford, Ferguson, Brooks,
& Hagadone, 2004; lnzlicht & Ben-Zeev, 2000; McIntyre, Paulson, & Lord, 2003;
Pronin, Steele, & Ross, 2003; Quinn & Spencer, 2001; Spencer, Steele, & Quinn, 1999;
Walsh. Hickey, & Duffy, 1999) follow the theme suggested here.

A considerable amount of evidence in other literatures supports the existence of
unfavorable stereotypes about women in the workplace (e. g., Crocker et al., 1998; Deaux
& LaFrance, 1998; Dipboye, 1985; Fiske, 1998; Heilman, 1983; 2001; Martel], Parker,
Emrich, & Crawford, 1998; Murrell & James, 2001). Specifically, existing research
suggests that women are stereotypically considered less competent and less likely to
succeed in traditionally masculine jobs (see Dipboye, 1985; Heilman, 1983; Heilman &
Haynes, in press), that they can become aware of this stereotype and its potential role in
others’ evaluation of them (Heilman & Alcott, 2001), and that this awareness can
ironically interfere with their performance. This interference may occur either via others
reactions’ towards them (e.g., Heilman, Wallen, Fuchs, & Tampkins, in press) or via
something akin to Merton’s (1948) notion of a self-fulfilling prophecy (e.g., Heilman &

Alcott, 2001).

26

( 'ansequem'es

Although stereotype threat is hypothesized to be relevant for a variety of
outcomes, most of the existing research has examined the effects of stereotype threat on
standardized test performance (Steele et a1, 2002. p. 391 ). Thus, the basic finding within
the stereotype threat literature (based on research-to-date) is that experiencing stereotype
threat negatively relates to standardized test perfonnance. More specifically, current
research suggests that those for whom unfavorable meta-stereotypes about performance
have been activated (e.g.. by bearing solo status) display lower test performance than
those for whom such a meta-stereotype has not been activated or does not exist. And, as
stated previously, Steele and Davies (2003) state that this effect “has been shown in
multiple groups, on many tests, in laboratory and real-life testing situations, in many
countries, by many different investigators, in published experiments that now approach
100 in number” (p. 312).

Earlier, I identified two studies that establish a link between solo status and
stereotype threat (Roberson et al., 2003; Stangor et al., 1998). Here, I discuss what these
studies found to be the consequences of stereotype threat. In addition to prompting
individuals to experience stereotype threat by manipulating anticipated solo status,
Stangor, et al. (1998) manipulated the expectations of success of participants and found
that participants who were given reason to have high expectations of success (via
feedback) reported higher performance expectancies, but only in the non-threat (i.e., non-
solo) conditions. This finding led the authors to conclude that stereotype threat (occurring
as a result of anticipated solo status) negatively impacts performance expectancies and

that these expectations are strong enough to overwhelm high expectations of success.

27

Roberson et al. (2003) examined the relationship threat between stereotype threat
(occurring as a result of bearing solo status) and various feedback-seeking behaviors of
African American professionals — namely. feedback seeking. feedback monitoring, and
feedback acceptance. The authors found that the experience of stereotype threat was
associated with increased use ofa monitoring strategy for seeking feedback and a greater
degree of feedback discounting (i.e.. dismissing feedback and/or viewing it as irrelevant).
The authors also note that feedback monitoring (vs. feedback seeking) is less useful than
direct inquiry for performance improvements and, coupled with discounting, may
“indicate pathways through which job performance is negatively affected” (p. 186).
Additionally, Roberson et al. suggest that these findings are consistent with those of the
solo status literature, which suggest that that solos are likely to have increased evaluative
concerns (e.g., Saenz, 1994).

There have also been two interesting hybrid studies that examined the dual-effects
of solo status and stereotype threat manipulations. lnzlicht and Ben-Zeev (2000, Study 1)
asked female undergraduates to take either a multiple-choice test comprised of either
math (i.e., threatening) or verbal (i.e., non-threatening) items from the Graduate Record
Exam (GRE) along with two other participants (actually confederates), who were either
of the participants’ own sex or of the opposite sex. Participants were informed that their
performance would be orally shared with the other participants. Thus, the experiment
used a 2 (solo/non-solo status) x 2 (stereotype threat/no threat condition) between
subjects design. The key dependent variable was test performance. The authors also
conducted a second study (lnzlicht & Ben—Zeev, 2000, Study 2), which was identical

except that (a) they manipulated the solo status of both male and female participants and

28

(b) they only used the math (and not the verbal) test. The results of the two studies
suggest that solo females experience performance deficits only on the math (but not the
verbal) test, whereas males perform equally well on the math test regardless of solo
status. Further investigation showed that females’ deficits are proportional to the number
of males in their group (i.e., solo females perfonned worse on the math test than non-solo
females).

In another hybrid study, Sekaquaptewa and Thompson (2003) used a procedure
similar to the one employed by Lord and Saenz (1985; see above) where they led male
and female undergraduates to believe that they were discussing information with three
other participants (actually videotaped confederates), who were either all of the
participants’ own sex or all of the opposite sex. Math information was presented in either
a threatening or non-threatening way, using different presentation methods (i.e., numbers
vs. terminology) and variable stereotype applicability (e.g., “Although gender differences
in test performance have been reported using traditional math materials, previous testing
has shown that men and women perform equally well on this type of material”; See
Spencer et a1, 1999). Thus, the experiment used a 2 (solo/non-solo status) x 2 (stereotype
threat/no threat condition) x 2 (participant gender) between subjects design. The key
dependent variable was performance (i.e., participants' answers to oral exam questions).
The authors found that “Both solo status and [the] stereotype threat [manipulation]
negatively influenced the performance of women but not men” and that “these factors
had an additive effect on women's performance: performance was lowest when both
factors were present and highest when both factors were absent” (Sekaquaptewa &

Thompson, 2003, pp. 72-73).

29

To summarize, stereotype threat (occurring as a result of bearing solo status) has
been shown — at least in laboratory settings — to negatively affect females’ performance
expectancies and to increase African Americans’ feedback monitoring and feedback
discounting. Additionally, stereotype threat (occurring as a result of other manipulations)
and solo status have been shown to have an additive and negative effect on females’
performance.

Potential Moderators

Steele and his colleagues (2002) review what they call the “most documented
moderators” of the relationship between stereotype threat and standardized test
performance: three situational factors (task difliculty, test diagnosticity, and stereotype
relevance) and three individual difference factors (domain identification, group
identification, and stigma consciousness) (p. 391). Definitions of each of these constructs
(as well as sample operationalizations) are provided in Table 4. Additional, more
recently examined individual difference variables include testosterone levels (Josephs,
Newman, Brown, & Beer, 2003) and coping sense of humor (Ford, Ferguson. Brooks. &
Hagadone, 2004).

The situational factors are theoretically irrelevant to the present paper in that they
are assumed present in a selection context (i.e., as the events that occur during and after a
selection decision is made; Gilliland, 1993). That is, the presence of a testing situation
implies that the relevant task is both difficult and perceived to be sensitive enough to

yield evaluative information and thus one that evokes stereotype threat (see Sackett,

2003; Steele & Davies, 2003).

30

800: .355 30:8 .3 “00930083 0: 9 30098 0:0 :0_:>» 8 3&0: 0:._t 5.05:0.8300 35$»:
3:302 :05 8 60:88: 095 :0u:0w .8380.“ :03? 8 E038 0:: 00380:: .moom £00235 ..w.0v

00:00:: W: 0930003 03:03: 0:: :0::>> 9 95% 308 0:: :33 8:982 0:0 :02? 8 8&0: 0: H EQQSSESS :30ng
005:0. 0:023 :05 :o Eowfifioo 00.3 5:03.28 003080: Maser/6:: :033 2 E030 0:: 005308

doc: is :0 .0:8m ..w.0v 5.827.200 0.0:: :8 0050:38: mug—o: 0058820: :033 3 08000: 0:.H ESSeSE03 SeEQQ

 

80:80:05 13220::

 

3:08 an $88 9 00930806 08 28:3. 3:: ESNEBENQ 0:9: 0::3 29 .002 in :0 :Om:o:< swov 0:0:
:0 00:08:80: 0:: 2 “550—0: 0008 E 5:30 99.5% 0 50:0 0930003 03:30: 0 :033 9 00.:w0w 0: H ”00:e>0~0- 35030:,0
03:50 go .oumo:w06-:o:. :o .0cmo:wm:o. 83 0:9 9 50:0 003 >05 30: 0:: 8:: 0:890:th 22 :0::0 .33 68:92.
0% 200% ..w.0v 20,6: 3:50 3800 $305852: mo 03300 00 08:08:: m: :08 0:: :033 8 0050: 0: H ”bwotwosmei Form
$238050”: 5 50: 80.33 mac 0:: :5 mac 50:00u 05 :0 8:08

:68 0:: 3033800 .000: J: :0 005%V swov w:_w:0:m:0 b0>£:moo m: :05 0:: :03? 9 00:3: 0:.H 530§Q «warm

 

35:00.: :9:0E:o:>:m

 

Q3431 .maméam «we. :3 :0 030% 59% 00::220Kx0k tag :0 339$ 0Nb00§fik0 80% 0%.? assumes: BESBR

v 030,—.

31

Existing research on individual difference variables suggests that one is more
likely to experience stereotype threat when he or she defines himself or herself in terms
of the task (Pronin et al., 2003), is conscious of the relevant stigma (Brown & Pinel,
2003), identifies with the group (Schmader, 20002), and has high testosterone levels
(Josephs et al., 2003). Conversely, one is less likely to experience stereotype threat when
she has a coping sense of humor (Ford et al., 2001). While each of these findings
highlight potentially fruitful areas of future research, they were published after the
hypothesis development phase of the present paper and thus will not be discussed further.

Potential Mediators

What are the mechanisms through which stereotype threat affects performance
(and related cognitions, affect, and behavior)? This remains an open question. Although
there has been some discussion of primary appraisal mediators (i.e., cognitions about the
threat itself), most research-to-date focuses on secondary appraisal mediators (i.e.,
cognitions about ability to cope with the threat). In fact, there is only published study
within the stereotype threat literature that assesses a primary appraisal variable (i.e.,
stereotype activation, Steele & Aronson, 1995, Study 3) and this study did not measure
its resulting effect (e.g., on performance), leaving it unclear whether or not it plays a
mediating role.

Two secondary appraisal mediators have received published empirical attention:
performance expectancies (Spencer, et al., 1999; Cadinu, Maass, Frigerio, Impagliazzo,
& Latinotti, 2003) and evaluative concerns (i.e., evaluation apprehension, anxiety;
Leyens, Desert, Croizet, & Darcis, 2000; Osborne, 2001; Spencer et al., 1999). There is

some support for the idea that performance expectancies mediate the relationship

32

between stereotype threat and test performance. Spencer et al. (1999) measured female
participants’ performance expectancies, but “nothing happened” (Steele et a1, 2002, p.
399). That is, the stereotype threat manipulation did not affect females’ performance
expectancies, and their performance expectancies did not relate to their actual
performance. However, more recently, Cadinu et al. (2003) found some support for the
mediating presence of performance expectancies. In Experiment 1, the authors found that
performance expectancies partially mediated the relationship between stereotype threat
and the math performance of female students for whom math was important. In
Experiment 2, they found performance expectancies partially mediated the relationship
between stereotype threat and the verbal performance of Afi'ican Americans living in
Italy. However, these results have a little twist such that the mediating effect was only
present for participants for whom the minority category membership (Black) was made
salient and not for those for whom the majority category membership (American) was
made salient. In the words of the authors, the reason for this specific finding is “not clear”
(p. 283) but they suggest “caution is needed when extending a general stereotype threat
model to high status group memberships” (p. 284).

There is some support for the suggestion that evaluative concerns act to mediate
the relationship between stereotype threat and test performance. The two laboratory
investigations yielded null results. Specifically, Spencer et a1. (1999) found that
stereotype threat had a negative on participants’ self-reported anxiety levels and on their
performance, but analyses did not reveal that anxiety played a mediating role. And
Leyens et al. (2000) found no evidence to suggest that anxiety played a mediating role in

the relationship between stereotype threat and males’ performance on an affective

33

processing test. However, Osborne (2001), in a field study, found that anxiety partially
mediated the relationship between stereotype threat (which was presumed present as a
function of test content) and Afi'ican Americans’ academic performance and women’s
math performance.

Steele et al. (2002) conclude their review and critique of this body of research
with what they admit is “not a fully satisfying resolution,” which is that, “depending on
the sensitivity of measures used, different multiples of these [mediating] processes will
be found to mediate different occurrences of this threat in the laboratory. And in real life,
multiple and varied mediation of this threat will almost certainly be the rule, not the
exception” (pp. 406-407). So, as stated previously, the mechanisms through which
stereotype threat affects test performance remain unspecified. However, it seems that
performance expectancies and evaluative concerns may play a role.

Summary

Here, I describe the solo status phenomenon from the perspective of the
stereotype threat literature (represented in Figure 2). As before, Steele and his colleagues
have not advanced this framework, but it represents how their theory and research may

apply to the present paper.

34

 

 

3508:0330...
00:080.:0: :00?

T

 

 

o
0
o
.
o
.
.

.l.l.l.l.l.l.l.l.l.._

 

2:00:00
036203:

 

 

86530098
00:85:20:

 

 

 

_.I.I. IIIII . IIIIIIII _

83:80:: 38:: 09:80:05 0:: .:0 03:00:80: 0:: 68,: 55880:: 080:0 20m 0: :t .N 0805‘

 

 

 

88:: 0%:080:m

 

 

 

mam;
20m

 

 

 

35

F irst, the stereotype threat literature suggests that bearing solo status prompts solos to experience
stereotype threat. Specifically. for females applying for positions historically occupied by men,

bearing solo status within their applicant pool activates an unfavorable meta-stereotype, the
content of which likely suggests that they are less competent and less likely to succeed in
traditionally masculine jobs. Second, this literature suggests that stereotype threat
(occurring as a result of bearing solo status) has negative consequences for the perceiver
(e.g., lower performance). Finally although this literature has yet to specify the exact
mechanisms through which these negative consequences occur, it provides tentative
support for the idea that solos’ performance expectancies and evaluative concerns may
each play a role.
Solo Status within a Selection Context

As stated previously, there is no published examination of the effects of bearing
solo status within a selection context on the solos themselves. Stated differently, there is
no published examination of the effects of bearing solo status within an applicant pool or
within a newly hired cohort on the solos themselves. And, more specifically, there is no
published examination of the effects of bearing solo status within an applicant pool on the
cognitions, affect, and/or behavior of newly hired females. Thus, to develop hypotheses
for the present paper, I draw from the three studies conducted by Heilman and her
colleagues (Heilman, 1980; Heilman & Alcott, 2001; Heilman & Blader, 2001) that I
argue collectively speak to this issue. It is important to note that only one of the three
studies includes a manipulation of the solo status of individuals within an applicant pool
(Heilman, 1980) and none of them examines the effect of bearing solo status within an
applicant pool on the solos themselves. Thus, it is necessary to explain how these studies

relate to the present paper.

36

The first study I review (Heilman. 1980) manipulates solo status within an
applicant pool and then examines the effects on others' personnel decisions. The results
of this study suggest that solo status activates detrimental sex stereotypes in the minds of
others, which I will use to argue that bearing solo status actives similar meta-stereotypes
in the minds of the solos themselves.

The second study (Heilman & Blader, 2001) manipulates solo status within a
newly hired cohort and then examines the effects on others’ perceptions of the solos (with
respect to reason for hire and level of competence). Thus. there are two features of
Heilman and Blader’s (2001) research that are different from the present study. The first
is an issue of context. The authors” focus is on bearing solo status within a cohort of
newly hired individuals, while the present focus is on bearing solo status within an
applicant pool (and then being hired for the job). As stated previously, some may argue
that the “why” of solo status in these two contexts is different such that solo status within
a newly hired cohort may be attributed to someone’s purposeful decision, while solo
status within an applicant pool may be attributed to chance. However, as stated
previously, the situations are similar once the solo applicant has been hired in that
bearing solo status in both contexts creates ambiguity about the reason for hire. Thus, I
am arguing that it makes sense to draw from Heilman and Blader’s (2001) work in
developing hypotheses for the present paper.

The second feature of Heilman and Blader’s (2001) research that is different from
the present study in that the authors focus on the others’ reaction to the solo, while the
present study focuses on the reactions of the solos themselves. However, as I will argue

later, Heilman and Blader’s (2001) results — which suggest that solo status actives the

37

perception of preferential selection in the minds of others — are relevant to the present
paper in that they lend support to the idea that solo status activates a similar meta-
perception in the minds of the solos themselves.

The third study I review (Heilman & Alcott, 2001) does not involve solo status at
all but rather manipulates women’s beliefs about whether or not others view them as
preferentially selected. Later, after reviewing the results of Heilman and Blader (2001), I
explain the relevance of Heilman and Alcott’s (2001) findings to the present study in
more detail. For now, I will simply point out that these findings connect to the present
research in that they focus on what I see as a consequence of bearing solo status within an
applicant pool for the solos themselves (the meta-perception of preferential selection).

In sum, none of the three studies reviewed here individually speaks to the effect
of bearing solo status within an applicant pool on the solos themselves. Rather, the results
of Heilman and Blader (2001; see also Heilman, 1980) suggest that women bearing solo
status within an applicant pool are subject to the assumption that they are less competent
and are beneficiaries of preferential selection, and the results of Heilman and Alcott
(2001) suggest that women who have been made aware that others view them as
beneficiaries infer decreased competence expectations and are vulnerable to lower
performance (relative to those not made aware that others view them as beneficiaries;
Heilman & Alcott, 2001). Later, I consider this pattern of findings in conjunction with
the solo status and stereotype threat literatures (which have examined the effects of
bearing solo status on the solos themselves, but not in a selection context), and, based on

this, suggest an implied but not yet untested model of the relationship between bearing

38

solo status within an applicant pool on newly hired females’ cognitions. affect. and
behavior.
Heilman (1980)

Heilman (1980) assessed the extent to which the sex composition of an applicant
pool influences others’ personnel selection decisions. She suggests. “increased
proportional representation of women in the applicant pool [minimizes] the use of
detrimental sex stereotypes and consequently [reduces] discriminatory selection
decisions” (p. 388). To guide further discussion, I present this idea as two implied
hypotheses: Solo (or minority) status prompts others to make unfair (unfavorable)
selection decisions, and this relationship is mediated by the use of the stereotype that
women are less capable of performing the job (e.g., because they are less ambitious and
more irrational) and thus less likely to succeed.

To empirically test these implied hypotheses. Heilman (I980) asked masters of
business administration students (half male and half female) to evaluate an individual’s
employment application. Specifically, she asked participants to assess the applicant’s
qualifications and potential as well as to provide a hiring recommendation. Heilman
(I980) assessed the use of sex stereotypes by asking participants to rate to applicants
along four adjective dimensions found to be associated with different widely shared
perceptions of male and female work orientations (e.g., Broverman, Vogel, Broverman,
Clarkson, & Rosenkrantz, 1972). They were: ambitious-unambitious; emotional-rational;
decisive-indecisive; and tough-soft.

The position for which the applicant was applying was “traditionally masculine:”

a managerial position in a large insurance company (Heilman, 1980, p. 393). Participants

39

were also given information about seven other individuals in the applicant pool (to use as
points of comparison) and led to believe that zero, one, two, three, or all of these other
applicants were female. Thus, the design was a 2 x 5 factorial, with the independent
variables being Subject Sex and Sex Composition of the Applicant Pool.

The first hypothesis was supported. Solo female applicants were judged less
favorably their non-solo counterparts. The second hypothesis also received at least partial
support. Namely, although Heilman (1980) states, “the idea that [stereotypic] impressions
mediate selection evaluations was supported by our findings” (p. 393), I contend that the
second hypothesis was only partially supported because no mediational analyses were
conducted. Rather, covariance analyses using the composite adjective score as a covariate
were conducted on each of the three selection decision variables. Results indicated that
when the adjective score was factored out, no effects involving the Sex Composition
variable were found to be significant. Thus, while these results are consistent with the
idea that stereotype use mediates the relationship between solo status and selection
decisions, these analyses do not explicitly test mediation.

In short, Heilman (I980) demonstrates that bearing solo status within an applicant
pool (for a traditionally masculine job) negatively affects selection decisions and suggests
that this relationship might be mediated by the use of stereotypes, such that solo status
makes sex more salient and in turn makes male and female evaluators more likely to use
detrimental sex stereotypes.

