Professional Documents
Culture Documents
Rothfuss Thesis Formatted 112817
Rothfuss Thesis Formatted 112817
WOMEN’S PERFORMANCE
A Thesis
MASTER OF ARTS
in
Psychology
(Industrial/Organizational)
by
FALL
2017
THE EFFECTS OF HOSTILE AND BENEVOLENT SEXISM ON MEN’S AND
WOMEN’S PERFORMANCE
A Thesis
by
Approved by:
____________________________
Date
ii
Student: Melanie Lauren Rothfuss
I certify that this student has met the requirements for format contained in the
University format manual, and that this thesis is suitable for shelving in the Library and
Department of Psychology
iii
Abstract
of
WOMEN’S PERFORMANCE
by
Statement of Problem
Although overt displays of sexism have decreased over the years, gender disparity
in the workplace still exists. Sexism research has traditionally focused on hostile
attitudes and behaviors towards women, but a more subtle and subjectively positive
form of sexism, called benevolent sexism, may help explain such gender disparity. The
present study examined the effects of both hostile and benevolent sexism on men and
women during a simulated job selection task, where participants imagined that they
benevolent, or neutral language. Specifically, the study assessed the effects of sexism
determine whether they would moderate the effect of sexism on task performance.
iv
Sources of Data
Sacramento State. They were recruited through Sacramento State’s psychology subject
pool and received research participation credits towards satisfying the department’s
Conclusions Reached
performance self-esteem for men or for women. Interestingly, the women in the study
perceived the benevolently sexist language as hostile, which may have helped protect
them against its negative effects. Further research is needed to understand the effects of
benevolent sexism and to determine which factors may protect women against its
harmful effects.
_______________________
Date
v
ACKNOWLEDGEMENTS
I would like to acknowledge some of the people who supported and encouraged
me during this thesis process. Thank you to my advisor, Dr. August, for guiding me
through my graduate studies, for always offering words of support, and for helping me
become a better researcher. Thank you also to my committee members: Dr. Harrison for
helping me find the inspiration for my thesis, and Dr. Bohon for your thoughtful
Jaime, and Lyla, for your hard work running sessions and your willingness to play the
“sexist recruiter.” Thank you to my boyfriend, Ryan, for being there through all the ups
and downs of this process and for being my rock. Finally, I would like to thank my
parents for your endless love and support, and for paving the way for me to achieve my
educational goals.
vi
TABLE OF CONTENTS
Page
Acknowledgements ...................................................................................................... vi
List of Figures................................................................................................................ x
Chapter
1. INTRODUCTION ..................................................................................................... 1
2. BACKGROUND ....................................................................................................... 6
3. METHOD ................................................................................................................ 15
Participants ........................................................................................................... 15
Materials ............................................................................................................... 15
Procedure .............................................................................................................. 20
4. RESULTS ................................................................................................................ 25
5. DISCUSSION.......................................................................................................... 36
vii
Implications and Future Research ........................................................................ 45
References ................................................................................................................... 90
viii
LIST OF TABLES
Tables Page
ix
LIST OF FIGURES
Figures Page
x
1
CHAPTER 1
INTRODUCTION
practices in employment settings against women, gender disparities still exist. The
passage of the Civil Rights Act of 1964, which prohibited discrimination based on sex in
addition to race, color, religion, or national origin, marked a major victory for women’s
rights. The Civil Rights Act provided equality to women “with respect to compensation,
terms, conditions, or privileges of employment” (HR 7152). Since the act was passed,
however, many seem to believe that sexism is no longer an issue; one common belief is
that because legislation provides women with equal rights, women have achieved gender
equality. While it is true that overt displays of sexism have decreased over the years
(Dovidio & Gaertner, 1998), gender inequality, especially in the workplace, is still a
The gender pay gap is one manifestation of sexism in the workplace. According to
the Bureau of Labor Statistics, women working full-time on average earn only 82% of
their male counterpart’s earnings. Compared to all men employed full-time, this number
is even lower for Black women (70%) and Hispanic women (64%). When controlling for
race, the largest earning differences exist between Asian men and Asian women (78%),
and White men and White women (81%). It is important to note that overall, Whites and
Asians’ salaries are around 45% higher than Latinos and Blacks. Thus, as income
increases so does the pay gap. Although salary inequities have improved since 1979,
2
when women earned only 62% of men’s earnings, the earnings ratio has not increased
significantly since 2004 and has continued to stay around 80 to 83% (US Department of
Labor, Bureau of Labor Statistics, 2016). Progress has therefore slowed and more work is
Another issue is that men tend to hold more powerful workplace positions than
women. While it is estimated that around 51% of all management jobs are held by
found in lower levels of their organizations (Blau et al., 2001). For example, in the
United States, women make up only 20% of board of directors and only 6% of CEOs
(Catalyst, 2017). Women are also more likely to have lower starting salaries and have
fewer salary increases with age compared to men (Goldberg, Finkelstein, Perry, &
Konrad, 2004). These occurrences may be explained by obstacles such as the “glass
ceiling,” where it is difficult for women to receive promotions above a certain level, and
“sticky floors,” where women are restricted from moving out of entry-level positions.
Together, these phenomena result in women holding positions with less status and power
than men.
An argument against the existence of the gender gap is that the statistics often do
not control for variables such as occupation, education levels, flex time, or experience.
One might argue that women on average earn less because they are less qualified for
higher ranking positions, or choose to work in industries that traditionally offer lower
pay. As reported by the Bureau of Labor Statistics, the three most common occupations
for women in 2016 were elementary and middle school teacher, registered nurse, and
3
secretary or administrative assistant. Part of the problem is that occupations that are
predominantly female, such as teachers and nurses, tend to be paid less than comparable
jobs that are held primarily by men (Spector, 2008). Women often have jobs that are
valued less and have lower status, and are consequently compensated at a lower rate
(Blau et al., 2001; Xu & Leffler, 1996). For instance, a job analysis with a comparable
worth study, which assesses a job’s contribution to the organization, would likely find
that nurses are underpaid with respect to physicians (Spector, 2008). However, even in
occupations that are predominantly female there is a pay gap. The US Labor statistics
show that female RNs make around 91% of their male counterparts’ earnings, and female
administrative assistants make around 85%. Support against the gender wage gap may
also come from one study of Dutch women, which found that the women were less
ambitious, less concerned about salary, and cared more about work-family balance than
men (Van Vienen & Fischer, 2002). While this finding may explain why women tend to
hold lower positions, it does not explain why women who aim for higher ranks have a
study found that when people are asked to describe characteristics of men, women, and
managers, their descriptions of men and managers overlap, but their descriptions of
women and managers do not (Schein, Mueller, Lituchy, & Liu, 1996). Women may
possessing stronger leadership qualities, and as a result appear better suited for the job.
Such gender stereotypes put women at a disadvantage for job advancement. Stereotypes
4
of women as delicate or less competent are also damaging. Research has found that
women are often given fewer challenging developmental work experiences than men,
despite expressing equal interest in having challenging work (King et al., 2012). Women
may also be passed over for demanding jobs or overseas assignments due to the belief
that women have more responsibilities at home. Evidence of gender discrimination in the
workplace is reflected in the number of individual charge filings with the Equal
reporting charge filings), the percentage of claims based on sex discrimination have
remained consistently around 29% to 30%, indicating a lack of progress over the past
fifteen years.
Sexism on the job can take a number of forms, including inappropriate jokes, demeaning
comments, exclusion from informal work events, or blatant harassment (Gutek, 2001).
Such experiences are damaging as they are associated with negative workplace
(Murrell, Olson, & Hanson-Frieze, 1995), as well as increased levels of physical and
harassment can also affect work performance. Women experiencing sexual harassment
have reported lower work motivation and higher levels of distraction, which can lead to
poor performance and lateness, absenteeism, and high turnover rates (Schneider, Swan, &
Fitzgerald, 1997). These outcomes in turn affect women’s ability for career advancement.
It is important to note that while women are typically viewed as the target of sexism, men
5
are also at risk of experiencing sexual harassment. One study found that almost 20% of
men report being the target of sexual harassment in the workplace at least once during
In summary, it is clear that work is still needed to achieve gender parity in the
workplace. Equal opportunity and pay for women is important not only because equality
is a fundamental right, but because utilizing women’s full range of skills in the workforce
would benefit organizations and society as a whole. Making women fully valued
members of the workplace has the benefit of increasing diversity, as women bring new
experiences, talents, and perspectives. Research has shown that gender diversity has the
(Hoogendoorn, Oosterbeek, & Van Praag, 2013). According to the Global Gender Gap
Report published by the World Economic Forum, gender equality also improves national
underutilized, then economic growth will be limited. It is therefore essential to study the
systems which enable gender discrimination so that we can determine ways in which it
can be combatted. One avenue worth examining is the effects of a more subtle form of
CHAPTER 2
BACKGROUND
With blatant displays of sexism decreasing over the years (Dovidio & Gaertner,
1998), gender disparity in the workplace may be explained, in part, by a less obvious
form of sexism called benevolent sexism. Benevolent sexism was first articulated by
Glick and Fiske (1996), who described a theory of sexism involving ambivalent attitudes
consist of both positive and negative views of women. Ambivalent Sexism Theory
describes the complementary relation between hostile and benevolent sexism, arguing
that together they reinforce traditional gender roles and maintain women’s inferior status
women, but this conception ignores the subjectively benevolent attitudes toward women
that support hostile sexism. Benevolent sexism is easily overlooked because it is more
subtle and therefore less likely to be recognized as sexism (Barreto & Ellemers, 2005;
Dardenne, Dumont, & Bollier, 2007). Benevolent sexism is less socially condemned and
as it includes chivalrous acts such as carrying a woman’s groceries or paying for her
dinner. These seemingly positive behaviors, however, suggest that women need to be
cherished and protected by men. While benevolent sexism is often associated with
7
positive intentions, it stems from gender stereotypes and supports ideas of masculine
Glick and Fiske define benevolent sexism as “a set of interrelated attitudes toward
women that are sexist in terms of viewing women stereotypically and in restricted roles,
but that are subjectively positive in feeling tone and also tend to elicit behaviors typically
categorized as prosocial or intimacy-seeking” (p. 491). They argue that benevolent and
hostile sexism work in conjunction to justify gender inequality; while hostile sexism
punishes women for challenging traditional gender roles, benevolent sexism rewards
those who embrace them. Benevolent sexism allows men to justify hostile sexism by
Women who embrace benevolent sexism by surrendering power are in return rewarded
with protection and security. Research shows that women who support benevolent sexism
are also more likely to accept hostile sexism (Sibley, Overall, & Duckitt, 2007).
