Who Goes to the Bargaining Table? The Influence of Gender and Framing
on the Initiation of Negotiation
Deborah A. Small
University of Pennsylvania
University of Maryland
Carnegie Mellon University
University of Maryland
Unlike typical negotiation experiments, these studies investigated when men and women initiate
negotiations in the absence of overt prescriptions to negotiate. Using a new experimental paradigm, the
authors showed that the framing of situations is a critical driver of gender differences in initiating
negotiations. Drawing on literature on language, power, and politeness, the authors argued that framing
situations as opportunities for negotiation is particularly intimidating to women, as this language is
inconsistent with norms for politeness among low-power individuals, such as women. By contrast,
framing situations as opportunities for asking is much less intimidating to women, as this language is
more polite and role-consistent. Consequently, gender differences in initiating negotiations persisted
when situations were framed as opportunities for negotiation yet were eliminated when situations were
framed as opportunities to ask. Moreover, primed power attenuated gender differences in aversive
reactions to the negotiation frame. In all, the studies presented begin to elucidate the reasons for gender
differences in initiating negotiations and further illustrate that such effects depend on the situation.
Keywords: negotiation, gender, framing
In virtually all experimental negotiation research, individuals
are instructed to conduct a negotiation. Often negotiators are given
explicit instructions—that is, to make offers and receive counter-
offers, to achieve as much value for themselves as possible, and so
forth. Although this research has illuminated the psychological and
behavioral processes that are related to negotiation strategies and
outcomes, there is little research on the perception of the negotia-
bility of situations when negotiation is not prescribed or on the
tendency to initiate negotiation more generally.
This void in the literature is significant because many situations
are not transparently negotiable. For example, although many
Americans probably know that the price of a car is negotiable,
most situations are far more ambiguous. Consider the case of a
university professor. Although it is typical to negotiate over salary
and teaching load when one is first hired, what is negotiable after
one is on the job is far less clear. In fact, numerous resources can
be negotiated throughout one’s faculty career, including reduced
teaching loads, summer support, increases in pay, office space,
computer upgrades, and so forth. However, these situations need to
be recognized as negotiable and capitalized on, often on an indi-
vidual basis. Otherwise, the status quo is likely to remain.
Indeed, recent changes in the workplace highlight the need for
individuals to initiate negotiations if a change or improvement of
circumstances is desired. Because of new emerging forms of
organization (flattened hierarchies, lower formalization, increased
participation), the steady decline of unionism (dropped from
20.1% of the workforce in 1983 to 13.5% in 2001), and increasing
rates of job turnover (39% of the workforce changed jobs between
May 2001 and May 2002; Bureau of Labor Statistics, 2002, 2006),
employees have opportunities to bargain that are unprecedented
(Rousseau, 2005). In her book on idiosyncratic deals, Rousseau
(2005) cited evidence that 30% of individuals graduating with a
master’s in business administration (MBA) negotiated customized
employment relationships (special deals), and 25% of health care
workers in a hospital negotiated special arrangements for them-
selves. These statistics suggest that if the flexibilities of contracts
are not made explicit, then individuals’ effectiveness and success
will depend on recognizing opportunities to negotiate for re-
Surprisingly, despite the employment trends described earlier,
which suggest that increasing opportunities for negotiations are
prevalent, there is a dearth of research on who initiates negotia-
tions and whether these opportunities are disproportionately cap-
italized on by certain people in general or by men or women in
particular. The question of whether and why gender differences
exist in the propensity to initiate negotiations is important, given
the persistent wage gap (Keaveny & Inderrieden, 2000), glass
ceiling (Tharenou, 2001), and the fact that women advance in their
careers at much slower rates than do men (Tharenou, 1999; Valian,
Deborah A. Small, Department of Marketing, Wharton School, Univer-
sity of Pennsylvania; Michele Gelfand and Hilary Gettman, Department of
Psychology, University of Maryland; Linda Babcock, Heinz School of
Public Policy and Management, Carnegie Mellon University.
Correspondence concerning this article should be addressed to Deborah
A. Small, Department of Marketing, 700 John M. Huntsman Hall, Univer-
sity of Pennsylvania, 3730 Walnut St., Philadelphia, PA 19104-6340;
Journal of Personality and Social Psychology Copyright 2007 by the American Psychological Association
2007, Vol. 93, No. 4, 600 – 613 0022-3514/07/$12.00 DOI: 10.1037/0022-35126.96.36.1990
Babcock, Gelfand, Small, and Stayn (2006) provided some
initial evidence that women are less likely to initiate negotiations.
They sampled 227 working adults and asked how recently they had
initiated negotiations. Men indicated that they had initiated nego-
tiations two to four times as recently as women. A separate study
comparing male and female master’s degree students found that
among graduating MBA students, 51.9% of men negotiated their
job offer, whereas only 12.5% of women did. Not surprisingly,
women received average annual starting salaries that were 8.5%
lower than those of men. Although there could be other explana-
tions for this gender salary gap, the possibility that some of the gap
could be attributed to gender differences in initiating negotiations
is intriguing. When projected across the course of women’s ca-
reers, this starting salary gap would be even more striking because
raises, bonuses, and other compensation are typically based on
initial salary (Gerhart & Rynes, 1991).
Nevertheless, although Babcock et al.’s (2006) studies provided
initial evidence for gender differences in the propensity to initiate
negotiation, they were field studies that had a number of limita-
tions. For example, the first study relied on retrospective reports
that could have been subject to biases, and the second study could
not control for gender differences that might have existed in the
graduate students’ bargaining leverage (e.g., other job offers).
Overview of the Current Research
The purpose of this research was to more systematically exam-
ine gender differences in the propensity to initiate negotiations and
the psychological mechanisms that account for such differences.
Our first aim was to replicate Babcock et al.’s (2006) research by
examining gender differences in a controlled setting using a newly
developed negotiation paradigm. In this method, participants
played a word game in exchange for cash and were subsequently
offered the minimum payment possible. We measured whether
participants negotiated higher payment from the experimenter.
This method simulates important advancement opportunities (ask-
ing for more money for a task) but allows us to control for the
variability in industry norms, job characteristics, work experience,
and so forth. By introducing this new method, we extended Bab-
cock et al.’s (2006) findings and broaden the scope of current
negotiation research from primarily examining what happens at the
negotiation table to examining who gets to the table in the first
Our second aim was to further explore the possible mechanisms
that account for gender differences in initiating negotiations. We
argue that although women are less likely to initiate negotiations
than men, it is possible that women are not unaware of such
opportunities to negotiate; rather, it is the framing of situations that
is a critical driver of gender differences in initiating negotiations,
with some frames exacerbating and some frames attenuating gen-
der differences. Drawing on literature on language, power, and
politeness (Areni & Sparks, 2005; Brown & Levinson, 1987;
Lakoff, 1975; Morand, 2000), we explored the notion that framing
situations as opportunities for negotiation is particularly intimidat-
ing to women, as this language implies a face-threatening act that
is inconsistent with norms for politeness among low-power indi-
viduals, such as women (Brown & Levinson, 1987). By contrast,
we expected that framing situations as opportunities for asking
would be much less intimidating to women, as this language is
indicative of a linguistic gesture of politeness and deference used
when attempting to get something from another, which is consis-
tent with low-power social roles (Brown & Levinson, 1987).
Studies 2– 4 illustrated support for these notions, showing that
gender differences do in fact persist when situations are framed as
opportunities for negotiation yet are dramatically reduced when
situations are framed as opportunities for asking. In Study 5, we
further explored these effects by showing that it is social power
that is a key driver of differential reactions of men and women to
these frames. We showed that when primed with power, gender
differences in reactions to opportunities to negotiate are greatly
attenuated. In all, the studies presented begin to elucidate the
reasons for gender differences in initiating negotiations and further
illustrate that such effects depend on the situation.
In what follows, we first situate our study of gender differences
in the propensity to initiate negotiations within the larger negoti-
ation literature, discussing how gender has typically been exam-
ined in negotiation research. We then provide more background on
our cognitive approach to gender differences, discussing the psy-
chology of framing in negotiations. Drawing on linguistic theory
of politeness (Brown & Levinson, 1987), we then describe the
theoretical rationale for why different frames might be more or less
intimidating—and produce less initiating of actual negotiation
behavior—than other frames. As these frames are intricately linked
to social power, we discuss why power might be a key explanation
for gender differences in reactions to different frames. We then
describe five studies that systematically examined gender, fram-
ing, and the initiating of negotiations using a newly developed
Gender Differences in Negotiation
Negotiation researchers have long been interested in gender
differences in negotiation. As early as 1975, Rubin and Brown
(1975) reviewed over 65 studies investigating the role of gender in
negotiation and bargaining. Since then, there have been numerous
studies of gender differences in negotiation styles, strategies, and
tactics (e.g., Clark, 1983; Kyl-Heku & Buss, 1996; Lind, Huo, &
Tyler, 1994; Maxwell, 1992; Putnam & Jones, 1982) and gender
differences in outcomes (e.g., Gerhart & Rynes, 1991; Mesch &
Dalton, 1989; Neu, Graham, & Gilly, 1988; Wachter, 1999;
Watson & Hoffman, 1996). Studies have also examined gender
differences in time to reach an agreement (Griffith, 1991; Neu et
al., 1988), negotiator goals or expectations (Major, Vanderslice, &
McFarlin, 1984; Stevens, Bavetta, & Gist, 1993), perceptions of
conflicts (e.g., Haferkamp, 1991–1992), and satisfaction (Papa &
Despite the quest to document gender differences in a wide
range of negotiation phenomena, however, findings have been
highly inconsistent. Rubin and Brown (1975) documented 21
studies that found men behaved more cooperatively than women in
bargaining experiments, 27 studies that concluded women behaved
more cooperatively than men, and 20 studies that found no differ-
ences among men and women. Other, more recent meta-analyses
(Stuhlmacher & Walters, 1999; Walters, Stuhlmacher, & Meyer,
1998) have shown that the magnitude of gender differences tends
to be small and that there are situational moderators of gender
GENDER, FRAMING, AND INITIATION OF NEGOTIATION
This inconsistency, however, has proven to be a breakthrough in
gender and negotiation research, as scholars have increasingly
advocated a more contextual and dynamic approach to gender
differences that documents the conditions under which gender
differences are exacerbated and attenuated in negotiation (Bowles,
Babcock, & McGinn, 2005; Gelfand, Major, Raver, Nishii, &
O’Brien, 2006; Kray, Galinsky, & Thompson, 2001, 2002). For
example, Bowles et al. (2005) showed that the roles that are
activated at the negotiation table (i.e., whether one is negotiating
for oneself or on the behalf of others) are important triggers for
gender differences. Likewise, Kray et al. (2002) showed that
gender differences in negotiation behavior are strongly affected by
cognitive constructs that are accessible at the negotiation table.