Heilman and Blader (2001)
Heilman and Blader (2001) assessed co-workers’ reactions to females bearing

solo status in a cohort of newly hired individuals. The authors hypothesized that being

40

exposed to a solo female in an otherwise male cohort would prompt co-workers to
perceive the solo as preferentially selected — i.e., hired (at least in part) on the basis of her
sex (versus her merit). The authors further hypothesized that this effect would not occur
for non-solo females or for males (whether they were solo or non-solo).

Although Heilman and Blader (2001) did not set out to empirically examine the
mechanisms underlying the proposed effect (i.e., that being exposed to a solo female in
an otherwise male cohort of new hires would prompt co-workers to perceive the solo as
preferentially selected), their basic reasoning is that seeing something “out of the
ordinary” prompts a need for explanation that is most easily fulfilled by assuming that the
solo was preferentially selected. More specifically, the authors claim that assuming that
one was preferentially selected “requires far less cognitive exertion than modifying group
stereotypes or devising new subgroup categories to account for the unexpected
occurrence” (p. 181). Moreover, the authors state that this assumption is “cognitively
efficient and, as a consequence, it predominates (Fiske & Taylor, 1991, pp. 13, 380—
381).” In this way, Heilman and Blader (2001) hint at the possibility that the assumption
of preferential selection somehow involves a stereotype, a point I will return to later.

To test the link between solo status (within a newly hired cohort) and co-workers’
perceived reason for hire, Heilman and Blader (2001) asked laboratory participants to
evaluate a fictional cohort of individuals recently accepted to a gender-neutral graduate
school program (also fictional and called “Urban Health Studies”). The research design
was a 3 X 3 factorial, with the independent variables being the Stated Admissions Policy
(preferential selection, merit-based, or ambiguous) and the target admittee's Proportional

Gender Representation among the total group of admittees for that year (solo woman,

41

non-solo woman, or solo man). The authors measured co-workers perceived reason for
hire by asking participants to rate the role they thought several different factors played in
the decision to admit the target student (played a large role/played no role at all), one of
which was the applicant's gender. Heilman and Blader (2001) also measured participants’
perceptions of the solos’ qualifications as well as their expectations of the solos’ future
success in the program.

As hypothesized, Heilman and Blader (2001) found that participants who were
given no explicit information about the selection process (i.e., those in the “ambiguous”
condition) perceived the solo females to be preferentially selected. Further, they found
that participants rated solo females as less qualified and less likely to succeed (compared
to males and non-solo females). These results are consistent with those of Craig and
Feasel (1998), who found that participants were more likely to perceive females (vs.
males) bearing solo status within a workgroup as selected for membership on the basis of
their sex (vs. ability). Thus, these studies jointly suggest that, in the absence of
information to the contrary, co-workers presented with newly hired female solos assume
that they were preferentially selected.

Heilman and Alcott (2001)

Heilman and Alcott’s (2001) study is somewhat tangential in that it is not an
examination of solo status. However, it is related to the previously reviewed study
(Heilman & Blader, 2001) — which demonstrates that co-workers assume that newly hired
female solos were preferentially selected — in that it examines how knowledge of this
assumption affects females’ task difficulty choice. Heilman and Alcott (2001) manipulate

participants’ knowledge of others’ assumption about their reason for hire (i.e.. merit or

42

sex). In the interest of presentational simplicity, I will refer to one’s belief that others
perceive them as selected (at least in part) on the basis of their sex as a meta-perception
of sex-based preferential selection. As before, this meta-perception is just a perception
and is not necessarily accurate.

Heilman and Alcott (2001) suggest that, for females with initially low self-
confidence to perform the relevant task, the awareness that others perceive them as
preferentially selected prompts them to assume that others perceive them as less
competent, which in turn makes them feel less competent, select less difficult tasks, and
have less positive affect when performing the task. This perspective implies the following
three key points

First, the authors hypothesize that, for females, the meta-perception of preferential
selection negatively relates to inferred competence expectations (i.e., one’s perception of
how capable he/she is perceived to be by others). Second, they suggest that, for females,
the meta-perception of preferential selection and information about task ability (i.e.,
whether one has been given positive feedback, negative feedback, or no feedback about
their ability to perform the task) have an interactive effect on task difficulty choice and
the favorability of one’s affective reactions. More specifically, they suggest that, in the
absence of favorable information about one’s task ability, females’ meta-perception of
preferential selection will negatively relate to task difficulty choice and to positive affect
(e.g., enjoying the task). Finally, the authors suggest that are at least two possible
mediating mechanisms. They suggest that self-efficacy6 (i.e., belief about one’s capability
to perform; Gist & Mitchell, 1992) acts as a mediator for individuals not provided with

information about their ability and that a construct called motivational forces (which they

 

6 Note that the authors use “information about task ability” and “self-efficacy” interchangeably.

43

do not define) acts a mediator for individuals provided with favorable information about
their ability. They also cite the stereotype threat literature and hint at the possibility that
“stigma awareness" and/or “stereotypes” may somehow play a role. Thus, similar to
Heilman and Blader (2001). Heilman and Alcott (2001) do not make any stereotype-
based hypotheses but imply that stereotypes are somehow relevant. However, in the
present case. the focus is on participants’ meta-stereotypes (i.e., perception of others’
stereotypic views of them), while, previously (Heilman, 1980; Heilman & Blader, 2001),
the focus was others’ other-stereotypes about the participants (which were solos). The
role of solos’ meta-stereotypes is worthy of further consideration, and is addressed it in
more detail later in the paper.

To empirically test these ideas, Heilman and Alcott (2001) conducted two studies
where they partnered female participants with male confederates to perform a task that
they were told was one at which men traditionally perform better. Study 1 was a 2 x 3
factorial design and Study 2 was a 2 x 2 factorial design. The authors manipulated the
same two independent variables for both studies — Implied Selection Method (merit or
preferential selection based on gender) and Information About Task Ability Level (high,
low, or not provided) — but in Study 2, the authors omitted the “low ability” condition.
The dependent measures were inferred competence expectations, task difficulty choice,
self-view of competence, and favorability of affective reactions.

The favorability of participants’ affective reactions while performing the task
were assessed with a general measure that Heilman and Alcott (2001) created. The
authors averaged self-reports of favorability of affect while working on the task (e.g.,

happy — not happy, not annoyed — annoyed), favorability of descriptions of the task on

44

affective dimensions (e.g.. pleasant -— unpleasant). and favorability of feelings about the
research session (e.g.. very/amiable -~ not at all favorable). All ratings were made using
9-point response scales.

The results of both studies provide support for Heilman and Alcott’s (2001)
perspective. First, they found that females’ meta-perception of preferential selection
negatively related to inferred competence expectations (regardless of information about
ability). Second, the results of Study 2 suggested that the relationship between females’
meta-perception of preferential selection and task difficulty choice was moderated by
information about task ability such that the two were positively related for those given
favorable ability information and negatively related for those given no ability
information. Third, intercell contrasts indicated that participants experienced less
favorable affect when the implied selection method was gender- (vs. merit-) based.

Finally, the results provide tentative support for the suggestion that, at least for
females with no information about their ability, the relationship between the meta-
perception of preferential selection and task difficulty choice was mediated by some sort
of self-perceived ability construct (self-efficacy and/or self-perceptions of competence).
Specifically, Heilman and Alcott (2001) conducted covariance analyses (similar to those
conducted by Heilman, 1980) and found results that were supportive of the mediating
role of self-perceptions of competence, but the authors did not conduct a full-scale
mediational analysis. The authors did not empirically test for the mediating presence of
“motivational forces.” Thus, in sum, Heilman and Alcott’s (2001) results suggest the

relationship between females’ meta-perception of preferential selection and task

45

difficulty choice is mediated by and depends on constructs that reflect one’s initial
feelings of capability.
Summary

Although they do not directly speak to the issue. the aforementioned studies
(Heilman, 1980; Heilman & Alcott, 2001; Heilman & Blader. 2001) may collectively
help to better understand the effects of bearing solo status within an applicant pool on the
cognitions, affect, and behavior of newly hired females. Namely, these results can be
extrapolated to derive four tentative propositions that collectively imply the framework
presented in Figure 3. As before, Heilman and her colleagues have not explicitly
advanced this perspective, but it rather it represents how her ideas and research on others’
reactions to solos within a selection context might apply to the solos themselves.

Three boundary conditions of Heilman and her colleagues’ research should be
kept in mind when interpreting the present framework. First, Heilman and her colleagues’
research focuses on females operating in traditionally masculine domains (e.g.,
leadership, management). Second, Heilman and her colleagues’ research focuses on
people’s inferences in the absence of a stated selection policy. Third, Heilman and her
colleagues’ research focuses on situations in which sex is somehow made salient (e.g., by
directly manipulating solo status).

The framework (and the proposed model that follows) is predicated on two key
assumptions, the logic of which are rooted in what is known as heuristic processing (i.e.,
the use of “cognitive shortcuts”) to judge a situation, to simplify a decision, or just to
manage uncertainty (see Kahneman, Slovic, & Tversky, 1982). Thus, the first assumption

is that, since many people engage in heuristic processing (even if unconsciously).

46

8:: 80:8 :0080008 :0 0:500 200.08 0:: 0:0 00:0:0008 0.:0_:0> 0::: :0: :0:3 00.00: m: :_ 0.
8:00:30 :0 0:0 .8008: 0 £0: 00 030:0 m: :_ 00 0000080 w:_:0_00:: 0.0300? 0::: :0: 80003 03080: 3:0 0: 80: :t N
030:? 0::: A8308: 0: 0000000 00: 08030808 m0=w00=00 :0: 0:0 88:01 _

_ ................... :

 

A0>00w0c
I 03:80:
80:008
0>_:00::<

 

 

 

 

M00:00_.::00.:_0m
0:::—

 

 

8050:0005
00:0:00800
0800::—

 

 

 

 

 

 

 

 

0065
032.00 :5

 

 

 

m00:0:00800

A
V

_:0_:00_0m 3:880:80 _000 8002000 8::3
00>_00:00.:_0m :0 8300800002 030:0 0.0m

 

 

 

A

 

 

 

 

 

 

 

 

 

 

 

 

N00:0:00800
0020800

.:_0m

 

 

 

I—O-o—O—O—l_l—l--I-I_t—I—I—O—t—o—u—I—n—n—o—o—I-o—o-~

:80008 .000w00=00 :0: 0:0 008002 03600800 0:: 80:: :0:080:0:0 030:0 0:00 0: 0. .m 0050:“:

47

solos believe that others (e.g., co-workers) make use of some sort of heuristic or
“cognitive shortcut” when exposed to what Heilman refers to “something out of the
ordinary” — e.g., a newly hired female who bore solo status in her applicant pool — and
thus are interested in “making sense of” the situation (Heilman & Blader, 200]; Heilman
& Haynes, in press).

The second assumption involves the source of the content of the heuristic (or
“cognitive shortcut”) that is believed to be present. Specifically, the model that follows is
based on the assumption that there exists a relationship between the actual and expected
content of the typical (or average) heuristic, lending support to the idea that solos expect
others to view them as less competent (than men) and to assume that they were
preferentially selected (Heilman, 1980; Heilman & Blader, 2001). Both assumptions are
important, as they allow me to use Heilman and her colleagues work on the reactions of
others (Heilman, 1980; Heilman & Blader, 2001), to generate hypotheses about one
expects the reactions of others to be as well, and, related, to connect the present line of
research (which focuses on the effects of bearing solo status) to that of Heilman and
Alcott (2001; which focuses on the effects of knowing that others assume you to be
preferentially selected).

With the boundary conditions explained and the assumptions noted, I can now
describe the four propositions that comprise the implied framework. First, Heilman and
her colleagues’ research can be extrapolated to suggest that newly hired females solos
hold the meta-perception of preferential selection and that this perception may occur
through meta-stereotypic conceptions of women as less competent and less likely to

succeed. Support for this proposition can also be derived from the results of Heilman

48

(1980), which suggest that others are less likely to hire solos (than non-solos) and that
this effect might occur via the use of sex stereotypes. And, based on the aforementioned
assumptions, support can also be derived from the results of Heilman and Blader (2001).
which suggest that others perceive female solos as preferentially selected and, moreover,
as less competent and less likely to succeed. Thus, these studies establish that, for others
(such as solos’ co-workers), there exists a link between solo status and the perception of
preferential selection, and that stereotype use may mediate this relationship. Thus, I am
suggesting that Heilman and her colleagues’ research might be extrapolated to suggest
the solos themselves are aware of this (similar to arguments made by Jussim, 1986).
Second, the results of Heilman and Alcott (2001) suggest that the meta-perception
of preferential selection negatively relates to desirable outcomes such as self-perceptions
of competence, inferred competence expectations, and task difficulty choice.
Unfortunately, as is shown in Figure 3, their primary analyses do not reveal the
relationships amongst these constructs. Third, Heilman and Alcott’s (2001) follow-up
analyses provide tentative support for the idea that self—perceptions of competence play a
mediating role in the relationship between the meta-perception of preferential selection
and cognitions, affect, and behavior. Finally, the results of Heilman and Alcott (2001)
suggest that favorable information about ability may buffer the negative effect of the
meta-perception of preferential selection on newly hired females cognitions, affect, and

behavior.

49

POINTS OF THEREOTICAL CONVERGENCE BETWEEN THE SOLO STATUS
AND STEREOTYPE THREAT LITERATURES

The solo status and stereotype threat literatures point to five broad points of
theoretical convergence, which imply the framework shown in Figure 4. The necessary
boundary conditions of stereotype threat still apply. Namely, there must be potential for
(in this case, unfavorable) evaluation by an out-group member in a valued domain, a
negative stereotype must exist (Steele, 1997), the target must be aware of the stereotype,
and the target must believe that the relevant entity (i.e., the one making the evaluation) is
aware of the stereotype.

First, these literatures converge to suggest that bearing solo status may prompt the
solo to experience stereotype threat. This point is most clearly supported by Roberson et
al., (2003), who empirically established a link between bearing solo status (within one’s
workgroup) and self-report measures of stereotype threat (see also Stangor et al., 1998).
However, it is also supported by the results of Niemann and Dovidio (1998), who
empirically established a link between bearing solo status (within one’s department) and
perceived distinctiveness, a construct which (as they measure it) “may be closely
associated with awareness of stigmatization” (p. 56). Support for the link between solo
status and stereotype threat can also be derived from the results of Cohen and Swim
(1995), who found that solos (within a workgroup) expect to be unfavorably stereotyped.
It can also be derived from the theoretical reasoning of Sekaquaptewa and Thompson

(2002), who suggest that negative stereotypic expectations somehow play a role in the

50

 

 

 

 

830:0: 000 .:00::0
.000000w00

 

 

0000205 .8050
:0 00000800

 

 

 

08380:: :08:: 0008880 000 030:0 0:00 0:: 00080: 00008088 38:88:: :0 00:00 .0. 0:030:

 

 

 

 

00::020>0.:_0m

 

 

 

 

02:00—05:30
3:0:

 

 

 

 

 

 

:003: 005080:m

 

 

 

 

 

030:0 20m

 

 

 

51

relationship between solo status and desirable cognitions and behavior.

An implication of this first point of convergence is that, when working from the
perspective of the solos themselves, the stereotype threat literature subsumes the solo
status literature such that the latter is a specific case of the former. That is, stereotype
threat can be prompted by many things including bearing solo status (see Table 3), but
bearing solo status reliably prompts stereotype threat (Roberson et al., 2003; Stangor et
al., 1998; see also Niemann & Dovidio, 1998). This idea is consistent with Kanter
(1977a), who suggests that salient minorities (e.g., tokens, solos) “are more easily
stereotyped than pe0ple found in greater pr0portions” (p. 211) and with Steele’s seminal
piece (Steele & Aronson, 1995) in which he refers to solo status as a factor that is “akin
to stereotype threat” (p. 798).

A second, related, implication of this first point of convergence is that the
framework implied by the solo status literature informs the stereotype threat literature and
visa versa. Support for this point can be derived from the fact that the two literatures do
not yield any conflicting hypotheses. Indeed, they instead yield nearly identical
frameworks (see Figures 1 and 2) that can likely be reconciled by refining their respective
terminologies (e.g., perceived distinctiveness). Moreover, many of the studies in the
present review draw from both bodies of work to support their arguments (e. g., lnzlicht &
Ben-Zeev, 2000; Niemann & Dovidio, 1998; Roberson et al., 2003; Sekaquaptewa &
Thompson; 2002; 2003, Stangor et al., 1998; Steele & Aronson, 1995).

Thus, the second point of theoretical convergence is that stereotype threat

(resulting from solo status) negatively affects desirable (i.e., performance-facilitating)

52

cognitions. affect. and behavior. This point is supported by the stereotype threat
literature. which has not examined stereotype threat’s affective consequences but has
linked stereotype threat (resulting from solo status) to lower performance expectancies
(Stangor et al.. 1998) and to non-optimal feedback-seeking behaviors (Roberson et al.,
2003). This point is also consistent with the solo status literature in that solo status has
been shown to produce a variety of undesirable cognitions, affect, and behavior (e.g.,
cognitive deficits, Saenz & Lord, 1989; lower performance, Lord & Saenz, 1985;
Sekaquaptewa & Thompson, 2003; lowerjob satisfaction, Neimann & Dovidio, 1998).

Third, these literatures converge to suggest that the relationship between
stereotype threat and cognitions and behavior is mediated by a construct relevant to one’s
self-evaluation as well as by a construct relevant to one’s perception of others’
evaluations of one’s self. This point is supported by the stereotype threat literature, which
suggests that the link between stereotype threat and test performance is mediated by
performance expectancies (Cadinu et al., 2003) and by evaluative concerns (Osborne,
2001). It is also supported by the solo status literature, which provides empirical support
for the idea that performance expectancies mediate the relationship between solo status
and performance (Sekaquaptewa & Thompson, 2003).

Initially, the idea that stereotype threat negatively relates to a construct relevant to
one’s self-evaluation may seem to be in conflict with the stereotype threat literature,
which, as stated previously, has found that those who experience stereotype threat need
not endorse the stereotype (e.g., Steele & Aronson, 1995). This issue has not been
explicitly discussed within the stereotype threat literature, but, briefly, I argue that

stereotype threat can negatively affect, for example, one’s self-evaluation of their ability

53

to receive a favorable evaluation in the relevant context while at the same time have no
effect on their overall perception oftheir ability. This seemingly paradoxical statement is
consistent with research on the self-protective properties of stigma (e.g., Crocker &
Major, 1989), which suggests that stigmatized individuals attribute negative feedback (or,
in this case, anticipated negative feedback) to prejudice against their group. This external
attribution is likely easier to make when one call into question the validity of the
performance measure by. for example, saying that a supervisor and/or a test is biased.
Finally, these literatures converge to suggest the relationship between stereotype
threat and cognitions, affect, and behavior is moderated by a construct relevant to one’s
initial self-evaluation. Research has not yet examined this moderator in conjunction with
the other variables in the model and thus its point of moderation is unclear. Having said
that, this general point is most clearly supported by the solo status literature, where
Cohen and Swim (1995) found that many of the negative effect of solo status (e.g.,
preference for a different group) were stronger for women with low (versus high)
confidence. This point is also more tangentially supported by the stereotype threat
literature, where Stangor et al. (1998) found that stereotype threat had the least
detrimental effect on the performance expectancies of individuals who were given reason

to believe they were good at the task they were about to perform.

54

PROPOSED MODEL

Here, using Heilman and her colleagues‘ research as a guide, I tailor the loose
framework shown in Figure 4 to develop a model of the relationship between bearing
solo status within an applicant pool and newly hired females’ cognitions. affect, and
behavior. This model is shown in Figure 5. I will also examine whether this model holds
after controlling for cognitive ability.

First, I propose that bearing solo status within an applicant pool prompts
stereotype threat. More specifically, bearing solo status prompts the activation of the
unfavorable meta-stereotype that women are less competent and less likely to succeed in
traditionally masculine jobs. General support for this link comes from the solo status and
stereotype threat literatures, which, as previously discussed, converge to suggest that solo
status activates stereotype threat.

Specific support for the proposed content of this particular meta-stereotype can be
derived from the research of Heilman and her colleagues. Heilman’s (1980) work implies
that females bearing solo status within an applicant pool are subjected to others’ use of
detrimental sex stereotypes stating that women are less capable of performing
traditionally masculine jobs. Similarly, Heilman and Blader (2001) suggest that females
bearing solo status within a newly hired cohort of individuals are assumed by co-workers
to be less competent and less likely to succeed. Thus, based on the framework present in
Figure 3 and the assumptions that accompany it, I am arguing that the solos themselves

are aware of this.

55

 

 

0000800000
:00::<
803000:
w::0_00m-0_00:0000
8:20 032.00 000:

 

 

 

v

 

0:00:00 0: 09:30
0: 000000000
8800 :0 00000800

 

 

 

>0000:0.:_0m

 

 

 

 

 

 

5000:8000

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

000:0:
0:
:000
.fi 000023 5002000
3 00:80:80 N: :0000 _ I 00005
0.: :0 t 005088m 020:0
00000800 0_0m
0:02
3
E
E
m: l.

