ORGANIZATIONALBEHAVIOR

Organizational Behavior and Human Decision Processes 95 (2004) 125–139

AND HUMANDECISION PROCESSES

www.elsevier.com/locate/obhdp

Affect from the top down: How powerful individuals� positiveaffect shapes negotiations

Cameron Andersona,* and Leigh L. Thompsonb

a Department of Management and Organizations, Stern School of Business, New York University,

44 West 4th Street, KMC 7-59, New York, NY 10012, USAb Northwestern University, USA

Received 22 October 2003

Available online 24 June 2004

Abstract

We tested the hypothesis that the positive affect of powerful negotiators shapes the quality of negotiation processes and outcomes

more than the positive affect of less powerful negotiators. Findings from two studies supported the hypothesis: powerful individuals�trait positive affect was the best predictor of negotiators� trust for each other and of whether they reached integrative outcomes.

Positive affect predicted joint gains above and beyond negotiators� trait cooperativeness and communicativeness. However, positive

affect was unrelated to distributive outcomes; thus, there were no observed disadvantages of being positively affective.

� 2004 Elsevier Inc. All rights reserved.

Keywords: Power; Affect; Emotion; Mood; Negotiation; Trust; Individual differences; Personality traits

Introduction

The use of power and the experience of affect are

among the most fundamental aspects of social interac-

tions. It is not surprising, therefore, that the study of

negotiation behavior has begun to examine the effects of

power on negotiated outcomes (e.g., Lawler & Yoon,

1993; Mannix, 1993a, 1993b; Pinkley, Neale, & Bennett,

1994) and how affect shapes and is shaped by negotia-

tions (for reviews, see Barry, Fulmer, & Van Kleef, inpress; Morris & Keltner, 2000). What is surprising,

perhaps, is that research on power and affect in negoti-

ation have proceeded independently of one another,

when in fact, they are often intimately related in social

interactions (Anderson & Berdahl, 2002; Tiedens, 2001).

In this paper, we tested the hypothesis that in power-

asymmetric negotiations, the positive affect of the

powerful negotiator shapes the process and outcomemore than the positive affect of the less powerful nego-

tiator. This hypothesis stems from recent analyses of the

effects of power in face-to-face interactions (Keltner,

* Corresponding author. Fax: 1-212-995-4235.

E-mail address: c.anderson@stern.nyu.edu (C. Anderson).

0749-5978/$ - see front matter � 2004 Elsevier Inc. All rights reserved.

doi:10.1016/j.obhdp.2004.05.002

Gruenfeld, & Anderson, 2003; Kramer, 1996), and from

research on the benefits of positive emotion in negotia-tions (e.g., Carnevale & Isen, 1986; Forgas, 1998).

Defining power

Along with many theorists, we define power as the

capacity to influence others (French & Raven, 1959;

Goldhamer & Shils, 1939; Keltner et al., 2003; Lewin,

1951). Thus, power is a relational variable, in that in-dividuals� power can be understood only in relation to

another person or a group (Emerson, 1962; Thibaut &

Kelley, 1959). Power vis-�a-vis others can derive from a

number of sources. For example, people might have

‘‘legitimate power,’’ in that they occupy a position of

authority (French & Raven, 1959; Pfeffer, 1992), ‘‘re-

ward power,’’ in that others depend them for valued

resources (Emerson, 1962; Lawler & Bacharach, 1987),or ‘‘coercive power,’’ in that they have the ability to

inflict harm on others (Bacharach & Lawler, 1980; Brett,

2001; French & Raven, 1959).

Likewise, in the specific context of negotiations,

power vis-�a-vis a counterpart can derive from a number

of sources. Negotiators might occupy a position of

2 We focused on positive affect, and not negative affect, for two

reasons. First, our primary focus is how parties of unequal power can

achieve integrative agreements, and theorists and researchers have

126 C. Anderson, L.L. Thompson / Organizational Behavior and Human Decision Processes 95 (2004) 125–139

authority relative to their counterpart (as when anacademic dean negotiates with a job candidate), their

counterpart might be dependent on them for resources

(as when the only potential buyer of a house negotiates

with a desperate seller), they might have the ability to

inflict harm on their counterpart (e.g., the ability to sue

for fraud), or their power might stem from a combina-

tion of these sources.

The role of power in negotiations

The study of power in negotiation has centered on

two key issues. An enduring question focuses on power

at the individual level, addressing how negotiators�power affects the outcomes they attain (e.g., Pinkley

et al., 1994). Another question focuses on power at the

dyad or group level, addressing how the relative powerdifference between parties affects both the process and

outcomes of negotiation (e.g., Lawler & Yoon, 1993;

Mannix & Neale, 1993). Our focus is the latter question.

Differences in power can drive a wedge between nego-

tiating parties, making it more difficult for them to reach

integrative or ‘‘win-win’’ agreements (e.g., Mannix,

1993a; Pinkley et al., 1994). When parties of unequal

power negotiate, the question arises as to whether thenegotiated outcome should reflect the difference in

power; negotiators with lower power tend to resist

agreements that reflect the power differences, and ne-

gotiators with higher power push for agreements that

distribute payoffs proportional to their power (Lawler &

Yoon, 1993; Mannix, 1993a). As a result, both parties

focus more on the distributive elements of the negotia-

tion and less on its integrative potential (Faley &Tedeschi, 1971; Mannix, 1993a). They use more com-

petitive and even coercive bargaining tactics (Lawler,

1992; Lawler & Bacharach, 1987; Lawler, Ford, & Ble-

gen, 1988) and ultimately, reach less integrative out-

comes. This has been documented in dyadic negotiations

(Lawler & Yoon, 1993; McAlister, Bazerman, & Fader,

1986; Pinkley et al., 1994), multi-party bargaining con-

texts (Mannix, 1993a, 1993b), and prisoner�s dilemmagames (Rekosh & Feigenbaun, 1966).1

Positive affect and integrative outcomes

How can negotiators of unequal power reach better

integrative outcomes? Research suggests that affective

processes might be highly beneficial in facilitating de-

velopment of integrative bargaining. The term ‘‘affect’’represents a broad category of affective processes, in-

cluding emotional experiences, moods, and trait or dis-

1 A smaller number of studies have found that parties with equal

power achieve lower joint gains than parties of unequal power. Mannix

(1993b) hypothesized that in some of these studies, higher joint gains

was produced by less powerful parties who had high aspirations.

positional affect (Thompson, 1998). Emotions are briefstates that involve cognitive, physiological, and behav-

ioral processes that help individuals quickly respond to

threats or opportunities; they are relatively short in

duration and are directed at specific events or stimuli

(Ekman, 1973; Frijda, 1994; Keltner & Haidt, 2001).

Moods, in contrast, have a longer duration, lasting

hours or days, and are less directly focused on anything

specific (Ekman, 1994; Frijda, 1994). Trait or disposi-tional affect reflects stable individual differences in the

tendency to experience and express certain emotions and

moods (Watson & Clark, 1984; Watson, Clark, &

Tellegen, 1988).

Researchers have found the experience and expres-

sion of positive-affective states such as excitement, en-

thusiasm, and happiness help stimulate the integrative

process (Allred, Mallozzi, Matsui, & Raia, 1997; Baron,1990; Carnevale & Isen, 1986; Forgas, 1998; Kramer,

Newton, & Pommerenke, 1993; Moore, Kurtzberg,

Thompson, & Morris, 1999). In a seminal study, Car-

nevale and Isen (1986) induced positive emotion in some

participants by having them read humorous cartoons

and giving them a small gift. Although the manipulation

of emotion was somewhat mild, they found that par-

ticipants in the positive-emotion condition were morelikely to communicate their priorities, perceive each

other�s interests accurately, and achieve high joint gains

than control participants. Subsequent studies that have

induced positive emotion through the use of pleasant

scents (Baron, 1990), humorous videos (Kramer et al.,

1993), or false performance feedback (Forgas, 1998)

have found similar results; positively emotional negoti-

ators have consistently bargained more integratively andcreated more value than non-emotional negotiators.2

There are a number of likely reasons why positive

affect facilitates integrative bargaining; some are ‘‘in-

trapersonal’’ in nature, in that they concern the way

emotion influences the person experiencing them, and

some are ‘‘interpersonal’’ in nature, in that they concern

the way emotional expression influences others (Van

Kleef, De Dreu, & Manstead, in press). In terms of in-trapersonal mechanisms, the experience of positive

emotion evokes a more pro-social and cooperative ori-

entation towards others (Baron, 1990; Forgas, 1998;

Isen & Levin, 1972; Levin & Isen, 1975; Rosenhan,

Salovey, & Hargis, 1981). Thus, when negotiators feel

positive emotion, they should be drawn away from a

drawn stronger links between positive emotion and integrative agree-

ments. Second, negative emotion is generally viewed as more multifac-

eted than positive emotion (e.g., Ekman, 1973; Lerner &Keltner, 2001),

which suggests the relation between negative emotion and integrative

agreements is quite complex, and that a full investigation of the effects of

negative emotions is beyond the ken of this paper.

C. Anderson, L.L. Thompson / Organizational Behavior and Human Decision Processes 95 (2004) 125–139 127

solely self-interested, competitive stance in which theyare looking only to maximize their own outcomes, and

drawn toward a cooperative mind-set in which they seek

to maximize others� outcomes as well as their own.