Gender differences were amplified when stereotypes of the mas-
culinity of negotiation were subtly primed (causing stereotype
threat), whereas gender differences were diminished when the
stereotype was overtly primed (causing women to react against the
stereotype). Gelfand et al. (2006) similarly argued that gender
differences in negotiation are exacerbated in situations that make
the relational self accessible and are reduced in situations in which
the relational self is inhibited or less applicable. In all, these newly
emerging social– cognitive approaches offer a promising and con-
textually rich perspective to gender and negotiation research.
In this article, we took a cognitive approach to gender differ-
ences in propensity to initiate negotiations. We theorized that
women’s feelings about, and behaviors of, initiating a negotiation
are partly due to the way in which that behavior is framed. To this
end, we drew on research in judgment and decision making (e.g.,
Kahneman & Tversky, 1984), in the cognitive tradition in negoti-
ation (e.g., Neale & Bazerman, 1991), and on language and power
(Brown & Levinson, 1987) and examined how subtle cues in the
environment affect the propensity to negotiate among men and
Research on framing effects has shown that judgments and
decisions are highly sensitive to varying descriptions of a prospect
(Kahneman & Tversky, 1984). For instance, framing a risk in
terms of gains versus losses (e.g., a 90% chance of survival vs. a
10% chance of death) should rationally not affect choices. How-
ever, people are more risk averse when a prospect is described as
a gain than when an identical prospect is described as a loss.
Framing effects can occur because different descriptions may be
associated with quite different construals and feelings (McFarland
& Miller, 1994). For example, Liberman, Samuels, and Ross
(2004) found that players were much more likely to defect in a
Prisoner’s Dilemma–type game when the experiment was de-
scribed to players as the “Wall Street Game” as opposed to the
“Community Game,” even though the payoff matrices were iden-
tical. Thus, the context of the game influenced participants’ con-
strual of the situation and their subsequent behavior.
In this research, we examined how differential frames, namely,
negotiating versus asking, affect the propensity to negotiate among
women and men. More specifically, we proposed that framing
situations as an opportunity for negotiation is threatening to
women, which inhibits their propensity to initiate a negotiation. In
contrast, however, describing the identical situation as an oppor-
tunity for asking is less threatening to women and reduces gender
differences in the propensity to initiate negotiations.
Scholars have long argued that language not only reflects a
representational function but also serves a social function (Brown
& Levinson, 1987; Goffman, 1967). In any encounter, speakers are
concerned with maintaining their own and others’ face and use
varying degrees of politeness strategies to avoid threatening others.
In their seminal work on politeness theory, Brown and Levinson
(1987) argued that the use of politeness was intricately tied to
social power: Individuals who hold low-power positions are more
inclined to use polite speech (such as carefully worded requests
rather than direct demands) to avoid imposing on others, as low-
power individuals do not have the appropriate status to impose on
others (Lakoff, 1975; Morand, 2000). Moreover, politeness is
particularly important when a low-power individual is engaging in
a face-threatening act, such as criticizing, disagreeing, or attempt-
ing to get something from another person (e.g., information, a
favor, resources), and is used as a way to “minimize or defray” that
threat (Morand, 2000, p. 237).
We contend that the concept of negotiation is one such face-
threatening act, as it implies demands being made from a position
of power. By contrast, asking conveys a weaker stance and is
considered a linguistic gesture of politeness used when attempting
to get something from another person. Simply put, asking implies
linguistic deference (Brown & Levinson, 1987), in which speakers
acknowledge restraint to minimize imposing on others, and is
particularly important for low-power individuals. For example, a
low-power person might say, “May I borrow a dollar?” rather than,
“Give me a dollar,” to minimize the face-threatening act of re-
questing money (Morand, 2000, p. 237). Drawing on this theory,
we reasoned that the language or frame of asking would be less
threatening than the frame of negotiating among women, given
they tend to have lower power in American society (Eagly &
Wood, 1982; Henley & LaFrance, 1984). We expected that the
language of negotiation would be less threatening to men, given
that it is congruent with their position of high power in society.
Indeed, women have been found to prefer more indirect, polite
speech, as compared with men (Holtgrave, 1997). By extension,
we expected women to react more positively to the framing of
indirect polite asking than the direct powerful frame of negotiating.
In Studies 2–4, we directly tested these notions and showed that
women perceive opportunities for negotiating as much more aver-
sive than opportunities for asking and that these frames affect
actual rates of initiating negotiations.
Of importance, in their classic work, Brown and Levinson
(1987) cautioned that women might not prefer polite language in
all situations but rather that it is their typically low-power status
that leads them to generally prefer politeness. However, power can
vary by situation. A number of recent empirical studies have
shown that situationally primed power is related to a diverse set of
approach-related behaviors, with an increased attention to rewards
in the environment and means for obtaining rewards (Anderson &
Berdahl, 2002; Galinsky, Gruenfeld, & Magee, 2003; see also
Keltner, Gruenfeld, & Anderson, 2003). For example, Galinsky et
al. (2003) showed that the situational experience of power leads
individuals to act against an annoying stimulus in the environment,
to take resources when they are available, and to take action in
competitive situations. Furthermore, power is associated with per-
ceived freedom, control, and influence (Keltner et al., 2003; Kip-
nis, 1972). Consistent with this, the possession of power increases
the propensity to negotiate (Magee, Galinsky, & Gruenfeld, 2007).
Therefore, it is critical to examine what happens when women feel
powerful, because power is one critical driver of politeness and
SMALL, GELFAND, BABCOCK, AND GETTMAN
thus may mitigate negative reactions to frames of negotiation. In
Study 5, we directly tested this notion by manipulating power
among men and women and examining their reactions to negoti-
ating versus asking frames. We reasoned that priming men and
women with power would attenuate differences in their reactions
The purpose of Study 1 was to develop an experimental para-
digm to examine gender differences in initiating negotiations and
to replicate and extend Babcock et al.’s (2006) research. This study
explored whether people would ask for more money in a situation
in which the option of negotiation was not made explicit. Partic-
ipants were told that they would receive cash payment ($3–$10)
for their participation in a word game task. Following the word
game, all participants were offered the minimum ($3) payment
from the experimenter. Participants were not informed of how the
payment was determined, nor were they given any performance
feedback. If they initiated a negotiation, they could receive more
money (up to $10), and if not, they received just $3. This situation
intentionally mirrors many real-world situations in which individ-
uals could negotiate and receive a better deal, but because the
negotiation opportunity is not explicit, only those who initiate
negotiations can realize a gain. We predicted that more men than
women would initiate a negotiation for higher payment. In addition
to testing our prediction about gender, this study served as a pilot
test to validate that the task (playing the word game) was gender
neutral. Specifically, we sought to make certain that men and
women did not differ in their perceived or actual performance in
Participants. A total of 74 individuals (35 men and 39
women) participated in exchange for cash payment of between $3
and $10. Participants’ ages ranged from 17 to 40 years old, with a
mean of 23 years.
Procedure. Instructions explained that participation included
four rounds of the word game “Boggle,” to be played alone in a
private cubicle. For each round, the participant was instructed to
shake a cube of lettered dice so that all letters fell into a grid at the
bottom of the cube. A timer was set, and participants searched the
grid of letters for as many words as possible until time was up.
They were instructed to search for and list all words formed from
letters that adjoined horizontally, vertically, or diagonally to the
left and right or up and down. According to the game rules, no
letter could be used more than once within a single word.
The four rounds of the game each lasted 3 min. When partici-
pants were finished with all four rounds, instructions reminded
them, “You have now completed four rounds of Boggle and will be
compensated between $3 and $10. Please indicate to the experi-
menter that you are finished, so that he or she can score your
rounds. Then you will be paid.”
The game Boggle is typically played competitively, and players
can gauge their own performance by comparing it with the per-
formance of others. However, when played autonomously, as in
our design, a player has no external cues to make this evaluation.
Moreover, because the maximum number of words possible in any
given round depends on how the lettered dice fall, the maximum
possible score can vary greatly from game to game, so there are no
evident reference points. These properties of the game limit self-
assessment to merely internal cues, which may differ by gender.
