OBSERVER RESPONSES TO WORKPLACE BULLYING: THE DYNAMIC
INFLUENCE OF RACE AND RELATIONAL DEMOGRAPHY
By
Brent Lyons

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF ARTS
Psychology
2010

ABSTRACT
OBSERVER RESPONSES TO WORKPLACE BULLYING: THE DYNAMIC
INFLUENCE OF RACE AND RELATIONAL DEMOGRAPHY
By
Brent Lyons
The present study explored the impact of relational demography on
observer responses to workplace bullying. To do so, we experimentally
manipulated the racial composition a working group and the race of a bullied
victim. Participants watched their workgroup bully a coworker at three time
points. At each time point, observers responded during the bullying incident (high
immediacy) and/or after the incident ended (low immediacy). The impact of racial
categorizations on observer responses depended upon the immediacy of the
response and time. Observers in groups that were mainly White and in groups that
were mainly White with a Black victim responded more riskily at first and
subsequent exposure to the bullying, suggesting that racial similarity and
stereotypes of prejudicial interactions are initial contextual cues for observer
responses. Relational demography did not influence observer responses after
repeated exposure to the bullying. Time moderates the impact of relational
demography on observer responses, though its effects differ depending on the
nature of the response.

Copyright by
BRENT LYONS
2010

TABLE OF CONTENTS

LIST OF TABLES……………………………………………………………….v
LIST OF FIGURES……………………………………………………………....vi
INTRODUCTION………………………………………………………………..1
METHOD………………………………………………………………………..20
RESULTS………………………………………………………………………..31
DISCUSSION……………………………………………………………………37
APPENDICES…………………………………………………………………....51
Appendix A: Procedural Materials……………………………………….52
Appendix B: Measures…………………………………………………...63
Appendix C: Tables and Figures…………………………………….…...70
REFERENCES…………………………………………………………………..75

iv

LIST OF TABLES

Table 1: Descriptive Statistics and Inter-Correlations………………………………….. 71
Table 2: Hiearchical Linear Modeling Results…………………………………………..72

v

LIST OF FIGURES

Figure 1: Interaction between SDO and Group Composition at Time 4………………...74

vi

INTRODUCTION
Bullying represents a serious problem in today‟s workforce: In 2007,
approximately 37% of U.S. employees reported being bullied while at work (Namie &
Namie, 2007). Workplace bullying is characterized as a significant source of stress for its
victims (Vartia, 2001). Humiliating and belittling remarks, gossiping, and rumours are
intimidating and frightening. Beyond negative outcomes typically reported by victims of
workplace aggression (for a review see Aquino & Thau, 2009; Bowling & Beehr, 2006)
bullying – that occurs on a persistent and long term basis – seems to be associated with
more severe health problems (Vartia, 2001). Effects of exposure to workplace bullying
have been shown to include social isolation, anger, anxiety, despair, depression
(Leymann, 1990), psychosomatic complaints, low sleep quality (Zapf, Knorz, & Kulla,
1996) and in some European samples, the development of posttraumatic stress disorder
(Leymann & Gustafsson, 1996). As a result of such stressors, victims of bullying report
lowered job satisfaction, increased absenteeism and turnover intentions (Hoel, Einarsen,
& Cooper, 2003; Keashly & Jagatic, 2003) and are often more likely to view leaving their
organization as a positive coping strategy (Zapf & Gross, 2001); victims are constantly
less able to cope with the daily task and interpersonal demands of their jobs (Einarsen,
2000).
Current research proposing organizational strategies to tackle workplace bullying
is scarce and has focused mainly on formal strategies and organizational policies
(Merchant & Hoel, 2003; Richards & Daley, 2003). It is, however, unclear whether
formal policies are actually effective in decreasing incidents of workplace bullying.
Victims are rarely likely to formally report victimization and are more likely to avoid the
1

aggressor(s) or deny the acts (Einarsen et al., 2003; Knapp, Faley, Ekeberg, & Dubois,
1997). As a result, organizational researchers have called for the examination of
alternative strategies –beyond victim self-reports – of intervention (Bowes-Sperry &
O‟Leary-Kelly, 2005).
An alternative mechanism thought to be helpful in attenuating workplace
aggression is observer (i.e., third party) responses to the aggression. Bowes-Sperry and
O‟Leary-Kelley (2005) formulated a theoretical framework postulating antecedents to
observer responses to sexual harassment. To our knowledge, no empirical research
examining observer responses to harassment or bullying has been published. Although,
research on organizational citizenship behaviors (OCBs) (Smith, Organ, & Near, 1983;
Organ, 1988; Organ, Podsakoff, & MacKenzie, 2006) has examined interpersonal forms
of coworker helping (i.e., OCB-I; Robinson & Bennett, 1995; Williams & Anderson,
1991), its focus is prosocial behaviors generally (e.g., helping with work duties, being
courteous during social interactions), not helping behaviors in response to acts of
aggression (e.g., consoling a victim or reporting the acts). However, researchers have
examined observer responding in child and adolescents populations with promising
outcomes: Based on naturalistic observations of school-yard bullying, peers intervened
rarely, but when they did, bullying behaviour ended quickly (Hawkins, Pepler, & Craig,
2001). As follows, coworker responding to bullying in the workplace is an understudied
but potentially important avenue of empirical research for organizational researchers.
Below, we clarify a definition workplace bullying along with a discussion of
issues related to its measurement. After which, we highlight Bowes-Sperry and O‟LearyKelley‟s (2005) conceptualization of observer responding to sexual harassment while
2

highlighting factors Bowes-Sperry and O‟Leary-Kelley propose will influence observer
responses, particularly factors related to social identity categorizations. We then discuss
theory and research on relational demography and how it can inform social identity
effects on observers‟ helping. Finally, we make our own hypotheses integrating relational
demography and Bowes-Sperry and O‟Leary-Kelley‟s theoretical framework in order to
predict observer responses to bullying.
Theoretical Background
Defining Bullying
Research on workplace bullying originated in Finland in the 1980‟s (Leymann,
1990) following the country‟s surge of interest in school-based bullying. The construct
rapidly developed interest in other Scandinavian and U.K. countries in the mid to late
1990‟s (Einarsen et al., 2003). In the U.S., researchers have explored a wide variety of
hostile workplace behaviours labelled as a myriad of constructs, such as workplace
deviance (Robinson & Bennett, 1995), workplace aggression (Baron & Neuman, 1996),
incivility (Andersson & Pearson, 1999), harassment (Bowling & Beehr, 2006), and
counterproductive work behaviors (Sackett & DeVore, 2001). In their review of these
constructs, Raver and Barling (2008) note that despite the seemingly endless list of labels
for negative interpersonal acts, such behaviours possess many commonalities. To
encompass most negative interpersonal acts, Raver and Barling adopt Neuman and
Baron‟s (2005) conceptualization of “workplace aggression” defined as behaviour
directed by an employee toward a coworker or series of coworkers with the intent to
harm the target(s) in ways the target(s) is/are motivated to avoid (also see Baron &
Neuman, 1996). Behaviors encompassed under the general term of “workplace
3

aggression” do, however, vary along a series of dimensions. In particular, workplace
bullying is characterized by its persistence over time and a power difference between the
actor and target (Raver & Barling, 2008). Research on workplace aggression has typically
been limited to episodic aggression without consideration for the persistent and long term
nature of bullying (Hoel, Rayner & Cooper, 1999; Leymann, 1996; Rayner & Keashly,
2005). Researchers in Europe and Scandinavia have used the term mobbing to describe
persistent aggressive acts by a group of coworkers towards a victim (Einarsen, 2000;
Leymann, 1990, 1996). Though unique in its focus on group behaviour, the term
“mobbing” tends to be used interchangeably with “bullying” (Einarsen, 2000) with
bullying more generally encompassing aggressive acts that involve either a group or a
single perpetrator.
The current study utilizes an agreed upon definition of workplace bullying:
Workplace bullying is defined as an actor or group of actors repeatedly harassing,
offending, socially excluding or negatively affecting an individual. A power imbalance is
created when the victim perceives he/she is unable to defend him/herself against the
persistent harmful behaviour. An incident is not bullying when the negative acts are
episodic and the actor and target are of equal power when in conflict (Lutgen-Sandvik et
al., 2007; Einarsen et al., 2003; Salin, 2003). Common examples of workplace bullying
include: persistent offensive remarks towards a victim, persistent criticism, the „silent
treatment‟ and persistent isolation or exclusion of the victim from his or her peer group
(Leymann, 1996; Zapf et al., 1996).
Workplace bullying has a series of defining characteristics that differentiate it
from other forms of workplace aggression. Bullying involves an individual as a target of
4

multiple negative acts by one or more actors on a persistent basis (Leymann, 1990; Zapf
et al., 1996); occurring frequently, usually at least twice a week (Salin, 2001). In order to
differentiate bullying from other social stressors at work, bullying is typically defined as
occurring over a period of six months (Zapf et al., 1996); six-months is frequently used in
the assessment of various psychological disorders and bullying is thought to lead to
severe psychological distress (Leymann, 1990). Finally, bullying involves the presence of
a power disparity between the victim and perpetrator(s). That is, victims of bullying, for a
variety of reasons, feel they are unable to overcome their experienced abuse (LutgenSandvik et al., 2007). The power imbalance can exist prior to the onset of bullying (e.g.,
the formal power structure of the organization; or informal sources of power such as
information, experience, or social dependence) or it can develop as the bullying escalates
(i.e., the emergence of a perceived power deficit occurs as victims are continuously
treated as inferior) (Einarsen et al., 2003). Power differentials can be formal, when a
supervisor withholds information from a subordinate, or informal, when a group of
employees begins to continuously exclude another coworker.
Measuring Bullying
Current understanding of the antecedents and outcomes of bullying and workplace
aggression is primarily based on research using self-reports of victim experiences, either
reporting on their perceived exposure to the behaviours or their perceived victimization
from the behaviors (Einarsen et al., 2003). Specifically, the Negative Acts Questionnaire
(NAQ) (Hoel & Cooper, 2000) and the Sexual Experiences Questionnaire (SEQ)
(Fitzgerald et al., 1997) are common inventories used to assess experiences with bullying
and sexual harassment, respectively. In both surveys, respondents are asked to indicate
5

how frequently they are on the receiving end of specific forms of negative acts (e.g.,
jokes of a sexual nature for SEQ and withholding of job-necessary information for NAQ).
Ilies, Hauserman, Scwochau, and Stibal (2003) argue that self-reports of victimization are
likely lower than is actually experienced by victims because respondents are hesitant and
less likely to report victimization as it is defined by the researchers in the survey.
Additionally, by utilizing retrospective accounts of victimization, researchers have not
been able to explore how victims and bystanders experience bullying as a dynamic
process. Additionally, an examination of bullying across time is imperative in order to
gage an understanding of the dynamic processes and mechanisms associated with
observer responses.
In the current study, bullying is operationalized as persistent negative acts by
actor(s) towards a target that occur over several months in situations where the target is
in a position of less power than the aggressor(s). The current study intends to develop an
understanding of how time influences the role of social identity factors in influencing
observers‟ responses to workplace bullying.
Observer Responses to Workplace Aggression
Researchers of developmental and school psychology are well aware of the
positive effects of observer responding in child and adolescent bullying (Frey et al.,
2005). Specifically, bystanders are present in 85% of school bullying incidents and
bystander efforts to stop the bullying are typically effective in a majority of instances
(Hawkins et al., 2001). The role observers play in intervention in workplace bullying is,
however, less clear. An estimated 57% of workplace aggression occurs in the presence of
observers (Glomb, 2002) and for every mistreatment in the workplace there are more
6

observers present than victims (Skarlicki & Kulik, 2005). Observers have power and can
be influential in changing situations (Clarkson, 1996). However, no empirical research
has investigated factors influencing observer responses to workplace bullying and how
the effects of these factors change as the bullying continues.
Observers are defined as individuals who witness workplace aggression
occurring but are not directly involved (Bowes-Sperry & O‟Leary-Kelly, 2005).
Observers can take action to stop the bullying or prevent its continuation. They can
respond to the negative acts in a variety of ways by reporting the incidents to formal
authorities, stopping an unfolding event, providing negative feedback to the perpetrators
regarding their behavior, or by consoling the victim. Bowes-Sperry and O‟Leary-Kelly
acknowledge that observers will have different motives for responding to workplace
aggression: Helping behaviors can be in-role (i.e., an expectation of the observer‟s job),
extra-role (e.g., OCBs), or observers can be motivated to benefit themselves, the victim,
the group, or the organization.
Bowes-Sperry and O‟Leary-Kelly (2005) categorize observer responses to sexual
harassment along two dimensions: 1) the immediacy of the intervention, and 2) the level
of involvement. Immediacy of intervention distinguishes between responses that occur
during the incident (high immediacy) versus those that occur after the incident (low
immediacy). For example, an observer may choose to intervene and disrupt the incident
as it is occurring (e.g., by taking the target away from the incident), or the observer may
choose to later advise or console the target after the incident has ended.
The second dimension, level of involvement, reflects the degree to which
observers immerse, or publicly embroil themselves in the incident (Bowes-Sperry &
7