Further, the experience of positive emotion increases

creativity and innovative thinking; people who feel even

mild levels of positive emotion can think about infor-

mation more flexibly, and can generate more innovative

ideas than non-emotional people (e.g., Ashby, Isen, &Turken, 1999; Carnevale & Isen, 1986; Isen, Daubman,

& Nowicki, 1987). Thus, positively affective negotiators

should be more effective integrative bargainers, as their

increased creativity should help them think of new issues

to add to the table and of novel ways they can trade

issues and resolve their differences (Bazerman & Neale,

1992; Carnevale & Isen, 1986; Fisher & Ury, 1991;

Thompson, 1990).In terms of interpersonal mechanisms, the display of

positive emotion communicates one�s trustworthiness

and cooperativeness to others (Frank, 1988; Fridlund,

1994; Knutson, 1996). When negotiators display posi-

tive affect, they signal that they are trustworthy and are

ready to cooperate—a signal that should lead other

parties to trust them more. The increased trust should

then trigger communication of interests and priorities(Carnevale & Isen, 1986), and to the discovery of com-

patible and tradable issues, giving negotiators more

opportunities to expand the resource pie (Bazerman &

Neale, 1992; Fisher & Ury, 1991; Thompson, 1990).

Positive affect from the top down

In power-imbalanced negotiations, we propose thatpowerful negotiators� positive affect will be more im-

portant to reaching integrative agreements.3 This hy-

pothesis derives from research on the emergence of trust

in power-asymmetric relationships, and on the role of

power in face-to-face interactions.

Trust in hierarchical relationships

Theorists examining interpersonal trust in hierarchi-cal relationships have argued that people with less

power occupy a vulnerable position, in that they are

susceptible to being exploited by others (Chen, Brock-

ner, & Greenberg, 2004; Kramer, 1996). Therefore, less

3 The hypothesis that powerful parties� positive emotion is more

important to reaching integrative deals, at first glance, might seem to

contradict a hypothesis proposed in previous studies—namely, that the

low-power party tends to drive integrative gains through their effort

and hard work (Kim, 1997; Mannix & Neale, 1993). However, this is

only an apparent contradiction. We suggest one factor that helps

parties of unequal power find integrative outcomes is powerful parties�positive emotion; we are not proposing that this is the only route to

integrative gains. Thus, while powerful parties� positive emotion is

likely to help, it is also possible that low-power parties� higher effort

also facilitates more integrative settlements (Mannix & Neale, 1993).

powerful individuals tend to be more anxious and con-cerned about issues of trust than people with power

(Mannix, 1993a; Tyler & Degoey, 1996; van den Bos,

Wilke, & Lind, 1998); they are more vigilant for indi-

cators of others� trustworthiness, and frequently look

for signals that they can trust others and that they will

not be exploited (e.g., Kramer, 1996).

In the context of power-imbalanced negotiations, this

suggests that negotiators with less power will be highlyresponsive to their counterpart�s positive affect. As we

have argued above, positive emotion signals one�strustworthiness and cooperativeness to others; when

people exhibit positive emotion, they are in effect saying

to others, ‘‘you can trust me.’’ Negotiators with less

power, therefore, should be highly sensitive to this trust-

building signal, and more likely to respond to it. Once

they begin trusting their more powerful counterpart,they should feel more comfortable communicating their

interests and priorities. In contrast, powerful negotiators

should be less responsive to their counterpart�s positiveaffect, because they would be less vigilant for signs of

others� trustworthiness. Their lower level of concern

about trust-related issues means they would not be

looking for those signs in the first place. Therefore, this

suggests powerful negotiators� positive affect will be abetter predictor of the trust that emerges between ne-

gotiating parties and a better predictor of the negotia-

tions� integrativeness.

Power in face-to-face interactions

Recent research has also shown that when individuals

have power in a face-to-face interaction, they tend to

influence others� thoughts, feelings, and behaviors muchmore than vice versa. For example, in a recent study of

decision-making in dyads (Anderson & Berdahl, 2002),

individuals randomly assigned to a more powerful po-

sition had a stronger influence over subordinates� atti-tudes during a 15-min interaction than vice versa.

Similarly, a study of non-verbal mimicry (Cheng &

Chartrand, 2003) found that individuals in a low-power

position unconsciously mimic the non-verbal behaviorsof those with power. Even in terms of emotional expe-

riences, less powerful individuals tend to ‘‘catch’’ the

emotional experiences of powerful individuals, whereas

less powerful individuals� emotions have little to no ef-

fect on powerful individuals (Anderson, Keltner, &

John, 2003).

Powerful individuals� impact on others in face-to-face

interactions makes sense. Individuals tend to be moreexpressive when they have power, which gives others

more opportunities to view their emotion (Anderson &

Berdahl, 2002; Hecht & LaFrance, 1998; Keltner et al.,

2003), and less powerful individuals are highly attentive

to those with power (Chance, 1967; Fiske, 1993; Keltner

& Robinson, 1997). In power-imbalanced negotiations,

this suggests powerful individuals� positive affect should

128 C. Anderson, L.L. Thompson / Organizational Behavior and Human Decision Processes 95 (2004) 125–139

have a stronger impact on the tone of the negotiation asa whole. When powerful negotiators are more positively

affective, they would invoke similar positive states in

others, which would lead to a more pro-social orienta-

tion among all negotiators, increased communication,

and better creative thinking. Collectively, the heightened

trust, motivation to cooperate, and better integrative

thinking should benefit their integrative outcomes.

The current research

Our main hypothesis in the current research was that

the positive affect of powerful negotiators will predict

integrative processes and outcomes more strongly than

the positive affect of less powerful negotiators. To test

this hypothesis, we focused on trait positive affect, or the

tendency to experience and express positive emotionssuch as excitement, enthusiasm, and interest (Watson &

Clark, 1997). As described above, previous research on

positive emotion in negotiation has experimentally ma-

nipulated positive emotion using pleasant scents or by

presenting individuals with humorous cartoons. Such

experimental methodologies are extremely valuable, as

they allow for greater environmental control and can

demonstrate the causal effects of positive affect on de-pendent variables of interest. However, in real-world

negotiations, it is likely that individual differences in

trait positive affect play a prominent role in influencing

individuals� behavior (Barry et al., in press; Watson &

Clark, 1997). As such, negotiation researchers have ar-

gued that an adequate psychological understanding of

negotiation needs to consider the role of personality and

individual differences in addition to the contextual de-terminants of negotiator behavior (Barry et al., in press;

Forgas, 1998; Thompson, 1990).

As a secondary aim, we also examined the influence

of positive affect on distributive outcomes. If powerful

individuals who are more positively affective achieve

larger integrative gains, it is important to understand if

their positive affect shapes how they allocate resources in

the relationship. One hypothesis is that positively af-fective people will not fare well in terms of their allo-

cation. Previous research has shown a link between

positive emotional experience and selflessness or altru-

ism (Isen & Levin, 1972; Rosenhan et al., 1981).

Therefore, positively affective people might make more

concessions than others, or be so trusting of others that

they communicate too much of their information, al-

lowing others to take advantage of them. A secondhypothesis is that positively affective people achieve

higher distributive gains. They might approach negoti-

ations with more energy and enthusiasm, thus putting

more effort into both increasing the size of the pie and

into gaining a big slice for themselves. Still another

hypothesis is that positive affect does not relate to dis-

tributive gains at all; whereas positive affect helps

increase trust and enhances communication betweenparties, it does not affect how they divide their increased

resources. Negotiation research has provided mixed re-

sults on this issue. Positive affect has been shown to

increase one�s own concession making (Baron, 1990)

and make opponents concede less (Van Kleef et al., in

press), which suggests positive affect might negatively

predict distributive outcomes. However, other work has

found no relation (Allred et al., 1997), or a positive re-lation (Forgas, 1998) between positive affect and dis-

tributive gains. Thus, we examine this issue as an open

research question.

Study 1

Methods

Participants

One hundred twenty masters of business administra-

tion (MBA) students enrolled in a course on negotia-

tions participated. Of these, 92 were in dyads in which

both members completed all measures (64 men, 28 wo-

men); thus, our analyses focused on those 92 partici-

pants. The study was conducted in the first week of thecourse. Participants were randomly paired together and

placed into roles using a random assignment software

program.

Positive affect

To measure participants� trait positive affect, we usedthe positive affect scale of the Positive and Negative

Affect Schedule, or PANAS (Watson et al., 1988), whichparticipants completed on-line in the week prior to the

beginning of the course. The PANAS scales are among

the most widely used affect measures, and have been

used in research on leadership (Watson & Clark, 1997),

small groups (Barsade, 2002), procedural justice (van

den Bos & Miedema, 2000), and goal setting (Sheldon &

Elliot, 1999) to name a few domains. The PANAS scales

can measure affect at multiple levels, including theemotion, mood, or trait levels; we focused on the trait

level. The scales have been shown to be internally con-

sistent, with as ranging in the .80s to .90s, and to be

consistent over time (Watson et al., 1988; Watson &

Walker, 1996). Further, the two trait-level affect scales

are largely uncorrelated with each other, supporting the

notion that positive and negative affect are two inde-

pendent or orthogonal dimensions (Watson et al., 1988).Participants reported how much they felt 10 emotions

‘‘in general, or on average’’: interested, excited, enthu-

siastic, proud, alert, strong, inspired, determined, at-

tentive, and active, on a scale from 1 (‘‘None’’) to 5

(‘‘Extreme’’). The scale showed satisfactory internal

consistency, coefficient a¼ .76; the mean score was 3.69

(SD ¼ :42). There was not a significant sex difference;

C. Anderson, L.L. Thompson / Organizational Behavior and Human Decision Processes 95 (2004) 125–139 129

the mean score for men was 3.68 (SD ¼ :42), and forwomen was 3.70 (SD ¼ :43).