Experimenters. Two experimenters (one man and one woman)
conducted the study; each presided over approximately half of the
sample. Experimenters were naive to the hypothesis of the study.
They were trained to standardize their interactions with partici-
pants and to adhere to specific scripts (discussed later) for offering
payment to each participant.
Dependent measure. Each participant waited approximately 3
min in a cubicle while the experimenter tallied the words. Then,
the experimenter approached the participant privately, held out $3,
and said, “Here’s $3. Is $3 OK?” If a participant asked for more,
then the experimenter gave the participant the requested amount up
to the $10 maximum. If the participant complained but did not ask
for more, then the experimenter did not offer more and the par-
ticipant received $3. If the participant asked further questions
about how the payment was determined, then the experimenter
explained that a full report of the research would be provided to
them when the study was complete. The crux of this payment
procedure was that any participant could receive $10, but only if he
or she initiated a negotiation with the experimenter. If, instead, the
participant accepted the offer or complained about the offer but did
not ask for a larger payment, then the payment remained at the
Additional measures. At the end of the study, participants
answered a series of questions designed to assess perceptions of
their performance in the game. Last, participants self-reported their
age, gender, and student/nonstudent status.
Results and Discussion
Of the 74 participants, 9 participants (12.2%) asked for more
money. This low rate of requesting more cash payment is not
surprising, given the relative lack of any available cues in the
situation informing participants that asking was acceptable or
strategically wise. However, this prime-free experimental context
provided a precise means of measuring base rates of initiating
negotiation in the absence of such cues and external motivating
factors. Because all 9 individuals who negotiated asked for the
maximum payment ($10) or for more than the maximum, all 9
received $10. Therefore, we used a binary dependent variable in all
analyses (asked for more: yes–no). All 9 participants (8 men, 1
woman) who requested more money asked one experimenter (the
Gender differences. Of the men, 23% requested more money,
whereas of the women, only 3% asked for more, which was a
(1, N ⫽ 74) ⫽ 7.11, p ⬍ .01. The gender
difference in asking for more remained significant when we ex-
cluded participants who interacted with the experimenter that no
(1, N ⫽ 33) ⫽ 4.15, p ⬍ .05.
These differences in the propensity to initiate negotiation were
not driven by actual or perceived gender differences in perfor-
mance. There was no gender difference in Boggle performance
⫽ 11.57, SD ⫽ 4.69; M
⫽ 10.29, SD ⫽ 5.89), t(71) ⫽
⫺1.02, p ⫽ .31,
⫽ .02. Moreover, men and women did not
differ in their perceptions of their own performance. In response to
the question, “Compared with the average student, how well do
GENDER, FRAMING, AND INITIATION OF NEGOTIATION
you think you did today in Boggle?” participants rated their
performance on a 5-point rating scale (ranging from 1, Much worse
than the average student,to5, Much better than the average
student). There was no gender difference (M
⫽ 3.12, SD ⫽
⫽ 2.89, SD ⫽ 0.91), t(68) ⫽⫺0.95, p ⫽ .35,
.01. Of interest, perceived and actual performance were correlated
for women (r ⫽ .39, p ⬍ .05) but not for men (r ⫽ .22, p ⫽ .21).
Neither perceived nor actual performance was correlated with
initiation of negotiation for either gender (r ⫽ .16, p ⫽ .36, and
r ⫽ .07, p ⫽ .70, respectively, for men, and r ⫽ .22, p ⫽ .18, and
r ⫽ .08, p ⫽ .62, respectively, for women).
To complete our analysis of this study, we estimated a logit
model where the dependent variable was whether the participant
asked for more money (a binary measure) and the independent
variables were gender (1 if male, 0 if female) and the continuous
measure of actual Boggle performance. This analysis indicated that
even when we controlled for Boggle performance, the effect of
gender remained significant and large, B ⫽ 2.36, p ⬍ .04. Because
the magnitude of coefficients in a logit model were difficult to
interpret (Greene, 1993), we translated the coefficients into odds
ratios. The odds ratio coefficient on gender indicated that when we
controlled for performance, men were 10.6 times as likely as
women to ask for more money.
Summary. This study showed a gender difference in the propen-
sity to initiate a negotiation. To our knowledge, it is the first labora-
tory study on negotiation in which participants were not informed
explicitly that negotiation was the prescribed behavior. As predicted,
men were far more likely to ask for money than were women, despite
no apparent gender differences in actual and perceived performance.
In this study, all of those who negotiated did so with a female
experimenter. It is possible that this was due to expectations that
women would provide more money than men. However, given that
the experimenters also may have varied on a number of characteristics
(e.g., personality, attractiveness, height, attire; Rosenthal, 1966), this
may not be a reliable finding. In the studies that follow, we trained
professional actors who were matched on attractiveness and height to
minimize potential experimenter effects, as this was not the focus of
the current theory and research.
In Study 1, we examined rates of initiating negotiation in a
situation with few cues about the negotiability of the situation.
Without any explicit cue to initiate a negotiation, the overall asking
rate was low (12.2%). In the next study, we examined the question
of how cues about the negotiability of the situation affect asking
rates of men and women. More specifically, in this experiment, we
used an identical paradigm as in Study 1, but we varied situational
ambiguity by providing information to some participants about the
possibility of initiating negotiation with the experimenter. We
reasoned that two possible patterns could emerge. On the one
hand, gender differences in the initiation of negotiation may de-
pend on the strength of the situation (Mischel, 1977). Strong
situations limit the number of acceptable behavioral patterns and
reduce individual variation. By contrast, individual differences are
more likely to emerge in weak than in strong situations (Mischel,
1977), because actors in weak situations rely on internal cues, such
as attitudes, traits, and values, to guide behavioral responses (Dyk-
man, Abramson, Alloy, & Hartlage, 1989; Snyder & Ickes, 1985).
The situation participants faced in Study 1 was notably weak—
lacking explicit cues to negotiate. It is possible that gender differ-
ences could be eliminated in a stronger situation or one in which
cues for the opportunity for a negotiation were more apparent.
On the other hand, an equally plausible alternative is that gender
differences would persist despite cues to negotiate. This could
occur because recognizing that an opportunity to negotiate exists
may not be the only factor inhibiting women from initiating a
negotiation. Women may fail to initiate a negotiation even when
cued that negotiation is possible because they find negotiation to
be particularly intimidating. Some evidence for this notion can be
found in the literature. For example, Babcock et al. (2006) asked
individuals to rate their feelings toward the prospect of initiating a
negotiation and reported that women found the prospect to be more
anxiety provoking than did men. Stevens et al. (1993) similarly
found that women had lower self-efficacy about negotiating as
compared with men. In all, it is possible that the prospect of
negotiating has negative connotations for women, suggesting that
gender differences would persist when cues to negotiate are given.
We investigated both possibilities in this study.
As in previous studies, participants were told that they would
receive cash payment ranging from $3 to $10 for their participation
in a word game task. Again, all participants were offered the
minimum ($3) from the experimenter with no feedback on their
performance. The control condition provided no cue (as in Study
1). In the treatment condition, we provided a cue to negotiate.
Participants. A total of 67 individuals (33 men and 34
women) participated in exchange for cash payment of between $3
and $10. Participants’ ages ranged from 18 to 46 years old, with a
mean of 21 years.
Procedure. All procedures and measures matched those in
Study 1, with additional instructions necessary for the manipula-
tion. The experimental manipulation consisted of two variations of
written instructions given to participants just after they finished the
Boggle task. In the control condition, instructions were identical to
those given in Study 1 following the Boggle task: “You have now
completed four rounds of Boggle and will be compensated be-
tween $3 and $10. Please indicate to the experimenter that you are
finished, so that he or she can score your rounds. Then you will be
paid.” Thus, the control condition instructions gave no indication
that negotiation was an option (nor did they explicitly state that it
was not an option). The payment negotiable cue condition con-
sisted of the following additional statement, “The exact payment is
negotiable.” Thus, this cue directly informed participants that
negotiation was a viable option.
When participants finished the Boggle task, a written announce-
ment placed in front of them while they waited for the experimenter,
who was scoring their rounds. The notice thanked them for partici-
When we controlled for perceived performance rather than actual
performance in the logit model, the odds ratio coefficient on gender
remained large and significant (B ⫽ 8.80, p ⬍ .05).
We also tested for gender differences in whether participants com
plained about the payment but did not follow up with a request for a higher
payment. We found no significant gender differences in complaining ( p ⬎
.87) in Study 1 or in the subsequent studies, so we do not discuss it further.
SMALL, GELFAND, BABCOCK, AND GETTMAN
pation and repeated the manipulated instructions above about the
nature of payment determination. The experimenter left participants
alone for 3 min so that they had sufficient time to review this message
and to consider possible negotiation strategies.
Experimenters. As in Study 1, two hypotheses-naive experi-
menters (one man and one woman) each conducted the study on
approximately half of the sample. For this study, we made efforts
to control for experimenter effects by hiring and training acting
students from a reputable drama program to play the role of the
experimenters. The two experimenters, despite different genders,
were similar in appearance (height, weight, and attractiveness).
They adhered to the same scripts for offering payment to all
participants. Nevertheless, we controlled for possible idiosyncratic
experimenter effects prior to examining our hypotheses.