O‟Leary-Kelly, 2005). Observers may become highly involved by inserting themselves
within the ongoing incident; or an observer may opt to become less involved and provide
support for the target, assistance that does not involve a strong public connection to the
issue. A key component of involvement is response riskiness. Observers that choose a
more confrontational style of responding (e.g., telling the aggressor to stop, directly
labeling the behavior) are at risk for pulling themselves into the conflict and such a
strategy is risky for observers compared to less confrontational styles (e.g., diverting the
victim‟s attention, consoling the victim). Observers choosing to engage in higher
involvement (i.e., more risky) behaviors are taking an active role in stopping the negative
acts. Less confrontational and more covert strategies are used by observers to assuage the
situation and provide support for those involved (2005). In their two-dimensional
typology of observer responding, Bowes-Sperry and O‟Leary-Kelley provide a
framework which researchers can use to evaluate observer responses to workplace
aggression. In the present study, we incorporated both response immediacy and
involvement into our operationalization of observers‟ responding. We assessed observer
response involvement (i.e., the riskiness or confrontational nature of the response) for
both high and low immediacy responses.
When choosing to become involved, observers use an array of behaviors. In order
to formulate their propositions predicting observer responding, Bowes-Sperry and
O‟Leary-Kelley (2005) drew from Latane and Darley‟s (1970) conceptualization of
bystander intervention. Latane and Darley‟s framework presents bystander intervention
as the last step in a decision-making process, where the observer decides on a form of
assistance to provide to the victim. Bowes-Sperry and O‟Leary-Kelley highlight several
8

situational factors that may influence variability in observer response behaviors as a part
of this decision-making process. They place particular emphasis on social identity
characteristics of the situation (e.g., the gender of the perpetrator and victim) as important
influences in observers‟ choice of whether and how to intervene. Of particular interest in
the present study are social identity – and more specifically, racial identity – influences
on observer response behaviors. In order to inform our hypotheses, we draw on theory
and research on relational demography.
Relational Demography
It is important to study the effects of surface level characteristics as influences on
interpersonal functioning in groups: New and short-lived groups such as task forces and
project teams are continuously forming and dismantling in organizations never reaching a
point where deep level characteristics become a strong determinant of interpersonal
interactions. Demographic characteristics of individuals (e.g., age, race, gender) have
long been considered important variables in psychological research (Tsui & O‟Reilly,
1989) and previous research has highlighted surface level characteristics as influential in
predicting helping behaviors (e.g., Dovidio et al., 1997; Levine & Crowther, 2008; Wispe
& Freshley, 1971; Piliavin, Rodin, & Piliavin, 1969). Beyond studying the independent
effects of demographic attributes on individual outcomes, the effects of demography can
also have a broader impact. Pfeffer (1982, 1983) argued that distributional properties of
an organization‟s demography are important for understanding the effects of
demographic attributes on unit outcomes. Organizational demography refers to the
composition of a social unit (e.g., a work group) in terms of the distribution of its
demographic attributes (e.g., the number of women at a law firm). Research that has
9

treated demography as a compositional property (e.g., Pfeffer, 1983) has measured
variance in demography of the unit and related that variance to unit outcomes. Kanter
(1977) termed the degree of demographic composition of a social unit as structural
integration, that can range from token representation (i.e., a lone woman in a group of
men), skewed representation (i.e., women are a large minority in a group of mainly men),
or equal representation in a balanced group. Demographic variation is also measured at
the individual level: Tsui and O‟Reilly use the term relational demography to refer to the
“comparative demographic characteristics of members of dyads or groups who are in a
position to engage in regular interactions” (p. 403). In essence, they propose that
knowledge of the similarity (or dissimilarity) of a given demographic attribute of
members of a unit provides greater insight into the individual members‟ characteristic
attitudes, behaviours, and how demography influences job outcomes.
Findings from past research exploring the effects of race on bystanders‟ helping
behaviors have been inconsistent (e.g., Piliavin et al., 1969; Wispe & Freshley, 1971).
Given that Bowes-Sperry and O‟Leary-Kelley (2005) suggest social identity
categorizations influence Latane and Darley‟s (1970) decision-making process of
observer responding, we explore race as an influence on observer responding to
workplace bullying in the current study. Drawing from the relational demography
framework, we contextualize individual race within a broader social unit, exploring the
impact of racial similarity between an observer and victim and the racial composition of
the victim and observer‟s working group on the observers‟ responses.
Theoretical Rationale behind Relational Demography

10

Theory used to inform relational demography research has traditionally stemmed
from the similarity-attraction paradigm (Byrne, 1971; for a review of how the similarityattraction paradigm relates to relational demography see Tsui, Egan, & O‟Reilly, 1992).
The similarity-attraction paradigm posits that similarity in attitudes is a source of
attraction between interacting individuals. Physical traits can be used as a basis for
inferring attitudes, beliefs and personality (Tsui et al., 1992). O‟Reilly, Caldwell, and
Barnett (1989) adopted the similarity-attraction paradigm to explain outcomes of
relational demography and described social integration as the explanatory mechanism.
Social integration is defined as the degree to which an individual is psychologically
linked to others in a group. A multifaceted phenomenon, it reflects an individual‟s
attraction to the group, satisfaction with other members of the group, and social
interaction among the group members (Katz & Kahn, 1978; O‟Reilly et al., 1989). Social
integration is thought to be influenced by relative similarity to other group members
through perceived similarity in terms of attitudes and beliefs. Individuals who are more
socially integrated perceive themselves as more similar to their group members, have
lower rates of turnover and higher satisfaction (O‟Reilly et al., 1989), and have higher
perceptions of inclusion (Pelled, Ledford, & Mohrman, 1999). However, Tsui et al., note
that the similarity-attraction paradigm assumes interaction amongst individuals, and
social identity theory (Tajfel & Turner, 1986) and self-categorization theory (Turner,
Hogg, Oakes, Reicher, & Wetherell, 1987) are used together to account for relational
demography effects without requiring social interaction and integration.
Social identity theory posits that individuals desire to maintain a positive self
identity (Tajfel & Turner, 1986) but in order for individuals to know how to feel about
11

others they need to first define themselves, for which they go through a process of selfcategorization (Turner et al., 1987). Through the self-categorization process, individuals
classify themselves and others into social categories based on relevant attributes (e.g.,
age, race, gender etc.,) allowing individuals to define themselves in terms of a social
identity. Self-categorization utilizes prototypes (i.e., a cognitive representation of
attributes characteristic of a particular group) that accentuate perceived similarities and
dissimilarities between referent in-group and out-groups, respectively (Hogg & Terry,
2000). The referent target is therefore no longer viewed as a unique individual, but as an
embodiment of the relevant prototype of the social identity, a process Hogg and Terry
call “depersonalization”. Bowes-Sperry and O‟Leary-Kelley (2005) propose that the
interpersonal similarity and sense of common fate characteristic of common in-group
categorizations give rise to a sense of we-ness, that Dovidio et al. (1997) define as “a
sense of connectedness or a categorization of another person as a member of one‟s own
group” (p. 102). Depersonalization of the self leads individuals to develop “empathetic
altruism” in which individuals feel obliged to support the needs and goals of similar
others (Tajfel & Turner, 1986). The sense of we-ness as a force behind in-group helping
behavior can thus be explained in terms of individuals‟ concerns for protection and
enhancement of their own self-concept.
If a social identity allows an individual to have a positive self-identity, individuals
thus perceive prototypical attributes of their social group as attractive (Tajfel & Turner,
1986). As follows, social units that contain similar others are more likely to be regarded
positively leading individuals to prefer homogeneous groups of similar others over
diverse groups of dissimilar others (O‟Reilly et al., 1989; Pelled et al., 1999). In addition
12

to social identity processes being motivated by the pursuit of a positive social identity
(Abrams & Hogg, 1988; Hogg & Abrams, 1990), self-categorization processes are
motivated by a need to reduce subjective uncertainty about one‟s self concept (Hogg &
Abrams, 1993). Self-categorization reduces uncertainty by transforming the self into a
prototype that prescribes perceptions, feelings, attitudes and behaviours (Goldberg,
Riordan, & Schaffer, 2010; Reid & Hogg, 2005). A social unit is psychologically
attractive to an individual to the extent that it is composed of others who are similar based
on the relevant demographic category (Tsui et al., 1992) and individuals interpret
themselves within that unit based on the categorizations in order to reduce subjective
uncertainty. Tsui et al., explored forms of organizational attachment as an outcome of
demographic composition. Individuals had higher attachment (i.e., less absences, higher
psychological commitment, and intent to stay) to social units composed of similar others.
Relational Demography and Helping Behaviors
The underlying link between similarity in relational demography and helping
behaviors is attachment. Researchers have demonstrated that individuals are more likely
to behave prosocially and engage in OCBs in units to which they perceive attachment.
Individuals are attached to a group if they share a psychological bond with that group
(Allen & Meyer, 1990; O‟Reilly & Chatman, 1986). Such individuals identify with and
internalize the group‟s attitudes, values, and goals as their own because they perceive the
group‟s attributes as congruent with their own (1986). Members of a unit who share its
goals and values view their membership as a relational exchange and act in reciprocation
to benefit the unit (O‟Reilly & Chatman, 1986; Morrison, 1994). Individuals feel more
committed and attached to groups to which they are demographically similar (O‟Reilly et
13

al., 1989; Tsui et al., 1992). Thus, racial categorizations can act as cues for attachment
and subsequent helping behaviors.
However, in order for individuals to base perceptions of attachment to a group on
a specific social identity (e.g., race) and the representation of that social identity within
the group, the relevant social category first needs to be salient in that particular context.
According to Hogg and Terry (2000), a social category that “fits” with a social context is
the category made salient in interpreting one‟s place within that context. Social categories
become salient because they “account for situationally relevant similarities and
differences among people and/or because category specifications account for contextspecific behaviours” (2000; p. 125). Once a category is made salient, contextually
relevant prototypes are used for accentuating in-group similarities and out-group
differences. Previous research has demonstrated that the demographic composition of a
social unit influences the salience of those demographic categories. For example, racial
and gender distinctions are more pronounced when a group lacks diversity and is
primarily composed of one race or gender (i.e., token or skewed groups) (Austin, 1997;
Kanter, 1977). It thus follows that in groups with pronounced dissimilarity in racial
representation, race is made salient, prototypes accentuate in-and out-group
categorizations, and race is more likely to be used by individuals when interpreting their
place within the group (e.g., perceived attachment).Therefore, individuals who are
racially similar to their groups perceive greater attachment to that group and are more
likely to behave in ways to benefit the group. Previous research has demonstrated that
individuals who perceive less attachment to their unit are less willing to cooperate with
others, make less sacrifices on others‟ behalves, behave less respectfully during social
14

interactions (Thau, Aquino, & Poortvliet, 2007), and perform less OCBs (Den Hartog, De
Hoogh, & Keegan, 2007).
The current study explores the effects of race and its representation within a
working group (i.e., racial composition) on observers‟ responses to workplace bullying.
Specifically, White observers responded to the bullying of a White or Black victim. The
working group, in which the bullying occurred, consisted of mainly White coworkers,
mainly Black coworkers, or a balanced number of White and Black coworkers. Observers
had opportunities to respond as the bullying continued over time. Finally, the role of
social dominance orientation and its interaction with victim race and group composition
was also explored.
Hypotheses
Relational Demography and Observer Responses to Workplace Bullying
Similarity in group membership fosters a sense of we-ness (Dovidio et al., 1997)
and empathetic altruism (Tafjel & Turner, 1986) promoting helping behaviors in times of
threat (Riordan & McFarlane Shore, 1997). Individuals act in ways to benefit their own
self-concept (1986) through the depersonalizing of others based on prototypical
categories (Hogg & Terry, 2000; Turner et al., 1987), such that individuals will act in
ways to benefit their in-group. In instances of mistreatment, it is expected that observers
will display more involved helping behaviors when they are racially similar to the victim
of bullying.

15

H1: Observers will display higher involvement (i.e., higher risk) in their
responses when the victim of bullying is racially similar to them (i.e., White) as
opposed to when the victim is racially dissimilar (i.e., Black).

Individuals are more attached to groups to which they are similar in some
meaningful way (e.g., a demographic attribute) (O‟Reilly et al., 1989; Pelled et al., 1999;
Tsui et al., 1989). Individuals who perceive attachment to a unit are more likely to behave
in ways to benefit the unit (Allen & Meyer, 1990; O‟Reilly & Chatman, 1986). As
bullying is a behavior that is disruptive to interpersonal functioning, individuals who are
attached to the group will be motivated to overcome the bullying. Additionally, groups
that are more homogeneous in their racial composition will provide a context by which
race is salient for individuals to perceive attachment to the group (Hogg & Terry, 2000).
Attachment and helping behavior will be higher as perceived similarity will be high. In
balanced groups, race is not a key means for interpretation of attachment (Austin, 1997).

H2: Observers will display lower response riskiness when they are racially
dissimilar to their working group (i.e., the group is mainly Black) than when the
group is racially balanced.

H3: Observers will display higher response riskiness when they are racially
similar to their working group (i.e., the group is mainly White) than when the
group is racially balanced.

16

However, in groups that are mainly Black, the mistreatment of a White member
will make race particularly salient for observers because dissimilarity will be enhanced
by prototypical categorizations of in and out-group membership (Hogg & Terry, 2000).
In such situations, the desire for enhancing one‟s self-concept will be accentuated (Tajfel
& Turner, 1986) fostering in group helping behaviors. However, when Whites are
dissimilar and the victim is Black, attachment is low and concerns for a positive selfconcept are not at play and helping behaviors will be at a minimum.

H4: When the working group is mainly Black, observers will display higher
response riskiness when the victim is White than when the victim is Black.

When the group is mainly White, observers will have higher perceived
attachment, but their responses will also be accentuated by the nature of the situation if
the victim is Black. Race, and prototypical conceptions of prejudice and discrimination
(i.e., perceptions of the prototypic perpetrators and victims of racism and sexism etc.,)
(Inman & Baron, 1996) will be a salient basis for interpretation of the context-specific
behaviors (Hogg & Terry, 2000). That is, a prototypical view of racism is a White
individual discriminating against a person of color. Inman and Baron assert that
individuals are particularly sensitive to and reactive to prototypical mistreatment.
Instances high in prototypical characteristics are more likely to trigger stereotypical
effects in interpretation and behavior (1996). Therefore, observers will be more sensitive
to discrimination and prejudice within groups with characteristics prototypical of such
mistreatment.
17

H5: When the working group is mainly White, observers will display higher
response riskiness when the victim is Black than when the victim is White.

Observer interpretations of a prototypically salient situation as a negative act and
their subsequent response may also be influenced by their tendencies to endorse
ideologies that legitimize group status differences. Social dominance orientation (SDO) is
a general attitudinal orientation reflecting whether one generally prefers intergroup
relations to be equal, or hierarchical along a superior-inferior dimension. Individuals high
on SDO desire that their in-group dominate and be superior to out-groups whereas those
lower on SDO favor hierarchy-attenuating ideologies (Pratto, Sidanius, Stallworth, &
Malle, 1994). It is suggested that members of traditionally high-status groups (e.g., White
men) use SDO as a device to maintain superior group status; SDO is predictive of
legitimizing ideologies endorsing prejudice and discrimination (Pratto et al., 1994).
Therefore, in situations prototypical of prejudice or discrimination, individuals high on
SDO may exhibit lower response behaviors in attempts to maintain their superior group
status. Thus, an interaction between SDO, victim race, and group composition is
proposed:

H6: When the working group is mainly White, observers low on SDO will display
higher response riskiness when the victim is Black than White, but observers high
on SDO will display higher response riskiness when the victim is White than
Black
18

Effects of Time on Observer Responses to Workplace Bullying
Bullying involves recurring incidents of mistreatment over a long period. With
increasing incidents of mistreatment, perceptions of persistence are likely to result.
Bowes-Sperry and O‟Leary-Kelley (2005) propose that if observers expect sexual
harassment to recur, they will be more likely to take subsequent action. Given that
observers view mistreatment as harmful (Skarlicki & Kulik, 2005), they are likely to feel
accountable for future incidents. As mistreatment persists, expectations of recurrence
solidify and feelings of accountability increase. It is likely that substantial dissonance will
be caused by continued inaction if observers witness more mistreatment (Bowes-Sperry
& O‟Leary-Kelley, 2005).