Procedure and role instructions

The study was conducted as a class exercise. Each

negotiating dyad comprised an owner-operator of a

service station that would like to sell the station, and a

representative of a large oil company (the Vice President

for Operations of ‘‘Texoil’’) that would like to buy it.Participants were given half an hour to read their role

instructions and prepare for the negotiation. Dyads were

given 30min to negotiate. They negotiated in private

rooms, out of earshot of other groups.

Station owners were told that it is imperative that

they complete this sale promptly, because they and their

spouse have been working to exhaustion and thus are

desperate to take a two-year vacation sailing around theworld. They were told the only other offer they have for

the station is not adequate for them to afford their va-

cation, and therefore, that the upcoming meeting with

the Texoil representative is highly important. Texoil

representatives were told their company is in the midst

of buying independent service stations and turning them

into mini service marts, and that they are interested in

buying this station for the right price. However, theircompany can build a new station if they do not reach a

deal. They were also told that it is a shame that these

owner-operators are leaving, because they have been an

excellent management team and Texoil needs good

managers as it expands.

We construed the Texoil representatives as more

powerful than the station owners. They had reward

power, as the station owners were more dependent onthem to reach a deal than vice versa, and legitimate

power, as the Texoil representative is the Vice President

of Operations for a large petroleum refining company,

and thus occupied a position of relative authority. To

confirm that the Texoil representatives had more power,

we asked 50 independent judges who read the case ma-

terials ‘‘Who has more power?’’ Of these 50 judges, 80%

(40 total) perceived the Texoil representative as morepowerful (v2 ð1;N ¼ 50Þ ¼ 18:00, p < :01).

When participants begin negotiating this exercise,

there seems to be no viable bargaining range between

the two parties� reservation prices: Texoil representatives

are authorized to pay up to $500,000, yet station owners

need at least $580,000 for the station. Specifically, sta-

tion owners need $413,000 to pay off their boat and to

cover their expenses for two years, $92,000 to pay thecapital gains taxes from selling the station, and $75,000

for ‘‘safety money,’’ that they will need when they return

from vacation and have to look for a new job. The

parties can overcome this problem, however, if station

owners communicate their interests with the Texoil

representative—namely, that they want to take a two-

year vacation, and that they will need a job upon return.

By doing so, the Texoil representative, as Vice Presidentof Operations, can add other issues to the agreement,

such as offering a job upon return from their vacation or

offering to pay for gas and boat maintenance during the

two years. The promise of a job offer, for example,

alleviates the need for $75,000 ‘‘safety money,’’ and the

station owner can sell for as low as $486,250, with

$413,000 paying off their boat and covering their ex-

penses for two years, and $73,250 paying off the (nowlower) capital gains taxes. An offer for gas and boat

maintenance alleviates the need for some of those funds

as well.

Whether station owners disclose their interests of

taking a vacation and sailing around the world and the

need for a job upon return, however, generally depends

on whether they trust the Texoil representative (Gold-

berg, 1997). Station owners typically believe that dis-closing such information leaves them vulnerable to

claiming tactics; they believe the Texoil representative

could use that information to infer that they are des-

perate, and thus push them toward selling for as little as

possible.

Negotiation outcomes

After participants completed the negotiation, dyadsreported whether or not they were able to reach an

agreement, and if they reached an agreement, they re-

ported the terms of their deal. Dyads were coded as ‘‘0’’

if they did not reach an agreement, ‘‘1’’ if they reached a

non-integrative agreement, and ‘‘2’’ if they reached an

integrative agreement. Non-integrative agreements were

those in which the sale price was the sole issue, and no

other issues were included. For example, if the stationowner sold the station for $490,000 and was not offered

a job upon return from vacation, this agreement was

considered non-integrative because it did not satisfy the

station owner�s interests maximally (they need a job

after their vacation) or the Texoil representative�s in-

terests maximally (they need good managers). Integra-

tive agreements, in contrast, were those that included

other issues that helped parties more effectively satisfytheir interests; for example, if the Texoil representative

offered the station owner a job offer upon return

from vacation, to pay the station owners for placing a

‘‘Texoil’’ advertisement on the sails, or to provide gas

and boat maintenance during the two-year vacation. In

these agreements, station owners would have adequate

funds for their vacation, and the Texoil representative

would have retained good managers in the long run. Toestablish reliability of the coding of agreements, a sec-

ond independent rater coded 20% of the negotiated

agreements using the same 0- to 2-point scheme; these

judgments were in 100% agreement with the original

codes.

Out of all 46 dyads, 20 (43.5%) did not reach agree-

ment, 14 (30.4%) reached a non-integrative agreement,

130 C. Anderson, L.L. Thompson / Organizational Behavior and Human Decision Processes 95 (2004) 125–139

and 12 (26.1%) reached an integrative agreement. Therewas no effect for sex on whether dyads reached an im-

passe, a non-integrative agreement, or an integrative

agreement, v2 ð2;N ¼ 46Þ ¼ 3:88.Across the dyads who reached a deal, integrative or

non-integrative, the average selling price for the gas

station was $489,608 (SD ¼ $36; 169:35). This average

price was lower than the mid-point between both

party�s reservation prices ($493,125), tð25Þ ¼ 3:28,p < :01. Therefore, Texoil representatives achieved bet-

ter distributive outcomes than station owners, in terms

of reaching an agreement that was farther from their

reservation price. There was no effect for sex on dis-

tributive outcomes (B ¼ �8:87, SE ¼ 21:94, b ¼ �:08;where men were coded as ‘‘1’’ and women as ‘‘2’’).

Data analyses

All analyses were conducted on the dyad level. As our

primary outcome variable was categorical (i.e., whether

or not a dyad�s outcome was integrative), we used lo-

gistic regression analyses in predicting integrative out-

comes. For each predictor variable, we report the

unstandardized regression coefficient and odds ratio.

The odds ratio reflects the amount by which the odds of

reaching agreement are multiplied for each additionalunit increase in a predictor variable. For example, in

predicting integrative agreements with powerful negoti-

ators� positive affect, an odds ratio of 2.0 means the odds

of reaching an integrative agreement are multiplied by

2.0 for every one-unit increase in powerful negotiators�positive affect.

To examine whether the more powerful negotiator�spositive affect predicted integrative outcomes better thanthe less powerful negotiator�s positive affect, we adapteda technique suggested by Cohen, Cohen, West, and

Aiken (2003) and outlined by Rindskopf (1984). This

technique involves comparing the residual from the

usual regression estimation (in which the powerful ne-

gotiator�s positive affect and the less powerful negotia-

tor�s positive affect are entered into the regression

separately) with the residual from a regression estima-tion in which the predictors being examined have been

combined (the powerful negotiator�s positive affect and

the less powerful negotiator�s positive affect are literally

added together). If the powerful negotiator�s positive

affect is a stronger predictor of integrative agreements,

the ‘‘unrestricted’’ model with both predictors entered

separately will have significantly less error variance than

the ‘‘restricted’’ model in which the predictors weresummed together (Rindskopf, 1984).4 Because we used

4 For users of structural equation modeling (SEM), this technique

is mathematically equivalent to comparing two models, one in which

the two paths from the predictors to the dependent variable are

determined by the model fitting program (i.e., allowed to vary), and

another in which the betas of the two paths are constrained to equality.

logistic regression, we adapted this technique by com-paring the model deviances between the two models,

rather than the residual sums of squares; this type of

likelihood ratio v2 test is similar to that in hierarchical

logistic regression analyses (Cohen et al., 2003).

Results and discussion

Integrative outcomes

Our main hypothesis was that powerful negotiators�positive affect would predict reaching an integrative

agreement more strongly than less powerful negotiators�positive affect. To test this hypothesis, we first focused

on predicting whether dyads reached an integrative

agreement rather than an impasse. The model deviance

of the restricted model, in which the powerful and less

powerful negotiators� positive affect were combined, was40.25 (v2 ð1;N ¼ 32Þ ¼ 2:095, p ¼ :07); the model devi-

ance of the unrestricted model, in which both predictors

were entered separately, was 35.85 (v2 ð2;N ¼ 32Þ ¼6:488, p < :05). The decrease in model deviances, or the

likelihood ratio v2 test (df ¼ 1), was 4.393, p < :05. Thisindicates powerful negotiators� positive affect was a

stronger predictor of reaching integrative agreement

than less powerful negotiators� positive affect, and thatour hypothesis was supported.

Fig. 1 illustrates this effect. As shown, the probability

of reaching an integrative agreement rather than an

impasse increased when the more powerful negotiator

was more positively affective. Specifically, the regression

coefficient for powerful negotiators� positive affect was

2.864, with an odds ratio or exp ðbÞ ¼ 17:53, p < :05.Therefore, the odds of reaching an integrative agreementwas multiplied by 17.53 for every unit increase in the

powerful negotiators� positive affect. However, the

probability of reaching an integrative agreement rather

Fig. 1. Study 1: The probability of reaching an integrative agreement

rather than an impasse increased when powerful individuals were more

positively affective.