Results and Discussion
Across the two conditions, 22.4% of participants asked for more
money. Figure 1 presents the proportion who asked by gender in
each condition. The graph shows a 13% gender difference in the
propensity to initiate negotiation in the control condition and a
42% gender difference in the negotiate condition.
To examine whether gender differences would depend on the
situational cues, we estimated a logit model where the binary
dependent variable was assigned a 1 if the participant asked for
more money anda0ifheorshedidnotask(see first column of
Table 1). The independent variables included a binary variable for
gender (coded as 1 if the participant was male and as 0 if the
participant was female), a binary variable that controlled for the
experimenter, and a binary variable for the experimental condition
(the condition that cued the participants to negotiate was coded as
1 and the control condition was coded as 0). The likelihood ratio
tests suggests that the overall model wais significant,
(3, N ⫽
67) ⫽ 19.4, p ⬍ .001. The coefficient on gender was significantly
different from zero ( p ⬍ .01), and the odds ratio coefficient
suggested that men were 8.3 times as likely to initiate negotiations
as women are. The coefficient on the experimental condition was
also large and significant (B ⫽ 12.1, p ⬍ .01), indicating the
participants did initiate more negotiations when explicitly cued to
do so. The experimenter effect was not significantly different from
zero ( p ⫽ .93).
Since 0% of the women asked for more in the control condition, we
could not test the interaction between gender and experimental con-
dition in the logit model because the model is not identified (i.e., there
is no variation in the dependent variable in this condition—it is 0 for
every female participant in the control condition). However, we could
examine the interaction by exploring whether there was an overlap
between the 95% confidence interval around the gender difference in
the percentage who asked in the negotiate condition and the 95%
confidence interval in the control condition. The confidence interval
for the gender difference in the negotiate condition was 0.42 ⫾
1.96(.15) ⫽ .13–.71, where 1.96 is the cutoff value for a 95%
confidence interval and .15 is the standard error of the gender differ-
ence. For the control condition, the 95% confidence interval was
0.13 ⫾ 1.96(.09) ⫽⫺.05–.31. Because these confidence intervals do
indeed overlap, we can conclude that the gender difference is not
different across the two conditions ( p ⬍ .05).
As in the previous study, there were no gender differences in
perception of Boggle performance, t(52) ⫽⫺1.54, p ⫽ .13,
⫽ 3.21, SD ⫽ 0.83; M
⫽ 2.85, SD ⫽ 0.93). Men
did perform better than women this time (M
⫽ 9.38, SD ⫽
⫽ 7.02, SD ⫽ 3.50), t(57) ⫽ 3.56, p ⫽ .06,
.06. In this study, there was no significant correlation between
perceived and actual performance for either men (r ⫽ .17, p ⫽ .41)
or women (r ⫽ .02, p ⫽ .94).
Because there were gender differences in actual Boggle perfor-
mance, it is important to control for performance in the equation of
If an interaction term between gender of participant and gender of
experimenter was incorporated in the model, then the interaction was not
significant ( p ⬎ .5).
Logit Model for the Initiation of Negotiation in Study 2
Gender (1 ⫽ male, 0 ⫽ female)
SE .77 .90
Odds ratio 8.33 12.93
Condition (1 ⫽ negotiate, 0 ⫽ control)
SE .86 .90
Odds ratio 12.14 10.53
B 0.06 0.04
SE .70 .78
Odds ratio 1.06 1.04
Odds ratio 0.94
Note. The dependent variable was whether participants asked for more
money; it was coded as 1 if participants negotiated for more money and 0
if they did not.
p ⬍ .01.
Figure 1. Percentages of men and women who asked for greater com-
pensation in Study 2.
GENDER, FRAMING, AND INITIATION OF NEGOTIATION
whether the participant initiated a negotiation. We did this in the
second column of Table 1. The coefficients on gender and exper-
imental condition remained large and statistically significant once
actual performance was controlled. As in Study 1, the results were
comparable if perceptions of Boggle performance were used in-
stead of actual Boggle performance.
Summary. This study replicated the gender differences in the
initiation of a negotiation found in Study 1. As predicted, provid-
ing cues about the negotiability of payment did indeed increase the
rates of asking. However, direct cues about the negotiability of
payment did not eliminate gender differences.
The majority of women did not negotiate, even when told
explicitly that payment was negotiable. This raises the question of
why asking rates were so low among women (relative to men) in
the negotiate cue condition in which negotiating was explicitly
suggested to them. One possibility is that women reacted with
discomfort and/or anxiety to the prospect of initiating negotiation
(Babcock et al., 2006), which prohibited them from behaving in
their own interest by negotiating for more money. We further
explored this issue in the following study.
The purpose of Study 3 was to explore men and women’s
perceptions and feelings about the prospect of negotiating. On the
basis of our results from Study 2, we reasoned that even when
information exists about the viability of negotiating for greater
payment, women might still opt out because of their negative
feelings about negotiating.
Although prior research documents that woman feel more ap-
prehension about negotiation than do men (Stevens et al., 1993;
Watson & Hoffman, 1996), an empirical question is whether
alternative frames, such as asking, carry the same negative con-
notations for women. On the basis of research on the language of
politeness, we hypothesized that framing situations as an oppor-
tunity for negotiation would be particularly threatening to women.
Relative to negotiating, asking may be perceived to be a state of
placing oneself in a submissive role (Brown & Levinson, 1987;
Morand, 2000), which is more consistent with extant gender roles
(Holtgrave, 1997; Lakoff, 1975). Thus, this frame may be viewed
less negatively among women.
Moreover, framing the initiation of negotiation in a different
way could alter behavior as well (Kahneman & Tversky, 1984;
Liberman et al., 2004; McFarland & Miller, 1994). If changing the
frame has no effect, then the prospect of asking for something for
oneself would be just as inhibitive for women as the prospect of
negotiating because, arguably, the two frames describe the same
behavior. However, for women, the prospect of asking might
provoke a less inhibitory response than the prospect of negotiating
because the meaning of the cue is far less strongly associated with
high-power behavior. To test this, we first explored gender differ-
ences in perceptions and feelings about the two frames in Study 3.
In Study 4, we examined behavioral effects resulting from the two
frames. Finally, in Study 5, we directly tested the notion that
gender differences in power help to explain the differential re-
sponse to asking versus negotiating among men and women.
Participants. A total of 108 individuals (62 men and 46
women) completed a survey as part of a series of surveys for which
they received course credit. Participants’ ages ranged from 18 to
34 years old, with a mean of 19 years.
Procedure. Each participant received one of two versions of a
short survey; one framed the prospect as “negotiating” and the
other as “asking.” Half of the participants (randomly assigned)
were instructed, “We are interested in your thoughts and feelings
about negotiating for things for yourself.” The other half were
instructed, “We are interested in your thoughts and feelings about
asking for things for yourself.” Participants then rated the degree
to which they expected negotiating/asking for things to be easy/
difficult, nonthreatening/scary, agonizing/fun, and overbearing/
reasonable. Each of these word pairs possesses a valence interpre-
tation: One endpoint is positive, and the opposite is negative. Each
pair of antonyms was presented as opposite ends of a 7-point rating
scale with extremely denoted at each endpoint, followed by quite
at Points 2 and 6, then slightly at Points 3 and 5, and neither at the
midpoint. After reverse scoring so that higher scores meant more
intimidating for all items, a factor analysis produced a single factor
that accounted for 58.50% of the variance (␣⫽.73). All items had
factor loadings above .70.
Demographic measures. After completing the rating task, par-
ticipants were asked to self-report their age, gender, and highest
Results and Discussion
The means by condition are presented in Figure 2. To compare
the effects of the frame across male and female participants, we
conducted an analysis of variance (ANOVA) with both gender and
frame as independent variables and average intimidation as the
dependent variable. The results revealed no main effect of gender,
F(1, 104) ⫽ 1.53, p ⫽ .22,
⫽ .02; a significant main effect of
frame, F(1, 104) ⫽ 21.30, p ⬍ .001,
⫽ .17; and a significant
interaction between gender and frame, F(1, 104) ⫽ 13.47, p ⬍ .01,
⫽ .12. Men did not differentiate between negotiating and
⫽ 3.71, SD ⫽ 0.78, vs. M
⫽ 3.54, SD ⫽
1.00), t(60) ⫽ 0.52, p ⫽ .72,
⫽ .01. However, consistent with
our expectations, women considered negotiating to be a great deal
more intimidating than asking (M
⫽ 4.14, SD ⫽ 1.04, vs.
⫽ 2.66, SD ⫽ 0.65), t(44) ⫽⫺5.49, p ⬍ .001,
Figure 2. Average intimidation by condition in Study 3. Scales ranged
SMALL, GELFAND, BABCOCK, AND GETTMAN
Thus, the feeling about the prospect depended on frame and
This study showed that men and women reacted differently to
the prospects of asking and negotiating. These results suggest that
the provision of external cues about the negotiability of the situ-
ation may affect women and men differently, depending on the
nature of the cue. If the prospect of asking is much less intimidat-
ing for women than the prospect of negotiating, then cues to ask
might be more effective at increasing rates of initiating a negoti-
ation for women. We tested this prediction in Study 4.