:0000: 0000080 .m. 0.5%“:

56

fl: For newly hired females, bearing solo status within their applicant

pool will prompt stereotype threat. More specifically, it will activate

unfavorable meta-stereotypic conceptions of women holding traditionally

masculine jobs.

Second, I propose that this particular type of stereotype threat (which suggests
that women hired for traditionally masculine jobs are less competent and less likely to
succeed) will prompt the activation of the meta-perception of preferential selection (in the
minds of the solos themselves). This general notion of a meta-stereotype—meta-
perception link is consistent with the larger literatures on stereotypes (e.g., Fiske &
Taylor, 1991) and meta-stereotypes (Vorauer et al., 1998, 2000), which respectively
suggest that stereotype-consistent inferences follow from stereotypes (as a means of
perpetuating them) and that meta-stereotypes drive meta-perceptions.

The more specific idea of a link between this particular meta-stereotype and the
meta-perception of preferential selection (in the minds of the solos) can be derived from
the work of Heilman and her colleagues, who have repeatedly hint at the presence of a
similar link in the minds of others. First, Heilman’s larger body of work on preferential
selection contains repeated suggestions of an association between the “stigma of
incompetence” and preferential selection (see Heilman, 1996; Heilman, et al., 1992,
Heilman, Block, & Stathatos, 1997). Second, Heilman (1980) found that the use of
detrimental sex stereotypes prompted others to make unfavorable selection decisions
about the solo, which may suggest that knowledge of the favorable selection decision
would prompt assumptions of preferential selection. Finally, this idea is consistent with

Heilman and Blader (2001), who suggest that assuming that a solo (in a newly hired

57

cohort) was preferentially selected “requires far less cognitive exertion than modifying
group stereotypes or devising new subgroup categories to account for the unexpected
occurrence” (p. 181). Similarly, based on the framework presented in Figure 3 and the
assumptions that accompany it, I am arguing that solos are aware of this and thus the
meta-perception of preferential selection follows from this particular form of stereotype
threat.

fl: For newly hired female solos, stereotype threat — and, more

specifically, unfavorable meta-stereotypic conceptions of women holding

traditionally masculine jobs — will prompt the meta-perception of

preferential selection.

Third, I propose that the meta-perception of preferential selection will negatively
affect cognitions, affect, and behavior. This suggestion is largely consistent with Heilman
and Alcott (2001). Thus, this hypothesis largely follows the lead of these authors and
proposes that the meta-perception of preferential selection will negatively affect task
difficulty choice, affect favorability, and performance. However, it also expands upon
Heilman and Alcott’s (2001) examination of affectively favorability by suggesting more
specifically that the meta-perception of preferential selection will negatively relate to
positive aroused affect (e.g., happiness) and positively relate to negative aroused affect
(e.g., anxiety). Moreover, I am incorporating Roberson’s (2003) ideas about feedback-
seeking behaviors. Specifically, Roberson et al. (2003) found stereotype threat to

negatively relate to feedback-seeking.

Hi The meta-perception of preferential selection will negatively affect
cognitions, affect, and behavior.

@: The meta-perception of preferential selection will
negatively affect task difficulty choice, such that the greater
the meta-perception of preferential selection, the less
difficult task they will choose.

58

H3b: The meta-perception of preferential selection
negatively will relate to performance.

H_3c_: The meta-perception of preferential selection
negatively relates to positive aroused affect (e. g.,
happiness) and positively relates to negative aroused affect
(e.g., anxiety).

H3d: The meta-perception of preferential selection will
negatively relate to feedback-seeking.

Fourth, I propose that the relationship between the meta-perception of preferential
selection and cognitions, affect, and behavior will be jointly mediated by individuals’
self-efficacy and perception of others’ confidence in her ability to perform. This
suggestion is consistent with the three streams of research reviewed above. As stated
previously, both the solo status and stereotype threat literatures suggest that stereotype
threat (occurring as result of solo status) affects cognitions, affect, and behavior through
its negative affect on constructs relevant to one’s self-evaluation (performance
expectancies; Cadinu et al., 2003; Sekaquaptewa & Thompson, 2003) and to one’s
perception of others’ evaluations of one’s self (evaluative concerns; Lord & Saenz, 1985;
Osborne, 2001; Saenz, 1994; Saenz & Lord, 1989). Thus, it makes sense to argue that any
meta-perceptions following from the relevant meta-stereotype will not disrupt this
process.

Similarly, Heilman and her colleagues (Heilman & Alcott, 2001) suggest that
meta-perception of preferential selection influence cognitions, affect, and behavior via
self-perceptions of competence and/or self-efficacy and also note that inferred
competence expectations may somehow play a role. Here, I am following these

researchers’ leads yet at the same time using what I see as the most conceptually and

59

methodologically appropriate constructs for this particular context: self-efficacy and
one’s perception of others’ confidence in her ability to perform. That is, I am
hypothesizing that the meta-perception of preferential selection negatively affects
cognitions, affect, and behavior through its negative effect on self-efficacy and on

perception of others’ confidence in one’s ability to perform.

H_4; The relationship between the meta-perception of preferential selection
and cognitions, affect, and behavior will be partially mediated by
individuals’ self-efficacy. More specifically, the meta-perception of
preferential selection will affect cognitions, affect, and behavior through
its negative effect on self-efficacy, which in turn has a positive effect on
cognitions, affect, and behavior.

Big: The relationship between the meta—perception of
preferential selection and task difficulty choice will be
partially mediated by individuals’ self-efficacy. More
specifically, the meta-perception of preferential selection
will affect task difficulty choice through its negative effect
on self-efficacy, which in turn has a positive effect on task
difficulty choice.

Mb: The relationship between the meta-perception of
preferential selection and feedback-seeking behaviors will
be partially mediated by individuals’ self-efficacy. More
specifically, the meta-perception of preferential selection
will affect feedback-seeking behaviors through its negative
effect on self-efficacy, which in turn has a positive effect
on feedback-seeking behaviors.

5g: The relationship between the meta-perception of
preferential selection and affect will be partially mediated
by individuals’ self-efficacy. More specifically, the meta-
perception of preferential selection will influence affect
through its negative effect on self-efficacy, which in turn
has a positive effect on affect.

km: The relationship between the meta-perception of
preferential selection and performance will be partially
mediated by individuals’ self-efficacy. More specifically,
the meta-perception of preferential selection will affect

60

performance through its negative effect on self-efficacy,
which in turn has a positive effect on performance.

115; The relationship between the meta-perception of preferential selection
and cognitions, affect, and behavior will be partially mediated by
individuals’ perception of others’ confidence in their ability to perform.
More specifically, the meta-perception of preferential selection will affect
cognitions, affect, and behavior through its negative effect on individuals’
perception of other’ confidence in their ability to perform, which in turn
has a positive effect on cognitions, affect, and behavior.

__HS_a: The relationship between the meta-perception of
preferential selection and task difficulty choice will be
partially mediated by individuals’ perception of others’
confidence in their ability to perform. More specifically,
the meta-perception of preferential selection will affect task
difficulty choice through its negative effect on perception
of others’ confidence in their ability to perform, which in
turn has a positive effect on task difficulty choice.

1150: The relationship between the meta-perception of
preferential selection and feedback-seeking behaviors will
be partially mediated by individuals’ perception of others’
confidence in their ability to perform. More specifically,
the meta-perception of preferential selection will affect
feedback-seeking behaviors through its negative effect on
perception of others’ confidence in their ability to perform,
which in turn has a positive effect on feedback-seeking
behaviors.

Q: The relationship between the meta-perception of
preferential selection and affect will be partially mediated
by individuals’ perception of others’ confidence in their
ability to perform. More specifically, the meta-perception
of preferential selection will influence affect through its
negative effect on perception of others’ confidence in their
ability to perform, which in turn has a positive effect on
affect.

fl5_c_l: The relationship between the meta-perception of
preferential selection and performance will be partially
mediated by individuals’ perception of others’ confidence
in their ability to perform. More specifically, the meta-
perception of preferential selection will affect performance

61

through its negative effect on perception of others’
confidence in their ability to perform, which in turn has a
positive effect on performance.

Finally, I propose that the relationship between the meta-perception of
preferential selection and self-efficacy will be moderated by one’s initial self-efficacy,
such that individuals with initially higher self-efficacy will experience less detrimental
effects to their self-efficacy as a result of the meta-perception of preferential selection
than those with initially lower self-efficacy. The general idea that initial feelings of
capability can buffer solo status’ negative effects is consistent with the three streams of
research reviewed above. As stated previously, the solo status and stereotype threat
literatures converge to suggest that the relationship between stereotype threat and
cognitions, affect, and behavior is moderated by a construct relevant to one’s initial self-
evaluation (Cohen & Swim, 1995; see also Stangor et al., 1998). Similarly, Heilman and
her colleagues (Heilman & Alcott, 2001, Study 2) suggest that the relationship between
the meta-perception of preferential selection and task difficulty is positive (as opposed to
negative) when participants were given favorable information about their ability.

Although each of these streams of research generally suggests that initial feelings
of capability can buffer solo status’ negative effects, none clearly specify the exact point
of moderation — i.e., the link where initial self-efficacy plays a role. Rather, these studies
collectively assess the interactive effect of different independent variables (e.g., solo
status; Cohen & Swim, 1995; Stangor, et al., 1998; meta-perception of preferential
selection; Heilman & Alcott, 2001) and positive feedback on a variety of outcomes (e.g.,
desire to change group composition, Cohen & Swim, 1995; performance expectancies,

Stangor, et al., 1998; task choice difficulty, Heilman & Alcott, 2001). Here, I am

62

measuring (rather than manipulating) initial feelings of capability using the construct of
self-efficacy and hypothesizing that it acts to buffer solo status’ negative effects on
cognitions. affect. and behavior by buffering the negative effect of the meta-perception of
preferential selection on one's perception of others’ confidence in one’s ability to
perform. That is, the degree to which the meta-perception of preferential selection
negatively relates to this variable will depend on one’s initial self-efficacy, with those
with initially high levels of initial self-efficacy having self-efficacy levels that are more
resistant to change. Additionally, initial self-efficacy is not independent of self-efficacy,
so I am also hypothesizing that the two will relate to each other.

H6: The relationship between the meta-perception of preferential selection

and one’s perception of others’ confidence in one’s ability to perform will

be moderated by one’s initial self-efficacy, such that individuals with

initially higher self-efficacy will experience less detrimental effects to

their self-efficacy as a result of the meta-perception of preferential

selection than those with initially lower self-efficacy.

HZ: Initial self-efficacy will relate to self-efficacy.

63

METHOD
Overview

This study involved recruiting students from two Midwestem college campuses to
participate in what they were told was an evaluation of a selection process. Publicity
efforts on both campuses invited students to apply for a fictitious job that could be
completed over the Internet and. if selected, to work on the “job" for a short time (about
20 minutes). Participants were told that they would have the opportunity to participate in
the first phase because it involved applying for the job, but that only those who were
“selected” would be asked to participate in the second phase because it actually involved
working on the job for a short time. In reality, all participants were told they were
selected and given the opportunity to participate in both phases.

The solo status manipulation was delivered at the same time as the favorable
selection decision. It involved indicating to female participants that they were the only
person of their gender in an otherwise male applicant pool. For this reason, it was
necessary to recruit both males and females. Thus, while I collected data from both males
and females, I did not analyze the data from males. All further discussion is of the sub-
sample of females.

When someone first accessed the site (www.participate-in-research.com), he or

 

she was directed to a page titled, “Introduction to the Study,” and then to a Consent
Form. After completing the Consent Form, participants began Phase One or “the
application phase,” which involved being asked to provide typical employment test and
demographic information. Upon completion of this information, the applicant logged off,

submitting her data for evaluation. Within a period of time ranging from 12 to 24 hours

64

(Monday through Friday), participants received an e-mail message (from selection-
decision@partieipate-in-research.com) notifying them that they had been selected for the
job and were given false information about the others in their applicant pool. More
specifically, participants were provided with a list of clearly male and/or clearly female
names in a manner that suggested that these fictitious individuals were the other people in
the applicant pool who were not selected for the job.

Female participants in the solo status condition were given a list of clearly male
names (e.g., Jonathon) while those in the non-solo status condition were given a list of
clearly male and clearly female names (e.g., Jason, Elizabeth, respectively). The list of
clearly opposite sex names was intended to make participants believe that they were the
only person of their gender in the applicant pool, while the list of mixed sex names was
intended to make them believe that there were people of both sexes in the applicant pool.

The second part of the study was Phase Two or “the job phase.” Participants who
chose to “accept” the job logged back onto the website and were asked to work on the job
for a short time (about 20 minute) and then to respond to various survey items. After
completing the task and the survey items, the participants signed off, sending the data to
the investigator.

Participant Recruitment and Sign- Up Process

Flyers were posted on both campuses and classified advertisements were placed
in both campus newspapers to recruit participants for the study. Both media advertised a
fictitious job that paid $25 for about an hour and a half of work and briefly explained that
this job was part of a research study designed to test an online selection system.

Interested parties were directed to the website for the details. The flyers also explained

65

the payment schedule. Participants were paid $10 for participating in Phase One (the
application phase) and another $15 for participating in Phase Two (the job phase).
Additionally, it was explained that a $10 bonus would be given to the top performers in
Phase Two, with the odds of winning equal to approximately one in ten.

Interested parties who visited the site were directed to a page titled, “Introduction
to the Study,” which explained that this was a research study and communicated a cover
story intended to make potential participants feel as if they were actually going to be part
of a selection process. The website stated that the purpose of the study was to examine
the feasibility of having campus recruiters conduct the first phase of a company’s
selection process online rather than on-site to save on expenses (e.g., airfare). Next, the
web page asked interested participants to click on a link that directed them to a page
titled, “Consent Form” that explained the activities included in the study in more detail
and reviewed their rights as a participant. The bottom of the Consent Form indicated that
students who wished to participate in the experiment could indicate this by typing their
name and student identification number into the fields provided. Those who were not
interested could simply exit the website.

Key Features of the Cover Story

The website stated the organization was interested in applicants for a “managerial
internship.” The reason that the internship was specifically a managerial one was because
I wanted the job to be perceived as at least somewhat masculine by the participants
themselves, as this was one of the boundary conditions of my prOposed model
(previously shown in Figure 5). I unfortunately failed to ask participants any questions

about whether they actually perceived the job to be a masculine one or not. The

66

implications of this oversight will be addressed in the discussion section. For now, it is
only necessary to mention that any effects of solo status in the present study are likely to
be weaker and thus more difficult to detect to the extent that the job was actually not
perceived to be a masculine one (Karakowsky & Siegel, 1999; see also Carr, 2001;
Dipboye, 1985; Heilman, 1983; 2001). That is, any effects of solo status detected within
the present study should be stronger in situations where the job is perceived to be more
masculine than the one in the present study.

There is some evidence to support the idea that the aforementioned managerial
internship was perceived by participants to be at least somewhat masculine. Although
management jobs are “neutral” in an objective sense, (Powell & Graves, 2003) they are
“masculine” in a subjective one such that management is still associated with men in the
minds of Americans (Powell, Butterfield, & Parent, 2002; Sabine & Sczesny, 2003).
However, to bolster the perception that this internship was a masculine one, the website
stated that the company with which the research team was working was a
“communication and technology organization.” This phrasing is meaningful because
“technology jobs” (i.e., scientific, technical, and engineering occupations; Bureau of
Labor Statistics, 1999) are objectively more frequently held by men and thus referred to
as “masculine” (Powell & Graves, 2003; US. Department of Labor, 2003).

The website also explained that, after the study was over, “actual managers” from
the organization would review the entire applicant pool to determine how qualified the
managers thought that the selected and unselected applicants were. Including these
“actual managers” in the cover story was intended to create the sense that the applicant

would be evaluated by others at a later point in time.

67

Procedure

At the beginning of Phase One (the application phase), participants were asked to
provide demographic data. The main information of interest here was the self-reported
gender of the participants, which was used along with the randomly assigned solo status
condition to determine the content of the e-mail message that communicated the selection
decision. Participants were also asked to provide other information (e.g., race, score on
their college entrance exam). After being given a short description of the job for which
they would be applying, participants were asked to respond to survey items used to
measure their level of initial self-efficacy to perform the job. At the end of Phase One,
participants were asked to spend 15 minutes taking a selection test. At the completion of
the selection test, they sent in the application by clicking the link provided to them.

Within a period ranging from 12 to 24 hours (Monday through Friday),
participants received an e-mail message notifying them that they were selected. This e-
mail message also contained the solo status manipulation in the form of the names of all
others in the applicant pool. The participants were instructed to continue to the second
phase by clicking on a link provided in the notification of hire e-mail.

At the beginning of Phase Two (the job phase), participants were asked to respond
for a second time to items used to measure their self-efficacy to perform the job. They
were also asked to estimate the extent to which they expected their manager to perceive
them as capable of performing the job. Next, participants were asked to select the
difficulty level of the job they were about to complete and then allowed 20 minutes to
work on the task. In reality, all participants were presented with the same version of the

task (regardless of their choice of difficulty). Then, after spending 20 minutes on the job,

68

participants were asked to respond to a variety of items used to measure stereotype threat,
the meta-perception of preferential selection, affect, and feedback-seeking. More
information on the order in which the measures were administered is provided shortly.
For now it is worth noting that the order in which 1 administered the measures did not
directly correspond to the proposed model (shown in Figure 5) because I did not want to
risk alerting participants to the nature of the study before collecting the outcome
information.
Study Design and Participants

This study was a one-factor design with two levels, solo status or non-solo status.
A total of 442 people visited the website and provided a name and an e-mail address, but
many of these cases were omitted from the data set. First, of the 442, a total of 167 were
males and two did not indicate their gender. Among the 273 females remaining in the
sample, about a third (n = 96) failed to click on the link provided in the notification of
hire e-mail and thus were paid only for participating in Phase One. The rest (N = 177),
who completed both phases and were paid accordingly, comprise the sample for the
present study. Of these 177 cases, six provided logically impossible responses to a subset
of items, reducing the number of eligible participants to 171.

The remaining participants’ age ranged from 18 to 51 with a median age of 20
and an average age of 20.85 (SD = 3.41). The majority (86%) of respondents were White
(n = 147). Remaining respondents reported being Black or African American (n = 16),
Asian (n = 3), Mexican American (n = 2), Biracial (n = 2), or “other” (n = 1).
Participants reported being reasonably comfortable (M = 3.77; SD = 1.54) with

computers on a 5-point Likert scale (very uncomfortable — very comfortable).

69

Among the 171 remaining participants, 25 reported being unable to submit their
job performance data or move forward with the study. The computer programmer (who
designed the website) suspected that these participants were using an older web browser
that did not meet the system requirements of the original program. The programming
code was revised and the problem eventually fixed, but not in time to make it appropriate
to ask these individuals to check back and finish the study. Thus, as shown in Table 5,
there are 143 to 146 cases per variable from the point of the work sample forward. The
lowest pairwise sample for each hypothesis is 143.

Table 5

Study Measures and Sample Sizes

 

 

Measuresa Sample size
Initial self-efficacy 171
Selection test 171
Self-efficacy 171
POCA" 171
Task difficulty choice 170
Performancec l 46
Affect 143
Stereotype Threat I43
MPPSd 144
Feedback-seeking 143
Listwise 136

 

8The measures are listed in order of
administration. bPOCA is an abbreviation for
“perceptions of others’ confidence in one’s own
ability” cParticipants were told that they were
going to (perform the job by completing a work
sample. MPPS is an abbreviation for “meta-
perception of preferential selection”

70

Results of analyses comparing cases that were included in the analyses versus
those who were not, at multiple stages, are provided in Appendix B. While no concerning
trends were observed. participants who moved on to the second phase of the study were
older. scored higher on the selection test, and were white to a greater proportion than
those who did not move on to the second phase. Additionally, those who successfully
submitted their work sample in Phase Two had selected a more difficult task.

With one degree of freedom, these 143 participants provide enough statistical
power (80%; Murphy & Myors, 1998) to detect an effect size of .46 via the regression-
based analyses used for the individual hypothesis tests. Previous, similar research has
yielded effect sizes ranging from —.63 to -.76 (e.g., Sekaquaptewa & Thompson, 2003).
This sample of 143 participants also provides sufficient power for the path analyses used
to completely and simultaneously test the 15 parameter in the model., as ten to 20 cases
per estimated parameter have been described as sufficient to ideal, respectively (e. g.,
Kline, 1998, p. 112; Tabachnick & Fidell, 2001, pp. 659-660).

