Now You See It, Now You Don't: Interests, Issues, and Psychological ...

Journal of Personality and Social Psychology 2010, Vol. 98, No. 5, 761–774

© 2010 American Psychological Association 0022-3514/10/$12.00 DOI: 10.1037/a0017879

Now You See It, Now You Don’t: Interests, Issues, and Psychological Distance in Integrative Negotiation Mauro Giacomantonio

Carsten K. W. De Dreu

University of Rome “Sapienza”

University of Amsterdam

Lucia Mannetti University of Rome “Sapienza” Negotiators are often advised to seek win–win agreements by focusing on interests (primary features) rather than issues (secondary features), but whether such advice is valid remains to be seen. Consistent with construal level theory (Y. Trope & N. Liberman, 2003), Experiments 1 and 2 show that negotiators focus on secondary features (issues) more than on primary features (interests) when psychological distance is low rather than high, and concomitant construal level is local and specific rather than global and abstract. Experiment 3 showed that high construal level promoted problem-solving behavior and therefore facilitated the achievement of win–win agreement, but only when integrative potential resided in underlying interests; when integrative potential resided in the issues, low construal level negotiators achieved higher joint outcomes. Thus, both low- and high-construal negotiators may achieve win–win agreements when such agreements require trade-offs at the level of issues, or at the level of underlying interests, respectively. Keywords: construal level, underlying interests, negotiation

uncover this so-called integrative potential and achieve such desirable “win–win” solutions (for reviews, see, e.g., Bazerman, Curhan, Moore, & Valley, 2000; Carnevale & Pruitt, 1992; De Dreu, Beersma, Steinel, & Van Kleef, 2007). Much of this work relates to an equally impressive volume of practitioner-oriented writings that offer exceedingly long lists of advice to further integrative agreement in either the business environment, in sociopolitical conflicts, or in diplomacy and trade negotiations (e.g., Fisher & Ury, 1981; Kelman, 2006; Lax & Sebenius, 1986; Walton & McKersie, 1965). In both these literatures, a recurrent theme is that to design integrative solutions, negotiators should explore and consider interests underlying issues (i.e., focus on reputation and security rather than on a contested portion of the Sinai Peninsula; Fisher & Ury, 1981; Rubin et al., 1994). Despite the common advice to explore and consider underlying interests, however, researchers know quite little about the psychological mechanisms that drive individuals toward or away from a focus on the issues versus interests—what is it that helps negotiators to orient their focus on underlying interests; why is it that issues are prominently considered yet interests ignored; when do negotiators look beyond the immediate and salient intrinsic issues value (e.g., commercial value) that are exchanged and instead consider the deeper underlying interests these issues putatively serve? The present research was designed to answer these questions. We invoke insights from construal level theory (CLT; Trope & Liberman, 2003) and propose that issues operate as secondary features in a negotiation situation that attract attention especially under a local construal level, when parties focus on the trees rather than the forest. Interests, in contrast, operate as primary features in a negotiation situation and attract attention especially under a global construal level. This proposition is the focus of the first two

To strike deals or to solve social conflict, individuals often resort to negotiation. Successful negotiation yielding mutually acceptable agreements may promote order and stability, increases feelings of self-efficacy, reduces the probability of future conflict, and stimulates economic prosperity (Rubin, Pruitt, & Kim, 1994). These potentially positive consequences are particularly likely when agreements are integrative, so that both parties achieve their aspirations to a greater extent than they would have through a simple fifty–fifty compromise. Consider, for example, the conflict in the mid-1970s between Egypt and Israel about the Israeli occupation of the Sinai Desert. Assisted by President Carter of the United States, President Sadat of Egypt and Prime Minister Begin of Israel were able to reach a mutually beneficial agreement: Instead of continuing to fight about how to divide the Sinai, parties addressed their respective underlying needs by returning the desert to its historical owner, Egypt, while simultaneously demilitarizing the area to provide Israel with its need for security (Fisher & Ury, 1981). Partly inspired by this and similar examples, a large volume of research has considered the conditions under which individuals

Mauro Giacomantonio and Lucia Mannetti, Department of Social and Developmental Psychology, University “Sapienza” of Rome, Rome, Italy; Carsten K. W. De Dreu, Department of Psychology, University of Amsterdam, Amsterdam, the Netherlands. We thank Luigi Leone for comments on an earlier version of the manuscript. Correspondence concerning this article should be addressed to Mauro Giacomantonio, Dipartimento di Psicologia dei Processi di Sviluppo e Socializzazione, Universita` di Roma “Sapienza”, Via dei Marsi 78, 00185, Rome, Italy. E-mail: [email protected] 761

762

GIACOMANTONIO, DE DREU, AND MANNETTI

studies reported herein. In the third study, we examined the implication that when integrative potential resides at the level of underlying interests, a global construal level with its concomitant focus on primary features such as interests leads to more integrative agreements; when integrative potential resides at the level of issues, a local construal level with its concomitant focus on secondary features, including issues, may promote integrative agreements more.

Issues and Interests in Negotiation Negotiations often involve one or more issues and one or more needs or concerns. People negotiate a higher salary (the issue) to satisfy their need for status (the underlying interest), and as buyers they try to bring down the price of the apartment (the issue) so they can afford a comfortable life and look good in the eyes of their friends (underlying interests). As these examples show, one single issue may serve one particular underlying interest (salary ⫽ status), or it may serve several underlying interests (lower price ⫽ comfortable life ⫹ looking good). Likewise, it is conceivable that several issues all derive from the same underlying need or that several issues derive from several different underlying needs. For example, a divorcing couple may negotiate the allotment of a fancy bike and an mp3 player. Each party may want both issues but for different reasons—the wife wants the bike because it reminds her of those good old days, whereas to the husband the bike is a vital means of transportation; and whereas the wife wants the mp3 player because the music helps her relax, the husband wants it because otherwise his former wife gets something she wants. The distinction between issues and interests yields two important insights. First, issues often have an attached value to which parties can agree—monetary worth, for example. Both the mp3 player and the fancy bike can be expressed in monetary terms, and doing so might make it easier to negotiate a deal that satisfies both parties (e.g., split the difference instead of having to haggle about who gets the expensive bike and who gets the less expensive mp3 player). This reasoning does not apply to needs and interests— rarely do parties agree on (or even think about) the monetary value of needs and interests. Needs and interests are “just there,” and some are more important than others. Sometimes parties seek to fulfill the same needs and interests (e.g., security, status); sometimes they are after quite different needs and interests (e.g., transportation and revenge vs. good memories and relaxation). The second insight is that whereas issues may be used to fulfill needs and interests, the reverse is not true. Issues can be transformed into monetary value, which can subsequently be used to satisfy specific needs and interests—the bike and the mp3 player may be sold at market value, the profit equally split among divorcees, with the wife using her share to buy an mp3 player, and the husband using his share to buy a bike. Obviously, the reverse does not work—the need for transportation cannot be used to acquire a bike, nor can a need for relaxation be sold at market value to finance an mp3 player. Put differently, although the importance of an issue depends also on the interest it serves (the mp3 player is more important for a person interested in music), the importance of an interest does not depend on the issues present in the negotiation. As such, interests drive the prioritization of issues, but issues do not drive the prioritization of interests; inter-

ests constitute superordinate, primary features of a negotiation, whereas issues are subordinate, secondary features. The two insights together imply that negotiators may not only distribute but also create (or destroy) value. Returning to the divorcing couple, the bike is worth more than the mp3 player (in economic terms), but given what they represent to both parties, value differences emerge. Whereas the husband may have no other means of transportation than the bike, the wife may satisfy her memories through other means than by looking at the bike—the bike is subjectively more valuable to the husband than to the wife. Likewise, the wife cannot easily get her music and concomitant relaxation elsewhere, whereas the husband may mellow down a bit or find other ways to get back at his former wife—to him, the mp3 player represents less subjective value than to her. Now, when husband and wife focus on their needs and interests, they may uncover creative agreements, like the one in which he gets the bike and she the mp3 player. Deals that integrate and reconcile underlying interests are usually of higher quality and more satisfying than deals in which parties split the (economic value) difference on both issues. Because a focus on underlying interests rather than on overt issues is so helpful in crafting high-quality agreements, negotiation research and advice often is about ways to bring out underlying interests, to find ways to reconcile interests rather than divide issues, and so on. As mentioned, researchers know relatively little about the psychological mechanisms that drive people to focus on either issues or interests, and we suspect that construal level is one such critical mechanism. This we elaborate on in the next section.

