Content uploaded by Marwan Sinaceur
Author content
All content in this area was uploaded by Marwan Sinaceur
Content may be subject to copyright.
Hot or Cold:
Is Communicating Anger or Threats More Effective in Negotiation?
Marwan Sinaceur
INSEAD
Gerben A. Van Kleef
University of Amsterdam
Margaret A. Neale
Stanford University
Hajo Adam
Northwestern University
Christophe Haag
EM Lyon
Is communicating anger or threats more effective in eliciting concessions in negotiation? Recent research
has emphasized the effectiveness of anger communication, an emotional strategy. In this article, we argue
that anger communication conveys an implied threat, and we document that issuing threats is a more
effective negotiation strategy than communicating anger. In 3 computer-mediated negotiation experi-
ments, participants received either angry or threatening messages from a simulated counterpart. Exper-
iment 1 showed that perceptions of threat mediated the effect of anger (vs. a control) on concessions.
Experiment 2 showed that (a) threat communication elicited greater concessions than anger communi-
cation and (b) poise (being confident and in control of one’s own feelings and decisions) ascribed to the
counterpart mediated the positive effect of threat compared to anger on concessions. Experiment 3
replicated this positive effect of threat over anger when recipients had an attractive alternative to a
negotiated agreement. These findings qualify previous research on anger communication in negotiation.
Implications for the understanding of emotion and negotiation are discussed.
Keywords: anger, negotiation, conflict, threat, emotion communication
The use of anger and the use of threats are two common
strategies in negotiation (e.g., Barry, 1999). Anecdotes suggest that
both are effective. For example, British Prime Minister Winston
Churchill was persuasive because of his ability to communicate
anger, according to French President Charles De Gaulle (De
Gaulle, 1954/1964, pp. 57–58). U.S. President John F. Kennedy
used threats effectively during the Cuban Missile Crisis, declaring
an ultimatum to the Soviets while appearing poised in a televised
address on October 22, 1962 (Schlesinger, 1965, pp. 832– 840).
Although anger and threats may both seem effective strategies in
negotiation, it is not clear whether the more emotional strategy or
the colder strategy is more effective. In this article, we examine
this question: Is communicating anger or threats more effective in
eliciting concessions in negotiation? In doing so, we bridge two
streams in negotiation and conflict research that evolved indepen-
dently: recent research on emotions and classic literature on
threats.
Anger Communication in Negotiation
Affect in general (e.g., Elfenbein, 2007; Forgas & George,
2001) and emotional communication in particular (Barry, 2008)
have become important areas in negotiation research. Of the var-
ious emotions that may be communicated in negotiations, anger is
one of the most prominent (Allred, 1999; Barry, 1999; Daly, 1991;
Van Kleef, Van Dijk, Steinel, Harinck, & Van Beest, 2008).
Drawing on the idea that emotional communication can be strate-
gic (Barry, 1999; Clark, Pataki, & Carver, 1996; Kopelman, Ro-
sette, & Thompson, 2006; Thompson, Nadler, & Kim, 1999),
recent research has investigated the effectiveness of anger com-
munication in negotiation. It has found that communicating anger
(compared to a control condition or to communicating happiness)
can be an effective strategy in eliciting concessions from a coun-
terpart in negotiation (e.g., Sinaceur & Tiedens, 2006; Steinel, Van
Kleef, & Harinck, 2008; Van Kleef & Coˆte´, 2007; Van Kleef & De
Dreu, 2010; Van Kleef, De Dreu, & Manstead, 2004a, 2004b,
2010).
This article was published Online First June 20, 2011.
Marwan Sinaceur, Department of Organizational Behavior, INSEAD,
Fontainebleau, France; Gerben A. Van Kleef, Department of Social Psy-
chology, University of Amsterdam, Amsterdam, the Netherlands; Margaret
A. Neale, Department of Organizational Behavior, Stanford University;
Hajo Adam, Department of Management & Organizations, Northwestern
University; Christophe Haag, Department of Management, EM Lyon,
Lyon, France.
This research was generously funded by the INSEAD Alumni Fund
(INSEAD Grant 2520-306) through an INSEAD R&D Committee Grant
that was awarded to Marwan Sinaceur.
Correspondence concerning this article should be addressed to Marwan
Sinaceur, INSEAD, Department of Organizational Behavior, Boulevard de
Constance, 77300 Fontainebleau, France. E-mail: marwan.sinaceur@
insead.edu
Journal of Applied Psychology © 2011 American Psychological Association
2011, Vol. 96, No. 5, 1018–1032 0021-9010/11/$12.00 DOI: 10.1037/a0023896
1018
Implicit in this positive effect of anger communication (vs. a
control) on concessions is the idea that anger conveys the threat of
an impasse or other detrimental consequences. For example, the
communication of anger was found to be effective because nego-
tiators use the emotion as information to appraise the counterpart’s
limit, inferring that a counterpart who communicates anger is close
to impasse (Van Kleef et al., 2004a); thus, communicating anger in
negotiation acts as a “signal that an impasse is particularly likely”
(Van Kleef & Coˆte´, 2007, p. 1558). Consistent with this notion,
communicating anger is effective only when the recipients of that
communication have poor alternatives to the negotiation and there-
fore have a lot to lose in case of an impasse (Sinaceur & Tiedens,
2006). Similarly, the communication of anger induces recipients to
make a conciliatory offer only when the costs associated with
rejection of the offer are high (Van Dijk, Van Kleef, Steinel, &
Van Beest, 2008).
These findings are in line with arguments that the communica-
tion of anger has a deterrent effect on recipients in negotiations
(Daly, 1991; Morris & Keltner, 2000). Anger “indicates potential
‘deal breakers’” (Morris & Keltner, 2000, p. 29); it communicates
that “negative consequences (such as an impasse) may ensue”
(Daly, 1991, p. 36). Hence, anger “conveys threatening messages”
(Friedman et al., 2004, p. 370). Anger thus serves the social
function of conveying a threat (Averill, 1982). At a basic level,
emotion theorists have argued that anger generally signifies a
threat to the recipients of that emotional expression (Aronoff,
Woike, & Hyman, 1992; Goos & Silverman, 2002; Knutson, 1996;
Lerner & Tiedens, 2006; Marsh, Ambady, & Kleck, 2005). Indeed,
it has been found that anger appears intimidating to recipients
(Clark et al., 1996). On the basis of these arguments that one
primary social function of anger is to convey a threat, we predict
the following:
Hypothesis 1: Perceptions of threat conveyed by anger me-
diate the positive effect of anger communication (vs. a con-
trol) on concessions in negotiation.
If anger indeed poses a threat, the question arises as to whether
it is the emotional communication of anger per se or the implied
threat conveyed by the anger that makes recipients concede. If
anger were effective because of the threat it implicitly contains,
then a direct threat may be even more effective in eliciting con-
cessions. When we make a direct comparison of the effects of
anger and threat rather than between anger and a control condition
(the latter being the focus of past research), we can better disen-
tangle the effects of anger from threat on concessions. In sum,
although anger is more effective than a control comparison point,
it may not be more effective than a direct threat.
The Role of Threats
A threat may be defined as a conditional statement, usually of
the form “if . . . then,” that mentions a negative consequence
associated with not complying, for example, “If you don’t make
further effort, we won’t reach any deal. Take it or leave it”
(Bacharach & Lawler, 1981; De Dreu, 1995; Lytle, Brett, &
Shapiro, 1999; Pruitt, 1981; Rubin & Brown, 1975; Schelling,
1960). As Pruitt (1981, p. 77) proposed, “a threat is a communi-
cation of intent to punish the other if the other fails to concede.”
Communicating a threat is often an effective strategy to elicit
concessions in conflict and negotiation (e.g., Faley & Tedeschi,
1971; Michener & Cohen, 1973; Pruitt, 1981; Rubin & Brown,
1975; Sinaceur & Neale, 2005; Tedeschi & Bonoma, 1977).
Although a threat may be accompanied by anger (Frank, 1988;
Morris & Keltner, 2000), it need not be. A threat may be delivered
in a cool, nonemotional way, resulting from the threatener’s care-
ful consideration of his or her options (Galinsky & Liljenquist,
2004; Schelling, 1960). Indeed, qualitative observation of negoti-
ations suggests that ultimatums can be presented in a matter-of-
fact way (Kwon & Weingart, 2004, p. 272). Similarly, prior
research has suggested that negotiators can make intimidating
statements (such as threats) without necessarily being viewed as
attacking (Lytle et al., 1999, p. 35). Hence, despite threats and
anger being often associated with one another (Pruitt, 1981; Te-
deschi & Bonoma, 1977), they are distinguishable. In the current
research, we thus compared the respective effects of anger and
threats on concession making.
Is Anger or Threat Communication More Effective?
One important distinction between communicating threats and
anger is that threats are associated with a negotiator being per-
ceived as confident and in control (Friedland, 1976). In contrast,
emotional expressions such as anger, it has been argued, reflect a
lack of poise (Dewey, 1895; Hebb, 1949), particularly in negoti-
ation (Adler, Rosen, & Silverstein, 1998; Jackall, 1988). Typically,
anger is positively correlated with impulsiveness (Dahlen, Martin,
Ragan, & Kuhlman, 2004; Ramı´rez & Andreu, 2006), while colder
forms of aggression are not (Ramı´rez & Andreu, 2006). Anger
communication connotes less composure (Underwood, Hurley,
Johanson, & Mosley, 1999) than nonemotional communication.
Thus, it is plausible that compared to hot or emotional anger, a
cold or nonemotional threat will be associated with more perceived
poise, that is, a greater perceived sense of confidence and control
over one’s own feelings and decisions (The Oxford Dictionary of
English, Soanes & Stevenson, 2005). Specifically, we hypothe-
sized that counterparts who issue threats would be perceived as
having greater poise than those who communicate anger. This
hypothesized difference mirrors a classic distinction in social
psychology between instrumental, calculated aggression and im-
pulsive, affective aggression (Berkowitz, 1974, 1993). Specifi-
cally, when aggression is marked by intense emotions such as
anger, it is perceived as more impulsive and less controlled.
In turn, we hypothesized that threats would be more effective
than anger communication in eliciting concessions because of the
greater poise conveyed by threats. A negotiator lacking poise may
be perceived as not in control, not so likely to follow through on
his or her words, or not having seriously thought about possible
consequences. In particular, compared to instrumental, calculated
aggression, impulsive, affective aggression such as anger is asso-
ciated with less conscious premeditation and less thinking about
consequences (Berkowitz, 1974, 1993; Kimble & Perlmuter,
1970). So, compared to a composed, threatening negotiator, an
angry negotiator may be less perceived as having premeditated his
or her actions and as having anticipated what steps to take if the
recipient were not to comply. Thus, recipients are less likely to
infer that a poised, threatening counterpart exaggerated his or her
actual willingness to follow through on the threat (Bacharach &
1019
ANGER AND THREATS IN NEGOTIATIONS
Lawler, 1981). The greater poise ascribed to a threatening ex-
presser makes it more likely that the threat will be viewed as
informative and real (Friedland, 1976).