H7: Observers will display higher response riskiness as time passes.

Across time, with increasing exposure to individuals from various racial groups,
attention paid to racial category distinctions decreases (Harrison, Price, & Bell, 1998;
Martins et al., 2003). Literature on individual cognitive processing of diversity (e.g.,
Austin, 1997), the contact hypothesis (e.g., Brewer & Brown, 1998), and on the symbolic
aspects of diversity (e.g., Riordan, 2000) all suggest that recurrent exposure to individuals
from out-group racial categories reduces the reliance individuals place on racial identity
when interacting with members of other social groups. That is, with increased exposure to
diversity, race becomes less important for interpersonal functioning. Group members
become desensitized to racial differences within their group as they come to develop a
19

deeper level understanding of their relationships and interactions with other group
members (Glaman, Jones, Rozelle, 1996). With increased exposure, interpretations of and
reactions to others focus more on deeper level differences and less on race.

H8: As time passes, differences in response behaviors as a function of race of the
victim and group composition will lessen

On a final note, Bowes-Sperry and O‟Leary-Kelley (2005) were not clear in their
differentiation between the riskiness of specific types of observer responses. For example,
it is not clear if engaging in a lower involvement strategy while the bullying is occurring
(high immediacy) is more or less risky than engaging in a high involvement strategy after
the bullying incident is over (low immediacy). We therefore assessed observer response
involvement separately at two levels: high immediacy and low immediacy. We
acknowledge that observer responses at high and low immediacy are likely to accrue
important practical and theoretical differences, though for the current study we view
those differences as exploratory in nature. As a result, the above hypotheses are general
to response riskiness and no specific hypotheses are made for differences in high and low
immediacy.
Method
Participants
Participants were recruited from an undergraduate psychology subject pool at a
large public university in the Midwestern United States. There were 226 participants
(mean age = 19.69; SD = 2.57) of which 181 (76%) were female and 44 (18%) were
20

male; 14 participants (6%) did not indicate their gender. Sample size was determined
based on a priori power analysis using G*Power (version 3.0.10) software (Faul,
Erdfeller, Lang, & Buchner, 2007) in order to detect a medium effect size employing the
traditional .05 criterion level of statistical significance. As per required for participation
in this study, 100% of participants identified their race as White. The mean amount of
participant work experience ranged from 2 to 2.5 years. Participants received course
credit for participating in the study.
Procedure
Participants were made aware of each stage in the study in the initial consent
form. Participants firstly completed an online survey containing the individual difference
measures: SDO; and prosocial orientation, trait empathy and perspective taking as
exploratory measures. Rosenberg‟s measure of self-esteem (Rosenberg, 1965) was
included in this online survey as a distracter measure. Between one and two days after
completing the online survey, an experimenter emailed participants to coordinate times
for them to complete the in-lab portion of the study. Effort was made to ensure that
participants completed the in-lab portion of the study between four and seven days after
their completion of the online survey. Completion of the online survey and the lab study
were separated in time to minimize the sensitive nature of the individual difference
measures influencing participant responses to the lab stimuli.
Prior to each lab session, an experimenter randomly assigned each participant to
one experimental condition. Once in lab, participants read instructions asking them to
think of themselves as a part of a workplace scenario (see Appendix A). In this scenario,
each participant was a member of a workgroup, along with five other coworkers engaging
21

in a series of weekly meetings that occur several weeks apart. The scenario provided a
brief history of the workgroup as well as a brief outline of the group‟s tasks. In the
scenario, participants were told they would observe videos of their group engaging in
meetings. Following the scenario description, participants were asked to study a diagram
consisting of photographs and demographic information (e.g., name, age, tenure) of the
members of their workgroup and they were told that the diagram was representative of
the location where each group member would be sitting during each meeting. Group
member demographic information depicted in the diagram remained consistent across all
conditions of the experiment, but the racial composition of the group and the victim‟s
race (a woman named Ashley) were manipulated with photographs across conditions.
Experimental conditions for this study consisted of a 2 (victim race: Black vs.
White) x 3 (group racial composition: mainly White vs. balanced vs. mainly Black)
between-participants design. Victim race and group composition were manipulated in the
group diagrams and videos of the group meetings. Photographs of the five other group
members within the diagrams and the corresponding actors in the videos were
manipulated according to each condition. It is important to note that a fully crossed
observer race design (i.e., Black and White observers) was not used for two reasons: a)
The interest of this study is in the reactions of traditionally White majorities to negative
acts in various diverse contexts, and b) to simplify procedures and data collection.
After studying the scenario and group diagram, participants responded to a series
of manipulation check questions to ensure that they had attended to the manipulations of
interest (i.e., their group‟s racial composition and the victim Ashley‟s race). Embedded in
a series of questions about their coworkers‟ age, tenure, gender, appearance, and
22

positions in the group diagram (see Appendix A), participants responded to two focal
questions: “Approximately how many of your fellow group members are White (or
Caucasian)?” and “Based on her picture, what is Ashley‟s race/ethnicity?”
Following the manipulation check, participants were cued to watch the first video
of their workgroup meeting. Each video was approximately two minutes in length and
participants were not told how many videos they would observe. The first video did not
depict acts of bullying as it was intended to accustom participants to the video stimuli and
response format. After each video ended, participants were prompted to respond to an
open-ended assessment of their response to the video (see Appendix A). After completing
these measures, participants were cued to watch the next video. Each participant observed
four videos: video two, three, and four depicted exclusionary bullying by the group
members of a female victim. Response formats were identical after each video.
Participants completed a final demographic inventory after completing measures
for video four (see Appendix B).
Stimuli Development
Bullying Behavior. Before the scripts for each of the four videos were created, the
authors wanted to ensure that the bullying depicted in the videos was consistent with how
workplace bullying is conceptualized in the literature. In order to do so, the primary
investigator created five descriptions of workplace scenarios (see Appendix A) each
depicting some form of workplace incivility. Some of the scenarios were more in line
with the definition of workplace bullying (see Lutgen-Sandvik et al., 2007; Einarsen et
al., 2003; Salin, 2003) than others. In an online survey, separate participants indicated the
extent to which each scenario was reflective of the definition of workplace bullying
23

provided to them on a seven point Likert-type scale (1 = “Not at all” to 7 = “Definitely”).
Participants were 37 students (mean age = 19.27; SD = 1.12) recruited from an
undergraduate psychology subject pool of a large public university in the Midwestern
United States. Of the sample, 30 (81%) identified their gender to be female, and 7 (19%)
male; 31 (92%) identified their race to be White, 2 (5%) as Black/African American, and
1 (3%) as Hispanic. Mean work experience for this sample ranged from 2 to 2.5 years.
Participants rated scenario two (M=6.27; SD=0.84) and four (M=6.19; SD=1.02) as
significantly more consistent with the definition of workplace bullying than scenarios one
(M=4.81; SD=1.49), three (M=4.59; SD=1.50), and five (M=4.84; SD=1.77) based on
repeated-measures ANOVA, F(1,36)=11.38, p=.001, and follow-up paired-sample t-tests,
ts(36)>3.71, ps<.001.
The scripts for videos two, three and four were based on behaviors described in
scenario four. Bullying behaviors depicted in the videos included social exclusion and
isolation by all group members of a lone female victim (i.e., Ashley) during the meetings.
Example bullying behaviors included: interrupting and ignoring Ashley‟s comments, not
holding the door open or passing her an agenda. We ensured that the bullying behaviors
depicted at times two, three, and four were similar in type (e.g., all videos involved
similar exclusion from group discussion) to ensure that variation in bullying type did not
confound with time. Specific bullying behaviors were repeated in each video in order to
capture the repetitive nature of bullying. The scripts used to create the videos can be
found in Appendix A.
Victim Attractiveness. In order to ensure that the attractiveness of the Black or
White victims did not influence participant responses to the bullying incidents, female
24

actors were matched for attractiveness across conditions. Each actress had a picture of her
face taken. Pictures were standardized to only include the face (from the shoulders up, no
smile) against a white background. A separate group of 20 White volunteers (mean
age=23.01, SD=2.13; 13 [65%] female, 7 [35%] male), who were acquaintances of the
primary investigator, rated each picture on attractiveness. Using a five-point Likert-type
scale, participants rated the extent to which they thought each person was attractive (1=
“Very unattractive” to 5 = “Very attractive”). Statistically controlling for participant
gender, repeated measures ANCOVA was used to compare the attractiveness ratings of
the Black and White actresses. The Black and White actresses whom did not differ
significantly in attractiveness were selected as the Black and White alternates for the
victim, Ashley.
Video Production. Four Black (two males, two females) and three White (one
male, two females) undergraduate students from a large public university in the
Midwestern United states were recruited to act in all of the videos. Four videos were
created for each of six conditions. Across all conditions, gender composition of the
workgroup was held constant: Each workgroup was composed of two male actors and
three female actors in addition to the gender of the participant. Great effort was made to
ensure that the actions and appearances of the actors were consistent across conditions
while filming.
Pilot Testing
The instructions, group diagrams, and videos were pilot tested to ensure that the
manipulations were strong enough to yield differences in response behaviors across

25

conditions. The stimuli yielded promising levels of difference so no iterative changes
were necessary before official data collection began.
Independent Variables – Level 2 (Between-Individual) Measures
Gender. Participants‟ gender was measured with a variable in the final
demographic inventory (coded as 0 = male, 1= female).
Group racial composition and victim race. Participant workgroups contained six
coworkers , the participant and five actors: one victim (i.e., Ashley) and four additional
perpetrating coworkers. Six conditions were created by varying (a) the racial composition
of the group: whether the group was mainly White, balanced or mainly Black (dummy
coded so balanced was the comparison group) and (b) the race of the victim: the victim
was either White or Black (coded as 0 = Black, 1 = White). In the mainly White condition,
four members of the group were White and two were Black. In the mainly Black
condition four members of the group were Black and two were White. In the balanced
condition, half of the group was White and half was Black. As participants were asked to
think of themselves as a part of the work group, the participants own race was included as
a part of the group racial composition. There were equal numbers of male and female
coworkers across all conditions.
Social Dominance Orientation (SDO). SDO was measured using fourteen
statements for which respondents indicated whether they had a positive or negative
feeling on a seven-point Likert-type scale ranging from “very negative” to “very
positive” (Pratto et al., 1994) (see Appendix A). Items assessed participant feelings about
statements regarding hierarchical systems where one group dominates another (e.g.,
“Sometimes other groups must be kept in their place”). Higher scores indicated higher
26

levels of SDO. For participants in this study (N=210), this scale had good internal
consistency reliability, α = .87.
Dependent Variables – Level 1(Within-Individual) Measures
The Level 1 dependent variable was measured at three points in time when
observers responded to the bullying incidents: Time 2 (T2), Time 3 (T3), and Time 4
(T4). Thus, there were three observations of behavioral responses for each participant.
Behavioral Response. Bowes-Sperry and O‟Leary-Kelley (2005) characterized
observer responses to sexual harassment along two dimensions: the immediacy of the
response and the level of involvement. In the present study, we incorporated both
dimensions into our conceptualization of observer responding. Participants were
prompted to respond to their group after viewing each video; they were asked to say
something to their group as the bullying was occurring during the meeting (high
immediacy) or have private follow-up conversations with each group member after the
bullying had ended (low immediacy), or both, capturing variation in the immediacy of
their response. Participants were instructed to respond with at least five sentences. Only
responses for T2, T3, and T4 were coded as bullying was not depicted at T1. Open-ended
response formats are preferable to behavioral checklists or ratings of response intentions
because they are less prone to common method biases and socially desirable responding
(Podsakoff, MacKenzie, Lee, & Podsakoff, 2003). Responses were coded by three
coders: two research volunteers who were blind to the intent of the study and the primary
investigator.
Additionally, in line with Bowes-Sperry and O‟Leary-Kelley‟s (2005)
conceptualization of involvement of observer response, each response was coded for its
27

riskiness to the observer. Within their instructions, coders were provided with a definition
of personal risk to the observer along with examples of behaviors associated with risky
responding (for coder instructions see Appendix A). Response riskiness codes included 0
(no attempt to help victim), 1 (non-confrontational with minimal or mild risk), and 2
(confrontational or assertive response with high risk). Non-responses, or responses not
related to the group‟s mistreatment of the victim, were coded as 0. Responses were
coded as 1 or 2 if there was clear intent by the observer to respond to the bullying.
Response that were not confrontational or assertive (e.g., polite) were coded as 1.
Responses that were confrontational or assertive (e.g., rude and aggressive) in that they
put the respondent at greater risk for negative retribution from fellow group members
were coded as 2.
Below is one example of a response that was coded as 0 for non-response:
“I'm glad to hear that we are making progress on the project for the client. What
are the finishing touches we can work on before submitting it? Let’s finish this up
to make the VP happy.”
Below are two examples of responses coded as 1, or as minimally risky to the
observer:
“Have you noticed that Ashley has been singled out of the team? She keeps
getting interrupted and it seems no one wants to sit by her, or hand out an agenda
to her. I would go talk to her and address that this has to stop now.”
“Ashley has something to say, you should listen to her. We need to improve by
including Ashley as a part of the group too. When management comes in, I'll let

28

them know that we did not work effectively as a group. We need to make some
changes.”
Below are two examples of responses coded as 2, or as highly risky to the
observer:
“I do not believe everyone in this meeting is being treated with respect and I do
not understand why. Most of you are participating in treating Ashley in a rude
manner. I don't understand why, but it needs to stop. I want an explanation for
this behavior. Can anyone give me one?”
“Why do you have to ignore everything that Ashley says? She may have a valid
point. Just give her a chance. Your attitude is unacceptable in the workplace.”
Inter-rater agreement was generally high amongst the three coders for both high
and low immediacy responses at times two, three and four. At time two, full agreement
amongst coders was 97% for high immediacy and 94% for low immediacy; at time three,
full agreement amongst coders was 86% for high immediacy and 91% for low
immediacy; at time four, full agreement amongst coders was 89% for high immediacy
and 95% for low immediacy. Coder disagreement was resolved by discussion and
agreement amongst coders facilitated by the primary investigator.
Exploratory Variables
In addition to the above variables hypothesized to be related to observer
responses, we also explored the predictive effects of three additional individual difference
(level 2) variables: Prosocial orientation, trait empathy, and trait perspective taking. We
believed these variables would have an effect on response riskiness but we made no
specific hypotheses about their expected relationships
29

Prosocial Orientation. Prosocial orientation was measured using a 16-item scale
for assessing individual differences in adult prosocialness (Caprara, Steca, Zelli, &
Capanna, 2005). On a seven-point Likert-like scale, ranging from “Never/almost never”
to “Always/ almost always”, respondents indicated the extent to which each statement
reflected their own feelings and behaviors (e.g., “I try to be close to and take care of those
who are in need”) (see Appendix A). Higher scores indicated higher levels of
prosocialness. For participants in this study (N=226), this scale had good internal
consistency reliability, α = .83.
Trait Empathy and Perspective Taking. The seven item perspective taking and six
item empathetic concern subscales from the Interpersonal Reactivity Index (IRI) (Davis,
1983) were used to assess trait perspective taking and empathy, respectively (see
Appendix B). Respondents were asked to indicate how well a series of statements
describes them on a five-point Likert-type scale ranging from (1) “Does not describe me
well” to (5) “Describes me well”. The perspective taking measure reports a tendency to
adopt another‟s psychological point of view in everyday life (e.g., “I sometimes try to
understand my friends better by imagining how things look from their perspective”). The
empathetic concern scale assesses tendencies to experience sympathy and compassion for
unfortunate others (e.g., “I often have tender, concerned feelings for people less fortunate
than me”). Higher scores indicated higher levels of empathetic concern and perspective
taking. For participants in this study, the perspective taking scale (N=213) had good
internal consistency reliability, α = .84, but the empathetic concern scale (N=214) had a
poor internal consistency reliability, α = .62.