hostile sexist attitudes at the individual level, and a very strong correlation (r = .81) when
comparing averages across nations (Glick et al., 2000, 2004). People who are benevolent
reinforces gender stereotypes and contributes to gender inequality (Barreto & Ellemers,
2005). For example, benevolent sexism encourages women to accept traditional gender
roles and discourages them from displaying “masculine” traits such as dominance or
8
assertiveness. This is particularly harmful in the workplace where these traits are desired
qualities for high status positions. As more policies and laws are enacted to protect
women against discrimination, it is possible that benevolent sexism may be the new
Both hostile and benevolent sexism have negative implications for women in the
workforce. As discussed earlier, extensive research has shown the harms of hostile
sexism, as it negatively affects job satisfaction and psychological well-being (Bond et.
al., 2004; Klonoff, Landrine, & Campbell, 2000; Murrell, Olson, & Hanson-Frieze,
1995), and is associated with fewer opportunities for promotion and professional growth
(Gutek, 2001). Less attention has been given to the effects of benevolent sexism,
however. Research has found that in the short term, exposure to benevolent sexism
lowers women’s performance on cognitive tasks (Dardenne et al., 2007; Dumont, Sarlet,
& Dardenne, 2010; Jones, Stewart, & King, 2014; Vescio et al., 2005). Benevolent
sexism also leads women to emphasize their relational rather than task-related qualities
(Barreto et al., 2010). Thus, benevolent sexism encourages women to perpetuate feminine
towards women, and more research is needed to understand the harmful effects of
benevolent sexism. Additionally, little research has been done examining sexism towards
men and its subsequent effects on performance. With increasing numbers of women
entering the workforce, it will be important to understand how sexism affects their ability
to perform tasks during selection procedures and daily work life. If women’s
9
fields and may be at an increased risk for experiencing sexism in the workplace (Roos &
Stevens, 2014).
The purpose of the present study was to determine how benevolent and hostile
sexism affect performance on a working memory job selection task. I examined the
effects of sexism from an opposite gender recruiter on male and female “job applicant”
performance, as well as the effects on their performance self-esteem and perceived self-
identification and amount of sexist attitudes moderate the effect of sexism on task
performance.
Effects on Performance
women’s work performance (Dardenne et al., 2007; Dumont et al., 2010; Jones et al.,
2014; Vescio et al., 2005). Dardenne et al. (2007) conducted a series of experiments in
which they exposed women to written or spoken benevolent and hostile language within
the context of a job selection scenario. They found that benevolent sexism harmed
performance on a subsequent spatial working memory task, but that hostile sexism did
not. Dumont et al. (2010) obtained similar results when they conducted a study on the
effects of hostile and benevolent comments during a job interview; again, women’s
performance suffered after hearing benevolent sexism but was unaffected by hostile
sexism. A third study found that women’s performance was negatively affected after men
10
in high status positions treated them in a benevolently sexist manner (Vescio et al., 2005).
The results of the study also showed that the women reacted with anger to the situation,
but that anger was not responsible for the decline in performance. One explanation is that
2007, 2013). When women attempt to suppress these thoughts, it places demands on their
working memory capacity, reducing their mental resources for completing the task. In
internalized.
It is less clear how sexism affects men, and little research has been done on the
effects of sexism on men’s performance. Vescio et al. (2005) found that benevolent
sexism actually improved men’s performance, as their resulting anger from the unfairness
of the situation drove them to become more competitive and perform better. In contrast, a
recent study examining hostile sexism found that men’s performance on subsequent math
and language tasks was not impaired (Koch, Konigorski, & Sieverding, 2014). This
disparity may be because hostile sexism is easily identified as sexism and as a result more
benevolent sexism will improve performance for men. Hostile sexism, however,
accomplish a task in a specific situation and to reach personal goals (Bandura, 1982).
performance (Bandura, 1982). One’s self-efficacy can therefore suffer from receiving
negative feedback on personal skills or abilities. Research shows that even suggesting
that someone may be incompetent can negatively affect his or her self-efficacy (Gist and
Mitchell, 1992; Schunk, 1983). This can have serious consequences, as lower self-
efficacy is associated with poorer performance (Costa, Serrano, & Salvador, 2016;
Hewitt, 2015).
suffers when people worry about their competency levels or feel frustrated about their
performance. For instance, hearing that an upcoming test will be difficult significantly
reduces performance self-esteem (Heatheron & Polivy, 1991). Poor performance on a test
also negatively affects performance self-esteem in the short term (Heatheron & Polivy,
1991). As a result, the performance dimension of self-esteem fluctuates more than global
self-esteem is more closely associated with global self-esteem for men then it is for
women (Guertin & Jourard, 1962). Men gain more of their self-esteem from their skills
12
and abilities, which is consistent with masculine gender role norms. Women’s
performance self-esteem also tends to be lower than men’s in general, even when their
Since benevolent sexism implies that women are incompetent and that they need
men’s help to succeed, it is likely that benevolent sexism is damaging to women's self-
efficacy and performance self-esteem. Benevolently sexist ideas may lead women to
question their abilities and lower their feelings of competence. Women’s self-efficacy
benevolently sexist language. This is consistent with research by Jones et al. (2014), who
found that women’s self-efficacy was lower after benevolent sexism but was unaffected
by hostile sexism. Dardenne et al. (2007) found that women’s performance self-esteem
was also lower after experiencing benevolent sexism than after experiencing hostile
sexism. Since hostile sexism is easily recognized as sexism, it may be easier for women
to disregard it, keeping their feelings of self-efficacy and performance self-esteem intact.
(2005) found that patronizing behavior did not diminish men’s feelings of self-
competence. Additionally, a study by Jones et al. (2014) found that men reported lower
benevolent sexism condition compared to the hostile and control conditions, but
13
condition.
Two factors that may influence the effect of sexism on performance are the
participant’s gender identification and degree of personal sexist attitudes. Because sexism
identification could affect one’s attitudes about sexism. Individuals who strongly identify
with their gender may view their gender in-group more positively, and therefore be better
able to disregard negative ideas about their gender group. Research has shown that highly
identified women are more likely to recognize discrimination (Major, Quinton, &
Schmader, 2003), and have fewer mental intrusions after experiencing sexism (Dardenne
et al., 2007).
Additionally, individuals with less sexist attitudes may be more educated about
sexism and therefore be more likely to recognize subtle signs of sexism. As a result, they
could be expected to experience fewer mental intrusions and perform higher on tasks than
those with more sexist attitudes. I predicted that gender identification and sexist attitudes
would act as moderators, such that those who identify more strongly with their gender
and those who have less sexist attitudes would be less affected by exposure to benevolent
sexism. Building on Hypothesis 1 (which states that women’s performance will suffer
due to benevolent sexism but that men’s performance will improve), I proposed the
following hypotheses:
14
H3a. Among women, gender identification will moderate the relation between
type of sexism exposure and performance, such that those with higher gender
sexism.
H3b. Among women, sexist attitudes will moderate the relationship between type
of sexism exposure and performance, such that those with less sexist attitudes will
H4a. Among men, gender identification will moderate the relationship between
type of sexism exposure and performance, such that those with higher gender
sexism.
H4b. Among men, sexist attitudes will moderate the relation between type of
sexism exposure and performance, such that those with less sexist attitudes will
CHAPTER 3
METHOD
Participants
Sacramento State. There was no inclusion or exclusion criteria for the study. The median
age was 20 years (SD = 3.77) with ages ranging from 18 to 40 years old. The majority of
the students were in their first, second, or third year of school, with about 38% being first
years, 21% being sophomores, and 32% being juniors. Only 8% were seniors and 0.8%
were continuing education students. For the participants’ ethnicity, approximately 39%
White, 13% identified as Black, 0% identified as American Indian or Alaskan Native, and
14% identified as other or selected multiple races. The majority of the participants
psychology subject pool and received research participation credits towards satisfying the
Materials
Vignette. Participants read a vignette describing the simulated job scenario. In the
vignette, participants were asked to imagine that they were currently unemployed and had
been applying to jobs for months without any luck. They then finally received a positive
email from a company, who asked them to come in to complete a work-related task as
16
part of the hiring process. The position was vaguely described as a retail sales position.