Note that we do not argue that poise is sufficient to elicit
concessions independent of the negative valence, or toughness, of
the strategy used. Poise in more neutral or positive contexts (e.g.,
poise conveyed by a charismatic or happy negotiator) might not
elicit concessions. Rather, we propose that, in the specific context
of value-claiming strategies such as anger and threats, which are
both seen as tough (De Dreu, 1995; Pruitt, 1981; Sinaceur &
Tiedens, 2006; Tedeschi & Bonoma, 1977; Van Kleef & De Dreu,
2010), relatively greater poise characterizes threats in comparison
to anger, which elicits relatively larger concessions in that context.
Overall, we predict the following:
Hypothesis 2: Threat communication is more effective than
anger communication in eliciting concessions in negotiation.
Hypothesis 3: The positive effect of threat communication
compared to anger communication in eliciting concessions is
mediated by poise ascribed to the threatener.
The Present Studies
To disentangle the effects of communicating anger from com-
municating threat, we conducted one pilot study and three exper-
iments. We conducted the pilot study to investigate whether people
can differentiate between anger and threat statements. Then, we
conducted three computer-mediated negotiation experiments in
which participants received either angry or threatening messages
from a simulated counterpart. In Experiment 1, the design included
two experimental conditions (anger communication vs. control) to
test the idea that anger induces concessions because it conveys an
implied threat. Drawing from that, in Experiments 2 and 3, we
made a direct comparison of anger to threat. Therefore, in Exper-
iment 1, the comparison was between anger and control (to test
that threat underlies the effect of anger), whereas, in Experiments
2 and 3, the comparison was between anger and threat (to test
whether a direct threat is more effective than anger). In Experiment
2, the design was a 2 (anger vs. threat communication) !2 (early
vs. late communication) factorial design. We wanted to test for a
positive effect of threat over anger at different points in a negoti-
ation because prior research found that people concede late, rather
than early, in negotiation (e.g., Moore, 2004; Pruitt & Drews,
1969; Yukl, 1974a). In Experiment 3, the design included two
experimental conditions (anger vs. threat communication) in a
situation in which recipients had an attractive alternative to the
current negotiation. We wanted to replicate the effect of threat over
anger in this situation because prior research found that having an
attractive alternative can affect negotiators’ reactions to the strat-
egies of the opponent (Komorita & Barnes, 1969; Yukl, 1974b).
Pilot Study
Phase 1
Fifty-three undergraduate students, graduate students, and pro-
fessionals participated in Phase 1 of the pilot study. They were
randomly recruited in the street near several urban university
campuses. Participants were volunteers participating on the spot,
and their educational background, age, and gender were not re-
corded. In a repeated-measures design, each participant rated a
series of 35 negotiation phrases on whether these expressed anger
and/or threats.
The phrases intended to communicate anger were taken from
previous research on anger in negotiation (Sinaceur & Tiedens,
2006; Van Kleef & Coˆte´, 2007; Van Kleef et al., 2004a, 2004b).
For example, phrases were “I’m fed up with this. You need to
make real efforts! It really starts to make me annoyed,” “This
begins to seriously get on my nerves!”, and “I am very angry with
your offer.” The phrases intended to communicate threat(s) were
adapted from previous negotiation research on threats (De Dreu,
1995; Shapiro & Bies, 1994; Sinaceur & Neale, 2005). For exam-
ple, phrases were “If you don’t make real efforts, we will reach no
agreement for this negotiation,” and “If you don’t modify your
offer, there will be consequences. It is up to you.” Participants
were told that the phrases were taken from real-life negotiations.
For each phrase, participants indicated the extent to which they
thought it expressed anger and/or a threat. (The item was worded
as follows: “For each phrase, indicate the extent to which you think
it expresses a threat, and the extent to which you think it expresses
anger. Each phrase can express both a threat and anger, express
neither a threat nor anger, express a threat only, or express anger
only.”) Participants’ perceptions of a phrase expressing anger
and/or threats were measured on 9-point scales (1 "little, 9"
much).
Paired-sample ttests for each of the 35 phrases probed whether
each phrase communicated more anger, more threat, or similar
levels of both. Sixteen phrases were rated as communicating more
anger than threats (all ts#2.06, all ps$.05), 15 phrases were
rated as communicating more threats than anger (all ts#3.08, all
ps$.005), and four showed no difference (all ts$2.00, all ps#
.05).
These paired-sample ttests suggested that participants were
often able to differentiate between anger and threat phrases. How-
ever, to increase confidence in our manipulation of anger versus
threat, we selected the five most differentiated angry phrases (all
ts#7.10, all ps$.001) and the five most differentiated threat-
ening phrases (all ts#9.10, all ps$.001). Then, we combined the
most differentiated angry phrases into two longer angry state-
ments. Similarly, we combined the most differentiated threatening
phrases into two longer threatening statements. This resulted in
two angry statements and two threatening statements of similar
length that are displayed in Table 1.
1
Phase 2
Thirty-four undergraduate students participated in Phase 2 of the
pilot study. They were selected in exactly the same way as in Phase
1, and their educational background, age, and gender were not
recorded. In a repeated-measures design, each participant rated the
two angry statements and the two threatening statements that had
1
The resulting angry statements were similar to the ones by Van Kleef
and his colleagues (2004a), except that they were longer. These were
combined with angry statements by Sinaceur and Tiedens (2006) to get
angry and threatening statements of similar length.
1020 SINACEUR, VAN KLEEF, NEALE, ADAM, AND HAAG
been created from the results of Phase 1. The instructions and
measures were the same as in Phase 1.
Table 1 presents means. Paired-sample ttests for the four
statements showed that each statement was rated as intended (all
ts#3.60, all ps$.001). Specifically, the two angry statements
were rated as more angry than threatening, and the two threatening
statements were rated as more threatening than angry. Thus, this
pilot study provides evidence that people can differentiate between
anger and threat statements. In Experiments 1–3, we used the
statements resulting from the pilot study to manipulate anger
versus threat.
Experiment 1
The goal of Experiment 1 was twofold. First, we wanted to test
the argument that the effectiveness of anger may lie in the threat
implied. Although the idea that anger conveys an implied threat
(e.g., of an impasse or other detrimental consequences) is consis-
tent with results from prior research (e.g., Sinaceur & Tiedens,
2006; Van Kleef et al., 2004a), this relationship and its causal role
have not been tested directly. Thus, in Experiment 1, we investi-
gated whether perceptions of threat causally underlie the effect of
anger on concessions.
Second, as can be seen in Table 1, the anger statements that
resulted from the pilot study included exclamation marks, whereas
the threat statements did not. Exclamation marks may indicate
different levels of intensity and appropriateness for anger, thus
causing different levels of concessions. To alleviate that potential
confound of our anger manipulation, Experiment 1 included two
versions of the anger statements: the statements that resulted from
the pilot study (see Table 1) and the same statements without
exclamation marks. To see if there were differences between these
two versions of the anger manipulation, we measured perceptions
of intensity and appropriateness in addition to examining conces-
sions.
Method
Participants. Thirty-eight master’s and undergraduate stu-
dents (21 women and 17 men; mean age "21.18 years, SD "
3.09) participated in the study for compensation (a voucher worth
five euros). They were randomly recruited in the street near several
urban university campuses. Their educational background was
either in literature (60.53%) or social science (39.47%). To ac-
count for the heterogeneity of the sample, we controlled for de-
mographics (i.e., educational background, age, and gender) in all
analyses. However, not including participant demographics as
controls did not change any of the results reported below.
Experimental design. Participants were randomly assigned
to one of two experimental conditions: anger communication ver-
sus control. This was done via a double-blind procedure in which
neither the participants nor the experimenter knew to which con-
dition each participant was assigned. We included two versions of
the anger manipulation to see if the presence of the exclamation
marks in the anger statements (exclamation marks present vs.
exclamation marks absent) mattered.
Procedure. For each session, four to six participants were
invited to the laboratory. On arrival, participants were seated in
separate cubicles in front of a computer. All materials were pre-
sented on the computer screen. Participants read that the purpose
of the experiment was to study negotiation in a situation in which
the parties could not see each other. They were led to believe that
they would engage in a computer-mediated negotiation with an-
other participant who was in one of the other cubicles. In reality,
the other participant’s behavior was simulated by the computer.
Negotiation task. All participants engaged in a negotiation
task that was an adapted version of the one used by Van Kleef and
colleagues (Steinel et al., 2008; Van Kleef & Coˆte´, 2007; Van
Kleef et al., 2004a, 2004b; Van Kleef, De Dreu, Pietroni, &
Manstead, 2006; see also De Dreu, 1995; De Dreu & Van Lange,
1995). The task captures main characteristics of real-life negotia-
tions. Specifically, there are multiple issues and multiple agree-
ment positions, there is only information about one’s own payoffs,
and it contains the typical offer and counteroffer sequence. Par-
ticipants learned that they would be assigned the role of either the
buyer or the seller of a consignment of mobile phones and that
their objective was to negotiate the price, the warranty period, and
the number of offered models for the phones.
After a short pause, during which the computer supposedly
assigned buyer and seller roles, all participants were told that they
had randomly been assigned the role of seller. They were then
presented with a payoff chart and were told that their objective was
to earn as many points as possible. For each of the three negoti-
ation issues there were nine possible agreement positions (Position
1 yielded the maximum number of points; Position 9 yielded the
minimum number of points). The payoff table showed the number
of points that participants could make depending on the agreement
Table 1
Pilot Study: Statements Used to Manipulate Threat Versus Anger
Statement
Anger perception
pretest (N"34)
Threat perception
pretest (N"34)
Anger statement: “I am very angry with your offer. This begins to seriously get on my nerves!” 6.18
a
(1.72) 4.39
b
(2.26)
Anger statement: “This negotiation really makes me angry. I’m fed up with this. You need to make real
efforts! It really starts to make me annoyed. All this is not serious!!!” 6.97
a
(1.70) 5.62
b
(1.79)
Threat statement: “If you don’t seriously modify your offer, there will be consequences. It is up to you . . .” 3.97
a
(2.35) 7.33
b
(1.95)
Threat statement: “I’m warning you. There will be no contract concluded. If you don’t make real efforts,
we will reach no agreement for this negotiation. Take it or leave it.” 3.56
a
(2.20) 6.65
b
(2.03)
Note. Column 1 presents the statements resulting from Phase 1 of the pilot study. Columns 2 and 3 present tests for these statements in Phase 2 of the
pilot study. Values across columns that do not share a subscript differ significantly at p$.001 according to paired-sample ttests. Standard deviations are
in parentheses.
1021
ANGER AND THREATS IN NEGOTIATIONS
position reached for each negotiation issue. To ease participants’
computations of point totals, a calculator was available in each
cubicle and each negotiation issue was given the same weight in
points. Furthermore, participants saw examples of computations of
the point totals that they would obtain if they achieved two specific
settlements: the best possible settlement for them and the worst
possible settlement for them. The payoff table for the counterpart
was not shown, and participants were told only that it differed from
their own.
To enhance involvement, participants were informed that the
points they earned would be converted to lottery tickets after the
negotiation. They were told that by earning more points, they
would collect more lottery tickets and increase their chance of
winning one of the prizes (i.e., MP3 players). To emphasize the
mixed-motive nature of the negotiation, participants were told that
only those who reached an agreement would participate in the
lottery. Thus, there was an incentive to earn as many points as
possible as well as an incentive to reach agreement. Participants
were then informed that the negotiation would end when an
agreement was reached or when time ran out.