30

Results
Manipulation Check
All participants included in the data analysis passed the manipulation check by
correctly indicating the racial composition of their work group and the race of the victim
(N=226).
Descriptive Statistics
Descriptive statistics and inter-correlations among the study variables are
provided in Table 1 of Appendix B. Readers should be aware of two caveats in
interpreting this matrix: Time is excluded from the matrix because it is constant across
individuals and interactions among the level 2 independent variables are also excluded.
High immediacy responses increased in riskiness with time F(2, 224)=59.48, p<.001.
Responses at T3 (t[225]=-8.08, p<.001) and T4 (t[225]=-10.75, p<.001) were
significantly more risky than Responses at T2, and responses at T4 were significantly
more risky than responses at T3, t(225)=-2.78, p<.01. On average, response riskiness
tended to fall below average, or was less than mildly risky, at T2 and T3. Though not
drastically, responses were on average above mild riskiness at T4. Low immediacy
responses also differed in riskiness with time F(2, 224)=5.51, p<.01: Responses at T2
(t[225]=2.81, p<.01) and T3 (t[225]=3.01, p<.01) were significantly more risky than
Responses at T4, but responses at T2 did not differ in riskiness than responses at T3,
t(225)=-0.87, p>.05. On average, low immediacy response riskiness tended to fall below
average, or was less than mildly risky, at all points in time. There appears to be no issues
of range restriction for high and low immediacy responses at T2, T3, or T4.
High Immediacy Response
31

Data were analyzed utilizing Hierarchical Linear Modeling (HLM 6.0;
Raudenbush, Bryk, Cheong, & Congdon, 2004). In HLM, Level 1 variables were
modeled to vary over time, in this case over the three time periods. The dependent
variables (within-individuals) were measured at three points in time and the Level 2
variables (between-individuals) were used to predict variation in the outcome of interest
at each time point. The Level 1 time variables (i.e., T2, T3, and T4) were dummy-coded.
In order to prevent over-paramaterization, the Level 1 intercept was not included as a part
of the model for either dependent variable and slopes for all three time points were
included instead (Raudenbush et al., 2004); the purpose of our analyses was to capture
the effect of each level 2 independent variable across time.
The results predicting riskiness of high immediacy responses are provided in the
left side of Table 2 in Appendix C. The hierarchical linear model was specified as using
the Level 2 variables to predict the slope of the behavioral response at T2, T3, and T4.
Additionally, as gender was not included as a part of the study manipulations and gender
similarity may play a role in observer responses, participant gender was included as a
covariate in all analyses.
At T2 victim race did not significantly predict response riskiness (B = 0.069, ns)
meaning that observers‟ immediate response riskiness did not differ depending on
whether the victim was Black or White. H1 was not supported. Inconsistent with H2,
observers‟ immediate responses did not differ in riskiness depending on whether the
group was mainly Black or balanced (B = -0.015, ns). H3 was supported in that observers
responded more riskily in groups that were mainly White compared to balanced (B =
0.381, p < .01). Inconsistent with H4, a victim race*group composition interaction was
32

not apparent in mainly Black groups (B = 0.090, ns). However, the significant main effect
of the mainly White group appeared to be partially qualified by the significant victim
race*group composition interaction. In support of H5, observers immediately responded
more riskily if the victim in a mainly White group was Black rather than White (B = 0.464, p < .05). Observer levels of SDO (B = -.023, ns) and its interaction with victim
race (B = -0.039, ns), group composition (Black group [B = -0.072, ns], White group [B =
0.173, ns]), or their interactions (victim race*Black group [B = 0.070, ns], victim
race*White group [B = -0.148, ns]) did not significantly predict highly immediate
response riskiness at T2. Therefore, H6 was not supported
At T3, victim race (B = 0.230, ns); group composition (Black group [B = -0.109,
ns], White group [B = 0.098, ns]); ]) and their interactions (victim race*Black group [B =
-0.253, ns], victim race*White group [B = -0.272, ns]); observer SDO (B=-.009, ns), and
SDO interactions with victim race (B = -0.050, ns), group composition (Black group [B =
-0.086, ns], White group [B = 0.022, ns]), or their interactions (victim race*Black group
[B = -0.126, ns], victim race*White group [B = 0.471, ns]) did not significantly predict
high immediacy response riskiness.
Similarly, at T4 victim race (B = -0.021, ns); group composition (Black group [B
= -0.176, ns], White group [B = 0.026, ns]) and their interactions. Also, observer SDO (B
= 0.047, ns) and its interactions with victim race (B = -0.032, ns), mainly White group [B
= -0.061, ns]), or their interactions (victim race*Black group [B = 0.094, ns], victim
race*White group [B = -0.074, ns]) did not significantly predict high immediacy response
riskiness. However, despite having no interactive effect with mainly White group,
observer SDO did significantly interact with mainly Black group (B = -0.606, p < .05):
33

Simple slopes analysis indicated that SDO (B = -0.316) negatively predicted response
riskiness in groups that were mainly Black, F(1, 74) = 6.61, p < .05, but had no predictive
ability in groups that were balanced, F(1, 69) = 0.05, ns, or mainly White, F(1, 71) =
0.09, ns. That is, individuals lower on SDO responded more riskily than individuals
higher on SDO in mainly Black groups at T4. Figure 1 of Appendix C contains a
graphical representation of this interaction.
H7 was partially supported for high immediacy responses. Analysis of slope
intercepts at T2, T3, and T4, indicates that subsequent time points, T3 (B = 1.027, p <
.05) and T4 (B = 1.018, p < .05) were positively related to response riskiness, whereas T2
(B = 0.478, ns) was not. Analysis of means (see Table 1) is consistent with this trend:
response riskiness at T3 (M=0.92) and T4 (M=1.04) are significantly higher than riskiness
at T2 (M=0.48).
Finally, H8 was also partially supported for high immediacy responses. Victim
race, group composition, and their interactions were only significant predictors of
response riskiness at T2, not at T3 and only the interaction between SDO and group
composition was significant at T4. Social category situational cues significantly
influenced response riskiness at T2 and not at later times.
Low Immediacy Response
The results predicting riskiness of the low immediacy response are provided in the
right side of Table 2 in Appendix B. An identical analytical strategy as used with high
immediacy was used for low immediacy responses.
At T2, victim race (B = 0.151, ns); group composition (Black group [B = -0.009,
ns], White group [B = 0.078, ns]) and their interactions (victim race*Black group [B = 34

0.061, ns], victim race*White group [B = -0.290, ns]); observer SDO (B = 0.145, ns) and
its interactions with victim race (B = -0.290, ns), group composition (Black group [B = 0.321, ns], White group [B = -0.333, ns]), or their interactions (victim race*Black group
[B = 0.498, ns], victim race*White group [B = 0.426, ns]) did not significantly predict
riskiness of observer low immediacy responses. Therefore, H1-6 were not supported for
the low immediacy response.
At T3, gender was a significant predictor of low-immediacy response riskiness (B
= 0.226, p < .05), in that females responded more riskily than males. Victim race was also
a significant predictor of low immediacy response riskiness (B = 0.418, p < .01):
observers responded more riskily if the victim was White than Black. Though observers‟
low-immediacy responses did not differ in riskiness depending on whether the group was
mainly Black or balanced (B = 0.103, ns), observers responded more riskily in groups
that were mainly White compared to balanced (B = 0.455, p < .01). This main effect
appeared to be partially qualified by the significant victim race*group composition
interaction, in that observers‟ low immediacy responses were more risky if the victim in a
mainly White group was Black rather than White (B = -0.793, p < .01). A victim
race*group composition interaction was not significant for mainly Black groups (B = 0.333, ns). Observer SDO (B = -0.161, ns) and its interactions with victim race (B =
0.253, ns), group composition (Black group [B = 0.252, ns], White group [B = -.119, ns]),
or their interactions (victim race*Black group [B = -0.222, ns], victim race*White group
[B = 0.080, ns]) did not significantly predict riskiness of observers‟ low immediacy
responses.

35

At T4, victim race (B = 0.038, ns); group composition (Black group [B = -0.032,
ns], White group [B = 0.136, ns]) ]) and their interactions (victim race*Black group [B =
0.023, ns], victim race*White group [B = -0.141, ns]); observer SDO (B = 0.320, ns) and
its interactions with victim race (B = -0.078, ns), group composition (Black group [B = 0.127, ns], White group [B = -0.347, ns]), or their interactions (victim race*Black group
[B = 0.016, ns], victim race*White group [B = -0.067, ns]) did not significantly predict
riskiness of observers‟ low immediacy responses.
H7 was not supported for the low immediacy response. As indicated by the
intercepts of the slopes at T2 (B = 0.343, ns), T3 (B = 0.625, ns), and T4 (B = -0.144, ns),
time was not related to riskiness of observer low immediacy responses.
Finally, H8 was partially supported for the low immediacy response. Victim race,
group composition, and their interactions were only significant predictors of response
riskiness at T3, not at T2 or T4. Similar to high immediacy responses, social category
situational cues do not appear to influence response riskiness after repeated exposure at
T4.
Additional Results
Riskiness for high and low immediate responses was also predicted with the
additional exploratory individual difference variables (prosocial orientation, trait
perspective taking, and trait empathy). Trait empathy had poor reliability (α = .62) with
this sample and the additional variables were not significant predictors of the outcomes of
interest. In order to maintain simplicity, the exploratory variables were excluded from the
final model reported in this paper.

36

DISCUSSION
In the current study we applied a relational demography framework to explore the
influence of racial identity on observer responses to workplace bullying. In order to do
so, we examined the impact of observers‟ racial similarity to a bullied victim and the
racial composition of the work group in which the bullying occurs, and how these factors
change in their impact over time. Observers viewed videos of their work group bullying a
victim at three points in time. They made open-ended responses to the group as each
incident of bullying was occurring and/or after each incident had ended. Results indicated
that racial similarity, group composition, the interaction between racial similarity and
group composition and social dominance orientation all play a role in influencing
observer responses. In particular, racial similarity to one‟s group and the prototypical
nature of a prejudicial interaction play an initial and continued role in observer responses.
Our results both support and extend propositions made by Harrison and colleagues (1998)
who suggested that time moderates the influences of social categories on coworker
interactions: the effects of demography on observer responses persist over time
depending on the nature of the racial context and the immediacy of the observers‟
response.
Summary of Findings
At first exposure to the bullying, group composition influenced observer
responses as the bullying was occurring (high immediacy), but not after the bullying
incident had ended (low immediacy): Observers responded more riskily when their group
was mainly White and when the victim was Black in a mainly White group. During the
second exposure to the bullying, racial similarity and group racial composition did not
37

influence observer responding as the bullying was occurring, but instead influenced low
immediacy responses after the bullying had ended: Observers responded more riskily
after the bullying had ended when the victim was White, when the group was mainly
White, and when the victim was Black in a mainly White group. During the third
exposure to the bullying, racial similarity and group racial composition did not influence
observer responding as the bullying was occurring or after the bulling had ended, except
that when the bullying was occurring individuals high on SDO had lower response
riskiness when in a mainly Black group.
Observer response riskiness also varied with time: As exposure to the bullying
increased, highly immediate responses while the bullying was occurring became more
risky, but low immediate responses after the bullying ended became less risky. The
results of this study indicate that observer responses are influenced by racial identity
categorizations among the individuals present in the immediate bullying situation. The
extent to which racial categorizations influence responses is influenced by time but the
effects of time differ depending on the immediacy of the responses. Our findings are in
line with predictions by Harrison and colleagues (1998) for high immediacy responses:
Relational demographic factors are impactful at initial interaction and become less
important after subsequent exposure; but for low immediate responses, relational
demography effects are not present at the initial interaction and take effect after
subsequent interaction. Overall, our study integrates the relational demography
framework with Bowes-Sperry and O‟Leary-Kelley‟s (2005) conceptualization of
observer responding, while both supporting and extending views of Harrison and