CLT According to CLT (Liberman & Trope, 1998; Liberman, Trope, & Stephan, 2007; Trope & Liberman, 2003), the psychological distance from an object or event influences the way information about that object or event is construed and, consequently, how that object or event is judged and evaluated. Psychological distance can be rooted in temporal distance (far or close in the future or past), spatial distance (far or close in space), social distance (outgroup vs. ingroup membership; strangers vs. friends), and hypotheticality (hypothetical vs. real). In theory, all four forms of distance influence mental functioning. When an object or event is psychologically distant (e.g., in another geographical continent), it cannot be experienced directly, and the lack of available information leads to high construal. The mental representation of the object is abstract, simple, and decontextualized. The emerging representation is similar to a schema that extracts the gist and organizes the knowledge integrating different aspects of the object or event around simple and general elements— one focuses on the forest rather than on the trees (Schul, 1983; Smith & Trope, 2006). But when an object or event is psychologically close (e.g., in the short-term future), it can be experienced more directly, and contextual and specific information is available. A more detailed, less schematic, and more contextualized representation is formed— one focuses on the trees rather than on the forest (Fo¨rster, Friedman, & Liberman, 2004). Thus, reading a book can be construed at an abstract level and in terms of a superordinate goal (e.g., gaining knowledge), or at a more concrete level and, in terms of subordinate means, implied in the action (e.g., following printed lines; Liberman & Trope, 1998;

CONSTRUAL LEVEL IN NEGOTIATION

Vallacher & Wegner, 1989). Or, applied to the present context, negotiation can be construed at an abstract level and in terms of a superordinate goal (e.g., meeting fundamental needs), or at a more concrete level and, in terms of subordinate means, implied in the action (e.g., exchanging demands and concessions on particular issues). The level of construal has important consequences for cognitive, motivational, and behavioral tendencies. When primed with a distant future, people categorize a set of items in a small number of broad categories; when primed with a near future, the same set of items is categorized in a large number of small categories (Liberman, Sagristano, & Trope, 2002). Fo¨rster et al. (2004) showed that temporal distance boosts creative performance, and Fo¨rster (2009) uncovered that people with high construal focus on similarities rather than on dissimilarities, whereas individuals with low construal tend toward the reverse. One particularly interesting feature of psychological distance that has received extensive attention is that it influences the relative weight of superordinate goals versus subordinate means (Sagristano, Trope, & Liberman, 2002; Trope & Liberman, 2000). Put differently, psychological distance determines whether the emphasis is on primary features (the forest; learning new knowledge; fundamental needs) or on secondary features (the trees; following print lines; issues in a negotiation). For example, Sagristano et al. (2002) showed that participants with a high level of construal preferred a gamble with a high payoff and low probability, whereas participants with a low construal level preferred gambles with a low payoff and high probability. These different preferences were due to the fact that probability is subordinate to payoff in choosing a gamble. Similarly, because cons are subordinate to pros in considering an action, participants generated more pro arguments than con arguments toward an activity planned in the distant future. The reverse was shown for an activity planned in the near future (Eyal, Liberman, Trope, & Walther, 2004; see also Fujita, Trope, Liberman, & Levin-Sagi, 2006).

Psychological Distance in Negotiation Psychological distance and concomitant construal levels may influence negotiation because of its influence on whether primary or secondary features of the negotiation are attended to. As explained, needs and interests represent superordinate, primary features in negotiation, and issues represent subordinate, secondary features. Following CLT, negotiators with high construal levels should focus on primary features and are directed toward underlying interests more than negotiators with low construal levels. Conversely, those with low construal levels should focus on secondary features of the negotiation and are directed toward specific demands and issues more than those with high construal levels. When the negotiation task is structured so that underlying interests can be integrated, high-construal level negotiators should be more likely to uncover this integrative potential and, all other things being equal, more likely to achieve mutually beneficial, integrative agreements. There is some initial support for this general idea. Henderson, Trope, and Carnevale (2006) conducted a series of studies showing that high temporal distance promoted a more global and structured approach toward the negotiation. High temporal distance led to an increased preference for simultaneous issue consideration, whereas

763

a piecemeal, fragmented approach to issues was preferred in the low-temporal distance condition. This “package deal” approach in the temporally distant condition fostered integrative agreements more than the “sequential deal” approach in the temporally near condition (e.g., Weingart, Bennett, & Brett, 1993). However, although not inconsistent with the idea that psychological distance orients toward primary features of the negotiation, a viable alternative explanation for the results obtained by Henderson and colleagues is that temporal distance in their studies fostered a cooperative mind-set and concomitant prosocial motivation. Indeed, there is evidence that psychological distance co-varies with a less competitive and more cooperative motivational orientation (De Dreu & Nijstad, 2008) and that a cooperative approach fosters “package deal” negotiation and integrative agreements (De Dreu, Weingart, & Kwon, 2000; Weingart et al., 1993).

Overview of the Present Study To further understanding of the psychological processes that orient negotiators to focus on interests versus issues, and to provide a test of CLT that high psychological distance promotes abstract construal and a focus on primary rather than on secondary features, we conducted three experiments. In the first two experiments, we tested the proposition that under high (low) psychological distance, negotiators focus on interests (issues) when making judgments and decisions about their negotiation. In Experiment 1, we examined fixed-pie perceptions—the tendency for negotiators to assume own and other’s interests to be diametrically opposed (Thompson & Hastie, 1990). We created a negotiation task in which a focus on issues would confirm such a fixed-pie perception, but a focus on underlying interests would lead people to uncover their fixed-pie perception to be erroneous. We measured fixed-pie perceptions prior to the negotiation and then again after participants had received information about their counterpart’s issue chart (or not, in the control condition). We predicted stronger revisions of fixed-pie perceptions when negotiators had a high rather than a low level of construal. In Experiment 2, we tested whether construal level influences the willingness to accept an interest-based offer as opposed to an issue-based offer. Offers were created so that they would be especially attractive when evaluated in terms of underlying interests or, alternatively, in terms of specific issues. We predicted greater satisfaction with, and willingness to accept, an interest-based offer under high rather than low construal level, and greater satisfaction with and willingness to accept an issue-based offer under low rather than high construal level. In Experiment 3, we engaged participants in actual face-toface negotiation to further our understanding of the behavioral implications of psychological distance and its concomitant focus on primary versus secondary features of the negotiation. This experiment is introduced in detail after we have reported Experiments 1 and 2.

Experiment 1 Method Design and participants. The design was a 2 ⫻ 2 factorial, with construal level (high vs. low) and information about other’s underlying interests (present vs. absent) as between-subjects vari-

764


ables. The main dependent variable was the (fixed-pie) perception of underlying interests before and after the manipulation of information about the counterpart. Eighty-one students of the University of Rome “Sapienza” participated on a voluntary basis and were randomly assigned to experimental conditions. Negotiation task. An adapted version of the task developed by Tro¨tschel and Gollwitzer (2007) was used. It is explicitly designed to distinguish between preferences for underlying interests and preferences for negotiation issues. Participants in this task negotiate in dyads, with one party being in the role of recruiter and the other in the role of candidate. In the present study, participants were tested individually, and all participants were in the role of the candidate. As can be seen in the top half of Table 1, there were eight issues to be negotiated: (a) annual salary, (b) location where the candidate will work, (c) annual raise, (d) moving expenses, (e) percentage of health insurance coverage, (f) job assignment, (g) weeks of vacation per year, (h) department destination. Three possible agreements were available for each issue (e.g., rate of annual raise: 6%, 9%, 12%), and each agreement option had a value to the participant ranging from 0 to 60 points (annual raise, moving expenses, vacation and department destination) or from 0 to 120 points (salary, location, health insurance, job assignment). Furthermore, participants read that these eight issues were grouped into four superordinate categories representing their underlying interests. Vacation and health insurance grouped into the category benefit; salary and annual raise grouped into the category money; location and moving expenses grouped into the category moving costs; and job assignment and department destination grouped into the category job type. Participants were told that “money” was the most important category to them (4 points), followed by “moving costs” (3 points), “benefit” (2 points), and “job type” (1 point). Thus, the overall value of an agreement was calculated by multiplying the points obtained on a specific issue, with the points associated with the category to which that issue belonged (e.g., an agreement of 60% of health insurance had the overall value of [60 ⫻ 2] ⫽ 120 points). To explain how to compute the overall outcome of an agreement, participants were provided with examples of the worst and the best agreement they could get. For instance, their best outcome was (120 ⫻ 4) ⫹ (120 ⫻ 3) ⫹ (60 ⫻ 4) ⫹ (60 ⫻ 3) ⫹ (120 ⫻ 2) ⫹ (120 ⫻ 1) ⫹ (60 ⫻ 2) ⫹ (60 ⫻ 1) ⫽ 1,800 points.1 In contrast to previous tasks used to study integrative negotiation (e.g., Pruitt & Lewis, 1975; for a discussion, see De Dreu & Carnevale, 2003), which clearly specified preferences across issues, the present task was intentionally designed so that negotiators could attribute preferences to each issue according to three different interpretations. First, a person could just focus on the interest values. As can be seen in Table 1, annual raise serves a more valuable underlying interest than location (4 vs. 3), thus resulting in an increased priority. Alternatively, however, a person could focus on the intrinsic value of the issue. In this case, location appears twice as valuable as annual raise (120 vs. 60). Third, a person could consider the joint function of the two sources of value. In this case, the person should prefer the location (360) versus annual raise (240). Put differently, the attractiveness of an issue derives independently from the intrinsic value of the issue itself, and the underlying interest it serves. This feature makes the task particularly suitable to test whether negotiators place more emphasis on interests or issues— different emphasis placed on