Participants also learned that an additional goal of the study was
to examine the effects of having versus not having information
about the counterpart’s reactions. They read that the computer had
randomly determined that they would receive information about
the counterpart’s reactions but that the counterpart would not
receive information about their reactions. This allowed us to in-
clude the counterpart’s anger communication with the counter-
part’s reactions, as described below. After a short pause, during
which the computer supposedly prepared the negotiation, partici-
pants were told that the buyer (i.e., the counterpart) would make a
first offer, which they could counter.
The negotiation began with the buyer (the computer) making a
first offer. Over six negotiation rounds, the buyer proposed the
following levels of agreement for price, warranty, and number of
models, respectively: 8 –7– 8 (Round 1), 8 –7–7 (Round 2), 8 – 6 –7
(Round 3), 7– 6 –7 (Round 4), 7–6–6 (Round 5), and 6–6–6
(Round 6). Prior research has shown that these specific proposals
have face validity and are seen as intermediate in cooperativeness
and competitiveness (De Dreu & Van Lange, 1995). Every time
after the simulated counterpart made an offer, the participant
responded to the simulated counterpart’s offer by making an offer
himself or herself, with the participant going from making an
initial offer to making a final offer. A proposal by the participant
was accepted if it equaled or exceeded the offer the computer was
about to make in the next round. If no agreement were reached
after the sixth round, the negotiation was interrupted (cf. De Dreu
& Van Lange, 1995; Van Kleef & Coˆte´, 2007).
Manipulation.
Anger communication versus control. In the angry counter-
part condition (N"19), participants received a message from the
alleged buyer disclosing his or her reactions before the last two
rounds (i.e., before Round 5 and before Round 6). These two
messages contained the manipulation of the buyer’s anger com-
munication. In the control condition (N"19), participants re-
ceived no messages about the buyer’s reactions.
Two versions of the anger manipulation. There were two
versions for the anger manipulation. One version used as the two
messages from the buyer (i.e., the computerized counterpart) the
two angry statements that resulted from the pilot study and that are
presented in Table 1, which included exclamation marks (N"10).
The other version used the same statements but without any
exclamation marks (N"9).
Dependent measures.
Outcome. The final offer made by participants in the nego-
tiation sequence was the main dependent variable. It was obtained
by adding together the levels of each issue in the final proposal:
price, warranty, and number of models. Lower (higher) numbered
options on these issues were better for participants (for the oppo-
nent). Hence, the final offer by participants represented a total
concessions index: the higher the final offer, the greater the par-
ticipants’ concessions (possible range "3–27, actual range "
8 –19). In all analyses on final offers, we controlled for partici-
pants’ initial offers (i.e., offers by participants at Round 1, before
any manipulation occurred) because research has documented that
negotiation outcomes are strongly influenced by initial offers
(Adair, Weingart, & Brett, 2007; Galinsky & Mussweiler, 2001;
Magee, Galinsky, & Gruenfeld, 2007).
Perception of threat. Perception of threat was the hypothe-
sized mediator of the effect of anger on concessions. This was
measured by having participants rate four items (“Did the coun-
terpart explicitly state a threat against you in the negotiation?”,
“Did the behavior of your counterpart implicitly convey a threat?”,
“To what extent does the adjective ‘threatening’ describe your
counterpart given the reactions he/she sent?”, and “To what extent
does the adjective ‘intimidating’ describe your counterpart given
the reactions he/she sent?”; 1 "little, 9"very much), which were
averaged into a single index (Cronbach’s %".92).
Perceptions of intensity and appropriateness. To see if there
were differences between the two versions of the anger manipu-
lation, we measured perceptions of intensity and appropriateness.
Perception of intensity was measured by having participants rate
two items (“Were your counterpart’s reactions intense?” and
“Were your counterpart’s reactions extreme?”; 1 "little, 9"very
much), which were averaged into a single index (Cronbach’s %"
.82). Perception of appropriateness was measured by having par-
ticipants rate two items (“Were your counterpart’s reactions
proper?” and “Were your counterpart’s reactions legitimate?”; 1 "
little, 9"very much), which were averaged into a single index
(Cronbach’s %".76).
Manipulation check.
Anger. To check the effectiveness of the anger communica-
tion manipulation, we measured perceptions of anger expressions
in the counterpart using three items on 9-point scales (“To what
extent did the counterpart express anger in the negotiation?”, “To
what extent did the counterpart express irritation in the negotia-
tion?”, and “To what extent did the counterpart express discontent
in the negotiation?”; 1 "little, 9"very much). These were
averaged into a single index (Cronbach’s %".95).
Results
As indicated before, we controlled for demographics in all
analyses, thus reporting analyses of covariance (ANCOVAs) be-
low. However, not including participant demographics as controls
did not change any of the results reported below.
Differences between the two versions of the anger manipu-
lation. We first examined differences between the two versions
of the anger manipulation for all dependent measures to see if the
1022 SINACEUR, VAN KLEEF, NEALE, ADAM, AND HAAG
presence versus absence of exclamation marks mattered. Specifi-
cally, we did so for the anger manipulation check, concessions
(i.e., final offers), perceptions of threat, perceptions of intensity,
and perceptions of appropriateness. There were no differences
between the two versions of the anger manipulation (all five Fs$
.24, all five ps#.62). On that basis, we collapsed the two versions
of the anger condition in subsequent analyses.
Anger manipulation check. An ANCOVA (strategy: anger
communication vs. control) on the anger manipulation check
showed that participants in the anger condition thought that the
counterpart expressed more anger (M"7.18, SD "1.43) than
control participants (M"2.76, SD "2.00), F(1, 33) "62.48, p$
.0001, &
2
".65.
Concessions. We examined the total amount of concessions
made by participants, which was operationalized as their final
offer. An ANCOVA (strategy: anger communication vs. control)
was conducted on final offers, controlling for participants’ initial
offers (i.e., offers by participants at round 1, before any manipu-
lation occurred). As expected, there was a main effect for partic-
ipants’ initial offers, such that participants who made small con-
cessions initially also made small concessions finally, F(1, 32) "
21.47, p$.0001, &
2
".40. Above and beyond that, there was a
main effect of strategy, such that participants made larger conces-
sions to an angry counterpart (M"15.02, SD "2.70) than to a
control counterpart (M"13.14, SD "2.89), F(1, 32) "6.47, p$
.02, &
2
".17. This positive effect of anger communication in
eliciting concessions replicates prior work.
Perceived threat implied by anger. We argued that recipi-
ents would perceive an angry counterpart as conveying a threat. An
ANCOVA (strategy: anger communication vs. control) on per-
ceived threat showed that, indeed, an angry counterpart was per-
ceived as threatening (M"5.19, SD "1.68) compared to a
control counterpart (M"1.65, SD "1.29), F(1, 33) "52.47, p$
.0001, &
2
".61.
Mediation analysis. Next, we examined whether perceived
threat mediated the positive effect of anger on concessions. We
followed the approach of Baron and Kenny (1986), controlling for
participants’ initial offers. The results of the analyses are presented
in Figure 1.
First, the anger versus control factor predicted participants’
concessions, '".34, t(32) "2.54, p$.02, such that anger
elicited larger concessions. Second, the anger versus control factor
also predicted participants’ perception of threat, '".79, t(33) "
7.24, p$.00001, such that an angry counterpart was perceived as
more threatening. Third, participants’ perception of threat was
associated with larger concessions, '".50, t(32) "4.32, p$
.0002. Fourth, a regression on concessions was conducted with the
anger versus control factor and participants’ perception of threat as
simultaneous predictors. In this regression, the effect of the anger
versus control factor was no longer significant, '"(.12, t(31) "
(0.67, p#.50, whereas the effect of participants’ perception of
threat remained significant, '".59, t(31) "3.23, p$.003. To
test the significance of the indirect effect (i.e., the path through the
mediator), we followed a bootstrapping procedure (Preacher &
Hayes, 2008, 2009), as recommended by Shrout and Bolger
(2002). We used Preacher and Hayes’s (2008, 2009) multiple
mediation model because it allows one to include covariates (but
using their simple mediation model yielded similar results). The
result of 1,000 resamples demonstrated that zero fell outside of the
95% confidence interval (CI) of the indirect effect (95% CI [1.42,
4.69]). Thus, recipients’ perception of threat mediated the positive
effect of anger on concessions.
Discussion
Experiment 1 provides direct evidence that anger elicits conces-
sions in negotiations because it conveys a threat (Hypothesis 1).
Although this idea has long been suggested by negotiation (e.g.,
Morris & Keltner, 2000) and emotion (e.g., Marsh et al., 2005)
theorists alike, this relationship had not been established by em-
pirical testing. Yet, if anger were effective because of the implied
threat, then it is plausible that a direct threat would be even more
effective in eliciting concessions than anger because negotiators
who display anger are likely to be perceived as lacking poise
(compared to negotiators who issue threats) and this perception is
likely to decrease recipients’ compliance. Hence, in Experiments
2–3, we made a direct comparison of anger and threat and explored
whether a threat would be more effective than anger.
In doing so, we used the angry and threatening statements that
resulted from the pilot study (see Table 1). Experiment 1 found no
difference between anger with exclamation marks and anger with-
out exclamation marks. Thus, it seemed unlikely that potential
differences between our angry and threatening statements were the
result of punctuation differences that could cause angry statements
to be seen as more intense and inappropriate. We returned to this
issue in Experiment 3.
Experiment 2
The goal of Experiment 2 was to directly compare the effects of
anger and threat on concession making in negotiation. As elabo-
rated earlier, we predicted that threats would be more effective
than anger (Hypothesis 2) and that this relative effect would be
mediated by perceived poise (Hypothesis 3).
In addition, we examined the role of timing in a more
exploratory fashion. Prior research has shown that timing af-
fects concessions in negotiation. Specifically, negotiators make
more concessions late rather than early in the negotiation (Ben-
ton, Kelley, & Liebling, 1972; Moore, 2004; Pruitt & Drews,
1969; Yukl, 1974a). This timing effect has led scholars to
propose that the threat of an impasse looms larger (Pruitt, 1981)
and that communicating threats could be more effective (Schell-
ing, 1960) late rather than early in negotiation. Empirical evi-
dence also supports the notion that the timing of a threat is
important: late explicit threats were found to elicit more con-
cessions than early explicit threats (Sinaceur & Neale, 2005).
We therefore included a timing manipulation in our design to
Perceived
threat
β = .79 *** β = .50 *** / β = .59 ***
Counterpart communicates Concessions
anger vs. Control counterpart by participants
β = .34 * / β = -.12 ns
Figure 1. Experiment 1: Mediation of the anger versus control effect by
perceived threat.
!
p$.05.
!!!
p$.001.
1023
ANGER AND THREATS IN NEGOTIATIONS
explore whether anger and threats are more effective when
issued later rather than earlier in a negotiation and to insure that
the predicted positive effect of threats over anger holds above
and beyond this timing effect.