38

colleagues regarding time as a moderator of the effects of relational demography on
coworker interactions.
Theoretical Implications
The above findings highlight the implications that social identity categorizations
have on helping behaviors and how characteristics of the situation can influence social
categorization effects. We believe that the results reported here can be interpreted through
the unique influence that time and response immediacy have on observers‟ interpretation
of their role within a newly interacting social unit.
The general trend in our results across the three bullying exposures was, as time
progressed more observers responded for high immediacy responses, but more observers
chose to do nothing in terms of low immediacy responses. As a result, high immediacy
response riskiness increased with time, while low immediacy response riskiness
decreased with time. According to Bowes-Sperry and O‟Leary-Kelley (2005), highly
immediate responses are intended to stop the aggression as it is occurring. High
immediate responders use a variety of behaviors (e.g., remove the victim from the
situation, ask the perpetrator to stop), but the behaviors are thought to be more effective
than low immediate responses in preventing any additional harm from continuing. When
observers respond after the bullying has ended, their role in preventing additional harm is
less clear. Low immediate responders can choose to console the victim, report the
incident, retaliate or offer suggestions to the aggressor(s) (Bowes-Sperry & O‟LearyKelley, 2005) but the result of their responses are not direct. Results in the present study
suggest that, with increased exposure, expectations of recurrence may increase response
riskiness for responses thought to be more effective in preventing additional harm.
39

During initial exposure to the bullying, our results support arguments made by
Harrison et al., (1998) for high immediate responses: Individuals appear to use superficial
categorizations of their group members as the basis for their response behaviors.
Observers possibly base their attachment to their group on racial similarity (O‟Reilly et
al., 1989; Pelled et al., 1999; Tsui et al., 1989) and thus behave in ways to benefit their
group (Allen & Meyer, 1990; Morrison, 1994; O‟Reilly & Chatman, 1986) particularly in
novel and uncertain situations. Group composition acted as an important determinant of
observer response riskiness at early exposure: mainly White groups had higher response
riskiness, whereas victim race and mainly Black groups had no significant effects. One
possible explanation for these effects is that at first exposure to the bullying observers are
new to the social unit and are unsure about their role within the group. In order to reduce
subjective uncertainty about their self-conceptualization of their place within their new
social unit, observers initially self-categorize based on race, conferring confidence in how
to behave and interpret their social environment (Goldberg et al., 2010; Hogg & Terry,
2000; Reid & Hogg, 2005). At initial exposure to bullying that is clearly disruptive to
group interpersonal functioning, race is salient as a cue for attachment to the group, and
observers racially similar to their group are motivated to overcome the bullying.
Additionally, during initial group interactions, group characteristics as a whole, and not
characteristics of individual group members, serve as a cue for attachment interpretations
that drive helping behaviors; at initial exposure, it thus follows that individual
characteristics of victims (i.e., victim race) may not play an important role in helping
behaviors. However, at initial exposure the interaction between group composition and
victim race appears to activate prototypical conceptualizations of prejudicial treatment
40

that includes Whites discriminating against a person of color (Inman & Baron, 1996).
Observers appeared to be responsive to such contexts acting to overcome the prejudicial
treatment they perceived. Despite racial similarity to the victim not influencing
interpretational cues at initial exposure, the stereotypical nature of the interaction when
the victim is Black in a White group appears to be particularly salient and influential in
observer responding at initial exposure to the bullying. Finally, racial dissimilarity to
one‟s group (i.e., a White observer in a mainly Black group) appears to be less important
than racial similarity to one‟s group (i.e., a White observer in a mainly White group) in
influencing helping behaviors across all time points. Future research will benefit from
investigating the differential influence of demographic similarity versus dissimilarity in
group members‟ cues for group interpersonal functioning.
Responses during initial exposure to the bullying were generally more risky for
low immediate responses than high immediate responses. This suggests that when
observers are new to a group they are less willing to directly involve themselves in the
bullying as it is occurring and are more likely to respond riskily in a less direct manner
after the bullying incident has ended. Indeed, a large proportion of observers engaged in
minimally risky low immediacy responses at initial exposure. Observers thus used this as
an opportunity to alleviate the conflict in a less direct manner. Given the majority of
observers who responded in this fashion, differences based on group composition or
victim race were less likely. At initial exposure, conflict is evident and there is a clear
expectation that some response is required even though the effectiveness of such a
response is not clear. One possibility underlying differential impact of relational
demography on high and low immediate responding could be that observers‟ responses
41

rely less on racial identity cues to aid in their interpretation and behavior within the social
unit when they are more certain about the role they should take. Consistent with theory
proposed by Hogg and Terry (2000), in order to reduce subjective uncertainty,
individuals rely on social category distinctions to aid their interpretation of themselves
within the group. At initial exposure, a low risk response after the bullying has ended is
perceived as less direct, intrusive, and more role-appropriate for a new group member; an
unambiguous interpretation requiring less emphasis on surface cues to infer appropriate
behavior. In the high immediate condition at first exposure, the role of a direct responder
as the bullying is occurring may be less clear because such a response could have a
multitude of implications for the responder (e.g., others may perceive him/her as a leader,
he/she may be targeted as the next victim). After subsequent exposure to the bullying,
however, expectations of recurrence may increase perceptions of responsibility and more
risky high-immediate responses are seen as role-appropriate. After subsequent exposure,
implications of less direct-low immediate responses are less clear, by comparison,
fostering a context where social categorization will influence observer responses, as is
evident in our results at second exposure to the bullying.
During second exposure to the bullying, racial identity categorizations did not
influence high immediacy responses but played an important role in low immediacy
responses. Similar to high immediacy responses during first exposure, being in a mainly
White group and a mainly White group with a Black victim predicted more risky
responses by observers and, additionally, if the victim was White observer responses
were more risky compared to if the victim was Black. Consistent with above
explanations, uncertainty regarding observer role-appropriate behavior in low immediacy
42

responses fosters greater reliance on social category distinctions on which helping
behaviors are based. As observers were able to respond to group members individually
after bullying incidents had ended, characteristics of individual group members played a
more important role in helping behaviors compared to when responding to the group as a
whole. At a dyadic level of responding, when response outcomes are uncertain, the sense
of we-ness (Dovidio et al., 1997) and empathetic altruism (Tafjel & Turner, 1986) of
racial in-group categorizations prompt individuals to respond in helpful ways (Riordan &
McFarlane Shore, 1997). As is evident in our results, attention to racial category
distinctions persists past initial exposure to bullying. Racial category effects subside after
initial exposure for high immediacy responses but emerge after subsequent exposure for
low immediacy responses, with our results suggesting that dyadic in-group favoritism
based on race also emerges with this trend. However, at third exposure, reliance on racial
categorizations is minimal, consistent with research finding that after increased exposure
to diversity, race becomes less important for interpersonal functioning (Glaman, Jones, &
Rozelle, 1996; Martins et al., 2003).
After repeated exposure to diverse others, interpretations of and reactions to
behavior are based more on deeper level differences and less on race (Harrison et al.,
1998), as the results of the current study would suggest. During third exposure to the
bullying, a predominant proportion of observers responded more riskily as the bullying
was occurring, and a large proportion of observers did not respond at all after the bullying
had ended. It is likely that after repeated exposure to the bullying observers are
particularly likely to feel accountable for future mistreatment (Bowes-Sperry & O‟LearyKelley, 2005) and will respond in ways to attenuate the negative effects of the behavior
43

more directly. With role-appropriate behaviors more clear, reliance on racial category
distinctions are less important to reduce subjective uncertainty. One interesting finding
during third exposure was that observers high on SDO responded in a less risky way
during the bullying compared to low SDO observers. If high immediate responses are
perceived as more effective after repeated exposure, individuals who feel less attached to
the group may be less willing to reduce the conflict effectively. In particular for Whites,
racial dissimilarity can accentuate feelings of exclusion (O‟Reilly et al., 1989) because it
is inconsistent with their conceptualization of dominance within organizations (Pelled et
al., 1999; Riordan & McFarlane Shore, 1997; Tsui et al., 1992). Individuals high on SDO
tend to favor hierarchical legitimizing ideologies (Pratto et al. 1994). Whites high on
SDO in groups that are mainly Black are less likely to behave effectively to reduce
mistreatment in a group that is unlike them or predominantly composed of Blacks. Direct
and more risky responses are less appealing to observers concerned with maintaining the
dominance of their salient social identity.
The current study provides evidence for the role of social category group
membership in observer responses to workplace aggression, building on few previous
studies that have explored the role of social categorization on bystander intervention (e.g.,
Levine, Cassidy, Brazier, & Reicher, 2002; Levine & Crowther, 2008; Levine, Prosser,
Evans, & Reicher, 2005). Consistent with these previous studies, our results demonstrate
that being racially similar to fellow observers promotes risky responding. Our research
also suggests that racial similarity to other observers and the race of the victim interact to
promote helping when the interaction is prototypical of a prejudicial interaction.
However, our study extends previous research that has primarily looked at single acts of
44

violence or aggression by exploring the role of victim race and racial similarity across
time and at different response immediacies. This is important in two respects. Firstly, our
high immediacy results fit with the popular injunction that people base behaviors on
social categorizations early on in interactions, but these effects diminish with increased
exposure. Secondly, the fact that this held for high immediate responses, but not low
immediate responses, suggests that other, situation-specific, norms and expectations (e.g.,
effectiveness of a response strategy) may come to shape the behavior of observers and
that these effects differ with time. Finally, we also provide evidence that SDO interacts
with racial categorizations to influence observer responses in a manner consistent with
Whites‟ concern for their dominance in organizations (Pelled et al., 1999; Tsui et al.,
1992).
Surprising in our results was that neither trait perspective taking nor prosocial
orientation predicted observer response riskiness across time. We believe that these
findings may be due to how we conceptualized our dependent variable. Indeed, previous
research linking perspective taking (e.g., Kamdar, McCallistar, & Turban, 2006) and
prosocial orientation (Caprera, Steca, Zelli, & Capanna, 2005) to helping behavior has
analyzed helpers‟ general tendency to take a perspective or be helpful. Perspective taking
and prosocial orientation may relate to helpful tendencies but may not predict the specific
nature of helping behaviors, such as response riskiness. In light of these findings, future
research is clearly needed to delineate the effects of individual differences on observer
responses to workplace bullying.

45

Practical Implications
Organizations stand to suffer considerable losses when employees are victimized
by bullying. Physical and psychological well-being (Leymann, 1990; Leymann &
Gustafsson, 1996; Zapf, Knorz, & Kulla, 1996) tends to be lower and withdrawal
behaviors (e.g., absenteeism and turnover intentions) tend to be higher (Hoel et al., 2003;
Keashly & Jagatic, 2003; Zapf & Gross, 2001) for victims of bullying. As victims of
bullying are rarely likely to confront the aggressor(s) themselves (Einarsen et al., 2003;
Knapp, Faley, Ekeberg, & Dubois, 1997) it is in organizations‟ best interests to
understand factors (e.g., demographic similarity) that promote third party observer
responses to bullying. Toward this end, our findings suggest a couple of practical
implications. First, organizations should pay particular attention to situations wherein
employees are demographically dissimilar to a large portion of their coworkers or on
interactions that are not stereotypically prejudicial. In such situations, diversity training
provided by competent diversity experts may prove especially valuable. This training
may be used to heighten awareness amongst all parties of the potential for
misunderstandings to be interpreted in social category terms. Moreover, educating
participants about how bullying is manifested in the workplace (e.g., identifying
examples of continued exclusionary aggression) should help to create a more objective
and uniform standard for by which observers will respond to bullying.
Second, we encourage organizations to work on creating highly prosocial work
climates (Brief & Motowidlo, 1986). Creating a climate that punishes aggression and
rewards interpersonal support may make observers more likely to put an end to
mistreatment. Training that includes perspective taking may be one strategy that would
46

promote helpful responses of others through increased empathy and by debiasing social
category and in-group favoritism influences on observer responses (Galinsky &
Moskowitz, 2000). Additionally, by training employees about effective response
behaviors, responses may be quicker to the unfolding aggression (a high proportion of
observers did not respond during initial exposure in the current study), sooner
diminishing its continued harm for the victim and organization.
Limitations and Future Research
The findings of the current study should be further validated and extended in a
variety of ways. As with all research, our study has several limitations noted in the
paragraphs below.
One limitation of our study is that it examined observer responses to workplace
bullying in a laboratory rather than a field setting. We opted to test our hypotheses using
a laboratory experiment because past studies assessing observer responses to workplace
aggression have typically been correlational and based on retrospective reports of
responses. Conducting an experimental study allowed us to test relationships between
variables – while maintaining internal validity – that would otherwise be impossible to
manipulate in a field setting. The nature of our experimental design also provided a
strong basis for examining causality while also avoiding common method bias and
confounding effects: We separated our assessment of SDO and observer responses in
time, and our dependent variable was assessed using a behavioral measure with openended responses. We also presented participants with videos scenarios instead of written
scenarios to depict the workplace bullying incidents. Researchers of violence (e.g., Noel
et al., 2008) prefer using video vignettes because participants feel more engaged and
47

respond more realistically than when presented with written scenarios that are typically
used in organizational research of observer responding. We made attempts to ensure that
the experimental design was similar to real world workplace scenarios but further
research is required to assess the external validity of our findings.
In addition to the efforts we made to enhance the design of our study, future
research will help overcome other limitations associated with our study. First, we only
allowed participants to observe their fellow group members for short periods of time.
Although, we believe that is important to examine interpersonal relationships in shortlived work groups, future research is needed to examine the influence of race and
relational demography on observer responses amongst coworkers that have longer
interactional histories. Second, Hopkins and Reicher (1996) argue that experimental
studies have limitations because they predefine the nature of the social context forcing
observers to be passive, powerless to alter or advance the social construction. In our
study, participants were not able to actively engage with their work group and alter the
outcome of the bullying. Future research will benefit from utilizing other methodologies
(e.g., interviews eliciting people‟s own accounts of observer responses) to develop a
better understanding of how observers actively involve themselves in the situation. Third,
our scenario focused on one form of workplace bullying, exclusion, which may result in
different observer behaviors than other forms of bullying, such as more direct verbal
derogation. Fourth, in order to simplify the design of our study we did not manipulate the
gender of the bullied victim and only used a female victim. Early research has indicated
that men are stereotyped as more assertive and aggressive than women (Eagly &

48

Mladinic, 1989), and such stereotypes may have implications for observers‟ willingness
to help a male victim.
Another limitation of our study is that the sample was comprised of relatively
young college-aged students that were primarily female. Although this population is not
typical of the working class, young people are becoming a large part of the working
population and college students often work part-time jobs or have several years of work
experience. It is thus important for organizations to manage young employees and
understand their responses to workplace aggression. Future research should seek to
examine the effects of similar variables on samples with varying demographics like age,
gender and work experience.
A final limitation to our study is more theoretical than methodological. Our
experiment did not examine the precise mechanisms whereby social categorizations
influence observer responses. In our theoretical conceptualization we speculate that
psychological attachment (Allen & Meyer, 1990; O‟Reilly & Chatman, 1986; O‟Reilly et
al., 1989; Tsui et al., 1989) to the social unit and expectations of recurrence (BowesSperry & O‟Leary-Kelley, 2005) are mediators between racial categorizations and helpful
observer responses. However, we did not test such processes. Additional mediating
variables thought to underlie social category relationships include: Awareness of
common fate, perceived similarity, and collective responsibility (Levine et al., 2002), and
people‟s internalized sense of their membership to a particular social category (Crowther
& Levine, 2008; Tajfel & Turner, 1979).
Conclusion
With this research, we have offered an integration of literature on observer
49

responses to workplace aggression with literature on relational demography, and we
presented initial evidence illustrating the dynamic influence of racial categorizations on
the observer response process. This evidence highlights the influential role of
demographic similarity to coworkers and stereotypical nature of interactions, which are
overlooked situational characteristics that impact the nature of observer response
behaviors. Greater attention to group demography could enhance research in this domain,
especially when coupled with a consideration of potential mediating mechanisms linking
racial categorizations to response behaviors. We strongly encourage other scholars pursue
this line of inquiry, to join us in further investigating workplace aggression as a dynamic
process and considering the dynamic impact of antecedents of observer responses.