In this experiment, we returned to the method used in Studies 1
and 2 in which participants could initiate a negotiation for (and
receive) greater cash payment for their participation in a word
game task after being offered the minimum ($3 from a range of
$3–$10). However, this study consisted of three conditions: (a)
control, (b) negotiating cue, and (c) asking cue. Study 3 provided
suggestive evidence that men and women have different percep-
tions and feelings about engaging in negotiating versus asking. In
the current study, we examined the behavioral consequences of
these differing perceptions and feelings. Specifically, we tested the
prediction that cuing to ask would increase rates of the initiation of
a negotiation for more money among women (more so than cuing
to negotiate), thereby reducing the gender gap.
Participants. A total of 153 individuals (81 men and 72
women) participated in exchange for cash payment of between $3
and $10. Participants’ ages ranged from 18 to 56 years old, with a
mean of 22 years.
Procedure. All procedures and measures matched those in
Studies 1 and 2. We selected two new hypotheses-naive experi-
menters (one man and one woman) closely matched on appearance
and attire to control for experimenter effects as much as possible.
As in Study 2, the experimental manipulation consisted of provid-
ing distinct cues regarding the viability of initiating a negotiation
with the experimenter for more money. Because the rate of initi-
ating a negotiation with the experimenter remained relatively low
even with the cue to negotiate in Study 2, we added an additional
sentence to the two treatment conditions, informing participants
that many participants negotiate/ask for higher payment. We rea-
soned that this social comparison information would likely boost
rates of asking for more compensation. It is important to note that
this was done across both the asking and negotiating conditions.
The first condition was a control condition, which replicated the
exact instructions in Studies 1 and 2 (no cue). In the second
condition, participants read a negotiating cue: “You will be com-
pensated between $3 and $10. The exact payment is not fixed, and
you can negotiate for more if you want. Many participants nego-
tiate for a higher payment.” In the third condition, participants read
an asking cue: “ You will be compensated between $3 and $10.
The exact payment is not fixed, and you can ask for more if you
want. Many participants ask for a higher payment.” As in Studies
1 and 2, the dependent variable was whether participants initiated
a negotiation with the experimenter for greater compensation.
Results and Discussion
Across the three conditions, 51.0% of participants asked for
more money. There were no experimenter effects. Of the partici-
pants who interacted with a male experimenter, 52.9% asked for
more, whereas of those who interacted with a female experimenter,
49.4% asked for more,
(1, N ⫽ 153) ⫽ 0.18, p ⫽ .67.
Gender differences. Figure 3 displays the proportion of par-
ticipants who asked by gender and frame. In the control condition,
the percentage of men who initiated negotiations was 21% higher
than the percentage of women. In the negotiate condition, this
percentage difference was 25%. By contrast, in the ask condition,
the percentage of women who initiated negotiations was 4% higher
than the percentage of men who negotiated. Both men and women
found approximately 12 words per Boggle round on average
⫽ 12.75, SD ⫽ 7.33; M
⫽ 12.35, SD ⫽ 6.15),
t(151) ⫽ 0.36, p ⫽ .72.
To test the hypothesis that gender differences would depend on
the situational cues, we conducted a logit regression with a dichot-
omous dependent variable assigneda1iftheparticipant asked for
more money anda0ifthey did not ask (see Table 2). The
independent variables included dummy variables for gender, the
negotiate cue, and the ask cue, as well as interaction terms of
negotiate cue with gender, ask cue with gender, and the control
variables (i.e., experimenter and actual Boggle performance).
Overall, the model was highly significant, as indicated by the
likelihood ratio test,
(7, N ⫽ 153) ⫽ 53.77, p ⬍ .001.
For this model, the coefficient on gender was positive as pre-
dicted but not significant at the .05 level (B ⫽ 2.12, p ⫽ .06; odds
ratio ⫽ 8.36). Because of the interactions in the model, this
coefficient illustrated the gender difference in the control condition
and thus the findings generally replicated those of Study 1, which
showed that women initiated negotiations less than men when
given no situational cues. Both dummy variables representing the
two cues were significant (B ⫽ 3.53, p ⬍ .001, for negotiate cue;
B ⫽ 4.20, p ⬍ .001, for ask cue). Because women were coded as
0 in the gender dummy variable, these coefficients illustrated the
effects of each cue on women’s propensity to initiate a negotiation
relative to the control condition.
We next tested whether gender differences diverged across the
control frame and the negotiate frame. The test for this is the
interaction between gender and the negotiating cue dummy, and it
Figure 3. Percentage of participants who asked for greater compensation
in Study 4.
GENDER, FRAMING, AND INITIATION OF NEGOTIATION
was nonsignificant (B ⫽ – 0.84, p ⫽ .52), indicating that, consis-
tent with Study 2, gender differences did not change even in the
presence of this cue (see Figure 3). However, consistent with the
hypothesis that cuing to ask reduces the gender gap, there was a
significant interaction between gender and the asking cue dummy
(B ⫽⫺2.53, p ⬍ .05). As can be seen in Figure 3, this reflects the
fact that the gender gap was eliminated in the asking cue condition,
despite a significant gender difference in the control condition.
Neither of the control variables (performance and experimenter)
was significant (B ⫽ 0.04, p ⫽ .20, and B ⫽⫺0.00, p ⫽ .99,
Another way to examine the data is through chi-square tests
comparing men and women within each experimental condition.
Men asked for higher payments significantly more frequently than
did women in the control condition,
(1, N ⫽ 52) ⫽ 4.31, p ⬍.04,
and more in the negotiate cue condition,
(1, N ⫽ 50) ⫽ 3.91,
p ⬍ .05, replicating our previous findings. However, cuing to ask
completely eliminated the gender effect,
(1, N ⫽ 51) ⫽ 0.30,
p ⫽ .58.
In summary, gender differences in the initiation of negotiation
appear to depend on the frame of a cue to ask for greater com-
pensation. That is, cuing to negotiate boosted rates of initiating a
negotiation among men and women, but the gender gap persisted.
However, cuing to ask rather than to negotiate enabled women to
act in a manner contrary to the typical pattern. Thus, introducing a
new framing of the behavior was sufficient to eliminate the gender
gap in the initiation of negotiation.
Having shown that the way in which negotiations are framed
affects the rates at which men and women initiate negotiations, we
set out to examine what explains the differential response to asking
versus negotiating among men and women in Study 5. As noted in
the introduction, we speculated that women’s relative lack of
power in society (Eagly & Wood, 1982; Henley & LaFrance,
1984) might lead them to view the prospect of negotiating as
intimidating. If true, then priming women with power might make
negotiating less intimidating.
To examine whether power moderates the interaction between
gender and cue, we conducted a study that first primed half of our
participants to feel powerful and then asked them to reflect on their
thoughts and feelings about the prospect of either asking or nego-
tiating. We predicted a three-way interaction of gender, framing,
and power. Specifically, in the absence of the power prime, we
expected that our results would replicate those in Study 3, in which
women felt more intimidated by the prospect of negotiating than
by the prospect of asking and no framing effect was apparent for
men. We predicted that the power prime would not affect men but
would eliminate the framing effect for women by reducing their
intimidation toward the prospect of negotiation.
Participants. A total of 149 individuals (70 men and 79
women) completed a short survey as part of a series of surveys for
which they received course credit. Participants’ ages ranged from
18 to 60 years old, with a mean of 25 years.
Procedure. We manipulated both power and frame of the
prospect between participants. For the power induction, half of the
participants were primed to experience power and the other half
engaged in a control prime. The power prime was adapted from
Galinsky et al. (2003) and involved having participants recall a
situation in which they possessed power over someone else. To
minimize heterogeneity in the types of instances different people
recalled, we tried to be specific about the meaning of power. The
instructions were as follows:
Please recall a particular incident in which you had power over
another individual or individuals. By power, we mean a situation in
which you had control and influence over others. Please describe this
situation in which you had power—what happened, how you felt, and
In the control condition, instructions read:
Please describe the way you typically spend your evenings. Begin by
writing down a description of your activities, and then figure out how
much time you devoted to each activity. Examples of things you might
describe include eating dinner, studying for a particular exam, hang-
ing out with certain friends, watching TV, and so on.
Next, each participant received one of two versions of a short
survey, which were the same manipulations that we used in Study
3. Instructions for half of the participants (randomly assigned)
read, “We are interested in your thoughts and feelings about
negotiating for things for yourself.” Instructions for the other half
read, “We are interested in your thoughts and feelings about asking
for things for yourself.” Participants then rated the degree to which
they expected negotiating/asking for things to be easy/difficult,
nonthreatening/scary, agonizing/fun, and overbearing/reasonable.
These word pairs were used in Study 3, and, as in the former study,
each of these word pairs possesses a straightforward valence
interpretation: One endpoint is positive, and the opposite is nega-
tive. Each pair of antonyms was presented as opposite ends of a
7-point rating scale with extremely denoted at each endpoint,
When we included an interaction term between gender of the partici
pant and gender of the experimenter in the model, an unexpected pattern
emerged: In the ask and negotiate cue conditions, participants were more
likely to negotiate with a same-sex experimenter ( ps ⬍ .05). This is not
consistent with the previous studies reported, and we elaborate on gender
of the experimenter in the General Discussion section.
Logit Model for the Initiation of Negotiation in Study 4
Variable BSEOdds ratio
Gender (1 ⫽ male, 0 ⫽ female) 2.12
Ask cue condition 4.20
Negotiate cue condition 3.53
Gender ⫻ Ask Cue Condition ⫺2.53
Gender ⫻ Negotiate Cue
Condition ⫺0.84 1.31 0.43
Experimenter ⫺0.00 0.40 1.00
Boggle performance 0.04 0.03 1.04
Note. The dependent variable was whether participants asked for more
money; it was coded as 1 if participants negotiated for more money and as
0 if they did not.
p ⬍ .10.
p ⬍ .05.
p ⬍ .001.