Method
Participants. Two-hundred and seven master’s and under-
graduate students (109 women and 98 men; mean age "21.12
years, SD "2.76) participated in the study for compensation (a
voucher worth five euros). They were randomly recruited in the
street near several urban university campuses. Their educational
backgrounds were broad, including areas such as medicine, finan-
cial management, economics, literature, history, and law. We
coded educational background as either emphasizing quantitative
skills (e.g., medicine) or not (e.g., literature). To account for the
heterogeneity of the sample, we controlled for demographics (i.e.,
educational background, age, and gender) in all analyses. How-
ever, not including participant demographics as controls did not
change any of the results reported below.
Experimental design. The experimental design was a 2
(strategy: anger vs. threat communication) !2 (timing: early vs.
late communication) factorial, between-subject design. Partici-
pants were randomly assigned to experimental conditions accord-
ing to a double-blind procedure in which neither the participants
nor the experimenter knew to which condition each participant was
assigned.
Procedure. The procedure was identical to the one used in
Experiment 1. Again, participants negotiated with a simulated
counterpart in a sequence of six rounds.
2
Manipulations.
Anger versus threat communication. Before two successive
rounds (the timing manipulation below indicates which rounds)
out of the six negotiation rounds, participants received a message
from the buyer (i.e., the computerized counterpart) disclosing his
or her reactions. These messages contained the manipulation of the
buyer’s anger versus threat communication. For the manipulation,
we used the two angry statements and the two threatening state-
ments that resulted from the pilot study (see Table 1). In the angry
counterpart condition (N"104), participants received the two
angry statements as messages from the counterpart before two
successive rounds. In the threatening counterpart condition (N"
102), participants received the two threatening statements as mes-
sages from the counterpart before two successive rounds.
Timing. In the early condition (N"107), participants re-
ceived the messages from the alleged buyer, which contained the
anger or threat statements, before Round 2 and before Round 3 in
the negotiation sequence. In the late condition (N"99), partici-
pants received the same messages before Round 5 and before
Round 6 in the negotiation sequence. This manipulation of timing
is consistent with that of prior research (e.g., Pruitt & Drews, 1969;
Yukl, 1974a).
Dependent measures.
Outcome. As in Experiment 1, the final offer by participants
represented the total amount of concessions they made across the
three negotiation issues (possible range "3–27, actual range "
3–22).
Perception of poise. Perception of poise between threat and
anger was the hypothesized mediator. This was measured by
having participants rate five items, based on the definition of poise
as perceived sense of confidence and control (The Oxford Diction-
ary of English, Soanes & Stevenson, 2005; “To what extent does
the adjective ‘self-confident’ describe your counterpart given the
reactions he/she sent?”, “Did you think that your counterpart had
control of his/her emotions?”, “Did you think that your counterpart
was sure of his/her decisions?”, “Did your counterpart seem to lose
his/her grip during the negotiation?” [reverse scored], and “Did
your counterpart seem to control his/her impulses?”; 1 "little, 9"
very much), which were averaged into a single index (Cronbach’s
%".72). A confirmatory factor analysis verified that these items
loaded on one factor (it yielded only one factor with an eigenvalue
greater than 1, i.e., 2.39, which explained 47.77% of the total
variance; also, all items loaded positively on that factor, with all
factor loading ls#.57).
Manipulation checks.
Anger. To check the effectiveness of the anger communica-
tion manipulation, we measured perceptions of anger expressions
in the counterpart using the same three items as in Experiment 1.
These were averaged into a single index (Cronbach’s %".80).
Threat. To check the effectiveness of the threat communica-
tion manipulation, we measured perceptions of threats communi-
cated by the counterpart using two items on 9-point scales (“Did
the counterpart explicitly state a threat against you in the negoti-
ation?” and “Did the counterpart explicitly threaten to take steps
against you?”; 0 "not at all, 8"very much). These were
averaged into a single index (Cronbach’s %".86). To further
probe the threat and anger manipulations, we added a dichotomous
item about perceptions of cold versus more emotional threats (“If
you answered 1 or more to the prior question, did the counterpart
express his/her threats in a cold way or in an emotional way?”),
which could be answered by either “In a cold way” or “In an
emotional way.”
Timing. To check the effectiveness of the timing manipula-
tion, we used dichotomous items for anger (“If you think the
counterpart expressed anger during the negotiation, when did he or
she express anger?”) and for threat (“If you think the counterpart
explicitly stated a threat, when did he or she state a threat?”).
These could be answered by either “Start of the negotiation” or
“End of the negotiation.”
Results
As indicated before, we controlled for demographics in all
analyses, thus reporting ANCOVAs below. However, not includ-
ing participant demographics as controls did not change any of the
results reported below.
Manipulation checks.
Anger. A 2 (strategy: anger vs. threat communication) !2
(timing: early vs. late communication) ANCOVA on the anger
manipulation check showed that participants in the anger condition
thought that the counterpart expressed more anger (M"7.60,
SD "1.85) than participants in the threat condition (M"5.81,
SD "1.86), F(1, 200) "47.58, p$.001, &
2
".19. There were
no main or interaction effects due to timing.
2
One participant made an initial offer that exceeded the offer by the
computer in round 2, so her negotiation ended before any manipulation and
she could not be included in the analyses.
1024 SINACEUR, VAN KLEEF, NEALE, ADAM, AND HAAG
Threat. A 2 (strategy: anger vs. threat communication) !2
(timing: early vs. late communication) ANCOVA on the threat
manipulation check showed that participants in the threat condition
thought that the counterpart communicated more threats (M"
4.80, SD "2.61) than participants in the anger condition (M"
2.99, SD "2.60), F(1, 200) "24.74, p$.001, &
2
".11. There
were no main or interaction effects due to timing.
In addition, we conducted a logistic regression on the dichoto-
mous check about perceptions of cold versus emotional threats
with the anger versus threat factor, the timing factor, and their
interaction as predictors (all participants were asked to answer this
check). This indicated that participants in the threat condition
found their counterpart’s threats to be cold rather than emotional
compared to participants in the anger condition, B"1.86,
Wald(1) "15.66, Ex(B)"0.16, p$.001. There were no main or
interaction effects due to timing.
Timing. A logistic regression on the timing of anger manip-
ulation check with the anger versus threat factor, the timing factor,
and their interaction as predictors indicated that participants in the
early condition thought that anger, if any, had been expressed
earlier rather than later compared to participants in the late con-
dition, B"4.69, Wald(1) "29.64, Ex(B)"109.24, p$.001.
There were no other effects. A similar logistic regression on the
timing of threat manipulation check indicated that participants in
the early condition thought that threats, if any, had been expressed
earlier rather than later compared to participants in the late con-
dition, B"4.00, Wald(1) "36.74, Ex(B)"54.41, p$.001.
3
Concessions. We examined the total amount of concessions
made by participants. A 2 (strategy: anger vs. threat communica-
tion) !2 (timing: early vs. late communication) ANCOVA was
conducted on final offers, controlling for participants’ initial offers
as in Experiment 1. The means are displayed in Figure 2. As was
also the case in Experiment 1, there was a main effect for partic-
ipants’ initial offers, F(1, 198) "93.89, p$.001, &
2
".32.
Above and beyond that, there was a main effect of strategy, such
that participants made larger concessions to a threatening counter-
part (M"15.03, SD "2.77) than to an angry counterpart (M"
13.98, SD "2.76), F(1, 198) "7.18, p$.01, &
2
".04. This
positive effect of threat over anger communication in eliciting
concessions supports Hypothesis 2. Consistent with prior research,
there was also a main effect of timing, as participants made larger
concessions to a counterpart using strategies late (M"14.89,
SD "2.75) than to a counterpart using strategies early (M"
14.12, SD "2.75), F(1, 198) "3.92, p$.05, &
2
".02, in the
negotiation. No interaction qualified the strategy and timing ef-
fects, F(1, 198) ".005, ns.
Perceived poise. We hypothesized that recipients would per-
ceive a threatening counterpart to be more poised than an angry
counterpart. A 2 (strategy: anger vs. threat communication) !2
(timing: early vs. late communication) ANCOVA on perceived
poise showed that participants ascribed more poise to a threatening
counterpart (M"3.36, SD "1.52) than to an angry counterpart
(M"2.60, SD "1.52), F(1, 200) "12.73, p$.001, &
2
".06.
There were no main or interaction effects due to timing on per-
ceived poise.
Mediation analysis. Next, we examined whether recipients’
perception of the counterpart’s poise mediated the positive effect
of threat over anger in eliciting concessions. We followed the
approach of Baron and Kenny (1986), controlling for participants’
initial offers and including the timing factor (and its interaction
term) in all regressions. The results of the analyses are presented
in Figure 3.
First, the anger versus threat factor predicted participants’ con-
cessions, '".15, t(198) "2.69, p$.01, such that threats elicited
more concessions than anger. Second, the anger versus threat
factor also predicted participants’ perception of the counterpart’s
poise, '".24, t(200) "3.53, p$.001, such that participants
ascribed more poise to a threatening counterpart than to an angry
counterpart. Third, perceived poise was associated with larger
concessions, '".18, t(200) "3.24, p$.002. Fourth, a regres-
sion on concessions was conducted with the anger versus threat
factor and participants’ perception of the counterpart’s poise as
simultaneous predictors. In this regression, the effect of the anger
versus threat factor was reduced, '".11, t(197) "2.01, p".045,
whereas the effect of participants’ perception of the counterpart’s
poise remained significant, '".15, t(197) "2.70, p".007. To
test the significance of the indirect effect (i.e., the path through the
mediator), we followed a bootstrapping procedure (Preacher &
Hayes, 2008, 2009). The result of 1,000 resamples demonstrated
that zero fell outside of the 95% CI of the indirect effect of
perceived poise (95% CI [0.09, 0.57]). Thus, recipients’ perception
of the counterpart’s poise partially mediated the positive effect of
threat over anger, supporting Hypothesis 3.
Discussion
Experiment 2 provides evidence that communicating threats is a
more effective strategy than communicating anger. Specifically,
participants made more concessions to a counterpart communicat-
ing threats than to a counterpart communicating anger (Hypothesis
3
There was a marginal effect for the interaction, B"1.71, Wald(1) "
3.70, Ex(B)"0.18, p".054, such that participants in the anger condition
(who did not receive threat communications) had more difficulty discrim-
inating between early and late threat compared to participants in the threat
condition, as all participants were asked to answer the timing checks and
anonapplicable choice was not offered.
12
13
14
15
16
ETALYLRAE
Total concessions (last offer made)
ANGER
COMMUNICATION
THREAT
COMMUNICATION
Figure 2. Experiment 2: Total of concessions made by participants. A
higher score indicates larger concessions.
1025
ANGER AND THREATS IN NEGOTIATIONS
2). Also, timing mattered such that value-claiming strategies (i.e.,
communicating threat or anger) were generally more effective later
rather than earlier in the negotiation. This timing effect is consis-
tent with prior literature on negotiation (Moore, 2004; Pruitt, 1981;
Sinaceur & Neale, 2005; Yukl, 1974a) and other social interactions
(Frantz & Bennigson, 2005). Yet the positive effect of threat over
anger held above and beyond this timing effect. Furthermore,
Experiment 2 tested the proposed mechanism and showed that
recipients’ perception of the counterpart’s poise mediated the
positive effect of threat versus anger on concessions (Hypoth-
esis 3).