This job was chosen for the study because of its relevance to undergraduate students, and
because approximately equal numbers of men and women hold the position (Bureau of
Labor Statistics, 2016). In addition to the vignette's brief backstory describing their
application process, it also explained in present terms that today is the day of the
interview and that they are currently in a room with the other job applicants nervously
Modified Lag Task. Participants completed the Modified Lag Task (Shelton,
Metzger, & Elliott, 2007) as a measure of working memory. The MLT is a version of
Dobbs and Rule’s (1989) lag task which has been adapted and validated for group
administration. The task was presented using PowerPoint and included two practice
sections and one test section. During the task, participants were shown a list of
individually-presented words on PowerPoint slides and were then presented with a probe
to recall a word in a certain position of the list (either Lag 0, Lag 1, Lag 2, or Lag 3). Lag
0 referred to the last word in the list, Lag 1 referred to the second to last word, or “one-
back” from the end of the list, and so on. Participants were shown 40 lists of either four
or six randomly ordered words. The task consisted of five trials of each of the eight
answer sheet and a total lag score was calculated using the following formula, as per
instructions included with the MLT: Lag score = (Lag 0 # correct *1) + (Lag 1 #
Burgess, and Bosson (2015), was used by the recruiter when speaking to the participants.
That manipulation was chosen because Salomon et al. (2015) had previously tested it
between the benevolent, hostile, and control conditions. The benevolent sexist language
used in the present study was slightly altered from their language to sound more
benevolent and less hostile. The revised benevolent statement was tested prior to
conducting the study and the results of the pilot study are discussed in Appendix P. A
manipulation check was also included in the present study to assess participant’s
indicator of mental intrusions. The subscale consists of seven items that measure self-
esteem in respect to task performance. Sample items include “I feel frustrated or rattled
about my performance” and “I feel confident that I understand things.” Each item is
rated on a nine-point scale ranging from 1 (not at all) to 9 (totally). In this study,
Cronbach’s alpha for the Performance Self-Esteem Subscale was found to be .78.
Scale developed by Jerusalem and Schwarzer (1992). The scale consists of 10 items
measuring one’s ability to cope with various challenges. Sample items include “I can
solve most problems if I invest the necessary effort” and “I am confident that I could deal
efficiently with unexpected events.” Each item is rated on a four-point scale ranging
18
from 1 (not at all true) to 4 (very true). Cronbach’s alpha for the General Self-Efficacy
measure of sexist attitudes towards women was obtained using the Ambivalent Sexism
Inventory (ASI). The inventory, developed by Glick and Fiske (1996), consists of two 11-
(hostile or intimate). Sample items include “Women are too easily offended” and
“Women, as compared to men, tend to have a more refined sense of culture and good
taste.” Each item is rated on a scale from 0 (disagree strongly) to 5 (agree strongly).
Cronbach’s alpha for the Ambivalent Sexism Inventory as used in this study was
determined to be .85.
were measured using the Ambivalence toward Men Inventory (AMI) developed by Glick
and Fiske (1999). This inventory provides a nice complement to the ASI, as it assesses
ambivalence towards men on the same dimensions of power, gender differentiation, and
heterosexuality. Like the ASI, each item is rated on a scale from 0 (disagree strongly) to 5
(agree strongly). Sample items include “Men are mainly useful to provide financial
security for women” and “Men would be lost in this world if women weren’t there to
guide them.” In this study, Cronbach’s alpha for the Ambivalence toward Men Inventory
to assess gender identification among women. The gender identification questions were
altered for the present study to be inclusive to men as well as women. The three items of
the scale as used for this study were “I have many characteristics in common with my
important reflection of who I am.” Each item was rated on a scale of 1 (do not agree) to
9 (totally agree), with a minimum score of 3 and maximum score of 27. Cronbach’s alpha
about their age, gender, sexual orientation, ethnicity, and year in school.
sexism manipulation was successful. The form included the sexist statement participants
heard from the recruiter, along with definitions of hostile and benevolent sexism. The
definitions and examples were derived from Glick and Fiske’s ASI and AMI, and were
different for male and female participants such that the two types of sexism were
described in terms referring to either men or women. Participants rated how benevolent
and hostile they found the statement on scales of 0 (not at all hostile/benevolent) to 4
(very hostile/benevolent). They also rated the extent to which they believed the recruiter
Procedure
The study was a 2x3 between-subjects design, with two levels of gender (men and
women) and three levels of sexism (benevolent, hostile, or control). Men and women
were recruited separately so that subject groups were comprised of members of only one
gender. Participants were randomly assigned into one of the three sexism conditions. The
study was conducted in two parts: the first consisted of online questionnaires and the
second was an experimental session involving a simulated job scenario. All participants
were required to complete both parts of the study in order to receive research credit.
Participants enrolled in the study online and signed an online consent form. They
then completed the following online questionnaires which were administered in random
order: the Ambivalent Sexism Inventory (ASI), the Ambivalence toward Men Inventory
(AMI), and a gender group identification questionnaire. At the end of the questionnaires
they were also asked to answer a demographics survey. These surveys were completed
online prior to the experimental session to minimize potential priming effects of the
questionnaires. Participants were also asked to record their assigned five digit SONA ID
so that their questionnaire responses could be linked to their performance during the
experimental session. At the end of the session, participants were directed to sign up for a
When students entered the research room and were seated, the researcher
distributed consent forms to sign. The researcher recorded each participant’s SONA ID
21
and assigned them a random subject ID, which was later used as a key to link each
Part 1. Participants were handed their subject ID on a post-it note and were instructed to
After consent, participants read a vignette explaining the simulated job scenario.
As described above, the vignette asked participants to imagine that they applied for a
retail sales position and that they were there to complete a job selection task as part of the
interview process. After allowing enough time to read through the vignette, the researcher
announced that the recruiter would be in in a few minutes, and left the room. A research
assistant wearing professional clothes then entered and introduced himself/herself as the
recruiter. The recruiter was male for female participants and female for male participants,
and was blind to the study condition until immediately before introducing the sexism
manipulation.
thanking the participants for their interest in the position, describing the job briefly, and
introducing them to the job selection task. The recruiter vaguely described the job as a
retail sales position requiring many of the skills learned during the undergraduate years.
The job description was intentionally vague to avoid introducing stereotype threat and to
After the recruiter’s introduction, the recruiter began the PowerPoint job selection
task (the Modified Lag Task). The task began with one example item and two sets of
practice items: one set of “easy” items (consisting of Lag 0, Lag 1, Lag 2, and Lag 3
22
items) and one set of “hard” items (consisting of Lag 4, Lag 5, and Lag 6). The recruiter
presented the first set of “easy” practice items and checked that the participants
understood the task by quizzing the participants and demonstrating the correct answers on
the Power Point presentation. They also provided opportunities for the participants to ask
Once the first set of practice items had been completed, the deception began. The
recruiter looked at his or her phone, acted surprised, and pretended that he or she had just
received a phone call from his or her boss. The recruiter excused him/herself from the
room supposedly to call back the boss, but before leaving the room, the recruiter started
the second set of practice items which ran automatically through PowerPoint. Once
outside the room, the recruiter met with the researcher to be informed of the study
condition. Meanwhile, the participants completed the second set of practice items
After a few minutes, the recruiter reentered the room and said, “I’m sorry, my
boss needs me to cover a coworker’s shift so I need to end the session a little early.” The
recruiter then said one of the following statements, depending on the study condition:
Hostile sexism (in a dismissive tone): “Girls/guys aren’t good at this task
anyway, so I’m going to get rid of the hard section to save some time. I’m willing
to sacrifice a little data so I don’t have to hear another girl/guy complain about
last section, so I’m going to go ahead and get rid of it for you. To help you all out,
23
I’m willing to sacrifice a little data so you don’t have to complete a really hard
section.”
Control (in an informative voice): “So I’m going to have the computer randomly
get rid of one of the sections. Looks like it got rid of the hard section.”
Immediately after, without allowing time for the participants to interject, the
recruiter began the PowerPoint presentation with the MLT test items. The recruiter
instructed participants to record their responses on their answer sheet under Section 1
(Lags 0, 1, 2, and 3), and mentioned that they could cross out the response area for
Section 2 (Lags 4, 5, and 6) if they wished. This instruction was included to ensure
participants understood that they would not be completing Section 2, or the “hard”
section.
At the end of the task, participants were given two questionnaires to complete: the
Performance Self-Esteem scale and the General Self-Efficacy scale. The questionnaires
were labeled with each participant’s subject ID. Participants were instructed not to put
their name or any other identifying mark on the questionnaires. The recruiter left the
room right after distributing the questionnaires, grabbing the laptop and their bags on the
After participants completed the questionnaires, the researcher entered the room
and collected their packets. The participants were handed a Manipulation Check sheet
where they were asked to rate how benevolent/hostile they found the recruiter’s
Once the Manipulation Check sheet was completed and collected, the researcher
thoroughly debriefed the participants and apologized for the deception. The researcher
then handed out the debriefing page for participants to keep, answered any questions, and
invited the recruiter into the room to apologize for the deception as well. Participants
were thanked for their time and given research credits for their participation.