C. Anderson, L.L. Thompson / Organizational Behavior and Human Decision Processes 95 (2004) 125–139 131

than an impasse did not increase when the less powerfulnegotiator was more positively affective. The regression

coefficient for less powerful negotiators� positive affect

was )1.14, exp ðbÞ ¼ :32, n.s.We also predicted whether, among the dyads who

reached an agreement, negotiators reached an integrative

rather than a non-integrative agreement. The model de-

viance was 31.29 in the restricted model (v2 ð2;N ¼26Þ ¼ 4:597, p < :05) and 31.19 in the unrestricted model(v2 ð2;N ¼ 26Þ ¼ 4:698, p < :05). The decrease in devi-

ance, or the likelihood ratio v2 test (df ¼ 1), was 0.10,

n.s. Therefore, positive affect did significantly predict

reaching integrative rather than non-integrative agree-

ments. However, powerful negotiators� positive affect

was not more important than less powerful negotiators�positive affect in reaching integrative rather than non-

integrative agreements. The positive affect of both ne-gotiators was equally important.

Distributive outcomes

To examine distributive outcomes, we focused on

dyads that reached an agreement (integrative and non-

integrative). Using dyad-level analyses, we regressed the

sale price of the station on both negotiators� positiveaffect. We did not find a relation between powerful ne-gotiators� positive affect and the sale price, B ¼ �21:67(SE ¼ 20:15, n.s.), or between less powerful negotiators�positive affect and the sale price, B ¼ �24:82 (SE ¼17:39, n.s.). Therefore, positive affect was not helpful orharmful in negotiators� attempts to achieve better dis-

tributive gains.

Additional analyses

Although the PANAS positive affect scale was de-

signed to measure the overarching construct of positive

affect (Watson et al., 1988), it is possible that the scale

subsumes multiple, smaller components of positive

affect. For example, two components seem to emerge

from a cursory examination of the scale items: one

characterized by enthusiasm (which would include items

such as ‘‘excited,’’ ‘‘enthusiastic,’’ ‘‘interested,’’ and‘‘inspired’’), the other characterized by surgency (which

would include items such as ‘‘determined,’’ ‘‘strong,’’

and ‘‘proud’’). If the PANAS positive affect scale does

subsume multiple components, do different components

have different effects on negotiation processes and out-

comes? For example, it is possible that powerful nego-

tiators� enthusiasm helps facilitate integrative bargaining

more than their surgency, and that surgency helps ne-gotiators achieve higher distributive gains, whereas en-

thusiasm does not.

To examine these questions, we conducted a factor

analysis with VARIMAX rotation. The scree test indi-

cated a single, general factor, rather than multiple fac-

tors, suggesting the PANAS positive affect scale

measured a single broad construct. To be thorough,

however, we conducted a second factor analysis andforced a two-factor solution. In this analysis, the two

components of enthusiasm (which included excited, en-

thusiastic, interested, inspired, and active) and surgency

(which included determined, strong, proud, alert, and

attentive) emerged. We thus constructed a measure for

each component based on the respective items.

To examine whether powerful negotiators� enthusi-asm was more important than their surgency in pre-dicting integrative outcomes, we again used the analysis

outlined by Rindskopf (1984). The model deviance of

the unrestricted model, in which the powerful negotia-

tors� enthusiasm and surgency were entered separately,

was not significantly lower than the model deviance of

the restricted model, in which the powerful negotiators�positive affect was entered as a single construct, v2

ðdf ¼ 1Þ ¼ 2:22, n.s. Therefore, powerful negotiators�enthusiasm and surgency were equally important to fa-

cilitating integrative outcomes. To examine whether

negotiators� surgency helped them achieve higher dis-

tributive gains, we regressed relative outcomes on both

negotiators� surgency. We did not find a relation be-

tween powerful negotiators� positive affect and their

distributive gains, B ¼ 17:90 (SE ¼ 15:71, n.s.), or be-

tween less powerful negotiators� positive affect and theirdistributive gains, B ¼ �20:86 (SE ¼ 16:82, n.s.).

Therefore, we did not find that surgency was helpful to

achieving better distributive gains.

These analyses suggest positive affect operated as a

single unified construct, rather than as a constellation of

distinct components. Negotiators� responses to the items

hung together in a single factor, and even when we split

the positive affect construct into two separate compo-nents, enthusiasm and surgency, we found they had

consistent relations with negotiation outcomes.

Summary

The results of Study 1 replicated previous work in

finding that higher levels of positive affect in negotiations

were related to more integrative agreements. However,

the findings also extend that work in two importantways. First, they suggest that trait or dispositional affect

benefits integrative bargaining, which contributes to our

understanding of individual differences in negotiations.

Second, they suggest that positive affect originating from

powerful negotiators is more important to the quality of

the negotiated outcome. High levels of positive affect

originating from powerful negotiators made it more

likely that dyads reached an integrative deal rather thanan impasse, whereas less powerful negotiators� positiveaffect had no effect. Moreover, higher levels of positive

affect did not harm negotiators� distributive gains in the

negotiation—thus, there was no observed disadvantage

to being positively affective.

It is interesting that the positive affect of both parties

was equally important in predicting whether dyads

132 C. Anderson, L.L. Thompson / Organizational Behavior and Human Decision Processes 95 (2004) 125–139

reached integrative, rather than non-integrative agree-ments. This could have been due to the small sample size

in comparing integrative and non-integrative agree-

ments (N ¼ 26), or it could have been due to the

different factors that contribute to integrative and non-

integrative agreements. Integrative agreements in this

negotiation required more than just the development of

mutual trust; it also required the ability to think crea-

tively and come up with solutions that are not imme-diately apparent. Thus, it is possible both powerful and

less powerful negotiators� positive affect might have fa-

cilitated creative thinking, and that both made the dif-

ference between integrative and non-integrative

agreements.

Study 2

Study 2 was designed to extend the findings of Study

1 in a number of ways. First, we wanted to show that

there is something uniquely important about positive

affect that helps negotiators reach integrative outcomes.

It is possible, for example, that the relation between

positive affect and joint outcomes in Study 1 was spu-

rious, caused by unmeasured variables such as trait co-operativeness or communicativeness. More cooperative,

altruistic people might tend to feel more positive affect in

negotiations and tend to achieve higher joint gains be-

cause of their motivation to increase all parties� well-being. Or, more communicative and sociable people

might feel more positive affect in negotiations and tend

to achieve higher joint gains, because their better com-

munication leads to greater understanding and thusmore opportunities to expand the pie. To help rule out

these third-variable explanations, we measured negoti-

ators� Agreeableness and Extraversion. Agreeableness

involves traits such as cooperativeness, altruism, and the

tendency to trust others (John & Srivastava, 1999).

Extraversion includes traits such as talkativeness and

sociability (John & Srivastava, 1999, p. 121).5

Second, we have argued that positive affect helpsnegotiators reach integrative outcomes because it helps

increase trust between parties. In Study 2, we directly

examined whether trust was related to positive affect,

and whether it plays a role in mediating the relation

between negotiators� positive affect and their integrative

outcomes.

5 Although one study found no effect for Agreeableness or for

Extraversion on integrative outcomes (Barry & Friedman, 1998), we

were hesitant to conclude, based on that one study, that these two

personality traits do not facilitate integrative outcomes. As Agree-

ableness includes trust for others and a cooperative orientation, and

Extraversion is related to better social skills and communicativeness,

these traits would seem highly beneficial for the integrative process.

Third, we used a negotiation simulation in Study 2that typically allows for greater variation in integrative

outcomes. Our main dependent variable in Study 1 was

whether or not participants were able to reach an inte-

grative agreement, and thus there was no variation in

the extent to which the deals could be integrative. In

Study 2, negotiators bargained over a number of issues

that were worth various points to them, which allowed

for more variation in integrativeness.Finally, in Study 1, we only examined the distributive

outcomes of individuals who reached an agreement,

because the distributive outcome of those who ended in

impasse was ambiguous (do you consider their best al-

ternative to the negotiation their distributive outcome,

and if so, which alternative to the negotiation is their

best?). This limited us to examining the distributive

outcomes of a possibly skewed subset of the sample; onecould argue the negotiators who reached deals achieved

better distributive outcomes than those who failed to

reach agreement. In Study 2, we used a negotiation

simulation in which all participants typically reach

agreement (Pinkley et al., 1994), which allowed us to

examine the distributive gains in a wider sample.

Methods

Participants

One hundred four masters of business administration

(MBA) students who were enrolled in a course on ne-

gotiations participated, a different set of students than

those in Study 1. Of these, 76 were in dyads that com-

pleted all measures (57 men, 19 women). Therefore, our

analyses focused on these 76 participants. The study wasconducted in the second week of the course. As in Study

1, participants were randomly placed into dyads and

into roles.

Positive affect

In the week prior to the beginning of the course,

participants completed the positive affect scale of the

PANAS in an on-line survey. They reported how muchthey felt each emotion ‘‘in general, or on average.’’ The

scale again showed satisfactory internal consistency,

coefficient a¼ .80; the mean score was 3.65 (SD ¼ :48).There were no effects for sex; the mean score for men

was 3.60 (SD ¼ :50), and for women was 3.79

(SD ¼ :40).

Agreeableness and extraversion

Prior to the course, participants also completed the

Agreeableness and Extraversion scales of the Big Five

Inventory (BFI; John, Donahue, & Kentle, 1991). For

the Agreeableness measure, participants rated how

much they agreed with nine statements such as, ‘‘I see

myself as someone who likes to cooperate with others,’’

on a scale from 1 (‘‘Strongly disagree’’) to 5 (‘‘Strongly

6 We also gave some candidates and some recruiters alternatives to

reaching a deal in the negotiation (or BATNAs). For example, some

candidates were told they had an equally attractive job offer.