One shortcoming of Experiment 2 is that participants did not
have an alternative to the negotiation, so they may have per-
ceived strong pressure to reach an agreement. Alternatives to
the negotiation are an important source of power in negotiation
and are therefore highly relevant to how much a negotiator
concedes (Mannix, Thompson, & Bazerman, 1989; Pinkley,
Neale, & Bennett, 1994; White, Valley, Bazerman, Neale, &
Peck, 1994; Wolfe & McGinn, 2005). Thus, the fact that par-
ticipants did not have an alternative potentially limits our
contribution. Would the positive effect of threats over anger
extend to a situation in which participants have an attractive
alternative to the negotiation?
On the one hand, recipients who have an attractive alternative
to the negotiation have little motivation to pay attention to or to
react to the counterpart’s displays of emotions, especially the
counterpart’s displays of anger. Having an attractive alternative
is related to low epistemic motivation (e.g., Van Kleef et al.,
2004b) and less experienced dependency (e.g., Van Kleef &
Coˆte´, 2007) toward another’s emotions in negotiation. In turn,
the positive effect of an opponent’s anger on recipients’ con-
cession making is reduced (as compared to happiness or neutral
emotion; Sinaceur & Tiedens, 2006; Van Kleef & Coˆte´, 2007;
Van Kleef et al., 2004b, 2006) when recipients have an attrac-
tive alternative. Thus, it is possible that anger would be less
effective than threats in this case.
On the other hand, having an attractive alternative makes a
negotiator generally less susceptible to the tactics of the opponent
(Bacharach & Lawler, 1981; Komorita & Barnes, 1969; Yukl,
1974b). In particular, recipients who have an attractive alternative
are less affected by tough moves from the counterpart (Komorita
& Barnes, 1969), which might include the communication of anger
as well as the communication of threats. Hence, it is theoretically
possible that the effects of threats would not differ from those of
anger when recipients have an attractive alternative. On the basis
of this latter argument, we thought that it was important to repli-
cate the difference between anger and threats when recipients have
an attractive alternative. We conducted Experiment 3 with that
goal in mind, giving all participants an attractive alternative to the
negotiation.
Experiment 3
The main goal of Experiment 3 was to compare the effects of
anger and threats on concession making when recipients had an
attractive alternative to the negotiation—a source of power in
negotiation (e.g., Pinkley et al., 1994). Thus, in Experiment 3, we
gave all participants an attractive alternative to the negotiation and
examined the resulting effects of anger versus threats on partici-
pants’ concessions. Another goal of Experiment 3 was to rule out
two possible alternative explanations for the difference between
anger and threats. In Experiment 1, we measured perceptions of
intensity and appropriateness and found that these were not af-
fected by the presence versus absence of exclamation marks in the
angry statements. However, it is still possible that, in comparison
to threats, anger might be seen as overly intense and/or more
inappropriate, and that these perceptions drive the positive effect
of threats over anger. Thus, in Experiment 3, we again measured
perceptions of intensity and appropriateness and examined
whether either mediated our hypothesized effect.
Method
Participants. Fifty-eight master’s and undergraduate stu-
dents (26 women and 32 men; mean age "21.14 years, SD "
2.30) participated in the study for compensation (a voucher worth
five euros). They were recruited in the same manner as in prior
experiments. Their educational background was diverse and was
coded as either emphasizing quantitative skills (e.g., economics) or
not (e.g., history). As before, we controlled for demographics (i.e.,
educational background, age, and gender) in all analyses, thus
reporting ANCOVAS below. However, not including participant
demographics as controls did not change any of the results re-
ported below.
Experimental design. Participants were randomly assigned
to one of two experimental conditions: anger communication (N"
33) versus threat communication (N"25). Assignment to condi-
tions occurred via a double-blind procedure.
Procedure. The procedure was identical to the one used in
Experiments 1 and 2, with two exceptions. First, participants
negotiated over four rounds rather than six. Thus, over the four
negotiation rounds, the buyer (i.e., the computer) proposed the
following levels of agreement for price, warranty, and number of
models, respectively: 8 –7– 8 (Round 1), 8 –7–7 (Round 2), 8 – 6 –7
(Round 3), and 7– 6 –7 (Round 4). As indicated before, prior
research has shown that these specific proposals are seen as
intermediate in cooperativeness and competitiveness (De Dreu &
Van Lange, 1995).
Second, all participants were given an attractive alternative to
the negotiation (whereas no information about alternatives had
been given before). In this study, the alternative to the negoti-
ation was set at 16 for all participants, an outcome that was
clearly better than the last offer made by the computer in Round
4 or the last round of the negotiation (the lower the offer, the
better it was for participants). Thus, all participants read in their
instructions prior to the negotiation that they had another offer
that was worth 16. Specifically, they were told that they had
Perceived
poise
β = .24 *** β = .18 ** / β = .15 **
Counterpart communicates Concessions
threats rather than anger by participants
β = .15 ** / β = .11 *
Figure 3. Experiment 2: Mediation of the threat versus anger effect by
perceived poise.
!
p$.05.
!!
p$.01.
!!!
p$.001.
1026 SINACEUR, VAN KLEEF, NEALE, ADAM, AND HAAG
already received an offer from another, important buyer for the
same consignment of mobile phones, which was worth 16. They
were told they could end the negotiation with their current
counterpart at any time and accept their valuable alternative
(e.g., Pinkley et al., 1994). They were reminded of these in-
structions right before starting the negotiation.
Manipulations. The statements for the manipulations were
identical to those in Experiment 2. Participants received these
statements before Round 3 and before Round 4, which were now
the later rounds.
Dependent measures.
Outcome. As before, the final offer by participants repre-
sented the total of concessions they made across the three negoti-
ation issues (possible range "3–27, actual range "9 –16).
Perception of poise. We measured perception of poise using
a combination of items used in Experiment 2 and some new
items to ensure that our mediation was not limited to one
particular set of items. Three items were the same as in Exper-
iment 2 (i.e., “To what extent does the adjective ‘self-confident’
describe your counterpart given the reactions he/she sent?”,
“Did you think that your counterpart was sure of his/her deci-
sions?”, and “Did your counterpart seem to lose his/her grip
during the negotiation?” [reverse scored]; 1 "little, 9"very
much), and two items were new (i.e., “To what extent did your
counterpart have a good head on his/her shoulders given the
reactions he/she sent?” and “Did the counterpart seem to care-
fully weigh his/her decisions?”; 1 "little, 9"very much).
These five items were averaged into a single index (Cronbach’s
%".74). A confirmatory factor analysis verified that these
items loaded on one factor (it yielded only one factor with an
eigenvalue greater than 1, i.e., 2.47, which explained 49.41% of
the total variance; also, all items loaded positively on that factor
with all factor loading ls#.60).
Perceptions of intensity and inappropriateness. We also
wanted to rule out perceptions of intensity and inappropriateness
as potential alternative explanations associated with our anger
versus threats manipulations. Perception of intensity was measured
using two items (“Were your counterpart’s reactions intense?” and
“Were your counterpart’s reactions exaggerated?”; 1 "little, 9"
very much), which were averaged into a single index (Cronbach’s
%".73). Perception of inappropriateness was measured using two
items (“Were your counterpart’s reactions inappropriate?” and
“Were your counterpart’s reactions rude?”; 1 "little, 9"very
much), which were averaged into a single index (Cronbach’s %"
.82).
Checks.
Bargaining power. To check that our alternative instructions
elicited a good bargaining power perception, we measured partic-
ipants’ perception of their bargaining power using a dichotomous
item (“At the start of the negotiation, did you think that you were
in a good bargaining position or in a bad bargaining position facing
the counterpart?”), which could be answered by either “In a good
bargaining position” or “In a bad bargaining position.”
Anger and threat. The manipulation checks for anger com-
munication (%".87), for threat communication (%".86), and for
perceptions of cold versus more emotional threats were the same
as in Experiment 2.
Results
Checks.
Bargaining power. We first checked that most participants
thought they were in a good bargaining position. Indeed, 47
participants (81%) thought they were so, which was significantly
greater than chance, )
2
(1, N"58) "22.34, p$.0001. We report
below results with all participants throughout all analyses. How-
ever, it is important to note that not including the 11 participants
(19%) who failed the bargaining power check did not change any
of the results reported below.
Anger. An ANCOVA (strategy: anger vs. threat communica-
tion) on the anger manipulation check showed that participants in
the anger condition thought that the counterpart expressed more
anger (M"8.34, SD "0.76) than did participants in the threat
condition (M"6.26, SD "1.76), F(1, 53) "42.08, p$.0001,
&
2
".44.
Threat. An ANCOVA (strategy: anger vs. threat communi-
cation) on the threat manipulation check showed that participants
in the threat condition thought that the counterpart expressed more
threats (M"4.37, SD "2.15) than did participants in the anger
condition (M"2.61, SD "2.54), F(1, 53) "7.57, p$.01, &
2
"
.12. In addition, we conducted a logistic regression on the dichot-
omous check about perceptions of cold versus emotional threats
with the anger versus threat difference as a factor. This indicated
that participants in the threat condition found their counterpart’s
threats to be cold rather than emotional compared to participants in
the anger condition, B"2.54, Wald(1) "14.01, Ex(B)"12.68,
p$.0005.
Concessions. Next, we examined concessions by partici-
pants. An ANCOVA (strategy: anger vs. threat communication)
was conducted on final offers, controlling for participants’ initial
offers as before. The means are displayed in Figure 4. As expected,
there was a main effect for participants’ initial offers, F(1, 52) "
16.98, p$.001, &
2
".25. Above and beyond that, there was a
main effect of strategy, such that participants made larger conces-
sions to a threatening counterpart (M"15.10, SD "1.12) than to
12
13
14
15
16
ANGER THREAT
Total concessions (last offer made)
Figure 4. Experiment 3: Total of concessions made by participants when
participants were given an attractive alternative. All participants were
given an attractive alternative to the negotiation. A higher score indicates
larger concessions.
1027
ANGER AND THREATS IN NEGOTIATIONS
an angry counterpart (M"14.29, SD "1.57), F(1, 52) "5.97,
p$.02, &
2
".10. This latter result is consistent with the main
result from Experiment 2, yet it extends this to a situation when
participants were given an attractive alternative to the negotiation.
Perceived poise. We argued that recipients would perceive a
threatening counterpart to be more poised than an angry counter-
part. An ANCOVA (strategy: anger vs. threat communication) was
conducted on perceived poise. It showed that participants ascribed
more poise to a threatening counterpart (M"4.08, SD "0.83)
than to an angry counterpart (M"2.57, SD "1.61), F(1, 53) "
19.23, p$.0001, &
2
".27.
Mediation by perceived poise. Then, we examined whether
recipients’ perception of the counterpart’s poise mediated the
positive effect of threat over anger in eliciting concessions. We
followed the approach of Baron and Kenny (1986), controlling for
participants’ initial offers as in Experiments 1 and 2. The results of
the analyses are presented in Figure 5.