50

APPENDICES

51

APPENDIX A
Procedural Materials

52

Scenario

CASE: Working in groups (or teams) is a common practice in many organizations. There
is no doubt that you have experience working with a group of people to complete a task,
whether it is a project at work or an assignment for school. In order for groups to be
effective they need to meet periodically in order to remain organized and focused.
In this study, you will be a part of a group of six coworkers who are working together to
complete a project over a period of several months. Your group‟s task is to create a
proposal for one of your company‟s major clients. You and your five fellow group
members hold meetings on a weekly basis to review what has been done and to plan what
needs to be done for the next week. There is no leader for this project and it is expected
that every member of the group work together to complete the task. Meetings are
relatively informal and usually involve a review of objectives and completed tasks and an
informal discussion that involves brainstorming and planning for the next week‟s work.
Your group‟s current project is in the middle stages of completion.

53

Manipulation Check
INSTRUCTIONS: Based on the information in the scenario, pictures, and minibiographies answer the following questions. For each question please indicate the answer
you think is most correct. Feel free to flip back to the scenario and chart to help you
answer the questions.
1) How long have you been working in this organization (your tenure)?
a) One month

b) Four months

c) Six months

d) Two years

e) Three years

2) How long has Thomas been working in this organization (tenure)?
a) One month

b) Four months

c) Six months

d) Two years

e) Three years

3) How long has Ashley been working in this organization (tenure)?
a) One month

b) Four months

c) Six months

d) Two years

e) Three years

4) Of your fellow group members (including you), how many have been working at this
organization (tenure) longer than you have?
a) 1

b) 2

c) 3

d) 4

e) 5

5) Based on her picture, what is Ashley‟s race/ethnicity?
a) Asian

b) Black

c) White

d) Can‟t tell from the picture

6) Based on his picture, what is Thomas‟ race/ethnicity?
a) Asian

b) Black

c) White

d) Can‟t tell from the picture

7) Based on her picture, what is Jessica‟s race/ethnicity?
a) Asian

b) Black

c) White

d) Can‟t tell from the picture

8) How many of your fellow group members (including you) are White (or Caucasian)?
a) 1

b) 2

c) 3

d) 4

e) 5
54

9) Who is sitting directly beside Sam on HIS right-hand side?
a) Jessica b) Ashley c) Stacey d) Thomas e) No one, the seat is empty
10) Who is sitting directly beside Stacey on HER left-hand side?
a) Jessica b) Sam c) Stacey d) Thomas e) No one, the seat is empty
11) How many of your fellow group members (including you) are female?
a) 1

b) 2

c) 3

d) 4

e) 5

12) Based on her picture, approximately how old (in years) would you say Stacey is?
a) 18

b) 21

c) 26

d) 30

e) 40

13) Based on her picture, approximately how old (in years) would you say Jessica is?
a) 18

b) 21

c) 26

d) 30

e) 40

14) Based on his picture, approximately how old (in years) would you say Sam is?
a) 18

b) 21

c) 26

d) 30

e) 40

15) Which one of the following names is not one of your fellow group members‟?
a) Jessica b) Anna c) Ashley d) Sam e) Stacey
16) Which one of your fellow group members is wearing a yellow sweater in the picture?
a) Ashley b) Thomas c) Sam d) Stacey e) Jessica
17) Which one of your fellow group members is wearing a white collared shirt in the
picture?
a) Ashley b) Thomas c) Sam d) Stacey e) Jessica

55

Descriptions of Workplace Scenarios
INSTRUCTIONS: Carefully read the following definition of workplace bullying.
„Workplace bullying is defined as an actor or group of actors repeatedly harassing,
offending, socially excluding or negatively affecting one or more individuals. A power
imbalance is created when the victims perceive they are unable to defend themselves as
targets of persistent systematic anti-social behaviour. Workplace bullying involves at
least two negative acts, weekly or more often, for six or more months in situations where
targets find it difficult to defend against and stop abuse.‟
NEXT: Read each of the following workplace scenarios that involve groups of coworkers
interacting. Indicate the degree to which YOU feel each scenario reflects an incident of
workplace bullying as it is defined above.
Scenario #1
During a team meeting a group of six coworkers were discussing their current project.
When one coworker, Jamie, asked a question several of the other team members
interrupted Jamie by laughing and whispering under their breaths, „stupid question!‟
These comments were made directly towards Jamie.
Scenario # 2
Several times a week a group of six coworkers meet. The group usually engages in casual
conversation during the meeting discussing events that happened at work and gossip
about other coworkers. A member of this group is a coworker named Jamie. During these
conversations several of the group members often make sarcastic and condescending
remarks about Jamie‟s appearance and work habits directly towards Jamie. These group
members have been making these types of comments towards Jamie ever since the group
started meeting around six months ago.
Scenario # 3
A group of six coworkers meets regularly in a company conference room. Almost
always, individuals tend to wait until Jamie takes a seat, and then sit on the other side or
other end of the table from Jamie.
Scenario # 4
A team of coworkers meets weekly to discuss a current project. The meetings occur in the
company‟s conference room where the team of six convenes to reflect on tasks and future
plans. During meetings the coworkers tend to sit away from a coworker Jamie, often
leaving Jamie to sit alone at one end of the table. When documents are distributed to the
team, the group often neglects to give any to Jamie. When Jamie speaks up or asks
questions, other group members tend to ignore Jamie‟s comments. The coworkers have
56

treated Jamie in this way for most meetings since the team originally convened almost
eight months ago.
Scenario # 5
Several times a week a group of six coworkers meet. The group usually engages in casual
and light-hearted conversation discussing events that happened at work and gossip about
other coworkers. A member of this group that often engages in these conversations is a
coworker named Jamie. During these conversations several of the group members often
joke about Jamie‟s appearance and work habits and most of the group laughs, including
Jamie. These group members have been making these types of jokes about Jamie ever
since the meetings started six months ago.

57

Video Scripts
Video 1
Team members Stacey, Thomas, Ashley, Jessica, and Sam file into the conference room.
Thomas holds open the door for everyone. The group stands in a huddle around the table,
and a few converse in small talk, before sitting down at the table. Everyone sits together
as a group, with relaxed body language. The VP has not convened the meeting officially,
so the group continues talking amongst themselves:
-Stacey: (to Ashley) “I wonder if the nice weather we‟ve been having is going to stick
around for a while.”
-Ashley: “I think the weather channel said that we‟re going to get some rain on Thursday.
-Stacey: “Oh, that‟s too bad.”
-Everyone else in group nods to express agreement.
-Thomas: (addressing the group) “Does anyone know what this meeting is about?”
-Sam: “I think it might have to do with the reports due at the end of the week.”
-Jessica: (addressing the group) “I really hope it does not take too long.”
-Ashley: (to Jessica) “Yes, an early lunch would be nice.”
-Stacey: “It‟s one of the bigger clients, so the VP may want to spend extra time on this
project”
-Sam: “We‟ll have to wait and see.”
-Thomas: “We should probably get settled.”
-Everyone moves towards their initial seat (as indicated in the seating chart provided to
the participant)
-Jessica: (looks across the table and notices a stack of white sheets stapled together) “Oh,
this must be the agenda.”
-Everyone leans to grab the papers.
-Jessica hands two papers to Sam, who cannot reach the pile.
-Sam: (to Jessica) “Thank you.” Sam puts one of those papers in front of the participant.
-Everyone has a paper in front of him or her.

58

Video 2
Scene begins similarly to last with the team members filing into the conference room,
about to have a meeting. Ashley is the last to enter. This time, while Thomas still holds
the door open for the majority of the group, when Ashley approaches the door last, he
abruptly lets go as if not to notice her. The group stands in a huddle near the table with
Ashley just outside the huddle, close to the door as she entered last.
-Jessica: (addressing the group) “So, does anyone have anything interesting planned for
the weekend?”
-Thomas: “I‟m hoping to do some camping up north with some of my friends from
school. It will be nice to take a break from work”
-The group nods in agreement
-Ashley: “Well, I was hoping to…”
-Sam: Abruptly interrupts Ashley and turns his back to her, excluding her even more
from the huddle. He then says “This‟ll be a good weekend to laze around and catch-up on
some TV I haven‟t seen in a while.”
-Ashley moves closer to the group and speaks up attempting to be heard and says “That
sounds like…”
-Stacey abruptly interrupts Ashley and turns her back to her, excluding Ashley
completely from the huddle. She then says “That sounds like fun, I guess I won‟t be
seeing you at the gym then.” --Stacey then jokingly nudges Sam
-The group, except Ashley, chuckles.
-Sam: “We should probably get settled”
-The group moves towards the seats they were sitting at the last meeting. As Ashley takes
her seat, Thomas and Jessica leave their seats and move to the other side of the table to sit
beside Sam and the participant. Stacey moves two chairs over to sit beside the participant.
-Ashley sighs and sits down at the table, slumped down looking frustrated.
-Jessica: (looks across the table and notices a stack of white sheets stapled together)
“Here is the agenda for this week”
-Everyone leans to grab the papers.
-Jessica hands two papers to Sam, who cannot reach the pile.
-Sam: (to Jessica) “Thank you.” Sam puts one of those papers in front of the participant.
59

-Ashley stands up and retrieves her own agenda from where Jessica is sitting and then
returns to her seat.
Video 3
This scene begins similarly to last with the team members filing into the conference
room, about to have a meeting. Like before, Ashley is the last to enter. Again, Thomas
holds the door open for everyone else but not Ashley. The group stands in a huddle near
the table with Ashley just outside the huddle, close to the door as she entered last.
The group moves towards the seats they were sitting at the beginning of last meeting. As
Ashley takes her seat, Thomas and Jessica again leave their seats and move to the other
side of the table to sit beside Sam and the participant. Stacey moves two chairs over to sit
beside the participant.
Ashley sighs and sits down at the table, slumped down looking frustrated.
-Jessica: (noticing the stack of papers across the table). “Here is the agenda again.” And
she slides them to each group member sitting close to her. Ashley is too far away to get
one.
-Ashley: (to the group) “Could someone please pass me an agenda?”
-Sam: Ignoring Ashley‟s request, Sam turns his back to Ashley and promptly says to the
group: “I thought last week‟s meeting was very productive. I think the VP is happy with
our progress.”
-Ashley: (to the group) “Excuse me…”
-Thomas interrupts Ashley and turns his back, excluding Ashley even more from the
group: “Yeah, Sam. I feel like we‟ll be wrapping up with this project shortly.”
-Stacey: “Just a few more finishing touches and we‟ll be able to send the proposal to the
client.”
-No one looks up or makes any sort of effort to assist her, so Ashley gets out of her chair
and walks over to get a paper. No one acknowledges that Ashley has moved. After
picking an agenda, Ashley sits down next to Stacey, hoping to be included. Stacey turns
her shoulders away from Ashley.
-Jessica: (addressing the group) Jessica chuckles “Hopefully management agrees!”
-Thomas: (addressing the group) “Anyways let‟s take a look and see what is on the
agenda for today before the VP gets here.”
-The group proceeds to read the agenda.