Interestingly, having an alternative to the negotiation did not have a

significant impact on perceptions of power or on the relation between

participants� positive affect and their dyads� joint outcome or mutual

trust. This suggests that in this negotiation simulation, the BATNA

manipulation did not have its usual effect of creating a power

difference. This could have been because negotiators were uncertain as

to whether their opponent had a BATNA as well, or it could have been

because they did not yet understand the importance of BATNAs in

determining their power in negotiations.

C. Anderson, L.L. Thompson / Organizational Behavior and Human Decision Processes 95 (2004) 125–139 133

agree’’). The items showed high internal consistency,coefficient a¼ .81; the mean score was 3.82 (SD ¼ :69).For the Extraversion measure, participants rated their

agreement with eight statements such as, ‘‘I see myself as

someone who is talkative,’’ on the same scale. The Ex-

traversion items also showed high internal consistency,

coefficient a¼ .84; the mean score was 3.46 (SD ¼ :78).Men did not differ from women on Agreeableness

(M ¼ 3:79 compared to M ¼ 3:93) or on Extraversion(M ¼ 3:41 compared to M ¼ 3:61).

Procedure and role instructions

Study 2 was also conducted as a class exercise. Each

negotiating dyad was comprised of a job candidate and

a recruiter in a job interview setting (Neale, 1997).

Participants received their role instructions before the

exercise began, and were given 40min to negotiate. Af-ter the negotiation, participants completed a question-

naire privately.

Participants� role materials described eight negotiable

issues, each of differing importance; importance was

reflected in how many points the issue was worth. The

issues were broken up into three types. First, distributive

issues were ones on which candidates and recruiters had

opposing preferences. For example, candidates wantedto start as early as possible, but recruiters wanted them

to start as late as possible, and this issue was worth the

same number of points to them. Second, tradable issues

were those in which candidates and recruiters had op-

posing preferences, but placed different values on them;

therefore, they could trade concessions on one issue for

gains on another. For example, candidates did not care

about their vacation time but cared a great deal aboutreceiving a large bonus; recruiters cared a great deal

about vacation time but cared less about giving larger

bonuses. Therefore, parties could trade a high bonus

amount for very little vacation time. Finally, compatible

issues were ones in which participants� preferences wereperfectly aligned with the other party. For example,

both parties wanted the candidate to work in San

Francisco.Previous research that has used this exercise has

found the recruiter to be perceived as more powerful by

negotiators than the job candidate (Allred et al., 1997;

Galinsky, Thompson, & Kray, 2003; Pinkley et al.,

1994). Job recruiters have legitimate power, because

they occupy a position of authority relative to the can-

didate, and they have reward power, in that in the real

world, job candidates are often more dependent on themto reach a deal than vice versa. To confirm that re-

cruiters were perceived as more powerful, participants

rated the item, ‘‘I think I had a great deal of power in

the negotiation,’’ on a scale from 1 (‘‘Strongly disagree’’)

to 7 (‘‘Strongly agree’’). Consistent with prior work,

recruiters saw themselves as more powerful, M ¼ 4:73(SD ¼ 1:16) than candidates, M ¼ 4:20 (SD ¼ 1:12),

t ¼ 1:58, p ¼ :06, though this effect was just below thep < :05 level.6

Negotiation outcomes

As expected, no dyads reached an impasse. Negotia-

tor�s own outcomes were scored by adding the points

they obtained across the eight issues. They could obtain

anywhere from )8400 to 13,200 points; the average was

6250.00 (SD ¼ 1362:30). To measure how integrativeagreements were, we focused on the commonly used

index of joint outcome, or the sum of dyad member�sown outcomes (e.g., Allred et al., 1997; Bazerman,

Magliozzi, & Neale, 1985; Ben-Yoav & Pruitt, 1984;

Mannix & Neale, 1993; Schurr, 1987). We conducted an

ANOVA, predicting joint outcome on participants� sex,their opponents� sex, and the interaction; for the full

model, F ð3; 37Þ ¼ :94, n.s. There was no effect of sex onjoint outcome.

To measure distributive outcomes, we calculated the

proportion of candidates� outcome to the total joint

outcome, so that the distributive and integrative out-

comes would be unconfounded. With this measure, the

higher the candidates� distributive outcome, the lower

the recruiters� distributive outcome. The mean distrib-

utive outcome was by definition¼ .50 (SD ¼ :10). Therewas no effect for sex on distributive outcome; in an

ANOVA, we predicted distributive outcomes with par-

ticipants� sex, their opponents� sex, and the interaction

(for the full model, F ð3; 37Þ ¼ :88). There was also no

effect for role on distributive outcome.

Mutual trust

Participants rated their agreement with three items,‘‘S/he cared about my interests and needs,’’ ‘‘S/he was

only looking out for him or herself,’’ and ‘‘I was weary

of trusting the other party,’’ on a scale from 1 (‘‘Strongly

disagree’’) to 7 (‘‘Strongly agree’’). After reverse-scoring

the latter two items and aggregating dyad members�ratings, the three items showed internal consistency,

a¼ .64, and were thus combined into one measure of

mutual trust (M ¼ 5:11, SD ¼ :82). Using Kenny and laVoie�s (1985) technique to test for dependence of data

within dyads, we found that trust was dependent in

dyads, F ð38; 39Þ ¼ 9:03, p < :01. Therefore, as they

134 C. Anderson, L.L. Thompson / Organizational Behavior and Human Decision Processes 95 (2004) 125–139

suggest, it was not appropriate to focus on eithernegotiator�s trust individually, and we focused on dyad-

level trust in all analyses. As expected, given the de-

pendence of trust in dyads, there was not a significant

effect for role on trust.

Results and discussion

Joint outcomes

To test whether powerful negotiators� positive affect

was a stronger predictor of joint outcomes than less

powerful negotiators� positive affect, we again used the

technique outlined by Rindskopf (1984). Specifically, we

predicted joint outcome with the sum of both negotia-

tors� positive affect, B ¼ 252:66 (SE ¼ 253:74), n.s. We

then predicted joint outcome with the more powerful

negotiators� (i.e., recruiters�) positive affect and the lesspowerful negotiators� (i.e., candidates�) positive affect asseparate independent variables. The residual sums of

squares in the restricted model was significantly larger

than in the second, unrestricted model, F ð1; 35Þ ¼ 4:17,

Table 1

Study 2: Predicting joint outcomes and distributive outcomes

Predictor variable Dyads� joint outco

Less powerful negotiators� Agreeableness 57.96 (258.08)

Less powerful negotiators� Extraversion )76.34 (230.90)

Less powerful negotiators� positive affect )309.92 (454.58)

Powerful negotiators� Agreeableness )49.34 (253.52)

Powerful negotiators� Extraversion )59.57 (240.61)

Powerful negotiators� positive affect 617.76� (342.70)

Note. Shown are regression coefficients. In the ‘‘Distributive outcome’’

portion of the dyad�s joint outcome. Thus, higher regression coefficients reflec

negative regression coefficients reflect a positive effect on the more powerful* p < :05.

Fig. 2. Study 2: The joint outcome dyads achieved was predicted by

powerful negotiators� positive affect more strongly than the less pow-

erful negotiators� positive affect.

p < :05. Therefore, this indicates that the unrestrictedmodel accounted for significantly more variance in joint

outcome (Cohen et al., 2003; Rindskopf, 1984). Illus-

trated in Fig. 2 are the regression lines for powerful

negotiators� positive affect (B ¼ 554:79, SE ¼ 284:75,p < :05) and for less powerful negotiators� positive affect(B ¼ �355:84, SE ¼ 384:76, n.s.). As shown, powerful

negotiators� positive affect was a significantly stronger

predictor than less powerful negotiators� positive affect.

Agreeableness and extraversion

Was there something unique about negotiators�positive affect above and beyond their Agreeableness

and Extraversion that helped reach higher joint gains?

Or did cooperativeness or talkativeness drive the

correlation between positive affect and integrative

outcomes? Table 1 displays full regression models inwhich we predicted joint outcome with negotiators�Agreeableness, Extraversion, and positive affect. As

shown, the relations between powerful parties� positiveaffect and joint outcomes did not change. Powerful

negotiators� positive affect was still significantly related

to integrative outcomes, whereas less powerful negotia-

tors� positive affect was unrelated. Further, for both

negotiators, Agreeableness and Extraversion did nothave a significant effect on joint outcomes. Thus, these

analyses suggest the relation between powerful parties�positive affect and joint gains was not due to Agree-

ableness or Extraversion. Further, they replicate Barry

and Friedman�s (1998) study that found no relation

between these personality traits and joint outcomes.

Though the two personality traits would seem likely

contributors to joint outcome, the evidence speaks to thecontrary.

Mutual trust

We first examined whether powerful negotiators�positive affect had a stronger relation to mutual trust

than less powerful negotiators� positive affect. Using

Rindskopf�s (1984) analysis, we again found that the

unrestricted model, in which both negotiators� positiveaffect were entered as separate predictors, had less error

me Less powerful negotiators� distributive outcome

0.02 (0.03)

0.00 (0.03)

0.01 (0.05)

0.00 (0.03)

0.04 (0.03)

)0.03 (0.04)

column, the dependent variable is the less powerful negotiators� pro-t a positive effect on the less powerful negotiators� distributive outcome;

negotiators� distributive outcome.