First, the anger versus threat factor predicted participants’ con-
cessions, '".28, t(52) "2.44, p$.02, such that threats elicited
more concessions than anger. Second, the anger versus threat
factor also predicted participants’ perception of the counterpart’s
poise, '".50, t(53) "4.39, p$.0001, such that participants
ascribed more poise to a threatening counterpart than to an angry
counterpart. Third, perceived poise was associated with larger
concessions, '".38, t(52) "3.30, p$.002. Fourth, a regression
on concessions was conducted with the anger versus threat factor
and participants’ perception of the counterpart’s poise as simulta-
neous predictors. In this regression, the effect of the anger versus
threat factor was no longer significant, '".13, t(51) "0.97, p#
.33, whereas the effect of participants’ perception of the counter-
part’s poise remained significant, '".31, t(51) "2.31, p".025.
To test the significance of the indirect effect (i.e., the path through
the mediator), we followed a bootstrapping procedure (Preacher &
Hayes, 2008, 2009). The result of 1,000 resamples demonstrated
that zero fell outside of the 95% CI of the indirect effect of
perceived poise (95% CI [0.12, 0.99]). Thus, recipients’ perception
of the counterpart’s poise mediated the positive effect of threats
compared to anger.
Perception of intensity. Next, we examined whether percep-
tion of intensity mediated the positive effect of threat over anger in
eliciting concessions. First, an ANCOVA showed that participants
in the threat condition found their counterpart’s reactions to be
marginally less intense (M"6.41, SD "1.44) than did partici-
pants in the anger condition (M"7.10, SD "1.89), F(1, 53) "
2.42, p".13, &
2
".04. Also, when a regression on concessions
was conducted with the anger versus threat factor and perception
of intensity as simultaneous predictors, the effect of the anger
versus threat factor remained significant (p$.04), whereas the
effect of perception of intensity was not significant (p#.15). A
bootstrapping procedure using 1,000 resamples demonstrated that
the indirect effect of perception of intensity was not significant
(95% CI [(0.37, 0.04]). Thus, the effect of threats compared to
anger was not mediated by perception of intensity.
Perception of appropriateness. Finally, we examined
whether perception of appropriateness mediated the positive effect
of threat over anger in eliciting concessions. First, an ANCOVA
showed that participants in the threat condition did not find their
counterpart’s reactions to be significantly less inappropriate than
did participants in the anger condition, F(1, 53) $1.00, p#.32.
Also, when a regression on concessions was conducted with the
anger versus threat factor and perception of appropriateness as
simultaneous predictors, the effect of the anger versus threat factor
remained significant (p".03), whereas the effect of perception of
appropriateness was marginal (p$.06). A bootstrapping proce-
dure using 1,000 resamples demonstrated that the indirect effect of
perception of appropriateness was not significant (95% CI [(0.03,
0.45]). Thus, the effect of threats compared to anger was not
mediated by perception of appropriateness.
Discussion
Experiment 3 extended the main result of Experiment 2 to a
situation where all participants were given an attractive alternative
to a current negotiation. In such a situation, participants again
made more concessions to a counterpart communicating threats
than to a counterpart communicating anger (Hypothesis 2). Also,
again, poise ascribed to the counterpart mediated the positive
effect of communicating threats rather than anger in eliciting
concessions (Hypothesis 3). Neither perceptions of intensity nor
perceptions of appropriateness mediated the positive effect of
threat over anger in eliciting concessions.
General Discussion
Three studies were conducted to disentangle the effects of
communicating anger from communicating threats in negotiations.
Prior research has examined either anger or threats but had not
examined these strategies in a way to separate out their respective
effects on concession making.
Experiment 1 provided evidence that anger communication con-
veys an implied threat. Specifically, Experiment 1 showed that
perceptions of threat mediated the effect of anger (vs. a control) on
concessions (Hypothesis 1). This finding supports the (hitherto
untested) argument made by both negotiation and emotion theo-
rists that anger conveys an implied threat in social interactions
(e.g., Clark et al., 1996; Morris & Keltner, 2000).
Drawing from that, Experiments 2 and 3 made a direct compar-
ison of anger and threat. Both these experiments showed that
issuing threats is a more effective negotiation strategy than com-
municating anger. Across Experiments 2 and 3, (a) participants
made more concessions to a counterpart communicating threats
than to a counterpart communicating anger (Hypothesis 2), and (b)
poise (being confident and in control of one’s own feelings and
decisions) ascribed to the counterpart mediated the positive effect
of communicating threats rather than anger on concessions (Hy-
pothesis 3).
Perceived
poise
β = .50 *** β = .38 ** / β = .31 *
Counterpart communicates Concessions
threats rather than anger by participants
β = .28 * / β = .13 ns
Figure 5. Experiment 3: Mediation of the threat versus anger effect by
perceived poise when participants were given an attractive alternative. All
participants were given an attractive alternative to the negotiation.
!
p$
.05.
!!
p$.01.
!!!
p$.001.
1028 SINACEUR, VAN KLEEF, NEALE, ADAM, AND HAAG
Experiment 2 also showed that the positive effect of communi-
cating threats compared to anger holds above and beyond the
timing of these communications. Specifically, these communica-
tions were more effective later rather than earlier in the negotia-
tion, consistent with prior research (e.g., Benton et al., 1972;
Moore, 2004). Yet, above and beyond that, threats elicited more
concessions than anger. In addition, Experiment 3 extended the
main result of Experiment 2 to a situation where all participants
were given an attractive alternative to the current negotiation.
Thus, Experiment 3 demonstrated that (a) threats are more effec-
tive than anger in eliciting concessions and (b) this relative differ-
ence is mediated by the greater poise conveyed by threats even
when negotiators have an attractive alternative to consider. Finally,
Experiments 1 and 3 indicate that our effects cannot be explained
in terms of differential perceptions of intensity or appropriateness,
thus ruling out important alternative explanations.
Theoretical Implications: Cognitive or Emotional
Communication?
Negotiation research has long been characterized by a cognitive
perspective (Neale & Bazerman, 1991). Even with the dominant
cognitive perspective, scholars were interested in the role of affect
in negotiation, focusing mostly on emotional experience (e.g.,
Carnevale & Isen, 1986; Forgas, 1998; Pillutla & Murnighan,
1996). Recently, researchers have started exploring emotional
communication in negotiation (e.g., Barry, 2008; Graham, Huang,
Clark, & Helgeson, 2008; Kopelman et al., 2006), particularly
anger communication (e.g., Sinaceur & Tiedens, 2006; Steinel et
al., 2008; Van Kleef & Coˆte´, 2007; Van Kleef & De Dreu, 2010;
Van Kleef et al., 2004a, 2004b, 2010). Drawing on the idea that
emotional communication conveys perceptions—specifically, that
recipients ascribe toughness to an angry expresser (Clark et al.,
1996; Sinaceur & Tiedens, 2006; see Tiedens, 2001)—these re-
searchers found that communicating anger generally elicits con-
cessions in negotiation. Thus, they highlighted the effectiveness of
anger communication as a strategy. However, in all prior studies,
the comparison was between anger and relatively softer strategies,
such as communicating positive (i.e., happiness) or no emotion
(i.e., a control condition). That anger elicited more concessions
than softer strategies might be unsurprising because acting tough
generally elicits concessions (cf. Pruitt, 1981; Yukl, 1974b). Thus,
departing from prior work, the current research is the first to assess
the impact of anger by comparing it to another tough, value-
claiming strategy: threats. Specifically, we showed that anger
communication may convey not only self-serving perceptions such
as being tough but also more problematic perceptions such as
lacking poise. Accordingly, the emotional communication of anger
was less effective in eliciting concessions as compared to the
cooler expression of a threat.
In this way, our findings suggest that the effectiveness of anger
communication may lie not in the emotional tone but in the threat
implied. Thus, assessing the effectiveness of emotional communi-
cations (such as anger) against their cognitive equivalents (such as
threat) might be warranted. In line with other research (Galinsky,
Maddux, Gilin, & White, 2008; Sinaceur, 2010), our findings
suggest that it is important to distinguish between emotional and
cognitive components of a strategy or phenomenon in negotiation.
Specifically, we found that an emotional strategy can be less
effective than its cognitive equivalent in negotiation. Future re-
search could investigate how other emotion-based strategies fare
against their cooler equivalents.
Applied Implications: Managing the Relationship in
Negotiation
The current findings, we believe, speak to the importance of
managing the relationship with the counterpart in negotiation. The
mediation by perceived poise suggests that, although threats and
anger are both strategies aimed at making the counterpart yield,
threats might convey comparatively more regard for the recipient
through conveying comparatively less aggressive emotions. Typ-
ically, communicating anger can result in the recipient of that
communication feeling, in turn, angry (Allred, 1999; Friedman et
al., 2004; Hatfield, Cacioppo, & Rapson, 1992; Van Kleef et al.,
2004a).
In contrast, threats may not always be associated with perceived
aggressiveness (Kwon & Weingart, 2004, p. 272; Lytle et al.,
1999, p. 35; Rubin & Brown, 1975, pp. 285–286; Sinaceur &
Neale, 2005; Tedeschi & Bonoma, 1977). Along that line, it is
plausible that threats convey relatively less aggressiveness than
anger. To investigate this speculation, we performed an explor-
atory examination while conducting Experiment 2 and found that
threatening counterparts were viewed as less aggressive, more
agreeable, and less coercive than angry counterparts (measured
through single items; (all ps$.04; we should note, however, that
none of these perceptions mediated the positive effect of threats
over anger in eliciting concessions; only perceptions of poise did).
Hence, threats may not only be more effective but also limit the
negative reactions often associated with communicating anger. In
sum, threats are more effective than anger because of the greater
poise conveyed by threats, and parallel to that, they have the added
benefit of being perceived as less coercive.
This, in turn, speaks to how a negotiator might reduce toughness
on the person while being tough on the issues in negotiation
(Fisher & Ury, 1981; Steinel et al., 2008). Both threats and anger
are perceived to be aggressive acts in negotiation, but threats are
viewed as relatively less aggressive. Communicating about the
likelihood of an impasse or a sanction need not involve negative
emotions directed at the recipient. Ideally, a threat could be pre-
sented as mere information about the possibility of reaching an
impasse (Sinaceur, 2004; Schelling, 1960); for example, it may be
framed as a cautionary notice (Tedeschi, 1970). As Kwon and
Weingart (2004, p. 272) observed, ultimatums can be presented in
a matter-of-fact way, rather than in a hostile way. In general,
intimidation may differ from attack (Lytle et al., 1999, p. 35). In
sum, a strategy need not rely on communicating aggressive emo-
tions to be effective; communicating a colder message along with
the associated perception of a greater sense of confidence and
control may be bludgeon enough.