Solo Status Manipulation

In an effort to balance experimental control and external validity, I did not
directly inform participants of their solo status but rather I gave them information about
the gender of the fictitious others in their applicant pool (see Figure 6). Specifically,
participants in the solo status condition received an e-mail stating, “Human Resource
(HR) personnel from the communication and technology organization compared your
application to the most recent group of applications processed (6 males, 0 females).
Congratulations, HR decided to select you for this job.” The message listed all the

999

“applicants names along with their unfavorable selection decision, and the participant’s

71

names were always second from the bottom. Included in this e-mail was a link to the
study itself, and the content of the first page that participants saw upon clicking this link
was a replicate of the information contained in the e-mail message.

Participants in the control condition received a similar e-mail message linked to a
similar web page, but they were told that there were three males and three females in the

applicant pool and provided a list of male and female applicant names.

mm

 

 

.00 00 0: >000: 000 50% 00:3 :00: c :0: fl:.0000fi 0. 00 :0: :00 "00 m 0 .0 00 800.00000000
-0:-000mwofluummwwfimmwfl u:0:> .00500 0:: :0 Mam: 000000 000 mumameoo ou x003 000 0>m£ so»

 

00:00:00 xamo SumnMNfiam
0000000¢ H0800 0.0:mmflufluum0:
00:00:00 unoEHHz :00:
00:00:00 :uoxoxz 00000
00:00:00 “00:00:00 MHHMU
00:00:00 0000:000 00:00
00:00:00 Hmcmudm mmum
coflmflomo cofluomamm unmoflammd

.00: 0::: no: 50% 000H00 0: 000:000
.mcofluMHSumumcou .A00H080: m .00H0: m: 000000000 mcoflumoflammm :0 msoum 00000: 0008 000 00 coflumoflammm
050% 00000500 cofluMNflcmmuo >m0H000000 0:0 00:000HCSEEOU 000 50:: Hmacomuwm Amm: 00050000 00:50

00:0:000 cofiuumamm "000:050

Eoo.noummmmuIQH-0:00:000000000:0:000100H000H00 "Eon:
..... mmmmmmz HMCHmfluoutitu

0020000 0:00.00: 0:: 0: 800060000 0: :000 0w0mm00: 0000-0 :0 :00:000 .0 000%:

 

73

Measures
Self-Efficacy

Since participants’ “job” involved generating unique schedules, the purpose of the
self-efficacy measure was to assess how confident participants felt in their capability to
complete a certain number of unique schedules (e.g. eight) within the 15-minute time
limit. To do this, I administered a 16-item version of Maurer and Pierce's (1998) version
of Lee and Bobko’s self-efficacy (1994) scale that uses a 5-point Likert scale asking
participants to rate how strongly they agreed with a series of performance statements
(e.g., “I can complete eight or more schedules”). I administered this measure twice to get
participants’ initial self-efficacy as well as their self-efficacy after receiving the positive
selection outcome. This measure was reliable on both occasions (a = .96 and a = .95,
respectively). This scale appears in Appendix A.

Selection T es!

The selection test was actually a cognitive ability test. Specifically, I administered
the Raven’s Advanced Progressive Matrices (Raven’s; Raven, Court, & Raven, 1985), a
non-verbal test that is considered a general test of cognitive ability (Carroll, 1993; Snow,
Kyllonen, & Marshalek, 1984). The test consists of 36 matrix or design problems
arranged in ascending order of difficulty with the score being the total number of
problems solved correctly. As pointed out by Mayer and Hanges (2003), one of the
advantages of this test was that its ambiguous form can easily be explained to participants
as either an intelligence test or a test of some other construct (e. g., analytical and problem
solving skills). The instructions included language intended to convince participants of its

face validity: “To help us determine whether or not you are qualified to be a managerial

74

intern at this communication and technology organization, you are about to take a test
that measures your analytical and problem solving skills. Previous studies suggest that
this test is a valid predictor of job performance.” Participants’ average response (using a
5-point Likert scale) to the four items placed at the end of the study to assess the test’s
face validity (see Appendix A) was 2.97. This neutral value indicates that, on average.
participants neither agreed nor disagreed that the test was face valid. Participants were
asked to spend 15 minutes on the selection test, and a timer embedded in the computer
program told them when their time was up. The results of the present study indicate that
the Raven’s was reliable (KR20 = .89).
Perception of Others ’ C onfidence in One '3 Own Ability

Items were adapted from two sources to create a measure of the perception of
others’ confidence in one’s own ability (POCA). The first source was a measure created
by Heilman and Alcott (2001) to assess “inferred competence expectations” and the
second was a measure created by Tyler and Blader (2002a; 2002b) to assess others’
respect for one’s work (Tyler & Blader, 2002a). Heilman and Alcott (2001) asked
participants to respond to two items asking them to infer their teammate’s perception of
their competence and likelihood of being successful at the task as well and one item
asking them to provide their expected level of performance. All ratings were made on 9-
point response scales, with anchors ranging from “very much” to “not at all” for the first
two items and from “very well” to “very poorly” for the last one. The scale was reliable
(a = .86). Tyler and Blader (2002a; 2002b) asked participants to rate the extent to which
others (i.e., co-workers) respect the work they did and their ideas as well as the extent to

which they value their contributions at work. All ratings were made with a 5-point scale

75

(strongly agree — strongly disagree) and the scale proved to be reliable (a = .95; Tyler &
Blader, 2002a).

In the present study, I adapted items from each of these scales to measure
participants’ perception of the organizations managers’ beliefs about their capability. A
principal axis factor analysis indicated that this measure has one factor (Table 6). It was

also reliable (or = .93).

Table 6

Exploratory Factor Analysis Results for POCA” Measure Using Principal
Axis Factoring (N = I 43)

 

Itemal Factor Loading
1. Respect the work I do .86
2. Respect my ideas .83

3. Value what I would contribute to the organization .88
4. Think it would be difficult to replace me .69
5. Disapprove of how I do my work (R) .46
6. Appreciate my contributions to the organization .83

7. Perceive me as competent .77
8. Perceive me as likely to be successful on the job .85
9. Expect me to perform the job well .78
% of variance 61.02%

 

Note. (R) indicates that the item was reverse-coded.

3Each item began with the phrase, “This organization’s managers will.”

b POCA is an abbreviation for “perceptions of others’ confidence in one’s
own ability”

Work Sample Performance
Participants were asked to spend 20 minutes on a work sample in Phase Two and
were told that doing so was the “job phase” of the study, meaning it was one in which
their job performance would be evaluated. As before, a timer embedded in the computer

program told participants when their time was up. The work sample was an adaptation of

76

a class scheduling task originally developed for goal-setting research (Early & Kanfer,
1985; see also Q. Roberson, Moye, & Locke, 1999; Steele-Johnson, Beuregard, Hoover
& Schmidt, 2000). The author had previously (Carr, 2001) adapted this task for use in a
similar study (e.g., it was also conducted online). It produced variability in participants’
performance (M = 3.54; SD = 3.42) and performance on it related to various other
constructs, such as self efficacy (r = .22, p < .001), self-evaluations of performance (r =.
67, p < .001), and the perceived fairness of the selection decision (r = .18, p < .001 ).

This class scheduling task was adapted to fit with the present study. In this case,
the task was presented as an inventory scheduling task (rather than an academic course
scheduling task [Early & Kanfer, 1985] or as a conference room scheduling task [Carr,
2001]). Participants were told that the organization was investigating the quality and
reliability of their suppliers and, for this reason, wanted the managerial interns to develop
as many unique inventory schedules as possible. Participants were provided with the
names of twelve suppliers, each with several possible inventory times, and a list of the
schedule constraints. Each schedule needed to be unique and contain five different
supplier codes within the same month was “valid.” More information about this task is in
Appendix C.

It is important to mention that over half the participants (n = 88) had zero valid
schedules, indicating a floor effect. This was unexpected because there was no such
effect in Carr (2001). Later, I address possible explanations for why there was a floor
effect in this study and not in the other. For now, it is worth noting that this floor effect

means that the performance data collected in the present study must be interpreted with

77

caution. Moreover, it is important to keep in mind that participants likely were able to
infer (at least to some extent) that they were not doing well on the task.
Affect

Affect was measured using Davis’s (2001) four-factor scale, which was adapted
from Spielberger’s State-Trait Anxiety Inventory (Spielberger, 1983). The scale measures
four related components of affect (positive aroused, negative aroused, positive calm, and
negative calm) and it is consistent with existing perspectives on the dimensionality of
affect (see Bagozzi, 2003; Frijda, 1999; c.f., Russell & Carroll, 1999). The items asked
participants what types of affect they were experiencing — i.e., “how they felt” — while
performing the task (e.g., happy, anxious, etc.) All ratings were made with a 5-point
Likert scale (strongly disagree — strongly agree).

The results of a principal axis factor analysis with oblique rotation (see Table 7)
are supportive of the four-factor solution: (a) Negative calm (seven items; e.g.,
“disappointed”), (b) Positive aroused (five items; e.g., “enthusiastic”), (c) Positive calm
(five items; e.g., “relaxed”), and (d) Negative aroused (four items; e.g., “tense”). As
shown in Table 8, the scales relate highly to each other, indicating a low level of
discriminant validity, but are reliable (with internal consistencies ranging from .89 to
.93).

However, because the meta-perception of preferential selection was only
hypothesized to relate to positive aroused affect (negatively) and negative aroused affect

(positively), only these two scales were used in the subsequent analyses.

78

 

Table 7

Results ofPrincipaI Axis Factor Analysis with Oblique Rotation for Affect

 

 

 

Measure
Factor Loadings

No Item Negative Positive Negative Positive

' Content Calm Calm Aroused Aroused
8 Discouraged 0.81 -0.23 0.17 -0.28
6 Disheartened 0.75 -0.12 0.33 -0.06
10 Dissatisfied 0.74 -0.29 -0.06 -0.27
1 Disappointed 0.73 -0.28 -0.03 —0. 10
2 Frustrated 0.67 -0.29 0.3 5 -0. 16
21 Annoyed 0.62 0.05 0.16 -0. l 7
7 Discontent 0.73 -0.13 0.40 -0.15
20 Content -0.33 0.79 -0.06 0.16
16 Calm -0. l 9 0.79 -0.31 -0.02
1 8 Delighted -O.1 8 0.75 -0. l 5 0.45
19 Relaxed 0.08 0.75 -0.46 0.13
1 5 Secure -0.45 0.67 -0.05 0.26
5 “On edge” 0.23 -O.l4 0.86 -0.13
3 Anxious 0.07 -0.20 0.83 0.16
4 Tense 0.33 -0.15 0.80 0.02
9 Stressed 0.50 -0.37 0.51 -0.25
1 1 Eager -0. l 8 -0.1 l 0.09 0.79
14 Excited -0.28 0.35 0.01 0.77
1 7 Enthusiastic -0.16 0.49 -0.03 0.71
l 3 Happy -0.29 0.52 -0.13 0.65
l 2 Cheerful -0.25 0.58 0.19 0.59
Eigenvalues 4.77 4.34 3.18 3.15

% of variance 22.7% 20.66% 15.16% 15.02%

 

Note. N = 142, Factor loadings over .40 appear in bold.

79

-M- manna-navy... — 'n—o aw -,-- ~r. wvl~"’.~ .... A--. .. -77 .. - 7‘ - . — <,__.,....'~..---.—.-._-~ -~».v ”-

Table 8

lntercorrelations and Internal Consistency Reliabilities for Affect

 

Factors
1 2 3 4
1. Positive Aroused .93
2. Positive Calm .67 .89
3. Negative Aroused -.25 -.56 .90
4. Negative Calm -.48 -.60 .72 .89

 

Note. All correlations off diagonal significant at the 0.01 level (2-tailed).
Reliability coefficients shown on the diagonal.

 

Task Difficulty Choice

This one-item measure asked participants to choose how difficult they would like
their work sample (i.e., “job”) to be. Participants had three response options: easy,
moderate, and difficult (coded as one, two, and three respectively). All three tasks were in
reality the same task and thus of the same difficulty level, so it was important to carefully
consider the best way to handle participants’ perception of the relationship between task
difficulty choice and eligibility for cash prizes. For example, if participants were told,
“The top 10% of performers in each condition will receive a $10 prize,” the participants
may have been angry when they found out (at the end of the study) that all tasks were
identical in terms of difficulty level. Thus, participants were just told that this choice
would be “taken into account” when awarding cash prizes. Most participants either chose
the moderately difficult task (n = 87) or the easy task (n = 61), although some chose the

most difficult one (n = 23).

80

Feedback-Seeking

The items of this measure asked participants how important it was for them to
receive feedback on their performance on the task. Specifically. three items (adapted
from Carr & Deshon, 2001) asked participants to rate the extent to which it was important
to have information about how well others performed on the task, about how to improve
their understanding of the task, and about how well they performed relative to others. All
ratings were made with a 5-point Likert scale (strongly disagree — strongly agree). This
measure was reliable (or = .88).

Stereotype Threat

This measure of stereotype threat was adapted from a previous measure of the
construct (Roberson et al., 2003; a = .77; see Appendix A). The first adaptation was
straightforward: The word “sex” replaced the word “race” in all items. For example, I
changed “My race does not...” to “My sex does not...” The second adaptation, which
was more complicated than the first, requires the reader to recall that stereotype threat
does not involve self-stereotypes (i.e., those pertaining to one’s own group; c.f., Wheeler
& Petty, 2001) but rather meta-stereotypes. That is, while an item assessing a self-
stereotype might say, “Cognitive ability tests may be easier to perform for people of my
race,” an item assessing a meta-stereotype might say, “Others will think cognitive ability
tests are easier. .

The third adaptation follows from the second such that I revised the items so that
they specifically referred to “this organization’s managers,” which I “introduced” to the
participants in the beginning of the study (i.e., when I communicated the cover story).

Thus, I adapted the items so that, instead of the generic “others” statement, they refer to

81

“this organization’s managers.” A specific example (that extends the aforementioned
generic example) of this (taken directly from the present study) is that I changed
“Cognitive ability tests may be easier to perform for people of my race” to “This
organization’s managers will think my job is easier for people of my sex to perform.”
This five-item scale had a very low coefficient alpha ((1 = .49). An inspection of
the corrected item-total correlations revealed that item two. “This organization’s
managers will think my job is easier for people of my sex to perform,” had a value of -
.62. This may have been because participants may have misinterpreted the item. Thus, I
dropped the item, leaving a four item scale with good reliability (or = .86). A principal

axis factor analysis yielded one factor (see Table 9).

Table 9

Exploratory Factor Analysis Results for F inal Stereotype Threat Measure Using Principal
Axis Factoring (N = 143)

 

Item Factor
Loading

1. This organization’s managers will think I have less ability to perform

. . .90
thlS Job because of my sex.
2. This organization’s managers will expect me to perform poorly because 92
of my sex. '
3. People of my sex will face biased evaluations from this organization’s 91
managers. '
4. My sex will not affect managers’ perceptions of my ability to perform 4

. . . 0
thlS jOb (R)
Eigenvalue 2.65
% of variance 66.33%

 

Note. (R) indicates that the item was reverse-coded.

82

Meta-Perception of Preferential Selection

This measure asked participants to rate the extent to which they believed the
organization’s managers would assume that the participant had been hired based on merit
(vs. sex). 1 adapted the items from Heilman and Blader’s (2001) measure of “assumed
gender-based preferential selection.” Heilman and Blader (2001) asked participants to
rate the extent to which participants thought several different factors played a role in the
decision to admit the target student, one of which was the applicant’s gender (played a
large role/played no role at all; p. 190). The measure used for the present study was also
a derivation of Heilman and Alcott’s (2001) manipulation of “implied selection method,”
where the authors varied information given to participants intended to “reveal” to them
their (confederate) teammates’ perception of the reason they, the participants. were
selected (i.e., on the basis of merit or sex).

Specifically, participants in the present study were asked to use a 5-point Likert
scale to rate the extent to which they agreed with five statements about the basis on which
the organization’s managers would think they were selected. The measure had two “key
items” and three “distracter items.” The “key items” read as follows: “This organization’s
managers will think that my sex played a large role in Human Resource’s decision to
select me for this job” and “This organization’s managers will think that my performance
on the selection test played a large role in Human Resource’s decision to select me for
this job.” The “distracter items” corresponded to the aforementioned demographic
information that was requested from participants (e.g., race, college admissions test

score).

83

A programming error that went undetected until all but 15 of the female
participants had completed the study resulted in the remaining participants’ responses to
the key item (“This organization’s managers will think that my gender played a large role
Human Resources’ decision to select me for this job”) not being recorded. Analyses were
conducted on the remaining item, which asked what role the organization’s managers
would think that the applicant’s performance on the selection test played a large role in
HR’s decision to select her for the job. This error is an obvious limitation of the study.
the implications of which will be addressed in the discussion section. For now, it means
that I will refer to MPPS as “the belief that one was not selected on the basis of the
selection test,” as this more accurately describes the remaining item.

The remaining item of the meta—perception of preferential selection measure
(“This organization’s managers will think that I was selected on the basis of my
performance on the selection test”) was reverse-coded to remain consistent with
hypotheses about the meta-perception of preferential selection (see Figure 5). It is an
open question as to whether responses to this (reverse-coded) item would have correlated
positively with those to the aforementioned gender one (to which only 15 responses were
recorded). However, with the very small and under-powered pairwise sample of 15, 1
found the correlation between these two items to be non-significant but positive, r = .31,
p > .05. Based on the magnitude of this correlation, I believe that the full sample’s
responses to the retained item would have significantly (and positively) related to the lost

gender item.

84

RESULTS
Manipulation Check

The solo status manipulation check was one item that asked participants to rate
the extent to which they agreed with the statement, “There were other people of my
gender in the applicant pool.” The item used a 5-point Likert response format where 1
indicated “strongly disagree” and 5 indicated “strongly agree.” To investigate if
participants in the solo condition were less likely to agree with this statement than those
in the non-solo condition, a t-test was conducted. Participants in the solo status condition
were less likely to agree with this statement (M = 2.34, SD = 1.41) than those in the non-
solo status condition (M = 4.19, SD = .95), t (144) = 9.15, p < .001. An examination of
the response distributions is consistent with this. The proportion of participants endorsing
a 3 or higher (i.e., “neither agree nor disagree”) was 37% for the solo status condition
versus 78% for the non-solo status condition. Thus, these results suggest that the solo
status manipulation was effective.

Tests of Hypotheses and Model

Table 10 presents the means and standard deviations of the major study variables
(overall and by condition). Table 11 presents the reliability coefficients and
intercorrelations of the major study variables. The results are presented in two sections.
First, I present result of analyses that specifically test each hypothesis in isolation.
However, these hypotheses may be better understood as an interrelated set of complex
indirect effects. Therefore, the second section presents a series of path analyses that
provide an overall evaluation of the model proposed here that accounts for the

interrelationships among all the variables of interest.

85

Table 10

Means And Standard Deviations of Major Study Variables

 

 

 

Overall Solo Condition Non-solo condition
n M SD n M SD n M SD
1. Solo status 171 n/a n/a 90 n/a n/a 81 n/a n/a
2. Initial self-efficacy 171 3.29 .89 90 3.31 .91 81 3.27 .87
3. Selection test 165 18.19 6.34 87 18.54 6.87 78 17.79 5.71
4. Self-efficacy 171 3.49 .86 90 3.43 .94 81 3.56 .76
5. POCAa 171 4.19 .71 90 4.15 .82 81 4.24 .56
6. TDCb 170 1.77 .67 89 1.80 .71 81 1.75 .62
7. Performancec 146 .91 1.49 77 .88 1.59 69 .94 1.37
8. Positive aroused 143 2.86 .91 76 2.88 .94 67 2.89 .88
9. Negative aroused 143 3.39 1.02 76 3.35 1.03 67 3.43 1.02
10. Stereotype threat 143 2.07 .87 76 2.02 .84 67 2.12 .91
11. MPPSd 144 2.04 1.11 77 2.22 1.19 67 1.84 .98
12. Feedback-seeking 143 3.47 1.12 76 3.53 1.13 67 3.41 1.11

 

Note. All scales on a 5-point Likert scale (where 1=strongly disagree and 5=strongly agree) except the
selection test, task difficulty choice, and performance. The selection test score represents number of
problems correct out of 36; Performance represents number of unique valid schedules generated. Task

difficulty choice ranges from 1 to 3, with 1 being the easiest and 3 being the most difficult. n = sample size;

M = mean; SD = standard deviation; n/a = not applicable.