25
CHAPTER 4
RESULTS
Manipulation Check
On the Manipulation Check form, participants were shown the sexist language
(either the hostile, benevolent, or control statement) that they heard from the recruiter
along with definitions of hostile and benevolent sexism. On two separate scales, they then
rated how hostile or benevolent the statement seemed to them. Finally, they rated the
extent to which they believed the recruiter during the session. The descriptive statistics
Table 1
Results of manipulation check
the manipulation check was partially successful. Female participants in the hostile
condition perceived the hostile statement to be significantly more hostile (M = 3.04) than
female participants in the control condition (M = 1.24, p < .001). Women in the hostile
26
those in the benevolent condition (M =3.00, p = .02), but not compared to the control
2.04) and control conditions (M = 1.68, p = .02 and p < .001, respectively). However,
women in the benevolent condition also perceived their statement to be more hostile (M =
were detected for the male participants for any of the conditions. A Bonferroni corrected
In summary, women perceived the hostile condition as more hostile and the
benevolent condition as more benevolent relative to the control condition. However, the
benevolent condition was also perceived as more hostile than the control condition. The
sexism manipulation did not appear to have an effect for men. Regardless of condition,
the men rated their statement as approximately equally hostile and benevolent. The
Across study conditions, participants indicated that they moderately believed the
recruiter during the session (M = 2.17, SD = 1.10). Only 6.8% of participants recorded
that they not at all believed the recruiter, compared to 15.5% who recorded that they very
Table 2. The ASI and AMI, both measures of sexism, were highly associated with a
27
correlation coefficient of r (131) =.76, p < .001. Contrary to expectations, the ASI was
positively correlated with gender identification, r (131) = .25, p =.004, such that
individuals who identified more strongly with their gender were more likely to possess
sexist attitudes. Performance on the working memory task and performance self-esteem
self-esteem was more likely to be higher among those who performed better on the task.
Finally, performance self-esteem and self-efficacy were positively related, r (131) = .49,
p < .001, with individuals high in performance self-esteem being more likely to score
high in self-efficacy. Self-efficacy and performance on the task did not have a significant
correlation, however.
Table 2
Descriptive statistics and inter-correlations of study variables
Variable M SD 2 3 4 5 6
1. ASI 2.28 0.72 .76* .25* -.06 .05 .12
2. AMI 2.03 0.74 .18_ -.09 -.06_ .03
3. Gender identification 7.52 1.36 -.09 .17 .14
4. Performance 27.01 4.29 _.24* .02
5. Performance self- 32.80 3.72 _.49*
esteem
6. Self-efficacy 50.71 18.05_
Note. N=131
*p < .01.
28
Tests of Hypotheses
Hypothesis 1
The first hypothesis was that benevolent sexism would harm performance for
women, but benevolent sexism would improve performance for men. Hostile sexism, on
the other hand, would not affect performance for either women or men. To test the
hypothesis, a two-way analysis of variance (ANOVA) was performed. The results of the
ANOVA revealed that there were no significant main effects of condition or gender on
task performance. The interaction between condition and gender of participant was also
non-significant, F (2, 131) = .089, p = .92, η2 = .001. Therefore, this hypothesis was
partially supported. As predicted, performance on the task was not affected by exposure
to hostile sexism for either men or women. In addition, it also was not affected by
exposure to benevolent sexism. Performance score means for each of the conditions are
displayed in Table 3.
Table 3
Effect of sexism on performance
Hypothesis 2
The second hypothesis was that women’s performance self-esteem and self-
efficacy would be lower in the benevolent sexism condition, but men’s self-esteem and
analysis of variance (MANOVA) was performed to test the hypothesis. The results of the
omnibus MANOVA revealed that there were no significant main effects of gender or
interactions were also not significant, F (2, 131) = .094, p = .91, η2 = .001, and F (2, 131)
predicted, hostile sexism did not affect performance self-esteem or self-efficacy for men
or for women. Benevolent sexism also did not have an impact on men’s scores. In those
respects, the hypothesis was supported. However, benevolent sexism was expected to
negatively affect women’s performance self-esteem and self-efficacy scores, which was
not supported. Performance self-esteem and self-efficacy score means for each condition
Table 4
Effect of sexism on performance self-esteem and self-efficacy
Performance
Gender Sexism Condition Self-Efficacy
Self-Esteem
Women Hostile 27.39 (.90) 33.52 (.77)
Benevolent 26.96 (.90) 33.65 (.77)
Control 25.88 (.87) 32.88 (.74)
Men Hostile 27.75 (.97) 32.95 (.83)
Benevolent 27.35 (.97) 32.40 (.83)
Control 26.95 (.97) 31.15 (.83)
Note. Means are presented with standard deviations in parentheses.
30
Hypothesis 3
The third hypothesis was that gender identification and sexist attitudes would
have protective effects against sexism for women. I predicted that those who identify
more strongly with their gender and those who possess fewer sexist beliefs would be less
gender identification and sexist beliefs for women, I performed two path analyses.
Hostile and benevolent sexism were coded as -1 and 1, respectively. Since two measures
of sexist attitudes were used in the study (the ASI and the AMI), the two sets of scores
were initially averaged into one measure of sexist attitudes for this analysis. However, the
model was non-significant so the analysis was re-run a second time using only AMI
scores and a third time using only ASI scores. In all three cases the model was
overidentified, so the regression coefficient for the relationship between sexism condition
moderator) are displayed below. The model’s Chi Square was 8.04 which was significant
at p = .004, indicating poor model fit. As reflected in the regression coefficients, neither
the sexism condition nor gender identification score significantly predicted performance.
Since strength of gender identification did not significantly affect the relation between
condition and performance, gender identification was not a moderator for women.
Figure 1
Path model for women with gender identification as a moderator
The models for Hypothesis 3b (with sexist attitudes as the moderator) are
displayed below. The model with the combined ASI and AMI score had a significant Chi
Square of 59.97, p = .000. Additionally, the model with only the AMI had a Chi Square
value of 59.93, which again was significant at p = .000. The model with the ASI only also
had a significant Chi Square of 59.33, p = .000. Based on the Chi Square values and other
fit indices, the models did not have good fit. Additionally, the regression weights for all
models were non-significant, indicating that neither sexism condition nor sexist beliefs
predicted performance. Sexist attitudes also did not act as a moderator between condition
Figure 2
Path model for women with sexist attitudes as a moderator
b. AMI only
c. ASI only
33
Hypothesis 4
Hypothesis 4 was that gender identification and sexist beliefs would minimize the
effects of exposure to sexism for men; specifically, those who identify more strongly with
their gender and those who possess fewer sexist beliefs would be less affected by
approach for the analysis was the same as for Hypotheses 3a and 3b. As before, the path
analysis with sexist beliefs as a moderator was first performed with the combined ASI
and AMI score, and then with the ASI only and AMI only.
The model for Hypothesis 4a, which had gender identification as a moderator, is
displayed below. The model’s Chi Square was 13.40 and was significant at p = .000,
reflecting poor fit to the data. The regression weights were also non-significant,
gender identification. Since strength of gender identification did not significantly affect
the relation between sexism condition and performance, it was not shown to be a
Figure 3
Path model for men with gender identification as a moderator
The results of the path analyses for Hypothesis 4b are displayed below. For the
model with the combined ASI and AMI score, model fit was an issue with a significant
Chi Square of 53.07, p = .000. In the ASI only model, the Chi Square was also significant
at 49.23, p = .000. The AMI only model had a significant Chi Square as well at 51.07, p =
.000. Sexist beliefs was a significant predictor of performance in the ASI only model, r =
.57, p < .001; however, none of the other relations were significant in any of the other
models. Sexist beliefs did not function as a moderator in the model since it did not
significantly change the relation between sexism condition and performance. Hypothesis
Figure 4
Path model for men with sexist atittudes as a moderator
b. ASI only
c. AMI only
36
CHAPTER 5
DISCUSSION
The purpose of the present study was to determine the effects of hostile and
benevolent sexism on men’s and women’s performance during a job selection task. The
analyses were based on 71 women and 60 men who were exposed to either benevolent
sexism, hostile sexism, or a neutral statement prior to completing a working memory test.
The results of the study indicated that Hypotheses 1 and 2 were partially supported, and
women’s performance but would improve men’s performance, and that hostile sexism
across the study conditions. Neither women’s nor men’s performance were affected by
the sexism conditions. Hypothesis 1 was therefore partially supported, because hostile
sexism did not affect performance for either men or women. Prior research also shows
that hostile sexism does not impact performance (Dardenne et al., 2007; Dumont et al.,
2010; Koch et al., 2014). The lack of effect on performance may be because hostile
internalized.
However, the findings for benevolent sexism are in contrast to previous research,
which has consistently found that benevolent sexism is damaging to performance for
women (Dardenne et al., 2007; Dumont et al., 2010; Jones et al., 2014; Vescio et al.,
37
2005). One explanation for the discrepancy in results may be related to task differences.