Limitations
A limitation of our experiments is that they did not involve
face-to-face interaction. The computer-mediated procedure was
ideal for the controlled manipulations of anger and threats and thus
allowed us to provide a rigorous first empirical test of our hypoth-
eses. This experimental control resulted in greater internal validity
1029
ANGER AND THREATS IN NEGOTIATIONS
as the differences between the conditions existed only in how the
counterpart’s reaction was phrased. However, this came at the
expense of mundane realism; for example, our manipulation of
anger involved only verbal (rather than nonverbal) cues. Hence,
caution must be exercised in generalizing from the results. We are
reassured by the consistency of prior findings on anger communi-
cation in negotiation that used the same computerized procedure as
ours (Van Kleef et al., 2004a, 2004b; Van Kleef & Coˆte´, 2007)
with those of face-to-face experiments that used both verbal and
nonverbal cues of anger (Sinaceur & Tiedens, 2006) and field
studies (Van Kleef et al., 2006). Also, the timing effect we found
is consistent with prior negotiation research (e.g., Benton et al.,
1972; Moore, 2004; Sinaceur & Neale, 2005; Yukl, 1974a).
However, we should note that the fact that everything in our
studies was conveyed in writing via computer might have
limited the effectiveness of anger as an emotional strategy
compared to threats. It is possible that in face-to-face negotia-
tions, the emotional tone of anger would be relatively more
powerful than cold threats. As a first attempt to explore this
question, we collected informal, preliminary data in the classroom.
We asked 28 economics undergraduates to role-play a two-party,
face-to-face negotiation exercise similar to that of Sinaceur and
Tiedens (2006, Experiment 2). Specifically, this negotiation exer-
cise was about the terms of a prospective business venture and
included four issues: one distributive issue, two integrative issues,
and one congruent issue (adapted from Sinaceur, 2010). All par-
ticipants were instructed to maximize the points they could reach
(possible range "0 –14,000 points). Participants negotiated in
pairs face to face. Within each pair, one negotiator was advised to
communicate either anger or threats. The instructions for anger
communication were the same as in Sinaceur and Tiedens’s Ex-
periment 2 (2006), and the instructions for threats communication
were adapted from it. Finally, all recipients of these anger or threat
communications were given an attractive alternative to the nego-
tiation (worth 8,000 points). We examined value claiming, which
was the percentage of the total points that one party achieved for
him- or herself. We found that participants advised to use threats
claimed more value (M"51.56%, SD "5.40%) than did partic-
ipants advised to use anger (M"46.38%, SD "3.18%), F(1,
13) "5.25, p".041, &
2
".30. These informal, preliminary data
make sense to the extent that prior research on anger communica-
tion in negotiation usually showed consistent results across proce-
dures. Nonetheless, caution must be exercised in generalizing the
results as the data were informal and the sample was quite small.
Thus, future research should systematically investigate whether the
effect of threats compared to anger generalizes to face-to-face
settings. Furthermore, it is important to note that our finding that
poise (rather than intensity, appropriateness, agreeableness, or
aggressiveness) accounts for the effectiveness of threats relative to
anger in eliciting concessions may also be a function of our
computerized procedure in which participants were exposed to
written rather than verbal communication.
Finally, the current research focused on comparing anger to a
cold threat. A question for future research is whether combining
anger and threat might be even more effective than issuing a cold
threat. On the one hand, it is plausible that anger would add
credibility to a cold threat, such that a threat conveyed with anger
may be even less easily dismissed than a threat without anger. On
the other hand, adding anger might not make a threat more effec-
tive insofar as the effectiveness of anger already lies in conveying
an implied threat. Future research is needed to settle this empirical
question.
Conclusion
To conclude, three experiments disentangled the effects of com-
municating anger from communicating threat in negotiations. Ex-
periment 1 showed that the reason why anger communication (vs.
a control) elicits concessions is that it conveys an implied threat.
Drawing from that, Experiments 2–3 made a direct comparison of
anger and threat and showed that (a) a direct threat was even more
effective than anger in eliciting concessions and (b) this relative
difference between threat and anger was mediated by recipients’
ascribing more poise to a threatening counterpart than to an angry
one. Thus, a cold strategy can be more effective than an emotional
strategy, a conclusion that qualifies prior research on anger com-
munication and has important implications for an understanding of
the dynamics of emotions, relationships, and concessions in nego-
tiations. Calmly issuing a threat matter of factly appears to be more
effective than getting angry. This suggests that getting angry may
be unnecessary.
References
Adair, W. L., Weingart, L., & Brett, J. (2007). The timing and function of
offers in U.S. and Japanese negotiations. Journal of Applied Psychology,
92, 1056 –1068. doi:10.1037/0021-9010.92.4.1056
Adler, R. S., Rosen, B., & Silverstein, E. M. (1998). Emotions in negoti-
ation: How to manage fear and anger. Negotiation Journal, 14, 161–180.
Allred, K. G. (1999). Anger and retaliation: Toward an understanding of
impassioned conflict in organizations. In R. J. Bies, R. J. Lewicki, &
B. H. Sheppard (Eds.), Research on negotiation in organizations (Vol. 7,
pp. 27–58). Greenwich, CT: JAI Press.
Aronoff, J., Woike, B. A., & Hyman, L. M. (1992). Which are the stimuli
in facial displays of anger and happiness? Configurational bases of
emotion recognition. Journal of Personality and Social Psychology, 62,
1050 –1066. doi:10.1037/0022-3514.62.6.1050
Averill, J. R. (1982). Anger and aggression. New York, NY: Springer-
Verlag.
Bacharach, S. B., & Lawler, E. J. (1981). Bargaining: Power, tactics, and
outcomes. San Francisco, CA: Jossey-Bass.
Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable
distinction in social psychological research: Conceptual, strategic, and
statistical considerations. Journal of Personality and Social Psychology,
51, 1173–1182. doi:10.1037/0022-3514.51.6.1173
Barry, B. (1999). The tactical use of emotion in negotiation. Research on
Negotiation in Organizations, 7, 93–121.
Barry, B. (2008). Negotiator affect: The state of the art (and the science).
Group Decision and Negotiation, 17, 97–105.
Benton, A. A., Kelley, H. H., & Liebling, B. (1972). Effects of extremity
of offers and concession rate on the outcomes of bargaining. Journal of
Personality and Social Psychology, 24, 73– 83. doi:10.1037/h0033368
Berkowitz, L. (1974). Some determinants of impulsive aggression: Role of
mediated associations with reinforced aggression. Psychological Re-
view, 81, 165–176. doi:10.1037/h0036094
Berkowitz, L. (1993). Aggression: Its causes, consequences, and control.
New York, NY: McGraw-Hill.
Carnevale, P. J., & Isen, A. (1986). The influence of positive affect and
visual access on the discovery of integrative solutions in bilateral nego-
tiations. Organizational Behavior and Human Decision Processes, 37,
1–13. doi:10.1016/0749-5978(86)90041-5
1030 SINACEUR, VAN KLEEF, NEALE, ADAM, AND HAAG
Clark, M. S., Pataki, S. P., & Carver, V. H. (1996). Some thoughts and
findings on self-presentation of emotions in relationships. In G. J. O.
Fletcher & J. Fitness (Eds.), Knowledge structures in close relation-
ships: A social psychological approach (pp. 247–274). Mahwah, NJ:
Erlbaum.
Dahlen, E. R., Martin, R. C., Ragan, K., & Kuhlman, M. M. (2004).
Boredom proneness in anger and aggression: Effects of impulsiveness
and sensation seeking. Personality and Individual Differences, 37,
1615–1627. doi:10.1016/j.paid.2004.02.016
Daly, J. P. (1991). The effects of anger on negotiations over mergers and
acquisitions. Negotiation Journal, 7, 31–39.
De Dreu, C. K. W. (1995). Coercive power and concession making in
bilateral negotiation. Journal of Conflict Resolution, 39, 646 – 670. doi:
10.1177/0022002795039004003
De Dreu, C. K. W., & Van Lange, P. A. M. (1995). The impact of social
value orientations on negotiator cognition and behavior. Personality and
Social Psychology Bulletin, 21, 1178 –1188. doi:10.1177/
01461672952111006
De Gaulle, C. (1964). The complete war memoirs of Charles de Gaulle.
New York, NY: Simon & Schuster. (Original work published 1954)
Dewey, J. (1895). The theory of emotion (II): The significance of emo-
tions. Psychological Review, 2, 13–32. doi:10.1037/h0070927
Elfenbein, H. A. (2007). Emotion in organizations: A review in stages. In
A. Brief & J. Walsh (Eds.), Academy of Management annals (Vol. 1., pp.
371– 457). Amsterdam, the Netherlands: Elsevier.
Faley, T., & Tedeschi, J. T. (1971). Status and reactions to threats. Journal
of Personality and Social Psychology, 17, 192–199. doi:10.1037/
h0030463
Fisher, R., & Ury, W. L. (1981). Getting to yes: Negotiating agreement
without giving in. Boston, MA: Houghton Mifflin.
Forgas, J. P. (1998). On feeling good and getting your way: Mood effects
on negotiator cognition and bargaining strategies. Journal of Personality
and Social Psychology, 74, 565–577. doi:10.1037/0022-3514.74.3.565
Forgas, J. P., & George, J. M. (2001). Affective influences on judgment,
decision making, and behavior in organizations: An information pro-
cessing perspective. Organizational Behavior and Human Decision Pro-
cesses, 86, 3–34. doi:10.1006/obhd.2001.2971
Frank, R. (1988). Passions within reason: The strategic role of the emo-
tions. New York, NY: Norton.
Frantz, C. M., & Bennigson, C. (2005). Better late than early: The influ-
ence of timing on apology effectiveness. Journal of Experimental Social
Psychology, 41, 201–207. doi:10.1016/j.jesp.2004.07.007
Friedland, N. (1976). Social influence via threats. Journal of Experimental
Social Psychology, 12, 552–563. doi:10.1016/0022-1031(76)90034-2
Friedman, R., Anderson, C., Brett, J., Olekalns, M., Goates, N., & Lisco,
C. C. (2004). The positive and negative effects of anger on dispute
resolution: Evidence from electronically mediated disputes. Journal of
Applied Psychology, 89, 369 –376. doi:10.1037/0021-9010.89.2.369
Galinsky, A. D., & Liljenquist, K. A. (2004). Putting on the pressure: How
to make threats in negotiations. Negotiation, 12, 1–5.
Galinsky, A. D., Maddux, W. W., Gilin, D., & White, J. B. (2008). Why
it pays to get inside the head of your opponent: The differential effects
of perspective-taking and empathy in negotiations. Psychological Sci-
ence, 19, 378 –384. doi:10.1111/j.1467-9280.2008.02096.x
Galinsky, A. D., & Mussweiler, T. (2001). First offers as anchors: The role
of perspective-taking and negotiator focus. Journal of Personality and
Social Psychology, 81, 657– 669. doi:10.1037/0022-3514.81.4.657
Goos, L. M., & Silverman, I. (2002). Sex related factors in the perception
of threatening facial expressions. Journal of Nonverbal Behavior, 26,
27– 41.
Graham, S. M., Huang, J. Y., Clark, M. S., & Helgeson, V. S. (2008). The
positives of negative emotions: Willingness to express negative emo-
tions promotes relationships. Personality and Social Psychology Bulle-
tin, 34, 394 – 406. doi:10.1177/0146167207311281
Hatfield, E., Cacioppo, J. T., & Rapson, R. L. (1992). Primitive emotional
contagion. In M. S. Clark (Ed.), Emotion and social behavior (pp.
151–177). Thousand Oaks, CA: Sage.
Hebb, D. O. (1949). The organization of behavior: A neuropsychological
theory. New York, NY: Wiley.