issues or interests produce different perceptions of the task, which should lead to different behaviors (we return to this especially in Experiments 2 and 3). Procedure and manipulation of independent variables. The procedure closely followed the one used by De Dreu, Koole, and Steinel (2000, Experiment 2). Upon arrival in the laboratory, participants received a booklet containing instructions regarding the negotiation task, background information, and the task intended to manipulate construal level. Construal level was manipulated using a procedure similar to the one used by Liberman et al. (2002; see also Fo¨rster et al., 2004). Participants were asked to write down 10 activities or events in which they could have been involved in the near future (Monday next week) or in the distant future (Monday next year). Initial work in our laboratory has shown that this manipulation induces a narrow versus a more global construal level and influences negotiation processes and outcomes in a task similar to the one used here (De Dreu, Giacomantonio, Shalvi, & Sligte, 2009). Upon completion of the construal level manipulation, participants filled out the Time 1 measure of fixed-pie perception. This Time 1 measure served as a baseline to verify that in the absence of any information, construal level does not influence fixed-pie perceptions. Following the baseline assessment, information availability was manipulated by providing half the participants with a chart containing information on their counterpart’s payoff on both issues and underlying interests. The counterpart’s issue chart (see Table 1, bottom half) was structured so that integrative potential exclusively resided at the underlying interests. As can be seen in Table 1, the candidate (the participant; top chart) and his or her recruiter (bottom chart) had perfectly opposed preferences on the issues but differed in the underlying interests—whereas participants valued “money” most and “job type” least, the recruiter valued “money” least and “job type” most. In other words, participants would be confirmed in their fixed-pie assumption when they either did not receive or process information on their counterpart or focused on issues only. They would revise their initial fixed-pie perception only when they would receive information on their counterpart and process the information on underlying interests rather than issues. Participants in the information availability condition were allowed to study this chart for 2 min, after which it was taken away and replaced by a 5-min filler task unrelated to the present experiment. In the no-information condition, participants were not 1 It might be helpful to note that in designing this task, we were inspired by the logic used in classical tasks adopted to study lay decisions under uncertainty. Imagine, for example, a person asked to express his or her preference about one of the following bets: (a) p ⫽ .3 to get an outcome of €100 with p ⫽ .7 to get €0; (2) p ⫽ .4 to get €50 with p ⫽ .6 to get €0. If we assume that probability and payoffs do not always combine multiplicatively in the perception of the lay decision maker, and therefore, under certain conditions, these parameters have asymmetric weight in determining the attractiveness of gambles (Sagristano et al. 2002), then a person could rely (mainly) on probability in making a choice. In this case, the second bet will be the most likely to be chosen ( p ⫽ .4 vs. p ⫽ .3 to get something). In contrast, a person could based the choice on the payoff, thus assigning greater value to the first bet (€100 vs. €50). Finally, a person could base the choice on the expected value (EV). In this case, the first bet is the most opportune (EV ⫽ 30 vs. 20).

80% (10) 60% (60) 40% (120) 100% (00) 80% (30) 60% (60) 12% (00) 9% (30) 6% (60) Torino (10) Milano (60) Genova (120) 30.000 (10) 28.000 (60) 26.000 (120)

Note. Assign. ⫽ Assignment; Dep. ⫽ Department. Numbers in parentheses refer to points earned by the participant on issues. To obtain the overall value earned, these points should be multiplied by the interests’ values.

(very low) (low) (high) (very high) Money Moving Benefit Job type Dep. A (00) Dep. B (30) Dep. C (60) 5 weeks (00) 4 weeks (30) 3 weeks (60)

Department destination Vacation Job assignment Health insurance Location Annual salary

Annual raise

Moving expenses

Assign. A (10) Assign. B (60) Assign. C (120)

Issues Category

Recruiter interests charts Recruiter issue charts

Salary, Annual Raise Location, Moving Expenses Vacation, Health Insurance Job Assign., Dep. Destination

1 2 3 4

Value

(very high) (high) (low) (very low) Salary, Annual Raise Location, Moving Expenses Vacation, Health Insurance Job Assign., Dep. Destination Money Moving Benefit Job type Dep. A (60) Dep. B (30) Dep. C (00) 5 weeks (60) 4 weeks (30) 3 weeks (00) 80% (120) 60% (60) 40% (10) Torino (120) Milano (60) Genova (10) 30.000 (120) 28.000 (60) 26.000 (10)

12% (60) 9% (30) 6% (00)

100% (60) 80% (30) 60% (00)

Assign. A (120) Assign. B (60) Assign. C (10)

Issues Vacation Job assignment Health insurance Moving expenses Annual raise Location Annual salary

Candidate issue charts

Table 1 Issue and Interests Charts for the Candidate (Participant’s Role) and Recruiter (Experiment 1)

Department destination

Category

Candidate interests charts

4 3 2 1

Value


765

shown any information about their counterpart’s issue chart, and after the Time 1 measure of fixed-pie perceptions, he or she proceeded immediately with the filler task. In both conditions, after the filler task, participants were asked to complete the Time 2 fixed-pie measure and also answered several questions (see the Dependent variables section). They were then thanked and fully debriefed. Dependent variables. Fixed-pie perceptions were measured before and after the presentation of the counterpart’s payoff chart, using the same procedure used earlier by De Dreu et al. (2000; see also Thompson & Hastie, 1990). Both times participants were given an empty chart regarding the counterpart’s underlying interests and were asked to fill in the points and weights they thought their counterpart had been given. Fixed-pie perceptions were calculated by summing the absolute difference between the rank order attributed to the partner and the participant’s own underlying interests. If the participant attributed to the counterpart exactly the same ranking he or she had, then the score would be 0 (i.e., [4 ⫺ 4] ⫹ [3 ⫺ 3] ⫹ [2 ⫺ 2] ⫹ [1 ⫺ 1] ⫽ 0), thus reflecting a perfect fixed-pie perception. In contrast, if participants had a perfectly accurate perception (i.e., understood the integrative potential in the task), then their score would be 8 (i.e., [4 ⫺ 1] ⫹ [3 ⫺ 2] ⫹ [2 ⫺ 3] ⫹ [1 ⫺ 4] ⫽ 8). Note that a high score of 8 could also be achieved when perceptions are not perfectly accurate (e.g., [4 ⫺ 2] ⫹ [3 ⫺ 1] ⫹ [2 ⫺ 4] ⫹ [1 ⫺ 3] ⫽ 8). Therefore, before summing, absolute differences were squared so that more weight was assigned to higher absolute differences that are more indicative of integrative perception. Put differently, a lower score is indicative of a fixed-pie perception; a higher score is indicative of an accurate (integrative potential) understanding of the task. To check the adequacy of the construal level manipulation, participants were asked to rate the events, actions, or feelings they experienced during the manipulation task using four semantic differential items on scales ranging from 1 to 7 (Burrus & Roese, 2006): (a) important–not important, (b) high priority–low priority, (c) central in life–secondary in life, (d) long-term goal–short-term goal. Ratings were averaged (␣ ⫽ .68), and lower (higher) scores indicate a lower (higher) level of construal. Finally, participants were asked how difficult it was to complete the perception measures with one item: “How difficult was it to fill in the empty charts?” (ranging from 1 [not at all] to 6 [extremely]).

Results and Discussion Manipulations checks. A 2 ⫻ 2 analysis of variance (ANOVA; Construal Level ⫻ Information) showed that the highconstrual condition produced higher ratings on the manipulation check scale (M ⫽ 4.9) than the low-construal condition (M ⫽ 4.3), F(1, 77) ⫽ 4.31, p ⬍ .05, ␩2 ⫽ .04. No other effects were significant. Two 2 ⫻ 2 ANOVAs on the self-reported difficulty in filling the fixed pie at Time 1 and Time 2 revealed no effects (all Fs ⬍ 1). Fixed-pie perceptions. We submitted fixed-pie perceptions assessed at Time 1 and Time 2 to a 2 ⫻ 2 ⫻ 2 (Construal Level ⫻ Information ⫻ Time of Measurement) ANOVA, with construal level and information as between-subjects factors and time of measurement as a within-subjects factor. Results revealed a main effect for time of measurement, F(1, 77) ⫽ 34.62, p ⬍ .001, ␩2 ⫽ .31; an interaction between time and construal level, F(1, 77) ⫽