The present study used a verbal working memory task which was chosen to avoid
introducing stereotype threat. Dardenne et al. (2007), who found a significant decline in
performance after benevolent sexism, used a spatial working memory task that required
domain, it could have been intimidating to some of the women and as a result amplified
the effect of sexism. In contrast, a study by Salomon et al. (2015) used a word association
task instead of a working memory task and did not find differences in performance
between benevolent and hostile sexism, although they did find differences in
Another potential explanation for the lack of effect of sexism may be that the
women were able to identify benevolent sexism as actually sexist rather than tacitly
supporting them in some way. The results of the manipulation check support this, since
the women rated the benevolent statement as both benevolent and hostile. If they were
offended by the statement and interpreted it as hostile, according to prior research, their
performance likely would have been preserved. Male and female participants also scored
below the midpoints of the ASI and AMI scales, indicating low endorsement of sexist
attitudes. Being aware of the dangers of sexism may have protected the participants
This finding contrasts with Vescio et al. (2005), who found that benevolent sexism
38
improved men’s performance on a masculine task. Differences in tasks between the two
studies may have played a role in the discrepancy. However, it is hard to draw
conclusions since few other studies have been done to assess men’s responses to
benevolent sexism. It is possible that men are less affected by benevolent sexism because
they are not traditionally stereotyped as incompetent, and as a result are better able to
disregard benevolent sexism. Men may simply not be affected by sexism in the same way
as women, because they are more likely than women to be the perpetrators rather than the
targets of sexism (Einarson et al., 2003). More research is clearly needed to understand
esteem would be lower after benevolent sexism, but that men’s would be unaffected, was
esteem across the study’s six conditions. The results for women did not support the
hypothesis, and were in contrast to prior findings that benevolent sexism lowers women’s
self-efficacy (Jones et al., 2014), and performance self-esteem (Dardenne et al., 2007).
However, the hypothesis that men’s self-efficacy and performance self-esteem would not
be affected was supported. This finding is consistent with one study’s finding that
patronizing behavior does not affect men’s self-efficacy (Vescio et al., 2005).
Since benevolent sexism can subtly imply that women are incompetent, I
perhaps the female participants were able to recognize the language as sexist, and were
39
therefore able to disregard it. This interpretation is consistent with research showing that
hostile sexism does not impact women’s performance self-esteem (Dardenne et al., 2007)
or self-efficacy (Jones et al., 2014), possibly because it is easily identified as sexist. The
present study also found that hostile sexism did not affect these domains. Another
possibility is that benevolent sexism simply does not affect performance self-esteem or
indicating that individuals with poor performance self-esteem tended to have lower
scores on the test. Knowing that performance was not affected by either sexism
condition, it is not surprising that performance self-esteem was also not affected by
exposure to sexism.
that sexist beliefs and degree of gender identification would act as moderators in the
relation between sexism condition and performance. I expected that women with low
levels of sexist attitudes and high levels of gender identification would be less affected by
benevolent sexism and perform better than women with high levels of sexist attitudes and
low levels of gender identification. I predicted the same for men, except that they would
have lower performance. Since there was no relation between sexism and performance
for either men or women, the moderation model was also not significant. Sexist beliefs
and gender identification did not moderate the relation between sexism condition and
performance.
The finding that gender identification did not act as a moderator was in contrast to
research by Dardenne et al. (2007). One interesting difference was that in their study,
40
strength of gender identification was associated with lower levels of sexist attitudes. In
the present study, the reverse was true; women who identified more strongly with their
gender group had higher levels of sexist attitudes. The same gender identification survey
was used for both studies, but it is possible that cultural differences played a role. The
Islander, or White, while Dardenne et al. (2007)’s study was conducted with French
women. The participant groups may have interpreted the questions in different light due
Limitations
The study’s findings should be considered with regard to its limitations. Since this
generalizability are a concern. It is difficult to know how the results could generalize to
the job selection process in an actual workplace. For example, the Modified Lag Task,
used as a common job selection task, since selection tasks tend to target skill sets for
specific positions. Scores on the task may therefore not reflect performance on an actual
job selection task. Additionally, the job scenario created for the study cannot capture the
full range of complexities of the job selection process, since there are a number of
psychology students at Sacramento State, with the majority of the students being in their
41
first, second, and third years of school. Participants were required in their Psychology
courses to participate in research, and were motivated to earn research credits rather than
to get a job. As a result, it is possible that their motivation for performing may have been
low. Participants also self-selected into the study, so the study may have attracted
participants with certain characteristics, such as those who are more outgoing,
competitive, or looking for job experience. Additionally, the ASI and AMI surveys in
Part 1 may have turned away some participants who were offended by the questions or
were uninterested in participating in a study related to sexism. For all of these reasons,
the participant pool is not a representative sample of the typical job candidate population.
Limitations related to group dynamics also exist. Participant groups were divided
into only men or only women, which would rarely occur in the American workplace. As
a result, interactions between male and female participants were absent from the study,
and the effect of the homogenous gender pool is unknown. Participants may have felt
strength in numbers by being in groups of all women or all men. A woman would likely
have a different experience if she was the only woman in the candidate pool than she
would if she were with a group of women as in this study. Additionally, the size of the
participant groups varied across sessions. While the aim was to have five to six
participants per session, a few sessions had only one or two participants due to no shows
The results of the manipulation check also point to some limitations. The analysis
of the manipulation check indicated that female participants perceived the benevolent
statement as both benevolent and hostile, and male participants did not perceive a
42
difference between any of the conditions. These results for benevolent sexism were
surprising, since the sexism manipulation was previously validated for female
participants by Salomon et al. (2015). The present study even toned down Salomon et al.
Since the study was conducted shortly after the 2016 presidential election, when
tensions were high and sexism was on everyone’s minds, the female participants in this
study may have been more sensitive to sexist language. The election was plagued with
sexism, with sexism being demonstrated by President Donald Trump and his base, and
targeting not only presidential candidate Hillary Clinton but other public female figures
and women in general. For example, conservative pundits criticized Hillary Clinton for
being shrill (Khazan, 2016) and not smiling enough, and argued that she was only
successful because of her marriage to Bill Clinton (Bock, Byrd-Craven, & Burkley,
2017). Donald Trump, in addition to making a number of sexist and abusive remarks, was
accused of sexual assault by over a dozen women (Desjardins, 2017). As a result of this
heated political climate, the female participants could have reacted more strongly to the
possible that participants were confused by the definitions of hostile and benevolent
sexism and failed to understand the difference between the two. Benevolent sexism is
more difficult to identify than hostile sexism and is a newer concept to many
somewhat benevolent and somewhat hostile, suggesting that there may have been an issue
43
with the manipulation check form. Reading the sexism definitions may have influenced
the participants’ decision to rate the control statement as somewhat sexist. Another
potential issue is that the manipulation check form did not capture the experience of
hearing the sexist language during the session. While the results of the manipulation
check indicate that participants perceived the written statements as hostile and
benevolent, the experience of hearing the language may have been different. The
language may have been too subtle to process during the session and as a result not strong
enough to have an effect on performance. Due to the results of the manipulation check, I
recommend that future studies further validate the sexism manipulation used in the study.
Finally, other limitations may exist due to slight variations between study
sessions. While the recruiters followed a script during the sessions and were trained not
to react to the participants, it is difficult to execute any scenario identically each time.
language, may have also affected the results of some sessions, although sessions where
reactions interfered with the protocol were excluded from the analysis. The participants’
reactions and anecdotal remarks were recorded by the research assistants after the
Strengths
Despite its limitations, the study had a number of strengths due to its high degree
of experimental control. Participants were randomly assigned to the conditions and the
sessions were conducted in research rooms with similar layouts to standardize the setting.
The recruiter was also kept blind to the study condition for the period prior to delivering
44
the sexist language. This ensured that the recruiter did not act differently toward
participants in the various conditions during the start of the experimental session. All
participants therefore had the same experience before the sexism manipulation was
introduced.
Another strength of the study was that online questionnaires were used to collect
data prior to the experimental session. The online questionnaires included questions about
sexist attitudes and beliefs, gender identification, and demographic information. Such
questions could potentially prime participants to think about sexism or other personal
them online several days prior to the session avoided the potential priming effects.
The study also attempted to create a gender neutral job scenario to avoid potential
stereotype threats. The job was described as a retail sales position requiring skills learned
during the undergraduate years. As mentioned earlier, the position was chosen since it is
relevant to undergraduate students, and approximately equal numbers of men and women
hold the position across the United States. The working memory task was also selected
because I believed it to be a task in which both men and women could feel competent.
Because male and female participants were recruited and assigned to conditions
separately, it was possible to recruit an approximately equal number of men and women
for the study. This would not have been possible to achieve had they been recruited
psychology subject pool. Keeping the gender groups separate also made it possible to
control the gender make up of groups as this factor could affect perceptions of sexism.