Jackall, R. (1988). Moral mazes: The world of corporate managers. Ox-
ford, England: Oxford University Press.
Kimble, G. A., & Perlmuter, L. C. (1970). The problem of volition.
Psychological Review, 77, 361–384. doi:10.1037/h0029782
Knutson, B. (1996). Facial expressions of emotion influence interpersonal
trait inferences. Journal of Nonverbal Behavior, 20, 165–182. doi:
10.1007/BF02281954
Komorita, S. S., & Barnes, M. (1969). Effects of pressures to reach
agreement in bargaining. Journal of Personality and Social Psychology,
13, 245–252. doi:10.1037/h0028278
Kopelman, S., Rosette, A. S., & Thompson, L. (2006). The three faces of
Eve: Strategic displays of positive, negative, and neutral emotions in
negotiations. Organizational Behavior and Human Decision Processes,
99, 81–101.
Kwon, S., & Weingart, L. R. (2004). Unilateral concessions from the other
party: Concession behavior, attributions, and negotiation judgments.
Journal of Applied Psychology, 89, 263–278. doi:10.1037/0021-
9010.89.2.263
Lerner, J. S., & Tiedens, L. Z. (2006). Portrait of the angry decision maker:
How appraisal tendencies shape anger’s influence on cognition. Journal
of Behavioral Decision Making, 19, 115–137.
Lytle, A. L., Brett, J. M., & Shapiro, D. L. (1999). The strategic use of
interests, rights, and power to resolve disputes, Negotiation Journal, 15,
31–51. doi:10.1111/j.1571-9979.1999.tb00178.x
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, 200 –212. doi:10.1177/
0146167206294413
Mannix, E. A., Thompson, L. L., & Bazerman, M. H. (1989). Negotiation
in small groups. Journal of Applied Psychology, 74, 508 –517. doi:
10.1037/0021-9010.74.3.508
Marsh, A. A., Ambady, N., & Kleck, R. E. (2005). The effects of fear and
anger facial expressions on approach- and avoidance-related behaviors.
Emotion, 5, 119 –124.
Michener, H. A., & Cohen, E. D. (1973). Effects of punishment magnitude
in the bilateral threat situation: Evidence for the deterrence hypothesis.
Journal of Personality and Social Psychology, 26, 427– 438.
Moore, D. A. (2004). Myopic prediction, self-destructive secrecy, and the
unexpected benefits of revealing final deadlines in negotiation. Organi-
zational Behavior and Human Decision Processes, 94, 125–139. doi:
10.1016/j.obhdp.2004.04.001
Morris, M. W., & Keltner, D. (2000). How emotions work: An analysis of
the social functions of emotional expression in negotiations. Research in
Organizational Behavior, 22, 1–50.
Neale, M. A., & Bazerman, M. H. (1991). Cognition and rationality in
negotiation. New York, NY: Free Press.
Pillutla, M. M., & Murnighan, J. K. (1996). Unfairness, anger, and spite:
Emotional rejections of ultimatum offers. Organizational Behavior and
Human Decision Processes, 68, 208 –224. doi:10.1006/obhd.1996.0100
Pinkley, R. L., Neale, M. A., & Bennett, R. J. (1994). The impact of
alternatives to settlement in dyadic negotiation. Organizational Behav-
ior and Human Decision Processes, 57, 97–116. doi:10.1006/
obhd.1994.1006
Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resampling
strategies for assessing and comparing indirect effects in multiple me-
diator models. Behavior Research Methods, 40, 879 – 891. doi:10.3758/
BRM.40.3.879
Preacher, K. J., & Hayes, A. F. (2009). SPSS and SAS macro for bootstrap-
ping specific indirect effects in multiple mediation models. Retrieved
1031
ANGER AND THREATS IN NEGOTIATIONS
from http://www.comm.ohio-state.edu/ahayes/SPSS%20programs/
indirect.htm
Pruitt, D. G. (1981). Negotiation behavior. New York, NY: Academic
Press.
Pruitt, D. G., & Drews, J. L. (1969). The effect of time pressure, time
elapsed, and the opponent’s concession rate on behavior in negotiation.
Journal of Experimental Social Psychology, 5, 43– 60. doi:10.1016/
0022-1031(69)90005-5
Ramı´rez, J. M., & Andreu, J. M. (2006). Aggression, and some related
psychological constructs (anger, hostility, and impulsivity): Some com-
ments from a research project. Neuroscience and Biobehavioral Re-
views, 30, 276 –291. doi:10.1016/j.neubiorev.2005.04.015
Rubin, J. Z., & Brown, B. R. (1975). The social psychology of bargaining
and negotiation. New York, NY: Academic Press.
Schelling, T. C. (1960). The strategy of conflict. Cambridge, MA: Harvard
University Press.
Schlesinger, A. M. (1965). A thousand days. Boston, MA: Houghton
Mifflin.
Shapiro, D. L., & Bies, R. J. (1994). Threats, bluffs, and disclaimers in
negotiations. Organizational Behavior and Human Decision Processes,
60, 14 –35. doi:10.1006/obhd.1994.1073
Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and non-
experimental studies: New procedures and recommendations. Psycho-
logical Methods, 7, 422– 445. doi:10.1037/1082-989X.7.4.422
Sinaceur, M. (2004). Utilisation de la menace en ne´ gociation [The use of
threats in negotiations]. Revue Francaise de Gestion, 30(153),101–121.
doi:10.3166/rfg.153.101-121
Sinaceur, M. (2010). Suspending judgment to create value: Suspicion and
trust in negotiation. Journal of Experimental Social Psychology, 46,
543–550. doi:10.1016/j.jesp.2009.11.002
Sinaceur, M., & Neale, M. A. (2005). Not all threats are created equal:
How implicitness and timing affect the effectiveness of threats in nego-
tiations. Group Decision and Negotiation, 14, 63– 85. doi:10.1007/
s10726-005-3876-5
Sinaceur, M., & Tiedens, L. Z. (2006). Get mad and get more than even:
When and why anger expression is effective in negotiations. Journal of
Experimental Social Psychology, 42, 314 –322. doi:10.1016/j
.jesp.2005.05.002
Soanes, C., & Stevenson, A. (Eds.). (2005). The Oxford dictionary of
English (rev. ed.). Oxford, England: Oxford University Press.
Steinel, W., Van Kleef, G. A., & Harinck, F. (2008). Are you talking to
me?! Separating the people from the problem when expressing emotions
in negotiation. Journal of Experimental Social Psychology, 44, 362–369.
doi:10.1016/j.jesp.2006.12.002
Tedeschi, J. T. (1970). Threats and promises. In P. G. Swingle (Ed.), The
structure of conflict (pp. 155–191). New York, NY: Academic Press.
Tedeschi, J. T., & Bonoma, T. V. (1977). Measures of last resort, coercion
and aggression in bargaining. In D. Druckman (Ed.), Negotiations:
Social-psychological perspectives (pp. 213–241). Beverly Hills, CA:
Sage.
Thompson, L. L., Nadler, J., & Kim, P. H. (1999). Some like it hot: The
case for the emotional negotiator. In L. L. Thompson, J. M. Levine, &
D. M. Messick (Eds.), Shared cognition in organizations: The manage-
ment of knowledge (pp. 139 –161). Mahwah, NJ: Erlbaum.
Tiedens, L. Z. (2001). Anger and advancement versus sadness and subju-
gation: The effect of negative emotion expressions on social status
conferral. Journal of Personality and Social Psychology, 80, 86 –94.
doi:10.1037/0022-3514.80.1.86
Underwood, M. K., Hurley, J. C., Johanson, C. A., & Mosley, J. (1999). An
experimental, observational investigation of children’s responses to peer
provocation: Developmental and gender differences in middle child-
hood. Child Development, 70, 1428 –1446. doi:10.1111/1467-
8624.00104
Van Dijk, E., Van Kleef, G. A., Steinel, W., & Van Beest, I. (2008). A
social functional approach to emotions in bargaining: When communi-
cating anger pays and when it backfires. Journal of Personality and
Social Psychology, 94, 600 – 614. doi:10.1037/0022-3514.94.4.600
Van Kleef, G. A., & Coˆ te´ , S. (2007). Expressing anger in conflict: When
it helps and when it hurts. Journal of Applied Psychology, 92, 1557–
1569. doi:10.1037/0021-9010.92.6.1557
Van Kleef, G. A., & De Dreu, C. K. W. (2010). Longer-term consequences
of anger expression in negotiation: Retaliation or spillover? Journal of
Experimental Social Psychology, 46, 753–760. doi:10.1016/j
.jesp.2010.03.013
Van Kleef, G. A., De Dreu, C. K. W., & Manstead, A. S. R. (2004a). The
interpersonal effects of anger and happiness in negotiations. Journal of
Personality and Social Psychology, 86, 57–76. doi:10.1037/0022-
3514.86.1.57
Van Kleef, G. A., De Dreu, C. K. W., & Manstead, A. S. R. (2004b). The
interpersonal effects of emotions in negotiations: A motivated informa-
tion processing approach. Journal of Personality and Social Psychology,
87, 510 –528. doi:10.1037/0022-3514.87.4.510
Van Kleef, G. A., De Dreu, C. K. W., & Manstead, A. S. R. (2010). An
interpersonal approach to emotion in social decision making: The emo-
tions as social information model. Advances in Experimental Social
Psychology, 42, 45–96. doi:10.1016/S0065-2601(10)42002-X
Van Kleef, G. A., De Dreu, C. K. W., Pietroni, D., & Manstead, A. S. R.
(2006). Power and emotion in negotiation: Power moderates the inter-
personal effects of anger and happiness on concession making. Euro-
pean Journal of Social Psychology, 36, 557–581. doi:10.1002/ejsp.320
Van Kleef, G. A., Van Dijk, E., Steinel, W., Harinck, F., & Van Beest, I.
(2008). Anger in social conflict: Cross-situational comparisons and
suggestions for the future. Group Decision and Negotiation, 17, 13–30.
doi:10.1007/s10726-007-9092-8
White, S. B., Valley, K., Bazerman, M. H., Neale, M. A., & Peck, S.
(1994). Alternative models of price behavior in dyadic negotiations:
Market prices, reservation prices, and negotiator aspirations. Organiza-
tional Behavior and Human Decision Processes, 57, 430 – 447. doi:
10.1006/obhd.1994.1023
Wolfe, R. J., & McGinn, K. L. (2005). Perceived relative power and its
influence on negotiations. Group Decision and Negotiation, 14, 3–20.
doi:10.1007/s10726-005-3873-8
Yukl, G. A. (1974a). Effects of situational variables and opponent conces-
sions on a bargainer’s perception, aspirations, and concessions. Journal
of Personality and Social Psychology, 29, 227–236. doi:10.1037/
h0036013
Yukl, G. A. (1974b). Effects of the opponent’s initial offer, concession
magnitude, and concession frequency on bargaining behavior. Journal of
Personality and Social Psychology, 30, 323–335. doi:10.1037/h0036895
Received March 9, 2009
Revision received March 18, 2011
Accepted March 21, 2011 !
1032 SINACEUR, VAN KLEEF, NEALE, ADAM, AND HAAG