766

5.91, p ⬍ .05, ␩2 ⫽ .07; and an interaction between time and information, F(1, 77) ⫽ 30.52, p ⬍ .001, ␩2 ⫽ .28. These effects were all qualified by the predicted three-way interaction among time, construal level, and information, F(1, 77) ⫽ 5.86, p ⬍ .05, ␩2 ⫽ .07. Figure 1 shows that in the no-information condition, construal level had no effects on perceptual accuracy at Time 1 or at Time 2. In the information available condition, participants were more accurate at Time 2 than at Time 1, but this effect was considerably stronger in the high-construal level condition, F(1, 77) ⫽ 72.2, p ⬍ .001, ␩2 ⫽ .48, than in the low-construal level condition, F(1, 77) ⫽ 9.61, p ⬍ .01, ␩2 ⫽ .11. As a result, participants in the high-construal level condition were more accurate at Time 2 than those in the low-construal level condition, F(1, 77) ⫽ 3.77, p ⬍ .06 (approached significance).

earlier findings and showed that negotiators provided with information about their counterpart’s preferences revised their fixed-pie perceptions and became more accurate than those not provided with such information. More important, however, we found that participants with a high construal level benefitted more from the information provided than those with a low construal level. That is, after information had been given to them, participants became more accurate especially when they had a high rather than a low construal level. Although findings thus far are consistent with our general predictions, they tell only half of the story. Whereas we showed that low-construal level participants were less concerned about underlying interests than high-construal level participants, this does not necessarily mean that low-construal level participants are more concerned about low-level features of the situation such as negotiation issues. To examine this, in Experiment 2, we investigated the extent to which participants were willing to accept an offer that met their preferences on underlying interests (henceforth, interestbased offer) or, in the other condition, met preferences on issues (henceforth, issue-based offer). We ensured that the objective value of both offers was the same, but among those focusing on

Experiment 2 Providing negotiators with information about their counterpart’s preferences and priorities has been shown to increase perceptual accuracy (e.g., De Dreu et al., 2000) and facilitates the development of integrative, mutually beneficial agreements (e.g., Thompson, 1991). The results of Experiment 1 are consistent with these

Information Available 18

Accuracy

16 14

High Construal Level

12 10 8

Low Construal Level

6 4 2 0

Time 1

Time 2

Fixed Pie Measure

No Information Available 18

Accuracy

16 14

High Construal Level

12 10 8

Low Construal Level

6 4 2 0

Time 1

Time 2

Fixed Pie Measure

Figure 1. Perceived accuracy of the counterpart’s preferences on underlying interests as a function of time of measurement, information, and construal level in Experiment 1.

CONSTRUAL LEVEL IN NEGOTIATION 1 (low) 2 (medium) 3 (high) Salary, Annual Raise Location Vacation, Insurance

The issue-based offer (top half of the table) and the interest-based offer (bottom half of the table) appear in boldface type. Note.

5.000 (60) 4.000 (45) 3.000 (30) 2.000 (15) 1.000 (00) 3 weeks (60) 2.5 weeks (45) 2 weeks (30) 1.5 weeks (15) 1 week (00) 30.000 (120) 28.000 (90) 26.000 (60) 24.000 (30) 22.000 (00)

15% (60) 12% (45) 9% (30) 6% (15) 3% (00)

100% (120) 80% (90) 60% (60) 40% (30) 20% (00)

Interest-based offer Roma (80) Milano (60) Napoli (40) Torino (20) Bolzano (00)

Money Destination Benefit

1 (low) 2 (medium) 3 (high) Salary, Annual Raise Location Vacation, Insurance Money Destination Benefit 5.000 (60) 4.000 (45) 3.000 (30) 2.000 (15) 1.000 (00) Issue-based offer Roma (80) Milano (60) Napoli (40) Torino (20) Bolzano (00) 100% (120) 80% (90) 60% (60) 40% (30) 20% (00) 3 weeks (60) 2.5 weeks (45) 2 weeks (30) 1.5 weeks (15) 1 week (00) 30.000 (120) 28.000 (90) 26.000 (60) 24.000 (30) 22.000 (00)

15% (60) 12% (45) 9% (30) 6% (15) 3% (00)

Category Merit bonus Location Insurance Annual raise Vacation Salary

Candidate issue charts

Table 2 Issue and Interest Charts for the Candidate (Participant’s Role) in Experiment 2

Design and participants. The design was a 2 ⫻ 3 factorial, with offer (interest based vs. issue based) and construal level (high vs. low vs. control) as between-subjects factors (the no-construal level control condition was added to explore default construal level among negotiators). The main dependent variable was the participant’s satisfaction with the offer. Ninety-four students at the University “Sapienza” of Rome participated in the experiment and were assigned randomly to experimental conditions. Task, procedure, and manipulation of independent variables. The procedure, the task structure, and the manipulation of construal level were identical to that of Experiment 1, except that there were five rather than three levels of agreement per issue and six rather than eight issues. This enabled the manipulation of the offer. Specifically, participants received the opening offer of their recruiter (who was presumably sitting in one of the other cubicles in the laboratory), and they were asked to indicate their willingness to accept or reject it. In the interest-based offer condition, the proposal maximized the value to the candidate on the issues belonging to the most important underlying interest (benefit). More specifically, the offer provided an outcome higher than compromise on vacation and insurance. At the same time, participants got a low value (below compromise) on the issues belonging to the low-value interests (money, salary, annual raise, merit bonus). They received a compromise offer with regard to the destination issue (see Table 2). In the issue-based offer condition, the proposal had an opposite “shape” meant to maximize the value on the issues under money (the least important underlying interest) and minimized the value on the issues under benefit (the most important underlying interest). Specifically, the proposal provided an outcome above compromise for salary, annual raise, merit bonus, and below compromise for vacation and insurance. Like in the interest-based offer condition, destination was characterized by a compromise offer. Although the interest-based and the issue-based offers differed in the extent to which they reflected underlying interests or issues, both had the same overall value of 500 points: For the interestbased offer: ([45 ⫻ 3] ⫹ [90 ⫻ 3] ⫹ [40 ⫻ 2] ⫹ 15 ⫽ 500 points); for the issue-based offer: (120 ⫹ [30 ⫻ 3] ⫹ 60 ⫹ [30 ⫻ 3] ⫹ [40 ⫻ 2] ⫹ 60 ⫽ 500 points). However, when participants considered only the value provided by the offer on issues without little or no concern for underlying interests, the two offers looked very different. In that case, the issue-based offer appears to be much better (120 ⫹ 30 ⫹ 60 ⫹ 30 ⫹ 40 ⫹ 60 ⫽ 340 points) than the interest-based offer (45 ⫹ 90 ⫹ 40 ⫹ 15 ⫽ 190 points). Conversely, if participants focus on interests rather than on issues, the interest-based offer provided greater value (i.e., ([45 ⫻ 3] ⫹ [90 ⫻ 3] ⫽ 405 points vs. [30 ⫻ 3] ⫹ [30 ⫻ 3] ⫽ 180 points on the most valuable issues with reference to underlying interests). Dependent variables. The main dependent variable was satisfaction with the offer, measured as in Henderson et al. (2006, Study 2). Sample items include “How unsatisfied/satisfied would you be with this proposal (ranging from 1 [very unsatisfied] to 6

Candidate interest charts

Method

Issues

Value

interests (presumably in the high-construal level condition), the interest-based offer should appear more attractive, whereas to those focusing on issues (presumably in the low-construal level condition), the issue-based offer should appear more attractive.

767


768

[very satisfied]) and “How unlikely/likely would you be to reject this proposal?” (ranging from 1 [very unlikely] to 6 [very likely]). Ratings were averaged (␣ ⫽ .80). The manipulation check of the construal level manipulation was the same as in Experiment 1 (␣ ⫽ .73). The control group did not do any task, and, consequently, participants in this condition did not fill out the manipulation check. To check the adequacy of the offer manipulation, participants were asked to rate “how coherent was your partner’s proposal with your underlying interests (money, benefit, location)?” (1 ⫽ totally incoherent, 6 ⫽ totally coherent).

Results and Discussion Manipulations checks and information-processing motivation. A 2 ⫻ 2 ANOVA (Construal Level ⫻ Type of Offer) showed that the high-construal condition produced higher ratings on the manipulation check scale than the low-construal condition (M ⫽ 5.0 vs. M ⫽ 4.2), F(1, 60) ⫽ 4.54, p ⬍ .05, ␩2 ⫽ .07. No other effects were significant. A 3 ⫻ 2 ANOVA further showed that compared with participants in the issue-based offer condition, those in the interest-based offer condition perceived the offer as more coherent with their underlying interests (M ⫽ 4.0 vs. M ⫽ 3.3), F(1, 88) ⫽ 6.78, p ⬍ .01, ␩2 ⫽ .072. No effects involving construal level were significant. Satisfaction with offer. A 3 ⫻ 2 ANOVA revealed the predicted interaction between construal level and type of offer, F(2, 87) ⫽ 7.33, p ⬍ .01, ␩2 ⫽ .14. Figure 2 shows that participants in the high-construal condition preferred interest-based offers more than issue-based offers, F(1, 87) ⫽ 7.31, p ⬍ .01, ␩2 ⫽ .078. The opposite pattern emerged in the low-construal level condition, in which the issue-based offer was preferred more than the interestbased offer, F(1, 87) ⫽ 6.93, p ⫽ .01, ␩2 ⫽ .074. In the control condition, ratings of satisfaction did not change as a function of the offer presented to participants (M ⫽ 3.80 vs. M ⫽ 3.48; F ⬍ 1).