C. Anderson, L.L. Thompson / Organizational Behavior and Human Decision Processes 95 (2004) 125–139 135

variance than the restricted model that combined thetwo negotiators� positive affect into one predictor,

F ð1; 35Þ ¼ 4:38, p < :05. In the unrestricted model, the

unstandardized regression coefficients for less powerful

negotiators� positive affect was ).02 (SE ¼ :37,b ¼ �:01, n.s.) and for more powerful negotiators� po-sitive affect was .73 (SE ¼ :32, b ¼ :45, p < :05). There-fore, more powerful negotiators� positive affect was a

stronger predictor of mutual trust.To test whether mutual trust was a mediator, we had

to satisfy four conditions (Baron & Kenny, 1986). First,

powerful individuals� positive affect needed to be related

to joint outcomes, which we have already shown. Sec-

ond, powerful individuals� positive affect needed to be

related to mutual trust, which we have also shown.

Third, mutual trust needed to be related to joint out-

comes while controlling for recruiters� positive affect,and fourth, the relation between recruiters� positive af-

fect and joint outcome needs to be reduced when taking

into account the indirect effect of group trust. Regress-

ing joint outcome on both powerful individuals� positiveaffect and mutual trust, we found the regression coeffi-

cient for mutual trust was 250.30 (SE ¼ 230:50, b ¼ :25,p ¼ :14), whereas the regression coefficient for powerful

individuals� positive affect was 98.58 (SE ¼ 378:74,b ¼ :06, p ¼ :40). Thus, when controlling for mutual

trust, the effect for powerful individual�s positive affect

almost disappeared; however, we can only conclude that

mutual trust at most partially mediated the effect of

powerful parties� positive affect on joint outcome, be-

cause the effect of mutual trust approached significance.

Distributive outcomes

We regressed, on the dyad-level, both negotiators�Agreeableness, Extraversion, and positive affect on

candidates� distributive outcomes (i.e., their proportion

of the total joint outcome). Shown in the last column of

Table 1 are the regression coefficients. For both less

powerful and more powerful negotiators, none of the

predictor variables had a significant relation to distrib-

utive outcomes. Thus, similar to Study 1, positive affectdid not relate to distributive gains. Positive affect did not

help or hinder obtaining a large slice of the pie.

Additional analyses

As in Study 1, we examined whether the PANAS

positive affect scale was comprised of two components,

enthusiasm and surgency, and whether these different

components had different effects on the negotiation. In afactor analysis with VARIMAX rotation, the scree test

again indicated a single, general factor, rather than

multiple factors. In a second factor analysis where we

forced a two-factor solution, however, the two compo-

nents of enthusiasm and surgency emerged again. In

examining whether powerful negotiators� enthusiasm

was more important than their surgency in predicting

integrative outcomes, the residual sums of squares of theunrestricted model was not significantly less than the

residual sums of squares of the restricted model,

F ð1; 34Þ ¼ :29, n.s. Therefore, powerful negotiators�enthusiasm and surgency were again equally important

to facilitating integrative outcomes. In examining whe-

ther negotiators� surgency helped them achieve higher

distributive gains, we did not find a relation between

powerful negotiators� surgency and their distributivegains, B ¼ �:01 (SE ¼ :028, n.s.), or between less pow-

erful negotiators� surgency and their distributive gains,

B ¼ �:03 (SE ¼ :04, n.s.). Therefore, we did not find

that surgency was helpful to achieving better distributive

gains. As in Study 1, these follow-up analyses suggest

that positive affect operated as a single construct rather

than as a constellation of distinct constructs.

Summary

The findings from Study 2 replicated those from

Study 1, but also extended them in important ways.

Powerful negotiators� positive affect predicted joint

gains more strongly than less powerful negotiators� po-sitive affect. However, the findings from Study 2 also

helped rule out an alternative explanation: that the re-

lation between powerful negotiators� positive affect andjoint gains was due to third variables such as trait co-

operativeness or communicativeness. When we con-

trolled for Agreeableness and Extraversion, the findings

remained, suggesting that there is something uniquely

important about positive affect that facilitated integra-

tive agreements. Study 2 also found that powerful ne-

gotiators� positive affect predicted mutual trust more

strongly than less powerful negotiators� positive affect,and that mutual trust played a partially mediating role

in the relation between positive affect and joint out-

comes. Thus, this suggests that part of the reason why

powerful negotiators� positive affect leads to higher joint

outcomes is that it fosters trust in negotiations. Finally,

higher levels of positive affect again did not relate to

lower distributive outcomes, suggesting that positive

affect did not lead to a trade-off between integrative anddistributive gains.

General discussion

In two studies, we found that powerful negotiators�positive affect shaped bargaining processes and out-

comes more than less powerful negotiators� positive af-fect. Powerful negotiators� positive affect was the only

and best predictor of the trust that emerged during the

negotiation, and it was the only and best predictor of

integrative outcomes. Further, mediation analyses sug-

gested that the relation between powerful individuals�positive affect and integrative outcomes was partly due

to the mutual trust it fostered. Therefore, the findings

136 C. Anderson, L.L. Thompson / Organizational Behavior and Human Decision Processes 95 (2004) 125–139

suggest that positive affect had a trickle-down effect inthese negotiations; the more powerful negotiator defined

the tone of the negotiation as either positive and trust-

ing, or as more competitive and suspicious, which ulti-

mately shaped the quality of the agreement.

Moreover, there was something uniquely important

about positive affect that facilitated joint gains. When we

included negotiators� Agreeableness and Extraversion as

predictors, positive affect was the only significant pre-dictor of joint gain, whereas the effects for Agreeable-

ness and Extraversion hovered around zero. This is

consistent with prior work (Barry & Friedman, 1998),

and indicates that variables such as trait cooperativeness

or sociability did not cause people to be positively

affective and to achieve high joint gains.

One interesting question is why positive affect facili-

tated integrative agreements, whereas Agreeableness andExtraversion did not, even though they include traits

such as cooperativeness and the propensity to trust

others. One reason could lie in how easily other people

detect these personality characteristics. For example,

people with high positive affect might exhibit a lot of

overt behaviors that signal their higher levels of excite-

ment and enthusiasm (e.g., Ekman, 1973). Therefore,

their positive affect is more easily detectable by others,which causes others to view them as more cooperative

and trustworthy (e.g., Knutson, 1996). In contrast,

people high in Agreeableness and Extraversion, though

they may actually be more cooperative, altruistic, or

more trusting, might not exhibit behaviors that signal

those traits as clearly; therefore, Agreeableness and Ex-

traversion might not be as easily detectable. As evidence,

people are generally quite inaccurate in rating others�Agreeableness (Albright, Kenny, & Malloy, 1988). In

short, therefore, positively affective people might come

across as more cooperative than cooperative people do.

Finally, we did not find any relations between posi-

tive affect and distributive gains. Positive affect did not

seem to introduce a trade-off between integrative and

distributive outcomes; there was no observable disad-

vantage of being positively affective. At the same time,however, positive affect also did not lead people to be

more effective at distributive bargaining. Positively af-

fective individuals� increased attention, interest, and

energy did not propel them to taking the biggest slice of

it, but rather, it only facilitated a trusting atmosphere

and allowed for more integrative agreements.

Limitations and future directions

The main limitation of the current research is that it

leaves relatively open the question of why powerful ne-

gotiators� positive affect had a stronger effect on inte-

grative bargaining than less powerful negotiators�positive affect. In the Introduction, we proposed two

possible explanations. First, powerful negotiators�

positive affect might communicate to others that theyare trustworthy (Frank, 1988; Knutson, 1996), which

helps build trust in the negotiation and thus facilitate

communication. Second, powerful negotiators� positiveaffect might have a stronger effect on their counterparts�thoughts, feelings, and behavior than vice versa

(Anderson et al., 2003; Cheng & Chartrand, 2003). Al-

though we obtained some supportive evidence for our

first explanation in Study 2, there seems to be othermechanisms responsible for the effect. Further, we did

not examine the second explanation in the current re-

search.

One additional explanation for our findings is that

powerful negotiators� positive affect attenuates the less

powerful negotiators� perception of the difference in

power, which in turn, helps mitigate some of the mis-

trust that typically exists in power-imbalanced negotia-tions. That is, less powerful negotiators might enter the

negotiation expecting their more powerful counterpart

to display high levels of anger, because anger is associ-

ated with power and status (Tiedens, 2001). Instead,

when their powerful counterpart displays more positive

affect, the less powerful negotiator might infer that the

power difference between them might not be that wide.

As the less powerful negotiators� begin to perceive thepower balance as more equal, they might be more open

to integrative bargaining (Lawler & Yoon, 1993). As

suggestive evidence for this point, in Study 2, powerful

negotiators� positive affect was correlated .20 with less

powerful negotiators� self-perceived power in the nego-

tiation. However, as we did not measure less powerful

negotiators� perceptions of the power balance (i.e., theirpower and their counterparts� power), this evidence isonly suggestive.

To further examine why positive affect is so beneficial

to integrative negotiations, future research should test

for emotional contagion effects, or whether less powerful

negotiators tend to ‘‘catch’’ the positive affect of pow-

erful negotiator. We expect that less powerful negotia-

tors will experience higher levels of positive emotion in a

negotiation with a powerful negotiator higher in traitpositive affect than in a negotiation with a powerful

negotiator lower in trait positive affect. However, we do

not expect powerful negotiators� positive emotional ex-

perience in a negotiation to be influenced by less pow-

erful negotiators� trait positive affect.