45
performance. To date, no previous studies have examined the separate effects of hostile
and benevolent sexism on men’s performance. The study therefore offered new insights
The study’s unanticipated finding that benevolent sexism did not affect
performance adds a new perspective to the existing literature on sexism in the workplace.
yet in the present study, I found no effect. As discussed, a possible explanation for the
discrepancy may be due to today’s tense political climate, which has highlighted the issue
of sexism and brought it to national attention. The study’s finding offers an optimistic
women may be more aware of it, and as a result more protected against its harmful
effects. Where benevolent sexism may have been perceived as subtle and harmless
before, it may now be viewed as outright offensive. While this is certainly not universally
true, it is possible that this study’s population of undergraduate women was more attuned
to benevolently sexist language. For instance, their sensitivity to the sexism was reflected
in their manipulation check ratings; women in the hostile condition perceived the
language to be hostile, but women in the benevolent condition perceived the language to
be both benevolent and hostile. Another possibility is that benevolent sexism simply does
not harm performance. Future research should further examine the effects of benevolent
populations in a range of regions. The present study was conducted with undergraduate
universities with different climates would react in other ways. It would be interesting to
assess political orientation and see if political beliefs moderate the relation between
sexism and performance. Research has shown that liberalism is associated with fewer
sexist attitudes (Cokley et al., 2010; Mosso, Briante, Aiello, & Russo, 2013; Ruthig,
Kehn, Gamblin, Vanderzanden, & Jones, 2017). There is also the need to study
workplace. One of the limitations of the present study, and many of the sexism studies
mentioned in this paper (Dardenne et al., 2007; Dumont et al., 2010; Salomon et al.,
2015; Vescio et al., 2005), was that they were conducted in a laboratory setting. Bringing
the research into the field would add significantly to the literature.
attempt to replicate the study under different conditions. The study looked at the impact
of sexism on groups of all men or all women. Future studies could examine the impact of
see how a woman in a group of men, or a man in a group of women, would react to sexist
language.
An alternative explanation for the lack of significant findings in the study may be
that sexism does not affect performance on the Modified Lag Task. Other research could
see if benevolent sexism affects performance on other types of tasks. Research so far has
47
shown that benevolent sexism harms working memory, and it is essential to see if it
affects other domains such as problem solving or creativity. This would have
relevant for day-to-day performance on the job. Additionally, there is a need to study the
the conditions inherent to the present study suggest that sensitivity to sexism may
possibly help diminish its negative effects on performance. Increasing the discussion
about sexism and its harmful effects would be a positive step toward reducing sexism in
the workplace. It is especially important to warn women about the dangers of accepting
benevolent sexism, and to teach men about the effects of their chivalrous and often well-
intentioned actions. Since the study did not specifically measure awareness of or
performance was preserved. Further research is needed to determine which factors may
APPENDIX A
INFORMED CONSENT
Pretend You’re A Job Applicant: Part 1
You are invited to participate in a research study which will involve completing several
online questionnaires. My name is Melanie Rothfuss, and I am a graduate student at
California State University, Sacramento, Psychology department. The purpose of this
research is to assess the usefulness of a new job selection task.
If you decide to participate, you will be asked to complete several questionnaires about
attitudes toward people. Research designs often require that the full intent of the study not be
explained prior to participation. Although we have described the general nature of the tasks
that you will be asked to perform, the full intent of the study will not be explained to you
until after the completion of the study. Your participation in the online portion of this study
will last about 30 minutes.
There are some possible risks involved for participants. Some of the questions may make you
feel uncomfortable; however, you may decline to answer any question. There are some
benefits to this research, particularly that participating may help increase your knowledge of
survey design. While you may not personally benefit from participating in this research, it is
hoped that the research may lead to a better understanding of the benefits of using selection
tests during the hiring process.
Your participation in this project is voluntary. You have the right not to participate at all or to
leave the study at any time without penalty or loss of benefits to which you are otherwise
entitled. Since this is a two-part study, you will receive 1.5 hours of credit toward satisfying
the Psychology Department’s research participation requirement after participating in Part 1
and Part 2 of the study.
Any information that is obtained in connection with this study and that can be identified with
you will remain confidential and will be disclosed only with your permission. Your name or
any other personally identifiable information will not be attached to any records, and you will
be assigned a number ID instead. The data obtained will be maintained in a safe, locked
location and will be destroyed after a period of three years after the study is completed.
If you have any questions about the research at any time, please email me at
melanierothfuss@csus.edu, or Dr. Rachel August at raugust@csus.edu. If you have any
questions about your rights as a participant in a research project please call the Office of
Research Affairs, California State University, Sacramento, (916) 278-5674, or email
irb@csus.edu.
Your participation and completion of this survey indicates that you have read and understand
the information provided above, that you willingly agree to participate, that you may
withdraw your consent at any time and discontinue participation at any time without penalty
or loss of benefits to which you are otherwise entitled, and that you are not waiving any legal
claims, rights or remedies.
50
APPENDIX B
(ONLINE QUESTIONNAIRE)
51
Listed below are a group of statements. Please rate your agreement with each of the
statements using the following scale.
0 1 2 3 4 5
Disagree Disagree Disagree Agree Agree Agree
strongly somewhat slightly slightly somewhat strongly
APPENDIX C
(ONLINE QUESTIONNAIRE)
53
Listed below are a group of statements. Please rate your agreement with each of the
statements using the following scale.
0 1 2 3 4 5
Disagree Disagree Disagree Agree Agree Agree
strongly somewhat slightly slightly somewhat strongly
APPENDIX D
(ONLINE QUESTIONNAIRE)
55
Please answer the following questions about yourself by indicating the extent of your
agreement using the following scale:
1 2 3 4 5 6 7 8 9
Do not Agree
agree totally
APPENDIX E
DEMOGRAPHICS
(ONLINE QUESTIONNAIRE)
57
Demographics Questionnaire
Age:______
Gender:
____ Female
____ Male
____ Other Identity: __________________________
Year in College:
___ First year
___ Sophomore
___ Junior
___ Senior
___ Continuing Ed
Ethnic Identity:
___ American Indian or Alaskan Native
___ Asian or Pacific Islander
___ Black or African American
___ Hispanic
___ White or Caucasian
___ Other: __________________________
Sexual Orientation:
___ Straight or heterosexual
___ Lesbian, gay, or homosexual
___ Bisexual
___ Other: __________________________
58
APPENDIX F
INFORMED CONSENT
Pretend You’re A Job Applicant: Part 2
You are invited to participate in a research study which will involve a simulated job
selection scenario, where you will complete a work-related task and fill out several
questionnaires. My name is Melanie Rothfuss, and I am a graduate student at California
State University, Sacramento, Psychology department. The purpose of this research is to
assess the usefulness of a new job selection task.
If you decide to participate, you will be asked to read a short story describing a simulated
job scenario. In the scenario, you will imagine that you applied for a retail sales position
and that the company invited you to complete a job selection task as part of the interview
process. You will be asked to complete a job selection task, which will involve recalling
certain words from a list. You will also complete a few short questionnaires so that we
can gain feedback on your experience. Research designs often require that the full intent
of the study not be explained prior to participation. Although we have described the
general nature of the tasks that you will be asked to perform, the full intent of the study
will not be explained to you until after the completion of the study.Your participation in
this portion of the study will last approximately one hour.
There are some possible risks involved for participants. It is possible that you may
experience psychological stress from the job scenario or by completing the job selection
task. Additionally, some of the questions may make you feel uncomfortable; however,
you may decline to answer any question. There are some benefits to this research,
particularly that participating may help increase your knowledge of research design.
While you may not personally benefit from participating in this research, it is hoped that
the research may lead to a better understanding of using selection tests during the hiring
process.
Your participation in this project is voluntary. You have the right not to participate at all
or to leave the study at any time without penalty or loss of benefits to which you may
otherwise be entitled. You will receive 1.5 hours of credit toward satisfying the
Psychology Department’s research participation requirement by participating in this
second part of the study.
Any information that is obtained in connection with this study and that can be identified
with you will remain confidential and will be disclosed only with your permission. Your
name or any other personally identifiable information will not be attached to any records,
and you will be assigned a number ID instead. The data obtained will be maintained in a
safe, locked location and will be destroyed after a period of three years after the study is
completed.
If you have any questions about the research at any time, please email me at
melanierothfuss@csus.edu, or Dr. Rachel August at raugust@csus.edu. If you have any
60
questions about your rights as a participant in a research project please call the Office of
Research Affairs, California State University, Sacramento, (916) 278-5674, or email
irb@csus.edu.
Your signature below indicates that you have read and understand the information
provided above, that you willingly agree to participate, that you may withdraw your
consent at any time and discontinue participation at any time without penalty or loss of
benefits to which you are otherwise entitled, and that you are not waiving any legal
claims, rights or remedies.
Signature Date
________________________ ___________________________
61
APPENDIX G
VIGNETTE
62
Back story
Imagine that you are currently unemployed and have been applying to jobs for months
without any luck. You are beginning to lose hope, but finally you receive an email about
a retail sales position at a company you applied to weeks ago. This is the first positive
email you’ve received back about a job position. The email says the company liked your
resume and would like you to come in for the next step of the hiring process. You learn
that they want you to complete a work-related task to see how well you might do at the
job. You’re a little nervous but you’re excited for the opportunity.
APPENDIX H
RECRUITER’S SCRIPT
64
Recruiter’s script
Introduction:
“Hi everyone, my name is _____ and I’m a recruiter for the company. Thank you for
coming today and for your interest in the retail sales position. We have reviewed your
resumes and we have narrowed the candidate pool down to the strongest applicants. We
are very excited to have you here today. We believe that each of you has the potential to
be a great fit with our company. As you know, this position requires many of the skills
you’ve learned in your undergraduate career. The next step of the hiring process, and the
reason you are here today, is to complete a short job selection task which will help us
assess these abilities. Your scores on this task will provide us with valuable information
to determine how you would perform on the job. After the task, we will review your
scores and invite a few of you back to be interviewed. Let me explain what you will be
doing today...” (Recruiter then goes over the instructions for the Modified Lag Task).
Hostile sexism (in a dismissive tone): “Girls/guys aren’t good at this task anyway, so
I’m going to get rid of the hard section to save some time. I’m willing to sacrifice
a little data so I don’t have to hear another girl/guy complain about how hard the
last section is.”