Experiment 3 The first two experiments showed that negotiators focus on issues when they have a low construal level and on underlying interests when they have a high construal level. In Experiment 1, this showed up in more adequate revisions of fixed-pie percep-

tions. In Experiment 2, this showed up in a stronger willingness to accept an interest-based offer among high-construal level participants, versus a stronger willingness to accept an issue-based offer among low-construal level participants. This finding is consistent with our hypothesis that high-construal level negotiators focus on interests and that low-construal level negotiators focus on issues. Furthermore, this finding is difficult to explain in terms of higher cooperative motivation among high-construal level negotiators because higher cooperative motivation should have produced greater satisfaction with and willingness to accept the counterpart’s offer regardless whether the offer is congruent with issues or with interests. The goal of Experiment 3 was twofold. First, we wanted to test the implications of our findings in Experiments 1 and 2 for dyadic-level performance. In Experiment 3, we engaged dyads in an interpersonal negotiation structured along the lines of the task used in Experiments 1 and 2. We manipulated construal level at the dyadic level (i.e., bringing both parties under the same construal level) and analyzed the quality of agreements dyads reached. Second, although the findings of the second experiment were difficult to explain in terms of enhanced cooperative motivation under high rather than low construal level, we wanted to definitively rule out this alternative explanation and to provide more direct evidence for our suspicion that negotiation parties focus on interests (issues) when they have a high (low) construal level. We achieved these goals by varying the level at which integrative potential resides. In one condition, it resided at the level of underlying interests and not at the level of issues. In the other condition, it resided at the level of issues and not at the level of underlying interests. If effects thus far were due to enhanced cooperative motivation among high-construal level negotiators, it should not matter at which level integrative potential resides, and we should only find a main effect of construal level on integrative agreements. But if our reasoning holds, high-construal level negotiators focusing on underlying interests should engage in more problem solving and achieve more integrative agreements when integrative potential resides at the level of interests rather than issues; lowconstrual level negotiators, in contrast, should engage in more problem solving and achieve more integrative agreements when integrative potential resides at the level of issues rather than interests.

Satisfaction with the Offer

Method 6 5 4 3

Interest-based Offer

2

Issue-based Offer

1 0

High

Low

Control

Construal Level Figure 2. Satisfaction with the offer as a function of construal level and type of offer in Experiment 2.

Design and participants. The design was a 2 ⫻ 2 factorial design with construal level (high vs. low) and location of integrative potential (underlying interests vs. issues) as between-dyads factors. The main dependent variable was joint outcome. One hundred sixty four participants from the University of Rome “Sapienza” were randomly assigned to dyads with the restriction that they were not acquainted prior to the experiment. Dyads were also randomly assigned to the four experimental conditions. Task and procedure. Upon arrival at the laboratory, participants received a folder with the instructions for an upcoming buyer–seller negotiation concerning the purchase of a car. The following five issues were described to participants: (a) price; (b) guarantee; (c) financing; (d) included accessories; (e) delivery time. Similar to the previous experiments, participants were told that the five issues were grouped into three superordinate under-


lying interests: (a) money (including price and financing); b) timing (including delivery time); (c) facilities (including guarantee and accessories). As in the previous studies, the overall value of an agreement was given by a joint function of the agreement value on a specific issue and the importance of the corresponding underlying interest. To enhance their understanding of the task, participants were provided with several examples of negotiation outcomes. Independent variables. Construal level was manipulated by asking participants, prior to the negotiation, to read a brief passage explaining how to think increasingly abstractly (see Freitas, Gollwitzer, & Trope, 2004, and Fujita et al., 2006, for the full version of the passage). Participants in the high-construal level condition were asked to consider why they would maintain and improve their health. After listing a reason, they were subsequently asked why they would pursue this goal. In total, participants had to answer four consecutive questions. A typical example of what participants wrote was: “Why to improve and maintain health?” (Step 1) “To look pretty”; (Step 2) “To improve my social life”; (Step 3) “To get a partner”; (Step 4) “To be happy.” In the low-construal level condition, participants first read the brief passage about how to think increasingly concretely (see Freitas et al., 2004, for the full version of the passage). Then they were asked to consider how they would maintain and improve their health. After listing a mean to reach this goal, they were asked to write down how they would do it. In total, participants had to answer four consecutive questions. A typical example of what participants reported was: “How to improve and maintain health?” (Step 1) “Avoiding junk food”; (Step 2) “Preparing my meal myself”; (Step 3) “Cooking some pasta”; (Step 4) “Putting pasta in the boiling water” (we tend to ignore culturally embedded beliefs about what is and is not healthy food). Participants in the control condition skipped this task, and they were involved in the negotiation just after the role instructions. Location of integrative potential was manipulated so that in the “integrative interests” condition, negotiators were provided with different rankings on underlying interests and diametrically opposed preferences on issues (money was seller’s most important interest, whereas facilities was buyer’s most important interest). Therefore, if negotiators exclusively focused on issues rather than on underlying interests, the task appeared as a zero-sum game because preferences about the five issues at the bargaining table were diametrically opposed. Consider, for example, the subjective value of an agreement of €11.000 (about U.S. $16,000) on price (see Table 3). This is a compromise agreement that, in case negotiators were not concerned with underlying interests, provided a value of 90 points for both parties. In case underlying interests were taken into consideration, the same agreement was more valuable to the seller (90 ⫻ 3 ⫽ 270 points) than to the buyer (90 ⫻ 1 ⫽ 90 points). Each negotiator could reach an overall outcome between 190 and 950 ([150 ⫻ 3] ⫹ [50 ⫻ 1] ⫹ [50 ⫻ 3] ⫹ [150 ⫻ 1] ⫹ [75 ⫻ 2]). A fully integrative agreement could be reached when the seller gets his or her way on price as well as on interest (money), and buyer gets his or her way on warrantee and optional (facilities) while delivery was split: ([150 ⫻ 3] ⫹ [10 ⫻ 1] ⫹ [50 ⫻ 3] ⫹ [30 ⫻ 1] ⫹ [45 ⫻ 2]) ⫹ ([30 ⫻ 1] ⫹ [50 ⫻ 3] ⫹ [10 ⫻ 1] ⫹ [150 ⫻ 3] ⫹ [45 ⫻ 2] ⫽ 1,460). In the “integrative issues” condition, participants were provided with identical rankings on underlying interests, but preferences on

769

issues were asymmetrical (price and optional were seller’s most important issues, whereas warrantee and interests were buyer’s most important issues). In this case, if negotiators exclusively focused on interests rather than on issues, the task appeared as a zero-sum game because preferences about the three issues at the bargaining table were diametrically opposed. Also in this condition, each negotiator could reach an overall outcome between 190 and 950. In contrast with the other condition, a fully integrative agreement could be reached when the seller gets his or her way on price and optional, and buyer gets his or her way on warrantee and interest while delivery was split: ([150 ⫻ 3] ⫹ [10 ⫻ 1] ⫹ [10 ⫻ 3] ⫹ [150 ⫻ 1] ⫹ [45 ⫻ 2]) ⫹ ([10 ⫻ 3] ⫹ [150 ⫻ 1] ⫹ [150 ⫻ 3] ⫹ [10 ⫻ 1] ⫹ [45 ⫻ 2] ⫽ 1,460). To sum up, in the integrative underlying interests condition, high joint outcomes can be achieved by making concessions mainly based on underlying interests. Conversely, in the integrative issues condition, high joint outcomes can be achieved by yielding appropriately on issues with little or no regard for underlying interests. Individuals did not receive their counterpart’s issue chart and were told not to exchange these issue charts during the negotiation. They were unaware of the integrative potential in the task, and through negotiation and the exchange of information, they had to uncover possibilities for trade-off and high joint gain (see De Dreu, Beersma, Stroebe, & Euwema, 2006, for a similar procedure). Dependent variables. Joint outcome (the main dependent variable) was obtained by summing the outcomes reached by the candidate and the recruiter. As mentioned previously, a postnegotiation questionnaire (De Dreu, Evers, Beersma, Kluwer, & Nauta, 2001) was included to assess cooperative problem solving with four items such as “I tried to find solutions that satisfy my own and the other party’s interests” and “I emphasized that we had to find a mutually optimal solution (1 ⫽ not at all, 6 ⫽ very much; ␣ ⫽ .76).