On a related point, we examined negotiators� trust foreach other after they knew of their outcome (and in

some cases, the other parties� outcome) in the negotia-tion. It is therefore possible that their trust was affected

by the final outcome, rather than vice versa. In the

future, researchers should also somehow obtain a mea-

sure of trust so that it is not affected by negotiators�knowledge of their outcomes.

Further, to manipulate power, we focused on a role

manipulation that might have created more than just

C. Anderson, L.L. Thompson / Organizational Behavior and Human Decision Processes 95 (2004) 125–139 137

power differences. For example, in Study 2, being des-ignated as a job candidate might have placed the indi-

vidual in a position of lower power, but it could also

have activated norms and standards of behavior unre-

lated to power; for example, it is commonly considered

inappropriate for job candidates to make the first offer

in a job offer negotiation. Thus, future work should use

manipulations of power that do not have potential

confounding factors. Although we manipulated alter-natives to the negotiation in Study 2 with that aim in

mind (Pinkley et al., 1994), this manipulation was inef-

fective in creating differences in power. Although we are

not sure why this manipulation was less effective than it

was in prior research, we believe this speaks only to its

own ineffectiveness, and not to the lack of validity of our

hypotheses.

Finally, the current work, like much of the researchon negotiation, used students in negotiation classes who

were involved in a course exercise. This obviously limits

the generalizability of the findings because the stakes

were lower than they would typically be in the real

world, and because the levels of affect were likely less

extreme. Others have pointed out that the range and

intensity of affect simply cannot be adequately captured

in classroom exercise settings, and thus it is important toconduct complementary research in real-world negotia-

tion settings (Barry et al., in press). Some current work

has begun to study the effect of affect in real-world

dispute situations (Friedman et al., 2004), a direction

that more researchers should take in negotiation con-

texts.

Theoretical implications

The current research has theoretical implications for

a number of domains. In the study of negotiations,

many theorists have pointed out that we know very little

of how individual differences shape the bargaining pro-

cess (Barry et al., in press; Bazerman & Neale, 1992;

Thompson, 1990). Although intuition tells us that indi-

viduals� stable and enduring behavioral tendenciesshould influence their negotiation behavior, the findings

are so inconsistent and sometimes contradicting that

many researchers have wondered whether individual

differences do in fact matter (Lewicki & Litterer, 1985).

The current studies may help shed some light on this

domain in two ways. First, they suggest that individual

differences in affective processes are an important focus

for future research. People�s tendency to experience andexpress certain emotions, as our findings suggest, might

have a profound effect on the way they approach ne-

gotiations, the way others perceive them, and ultimately

the negotiation outcomes they reach. Second, they sug-

gest individual differences might matter in some roles

more than in others; therefore, it is worth considering

whose personality is being measured. For example,

individuals� personality traits might matter little whenthey are in positions of low power, because the situa-

tional constraints of being lower in power override and

wash away any impact of personality traits (Chen, Lee-

Chai, & Bargh, 2001). Conversely, individuals� person-ality traits might matter a great deal in situations where

individuals have much power, because powerful indi-

viduals have the freedom to ‘‘call the shots’’ as they

wish, and therefore their traits have the capacity toshape their own behavior and the social environment.

Extending the work on affect in negotiations, the

current findings also provide a fuller understanding of

why affect matters. Previous work has shown that po-

sitive affect helps facilitate integrative agreements be-

cause it shapes negotiators� perceptions of others�interests (Allred et al., 1997) and their planned use of

strategies (Forgas, 1998). The current work shows thatpositive affect facilitates integrative outcomes also be-

cause it engenders trust between individuals. This find-

ing is consistent with a social-functionalist approach to

emotions, which argues that emotions help communi-

cate one�s intentions and likely future behavior (e.g.,

Frank, 1988; Morris & Keltner, 2000).

The current work also extends research on power by

showing that powerful individuals shape their socialenvironments in ways that are consistent with their

personality traits. Personality researchers have been in-

terested for some time in what they call the proactive

person–environment interaction (Caspi & Bem, 1990).

This refers to the process by which individuals select and

construct their own particular social environments.

What this work finds is that individuals� personalitiesdetermine the way they interact with the social world,which in turn makes their social environment more

consistent with their personality. For example, dis-

agreeable people tend to shape their interactions with

others in ways that make them more hostile, suspicious,

and contentious (Caspi & Bem, 1990). Consistent with

recent research (Anderson et al., 2003), our findings

suggest individuals with power have a particularly

strong capacity to construct their social environments.Their personality dispositions tend to more strongly

impact their social worlds than less powerful individuals�personality dispositions.

References

Albright, L., Kenny, D. A., & Malloy, T. E. (1988). Consensus in

personality judgments at zero acquaintance. Journal of Personality

and Social Psychology, 55, 387–395.

Allred, K. G., Mallozzi, J. S., Matsui, F., & Raia, C. P. (1997). The

influence of anger and compassion on negotiation performance.

Organizational Behavior and Human Decision Processes, 70, 175–

187.

Anderson, C., & Berdahl, J. L. (2002). The experience of power:

Examining the effects of power on approach and inhibition

138 C. Anderson, L.L. Thompson / Organizational Behavior and Human Decision Processes 95 (2004) 125–139

tendencies. Journal of Personality and Social Psychology, 83, 1362–

1377.

Anderson, C., Keltner, D., & John, O. P. (2003). Emotional conver-

gence between people over time. Journal of Personality and Social

Psychology, 84, 1054–1068.

Ashby, F. G., Isen, A. M., & Turken, A. U. (1999). A neuropsycho-

logical theory of positive affect and its influence on cognition.

Psychological Review, 106, 529–550.

Bacharach, S. B., & Lawler, E. J. (1980). Power and politics in

organizations. San Francisco: Jossey-Bass.

Baron, R. A. (1990). Environmentally induced positive affect: Its

impact on self-efficacy, task performance, negotiation, and conflict.

Journal of Applied Social Psychology, 20, 368–384.

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.

Barry, B., & Friedman, R. A. (1998). Bargainer characteristics in

distributive and integrative negotiation. Journal of Personality and

Social Psychology, 74, 345–359.

Barry, B., Fulmer, I. S., & Van Kleef, G. A. (in press). I laughed, I

cried, I settled: The role of emotion in negotiation. In M. Gelfand

& J. M. Brett (Eds.), Culture and negotiation: Integrative ap-

proaches to theory and research.

Barsade, S. G. (2002). The ripple effects: Emotional contagion and its

influence on group behavior. Administrative Science Quarterly, 47,

644–675.

Bazerman, M. H., Magliozzi, T., & Neale, M. A. (1985). Integrative

bargaining in a competitive market. Organizational Behavior and

Human Performance, 34, 294–313.

Bazerman, M. H., & Neale, M. A. (1992). Negotiating rationally. New

York: Free Press.

Ben-Yoav, O., & Pruitt, D. (1984). Resistance to yielding and the

expectation of cooperative future interaction in negotiation.

Journal of Experimental Social Psychology, 20, 323–353.

Brett, J. M. (2001). Negotiating globally: How to negotiate deals,

resolve disputes, and make decisions across cultural boundaries?. San

Francisco, CA: Jossey-Bass.

Carnevale, P. J., & Isen, A. M. (1986). The influence of positive affect

and visual access on the discovery of integrative solutions in

bilateral negotiation. Organizational Behavior and Human Decision

Processes, 37, 1–13.

Caspi, A., & Bem, D. J. (1990). Personality continuity and change

across the life course. In P. A. Lawrence (Ed.), Handbook of

personality: Theory and research (pp. 549–575). New York:

Guilford Press.

Chance, M. R. A. (1967). Attention structure as the basis of primate

rank order. Man, 2, 503–518.

Chen, Y., Brockner, J., & Greenberg, J. (2004). When is it ‘‘a pleasure

to do business with you’’? The effects of status, outcome favorability,

and procedural fairness. In press.

Chen, S., Lee-Chai, A. Y., & Bargh, J. A. (2001). Relationship

orientation as a moderator of the effects of social power. Journal of

Personality and Social Psychology, 80, 173–187.

Cheng, C. M., & Chartrand, T. L. (2003). Self-monitoring without

awareness: Using mimicry as a nonconscious affiliation strategy.

Journal of Personality and Social Psychology, 85, 1170–

1179.

Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied

multiple regression/correlation analysis for the behavioral sciences

(3rd ed.). Hillsdale, NJ: Lawrence Erlbaum.

Ekman, P. (1973). Universals and cultural differences in facial

expressions of emotion. Nebraska Symposium on Motivation, 207–

283.

Ekman, P. (1994). Moods, emotions, and traits. In P. Ekman & R. J.

Davidson (Eds.), The nature of emotion: Fundamental questions (pp.

56–58). New York: Oxford University Press.

Emerson, R. M. (1962). Power-dependence relations. American Socio-

logical Review, 27, 31–40.

Faley, T., & Tedeschi, J. T. (1971). Status and reactions to threats.

Journal of Personality and Social Psychology, 17, 192–199.

Fisher, R., & Ury, W. (1991). Getting to yes. Boston: Houghton-

Mifflin.

Fiske, S. T. (1993). Controlling other people: The impact of power on

stereotyping. American Psychologist, 48, 621–628.

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.

Frank, R. H. (1988). Passions within reason: The strategic role of

emotions. New York: Norton.

French, J., & Raven, B. (1959). The bases of social power. In D.