'POCA is an abbreviation for “perceptions of others’ confidence in one’s own ability” t’I‘DC is an
abbreviation for task difficulty choice cParticipants were told that they were going to perform the job by
completing a work sample. dMPPS is an abbreviation for “meta-perception of preferential selection” but, due

to a programming error, more accurately refers to “the belief that one was not selected on the basis of

selection test”

86

.0000 0000200
.00 .0000 000 00 0000200 000 003 000 .000 00:00 000.. 00 000000 00005000 0000 .0000 wEEESon 0 00 0:0 .000 :0000200 000000000 00
00000000000018 0000300000 00 m0 mam—20 030000 0:03 0 300000000 >0 00: 000 00000000 00 wEow 0003 0000 0000 0.0. 0003 0000900000.
02000 9000-00 0.000 00.0 0000300000 00 0 00.0... 30:00 030 0.000 5 000000000 .0050 .00 0000000000.. 00.0 0000300000 00 00 0.000..

50:30.50 _0>0_ no. 00. 00 0002-0030. 000 n _. 00 _0000 .8 0000 000000 0000—0000 0000300 00. 00 03000 30205000 0:00:00 .0002
mm. mm.- 2. mo. 8.. 0o. S.- _N. Q. 00. mo. co. 9: w0§00000000000m .m—

0\0 .1. 00. 0o: 00.- no: wmr wmr 00.. N0; 5. 3: $002 .2

0w. 0m. of- mo.- co. m_.- 2.. N_.- 0N: 00: 92 000-0.: 09300000m .00

ow. 0m.- mm; 3.- 0o.- 0o. N0- oo. 00.- 9: 0000000 0>00w0z .o

a. 00. 8. 8. E. 2. S. 8.- 0: 0380000800
00 cm. E. 8. R. 2. S.- 0: .8000000000

00 E. 00. 0.. mm. 8. E .0000

3. 00. $.- :.. 2.- E E0000

8. e. K. 8.- E 0800-0000

00. 3. 8. 02 028002000

00. 8. E 08000002005 .0

0E fihfi 03.30 20m;

 

N0 Z 00 o w n 0 m 0 m N 0 0

 

0.000020; 0.05% 00:03.0 00020000000005 000 0020000§00U b.50000m

_ 0 200.0-

87

Individual Hypothesis Tests

To test Hypothesis 1 (which predicted a positive relationship between solo status
and stereotype threat), I conducted a one-tailed independent-samples t-test on the
stereotype threat composite score. Solo participants’ responses (M = 2.02, SD = .84)
were not significantly different than the non-solo participants (M = 2.12, SD = .91), t
(14]) = .738. p > .05. Thus, there was no support for Hypothesis 1.

To test Hypothesis 2 (which suggests a positive relationship between stereotype
threat and the meta-perception of preferential selection), I computed a correlation
between perceptions of stereotype threat and the meta-perception of preferential selection
scale. The correlation revealed that perceived stereotype threat did not significantly relate
to meta-perception of preferential selection (which, due to a programming error, more
accurately refers to “the belief that one was not selected on the basis of selection test”), r
= .14, p > .05, one tailed. Thus, there was no support for Hypothesis 2.

To test Hypothesis 3 (which suggests that the meta-perception of preferential
selection will negatively relate to cognitions, affect, and behavior), I examined the
correlations between the meta-perception of preferential selection and the following
dependent variables: task difficulty choice, feedback-seeking, positive aroused affect,
negative aroused affect, and work sample performance (i.e., score on the scheduling
task). As shown in Table 11, the meta-perception of preferential selection significantly
negatively related to feedback-seeking (r = - .33, p < .001). That is, consistent with
Hypothesis 3, the higher the meta-perception of preferential selection (which, due to a
programming error, more accurately refers to “the belief that one was not selected on the

basis of selection test”), the less likely the participants were to report wanting feedback.

88

However, the meta-perception of preferential selection did not significantly relate to the
remaining dependent variables. Thus, there was partial support for Hypothesis 3.

Hypothesis 4 suggests that that the relationship between one’s meta-perception of
preferential selection and cognitions, affect, and behavior is partially mediated by the
meta-perception of preferential selection. Hypothesis 5 suggests the same relationship is
partially mediated by POC A (one’s perceptions of others’ confidence in one’s own ability
to perform). The first step of testing each of these mediation hypotheses was to show that
the meta-perception of preferential selection (which, due to a programming error, more
accurately refers to “the belief that one was not selected on the basis of selection test”),
relates to the dependent variable of interest (Baron & Kenny, 1986). As stated
previously, the meta-perception of preferential selection significantly related only to
feedback-seeking (r = - .33, p < .001), so subsequent steps focused only on this
relationship.

The second step of testing the mediation hypotheses was to show that the
proposed mediators (i.e., self-efficacy, POCA) relate to the independent variable (i.e., the
meta-perception of preferential selection). As shown in Table 11, the meta-perception of
preferential selection significantly negatively related to self-efficacy (r = -.28, p < .05)
and to POCA (r = -.28, p < .05), so this second step was satisfied for both hypotheses.

The third step of testing the mediation hypotheses was to show that the proposed
mediators relate to the dependent variable while the independent variable was held
constant. As shown in Tables 12a and 12b, the relationships between feedback-seeking
and the two mediators (i.e., self-efficacy, POCA) were no longer significant after

controlling for the meta-perception of preferential selection. Not satisfying this third step

89

precludes one from going on to the fourth step. Thus. based on these analyses, there is no

support for Hypotheses 4 and 5.

Table 12a

The Relationship Between Self-Efficacy and F eedback-Seeking,
Controlling for MPPS”

Independent variables B Standard Error t-statistic p R2
MPPS -.33 .084 -3.88 O

Self-efficacy .004 .1 1 .40 .69 .1 1

Note. Dependent variable: Feedback-seeking; degrees of freedom = 2; sample size=14 l;
B = unstandardized beta weight; R2 = variance explained.

' MPPS is an abbreviation for “meta-perception of preferential selection”, but due to a
programming error is more accurately described as “the belief that one was not selected
on the basis of the selection test”

Table 12b

The Igilationship Between POCA" and F eedback-Seeking, C ontrollingfor
MPP

Independent variable B Standard Error t-statistic p R2
MPPS -.29 .08 -3.53 .001

POCA .22 .13 1.68 .10 .13

Note. df= 2; sample size=14 l; B = unstandardized beta weight; a = significance level; R2
= variance explained.
“POCA is an abbreviation for “perceptions of others‘ confidence in one’s own ability”

To test Hypothesis 6 (which suggests that the relationship between one’s meta-
perception of preferential selection and POCA is moderated by initial self-efficacy), it
was again necessary to conduct regression analyses. As shown in Table 13, the

interaction term was not significant, meaning there is no support for Hypothesis 6.

90

Table 13

The Interactive Effect of MPP.” and Initial Self-Ellieacy on POCA”

 

Step Independent variable B Standard Error t-statistic p R2 AR2
1 MPPS -.16 .053 -3.30 .001
Initial self-efficacy .32 .048 5.14 .000 .23
2 Interaction term .08 .048 1.71 .090 .25 .02

 

Note: df= 3; sample size = 144; B = unstandardized beta weight; a = significance level; R2 = variance
explained.

“ MPPS is an abbreviation for “meta-perception of preferential selection” but due to a programming error is
more accurately described as “the belief that one was not selected on the basis of the selection test” bPOC A
is an abbreviation for “perceptions of others’ confidence in one's own ability”

In sum, there is only very limited support for the individual hypothesis tests.
However, the overall model from which these hypotheses were derived is comprised of
complex indirect effects that involve multiple variables and for this reason it makes sense
to interpret my hypotheses within the context of the full path model (described below) to
test the component path coefficients and overall model fit (Cohen & Cohen, 1983; Kline,
1998)

Model Test

One of the benefits of a path analyses is that it allows a complete and
simultaneous test of all relationships in the model. This is particularly useful in the
present case, where I have hypothesized complex indirect effects with multiple variables.
Path analysis is a means of testing structural models with observed variables only, which
is why I first assessed the quality of my measures by conducting factor analyses (when
relevant) and reliability estimates.

Any structural equation modeling (SEM) technique, which includes path analysis,

involves six basic steps (Kline, 1998, p. 48-50). The first step was to specify the model. I

91

replicated the basic model (see Figure 4) across the seven cognitions. affect. and behavior
variables (i.e., task difficulty choice, feedback-seeking, performance on the work sample.
positive aroused affect, negative aroused affect). For example, I tested the effect of solo
status on task-difficulty choice using the path model shown in Figure 7. Similar path
models were generated for the remaining variables. It is important to note the redundancy
in this approach (because the “back half” of the models is the same).

The model was comprised of three exogenous variables and five endogenous
ones. In SEM, the term exogenous variable is used instead of “predictor” or
“independent” variable and refers to variables that do not have hypothesized causes.
Endogenous variables, on the other hand, are represented as the effects of other variables
in the model. In this way, endogenous variables are similar to a dependent variable or a
criterion but are represented as causes of other endogenous variables (Kline, 1998, p. 16).
In the present case, the exogenous variables were the experimental manipulation (solo
status), participants’ initial self-efficacy, and the interaction term (initial self-efficacy ><
meta-perception of preferential selection). (While I am including an interaction term in
the model, it should be noted that such continuous-variable interactions can be
problematic in SEM, as assumptions are made regarding the appropriateness of reliability
estimates and the presence of multivariate normality that are ofien violated by using a
product term; e.g., see Mossbruger, Schermelleh-Engel, & Klein, 1997) The five
endogenous variables were stereotype threat, the meta-perception of preferential
selection, participants’ self-efficacy, participants’ POCA, and task difficulty choice. Each
of these five endogenous variables has an error term, which is called a disturbance and is

represented here with an uppercase D. Disturbance terms represent variance in an

92

 

300000

3....-.. .000
300000-0000 .00 0000000000002

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

000000
.. 58.00.03 0000
e 8 0: e e
0003
8.8.0000 00 50000 i 00000—00 0000000
\ 00 000000000 3 00000000000 0: 0000-0 0 E 00005
0000000 50:00-00 0000- 000000 .00 000000000 0: 005%0000 Al 0500005 MWW
“0002
3

 

 

 

 

 

>0000-00-00m l E E
.Ilve m:

 

 

 

00008 00030000 00900m N 000.00%

 

93

endogenous variable that is unaccounted for by the variables specified as their cause. The
absence of a specified association between disturbances reflects the presumption of
independence of unmeasured causes (a presumption that can later be verified when
testing for model fit). This path model was recursive (as opposed to non-recursive)
because its disturbances were uncorrelated and all its causal effects are unidirectional
(Kline, 1998, p. 105).

The second step was to determine whether the model was identified. A model is
identified if it is theoretically possible to derive a unique estimate of every model
parameter (Kline, 1998, p. 49). A simple method for determining this was to make sure
that the number of parameter estimates does not exceed the degrees of freedom. In the
present case, I have 36 pieces of information ((8 X9)/2). I will be estimating 18 parameters
(10 paths, 3 variances, and 5 disturbances). Thus, my model is theoretically identified.

The third and fourth steps involve data collection and analysis. Specifically, the
third step involves selecting measures of the variables represented in the model and
collecting the data, the details of which is described above. The fourth step was to
analyze the model, which I did using AMOS 4.01 (Arbuckle & Wothke, 1999).

The fifth step was to evaluate model fit. All decisions about model fit were made
using the combined criteria suggested by Hu and Bentler (1999). They recommended two
possible combination rules, the first of which is to reject the model when RMSEA > .06,
and SRMR > .09. The second combination rule is to reject the model when TLI < .96,
and SRMR > .09. Following these suggested rules, I report the Standardized Root Mean
Residual (SRMR), the Root Mean Square Error of Approximation (RMSEA), and the

Tucker-Lewis Index (TLI) as indices of fit throughout this paper. Additionally, I report

94

and interpret the chi-square ()8) index. As shown in Table 14, the initial fit for each
model was moderate to poor. Specifically, the average SRMR was .10 (SD = .004), the
average RMSEA was .12 (SD = .01), the average TLI was .69 (SD = .03), and the
average 12 was 56.34 (SD = 5.07). Neither of the combination rules was satisfied, and all
of the x2 indices were significant. Since the data did not fit the model well, the sixth step
was to respecify it.

Table 14

Initial Fit Statistics by Dependent Variable (N = I46)
SRMR RMSEA TLI x2 df p

 

Task difficulty choice .100 .1 12 .737 50.541 18 .000
Feedback-seeking . 100 .126 .691 59.575 1 8 .000
Performance on work sample .097 .112 .724 50.655 18 .000
Positive aroused affect .101 .118 .703 54.336 18 .000
Negative aroused affect .107 .128 .659 60.781 18 .000

 

Note. SRM R = Standardized root mean residual; RMSEA = Root mean square of approximation;
TLI = Tucker-Lewis Index; CH1 = Chi-square index; DF = degrees of freedom; p = Significance of
the chi-square test.

Respecifications Decision Points Common to All Models

Three types of respecification steps were common across models. First, based on
the modification indices (which approximate the amount by which the model’s overall 12
would decrease if a particular parameter were freely estimated; Kline, 1998, p. 134) and
the fact that multiple variables were self-report data, I allowed a portion of the error terms
to covary for all models. I allowed the error terms on self-efficacy and POCA to covary
for all the models, and I allowed the error terms on the dependent variable and stereotype
threat to covary for the models with negative aroused affect and feedback-seeking as the
dependent variables. Chi-square difference tests indicated that the fit for this second

iteration of models was significantly better. As shown in Table 15, overall fit was

95

Table 15

Fit Statistics For Second Iteration OfModels By Dependent Variable

 

SRMR RMSEA TLI x2 DF p

 

Task difficulty choice .091 .071 .895 29.282 17 .031
Feedback-seeking .086 .082 .870 31.483 16 .011
Performance on work sample .088 .071 .889 29.396 17 .031
Positive aroused affect .092 .081 .861 33.077 17 .011
Negative aroused affect .087 .065 .913 25.668 16 .061

 

Note. N = 146; SRMR = Standardized root mean residual; RMSEA = Root mean square of
approximation; TLI = Tucker-Lewis Index; CH1 = Chi-square index; DF = degrees of freedom; p =
Significance of the chi-square test.

moderate. The average SRMR was .09 (SD = .002), the average RMSEA was .07 (SD =
.006), the average TLI was .88 (SD = .02), and the average 12 was 29.87 (SD = 2.44).
Although the 12 difference tests indicated that the second iteration of models fit
significantly better than the original models, neither of the combination rules was
satisfied, and all of the x2 indices were significant (except for the one for negative aroused
affect, which was marginally significant).

The second respecification decision was with respect to stereotype threat. Across
all models, the relationship between solo status and stereotype thread was non-
significant, as was the relationship between stereotype threat and the meta-perception of
preferential selection. Further, the modification indices suggested that an indirect effect
be added from solo status to the meta perception of preferential selection. Together, this

indicates that the data do not support the hypothesized inclusion of stereotype threat as a

96

mediator. Therefore, models that did not include stereotype threat were compared against
the models that did include stereotype threat. A x2 difference test was not appropriate
here (because the models were not nested), but, as shown in Table 16, the fit for this third
iteration of models improved and was moderate to good. Specifically, the average SRMR
was .072 (SD = .004), the average RMSEA was .054 (SD = .013), the average TLI was
.95 (SD = .022), and the average x2 was 15.704 (SD = 2.253). Moreover, the revised
models all had non-significant chi-square tests and they satisfied at least one of the two
combination rules (Hu & Bentler, 1999) for three of the five dependent variables (task
difficulty choice, performance on work sample, and negative aroused affect). The final
models (and their further respecifications) are described in the next section.

Table 16

Fit Statistics For Third Iteration 0f Models By Dependent Variable

SRMR RMSEA TLI 78 DF p

 

Task difficulty choice .069 .041 .972 13.669 11 0.252
Feedback-seeking .071 .064 .937 17.445 11 0.095
Performance on work sample .071 .054 .950 14.689 11 0.153
Positive aroused affect .078 .073 .912 19.456 11 0.053
Negative aroused affect .068 .035 .978 12.998 11 0.293

 

Note. N = 146; SRMR = Standardized root mean residual; RMSEA = Root mean
square of approximation; TLI = Tucker-Lewis Index; CHI = Chi-square index; DF
== degrees of freedom; p = Significance of the chi-square test.

Third, the initial self-efficacy X meta-perception of preferential selection
interaction term was marginally significant (p = .055) for all of the subsequently

discussed models. This interaction term was retained because (a) the models do not fit

97

better with it removed. and (b) it is consistent with the hypothesized model. Any
interpretations made regarding this variable are necessarily tenuous, however it is
theoretically interesting because it suggests that one’s initial self-efficacy may buffer the
negative effects of the meta-perception of preferential selection on POCA.
Unique Respect/ications by Dependent Variable

Task difficulty choice. In the third iteration of the model for task difficulty choice,
the paths from POC A and from the meta-perception of preferential selection to task
difficulty choice were not significant. Additionally, in testing a model with these path
removed, a chi square difference test indicated that model fit was not significantly worse
A38 (2) = 2.02, p > .05. Hence, the more parsimonious model (i.e., the one without the
path) was adopted (see Figure 8). This model has good fit, 7803) = 15.689, p > .05;
SRMR = .0725; RMSEA = .038; TLI = .976, but all indirect effects are non-significant.
Thus, these results suggest that bearing solo status within an applicant pool affects task
difficulty choice through its effect on the meta-perception of preferential selection and
self-efficacy and that it affects POCA through its effect on the meta-perception of
preferential selection.

Feedback-seeking. In the third iteration of the model for feedback-seeking, the
paths from POCA and from self-efficacy to feedback seeking were non-significant. First I
removed the path from self—efficacy to feedback-seeking. A 12 difference test suggested
that model fit was not significantly worse with this path removed, A752 (2) = .31, p > .05,
and thus the more parsimonious model was adopted. In this new model, the path from
POCA to feedback-seeking was marginally significant (rather than non-significant).

However, a )8 difference test indicated that model fit was not significantly worse with

98

this path removed, A36 (1) = 2.76. p > .05. and thus the more parsimonious model was
adopted (see Figure 9). This model satisfies one of the combination rules and has
moderate fit, x2 = 17.752, p > .05; SRMR = .076; RMSEA = .063; TLI = .938, but all
indirect effects are non-significant. Thus, these results suggest that bearing solo status
within an applicant pool affects feedback-seeking, self-efficacy, and POCA through its

effect on the meta-perception of preferential selection.

99

.553 E 853:3 .3050 .6 22883; n <00; ”:32 c2823 2: .8 £83
2: so 3823. 8: .83 use .2: “9:3 2:: mm contomou 38838 22: a cote wEEEEwoa a 8 26 :5 :2828 _atcefloa ho 5:383
-802 n mam—2 .8_. v 3 Emoc_:w_m 3.233.: m_ 523 .::2 530885 2: no.“ 3088 Go. v S EnocEwE 98 356508 =< .862

 

Amézaosaoaa BE:
Shop “83030::

 

 

 

 

xomowtm
m _ . hm. .bom
BEE

 

 

 

 

 

 

 

3.
86:0 <UOm

538.25 V2.5

 

 

 

 

 

 

 

 

mam: 1| 85m 2%

 

 

 

 

 

 

 

 

 

xomofiofiom

 

 

 

 

.AmeEB gamma“: BEEQEQE ESL 86:0 335% xmfi do.“ $88 53 _mcE .m «SEEK

100

 

 

.252
5:0225 u mam—zhm— 5:35 5 coconccou 22:0 do 5:380; u <UOm ”:32 22523. 2: mo v.53 2: so 3823. Ho:

83 25 :2: “0:2: 2:: mm .6358: 32838 EOE m_ 8:» wEEESon a 2 26 :5 5:023. 3:53.35 .3 2.2385980:
n mam—2 .8: v S 2:85ch x__m=_w3E m_ 523 6.22 5:252: 2: 5.: 8088 Go. v S EmoEcw_m 2m $56526 :< .262

 

afizaoeaoaa 3&5
Show E50835

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2. . 585m
R .22
4 Ea:
<oom
. E.
n-
memo—03-6333 mm: mam—2 AH 36% 2cm
3.-
585.2%

 

 

 

 

.AmemoB 2.288on “636398; 22.5 wcfioomiogvofl 8.: 3:08 :3 35.2 d Emmi

101

Work sample performance. In the third iteration of the model for work sample
performance, three paths to the dependent variable were not significant. Specifically, this
was true for the ones from the meta-perception of preferential selection, from self-
efficacy, and from POCA. Based on a series of )8 difference tests and observed changes
in path significances, I retained the path to the dependent variable from POCA and
removed the paths from the meta-perception of preferential selection and from self-
efficacy. A x2 difference test indicated that model fit was not significantly worse, A36 (2)
= 1.16, p > .05. Thus, the more parsimonious model was adopted (see Figure 10), which
has good fit, )8 = 15.846,p > .05; SRMR = .071; RMSEA = .039; TLI = .975.