Results and Discussion Impasses. Eighteen out of the 82 dyads did not reach an agreement. The impasses were distributed as follows: two in the high construal–integrative interests condition; six in the high construal level–integrative issues condition; nine in the low construal level–integrative interests condition; one in the low construal level–integrative issues condition. We performed a binomial logistic regression with construal level, location of integrative potential, and their interaction as predictors to test whether experimental conditions affected the distribution of impasses. In line with our reasoning, the analysis only yielded a significant interaction between the independent variables (OR ⫽ .379, CI ⫽ .188 –.764, p ⬍ .01), indicating that dyads in the high construal– integrative issues and in the low construal–integrative interests conditions were more likely to impasse than the other dyads. Joint outcomes. When computing a joint outcome index, impasses can be treated in several ways (for an elaborate discussion, see, e.g., Pruitt, 1981). First, no-agreements dyads can be excluded from the analyses, but this destroys randomization and the ability to draw causal inferences. Second, a joint outcome equal to 0 could be assigned to the no-agreements dyads, but this would create substantial heterogeneity of variance with reduced statistical power. Third, a value that does not create too much heterogeneity


770

Table 3 Issue and Interest Charts for the Seller and the Buyer in Experiment 3 Seller issue charts Price

Warrantee

Interest

Seller interest charts Option included

Delivery

Category

Issues

Value

Money Time Facilities

Price, Interests Delivery Warrantee, Optional

3 (high) 2 (medium) 1 (low)

Integrative interests 12.000 (150) 11.500 (120) 11.000 (90) 10.500 (60) 10.000 (30)

6 months (50) 12 months (40) 18 months (30) 24 months (20) 30 months (10)

10% (50) 8% (40) 6% (30) 4% (20) 2% (10)

1 (150) 2 (120) 3 (90) 4 (60) 5 (30)

5 4 3 2 1

weeks (75) weeks (60) weeks (45) weeks (30) week (15)

Buyer issue charts

Buyer interest charts

Price

Warrantee

Interest

Option included

12.000 (30) 11.500 (60) 11.000 (90) 10.500 (120) 10.000 (150)


10% (10) 8% (20) 6% (30) 4% (40) 2% (50)

1 (30) 2 (60) 3 (90) 4 (120) 5 (150)

5 4 3 2 1

Delivery

Category

Issues

Value




1 (low) 2 (medium) 3 (high)

Integrative issues Seller issue charts

Seller interest charts

Price

Warrantee

Interest

Option included

12.000 (150) 11.500 (120) 11.000 (90) 10.500 (60) 10.000 (30)


10% (50) 8% (40) 6% (30) 4% (20) 2% (10)

1 (150) 2 (120) 3 (90) 4 (60) 5 (30)

5 4 3 2 1

Delivery

Category

Issues

Value





Buyer issue charts

Buyer interest charts

Price

Warrantee

Interest

Option included

12.000 (10) 11.500 (20) 11.000 (30) 10.500 (40) 10.000 (50)


10% (30) 8% (60) 6% (90) 4% (120) 2% (150)

1 (10) 2 (20) 3 (30) 4 (40) 5 (50)

of variance yet reflects a suboptimal performance could be assigned when an impasse is reached. This third option is commonly used in the negotiation literature (De Dreu et al., 2006; Pruitt & Lewis, 1975) and adopted here as well. Accordingly, we assigned impasse dyads a joint outcome equal to the lowest joint outcome reached by any dyad (i.e., 920 points). A 2 ⫻ 2 (Construal Level ⫻ Location of Integrative Potential) ANOVA on joint outcome only showed the predicted interaction between construal level and location of integrative potential, F(1, 78) ⫽ 10.27, p ⬍ .01, ␩2 ⫽ .12. Simple effects showed that when integrative potential resided at the level of underlying interests, high-construal level dyads obtained a better joint outcome (M ⫽ 1221.57) than low-construal level dyads (M ⫽ 1100.53), F(1, 78) ⫽ 6.36, p ⬍ .05, ␩2 ⫽ .08. When integrative potential resided at the issue level, low-construal level dyads obtained a better joint outcome (M ⫽ 1211.5) than high-construal level dyads (M ⫽ 1122.08), F(1, 78) ⫽ 3.98, p ⬍ .05, ␩2 ⫽ .05 (see also Figure 3). Negotiation strategies. We analyzed questionnaire data at the dyadic level to deal with statistical interdependency (e.g., Kenny,

5 4 3 2 1

Delivery

Category

Issues

Value





Kashy, & Bolger, 1998). A 2 ⫻ 2 (Construal Level ⫻ Location of Integrative Potential) ANOVA on cooperative problem solving revealed a significant two-way interaction between construal level and location of integrative potential, F(1, 78) ⫽ 7.37, p ⬍ .01, ␩2 ⫽ .09. When integrative potential was located in the underlying interests, high-construal level dyads reported more problem solving than low-construal level dyads (M ⫽ 3.84 vs. M ⫽ 3.47), F(1, 78) ⫽ 4.19, p ⬍ .05, ␩2 ⫽ .05. When integrative potential was located in the issues, low-construal level dyads tended to engage in more problem solving than high-construal level dyads (M ⫽ 3.82 vs. M ⫽ 3.53), F(1, 78) ⫽ 3.19, p ⫽ .08 (marginal). Cooperative problem solving related to joint outcomes (r ⫽ .37, p ⬍ .01). To test for mediation, we regressed the joint outcomes on the dummy-coded main effect of construal level and integrative location and their interaction before and after cooperative problem solving had been controlled for. Results showed that the originally significant regression of joint outcome on the interaction of the independent variables (␤ ⫽ .34), t(78) ⫽ 3.20, p ⬍ .01, decreased when integrative we entered behavior in the model (␤ ⫽ .25),

CONSTRUAL LEVEL IN NEGOTIATION 1240

Joint Outcome

1220 1200

High Construal Level Low Construal Level

1180 1160 1140 1120 1100 1080

Interests

Issues

Location Integrative Potential Figure 3. Joint outcome as a function of construal level and location of integrative potential in Experiment 3.

t(78) ⫽ 2.37, p ⬍ .05. The association between cooperative problem solving and joint outcome held when the predictor was controlled for (␤ ⫽ .30), t(78) ⫽ 2.80, p ⬍ .01. Because the product of the unstandardized coefficients is not normally distributed, we used the z-prime method to test whether the change from simple to multiple regression was significant (MacKinnon, Lockwood, Hoffman, West, & Sheets, 2002). This was indeed the case: The effect of Construal Level ⫻ Integrative Location on joint outcome was significantly reduced when cooperative problem solving was controlled for (z⬘ ⫽ 1.95; p ⬍ .05).

Conclusions and General Discussion The present study revealed that high rather than low psychological distance and concomitant global rather than local processing modes substantially influence negotiator cognitions, behavioral tendencies, and outcomes. Specifically, negotiators under a global rather than a local processing mode revise their inadequate fixedpie perceptions more when provided with information about counterpart’s underlying interests (Experiment 1). And they were more likely to accept offers that appear uninteresting when considered in terms of issues but quite valuable when underlying interests are taken into account (Experiment 2). Experiment 3 further clarified that these effects can be parsimoniously understood in terms of the focus on secondary features of the negotiation (issues) under low psychological distance and local processing, yet on primary features of the negotiation (underlying interests) under high psychological distance and global processing. That is, Experiment 3 showed less integrative agreements under low rather than high construal level and concomitant local processing when integrative potential was located at the level of underlying interests, but more integrative agreements under low rather than high construal level and concomitant global processing when integrative potential was located at the level of overt issues. Together, findings contribute to researchers’ understanding of interpersonal negotiation, and they expand core facets of CLT to social exchange settings. These issues, along with some limitations and questions for new research, are addressed in the remainder of this section.

Theoretical Implications The implications for negotiation theory are twofold. First, we used a negotiation task that allowed us to manipulate the level at