Benevolent sexism (in a pleasant-yet-patronizing tone): “Girls/guys don’t like the last
section, so I’m going to go ahead and get rid of it for you. To help you all out, I’m
willing to sacrifice a little data so you don’t have to complete a really hard section
Control (in an informative voice): “So I’m going to have the computer randomly get rid
of one of the sections. Looks like it got rid of the hard section.”
65
APPENDIX I
Practice
Section 1 Section 2
A. D.
B. E.
C. F.
Test: Section 1
Lag 0, Lag 1, Lag 2, Lag 3
1. 21.
2. 22.
3. 23.
4. 24.
5. 25.
6. 26.
7. 27.
8. 28.
9. 29.
10. 30.
11. 31.
12. 32.
13. 33.
14. 34.
15. 35.
16. 36.
17. 37.
18. 38.
19. 39.
20. 40.
67
Test: Section 2
Lag 4, Lag 5, Lag 6
1. 21.
2. 22.
3. 23.
4. 24.
5. 25.
6. 26.
7. 27.
8. 28.
9. 29.
10. 30.
11. 31.
12. 32.
13. 33.
14. 34.
15. 35.
16. 36.
17. 37.
18. 38.
19. 39.
20. 40.
68
APPENDIX J
Word Bank
APPENDIX K
Please answer the following questions about yourself by indicating the extent of your
agreement using the following scale:
1 2 3 4 5
Not at all A little bit Somewhat Very much Extremely
_______ 6. I feel that I have less scholastic ability right now than others.
APPENDIX L
Please answer the following questions about yourself by indicating the extent of your
agreement using the following scale:
APPENDIX M
WOMEN – HOSTILE
Girls aren’t good at this task anyway, so I’m going to get rid of the hard section to save some time. I’m
willing to sacrifice a little data so I don’t have to hear another girl complain about how hard the last
section is.
Hostile sexism involves overt hostility and disrespect towards women. It reflects the idea that women are
incompetent and inferior to men. Some examples include beliefs that women are easily offended, that they
make unreasonable demands, and that men are smarter than women.
With this definition in mind, please rate the above statement according to how hostile it seemed to
you, on a scale of 0 to 4.
0 1 2 3 4
Not at all hostile Somewhat hostile Moderately hostile Hostile Very hostile
Benevolent sexism involves ideas about women that sound positive but are actually harmful. Benevolent
sexism is more socially accepted and often is not recognized as sexism. However, it helps maintain gender
roles by reinforcing gender stereotypes, such as the belief that women are weak and need to be protected by
men. Some examples include beliefs that women need men for financial support, that women should be
cherished by men, and that women should be kind and nurturing.
With this definition in mind, please rate the above statement according to how benevolent it
seemed to you, on a scale of 0 to 4.
0 1 2 3 4
______________________________________________________________________
To what extent did you believe the recruiter during this study?
0 1 2 3 4
WOMEN – BENEVOLENT
Girls don’t like the last section, so I’m going to go ahead and get rid of it for you. To help you all out,
I’m willing to sacrifice a little data so you don’t have to complete a really hard section.
Hostile sexism involves overt hostility and disrespect towards women. It reflects the idea that women are
incompetent and inferior to men. Some examples include beliefs that women are easily offended, that they
make unreasonable demands, and that men are smarter than women.
With this definition in mind, please rate the above statement according to how hostile it seemed to
you, on a scale of 0 to 4.
0 1 2 3 4
Not at all hostile Somewhat hostile Moderately hostile Hostile Very hostile
Benevolent sexism involves ideas about women that sound positive but are actually harmful. Benevolent
sexism is more socially accepted and often is not recognized as sexism. However, it helps maintain gender
roles by reinforcing gender stereotypes, such as the belief that women are weak and need to be protected by
men. Some examples include beliefs that women need men for financial support, that women should be
cherished by men, and that women should be kind and nurturing.
With this definition in mind, please rate the above statement according to how benevolent it
seemed to you, on a scale of 0 to 4.
0 1 2 3 4
________________________________________________________________________
To what extent did you believe the recruiter during this study?
0 1 2 3 4
WOMEN – CONTROL
I’m going to have the computer randomly get rid of one of the sections. Looks like it got rid of the hard
section.
Hostile sexism involves overt hostility and disrespect towards women. It reflects the idea that women are
incompetent and inferior to men. Some examples include beliefs that women are easily offended, that they
make unreasonable demands, and that men are smarter than women.
With this definition in mind, please rate the above statement according to how hostile it seemed to
you, on a scale of 0 to 4.
0 1 2 3 4
Not at all hostile Somewhat hostile Moderately hostile Hostile Very hostile
Benevolent sexism involves ideas about women that sound positive but are actually harmful. Benevolent
sexism is more socially accepted and often is not recognized as sexism. However, it helps maintain gender
roles by reinforcing gender stereotypes, such as the belief that women are weak and need to be protected by
men. Some examples include beliefs that women need men for financial support, that women should be
cherished by men, and that women should be kind and nurturing.
With this definition in mind, please rate the above statement according to how benevolent it
seemed to you, on a scale of 0 to 4.
0 1 2 3 4
_______________________________________________________________________
To what extent did you believe the recruiter during this study?
0 1 2 3 4
MEN – HOSTILE
Guys aren’t good at this task anyway, so I’m going to get rid of the hard section to save some time. I’m
willing to sacrifice a little data so I don’t have to hear another guy complain about how hard the last
section is.
Hostile sexism involves overt hostility and disrespect towards men. Some examples include beliefs that
men fight for control in society, that they harass women when they’re in positions of power, and that men
act like children and would be lost without women.
With this definition in mind, please rate the above statement according to how hostile it seemed to
you, on a scale of 0 to 4.
0 1 2 3 4
Not at all hostile Somewhat hostile Moderately hostile Hostile Very hostile
Benevolent sexism involves ideas about men that sound positive but are actually harmful. Benevolent
sexism is more socially accepted and often is not recognized as sexism. However, it helps maintain gender
roles by reinforcing gender stereotypes, such as the belief that men are powerful and strong. Some
examples include the belief that men need to provide financially for women, that they are more willing to
take risks than women, and that men need women to take care of them.
With this definition in mind, please rate the above statement according to how benevolent it
seemed to you, on a scale of 0 to 4.
0 1 2 3 4
________________________________________________________________________
To what extent did you believe the recruiter during this study?
0 1 2 3 4
Guys don’t like the last section, so I’m going to go ahead and get rid of it for you. To help you all out,
I’m willing to sacrifice a little data so you don’t have to complete a really hard section.
Hostile sexism involves overt hostility and disrespect towards men. Some examples include beliefs that
men fight for control in society, that they harass women when they’re in positions of power, and that men
act like children and would be lost without women.
With this definition in mind, please rate the above statement according to how hostile it seemed to
you, on a scale of 0 to 4.
0 1 2 3 4
Not at all hostile Somewhat hostile Moderately hostile Hostile Very hostile
Benevolent sexism involves ideas about men that sound positive but are actually harmful. Benevolent
sexism is more socially accepted and often is not recognized as sexism. However, it helps maintain gender
roles by reinforcing gender stereotypes, such as the belief that men are powerful and strong. Some
examples include the belief that men need to provide financially for women, that they are more willing to
take risks than women, and that men need women to take care of them.
With this definition in mind, please rate the above statement according to how benevolent it
seemed to you, on a scale of 0 to 4.
0 1 2 3 4
_______________________________________________________________________
To what extent did you believe the recruiter during this study?
0 1 2 3 4
MEN – CONTROL
I’m going to have the computer randomly get rid of one of the sections. Looks like it got rid of the hard
section.
Hostile sexism involves overt hostility and disrespect towards men. Some examples include beliefs that
men fight for control in society, that they harass women when they’re in positions of power, and that men
act like children and would be lost without women.
With this definition in mind, please rate the above statement according to how hostile it seemed to
you, on a scale of 0 to 4.
0 1 2 3 4
Not at all hostile Somewhat hostile Moderately hostile Hostile Very hostile
Benevolent sexism involves ideas about men that sound positive but are actually harmful. Benevolent
sexism is more socially accepted and often is not recognized as sexism. However, it helps maintain gender
roles by reinforcing gender stereotypes, such as the belief that men are powerful and strong. Some
examples include the belief that men need to provide financially for women, that they are more willing to
take risks than women, and that men need women to take care of them.
With this definition in mind, please rate the above statement according to how benevolent it
seemed to you, on a scale of 0 to 4.
0 1 2 3 4
________________________________________________________________________
To what extent did you believe the recruiter during this study?
0 1 2 3 4
APPENDIX N
DEBRIEFING SHEET
82
Debriefing
Purpose
The main purpose of this study is to investigate how exposure to benevolent and hostile
sexism affects performance on a working memory job selection task. Hostile sexism is
what one traditionally thinks of with sexism, as it involves overt hostility and disrespect
towards the target. Benevolent sexism, on the other hand, is more subtle and less likely to
be recognized as sexism. When benevolent sexism is targeted toward women, it involves
the belief that women should be protected and cherished by men. When benevolent
sexism is targeted toward men, it involves the belief that men should be powerful and
strong. The study also examines how exposure to sexism affects performance self-esteem
and perceived self-efficacy (or confidence in one’s capabilities). Finally, the study looks
at the degree to which participants identify with their gender and their degree of sexist
beliefs to see if these variables influence the effect of sexism on task performance.
experimental groups by looking at total scores on the working memory task and the
questionnaires.