Next, I ran a model where I controlled for the effect of cognitive ability (assessed
via the selection test) on initial self-efficacy and on work sample performance. Model fit
was still good, )8 = 23.056, p > .05; SRMR = .0745; RMSEA = .044; TLI = .959, but the
link between POCA and work sample performance became non-significant. Additionally,
all indirect effects are non-significant Thus, these results suggest that bearing solo status
within an applicant pool may affect performance through its effects on the meta-
perception of preferential selection and POCA, although these effects may become
insignificant once cognitive ability is accounted for. They also suggest that solo status
affects self-efficacy through its effect on the meta-perception of preferential selection.

Model fit was moderate, 78 = 21.406, p > .05; SRMR = .082; RMSEA = .067; TLI
= .924. Thus, these results suggest that bearing solo status within an applicant pool
affects positive aroused affect through its effect on the meta-perception of preferential
selection and on self-efficacy. They also suggest solo status affects POCA through its

effect on the meta-perception of preferential selection.

102

 

 

 

 

:32 .5523. 22.3 1.25
2.: :o 332% :0: mm? 25 EE.B__.5 2:: 8 3353: $2232“ 20:. m_ 5.5 wEEEEmoE a 2 as: :E .5533.

HERMES do :ocguPSQéB—z n mnE—z .Ecotoa o. .9252“ E 8:92.23 22:9,? 2.2333; n <25 .21.
v 5 Eauczfii 33592: m_ 52:3 .::2 coconfizz 2: .5: E33 $3. .. 5 ::...3_.::.u._.r. 9.2.. 1:33:33 :4. 2:2.

 

Amfiz..>.§c.a.:a 3%:
8:3 £03033:—

 

 

 

 

 

 

 

 

 

 

 

 

 

m _. .3355 1
S. 2%
4 EEE
j<UOm .
E.

 

 

 

 

c _.
u2_mE._o..._um

 

mam—2 Llll £53m 20m

 

 

 

 

:.

 

 

 

 

 

.,83_.E..2_om

 

 

 

 

 

2.1.5.1395 :BmmeE :oNEEEEm 2:3: eggsheram 3030.. 3:2: 2:3; 2:: .E 83.11;;

Positive amused affect. In the third iteration of the model for positive aroused
affect. three paths to the dependent variable were not significant. Specifically, this was
true for the ones from the meta-perception of preferential selection, from self-efficacy,
and from POC A. Based on a series of 12 difference tests and observed changes in path
significances. l retained the path to the dependent variable from self-efficacy and
removed the paths from the meta-perception of preferential selection and from POCA. A
12 difference test indicated that model fit was not significantly worse, A38 (2) = 2.05, p >
.05. Thus, the more parsimonious model was adopted (see Figure ll). Again, note that
all indirect effects are non-significant

Negative aroused affect. The results of this path analyses did not provide support
for the model with negative aroused affect for a dependent variable. The third iteration of
the model for negative aroused affect had moderate fit, but, despite various iterations, key
paths (e.g., the links between the two mediators and the relevant dependent variable)

were non-significant.

104

 

:32 .8528 2: (8 £83
2: :o 3823 Ho: 83 one :2: 3:3 2:: mm wontomon 32838 29: fl Stu wEEEEon a 2 26 :5 5523 3222305
.6 53859802 n man—2 FEB cozofiefi n mmmZLm— "Eaton 8 35% E50 Paco E 853:8 .Eofio c6 cocaoeom
“<00; x2. v 3 Emoc_cw_m bEEwEE 3 523 :52 :28885 2: c8 88x0 Go. v S anc_:w_m 0.8 356508 :< .232

 

6&23085058 3%:
:23 2030585

 

 

 

 

m _ . 58qu
2 cam
4‘ REE

 

 

 

 

 

 

 

 

an.

 

 

80+},
33:26, 03:8;

4:/

 

 

 

 

mam—2 Al 258m 20m

 

 

 

 

 

 

 

 

38:52am

 

 

 

 

 

.AmeEB 55352 nofiEmncfim 53$ 8th 33:05 @358 8» 6on 53 E25 .2 oSmE

105

Follow-up A nulys'cs

The results described thus far suggest that solo status may individually affect four
dependent variables: task difficulty choice. feedback-seeking. work sample performance.
and positive aroused affect. For ease of interpretation. I created an overall model that
incorporated each of the four aforementioned dependent variables in which I controlled
for the effects of cognitive ability on initial self-efficacy and on task difficulty choice. As
before, the relationship between POC A and performance on the work sample fell out
once cognitive ability was controlled for and the interaction term was only marginally
significant (p = .055). The final overall model (shown in Figure 12) had moderate to good
fit, )6 = 45.482, p > .05; SRMR = .0863: RMSEA = .053; TU = .924, although all
indirect effects are non-significant. lt satisfied one of the combination rules and has a
non-significant )8 index. This model is used guide the discussion that follows.

This overall, post hoc model suggests differential pathways through which
bearing solo status within an applicant pool may influence the positive aroused affect,
task difficulty choice, feedback seeking. and FCC A (i.e., perceptions of others’ belief in
one’s capability to perform) of newly hired females. Specifically, solo status negatively
relates to the first two variables through its effect on the meta-perception of preferential
selection and to the second two variables through this as well as self-efficacy. This post
hoc model also suggests that one's initial self-efficacy may buffer the negative effects of
the meta-perception of preferential selection on POCA, although one must interpret the

marginally significant interaction term with caution.

106

 

 

 

.ccoton 8 >53“ :30 Paco

5 09:09.28 22:0 t8 c2683; n <00; ”v.32 .5523. 2: :0 £me 2: cc :2923. 8: 83 25 :2: 3:3 2:: 3 3953: E29208 29:

2 St» wEEEEon m o. 26 :5 :2823 32.222: .3 :o:a8§a-2~2 u mam—2 .Go. A 3 Emo:_cw_m-co: mm 533 $0850.23 BEE 9 big
yzxcwou :5: £8 2: :5 .8: v 5 ~535ch bEEwEE 2 £023 :52 5:09.25 2: 8.: 5088 Go. v 3 280$ch 98 256508 :< $82

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

an. E3“
03::on
3&2...§2.cv,:a 3%: 3.
E22 5:882:
m _.
5m. .3850
:8 3:5
<00; t
mm: .
ch
mafim
wciocmiomnuoom mam—2 20m
2.
4
w:
866 2356 V73 _
5. 5850.:om t
Hoots :83on 032mg 4

 

 

 

 

 

 

.GEwBB 2232on 35:3:56 5:5 3:06 58 =Eo>0 .mfi 2=wE

107

DISCUSSION

The objective of the present study was to draw from three streams of research
(i.e., on solo status, stereotype threat, and preferential selection) to generate a testable
mode 1 of the relationship between bearing solo status within an applicant pool and newly
hired females’ cognitions, affect, and behavior. Such a model was developed and then
tested Via regression-based individual hypothesis tests and path analyses. Regression
results provided little to no support for major hypotheses and path analyses yielded
moderate to poor fit statistics for initial models (see Table 14). Model respecification
Yield ed models for task difficulty choice, feedback-seeking. performance. and positive
aroused affect (see Figures 7 — 10). Models had moderate to good fit, although response-
1'65130 use issues are likely to have played a role in the path coefficients that were the
highest. Moreover, the significant results are likely because the redundant analyses
capitalize on chance. Thus, it is very important that the reader interpret these results with
cauti on. The overall, post hoc model (shown in Figure 12) had moderate to good fit (x2 =
45~4 82, p > .05; SRMR = .0863; RMSEA = .053; TLI = .924) and will be used to guide

the present discussion.

Key Findings

First, bearing solo status in a selection context (i.e., within an applicant pool)
prompts one to develop the meta-perception of preferential selection. The results of
Heil l'ljan and Blader (2001; see also Heilman, 1980) suggest that co-workers assume
females bearing solo status to be preferentially selected, and the results of the present

Study suggest that the solos themselves are aware of this. Notably, Heilman and Blader

108

manipulated solo status within a newly hired cohort, while the present study did so within
an applicant pool, suggesting that the link between solo status and preferential selection
occurs in the minds of different people across different types of selection contexts. The
central issue underlying this link likely concerns the “stigma of incompetence"
hypothesized to be associated with those who are (or are assumed to be) beneficiaries of
affirm ative action (e.g., Heilman & Haynes, in press). Based on the results of l leilman et
al., such a stigma appears to “tar” salient minorities and prompt thoughts of preferential
selection in the minds of themselves and others. The present study found some support
for thi s for women with respect to the meta-perception of preferential selection.

Second, the meta-perception of preferential selection (which, due to a
Programming error, more accurately refers to “the belief that one was not selected on the
basis of selection test”) negatively related to important variables: self-efficacy, POCA
(i.e - , one’s perception of others’ confidence in one’s ability to perform), positive aroused
affec t, task difficulty choice, and feedback-seeking. This general pattern of findings is
cOTIS istent with that of Heilman and Alcott (2001), who manipulated the meta-perception
0f preferential selection and found it to negatively relate to inferred competence
eXpe (:tations, ensuing task decisions, self-regard, and the favorability of affective
react ions. The results of the present study extend the known scope of impact to include
feed back-seeking, and also suggest differential paths by which the effects of the meta-
perc eption of preferential selection occur (see Figures 7-1 1).

Third, high initial self-efficacy may help foster some “resilient” reactions to the
meta‘ perception of preferential selection. Specifically, there was a marginally significant

(p-Q

- 055) interaction between initial self-efficacy and the meta-perception of preferential

109

select ion on POCA, such that individuals with high initial self-efficacy experienced a less
negative impact on POCA (vs. those with low initial self-efficacy). However, this
“bu ff‘ering” effect in the present study did not translate into other positive effects, as
POCA did not uniquely relate to any of the outcomes examined (see Figures 7-1 1).
Unexpected Findings

First, there was no significant relationship between solo status and stereotype
threat , and all models yielded higher fit statistics without the presence of the stereotype
threat variable. Roberson et al. (2003) found a link between these two variables and
Stan gor et al. (1998) assumed one was present, so its absence in the results of the present
study is surprising. The explanation likely lies in the measure of stereotype threat, the
qual i ty of which can be improved in future research, and the timing of this measurement,
Which was at the end of the study rather than at the point at which the threat was
hypothesized to occur. To reduce common method bias and the risk of alerting
part i Cipants to the nature of the study, it would be helpful if future self-report measures of
the: C onstruct were phrased less directly and/or if different (e.g., observational) methods
were developed.

Second, there was no significant relationship between POCA and the outcome
vari ables of interest (e.g., task difficulty choice, feedback-seeking, etc.). This absence is
unex pected because the proposed link is well-supported by the literature. In fact, each
stream of research reviewed here (i.e., on solo status, stereotype threat, and preferential
seleQ tion) suggested, in one form or another, that people’s perception of others’
eval nation of them would play a mediating role in the relationship between solo status

and the cognitions, affect, and behavior of the solos themselves. Moreover, this is

110

consi stem with the literatures on self-fulfilling prophecies (Merton, 1948) and, related, on
the Pygmalion effect (Eden, 2003). The most likely explanation for this lies in the
meas ure of FCC A, a point to which I return to later.

Third, there was no significant relationship between self-efficacy and
performance, which is unexpected because these two variables have been shown
repeatedly to be associated with one another (Stajkovie & Luthans, 1998). The simplest
explanation for its absence is range restriction, since approximately half of the
partic: ipants scored a zero on the performance measure. This floor effect of the
perfO mance measure is discussed in the next section.

Study Limitations

The strengths and weaknesses of laboratory experiments have been discussed
elsewhere (e.g., Kerlinger, 1992, pp. 367-369) and will not be repeated here. However,
the artificiality of any experimental research situation (including this one) is certainly
worth noting as well as the related lack of external validity. Moreover, the construct

Validity of some of the measures used is questionable (e.g., POCA, Stereotype Threat), a
POil’lt to which I return when highlighting directions for future research. Here, I describe
threQ key problems with the present study as well as their implications for interpreting its
resu 1 ts.

First, it is unclear whether participants perceived the job to be “masculine.” A
meEiSLJAre of this would have been helpful in interpreting the non-significant link between
5010 Status and stereotype threat (r = -.O6, p > .05), because, for example, I could examine
WhetI‘Aer the perceived masculinity acts as a moderator. As shown in Table 17,

part1 Q ipants do report being aware of stereotypes about women having less ability than

Ill

men i n traditionally masculine jobs and they do also report believing that the managers

who veould be evaluating them were aware of such stereotypes. This coupled with

evidc :1 cc that management is associated with men (e.g., Powell et al., 2002) provides

SuppO rts for the idea that study participants perceived the managerial internship to be at

least a somewhat masculine job, but, without any data, this unfortunately remains an

untested assumption. If it is a false one, this may explain the small magnitude (Cohen,

1988 ) ofthe bivariate correlations that include solo status (see Table 11).

Table 17

L "A‘onal Awareness of and Perception of Managers ’ Awareness of Stereotype

 

x Item n M SD
1 - 'I‘here is a stereotype about women having less ability than men 144 4.10 .88
i n traditionally masculine jobs
2 - There is a stereotype about women being less likely to succeed

144 3.87 1.10
th an men in traditionally masculine jobs
3 - Managers at this organization think there is a stereotype about

145 3.39 1.08
Women having less ability than men in traditionally masculine jobs
4 - Managers at this organization think there is a stereotype about
Women being less likely to succeed than men in traditionally 143 3.41 1.10

rrlzisculinejobs
N\

l) 04:8. All items were placed at the end of the study, began with phrase "Regardless of my personal
e l l efs. . .” and used a 5-point Likert scale where l = strongly disagree and 5 = strongly agree.

Second, the performance measure had a serious floor effect, and thus it is possible

tbelt participants' realization of their poor performance affected their response to

112

subsequent measures (e.g. affect. stereotype threat). One possible explanation for this
lloc3r effect is that the interface was somewhat clumsy. requiring that participants
su'i tch between multiple screens on their computer in order to complete the work
53:11 ple. While this mimics how individuals manage many paper-and-pencil tasks, it
it) a 3.! not have transferred well into an online format (see Appendix C). Another possible
exp 1 anation is the relatively low ($10) incentive offered for high performance.
Ad (1 itionally. participants were informed that they would be mailed payment at a later
dat e - These factors together — the low incentive, the interface, and the anticipated delay
in receiving the incentive by mail at a later date, might not have sufficiently motivated
Participants.

Additionally. it is possible that the self-efficacy measure that preceded the task set
unrealistically high performance expectations, making them particularly susceptible to
experiencing frustration and withdrawing effort. The final item of the self-efficacy
me asure read, “I will be able to complete 16 schedules,” a level of performance that is
high 1y unlikely based on previous research (e.g., Carr, 2001; Q.M. Roberson et al., 1999)
but _perhaps became a goal for those who provided a response greater than or equal to
“He LJtIal” (approximately 40%). Currently, applied researchers measure confidence in
one ’ s ability to perform a task in a variety of ways (e.g., see Donovan & Williams, 2003;
Phi l lips, Hollenbeck, & Ilgen, 1996; Schmidt et al., 2004). Perhaps a fruitful area of
futul‘e research is to develop appropriately anchored measures of self-efficacy (see
KaI-_11’1eman, 2003; Schwarz, 1999).

Finally, I failed to record responses from the majority (92%) of respondents for

o . . . . .
ne of the two items used to measure the meta-perception of preferential selection. This

113

limitation is signilicant because the one remaining item leaves open the question of
reliability. Further. as noted throughout the manuscript. one could argue that the results of
the present study do not suggest that solo status prompts the meta-perception of
preferential selection but rather that it prompts the perception that the reason for hire is
anything other than perfomiance on the selection test. Future research with a more fully
operational measure will help better understand the relationship between the meta-
perception of preferential selection and the other variables of interest.
Directions for Future Research

Future examinations of solo status. stereotype threat, and preferential selection
should focus less on the environmental condition of solo status, which has variable
contexts (e.g.. work group. applicant pool) and perspectives (e.g., co-workers, solos
themselves). and more on the psychologically meaningful variables of salience,
distinctiveness, and being “out of the ordinary” (Neimann & Dovidio, 1998; Heilman &
Blader, 2001). Doing so would facilitate connection with other literatures, such as
judgment-and-decision-making (e.g.. Kahneman, 2003), organizational justice (e.g., van
den Bos & Lind. 2002), and positive organizational scholarship (e.g., Cameron, Dutton,
& Quinn, 2003), which among other things might help clarify potentially informative
relationships amongst uncertainty. heuristic processing, stereotyping, identity, and the
meaning of work.

Additionally, while there has been a burgeoning interest in constructs related to
self-evaluation and others’ evaluation of ourselves, these constructs and their associated
measures need to be better validated before developing a theory that encompasses them.

Such investigations might include more qualitative and observational methods of data

114

collection. The present study may best illustrate the need for this type of research. as the
questionable construct validity of key measures (e.g.. stereotype threat. POCA) as well
the heavy reliance on self-report data made it difficult to disentangle common method
bias and affective reactions from more meaningful relationships. Useful research to
consult for such efforts may be that on social identity (Brewer. I991; Tajfel & Turner.
1986) and on the self-protective. attributional and emotional response patterns associated
with processing feedback (see C rocker & Major. I989: Fridja, 1999: Weiner. Russell. &
Lerman, 1978; see also Kluger & DeNisi. I996: llgen & Davis. 2001).

Finally, future research on solo status, stereotype threat. and preferential selection
would benefit from broadening the content and level of variables of interest to include
those related to social capital - the existence of positive interpersonal relationships
among individuals (Baker, 2000; Coleman. 1998). Much of the research reviewed here
suggests the damaging effects of solo status are in part because of others" unfavorable
reactions to the solo (perceived or feared). Perhaps the existence of just a few high
quality connections (i.e.. ones that are flexible, strong. and resilient; Dutton & Heaphy,
2003, p. 263) would ameliorate any negative effects of solo status. For example, such
connections may open key channels of communication so that solos (as well as other non-
solos) have access to power structures and system knowledge and create mechanisms for
sponsorship and peer alliances (Kanter, 1977; Ragins, 1991). Given recent research
suggesting that gender stereotypes can prompt bias in evaluative judgments of women
even when these women have proved themselves to be successful and demonstrated their

competence (Heilman, Wallen, Fuchs, & Tamkins. 2004), the cultivation of a few high

115

quality connections might facilitate success in nontraditional areas even if only by
keeping others’ evaluation in check.
Concluding Comment

The good news is that there are an increasing number of women who are gaining
employment in jobs historically occupied by men (e.g.. high status management
positions; Federal Glass Ceiling Commission. 1995; Powell. et al. 2002). The had new is
that, in such cases, if these women are the only female in an otherwise male applicant
pool, they may have the meta-perception of preferential selection. However. as steps are
taken to ameliorate the perception of preferential selection and to support solos both in
terms of personal growth and upward movement. hopefully the landscape will continue to
change such that women and men are more equally distributed across high-status jobs
(see Eagly & Steffen, 1984; Federal Glass Ceiling Commission. 1995; Powell & Graves.

2003).

116

Appendices

117

Appendix A

Measures
Self-efficacy

I. I will be able to complete 1 schedule.
2. I will be able to complete 2 schedules.
3. I will be able to complete 3 schedules.
4. I will be able to complete 4 schedules.
5. I will be able to complete 5 schedules.
6. I will be able to complete 6 schedules.
7. I will be able to complete 7 schedules.
8. I will be able to complete 8 schedules.
9. I will be able to complete 9 schedules.
10. I will be able to complete 10 schedules.
1]. Iwill be able to complete 11 schedules.
12. I will be able to complete 12 schedules.
13. I will be able to complete 13 schedules.
14. I will be able to complete 14 schedules.
15. I will be able to complete 15 schedules.
I6. I will be able to complete 16 schedules.

Perceived face validity of the selection test

I
2.
3.
4

The content of the selection test reflected the content of the job.

The actual content of the test clearly related to the job.

The content of the selection test seemed predictive of future performance on the job.
I am confident that the test predicts how well people will perform on the job.

Stereotype threat measure used by Roberson et al., 2003

MAWNe

Some people feel I have less ability to perform my job because of my race.
My job may be easier to perform for people of my race [Reverse-coded].
Other people expect me to perform poorly because of my race.

In my organization, people of my race often face biased evaluations.

My race does not affect people’s perceptions of my ability to perform my job [Reverse-
coded].

Stereotype threat measure used in the present study

I.

MAW

This organization’s managers will think I have less ability to perform this job because of
my sex.

This organization’s managers will think my job is easier for people of my sex to perform
[Reverse-coded but dropped from analyses].