SMALL, GELFAND, BABCOCK, AND GETTMAN
followed by quite at Points 2 and 6, then slightly at Points 3 and 5,
and neither at the midpoint. Agonizing/fun and overbearing/
reasonable were reverse scored so that higher scores on all items
reflected greater intimidation to the prospect. A factor analysis
produced a single factor (␣⫽.67). All items had factor loadings
Demographic measures. After completing the task, partici-
pants were asked to self-report their age, gender, and highest
Perceptions of power. To examine whether men and women
felt equally powerful as a result of the power prime, participants
were asked to recall what they wrote about at the beginning of the
study and to recall how powerful they felt in that particular
instance on a 7-point scale ranging from 1 (completely powerless)
to7(completely powerful). The results showed that when primed
with power, men and women reported equivalent levels of expe-
rienced power, t(69) ⫽⫺0.81, p ⫽ .42, as was intended by the
Results and Discussion
As in Study 3, we averaged responses to the four items listed
above to create an average intimidation to the prospect score. To
test whether the experience of power moderates the interaction
between gender and the cues, we first conducted an ANOVA with
power, frame (ask/negotiate), and gender as the independent vari-
ables and average intimidation as the dependent variable. Results
revealed a main effect of power only at a .10 alpha level, F(1,
141) ⫽ 3.11, p ⬍ .08,
⫽ .02, no main effect of gender, F(1,
141) ⫽ 2.05, p ⫽ .16,
⫽ .01, and a main effect of frame, F(1,
141) ⫽ 4.22, p ⬍ .05,
⫽ .03, indicating that negotiating was
perceived to be more intimidating than was asking. However, these
main effects were qualified first by a two-way interaction between
power and gender, F(1, 141) ⫽ 10.81, p ⬍ .01,
indicating that power decreased intimidation for women but not for
men. Furthermore, there was a two-way interaction between frame
and gender, F(1, 141) ⫽ 15.98, p ⬍ .01,
⫽ .10, indicating that
the negotiate cue was more intimidating than the ask cue for
women but had no such effect on men. The final two-way inter-
action between power and frame was also significant, F(1, 141) ⫽
9.22, p ⬍ .01,
⫽ .06, indicating that power reduced the
intimidation of the negotiate cue but not the ask cue. However, all
of these two-way interactions were qualified by a three-way inter-
action among power, frame, and gender, F(1, 141) ⫽ 8.38, p ⬍
⫽ .06. This three-way interaction reflects the fact that the
effects of framing of the prospect depended on both power and
gender and on the interaction of power and gender.
Figure 4 presents the means across gender, frame, and power
prime. Consistent with what we found in Study 3, in the conditions
that did not manipulate power, women found the prospect of
negotiating to be much worse than the prospect of asking
⫽ 5.10, SD ⫽ 0.93, vs. M
⫽ 3.33, SD ⫽ 0.89),
t(38) ⫽⫺6.01, p ⬍ .01,
⫽.49. Also consistent with Study 3,
men did not distinguish between these frames in these conditions
⫽ 3.38, SD ⫽ 1.03, vs. M
⫽ 3.65, SD ⫽ 1.16),
t(33) ⫽ 0.70, p ⫽ .49,
⫽.02. When power was primed,
consistent with our predictions, women no longer differentiated
between negotiating and asking (M
⫽ 3.48, SD ⫽ 0.82, vs.
⫽ 3.46, SD ⫽ 0.50), t(33) ⫽⫺0.08, p ⫽ .94,
⫽ .00, and
neither did men (M
⫽ 3.59, SD ⫽ 0.83, vs. M
SD ⫽ 0.83), t(33) ⫽ 1.10, p ⫽ .28,
To consider the data in a different way, one could say that
women who were cued to negotiate found the prospect signifi-
cantly less intimidating when they were first primed to experience
power than when they were in their natural state (M
SD ⫽ 0.82, vs. M
⫽ 5.10, SD ⫽ 0.93), t(43) ⫽ 6.19, p ⬍ .01,
⫽.47. However, power did not affect women’s level of intim
idation toward the prospect of asking (M
⫽ 3.46, SD ⫽ 0.50,
⫽ 3.33, SD ⫽ 0.89), t(32) ⫽⫺0.53, p ⫽ .60,
.01. The power prime did not affect men’s level of intimidation for
either the prospect of negotiating (M
⫽ 3.59, SD ⫽ 0.83, vs.
⫽ 3.38, SD ⫽ 1.03), t(31) ⫽⫺0.63, p ⫽ .53,
or asking (M
⫽ 3.59, SD ⫽ 0.83, vs. M
⫽ 3.65, SD ⫽
1.16), t(35) ⫽⫺0.73, p ⫽ .47,
This study extends the evidence of the previous two studies by
highlighting the role of power in moderating gender differences in
reactions to the prospects of asking and negotiating. When women
are primed to experience power, their aversion to negotiating is
diminished such that they react much more like men typically do.
Put differently, these results show that women’s intimidation at the
prospect of negotiating relative to the prospect of asking can be
Negotiate/Control Negotiate/Powerful Ask/Control Ask/Powerful
Figure 4. Average intimidation by condition in Study 5. Scales ranged from 1 to 7.
GENDER, FRAMING, AND INITIATION OF NEGOTIATION
attributed to the vital role of power in thoughts and feelings related
In this article, we developed a new paradigm to study gender
and the initiation of negotiations and through five studies program-
matically illustrated how gender and frames differentially affect
the initiation of negotiations. Study 1 illustrated that there were
gender differences in the initiation of negotiation despite a lack of
gender differences in actual or perceived performance in the game.
Study 2 illustrated that cuing participants about the negotiability of
compensation also had an effect on initiation of negotiations. We
found a main effect of gender (women initiated less than men) and
a main effect of cuing to negotiate but no interaction between
gender and the cue to negotiate. Cuing women to negotiate in-
creased their rates of initiating a negotiation, but the gender gap
remained significant. In Study 3, we explored why women might
not be as responsive to cues for negotiation by examining the
notion that women would find asking to be less intimidating than
negotiating. As predicted, women had a much more negative view
of negotiating for things than of asking for things, whereas men did
not differentially respond to the two frames. Thus, we reasoned
that the prospect of negotiating may inhibit women from initiating
negotiation more so than the differently framed prospect of asking.
We tested the behavioral implications of these differences on
initiating behavior in Study 4, using the method of Studies 1 and
2. Here, we varied the description of negotiability in the cues by
telling some participants that they may negotiate for a higher
payment and others that they may ask. As predicted, we found that
the effect of gender interacts with cues— gender effects are present
when cued to negotiate but disappear when cued to ask. Finally,
Study 5 examined the psychological mechanism that helps to
explain differential reactions to negotiating versus asking. As
predicted, intimidating reactions to negotiating as compared with
reactions to asking were diminished among women when power
Theoretical and Practical Implications
This research makes a number of contributions to the literature.
First, we broadened negotiation theory and research from an ex-
amination of what happens at the negotiation table to who gets to
the table in the first place. This research begins to build theory
regarding a neglected area of negotiation research and at the same
time helps to bridge the gap between the practical reality of
everyday negotiations and how they are studied in the laboratory.
To be sure, our paradigm is only one way to examine the propen-
sity to initiate negotiations. Future research should now develop
new methodologies that use different tasks to complement this task
that can further develop this as an area of scientific inquiry. As
Gigerenzer (1991) has reminded us, often it is with the develop-
ment of scientific tools—such as new negotiation tasks in this
context—that we start to build new theories regarding psycholog-
ical phenomena. In this spirit, our method offers an alternative tool
to begin looking at different questions.
Second, this article also makes a contribution to the psychology
of gender in negotiation. By testing who initiates a negotiation and
when, we address a notable gap in the negotiation literature. We
contend that the initiation of negotiation warrants consideration
because bargaining fails to transpire if opportunities go unrecog-
nized and if individuals do not initiate negotiations. This is critical,
given that in today’s workplace, idiosyncratic deals— or deals that
arise spontaneously between employers and employees—are be-
coming much more common (Rousseau, 2005). If women are
disproportionately unlikely to initiate negotiations, then this could
be one possible cause of gender inequities.
Third, this research also contributes to the gender and negotia-
tion literature by advancing a more contextual perspective on
gender and negotiation. Although we found main effects of gender
on propensity to negotiate, we also explored a more fine-grained
contextual analysis of gender and negotiation by examining how
subtle framing can affect gender differences in negotiation. We
expanded on research in cognitive psychology (e.g., Kahneman &
Tversky, 1984) and research in the cognitive tradition in negotia-
tion (e.g., Neale & Bazerman, 1991) by showing that framing
affects initiating negotiations and, furthermore, that framing ef-
fects vary by gender. We illustrated that women associated the
framing of negotiation with much more negative connotative
meaning than the framing of asking (Study 3). As a result, we
found that framing the situation in terms of asking eliminated
gender differences in initiation negotiations.
More generally, this research clearly illustrates that gender
effects in negotiation are not static (see also Bowles et al., 2005;
Gelfand et al., 2006; Kray & Babcock, 2006; Kray & Thompson,
2005; Stuhlmacher & Walters, 1999; Walters et al., 1998) but
rather are highly malleable depending on cues in the situational
context. This article builds on the new wave of research that takes
a social– cognitive and contextual perspective on gender differ-
ences in negotiation. From a methodological point of view, our
results illustrate that the term negotiation is not gender neutral.