Cartwright (Ed.), Studies in social power. Ann Arbor, MI: Institute

for Social Research.

Fridlund, A. J. (1994). Human facial expression: An evolutionary view.

New York: Academic Press.

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.

Frijda, N. H. (1994). Varieties of affect: Emotions and episodes,

moods, and sentiments. In P. Ekman & R. J. Davidson (Eds.), The

nature of emotion: Fundamental questions (pp. 59–67). New York:

Oxford University Press.

Galinsky, A. D., Thompson, L., & Kray, L. J. (2003). Role power and

negotiating power: A test of threat and recovery through self-

affirmation. Working paper.

Goldberg, S. (1997). Texoil. In J. M. Brett (Ed.), Teaching materials for

negotiations and decision making. Evanston, IL: Northwestern

University, Dispute Resolution Research Center.

Goldhamer, H., & Shils, E. A. (1939). Types of power and status.

American Journal of Sociology, 45, 171–182.

Hecht, M., & LaFrance, M. (1998). License or obligation to smile: The

effect of power and sex on amount and type of smiling. Personality

and Social Psychology Bulletin, 24, 1332–1342.

Isen, A. M., Daubman, K. A., & Nowicki, G. P. (1987). Positive affect

facilitates creative problem solving. Journal of Personality and

Social Psychology, 52, 1122–1131.

Isen, A. M., & Levin, P. F. (1972). Effect of feeling good on helping:

Cookies and kindness. Journal of Personality and Social Psychol-

ogy, 21, 384–388.

John, O. P., Donahue, E. M., & Kentle, R. (1991). The ‘‘Big Five’’

Inventory: Versions 4a and 54. Technical Report, Institute of

Personality and Social Research, University of California, Berke-

ley, CA 94720.

John, O. P., & Srivastava, S. (1999). The Big Five trait taxonomy:

History, measurement, and theoretical perspectives. In L. A. Pervin

& O. P. John (Eds.), Handbook of personality: Theory and research

(2nd ed., pp. 102–138). New York: Guilford Press.

Keltner, D., Gruenfeld, D. H., & Anderson, C. (2003). Power,

approach, and inhibition. Psychological Review, 110, 265–284.

Keltner, D., & Haidt, J. (2001). Social functions of emotions. In T. J.

Mayne & G. A. Bonanno (Eds.), Emotions: Current issues and

future directions. New York: Guilford Press.

Keltner, D., & Robinson, R. J. (1997). Defending the status quo:

Power and bias in social conflict. Personality and Social Psychology

Bulletin, 23, 1066–1077.

Kenny, D. A., & la Voie, L. (1985). Separating individual and group

effects. Journal of Personality and Social Psychology, 48, 339–348.

Kim, P. H. (1997). Strategic timing in group negotiations: The

implications of forced entry and forced exit for negotiators with

unequal power. Organizational Behavior and Human Decision

Processes, 71, 263–286.

Knutson, B. (1996). Facial expressions of emotion influence interper-

sonal trait inferences. Journal of Nonverbal Behavior, 20, 165–182.

C. Anderson, L.L. Thompson / Organizational Behavior and Human Decision Processes 95 (2004) 125–139 139

Kramer, R. M. (1996). Divergent realities and convergent disappoint-

ments in the hierarchical relation. In R. M. Kramer & T. R. Tyler

(Eds.), Trust in organizations: Frontiers of theory and research (pp.

216–245). Thousand Oaks, CA: Sage.

Kramer, R. M., Newton, E., & Pommerenke, P. L. (1993). Self-

enhancement biases and negotiator judgment: Effects of self-esteem

and mood. Organizational Behavior and Human Decision Processes,

56, 110–133.

Lawler, E. J. (1992). Power processes in bargaining. The Sociological

Quarterly, 33, 17–34.

Lawler, E. J., & Bacharach, S. B. (1987). Comparison of dependence

and punitive forms of power. Social Forces, 66, 446–462.

Lawler, E. J., Ford, R. S., & Blegen, M. A. (1988). Coercive capability

in conflict: A test of bilateral deterrence versus conflict spiral

theory. Social Psychology Quarterly, 51, 93–107.

Lawler, E. J., & Yoon, J. (1993). Power and the emergence of

commitment behavior in negotiated exchange. American Sociolog-

ical Review, 58, 465–481.

Lerner, J., & Keltner, D. (2001). Fear, anger and risk. Journal of

Personality and Social Psychology, 81, 146–159.

Levin, P. F., & Isen, A. M. (1975). Further studies on the effect of

feeling good on helping. Sociometry, 38, 141–147.

Lewicki, R., & Litterer, J. (1985). Negotiation. Homewood, IL: Irwin.

Lewin, K. (1951). In Dorwin Cartwright (Ed.), Field theory in social

science: Selected theoretical papers. Oxford, England: Harpers.

Mannix, E. A. (1993a). Organizations as resource dilemmas: The

effects of power balance on coalition formation in small groups.

Organizational Behavior and Human Decision Processes, 55, 1–22.

Mannix, E. A. (1993b). The influence of power, distribution norms and

task meeting structure on resource allocation in small group

negotiation. The International Journal of Conflict Management, 4,

5–23.

Mannix, E. A., & Neale, M. A. (1993). Power imbalance and the

pattern of exchange in dyadic negotiation. Group Decision and

Negotiation, 2, 119–133.

McAlister, L., Bazerman, M. H., & Fader, P. (1986). Power and goal

setting in channel negotiations. Journal of Marketing Research, 23,

228–236.

Moore, D. A., Kurtzberg, T. R., Thompson, L. L., & Morris, M. W.

(1999). Long and short routes to success in electronically mediated

negotiations: Group affiliations and good vibrations. Organiza-

tional Behavior & Human Decision Processess, 77, 22–43.

Morris, M. W., & Keltner, D. (2000). How emotions work: The social

functions of emotional expression in negotiations. In B. M. Staw &

R. I. Sutton (Eds.), Research in organizational behavior (Vol. 11,

pp. 1–50). Amsterdam: JAI.

Neale, M. A. (1997). New recruit. In J. M. Brett (Ed.), Teaching

materials for negotiations and decision making. Evanston, IL:

Northwestern University, Dispute Resolution Research Center.

Pfeffer, J. (1992). Managing with power: Politics and influence in

organizations. Boston, MA: Harvard Business School Press.

Pinkley, R. L., Neale, M. A., & Bennett, R. J. (1994). The impact of

alternatives to settlement in dyadic negotiation. Organizational

Behavior and Human Decision Processes, 57, 97–116.

Rekosh, J. H., & Feigenbaun, K. D. (1966). The necessity of mutual

trust for cooperative behavior in a two-person game. The Journal of

Social Psychology, 69, 149–154.

Rindskopf, D. (1984). Linear equality restrictions in regression and

loglinear models. Psychological Bulletin, 96, 597–603.

Rosenhan, D. L., Salovey, P., & Hargis, K. (1981). The joys of

helping: Focus of attention mediates the impact of positive affect

on altruism. Journal of Personality and Social Psychology, 40,

899–905.

Schurr, P. H. (1987). Effects of gain and loss decision frames on risky

purchase negotiations. Journal of Applied Social Psychology, 22,

386–407.

Sheldon, K. M., & Elliot, A. J. (1999). Goal striving, need satisfaction,

and longitudinal well-being: The self-concordance model. Journal

of Personality and Social Psychology, 76, 482–497.

Thibaut, J. W., & Kelley, H. H. (1959). The social psychology of

groups. New York: Wiley.

Thompson, L. (1990). An examination of naive and experienced

negotiators. Journal of Personality and Social Psychology, 59, 82–

90.

Thompson, L. L. (1998). The mind and heart of the negotiator. Upper

Saddle River, NJ: Prentice-Hall.

Tiedens, L. Z. (2001). Anger and advancement versus sadness and

subjugation: The effect of negative emotion expressions on social

status conferral. Journal of Personality and Social Psychology, 80,

86–94.

Tyler, T. R., & Degoey, P. (1996). Trust in organizational authorities:

The influence of motive attributions on willingness to accept

decisions. In R. M. Kramer & T. R. Tyler (Eds.), Trust in

organizations: Frontiers of theory and research (pp. 331–356).

Thousand Oaks, CA: Sage.

van den Bos, K., & Miedema, J. (2000). Toward understanding why

fairness matters: The influence of mortality salience on reactions to

procedural fairness. Journal of Personality and Social Psychology,

79, 355–366.

van den Bos, K., Wilke, H. A. M., & Lind, E. A. (1998). When do we

need procedural fairness? The role of trust in authority. Journal of

Personality and Social Psychology, 75, 1449–1458.

Van Kleef, G. A., De Dreu, K. W., & Manstead, A. S. R. (in press).

The interpersonal effects of anger and happiness in negotiations.

Watson, D., & Clark, L. A. (1984). Negative affectivity: The

disposition to experience aversive emotional states. Psychological

Bulletin, 96, 465–490.

Watson, D., & Clark, L. A. (1997). Extraversion and its positive

emotional core. In R. Hogan & J. A. Johnson (Eds.), Handbook of

personality psychology (pp. 767–793). San Diego, CA: Academic

Press.

Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and

validation of brief measures of positive and negative affect: The

PANAS scales. Journal of Personality and Social Psychology, 54,

1063–1070.

Watson, D., & Walker, L. M. (1996). The long-term stability and

predictive validity of trait measures of affect. Journal of Personality

and Social Psychology, 70, 567–577.