This organization’s managers will expect me to perform poorly because of my sex.
People of my sex will face biased evaluations from this organization’s managers.

My sex will not affect managers’ perceptions of my ability to perform this job [Reverse-
coded].

118

Appendix B

Attrition Analyses

I conducted two series of analyses in an attempt to determine whether participants’ attrition was
systematic. The first set compared participants who satisfactorily attempted to complete the full
study (n = 171) with those who did not (n = 102). Independent samples t-tests were conducted on
responses to demographics and major study variables on which data were collected from all
participants. As shown in the table below, the participants who provided useable Phase Two data
were significantly older (d = .31) and performed significantly better on the selection test (d = .53)
than non-Phase Two respondents. Not shown in the table, the proportion of racial group
membership did not significantly differ between the two samples for any of the minority
subgroups (p’s > .05, two-tailed). However, the proportion of whites in the phase two sample
(86%) significantly differed from that proportion in the remainder (68%; ((271) = -3.60, p <
.001). An alternative expression of this is that fewer total minorities were included in the Phase
Two sample than the non-Phase Two sample.

 

 

 

Phase two sample Non-phase two sample
Variable n M SD n M SD t (if p
Age 171 20.86 3.43 95 19.99 1.87 -2.30 264 .022
S010 171 0.53 0.50 100 0.56 0.50 0.56 269 .578
Campus 171 1.54 0.50 102 1.55 0.50 0.29 271 .772
Initial self-efficacy 171 3.3 0.89 54 3.21 0.74 -0.68 223 .498
Selection test 165 18.16 6.35 39 14.7 6.79 -3.05 202 .003

The second set of analyses utilized only those participants who satisfactorily attempted to
complete Phase Two of the study (n = 171). These analyses compared participants who were able
to submit and proceed past their work sample with those who were not able to do so. Independent
samples t-tests compared these two groups on self-efficacy, POCA (i.e., perceptions of others’
confidence in one’s own ability), and task difficulty choice in addition to the variables assessed in
the first set of analyses. As shown in the table below, those who completed the work sample had
selected a more difficult task than those who did not (d = .49). There were no significant
differences in attrition by racial subgroups (p’s > .05, two-tailed).

 

 

Work sample Remainder

Variable n M SD n M SD t df p

Age 145 20.81 3.39 25 21.12 3.68 0.41 168 .681
Solo 145 0.53 0.50 25 0.52 0.51 -0.10 168 .919
Campus 145 1.52 0.50 25 1.60 0.50 0.70 168 .485
Initial self-efficacy 145 3.28 0.85 25 3.42 1.12 0.75 168 .455
Selection test 140 18.34 6.35 24 17.13 6.40 -0.87 162 .387
Self-efficacy 145 3.46 0.84 25 3.70 0.97 1.29 168 .200
POCA 145 3.91 0.56 25 4.02 0.53 0.90 168 .368

Taskdifficulty 145 1.81 0.67 24 1.50 0.59 -2.17 167 .031

119

Appendix C

Detailed Description of the Task

The communication and technology organization is interested in investigating the quality
and reliability of their suppliers. For this reason, they want you to develop as many
unique inventory schedules as possible, with each schedule containing five different
suppliers within the same month.

Instructions

Here is a link to your twelve suppliers and their possible inventory times (each with a
different code) [See below]. Also, note the following schedule constraints:

1.

2.

A completed inventory schedule will include the Supplier Code and the Inventory
Time Code

Each inventory schedule must have 5 different suppliers schedule in the same
month

Each inventory schedule must be unique -- that is., it cannot duplicate another
inventory schedule.

Three of the suppliers have dual locations (Dallas/Irving, Los Angeles/El
Segundo, Tampa/Ybor City). When you order from one of these suppliers, you
must list both locations and you must list them separately.

Example of a correct schedule

 

 

 

 

 

 

 

SUPPLIER SUPPLIER INVENTORY INVTEfJEORY
NAME CODE TIME CODE
Birmingham BIR April 1-4 A
Atlanta ATL April 8-13 B
Tampa: Ybor City TAM-Y April 16-18 L
Tampa TAM April 20-24 D
Seattle SEA April 29-30 J

 

 

 

 

 

The schedule contains the Supplier Code and the Inventory Time Code

This schedule contains non-overlapping dates.

The schedule has 5 different suppliers in the same month

This schedule contains an order from one of the suppliers with dual locations
(Tampa/Ybor City) and therefore correctly lists both of these suppliers and lists
them separately.

120

Examples of incorrect schedules

 

 

 

 

 

 

 

 

 

 

SUPPLIER SUPPLIER INVENTORY INZESEfRY
NAME CODE TIME CODE
Birmingham BIR April 1-4 A
Boston BOS April 1-2 A
Atlanta ATL April 8-14 B
Tampa: Ybor City TAM-Y April 16—18 L
Tampa TAM April 20-24 D

 

 

- The schedule contains the Supplier Code and the Inventory Time Code

This schedule does not contain non-overlapping dates.

The schedule has 5 different suppliers in the same month

This schedule contains an order from one of the suppliers with dual locations
(Tampa/Ybor City) and therefore correctly lists both of these suppliers and lists
them separately.

 

 

 

 

 

 

 

 

 

 

SUPPLIER SUPPLIER INVENTORY qugngRY
NAME CODE TIME CODE
Birmingham BIR April 1-4 A
Birmingham BIR April 10-13 C
Tampa: Ybor City TAM-Y April 16—18 L
Tampa TAM April 20—24 D
Seattle SEA April 29-30 J

 

 

The schedule contains the Supplier Code and the Inventory Time Code

This schedule contains non-overlapping dates.

The schedule DOES NOT HA VE 5 different suppliers in the same month
This schedule contains an order from one of the suppliers with dual locations
(Tampa/Ybor City) and therefore correctly lists both of these suppliers and lists
them separately.

 

 

 

 

 

 

 

 

 

 

 

SUPPLIER SUPPLIER INVENTORY INXESESRY
NAME CODE TIME CODE
Birmingham BIR June 1-4 A
Atlanta ATL April 8-13 B
Tampa: Ybor City TAM-Y April 16-18 L
Tampa TAM April 20-24 D
Seattle SEA April 29—30 J

 

The schedule contains the Supplier Code and the Inventory Time Code

This schedule contains non-overlapping dates.

The schedule DOES NOT HA VE 5 different suppliers in the same month
This schedule contains an order from one of the suppliers with dual locations
(Tampa/Ybor City) and therefore correctly lists both of these suppliers and lists
them separately.

121

 

 

 

 

 

 

 

 

 

 

 

SUPPLIER SUPPLIER INVENTORY INYrEINIEEORY
NAME CODE TIME CODE
Birmingham BIR April 1-4 A
Atlanta ATL April 8-13 B
Boston 808 April 16—18 F
Tampa TAM April 20-24 D
Seattle SEA April 29-30 J

 

- The schedule contains the Supplier Code and the Inventory Time Code

- This schedule contains non-overlapping dates.

- The schedule has 5 different suppliers in the same month

- This schedule contains an order from one of the suppliers with dual locations
( T ampa/Ybor City) and therefore INC ORRE C TL Y lists only ONE of these

suppliers.

122

CODING INFORMATION

 

 

 

 

 

 

 

 

 

 

 

SUPPLIER SUPPLIER TIME TIME TIME

LOCATION CODE (MONTH) (DATE) CODE
Atlanta ATL APR or JUN 8—13 B
Atlanta ATL APR or JUN 16-21 C
Atlanta ATL APR or JUN 22—27 D
Atlanta ATL MAY or JUL 1—6 B
Atlanta ATL MAY or JUL 8—13 F
Atlanta ATL MAY or JUL 14—20 G
Atlanta ATL MAY or JUL 20—26 H
Atlanta ATL AUG 1—11 I
Atlanta ATL AUG 12-13 J
Birmingham BIR APR or JUN 1-4 A
Birmingham BIR APR or JUN 5-8 B
Birmingham BIR APR or JUN 10—13 C
Birmingham BIR MAY 20—24 I
Birmingham BIR MAY 25—28 J
Birmingham BIR MAY or JUL 1—4 E
Birmingham BIR MAY or JUL 5-8 F
Birmingham BIR MAY or JUL 10—13 G
Birmingham BIR MAY or JUL 16—20 H
Boston BOS APR, JUN, or AUG 1—2 A
Boston BOS APR, JUN, or AUG 3-6 B
Boston BOS APR, JUN, or AUG 7—8 C
Boston BOS APR, JUN, or AUG 10-11 D
Boston BOS APR, JUN, or AUG 12—15 E
Boston BOS APR, JUN, or AUG 16-18 F
Boston BOS APR, JUN, or AUG 19—21 G
Boston BOS APR, JUN, or AUG 22-24 H
Boston 803 APR, JUN, or AUG 25—27 I
Boston 808 JUN 29—30 J
Chicago CHI APR, JUN, or AUG 1-2 A
Chicago CHI APR, JUN, or AUG 3—6 B
Chicago CHI APR, JUN, or AUG 7—8 C
Chicago CHI APR, JUN, or AUG 10—11 D
Chicago CHI APR, JUN, or AUG 12-15 E
Chicago CHI APR, JUN, or AUG 16-18 F
Chicago CHI APR, JUN, or AUG 19-21 G
Chicago CHI APR, JUN, or AUG 22—24 H
Chicago CHI APR, JUN, or AUG 25—27 I
Chicago CHI JUN 29-30 J
Dallas DAL APR or JUN 5-8 B
Dallas DAL APR or JUN 10-13 C
Dallas DAL APR or JUN 14—18 D

123

 

 

 

 

 

 

 

 

 

 

 

Dallas DAL APR or JUN 19—23 E
Dallas DAL APR or JUN 24—27 F
Dallas DAL MAY or JUL 5—8 G
Dallas DAL MAY or JUL lO—l3 H
Dallas DAL MAY or JUL 14—18 I
Dallas DAL MAY or JUL l9—23 J
Dallas: Irving DAL—I JUN 3—6 K
Dallas: Irving DAL-I JUN 16—18 L
Dallas: Irving DAL—I JUL 3-6 M
Dallas: Irving DAL—I JUL 16—18 N
Detroit DET APR, JUN, or AUG l-2 A
Detroit DET APR, JUN, or AUG 3—6 B
Detroit DET APR, JUN, or AUG 7—8 C
Detroit DET APR, JUN, or AUG 10—11 D
Detroit DET APR, JUN, or AUG 12—15 E
Detroit DET APR, JUN, or AUG 16-18 F
Detroit DET APR, JUN, or AUG 22—24 G
Detroit DET APR, JUN, or AUG 25—27 H
Detroit DET APR, JUN, or AUG 28—29 I
Detroit DET JUL 29—30 J
Kansas City KAN APR or JUN 1—4 A
Kansas City KAN APR or JUN 5—8 B
Kansas City KAN APR or JUN 10—13 C
Kansas City KAN APR or JUN 14—18 D
Kansas City KAN APR or JUN 19-23 E
Kansas City KAN MAY or JUL 1—4 F
Kansas City KAN MAY or JUL 5—8 G
Kansas City KAN MAY or JUL lO—13 H
Kansas City KAN MAY or JUL 14—18 I
Kansas City KAN MAY or JUL 19—23 J
Los Angeles LOS APR 1—6 A
Los Angeles LOS MAY 16—21 B
Los Angeles LOS JUN 1-6 C
Los Angeles LOS JUL 16—21 D
Los Angeles LOS AUG 1—6 F
Los Angeles:
El Segundo LOS—E APR 3—6 G
Los Angeles:
El Segundo LOS-E APR 7—8 H
Los Angeles:
El Segundo LOS—E MAY 1—2 K
Los Angeles:
El Segundo LOS—E MAY 3-6 L
Los Angeles:
El Segundo LOS—E JUN 7-8 J
Los Angeles:
El Segundo LOS—E JUL 3—6 M

124

 

Los Angeles:

 

 

 

 

 

 

 

El Segundo LOS—E JUL 7—8 N
Los Angeles:

El Segundo LOS—E AUG 7—8 0
Los Angeles:

El Segundo LOS—E AUG lO—ll P
New York NEW APR or JUN 1-4 A
New York NEW APR or JUN 5~8 B
New York NEW APR or JUN 10-13 C
New York NEW APR or JUN l7—21 D
New York NEW APR or JUN 22-26 E
New York NEW MAY 29—30 J
New York NEW MAY or JUL 3—8 F
New York NEW MAY or JUL 8—11 G
New York NEW MAY or JUL 12—17 H
New York NEW MAY or JUL 17—21 I
Seattle SEA APR, JUN, or AUG 1—2 A
Seattle SEA APR, JUN, or AUG 3-6 B
Seattle SEA APR, JUN, or AUG 7—8 C
Seattle SEA APR, JUN, or AUG lO—ll D
Seattle SEA APR, JUN, or AUG 12—15 E
Seattle SEA APR, JUN, or AUG 16—18 F
Seattle SEA APR, JUN, or AUG 19—21 G
Seattle SEA APR, JUN, or AUG 22—24 H
Seattle SEA APR, JUN, or AUG 25—27 I
Seattle SEA APR 29—30 J

Tampa TAM APR or JUN 8—11 B
Tampa TAM APR or JUN 16—20 C
Tampa TAM APR or JUN 20-24 D
Tampa TAM MAY 29—30 J
Tampa TAM MAY or JUL 5—8 E
Tampa TAM MAY or JUL 10-13 F
Tampa TAM MAY or JUL 14—18 G
Tampa TAM MAY or JUL 19-23 H
Tampa TAM MAY or JUL 25—28 I
Tampa: Ybor
City TAM—Y APR 3—6 K
Tampa: Ybor
City TAM-Y APR 16—18 L
Tampa: Ybor
City TAM—Y MAY 3-6 M
Tampa: Ybor
City TAM-Y MAY‘ 16—18 N
Tuscon TUS APR or JUN 1—4 A
Tuscon TUS APR or JUN 5—8 B
Tuscon TUS APR or JUN lO—l3 C
Tuscon TUS APR or JUN 14-18 D

 

125

 

 

 

 

Tuscon
Tuscon
Tuscon
Tuscon
Tuscon
Tuscon

 

TUS
TUS
TUS
TUS
TUS
TUS

 

APR or JUN
MAY or JUL
MAY or JUL
MAY or JUL
MAY or JUL
JUL

 

 

[TI

QHIO'T:

 

126

 

REFERENCES

Alexander, V. D., & Thoits, P. A. (1985). Token achievement: An examination of

proportional representation and performance outcomes. Social Forces. 64. 332-
340.

Allport, G. W. (1954). The nature of prejudice. Cambridge, MA: Addison-Wesley.

Ambady, N., Paik, S.K., Steele, J ., Owen-Smith, A., & Mitchell, J .P. (2003). Deflecting
negative self-relevant stereotype activation: The effects of individuation. Journal
of Experimental Social Psychology, 40, 401-408.

Amott, T., & Matthaei, J. (1996). Race, gender, & work: A multi-cultural economic
history of women in the United States. South End Press: Boston, MA.

Arbuckle, J. & Wothke, W. (1999). AMOS 4 user’s reference guide. Chicago:
Smallwaters Corporation.

Aronson, J ., Lustina, M. J ., Good, C., Keough, K., Steele, C. M., & Brown, J. (1999).
When White men can't do math: Necessary and sufficient factors in stereotype
threat. Journal of Experimental Social Psychology, 35, 29-46.

Ashmore, R. D. & Del Boca, F. K. (1981). Conceptual approaches to stereotypes and
stereotyping. In D. L. Hamilton (Ed.), Cognitive processes in stereotyping and
intergroup behavior, pp. 1—36. Hillsdale, NJ: Erlbaum.

Bagozzi, R. (2003). Positive and negative emotions in organizations. In K.S. Cameron,
J .E. Dutton, & R.E. Quinn (Eds), Positive organizational scholarship:
Foundations of a new discipline, pp. 176-193. San Francisco: Berrett-Koehler
Publishers, Inc.

Brigham, J. C. (1971). Ethnic stereotypes. Psychological Bulletin, 76, 15-33.

Broverman, I. K., Vogel, S. R., Broverman, D. M., Clarkson, F. E., & Rosenkrantz, PS.
(1972). Sex role stereotypes: A current appraisal. Journal of Social Issues, 28, 59-
78.

Brown, R.P. & Pinel, EC. (2003). Stigma on my mind: Individual differences in the

experience of stereotype threat. Journal of Experimental Social Psychology, 39,
626-633.

Burke, R]. (1996). Work experiences, stress, and health among managerial and
professional women. In M.J. Schabracq, J .A.M. Winnubst, & C.L. Cooper (Eds),
Handbook of Work and Health Psychology. New York

127

Cadinu, M., Maass, A., Frigerio, S., Impagliazzo, L., & Latinotti, S. (2003). Stereotype
threat: The effect of expectancy on performance. European Journal of Social
Psychology, 33, 267-285.

Carr, J .2. (2001). Investigating the effects of preferential selection on self-perceptions
from an organizational justice perspective. Unpublished master’s thesis, Michigan
State University, East Lansing, Michigan.

Carr, J 2., & DeShon, RP. (2002, April). Empirical validation of a process model of
goal orientation. In V. J. F ortunato (Chair), An examination of the motivational
consequences of goal orientation. Symposium conducted at the annual meeting of
the Society for Industrial and Organizational Psychology, Toronto, Canada.

Carroll, J. B. (1993). Human cognitive abilities: A survey of factor-analytic studies.
Cambridge, England: Cambridge University Press.

Chandler, M. (1999, October 4). Secrets of the SAT [Television series episode]. In M.
Sullivan (Executive Producer), Frontline. Boston: WGBH

Cohen, J. & Cohen, P. (1983). Applied multiple regression/correlation analysis for the
behavioral sciences. Hillsdale, N.J.: L. Erlbaum Associates.

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd edition).
Hillsdale, NJ: Erlbaum.

Cohen, L. L., & Swim, J. K. (1995). The differential impact of gender ratios on women
and men: Tokenism, self-confidence, and expectations. Personality and Social
Psychology Bulletin, 21, 876-884.

Craig, K. M., & Feasel, K. E. (1998). Do solo arrangements lead to attributions of
tokenism? Perceptions of selection criteria and task assignments to race and
gender solos. Journal of Applied Social Psychology, 28, 1810-1836.

Crocker, J. & Major, B. (1989). Social stigma and self-esteem: The self-protective
properties of stigma. Psychological Review, 96, 608-630.

Crocker, J., & McGraw, K. M. (1984). What's good for the goose is not good for the
gander: Solo status as an obstacle to occupational achievement for males and
females. American Behavioral Scientist, 27, 357-369.

C rocker, J ., Major, B., & Steele, C. (1998). Social stigma. In D.T. Gilbert & S.T. Fiske
(Eds), The handbook of social psychology, Vol. 2 (4th ed. ), pp. 504-553. New
York: McGraw-Hill.

Cronbach, L. J ., & Meehl, P. E. (1955). Construct validity in psychological tests.
Psychological Bulletin, 52, 281-302.

128

 

Davis, C. A. (2001). Reactions to performance feedback: The role of goal orientation and
self-regulatory focus. Unpublished master's thesis, Michigan State University.

Davis. P.G., Spencer. S.J., Quinn, D.M. & Gerhardstein, R. (2002). Consuming images:
How television commercials that elicit stereotype threat can restrain women

academically and professionally. Personality and Social Psychology Bulletin, 28.
1615-1628.

Deaux, K., & LaFrance, M. (1998). Gender. In D. T. Gilbert, S. T. Fiske, & G. Lindzey
(Eds), The handbook of social psychology (4th ed., Vol. 1, pp. 788-827). Boston:
McGraw-Hill.

Dipboye, R. (1985). Some neglected variables in research on discrimination in appraisals.
Academy of Management Review, 10, 1 16-27.

Donovan, J. & Williams, K.J. (2003). Missing the mark: Effects of time and causal

attributions on goal revision in response to goal-performance discrepancies.
Journal of Applied Psychology. 88, 379-390.

Earley, P. C. & Kanfer, R. (1985). The influence of component participation and role
models on goal acceptance, goal satisfaction, and performance. Organizational
Performance and Human Decision Processes, 36, 378-390.

Eccles, J. S. (1994). Understanding women's educational and occupational choices:
Applying the Eccles et a1. model of achievement-related choices. Psychology of
Women Quarterly, 18, 585-609.

Eden D. (2003). Self-fulfilling prophecies in organizations. In J. Greenberg (Ed.),
Organizational Behavior: The State of the Science (2nd ed., pp. 91-122).
Mahwah, NJ: Lawrence Erlbaum.

Federal Glass Ceiling Commission. (1995). Good for business: Making full use of the
nation ’s human capital. Washington, DC: Government Printing Office.