Given that much of the negotiation research in the laboratory is
framed with this term, it will be important to now explore gender
and bargaining behavior when given different frames.
The current research also adds to the literature by integrating
research on the psychology of power with research on gender and
negotiation. Scholars have previously argued that gender effects
may be, in part, due to differences in power (e.g., Watson, 1994).
However, these arguments have received scant empirical attention,
and moreover, research has yet to link the psychology of power to
gender differences in negotiation. We expanded the growing lit-
erature on power (Galinsky et al., 2003; Keltner et al., 2003) and
theorized that the prospect of negotiating may be implicitly asso-
ciated with situations in which one has power and thus the author-
ity to try to change the status quo, whereas the prospect of asking
is associated with being in a submissive position vis-a`-vis a deci-
sion maker. Given that women generally have less power than men
in society (Henley & LaFrance, 1984), we reasoned that the
prospect of negotiation—and thus being in a high-power role—
would feel particularly intimidating to women but the prospect of
asking would not. We therefore suspected that if we equalized
power among men and women, then frames for negotiation would
elicit less negative reactions among women, a notion that was
supported in Study 5. Put differently, the results illustrate that
placing women on the same playing field psychologically changes
the way that they perceive negotiations.
From a practical point of view, our results are also instructive.
First, the results suggest that there are multiple roads to increasing
SMALL, GELFAND, BABCOCK, AND GETTMAN
asking among women. The results suggest that interventions that
help women to relabel negotiation situations as opportunities to
ask may help to increase women’s propensity to ask to a level
equal to that of men. Framing situations as negotiations—which is
clearly the language used in negotiation training—might be less
effective. The results also point to an additional intervention that
would be useful for women, namely a focus on empowerment.
When women are primed to think about situations in which they
have had power, they are less intimidated by the prospect of
negotiating. Thus, we would argue that it is critical for negotiation
training for women to incorporate issues of power, lest the very
training that is designed to produce effective negotiators proves to
Limitations and Future Directions
These studies are not without limitations. As with other labora-
tory research, caution must be taken in generalizing the findings to
other samples and settings. At the same time, our method was
developed to reflect real-world elements of ambiguous negotiation
situations, and our results are consistent with our theory and
previous field-based results. However, because we focused on the
initiation of negotiation, our paradigm did not allow us to examine
the social outcomes of initiating a negotiation for men and women.
Although it can generally be argued that negotiating leads to better
outcomes than does not negotiating, there can also be costs asso-
ciated with negotiation (e.g., damaging a relationship), and under
certain circumstances, those costs could outweigh the benefits.
Furthermore, these costs may be more severe for women than
for men. Women who behave in a stereotypically masculine way
(e.g, authoritative, dominating) encounter social resistance (Eagly,
Makhijani, & Klonsky, 1992; Rudman, 1998). Women are often
punished for the same behaviors that are rewarded when per-
formed by men. Initiating a negotiation is one behavior that could
engender this sort of double standard. Bowles, Babcock, and Lai
(2007) found that evaluators penalized female job candidates for
initiating negotiations more so than they did male job candidates.
On the basis of this evidence of a backlash toward women who
initiate negotiations, it would be remiss to encourage women to
behave more like men, given the social environment in which they
Another potential limitation of the current work is that the
Boggle task was somehow not gender neutral. Our extensive pilot
testing, however, showed that men and women found Boggle to be
equally enjoyable, interesting, and motivating. Thus, it seems
unlikely that results were specific to the task. However, future
research should use additional tasks to triangulate on these find-
This research also only looked at effects of the gender of the
participants on initiating negotiations. Our research was not de-
signed to test for interactions between, for example, the gender of
participants and the gender of the targets, which would require
multiple experimenters for each gender. Future research should
examine how gender of the target and evaluators possibly interact
to predict asking. Other target characteristics, such as targets’
personality dispositions, might factor into individuals’ decisions to
initiate a negotiation. For example, people might be more willing
to initiate negotiations with people whom they perceive to be
highly agreeable, warm, or extraverted versus people whom they
perceive to be contentious, cold, or introverted. Likewise, the
attractiveness of targets would be an interesting variable to explore
vis-a`-vis asking behavior. In an ultimatum context, Solnick and
Schweitzer (1999) found that participants demanded more of at-
tractive people, suggesting that people may find attractive people
more approachable in our Boggle paradigm. Future research
should examine interactions among target characteristics (e.g.,
gender and attractiveness), given that these factors often interact in
complex ways in other contexts (e.g., Frieze, Olson, & Russell,
Future research could also extend the exploration of framing
negotiations beyond simply the negotiate and ask frames. There
are clearly other frames that might produce similar effects reported
in this research. Conceptually, we would expect that, to the extent
that frames reflect differential social power, they will produce
similar gender effects as found here. For example, low status has
been linked to communal characteristics (as opposed to agentic
ones), such as communality (Fragale, 2006) and interpersonal
warmth (Fiske, Cuddy, Glick, & Xu, 2002). We would expect,
therefore, that frames that reflect communality, such as coopera-
tion or collaboration, would show similar effects. Further, we have
also shown that priming with social power reduces gender effects
of frames. Situational factors that empower women, such as having
power based on position or expertise, should also produce similar
effects. For example, in situations in which women are in positions
with legitimate authority, we would expect gender differences in
propensity to initiate negotiation to diminish.
Another aspect of this article that could be further developed in
future research is the role of power, which we only introduced here
in our final study. We found that endowing participants with power
mitigated women’s intimidation about initiating negotiation when
it was framed as a negotiation, thereby eliminating the gender gap
under the negotiate frame. A future hypothesis to test is whether
inducing a feeling of powerlessness has the opposite effect on men,
that is, increasing their intimidation toward to the prospect of
negotiating. Furthermore, future tests involving power as a manip-
ulation should test this in the Boggle paradigm to examine its
influence on real behavior.
Finally, future research should examine gender and the propen-
sity to ask in organizational contexts and should particularly focus
on organizational context variables that might affect this phenom-
enon. For example, in job contexts in which negotiating is legiti-
mized as part of the job description (e.g., law), we might find that
there are fewer differences among men and women than when
there is ambiguity regarding legitimacy. Likewise, in organiza-
tional cultures and climates in which idiosyncratic deals are per-
ceived to be commonplace and acceptable, there should be fewer
differences in gender and the initiation of negotiation, given that
initiating negotiations may be a normal part of organizational life.
Social networks might moderate the effects found here. To the
extent that women are centrally located in social networks, they are
more likely to hear about opportunities for resources that exist, as
well as feel more empowered to initiate negotiations on their own
behalf. Although these research questions go beyond the scope of
the current article, we are now conducting research on these
variables to broaden the theory and capture the phenomenon of
initiating negotiations in its situational complexity.
In conclusion, this article broadens existing negotiation theory
and research by moving beyond what happens when people are at
GENDER, FRAMING, AND INITIATION OF NEGOTIATION
the negotiating table by focusing on an important necessary con-
dition: that individuals have gotten to the table in the first place.
Our results advance a social– cognitive perspective on gender and
negotiation by showing that gender interacts with framing to
predict initiation of negotiation and that such processes are based
within differential power that is cultivated among men and women.
Illuminating contextual fluidity of gender effects and their linkages
to power opens up a new way of studying gender in negotiation.
Anderson, C., & Berdahl, J. L. (2002). The experience of power: Exam-
ining the effects of power on approach and inhibition tendencies. Jour-
nal of Personality and Social Psychology, 83, 1362–1377.
Areni, C. S., & Sparks, J. R. (2005). Language power and persuasion.
Psychology and Marketing, 6, 507–525.
Babcock, L., Gelfand, M., Small, D., & Stayn, H. (2006). Gender differ-
ences in the propensity to initiate negotiations. In D. De Cremer, M.
Zeelenberg, & J. K. Murnighan (Eds.), Social psychology and economics
(pp. 239 –262). Mahwah, NJ: Erlbaum.
Bowles, H. R., Babcock, L., & Lai, L. (2007). Social incentives for gender
differences in the propensity to initiate negotiations: Sometimes it does
hurt to ask. Organizational Behavior and Human Decision Processes,
103, 84 –103.
Bowles, H. R., Babcock, L., & McGinn, K. L. (2005). Constraints and
triggers: Situational mechanics of gender in negotiation. Journal of
Personality and Social Psychology, 89, 951–965.
Brown, P., & Levinson, S. C. (1987). Politeness: Some universals in
language usage. New York: Cambridge University Press
Bureau of Labor Statistics, United States Department of Labor. (2002).
Union members in 2001. Retrieved December 8, 2005, from ftp://
Bureau of Labor Statistics, United States Department of Labor. (2006).
Job openings and labor turnover: December 2005 (Report USDL
06-225). Retrieved from http://www.bls.gov/news.release/archives/
Clark, M. L. (1983). Effect of confederate’s sex and status on competitive
behavior of male and female college students. The Journal of Psychol-
ogy, 113, 191–198.
Dykman, B. M., Abramson, L. Y., Alloy, L. B., & Hartlage, S. (1989).
Processing of ambiguous feedback among depressed and nondepressed
college students: Schematic biases and their implications for depressive
realism. Journal of Personality and Social Psychology, 56, 431– 445.