Contact Information
The results of this study will be available by May 2017. If you would like further
information about the study or have questions regarding the experiment, please contact
Melanie Rothfuss at melanierothfuss@csus.edu at your convenience.
Psychological Services
If you have experienced any personal distress caused by the content or materials in this
research and want to talk to someone, counseling services are available through the
Student Health Center free of charge. Please contact Psychological Services at 278-6416
for assistance.
Closing
Do you have any questions? (Respond appropriately.)
Thank you for participating!
84
APPENDIX O
PARTICIPANT REACTIONS
85
Participant Reactions
The recruiters took notes and logged written observations after each session. A
summary of the notes, broken down by gender and condition, are included in the table
below. Also indicated is whether the reactions excluded the session from the analysis.
Only sessions with extreme reactions (defined as reactions that interrupted the session
and prevented the recruiter from following protocol) were removed. Unless otherwise
specified, the comments from participants occurred during the debriefing. Since the
manipulation referred to the participants as “guys” and “girls,” the recruiters used the
Participant Included in
Condition Reactions
Gender analyses?
After hearing that girls aren’t good at the
hard section, one participant said, "Maybe
Yes
we shouldn't be doing this section of the test
either."
A number of girls seemed to act colder and
Hostile more distant toward the male recruiter after
he said the sexist language. For example, Yes
they stopped saying “thank you,” or smiling
and making eye contact during the session.
Women Some girls “snorted” or laughed after
Yes
hearing the sexism.
Several girls said they didn’t notice the
sexism, and were more concerned about the Yes
test getting cut short.
One common response was smiling
Benevolent Yes
awkwardly after hearing the sexism.
Two girls were confused because they
thought the “hard” section of the test was Yes
actually easier than the “easy” section.
One participant thought that the female
recruiter acted hostile towards them because
Men Hostile Y
her boss, who called her in to work, might
be a man. He thought she was frustrated
86
APPENDIX P
PILOT STUDY
88
Pilot Study
Prior to conducting the study, I carried out a pilot study to test the sexism
manipulation and determine whether changing Salomon et al.’s (2015) benevolent sexism
statement would be effective. Their original benevolent sexism statement was “Girls
don’t like the hard section, so I’m going to go ahead and get rid of it for you. I’m willing
to sacrifice a little data so I don’t make another girl upset about how hard the last section
is.” The last sentence was altered to “I’m willing to sacrifice a little data to help you girls
out so you don’t have to complete a really hard section” to sound more benevolent.
Thirteen female and 7 male undergraduate students participated in the study. The pilot
study began similarly to the full study; participants read the vignette and an opposite-sex
research assistant introduced themselves as the recruiter. Participants completed the first
half of the practice items, then the recruiter pretended to get a phone call. However, the
recruiter’s excuse for ending the session early during the pilot was that they forgot they
had a meeting across campus that they needed to get to. They then delivered the
benevolent language and left the room for a couple minutes to give the participants time
to process the sexist statement. When they returned, they passed out a manipulation
check sheet so participants could rate how benevolent or hostile they found the statement.
Participants were also asked to respond to the following questions: (1) “How did hearing
this statement make you feel?” and (2) “Were there any specific words or characteristics
of the recruiter that caused you to feel this way?” Sessions for the pilot study began the
morning after the 2016 Presidential Election, and due to the sensitive time for many
Results from the pilot study are displayed in the table below. The pilot study
found that participants reacted more strongly to the benevolent sexism than expected, so
the statement was further altered for the present study to sound more benevolent. The
manipulation check sheet was also revised to provide clearer definitions. Finally, the
recruiter’s excuse for ending the session early was changed, since male participants found
the recruiter disrespectful for prioritizing her meeting over their session and expressed
Table P
Results of manipulation check
Perceived Perceived
Gender Sexism Condition
Hostility Benevolence
Women Benevolent 2.46 (0.97) 2.69 (1.11)
Men Benevolent 1.14 (1.11) 2.00 (1.53)
Note. Means are presented with standard deviations in parentheses.
90
References
37, 122-147.
Barreto, M., Ellemers, N., Piebinga, L. & Moya, M (2010). How nice of us and how
dumb of me: the effect of exposure to benevolent sexism on women’s task and
Barreto, N., and Ellemers, N (2005). The burden of benevolent sexism: how it contributes
Blau, F. D., Ferber, M. A., & Winkler, A. E. (2001). The economics of women, men, and
Bock, J., Byrd-Craven, J., & Burkley, M. (2017). The role of sexism in voting in the 2016
Bond, M. A., Punnett, L., Pyle, J. L., Cazeca, D., & Cooperman, M. (2004). Gendered
Cokley, K. O., Tran, K., Hall-Clark, B., Chapman, C., Bessa, L., Finley, A., et al. (2010).
Predicting student attitudes about racial diversity and gender equity. Journal of
Dardenne, B., Dumont, M., & Bollier, T. (2007). Insidious dangers of benevolent sexism:
Dardenne, B., Dumont, M., Sarlet, M., Phillips, C., Balteau, E., Degueldre, C., Luxen, A.,
Salmon, E., Maquet, P., & Collette, F. (2013). Benevolent sexism alters executive
Desjardins, L. (2017). All the assault charges against Donald Trump, recapped. PBS
Dovidio, J. F., & Gaertner, S. L. (1998). On the nature of contemporary prejudice: the
Eberhardt, JL. Racism: The Problem and the Response, Sage, Newbury Park, CA,
3-32.
Dumont, M., Sarlet, M., & Dardenne, B. (2010). Be too kind to a woman, she’ll feel
Einarson, S., Hoel. H., Zapf, D. & Cooper, C. L. (2003). Bullying and emotional abuse in
Glick, P., & Fiske, S. T. (1996). The ambivalent sexism inventory: differentiating hostile
and benevolent sexism. Journal of Personality and Social Psychology, 70, 491-
512.
Glick, P., & Fiske, S. T. (1999). The ambivalence toward men inventory: differentiating
hostile and benevolent beliefs about men. Psychology of Women Quarterly, 23,
519-536.
Glick, P., Fiske, S. T., Mladinic, A., Saiz, J. L., Abrams, D., Masser, B., & López, W.
Glick, P., Lameiras, M., Fiske, S. T., Eckes, T., Masser, B., Volpato, C., & Wells, R.
(2004) Bad but bold: Ambivalent attitudes toward men predict gender inequality
Goldberg, C. B., Finkelstein, L. M., Perry, E. L., & Konrad, A. M. (2004). Job and
industry fit: The effects of age and gender matches on career progress outcomes.
Gutek, B. A. (2001). Women and paid work. Psychology of Women Quarterly, 25(4),
379.
Gutek, B.A., & Done, R. (2001). Sexual harassment. In R. Unger (Ed.), Handbook of the
Heatherton, T. F., & Polivy, J. (1991). Development and validation of a scale for
895-910.
Hebl, M. R., King, E. B., Glick, P., Kazama, S., & Singletary, S. (2007). Hostile and
298-313.
Hoogendoorn, S., Oosterbeek, H., & Van Praag, M. (2013). The impact of gender
Jones, K., Stewart, K., & King, E. (2014). Negative consequence of benevolent sexism on
Khazan, O. (2016). Would you really like Hillary more if she sounded different? The
King, E. B., Botsford, W. E., Hebl, M. R., Kazama, S., Dawson, J. F., & Perkins, A.
1866.
Klonoff, E. A., Landrine, H., & Campbell, R. (2000). Sexist discrimination may account
Koch, S. S., Konigorski, S., & Sieverding, M. (2014). Sexist Behavior Undermines
Major, B., Quinton, W. J., & Schmader, T. (2003). Attributions to discrimination and
Mosso, C., Briante, G., Aiello, A., & Russo, S. (2013). The role of legitimizing
Murrell, A. J., Olson, J. E., & Hanson-Frieze, I. (1995). Sexual harassment and gender
Pearsall, M. J., Ellis, A. P. J., & Evans, J. M. (2008). Unlocking the effects of gender
Power, G. N., Butterfield, D. A., & Parent, J. D. (2002). Gender and managerial
Sex Segregation from 2000 to 2010" Paper presented at the annual meeting of the
Ruthig, J. C., Kehn, A., Gamblin, B. W., Vanderzanden, K., & Jones, K. (2017). When
Salomon, K., Burgess, K. D., & Bosson, J. K. (2015). Flash fire and slow burn: women’s
Schneider, K.T., Swan, S., & Fitzgerald, L.F. (1997). Job-related and psychological
Shelton, J. T., Metzger, R. L., & Elliott, E. M. (2007). A group-administered lag task as a
Sibley, C. G., Overall, N. C., & Duckitt, J. (2007). When women become more hostilely
Spector, P. E. (2008). Industrial and organizational behavior. Hoboken, NJ: John Wiley
Van Vianen, A. E. M., & Fischer. A. H. (2002). Illuminating the glass ceiling: The role of
Vescio, T. K., Gervais, S. J., Snyder, M., & Hoover, A. (2005). Power and the creation of
Xu, W., & Leffler, A. (1996). Gender and race effects on occupational prestige,
(Eds.), Race, class, and gender: Common bonds, different voices (pp. 107-124).