Fiske, S. T. (1998). Stereotyping, prejudice, and discrimination. In D. T. Gilbert, S. T.
Fiske, & G. Lindzey (Eds), The handbook of social psychology (4th ed., Vol. 2,
pp. 357—41 1). Boston: McGraw—Hill.

F iske, S. T., & Taylor, S. E. (1991). Social cognition (2nd ed. ). New York: McGraw-
Hill.

F loge, L., & Merrill, D. M. (1986). Tokenism reconsidered: Male nurses and female
physicians in a hospital setting. Social Forces, 64, 925-947.

Fontaine, D. C., & Greenlee, S. P. (1993). Black women: Double solos in the workplace.
Western Journal of Black Studies, 1 7, 121 -125.

129

Ford. Ferguson. Brooks. & Hagadone (2004). Coping sense of humor reduces effects of
stereotype threat on women’s math performance. Personality and Social
Psychology Bulletin, 30. 643-653.

Frijda. NH. (1999). Well-being: The/aundations of hedonic psychology. New York:
Russell Sage Foundation.

Fuegen. K.. & Biemat. M. (2002). Reexamining the effects of solo status for women and
men. Personality & Social Psychology Bulletin, 28, 913-925.

Gilliland. SW. (1993). The perceived fairness of selection systems: An organizational
justice perspective. Academy of Management Review, 18, 694-734.

Goffman, I. (1963). Stigma. New York: Simon & Shuster, Inc.

Gonzales, P.M., Blanton, H. & Williams, K.J. (2002). The effects of stereotype threat and
double-minority status on the test performance of Latino women. Personality &
Social Psychology Bulletin, 28, 659-670.

Greenwald. A. G., & Banaji, M. R. (1995). Implicit social cognition: Attitudes, self-
esteem, and stereotypes. Psychological Review, 102, 4-27.

Hamilton, D. L. & Trolier, T. K. (1986). Stereotypes and stereotyping: An overview of
the cognitive approach. In J. F. Dovidio & S. L. Gaertner (Eds), Prejudice,
discrimination, and racism (pp. 127—163). Orlando, FL: Academic Press.

Heikes, E. J. (1991). When men are the minority: The case of men in nursing.
Sociological Quarterly, 32, 389-401.

Heilman, M. E. (1980). The impact of situational factors on personnel decisions
concerning women: Varying the sex composition of the applicant pool.
Organizational Behavior & Human Decision Processes, 26, 386-395.

Heilman, M. E. (1983). Sex bias in work settings: The Lack of Fit model. Research in
Organizational Behavior, 5, 269-298.

Heilman, M. E. (2001). Description and prescription: How gender stereotypes prevent
women's ascent up the organizational ladder. Journal of Social Issues, 5 7, 657-
674.

Heilman, M. E., & Alcott, V. B. (2001). What I think you think of me: Women's
reactions to being viewed as beneficiaries of preferential selection. Journal of
Applied Psychology, 86, 574-582.

Heilman, M. E., & Blader, S. L. (2001). Assuming preferential selection when the

admissions policy is unknown: The effects of gender rarity. Journal of Applied
Psychology, 86, 188-193.

130

 

Heilman. M.E.. & Haynes. M.C. (in press). Combating organizational discrimination:
Some unintended consequences. In A. Colella and R. Dipboye (Eds), The
Psychological and Organizational Bases of Discrimination at Work. Jossey-Bass.

Heilman. ME. Block. C .J. & Stathatos, P. (1997). The affirmative action stigma of
incompetence: Effects of performance information ambiguity. Academy of
Management Journal. 40. 603-625

Heilman. M.F... Wallen. A.S.. Fuchs. D., & Tampkins, M.M. (2004). Penalties for
success: Reactions to women who succeed at male gender-typed tasks. Journal of
Applied Psychologv. 89. 416-427.

Hu. L.. & Bentler. P. M. (1999). Cutoff criteria for fit indexes in covariance structure
analysis: Conventional criteria versus new alternatives. Structural Equation
Modeling. 6. 1-55.

Hyers. L. L.. & Swim. J. K. (1998). A comparison of the experiences of dominant and
minority group members during an intergroup encounter. Group Processes &
lntergroup Relations. 1, 143-163.

lnzlicht. M., & Ben-Zeev. T. (2000). A threatening intellectual environment: Why
females are susceptible to experiencing problem-solving deficits in the presence

of males Psychological Science, 11, 365-371.

Jussim, L. (1986). Self-fulfilling prophecies: A theoretical and integrative review.
Psychological Review, 93. 429-445.

Kahneman, D. (2003). Social experiences of research. American Psychologist, 58, 697-
720.

Kahneman, D., Slovic, P., & Tversky, A. (Eds) (1982). Judgment Under Uncertainty:
Heuristics and Biases. New York: Cambridge University Press.

Kanter, R. M. (1977a). Men and women of the corporation. New York: Basic Books.

Kanter, R. M. (1977b). Some effects of proportions on group life: Skewed sex ratios and
responses to token women. American Journal of Sociology, 82, 965-990.

Kanter, R. M. (1977b). Some effects of proportions on group life: Skewed sex ratios and
responses to token women. American Journal of Sociology, 82, 965—990.

Kanter, R.M. (1977a). Men and women of the corporation (pp. 164-264). New York:
Basic Books.

131

Karakowsky. l... & Siegel. J. P. (1999). The effects of proportional representation and
gender orientation of the task on emergent leadership behavior in mixed-gender
work groups. Journal o/‘Applied I’.s'_t'cltol()gy. 84. 620-631.

Katz. 1). & Braly. K. (1935). Racial prejudice and racial stereotypes. Journal of
Abnormal and Social Psychology. 30, 175-193.

Kenny. D. A. & DePaulo. B. M. (1993). Do people know how others view them? An
empirical and theoretical account. Psychological Bulletin, 114, 145-161.

Kerlinger. F. N. 1986. Foundations ofbehavioral research. San Francisco: Holt,
Rinehart. and Winston.

Kline, RB. (1998). Principles and practice ofstructural equation modeling. New York:
Guilford.

Lawrence. B. S. (1997). The black box oforganizational demography. Organizational
Science. 8. 1-21.

Lee. C., & Bobko, P. (1994). Self-efficacy beliefs: Comparison of five measures. Journal
oprplied Psychology, 79, 364-369.

Leyens, J .P., Desert. M., Croizet, J.C., & Darcis, C. (2000). Stereotype threat: Are lower
status and history of stigmatization preconditions of stereotype threat? Personality
and Social Psychology Bulletin, 26, 1 189-1 199.

Lord, C. G., & Saenz, D. S. (1985). Memory deficits and memory surfeits: Differential
cognitive consequences of tokenism for tokens and observers. Journal of
Personality & Social Psychology, 49, 918-926.

Martell, R. F., Parker, C., Emrich, C. G., & Crawford, M. S. (1998). Sex stereotyping in
the executive suite: "Much ado about something." Journal of Social Behavior &
Personality, 13, 127-138.

Mattis. MC. (1994). Organizational initiatives in the USA for advancing managerial
women. In M.J. Davidson & R.J. Burke (Eds), Women in Management: Current
Research Issues, pp. 241-276. London: Paul Chapman.

Maurer, T. J ., & Pierce, HR. (1998). A comparison of Likert scale and traditional
measures of self-efficacy. Journal of Applied Psychology, 83, 324-329.

Mayer, D. M., & Hanges, P. J. (2003). Understanding the stereotype threat effect with

"culture-free" tests: An examination of its mediators and measurement. Human
Performance, 16, 207-230.

132

McIntyre, R.B., Paulson. R.M. & Lord. CG. (2003). Alleviating women‘s mathematics
stereotype threat through salience of group achievements. Journal of
Experimental Social Psychology. 39. 83-90.

McKay, P. F., Doverspike. D.. Bowen-Ililton. D.. & Martin. O. D. (2002). Stereotype
threat effects on the Raven Advanced Progressive Matrices scores of African-
Americans. Journal of Applied Social Psychology. 32. 767-787.

Merton, R.K. (1948). The self-fulfilling prophecy. Antioch Review. pp. 193-210.

Milliken, F.J. & Martins. LL. (1996). Searching for common threads: understanding the

multiple effects of diversity in organizational groups. Academy of Management
Review, 2], 402-433.

Morrison, A.M. & von Glinow, M.A. (1990). Women and minorities in management.
American Psychologist, 45, 200-208.

Morrison, A.M. (1992). The New Leaders. Jossey-Bass. San F ranciso, CA.

Morrison, A.M., White, R.P., & van Velsor. E. (1987'). Breaking the Glass Ceiling.
Reading, MA: Addison-Welsley.

Mossburgger, H., Scherrnelleh-Engel, K., & Klein, A. (1997). Methodological problems
of estimating latent interaction effects. Methods of Psychological Research
Online, 2. Retrieved August 13, 2004, from http://wvapabst-publishers.de/mpr/

Murphy, Kevin R., and Myors, Brett. (1998). Statistical Power Analysis: A Simple and
General Model for Traditional and Modern Hypothesis Tests. Mahwah, NJ:
Lawrence Erlbaum.

Murrell, A. J ., & James, E. H. (2001). Gender and diversity in organizations: Past,
present, and future. Sex Roles, 45, 243-257.

Niemann, Y. F., & Dovidio, J. F. (1998). Relationship of solo status, academic rank, and
perceived distinctiveness to job satisfaction of racial/ethnic minorities. Journal of

Applied Psychology, 83, 55-71.

O'Brien, L.T., & Crandall, CS. (2003). Stereotype threat and arousal: Effects on women's
math performance. Personality & Social Psychology Bulletin, 29, 782-789.

Osborne, J. W. (2001). Testing stereotype threat: Does anxiety explain race and sex
differences in achievement? Contemporary Educational Psychology, 26, 291-310.

Ott. E. M. (1989). Effects of the male-female ratio at work: Policewomen and male
nurses. Psychology of Women Quarterly, 13, 41-57.

133

 

 

Oyserrnan, D., Harrison, K.. & Bybee. D. (2001). Can racial identity be promotive of
academic efficacy? International Journal of Behavioral Development. 25. 379-
385.

Phillips, J., Hollenbeck, J. R.. & Ilgen. D. R. (1996). The prevalence and prediction of
positive discrepancy creation: Examining a discrepancy between two self
regulation theories. Journal of'Applied Psychology. 81. 498-511.

Pinel, EC. (1999). Stigma consciousness: The psychological legacy of social stereotypes.
Journal of Personality and Social Psychology, 76. 1 14-128

Pollak, K. 1., & Niemann, Y. F. (1998). Black and White tokens in academia: A
difference of chronic versus acute distinctiveness. Journal of Applied Social
Psychology, 28, 954—972.

Powell, G. N., Butterfield, D. A., & Parent. J. D. (2002). Gender and managerial
stereotypes: Have the times changed? Journal ofManagement. 28. 177-193.

Powell, G.N. & Graves. L.M. (2003). Women and Men in Management. Thousand Oaks.
CA: Sage Publications.

Pronin, E., Steele, C.M., & Ross, L. (2003). Identify bifurcation in response to stereotype

threat: Women and mathematics. Journal of Experimental Social Psychology, 40,
152-168.

Quinn, D. M., & Spencer, S. J. (2001). The interference of stereotype threat with

women's generation of mathematical problem-solving strategies. Journal of Social
Issues, 57, 55-71.

Raven, J. C., Court, I. H. & Raven, J. ( I985). Manualfor Raven 's Progressive Matrices
and Vocabulary Scales. London: H. K. Lewis.

Riordan, C. M. (2000). Relational demography within groups: past developments,
contradictions, and new directions. Research in Personnel and Human Resources
Management, 19, 131-173.

Roberson, L., Deitch, E. A., Brief, A. P., & Block, C. J. (2003). Stereotype threat and
feedback seeking in the workplace. Journal of Vocational Behavior, 62, 176-188.

Roberson, Q. M., Moye, N. A., & Locke, E. A. (1999). Identifying a missing link
between participation and satisfaction: The mediating role of procedural justice
perceptions. Journal of Applied Psychology, 84, 585—593.

Russell, J .A., & Carroll, J.M. (1999). On the bipolarity of positive and negative affect.
Psychological Bulletin, 125, 3-30.

134

Sackett, P. R., DuBois, C. L., & Wiggins Noe, A. (1991). Tokenism in perfomiance
evaluation: The effects of work group representation on male-female and White-

Black differences in performance ratings. Journal of Applied Psychology. 76.
263-267.

Sackett, P. R., Hardison, C. M., & Cullen, M.J. (2004). On interpreting stereotype threat
as accounting for Black-White differences on cognitive tests. American
Psychologist, 59, 7-13

Sackett, PR. (2003). Stereotype threat in applied selection settings: A commentary.
Human-Performance, I 6, 295-309.

Sackett, P.R., Schmitt, N., Ellingson, J.E., & Kabin, MB. (2001). High-stakes testing in
employment, credentialing, and higher education: Prospects in a post-affirmative-
action world. American Psychologist, 56, 302-318.

Saenz, D. S. (1994). Token status and problem-solving deficits: Detrimental effects of
distinctiveness and performance monitoring. Social Cognition, 12. 61-74.

Saenz, D. S., & Lord, C. G. (1989). Reversing roles: A cognitive strategy for undoing
memory deficits associated with token status. Journal of Personality & Social
Psychology, 56, 698-708.

Salas, E., Cannon-Bowers, J. A., & Blickensderfer, E. L. (1997). Enhancing reciprocity
between training theory and practice: Principles, guidelines, and specifications. In

J. K. Ford & Associates (Eds), Improving training effectiveness in work
organizations (pp. 291-322). Hillsdale, NJ: LEA.

Schmader, T. (2002). Gender identification moderates stereotype threat effects on
women's math performance. Journal of Experimental Social Psychology, 38, 194-
201.

Schmader, T. (2002). Gender identification moderates the effects of stereotype threat on

women’s math performance. Journal of Experimental Social Psychology, 38, 194-
201.

Schmidt, A. M., Chambers, B. A., & DeShon, R. P. (2004). A goal discrepancy
perspective on the relationship between self-efficacy and performance.
Unpublished manuscript.

Schmitt, N., Clause, C. S., & Pulakos, E. D. (1996). Subgroup differences associated
with different measures of some common job-relevant constructs. International
Review of Industrial and Organizational Psychology, 11, 115 — 139.

Schwarz, N. (1999). Self-reports: How the questions shape the answers. American
Psychologist, 54, 93-105

135

 

Sczesny, S. (2003). A closer look beneath the surface: Various facets of the think-
manager-think-male stereotype. Sex Roles, 49, 353-363.

Secord, P. (1959). Stereotyping and favorableness in the perception ofNegro faces.
Journal of A bnormal and Social Psychology, 59, 309-315.

Sekaquaptewa, D., & Thompson, M. (2002). The differential effects of solo status on

members of hi gh- and low-status groups. Personality & Social Psychology
Bulletin, 28, 694-707.

Sekaquaptewa, D., & Thompson, M. (2003). Solo status, stereotype threat, and
performance expectancies: Their effects on women's performance. Journal of
Experimental Social Psychology, 39, 68-74.

Snow, R. E., Kyllonen, P. C. & Marshalek, B. (1984). The topography of ability and
learning correlations. In R. J. Stemberg(Ed.), Advances in the psychology of
human intelligence (Vol. 2), pp. 47—103. Hillsdale, NJ: Erlbaum.

Spencer, S. J ., Steele, C. M., & Quinn, D. M. (1999). Stereotype threat and women's math
performance. Journal of Experimental Social Psychology, 35, 4-28.

Spielberger, C. D. (1983). Manual for the State-Trait Anxiety Inventory (F orm Y). Palo
Alto, CA: Consulting Psychologists Press.

Stajkovic, A.D., & Luthans, F. (1998). Self-efficacy and work-related performance: A
meta-analysis. Psychological Bulletin, 124, 240-261 .

Stangor, C., Carr, 0, & Kiang, L. (1998). Activating stereotypes undermines task
performance expectations. Journal of Personality & Social Psychology, 75, l 191-
1197.

Steele, C. M. (1997). A threat in the air: How stereotypes shape intellectual identity and
performance. American Psychologist, 52, 613-629.

Steele, C. M. (1998). Stereotyping and its threat are real. American Psychologist, 53, 680-
681.

Steele, C. M., & Aronson, J. (1995). Stereotype threat and the intellectual test
performance of African Americans. Journal of Personality & Social Psychology,

69, 797-811.

Steele, C. M., & Davies, P. G. (2003). Stereotype threat and employment testing: A
commentary. Human Performance, 16, 311-326.

136

Steele, C.M., Spencer, S.J., & Aronson (2002). Contending with group image: The
psychology of stereotype and social identity threat. In M. Zanna (Ed.), Advances “‘
in Experimental Social Psychology, 34, (pp. 379-436). San Diego, Academic
Press.

 

Steele-Johnson, D., Beauregard, R. S., Hoover, P. B., & Schmidt, A. M. (2000). Goal
orientation and task demand effects on motivation, affect, and performance.
Journal of Applied Psychology, 85, 724- 738.

Stone, J., Lynch, C. 1., Sjomeling, M., & Darley, J. M. (1999). Stereotype threat effects
on Black and White athletic performance. Journal of Personality & Social
Psychology, 77, 1213-1227.

Suryanarayanan, M. (2002). An interview with Ms. Sunalini Menon, Chief Executive,
Coffeelab Private Limited, Bangalore. Retrieved July 22, 2004, from
http://www.hindu.com/thehinduigbs/0201/05300014.htm.

Tabachnick, B. G. & Fidell, LS. (2001). Using Multivariate Statistics (4th Edition).
Boston: Allyn Bacon.
Taps, J ., & Martin, P. Y. (1990). Gender composition, attributional accounts, and

women's influence and likability in task groups. Small Group Research, 21 , 471-
491.

Tyler, T.R. & Blader, S.L. (2002b). Terms of engagement: Why do people invest
themselves in work? In H. Sondak (Ed.), Toward phenomenology of groups and

group membership. Research on managing groups and teams (V 01. 4, pp. 115-
140). New York: Elsevier Science.

Tyler, T.R., & Blader, S.L. (2002). Terms of engagement: Why do people invest
themselves in work? In H. Sondrak (Ed.), Toward phenomenology of groups and
group membership. Research on managing groups and teams, 4, (pp. 115-140).
New York: Elsevier Science.

US. Department of Labor, Women’s Bureau (2002). Facts on working women, 02-01.
Retrieved April 14, 2004 from http://wwwdol.gov/wb/factsheets/nontra2003.htm.

US. Department of Labor, Women’s Bureau (2002). Facts on working women, 03-01.
Retrieved July 22, 2004 from http://wwwdol.gov/wb/factsheets/hot'mbsO3.htm.

Vorauer, J. D., & Claude, S. (1998). Perceived versus actual transparency of goals in
negotiation. Personality & Social Psychology Bulletin, 24, 371-385.

Vorauer, J. D., & Miller, D. T. (1997). Failure to recognize the effect of implicit social

influence on the presentation of self. Journal of Personality & Social Psychology,
73, 281-295.

137

Vorauer, J. D., Hunter, A. J., Main, K. J ., & Roy, S. A. (2000). Meta-stereotype
activation: Evidence from indirect measures for specific evaluative concerns
experienced by members of dominant groups in intergroup interaction. Journal of
Personality & Social Psychology, 78, 690-707.

Vorauer, J. D., Main, K. J ., & O'Connell, G. B. (1998). How do individuals expect to be
viewed by members of lower status groups? Content and implications of meta-
stereotypes. Journal of Personality & Social Psychology, 75, 917-937.

Walsh, M., Hickey, C., & Duffy, J. (1999). Influence of item content and stereotype

situation on gender differences in mathematical problem solving. Sex Roles, 41,
219-240.

Weiner, B., Russell, D., & Lerman, D. (1979). The cognition-emotion process in
achievement-related contexts. Journal of Personality and Social Psychology, 3 7,
121 1-1220.

Wheeler, S. C., & Petty, R. E. (2001). The effects of stereotype activation on behavior: A
review of possible mechanisms. Psychological Bulletin, 127, 797-826.

Yoder, J. D. (1991). Rethinking tokenism: Looking beyond numbers. Gender & Society,
5, 178-192.

Yoder, J. D. (1994). Looking beyond numbers: The effects of gender status, job prestige,
and occupational gender-typing on tokenism processes. Social Psychology
Quarterly, 5 7, 150-159.

Young, J. L., & James, E. H. (2001). Token majority: The work attitudes of male flight
attendents. Sex Roles, 45, 299-319.

138

   

 

 

MICHIGAN STATE UNIVERSITY LIBRARIES
lllllllllllllllllllllllllllllllllllllllllllll l
l
3 1293 02504 430

0