Eagly, A. H., Makhijani, M. G., & Klonsky, B. G. (1992). Gender and the
evaluation of leaders: A meta-analysis. Psychological Bulletin, 111,
Eagly, A. H., & Wood, W. (1982). Inferred sex differences in status as a
determinant of gender stereotypes about social influence. Journal of
Personality and Social Psychology, 45, 915–928.
Fiske, S. T., Cuddy, A. J. C., Glick, P., & Xu, J. (2002). A model of (often
mixed) stereotype content: Competence and warmth respectively follow
form perceived status and competition. Journal of Personality and
Social Psychology, 82, 878 –902.
Fragale, A. R. (2006). The power of powerless speech: The effects of
speech style and task interdependence on status conferral. Organiza-
tional Behavior and Human Decision Processes, 101, 243–261.
Frieze, I. H., Olson, J. E., & Russell, J. (1991). Attractiveness and income
for men and women in management. Journal of Applied Social Psychol-
ogy, 21, 1039 –1057.
Galinsky, A. D., Gruenfeld, D. H., & Magee, J. C. (2003). From power to
action. Journal of Personality and Social Psychology, 85, 453– 466.
Gelfand, J. J., Major, V., Raver, J., Nishii, L., & O’Brien, K. (2006).
Negotiating relationally: The dynamics of the relational self in negoti-
ation. Academy of Management Review, 31, 427– 451.
Gerhart, B., & Rynes, S. (1991). Determinants and consequences of salary
negotiations by male and female MBA graduates. Journal of Applied
Psychology, 76, 256 –262.
Gigerenzer, G. (1991). From tools to theories: A heuristic of discovery in
cognitive psychology. Psychological Review, 98, 254 –267.
Goffman, E. (1967). Interaction ritual: Essays on face-to-face interaction.
Oxford, England: Aldine.
Greene, W. H. (1993). Econometric analysis (2nd ed.). New York: Mac-
Griffith, C. E. (1991). Personality and gender as factors in interpersonal
negotiation. Journal of Social Behavior and Personality, 6, 915–928.
Haferkamp, C. J. (1991–1992). Orientations to conflict: Gender, attribu-
tions, resolution strategies, and self-monitoring. Current Psychology:
Research and Reviews, 10, 227–240.
Henley, N. M., & LaFrance, M. (1984). Gender as culture: Difference and
dominance in nonverbal behavior. In A. Wolfgang (Eds.), Nonverbal
behavior: Perspectives, applications, intercultural insights (pp. 351–
371). Lewiston, NY: Hogrefe.
Holtgrave, T. (1997). Styles of language use: Individual and cultural
variability in conversational directness. Journal of Personality and So-
cial Psychology, 73, 624 – 627.
Kahneman, D., & Tversky, A. (1984). Choices, values, and frames. Amer-
ican Psychologist, 39, 341–350.
Keaveny, T. J., & Inderrieden, E. J. (2000). Gender differences in pay
satisfaction and pay expectations. Journal of Managerial Issues, 12,
Keltner, D., Gruenfeld, D. H., & Anderson, C. (2003). Power, approach,
and inhibition. Psychological Review, 110, 265–285.
Kipnis, D. (1972). Does power corrupt? Journal of Personality and Social
Psychology, 24, 33– 41.
Kray, L., & Babcock, L. (2006). Gender in negotiations: A motivated
social cognitive analysis. In L. L. Thompson (Ed.), Negotiation theory
and research (pp. 203–223). Madison, CT: Psychosocial Press.
Kray, L. J., Galinsky, A., & Thompson, L. (2001). Battle of the sexes:
Gender stereotype confirmation and reactance in negotiations. Journal of
Personality and Social Psychology, 80, 942–959.
Kray, L. J., Galinsky, A. D., & Thompson, L. (2002). Reversing the gender
gap in negotiations: An exploration of stereotype regeneration. Organi-
zational Behavior and Human Decision Processes, 87, 386 –409.
Kray, L. J., & Thompson, L. (2005). Gender stereotypes and negotiation
performance: An examination of theory and research. In B. Staw & R.
Kramer (Eds.), Research on organizational behavior (Vol. 26, pp. 103–
182). Greenwich, CT: JAI Press.
Kyl-Heku, L. M., & Buss, D. M. (1996). Tactics as units of analysis in
personality psychology: An illustration using tactics of hierarchy nego-
tiation. Personality and Individual Differences, 21, 497–517.
Lakoff, R. (1975). Language and a woman’s place. New York: Harper &
Liberman, V., Samuels, S. M., & Ross, L. (2004). The name of the game:
Predictive power of reputations versus situational labels in determining
prisoner’s dilemma game moves. Personality and Social Psychology
Bulletin, 30, 1175–1185.
Lind, E. A., Huo, Y. J., & Tyler, T. R. (1994). And justice for all: Ethnicity,
gender, and preferences for dispute resolution procedures. Law and
Human Behavior, 18, 269 –290.
Magee, J. C., Galinsky, A. D., & Gruenfeld, D. H. (2007). Power, propen-
sity to negotiate, and moving first in competitive interactions. Person-
ality and Social Psychology Bulletin, 33, 1–13.
Major, B., Vanderslice, V., & McFarlin, D. B. (1984). Effects of pay
expected on pay received: The confirmatory nature of initial expecta-
tions. Journal of Applied Social Psychology, 14, 399 –412.
Maxwell, D. (1992). Gender differences in mediation style and their impact
on mediator effectiveness. Mediation Quarterly, 9, 353–364.
McFarland, C., & Miller, D. T. (1994). The framing of relative perfor-
SMALL, GELFAND, BABCOCK, AND GETTMAN
mance feedback: Seeing the glass as half empty or half full. Journal of
Personality and Social Psychology, 66, 1061–1073.
Mesch, D. J., & Dalton, D. R. (1989). Gender context in resolution of
American workplace conflicts. The Journal of Social Psychology, 129,
Mischel, W. (1977). The interaction of person and situation. In D. Mag-
nusson & N. S. Endler (Eds.), Personality at the crossroads: Current
issues in interactional psychology (pp. 333–352). Hillsdale, NJ: Erl-
Morand, D. A. (2000). Language and power: An empirical analysis of
linguistic strategies used in superior-subordinate communication. Jour-
nal of Organizational Behavior, 21, 235–248.
Neale, M. A., & Bazerman, M. H. (1991). Cognition and rationality in
negotiation. New York: The Free Press.
Neu, J., Graham, J. L., & Gilly, M. C. (1988). The influence of gender on
behaviors and outcomes in a retail buyer–seller negotiation simulation.
Journal of Retailing, 64, 427– 451.
Papa, M. J., & Natalle, E. J. (1989). Gender, strategy selection, and
discussion satisfaction in interpersonal conflict. Western Journal of
Speech Communication, 53, 260 –272.
Putnam, L. L., & Jones, T. S. (1982). Reciprocity in negotiations: An
analysis of bargaining interaction. Communication Monographs, 49,
Rosenthal, R. (1966). Experimenter effects in behavioral research. New
Rousseau, D. (2005). I-deals: Idiosyncratic deals employees bargain for
themselves. New York: Sharpe.
Rubin, J. Z., & Brown, B. R. (1975). The social psychology of bargaining
and negotiation. New York: Academic.
Rudman, L. A. (1998). Self-promotion as a risk factor for women: The
costs and benefits of counterstereotypical impression management. Jour-
nal of Personality and Social Psychology, 74, 629 –645.
Snyder, M., & Ickes, W. (1985). Personality and social behavior. In G.
Lindzey & E. Aronson (Eds.), Handbook of social psychology (3rd ed.,
Vol. 2, pp. 883–947). New York: Random House.
Solnick, S., & Schweitzer, M. E. (1999). The influence of physical attrac-
tiveness and gender on ultimatum game decisions. Organizational Be-
havior and Human Decision Processes, 79, 199 –215.
Stevens, C. K., Bavetta, A. G., & Gist, M. E. (1993). Gender differences in
the acquisition of salary negotiation skills: The role of goals, self-
efficacy, and perceived control. Journal of Applied Psychology, 78,
Stuhlmacher, A. F., & Walters, A. E. (1999). Gender differences in
negotiation outcomes: A meta-analysis. Personnel Psychology, 52, 653–
Tharenou, P. (1999). Is there a link between family structures and women’s
and men’s managerial career advancement? Journal of Organizational
Behavior, 20, 837– 863.
Tharenou, P. (2001). Going up? Do traits and informal social processes
predict advancing in management? Academy of Management Journal,
Valian, V. (1998). Why so slow? The advancement of women. Cambridge,
MA: MIT Press.
Wachter, R. M. (1999). The effect of gender and communication mode on
conflict resolution. Computers in Human Behavior, 15, 763–782.
Walters, A. E., Stuhlmacher, A. F., & Meyer, L. L. (1998). Gender and
negotiator competitiveness: A meta-analysis. Organizational Behavior
and Human Decision Processes, 76, 1–29.
Watson, C. (1994). Gender versus power as a predictor of negotiator
behaviour and outcomes. Negotiation Journal, 10, 117–127.
Watson, C., & Hoffman, L. R. (1996). Managers as negotiators: A test of
power versus gender as predictors of feelings, behavior, and outcomes.
Leadership Quarterly, 7, 63– 85.
Received December 15, 2005
Revision received January 16, 2007
Accepted February 5, 2007 䡲
GENDER, FRAMING, AND INITIATION OF NEGOTIATION