60

Video 4
This scene begins similarly to last with the team members filing into the conference
room, about to have a meeting. Like before, Ashley is the last to enter. Again, Thomas
holds the door open for everyone else but not Ashley. The group stands in a huddle near
the table with Ashley just outside the huddle, close to the door as she entered last.
-Thomas: (addressing the group) “Let‟s get settled and see what management has planned
for today‟s meeting.”
-This time, the group takes their seats away from Ashley (where they were sitting at the
end of last meeting). Ashley sits alone at her end of the table.
-Jessica: “Here are the agendas.” Jessica slides the agendas to everyone at the table
except Ashley.
-Sam: “We‟re so close to being done they probably want us to wrap up a couple issues.”
-Jessica: “Yeah, I guess. Or, make some final changes.”
-Ashley gets out of her chair and walks over to get an agenda. No one acknowledges that
Ashley has moved. After picking an agenda, Ashley sits down next to Sam, hoping to be
included. Sam turns his shoulders away from Ashley to exclude her.
-Stacey: “Did anyone come up with a way to improve this last part?” (She holds up her
paper, pointing to one of the bullet points.”
-Everyone looks at his or her own paper; Ashley seems to have an idea so she speaks up.
-Ashley: “Oh, yeah I thought that we could improve it if we…” but Ashley is interrupted
by Sam sitting next to her who scoots his chair away and turns his shoulders to the rest of
the group as he says:
-Sam: “I don‟t think that there‟s much to improve upon.”
-Jessica: “I agree, I think our communication as a group works well and we‟ll just wait
and see what management says.”
-Ashley looks like she disagrees and tries to speak up.
-Ashley: “I don‟t think that we do as good...” but she is again interrupted.
-Stacey: “So we can say that we don‟t need to change anything?”
-Ashley: “But…”
-Thomas interrupts Ashley again: “So, moving on to this next point…”
-The group looks down to read their agendas again.
61

Instructions for Coders
The point of this code is to assess the riskiness of each participant‟s response. Participants
who respond in a confrontational or assertive way would be more likely to receive
negative retribution from their co-workers (i.e., this is a more risky response choice). For
example, a participant who says something forthright or assertive during a conversation
(e.g., assertively tells someone to stop what they‟re doing or calls them a jerk) is at
greater risk for negative retribution from his/her coworkers than a participant who coyly
asks a co-worker why they did something (e.g., “Why did you not pass an agenda to
Ashley?”) or politely asks them to stop.
0 = No response (i.e., no attempt to help Ashley or stop the aggression)
1 = Minimally/mildly risky (e.g., “I could not help notice there was tension
between you and Ashley? What happened?” “What‟s going on with
everyone? “Excuse me, but we need to let everyone speak now!”, “Why are
you ignoring Ashley?”, “Why are you avoiding Ashley? We should stop it”
“We need to work together better, it is not ok to ignore Ashley”)
2 = Assertive and highly risky (e.g., “Stacey, quit being a bitch”, “you‟re a jerk”,
“you‟re fired!”, “you guys are acting like a bunch of kids”, “stop bullying
Ashley”)

62

APPENDIX B
Measures

63

Dependent Variable - Open-Ended Response
INSTRUCTIONS: You are in the room during this meeting and you now have an
opportunity to say something or ask questions to your fellow group members. Or, you can
wait for the meeting to end and have a follow-up conversation with each group member.
If you would like to say something or ask questions during the meeting respond to part
(A) below. If you would like to have a private follow-up conversation with one or all of
the group members individually, respond to part (B). If you choose to do both, respond to
both (A) and (B).
(A)
INSTRUCTIONS: You are in the room during this meeting and you now have an
opportunity to say something or ask questions to your fellow group members. Using the
space provided, please write AT LEAST FIVE sentences of what you want to say. Please
be specific as to who your comments are directed (i.e., a specific group member or the
entire group). The more detailed you are the better!
________________________________________________________________________
________________________________________________________________________
(B)
INSTRUCTIONS: Now assume the meeting for this week has ended. If you choose to,
you can have a private follow-up conversation with each of your fellow group members.
Using the space provided, please write AT LEAST FIVE sentences of what you want to
say in your follow-up conversation with each of the following group members. The more
detailed you are the better!
Follow-up conversation with Jessica:
________________________________________________________________________
________________________________________________________________________

64

Demographic Inventory
INSTRUCTIONS: Please answer the following demographic questions about yourself.
Recall that this information will NOT be used to identify you personally. Results will be
presented based upon grouped data and your participation is completely confidential.
1.

What is your gender? (please select one)

_____ Female

_____ Male

2.

What is your age? (please write in)

3.

Which of the following BEST describes your racial or ethnic background?
(please select one)

______

_____ Native American
_____ Asian
_____ Black (African or Caribbean)
_____ White/Caucasian
_____ Hispanic
_____ Other (please specify) ____________________________________
4.

Were you in born in the U.S.? (please select one)

_____ Yes ______No

If you answered “No,” please answer the following questions.
4a. If “No,” where were you born ___________________.
4b. If “No,” how long have you been in the U.S.? ______Years ______Months
5.

How many years (cumulative) of paid work experience in the labour market do
you have? (please select one)

_____ None

_____ 1.5 years – 2 years

_____ 0-6 months

_____ 2 years - 2.5 years

_____ 6 months – 1 year

_____ 2.5 years – 3 years

_____ 1 year – 1.5 years

_____ more than 3 years

In what job(s) did you gain this experience?
____________________________________

65

Social Dominance Orientation (Pratto, Sidanius, Stallworth, & Malle, 1994)
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.

Some groups of people are simply not the equals of others.
Some people are just more worthy than others.
This country would be better off if we cared less about how equal all people were.
Some people are just more deserving than others.
It is not a problem if some people have more of a chance in life than others.
Some people are just inferior to others.
To get ahead in life, it is sometimes necessary to step on others.
Increased economic equality.
Increased social equality.
Equality.
If people were treated more equally we would have fewer problems.
In an ideal world, all nations would be equal.
We should try to treat one another as equals as much as possible (All humans
should be
treated equally.)
It is important that we treat other countries as equals.

66

Prosocial Orientation (Caprara, Steca, Zelli, & Capanna, 2005)
1. I am pleased to help my friends/colleagues in their activities
2. I share the things that I have with my friends
3. I try to help others
4. I am available for volunteer activities to help those who are in need
5. I am emphatic with those who are in need
6. I help immediately those who are in need
7. I do what I can to help others avoid getting into trouble
8. I intensely feel what others feel
9. I am willing to make my knowledge and abilities available to others
10. I try to console those who are sad
11. I easily lend money or other things
12. I easily put myself in the shoes of those who are in discomfort
13. I try to be close to and take care of those who are in need
14. I easily share with friends any good opportunity that comes to me
15. I spend time with those friends who feel lonely
16. I immediately sense my friends‟ discomfort even when it is not directly
communicated to me

67

Empathetic Concern (Davis, 1983)
1. I often have tender, concerned feelings for people less fortunate than me.
2. Sometimes I don‟t feel very sorry for other people when they are having problems. (R)
3. When I see someone being taken advantage of, I feel kind of protective towards them.
4. Other people‟s misfortunes do not usually disturb me a great deal. (R)
5. When I see someone being treated unfairly, I sometimes don‟t feel very much pity for
them. (R)
6. I would describe myself as a pretty soft-hearted person.
Note: (R) indicates scored in reverse fashion.

68

Perspective Taking (Davis, 1983)
1. I sometimes find it difficult to see things from the “other guy‟s” point of view. (R)
2. I try to look at everyone‟s side of the disagreement before I make a decision.
3. I sometimes try to understand my friends better by imagining how things look from
their perspective.
4. If I‟m sure I‟m right about something, I don‟t waste much time listening to other
people‟s arguments. (R)
5. I believe that there are two sides to every question and try to look at them both.
6. When I‟m upset at someone, I usually try to “put myself in his shoes” for a while.
7. Before criticizing somebody, I try to imagine how I would feel if I were in their place.
Note: (R) indicates scored in reverse fashion.

69

APPENDIX C
Tables and Figures

70

Table 1
Descriptive Statistics and Inter-correlation of study variables
Variable
1 Gender (male = 0, female = 1)

1
---

2

3

4

5

6

7

8

9

2 Victim Race (Black = 0, White = 1)
3 Black Group (Black = 1, all else = 0)
4 White Group (White = 1, all else = 0)

0.08
0.00
0.11

--0.03
0.02

--0.51**

---

5 Social Dominance Orientation

-0.09

0.03

-0.01

0.07

---

6
7
8
9

High immediacy T2
Low immediacy T2
High immediacy T3
Low immediacy T3

-0.02
0.02
-0.08
0.15*

0.05
0.04
0.03
0.05

-0.04
0.01
-0.13*
-0.04

0.12
0.04
0.03
0.09

10 High immediacy T4
11 Low immediacy T4

0.00
-0.01

0.01
0.01

-0.08
-0.03

M

0.80

0.49

SD

0.40

0.50

10

11

-0.01
-0.07
0.01
-0.08

--0.05
0.11
0.12

--0.24**
0.24**

---0.02

---

0.02
0.04

-0.07
0.09

0.04
0.11

0.16* 0.34** 0.08
--0.20** 0.06 0.36** 0.13

0.35

0.33

2.04

0.48

0.70

0.92

0.70

1.04 0.55

0.48

0.47

0.56

0.56

0.57

0.66

0.66

0.59 0.66

---

Note: *p < .05, **p < .001. Because time is invariant between individuals it is excluded from the table. High immediacy and low
immediacy coded as 0=no involvement, 1=low involvement, 2=high involvement

71

Table 2
Hierarchical Linear Modeling results predicting response riskiness
High Immediacy
Variable
Time 2 slope, B1
Gender (male = 0, female =
1)
Victim race (Black = 0,
White = 1)
Black group (Black = 1, all
else = 0)
White group (White = 1, all
else = 0)

Low Immediacy

B
0.478

SE
0.421

T statistic
1.137

B
0.343

SE
0.436

T statistic
0.787

-0.056

0.100

-.554

0.020

0.104

0.197

0.069

0.140

0.496

0.151

0.144

1.048

-0.015

0.138

-.109

-0.009

0.142

-0.066

0.381

0.137

2.782**

0.078

0.142

0.551

-0.023

0.186

-0.124

0.145

0.193

0.749

SDO
Victim race*Black group

0.090

0.194

0.464

-0.061

0.201

-0.305

Victim race*White group

-.464

0.196

-2.372*

-0.290

0.203

-1.434

Victim race*SDO

-0.039

0.259

-.159

-0.290

0.257

-1.126

Black group*SDO

-0.072

0.261

-0.277

-0.321

0.270

-1.192

White group*SDO

0.173

0.246

0.703

-0.333

0.255

-1.308

Victim race*Black
group*SDO

0.070

0.355

0.196

0.498

0.368

1.355

Victim race*White
group*SDO

-0.148

0.345

-0.428

0.426

0.357

1.195

1.027

0.421

2.439*

0.625

0.436

1.434

-0.125

0.101

-1.248

0.226

0.104

2.180*

0.230

0.140

1.647

0.418

0.144

2.898**

-0.109

0.138

-0.789

0.103

0.142

0.726

0.098

0.137

0.713

0.455

0.142

3.211**

-0.009

0.186

-0.050

-0.161

0.193

-0.835

-0.253

0.194

-1.303

-0.333

0.201

-1.661

Time 3 slope, B2
Gender (male = 0, female =
1)
Victim race (Black = 0,
White = 1)
Black group (Black = 1, all
else = 0)
White group (White = 1, all
else = 0)
SDO
Victim race*Black group

72

Table 2 continued
Victim race*White group

-0.272

0.196

-1.387

-0.793

0.202

-3.914**

Victim race*SDO

-0.050

0.249

-0.202

0.253

0.257

0.982

Black group*SDO

-0.086

0.261

-0.331

0.252

0.257

0.982

White group*SDO

0.022

0.246

0.088

-0.119

0.255

-0.468

Victim race*Black
group*SDO

-0.126

0.355

-0.354

-0.222

0.357

-0.623

Victim race*White
group*SDO

0.471

0.345

1.366

0.080

0.368

0.217

1.018

0.421

2.417*

-0.144

0.436

-0.330

-0.010

0.101

-0.098

-0.016

0.104

-0.155

-0.021

0.140

-0.152

0.038

0.144

0.266

-0.176

0.137

-1.283

-0.032

0.142

-0.226

0.026

0.137

0.189

0.136

0.142

0.957

Time 4 slope, B3
Gender (male = 0, female =
1)
Victim race (Black = 0,
White = 1)
Black group (Black = 1, all
else = 0)
White group (White = 1, all
else = 0)

0.047

0.186

0.252

0.320

0.192

1.657

SDO
Victim race*Black group

0.094

0.194

0.484

0.023

0.201

0.116

Victim race*White group

-0.074

0.196

-0.380

-0.141

0.203

-0.698

Victim race*SDO

-0.032

0.249

-0.127

-0.078

0.257

-0.303

Black group*SDO

-0.606

0.261

-2.324*

-0.127

0.270

-0.472

White group*SDO

-0.061

0.246

-0.248

-0.347

0.254

-1.361

Victim race*Black
group*SDO

0.492

0.355

1.385

0.016

0.368

0.044

Victim race*White
group*SDO

0.121

0.344

0.350

-0.067

0.357

-0.189

Note: *p < .05, **p < .01. B = unstandardized hierarchical linear modeling coefficient

73

Figure 1. Results for the interaction between SDO and group racial composition
predicting high immediacy response riskiness at T4. High SDO is 1 SD above the
mean and low SDO is 1 SD below the mean.

74

REFERENCES

75

REFERENCES
Abrams, D., & Hogg, M.A. (1988). Comments on the motivational status of self-esteem
in social identity and intergroup discrimination. European Journal of Social
Psychology, 18, 317-334.
Allen, N.J. & Meyer, J.P. (1990). The measurement and antecedents of affective,
continuance and normative commitment to the organization. Journal of
Occupational Psychology, 63. 1-8.
Andersson, L. M., & Pearson, C. M. (1999). Tit for tat? The spiraling effect of incivility
in the workplace. Academy of Management Review, 24, 452-471.
Aquino, K. & Thau, S. (2009). Workplace victimization: Aggression from the target‟s
perspective. Annual Review of Psychology, 60, 717 – 741.
Austin, J.R. (1997). A cognitive framework for understanding demographic influences in
groups. International Journal for Organizational Analysis, 5, 342 – 359.
Baron, R. A., & Neuman, J. H. (1996). Workplace violence and workplace aggression:
Evidence on their relative frequency and potential causes. Aggressive Behavior,
23, 161-173.
Bowes-Sperry, L., & O'Leary-Kelly, A. M. (2005). To act or not to act: The dilemma
faced by sexual harassment observers. Academy of Management Review, 30, 288306.
Bowling, N. A., & Beehr, T. A. (2006). Workplace harassment from the victim‟s
perspective: A theoretical model and meta-analysis. Journal of Applied
Psychology, 91, 998-1012.
Brewer, M.B. & Brown, R.J. (1998). Intergroup Relations. In The handbook of social
psychology, Vols. 1 and 2 (4th ed.). (pp. 554-594). New York, NY, US: McGrawHill.
Brief, A.P. & Motowidlo, S.J. (1986). Prosocial organizational behaviors. Academy of
Management Review, 11, 710-725.
Byrne, D. (1971). The attraction paradigm. New York: Academic Press.
Caprara, G.V., Steca, P., Zelli, A., & Capanna, C. (2005). A new scale for measuring
adults‟ prosocialness. European Journal of Psychological Assessment, 21, 77-89.
Clarkson, P. (1996). The bystander (An end to innocence in human relationships?).
Philadelphia, PA: Whurr Publishers.

76

Davis, M.H. (1983). Measuring individual differences in empathy: Evidence for a
multidimensional approach. Journal of Personality and Social Psychology, 44,
113–126.