771

which integrative potential resides. Past work, building off the classic study by Pruitt and Lewis (1975), used a multi-issue task with integrative potential located in the value attached to the various issues. Whereas this task allows the researcher to vary the amount of integrative potential, it did not allow researchers to vary preferences on issues independently from the integrative potential inherently attached to those preferences. In the present study, we used an adapted version of the task designed by Tro¨tschel and Gollwitzer (2007). It does allow the researcher to vary these two aspects of the negotiation task independently—tougher preferences on some or all issues were, in principle, independent of the extent to which underlying interests were served. This is an important innovation for two reasons. First, most real-life negotiations involving multiple issues have a rather indirect direct linkage between the intrinsic value of an issue and the underlying interests it serves. More than tasks used in past work (e.g., De Dreu et al., 2006; Pruitt & Lewis, 1975), the currently employed negotiation task accounts for complex features of “real-life” negotiation (e.g., underlying interests), and thus enhances the external validity of negotiation research. Second, this new task allows researchers to examine factors, both old and new, that contribute to the negotiators’ focus on issues, on interests, or some combination. As such, the currently employed task opens up a series of questions that could not be answered with the tasks used in past work. For example, are negotiators with a prosocial motivation more likely to focus on interests, whereas those with proself motivation tend toward a focus on issues? Does power influence the tendency to focus on underlying interests versus issues as much as it influences the tendency to engage in global versus local processing (e.g., Smith & Trope, 2006)? Do certain emotions focus negotiators on basic needs and interests, whereas other feeling states orient them toward overt issues (and what are the effects of the partner’s emotions on the receiving negotiator’s tendency to consider own and others interests, or issues instead)? Clearly, these are but a few possible questions that are of critical importance to further understanding negotiation and that the present task allows to study in a more sophisticated way than have previously used negotiation tasks. Apart from making this methodological contribution, our study contributes to researchers’ understanding of the psychological mechanisms underlying integrative negotiation. We already knew that negotiators are more likely to uncover integrative potential when they exchange and process information about their own and other’s preferences and priorities (e.g., Pruitt & Lewis, 1975; Weingart, Hyder, & Prietula, 1996). We also knew a lot about the cognitive and motivational determinants of information processing and exchange fostering integrative agreements (for reviews, see, e.g., Bazerman et al., 2000; Carnevale & Pruitt, 1992; De Dreu et al., 2007). Yet despite a long list of books of advice on the importance of focusing on underlying interests rather than on overt issues (e.g., Fisher & Ury, 1981; Lax & Sebenius, 1986; Walton & McKersie, 1965), we knew exceedingly little about the psychological determinants of a negotiator focus on secondary features such as issues, and primary features such as underlying interests and needs. The present finding that increased psychological distance along with its more global processing mode focuses individuals on primary rather than on secondary features of the negotiation, and thereby facilitates interest-based negotiation, thus is an important step forward.

772


Our findings are consistent with those reported by Henderson et al. (2006) and De Dreu and colleagues (2009). In these studies, the authors found that when negotiations involved issues and its consequences far removed in time, rather than nearby, negotiators created more package deals, considered interrelations among issues to a greater extent, and created more integrative agreements. We extend these findings in several important ways. First, our findings suggest that it is construal level rather than just temporal distance that triggers aforementioned effects. Whereas past work manipulated psychological distance exclusively through temporal construal, we obtained conceptual replications using a variety of different ways to vary psychological distance and to induce a highversus low construal level. However, two ways to create variations in psychological distance were not considered here, and it remains to be seen whether these two would produce effects similar to those we found here. For example, social distance caused by negotiating with a member of the outgroup should raise construal level and thus facilitate the achievement of mutually beneficial agreements (provided integrative potential resides at the level of underlying interests). But negotiating with outgroup members also reduces cooperative motivation and increases competitive tendencies hurting rather than facilitating integrative negotiation (e.g., Halevy, 2008; Howard, Gardner, & Thompson, 2007). In a similar vein, feelings of power, and power asymmetry, are sources of social distance that should increase construal level (Smith & Trope, 2006). However, work on power in negotiation suggests power differences generally impede rather than facilitate the development of integrative, win–win agreements (Giebels, De Dreu, & Van de Vliert, 2000; Mannix & Neale, 1993; Pinkley, 1995). Thus, there is reason to believe that not all dimensions of psychological distance have similar effects, whereas all may introduce more abstract construal level and more global information processing, in some cases co-varying processes may overrule the potentially beneficial effects of abstract and global thinking. Future research is needed to address these and related issues. Second, our findings render less likely an explanation in terms of enhanced cooperative motivation under high temporal distance. As mentioned, past work could not rule out the possibility that high rather than low psychological distance and global rather than local processing fostered a cooperative motivation (e.g., De Dreu & Nijstad, 2008), which in turn promoted simultaneous rather than sequential processing of issues. Simultaneous issue processing facilitates integrative agreements yielding high joint outcome (e.g., Weingart et al., 1993). However, such an explanation cannot account for the present findings. It is difficult to see how, in Experiment 2, enhanced cooperative motivation among highconstrual level negotiators led to greater willingness to accept an interest-based offer from their counterpart, but to reduced willingness to accept an issue-based offer. Likewise, it is difficult to see how, in Experiment 3, enhanced cooperative motivation among high-construal level negotiators led to more integrative negotiation in the integrative-interest condition but not in the integrative issue condition. Third, and finally, our findings suggest that present and previous findings can be best understood in terms of two distinct processes that operate in parallel—a focus on primary rather than on secondary features of the negotiation, and a tendency to craft multiissue offers that emphasize interrelations among issues and interests and that facilitate logrolling. Provided integrative potential

resides at the level of underlying interests, negotiators with a high construal level thus do achieve more integrative agreements. This “cross-over” interaction is important not only because it provides a critical test of CLT, and rules out the possibility that findings were driven by enhanced cooperative motivation, but also because it provides first-time evidence for the fact that interest-based negotiation need not necessarily lead to integrative agreements, an erroneous assumption often made in both academic and practitioner-oriented writings. To the extent that parties have interests that are diametrically opposed, interest-based negotiation will be difficult and unlikely to result in mutually beneficial, win–win agreements. In our analysis, we referred to primary features of the negotiation as the distant goals negotiators have or adopt, essentially referring to their underlying interests (i.e., security needs vs. reputation concerns in the Sinai Peninsula dispute). The subordinate means to achieve these more distant goals were referred to as secondary features; it entails the tactics, specific demands, and positional commitments negotiators adopt and communicate (i.e., the claimed share of the Sinai Peninsula). However, the means-end distinction between issues and interests does not reflect all the possible distinctions between primary and secondary elements of a negotiation. As goals are typically organized into a hierarchy with more exterior interests serving more fundamental ones, superficial interests can be seen as secondary and subordinate to more basic interests. In Experiment 1, for example, the high concern for money of the candidate can be a superficial interest serving the more basic interest of getting married. Results of the present research strongly suggest that under high, rather than low, construal level, negotiators will place more emphasis on the most central interests. This could be very helpful in order to devise a creative solution to a negotiation because, as noted by Rubin et al. (1994), “If Party moves along the (hierarchical) tree far enough, it may locate an interest that can be easily bridged with an interest of Other” (p. 179). Likewise, negotiators could have multiple goals at the same time, with some of them being more important than others even though not organized hierarchically. For example, a negotiator could be concerned with obtaining as much economic value as possible (high importance) and with presenting him- or herself as a fair person (low importance). If the most important goal can be seen as a primary feature of the negotiation, then, according to CLT, we should expect the person in the example to be more competitive and to demand more under a high construal level. Our conclusions and their implications are constrained by the fact that we obtained insight into the negotiating process either without also observing outcomes (i.e., in Experiment 1 and 2) or by relying on self-reports obtained after outcomes had been reached (i.e., in Experiment 3). Future research could invest in online observation of negotiation behavior to provide more solid evidence for the role of multi-issue offer making and cooperative problem solving. Future research could also seek to connect within one study task-related cognitions and perceptual accuracy, offer making and acceptance, and integrative outcomes. Whereas present findings strongly point to such interrelations, studies allowing for path analyses would further illuminate researchers’ understanding of the ways psychological distance influences negotiation processes and outcomes.


Another issue for future research is to consider the interplay between construal level, on the one hand, and negotiator tendency to engage in more or less deep and systematic information processing. Recent work suggests that when negotiators engage in such deep processing, for example, because they are under process accountability or have low need for cognitive closure, they develop more accurate perceptions and are more likely to achieve integrative agreements (De Dreu & Carnevale, 2003; De Dreu, Nijstad, & Van Knippenberg, 2008). Deep and systematic processing is, in principle, independent of construal level— one can take a global construal level and process information either systematically or heuristically as much as one can take a local construal level and process information either systematically or heuristically. In terms of the present findings, it follows that one may process information about either primary or secondary features—interests or issues—in either a systematic and deliberate fashion or in a more heuristic and shallow way. The intriguing implication is that deep information processing should amplify the effects observed here— construal level influences integrative agreements especially when negotiators engage in deep, deliberate, and systematic information processing.

Concluding Thoughts In three experiments, we showed that interests underlying parties’ preferences on issues are more relevant in orienting considerations, cognitions, and strategic behavior of negotiators with a high rather than a low construal level. Conversely, when negotiators have a low construal level, strategic behavior is based on negotiation issues. We also demonstrated that the impact of construal level on dyads’ joint outcome was moderated by the location of integrative potential. That is, when integrative potential resided in the underlying interests, as in most cases, negotiators can benefit from construing information at high level. When integrative potential resided in the issues, construing information at low level appeared to be beneficial for high-quality agreements. One of the most common types of advice to uncover underlying interests in a negotiation is that negotiators should ask themselves a series of why questions about the negotiation (Fisher & Ury, 1981): Why is he or she taking this position? Why is the other party not yielding as much as I demand? Why is the other giving up that issue so easily? Our study suggests that asking such “why questions” is useful indeed, but not because of the potential answers one may get. Rather, asking such questions helps one to switch from a local to a more global construal level and processing mode, facilitating a focus on underlying interests. Our study also suggests that the advice offered by Fisher and Ury, and their followers, is valid and helpful to the extent only that integrative potential resides at the level of underlying interests—when this is not the case, asking “why questions” or shifting to a global processing mode may turn out to be quite counterproductive and leads negotiators away from, rather than toward, mutually beneficial win–win agreements.