Den Hartog, D.N., De Hoogh, A.H.B. & Keegan, A.E. (2007). The interactive effects of
belongingness and charisma on helping and compliance. Journal of Applied
Psychology, 92, 1131 – 1139.
Dovidio, J. F., Gaertner, S. L., Validizic, A., Matoka, K., Johnson, B., & Frazier, S.(
1997). Extending the benefit of recategorization: Evaluations, self-disclosure, and
helping. Journal of Experimental Social Psychology, 33,401-442.
Eagly, A.H. & Mladinic, A. (1989). Gender stereotypes and attitudes toward women and
men. Personality and Social Psychology Bulletin, 15, 543-558.
Einarsen, S. (2000). Bullying and harassment at work: Unveiling an organizational taboo,
In Transcending Boundaries: Integrating People, Processes, and Systems
Conference, 6-8 September, 2001 (pp. 7-14). Brisband, Queensland, Australia:
Griffith University School of Management.
Einarsen, S., Hoel, H., Zapf, D., & Cooper, C.L. (2003). Bullying and emotional abuse in
the workplace: International perspectives in research and practice. Taylor &
Francis: New York.
Faul, F., Erdfelder, E., Lang, A. & Buchner, A. (2007). G*Power 3: A flexible statistical
power analysis program for the social, behavioral, and biomedical sciences.
Behavior Research Methods, 39, 175 – 191.
Fitzgerald, L.F., Drasgow, F., Hulin, C.L., Gelfand, M.J., & Magley, V.J. (1997).
Antecedents and consequences of sexual harassment in organizations: A test of an
integrated model. Journal of Applied Psychology, 82, 578 – 589.
Frey, K.S., Hirschstein, M.K., Snell, J.L., Edstrom, L.V.S., MacKenzie, E.P., &
Broderick, C.J. (2005). Reducing Playground Bullying and Supporting Beliefs:
An Experimental Trial of the Steps to Respect Program. Developmental
Psychology, 41, 479 – 491.
Galinsky, A. D., & Moskowitz, G. B. (2000). Perspective taking: Decreasing stereotype
expression, stereotype accessibility and in-group favoritism. Journal of
Personality and Social Psychology, 78, 708-724.
Glaman, J., Jones, A. & Rozelle, R. (1996). The effects of co-worker similarity on the
emergence of affect in work teams. Group & Organization Management, 21, 192215.

77

Glomb, T. M. (2002). Workplace anger and aggression: Informing conceptual models
with data from specific encounters. Journal of Occupational Health Psychology,
7, 20-36.
Goldberg, C.B., Riordan, C., & Schaffer, B.S. (2010). Does social identity theory
underlie relational demography? A test of the moderating effects of uncertainty
reduction and self-enhancement on similarity effects. Human Relations, 63, 903926.
Harrison, D.A., Price, K.H., & Bell, M.P. (1998). Beyond relational demography: Time
and the effects of surface and deep level diversity on work group cohesion.
Academy of Management Journal, 41, 96-107.
Hawkins, D.L., Pepler, D. J., & Craig, W.M. (2001). Naturalistic observations of peer
intervention in bullying. Social Development, 10, 512 – 527.
Hoel, H., Einarsen, S., & Cooper, C.L. (2003) Organizational effects of bullying, In
Bullying and emotional abuse in the workplace: International perspectives in
research and practice. Taylor & Francis: New York.
Hoel, H., Rayner, C., & Cooper, C. L. (1999). "Workplace bullying." In International
Review of Industrial and Organizational Psychology, In C. L. Cooper & I. T.
Robertson (Eds.): 195-230. John Wiley & Sons.
Hogg, M.A., & Abrams, D. (1990). Social motivation, self-esteem, and social identity. In
D. Abrams & M.A. Hogg (Eds.) Social identity theory: Constructive and critical
advances: 28-47. London: Harvester: Wheatsheaf.
Hogg, M.A., & Abrams, D. (1993). Social Towards a single process uncertainty reduction
model of social motivation in groups. In M.A. Hogg & D. Abrams (Eds.) Group
motivation: Social psychological perspectives: 173-190. London: Harvester:
Wheatsheaf.
Hogg, M.A. & Terry, D.J. (2000). Social identify and self categorization processes in
organizational contexts. Academy of Management Review, 25, 121 – 140.
Hopkins, N., & Reicher, S. (1996). The construction of social categories and processes of
social change: Arguing about national identities. In G. Breakwell & E. Lyons
(Eds.), Changing European identities: Social psychological analyses of social
change (pp. 69-93). Oxford, UK: Buttenvorth-Heinemann.
Ilies, R., Hauserman, N., Scwochau, S., & Stibal, J. (2003). Reported incidence rates of
work-related sexual harassment in the United States: Using meta-analyses to
explain reported rate disparities. Personnel Psychology, 56, 607 – 631.
Inman, M.L. & Baron, R.S. (1996). Influence of prototypes on perceptions of justice.
Journal of Personality and Social Psychology, 70, 727 – 739.
78

Kamdar, D., McAllister, D.J., & turba, D.B. (2006). “All in a day‟s work”: How follower
individual differences and justice perceptions predict OCB role definitions and
behaviors. Journal of Applied Psychology, 91, 841-855.
Kanter, R.M. (1977). Some effects of proportions on group life: Skewed sex ratios and
responses to token women. American Journal of Sociology, 82, 965 – 990.
Katz, D., & Kahn, R. L. (1978). The social psychology of organizations, 2nd Ed. New
York, NY: John Wiley and Sons.
Keashly, L. (2001). Interpersonal and systemic aspects of emotional abuse at work: The
target‟s perspective. Violence and Victims, 16, 233 – 268.
Keashly, L. & Jagatic, K. (2003). US perspectives on workplace bullying, In Bullying and
emotional abuse in the workplace: International perspectives in research and
practice. Taylor & Francis: New York.
Knapp, D. E., Faley, R. H., Ekeberg, S. E., & Dubois, C. L. Z. (1997). Determinants of
target responses to sexual harassment: A conceptual framework. Academy of
Management Review, 22, 687–729.
Latane, B., & Darley, J. (1970). The unresponsive bystander: Why doesn't he help? New
York: Appleton-Century-Crofts.
Levine, M., Cassidy, C., Brazier, G., & Reicher, S. (2002). Self-categorization and
bystander non-intervention: Two experimental studies. Journal of Applied Social
Psychology, 7, 1452 – 1463.
Levine, M. & Crowther, S. (2008). The responsive bystander: How social group
membership and group size can encourage as well as inhibit bystander
intervention. Journal of Personality and Social Psychology, 95, 1429-1439.
Levine, M., Prosser, A., Evans, D., & Reicher, S. (2005). Identity and emergency
intervention: How social group membership and inclusiveness of group
boundaries shapes helping behavior. Personality and Social Psychology Bulletin,
31, 443–453.
Leymann, H. (1990). Mobbing and psychological terror at workplaces. Violence and
Victims, 5, 119 – 126.
Leymann, H. (1996). The content and development of mobbing at work. European
Journal of Work and Organizational Psychology, 5, 164 – 185.
Leymann, H. & Gustafsson, A. (1996). Mobbing at work and the development of posttraumatic stress disorders. European Journal of Work and Organizational
Psychology, 5, 251- 275.
79

Lutgen-Sandvik, P., Tracey, S.J., & Alberts, J.K. (2007). Burned by bullying in the
American workplace: Prevalence, perception, degree and impact. Journal of
Management Studies, 44, 837 – 861.
Martins, L.L., Milliken, F.J., Wiesenfield, B.M., & Sadalgo, S.R. (2003). Raciothnic
diversity and group members‟ experiences. Group and Organization Management,
28, 75 – 106.
Merchant, V. & Hoel, H. (2003) Investigating complaints of bullying, In Bullying and
emotional abuse in the workplace: International perspectives in research and
practice. Taylor & Francis: New York.
Namie, R. & Namie, G. (2007). The Workplace Bullying Institute U.S. Workplace
Bullying Survey. Retrieved Sept 25, 2008, from
http://bullyinginstitute.org/zogby2007/wbi-zogby2007.html.
Neuman, J.H., & Baron, R.A. (2005). Aggression in the workplace: A socialpsychological perspective. In S. Fox & P.E. Spector (Eds.), Counterproductive
work behavior: Investigations of actors and targets (pp. 13-40). Washington DC:
American Psychological Association.
Noel, N.E., Maisto, S.A., Johnson, J.D., Jackson, L.A., Goings, C.D., & Hagman, B.T.
(2008). Development and validation of videotaped scenarios: A method for
targeting specific participant groups, Journal of Interpersonal Violence, 23, 419436.
O‟Reilly, C.A., Caldwell, D.F., & Barnett, W.P. (1989). Workgroup demography, social
integration, and turnover. Administrative Science Quarterly, 34, 21 – 37.
O'Reilly, C. & Chatman, J. (1986). Organizational commitment and psychological
attachment: The effects of compliance, identification and internalization on pro-social
behaviour. Journal of Applied Psychology, 71,492-499.

Organ, D. W. (1988). Organizational citizenship behavior: The good soldier syndrome.
Lexington, MA: Lexington.
Organ, D. W., Podsakoff, P. M., & MacKenzie, S. B. (2006). Organizational citizenship
behavior: Its nature, antecedents, and consequences. Thousand Oaks, CA: Sage.
Pelled, L.H., Ledford, G.E., & Mohrman, S.A. (1999). Demographic dissimilarity and
workplace inclusion. Journal of Management Studies, 36, 1013 – 1036.
Pfeffer, J. (1982). Organizations and organization theory. Marshfield, Mass: Pitman
Publishing.
Pfeffer, J. (1983). Organizational demography. In L.L. Cummings & B.M. Staw (Eds.)
Research in organizational behavior, 5, 299-357. Greenwhich, Conn: JAI Press.
80

Piliavin, I.M., Rodin, J. & Piliavin, J.A. (1969). Good samaritanism: An underground
phenomenon? Journal of Personality and Social Psychology, 13, 289 – 299.
Podsakoff, P.M., MacKenzie, S.B., Lee, JY. & Podsakoff, N.P. (2003). Common method
bias in behavioral research: A critical review of the literature and recommended
remedies. Journal of Applied Psychology, 88, 879 – 903.
Pratto, F., Sidanius, J., Stallworth, L.M., Malle, B.F. (1994). Social dominance
orientation: A personality variable predicting social and political attitudes.
Journal of Personality and Social Psychology, 67, 741 – 763.
Raudenbush, S. W., Bryk, A. S., Cheong, Y. F., & Congdon, R. T., Jr. (2004). HLM 6:
Hierarchical linear and nonlinear modeling. Chicago: Scientific Software
International.
Raver, J.L., & Barling, J. (2008). Workplace aggression and conflict: Constructs,
commonalities, and challenges for future inquiry. In C.K.W. De Dreu and M.J.
Gelfand (Eds.), The psychology of conflict and conflict management in
organizations (pp. 211-244). Mahwah, NJ: Lawrence Erlbaum Associates.
Rayner, C., & Keashly, L. (2005). Bullying at work: A perspective from Britain and
North America. In S. Fox & P.E. Spector (Eds.), Counterproductive work
behavior: Investigations of actors and targets (pp. 271-296). Washington, DC:
American Psychological Association.
Reid, S.A. & Hogg, M.A. (2005). Uncertainty reduction, self-enhancement, and in-group
identification. Personality and Social Psychology Bulletin, 31, 804-817.
Riordan, C. M. (2000). Relational demography within groups: Past developments,
contradictions, and new directions. In G. R. Ferris (Ed.), Research in personnel
and human resources management (Vol. 19, pp. 131-173). Greenwich, CT: JAI.
Riordan, C.M. & McFarlane Shore, L. (1997). Demographic diversity and employee
attitude: An empirical examination of relational demography within work units.
Journal of Applied Psychology, 82, 342 – 358.
Richards, J. & Delay, H. (2003). Bullying Policy, In Bullying and emotional abuse in the
workplace: International perspectives in research and practice. Taylor & Francis:
New York.
Robinson, S. L., & Bennett, R. J. (1995). A typology of deviant workplace behaviors: A
multidimensional scaling study. Academy of Management Journal, 38, 555-572.
Rosenberg, M. (1965). Society and The Adolescent Self-Image. Princeton, N.J.: Princeton
University Press.
Sackett, P. R., & DeVore, C. J. (2001). Counterproductive behaviors at work. In N.
Anderson, D. S. Ones, H. K. Sinangil, & C. Viswesvaran (Eds.), Handbook of
81

industrial, work, and organizational psychology (Vol. 1, pp. 145-151). Thousand
Oaks, CA: Sage.
Salin, D. (2003). Ways of explaining workplace bullying: A review of enabling,
motivating and precipitating structures and processes in the work environment.
Human Relations, 56, 1213 – 1232.
Skarlicki, D. P., & Kulik, C. T. (2005). Third-party reactions to employee
(mis)treatment: A justice perspective. US: Elsevier Science/JAI Press.
Smith, C. A., Organ, D. W., & Near, J. P. (1983). Organizational citizenship behavior: Its
nature and antecedents. Journal of Applied Psychology, 68, 653-663.
Tajfel, H., & Turner, J.C. (1986). The social identity theory of intergroup behaviour. In S.
Worchel & W. Austin (Eds.), Psychology of intergroup relations (pp. 33-48).
Chicago: Nelson-Hall.
Thau, S., Aquino, K. & Poortvliet, P.M. (2007). Self-defeating behaviors in
organizations: The relationship between thwarted belonging and interpersonal
work behaviors. Journal of Applied Psychology, 92, 840 – 847.
Tsui, A.S., Egan, T., & O‟Reilly, C.A. (1992). Being different: Relational demography
and organizational attachment. Administrative Science Quarterly, 37, 549 – 579.
Tsui, A.S., & O‟Reilly, C.A. (1989). Beyond simple demographic effects: The
importance of relational demography in superior-subordinate dyads. Academy of
Management Journal, 32, 402-423.
Turner, J.C., Hogg, M.A., Oakes, P.J., Reicher, S.D., & Wetherell, M.S. (1987).
Rediscovering the social group: A self-categorization theory. Oxford: Blackwell.
Vartia, M. (2001). Consequences of workplace bullying with respect to well-being of its
targets and the observers of bullying. Scandinavian Journal of Work Environment
and Health, 27, 63 – 69.
Williams, L. J., & Anderson, S. E. (1991). Job satisfaction and organizational
commitment as predictors of organizational citizenship and in-role behaviors.
Journal of Management, 17, 601-617.
Wispe, L.G. & Freshley, H.B. (1971). Race, sex, and sympathetic helping behavior.
Journal of Personality and Social Psychology, 17, 59 – 65.
Zapf, D., & Gross, C. (2001). Conflict escalation and coping with workplace bullying: A
replication and extension. European Journal of Work and Organizational
Psychology, 10, 497-522.

82

Zapf, D., Knors, C., & Kulla, M. (1996). On the relationship between mobbing factors,
job content, the social work environment, and health outcomes. European Journal
of Work and Organizational Psychology, 5, 212 – 237.

83