References Bazerman, M. H., Curhan, J. R., Moore, D. A., & Valley, K. L. (2000). Negotiation. Annual Review of Psychology, 51, 279 –314. Burrus, J., & Roese, N. J. (2006). Long ago it was meant to be: The

773

interplay between time, construal and fate beliefs. Personality and Social Psychology Bulletin, 32, 1050 –1058. Carnevale, P. J., & Pruitt, D. G. (1992). Negotiation and mediation. Annual Review of Psychology, 43, 531–582. De Dreu, C. K. W., Beersma, B., Steinel, W., & Van Kleef, G. A. (2007). The psychology of negotiation: Principles and basic processes. In A. W. Kruglanski & E. T. Higgins (Eds.), Handbook of basic principles in social psychology (2nd ed., pp. 608 – 629). New York, NY: Guilford Press. De Dreu, C. K. W., Beersma, B., Stroebe, K., & Euwema, M. C. (2006). Motivated information processing, strategic choice, and the quality of negotiated agreement. Journal of Personality and Social Psychology, 90, 927–943. De Dreu, C. K. W., & Carnevale, P. J. (2003). Motivational bases of information processing and strategy in negotiation and social conflict. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 35, pp. 235–291). New York, NY: Academic Press. De Dreu, C. K. W., Evers, A., Beersma, B., Kluwer, E. S., & Nauta, A. (2001). A theory-based measure of conflict management strategies in the workplace. Journal of Organizational Behavior, 22, 645– 668. De Dreu, C. K. W., Giacomantonio, M., Shalvi, S., & Sligte, D. (2009). Getting stuck or stepping back: Effects of obstacles in the negotiation of creative solutions. Journal of Experimental Social Psychology, 45, 542– 548. De Dreu, C. K. W., Koole, S., & Steinel, W. (2000). Unfixing the fixed pie: A motivated information-processing account of integrative negotiation. Journal of Personality and Social Psychology, 79, 975–987. De Dreu, C. K. W., & Nijstad, B. A. (2008). Mental set and creative thought in social conflict: Threat rigidity versus motivated focus. Journal of Personality and Social Psychology, 95, 648 – 661. De Dreu, C. K. W., Nijstad, B. A., & Van Knippenberg, D. (2008). Motivated information processing in group judgment and decision making. Personality and Social Psychology Review, 12, 22– 49. De Dreu, C. K. W., Weingart, L. R., & Kwon, S. (2000). Influence of social motives on integrative negotiation: A meta-analytic review and test of two theories. Journal of Personality and Social Psychology, 78, 889 – 905. Eyal, T., Liberman, N., Trope, Y., & Walther, E. (2004). The pros and cons of temporally near and distant action. Journal of Personality and Social Psychology, 86, 781–795. Fisher, R., & Ury, W. (1981). Getting to yes: Negotiating agreement without giving in. New York, NY: Penguin Books. Fo¨rster, J. (2009). Cognitive consequences of novelty and familiarity: How mere exposure influences level of construal. Journal of Experimental Social Psychology, 45, 444 – 447. Fo¨rster, J., Friedman, R. S., & Liberman, N. (2004). Temporal construal effects on abstract and concrete thinking: Consequences for insight and creative cognition. Journal of Personality and Social Psychology, 87, 177–189. Freitas, A. L., Gollwitzer, P. M., & Trope, Y. (2004). The influence of abstract and concrete mindsets on anticipating and guiding others’ self-regulatory efforts. Journal of Experimental Social Psychology, 40, 739 –752. Fujita, K., Trope, Y., Liberman, N., & Levin-Sagi, M. (2006). Construal levels and self-control. Journal of Personality and Social Psychology, 90, 351–367. Giebels, E., De Dreu, C. K. W., & Van de Vliert, E. (2000). Interdependence in negotiation: Impact of exit options and social motives on distributive and integrative negotiation. European Journal of Social Psychology, 30, 255–272. Halevy, N. (2008). Team negotiation: Social, epistemic, economic, and psychological consequences of subgroup conflict. Personality and Social Psychology Bulletin, 34, 1687–1702. Henderson, M. D., Trope, Y., & Carnevale, P. (2006). Negotiation from a

774


near and distant time perspective. Journal of Personality and Social Psychology, 91, 712–729. Howard, S. E., Gardner, W. L., & Thompson, L. (2007). The role of the self-concept and the social context in determining the behavior of power holders: Self-construal in intergroup versus dyadic dispute resolution negotiations. Journal of Personality and Social Psychology, 93, 614 – 631. Kelman, H. (2006). Interests, relationships, identities: Three central issues for individuals and groups in negotiating their social environment. Annual Review of Psychology, 57, 1–26. Kenny, D. A., Kashy, D. A., & Bolger, N. (1998). Data analysis in social psychology. In D. T. Gilbert, S. T. Fiske, & A. Lindzey (Eds.), The handbook of social psychology (Vol. 1, pp. 233–265). Boston, MA: McGraw-Hill. Lax, D. A., & Sebenius, J. K. (1986). The manager as negotiator: Bargaining for cooperation and competitive gain. New York, NY: Free Press. Liberman, N., Sagristano, M., & Trope, Y. (2002). The effect of temporal distance on level of construal. Journal of Experimental Social Psychology, 38, 523–535. Liberman, N., & Trope, Y. (1998). The role of feasibility and desirability considerations in near and distant future decisions: A test of temporal construal theory. Journal of Personality and Social Psychology, 75, 5–18. Liberman, N., Trope, Y., & Stephan, E. (2007). Psychological distance. In A. W. Kruglanski & E. T. Higgins (Eds.), Social psychology: Handbook of basic principles (2nd ed., Vol. 2, pp. 353–383). New York, NY: Guilford Press. MacKinnon, D. P., Lockwood, C. M., Hoffman, J. M., West, S. G., & Sheets, B. (2002). A comparison of methods to test mediation and other intervening variable effects. Psychological Methods, 7, 83–104. Mannix, E. A., & Neale, M. A. (1993). Power imbalance and the pattern of exchange in dyadic negotiation. Group Decision and Negotiation, 2, 119 –133. Pinkley, R. L. (1995). Impact of knowledge regarding alternatives to settlement in dyadic negotiations. Journal of Applied Psychology, 80, 403– 417. Pruitt, D. G. (1981). Negotiation. New York, NY: Academic Press. Pruitt, D. G., & Lewis, S. A. (1975). Development of integrative solutions

in bilateral negotiation. Journal of Personality and Social Psychology, 31, 621– 630. Rubin, J. Z., Pruitt, D. G., & Kim, S. H. (1994). Social conflict: Escalation, stalemate, and settlement. New York, NY: McGraw-Hill. Sagristano, M. D., Trope, Y., & Liberman, N. (2002). Time-dependent gambling: Odds now, money later. Journal of Experimental Psychology: General, 131, 364 –376. Schul, Y. (1983). Integration and abstraction in impression formation. Journal of Personality and Social Psychology, 44, 45–54. Smith, P. K., & Trope, Y. (2006). You focus on the forest when you’re in charge of the trees: Power priming and abstract information processing. Journal of Personality and Social Psychology, 90, 578 –596. Thompson, L. L. (1991). Information exchange in negotiation. Journal of Experimental Social Psychology, 27, 161–179. Thompson, L. L., & Hastie, R. (1990). Social perception in negotiation. Organizational Behavior and Human Decision Processes, 47, 98 –123. Trope, Y., & Liberman, N. (2000). Temporal construal and time-dependent changes in preference. Journal of Personality and Social Psychology, 79, 876 – 889. Trope, Y., & Liberman, N. (2003). Temporal construal. Psychological Review, 110, 403– 421. Tro¨tschel, R., & Gollwitzer, P. M. (2007). Implementation intentions and the willful pursuit of prosocial goals in negotiations. Journal of Experimental Social Psychology, 43, 579 –598. Vallacher, R. R., & Wegner, D. M. (1989). Levels of personal agency: Individual variation in action identification. Journal of Personality and Social Psychology, 57, 660 – 671. Walton, R. E., & McKersie, R. (1965). A behavioral theory of labor negotiations: An analysis of a social interaction system. New York, NY: McGraw-Hill. Weingart, L. R., Bennett, R. J., & Brett, J. M. (1993). The impact of consideration of issues and motivational orientation on group negotiation process and outcome. Journal of Applied Psychology, 78, 504 –517. Weingart, L. R., Hyder, E. B., & Prietula, M. J. (1996). Knowledge matters: The effect of tactical descriptions on negotiation behavior and outcome. Journal of Personality and Social Psychology, 70, 1205–1217.

Received January 8, 2009 Revision received September 4, 2009 Accepted September 8, 2009 䡲