Abstract
A number of studies have shown that certain events that occur during a negotiation can alter its course. Referred to as “turning points,” these events are precipitated by actions taken either outside or inside the talks that have consequences for outcomes. In this article, we report the results of two experiments designed to examine the impacts of two types of precipitating actions, external and internal. In the first experiment, which focused on external actions, we found that crises — as opposed to breakthroughs — produced more movement in negotiations in which parties viewed the social climate positively (high trust, low power). We found that parties achieved less movement in negative social climates (low trust, high power).
In the second experiment, which focused on internal actions, we found that cooperative precipitants (factors inducing change) were more likely to occur when parties negotiated in the context of positive social climates. Negotiation outcomes were also influenced by the climate: we found better individual outcomes for negotiations that occurred in positive climates (high trust, cooperative orientations). Inboth experiments, the social climate of the negotiation moderated the effects of precipitating factors on negotiation outcomes. Perceptions of trust and power filter the way negotiators interpret actions that occur outside or are taken inside a negotiation, which can lead to agreements or impasses.
Introduction
In his analysis of international negotiation, I. William Zartman (2000) proposed the concept of “ripe moments.” These moments designate times in a conflict when parties are more likely to come to the negotiating table. A key precipitating factor is conflict escalation, also known as a “hurting stalemate.” Ripe moments have a process analogue in the concept of turning points (Druckman 2001 and 2004). Turning points describe moments in time when the negotiation process changes. Like ripe moments, they are often observed after periods of intense escalatory behavior (Druckman 1986 and 2001), but they may also be observed after prolonged periods of cooperation (Olekalns and Weingart 2008). The unifying theme in the discussion of ripe moments and turning points is the recognition that negotiations do not unfold smoothly. Instead, the negotiation process is discontinuous and punctuated by transitions in dominant strategy, which are themselves precipitated by observable events.
We can understand turning points in terms of several psychological and social processes that characterize the negotiation before, during, and after the turning point itself. In his analysis of the underlying impetus for departures from negotiation process, Daniel Druckman (2004) identified both individual and relational factors. Individual negotiator factors include changes in emotions or moods, the ways that information about other negotiators is processed, and surprise. At the relational level, such elements as synchronized responding, interdependent relationships between negotiators, power symmetries, and perceived trust can catalyze turning points.
The way that these factors interact with precipitants of turning points has not yet been explored. The experiments that we report in this article addressed this issue by testing whether individual and relational factors create interpretive filters that shape negotiators' responses to changes in the negotiation process.
A Turning Points Framework
In his analysis of turning points, Druckman (2001) distinguished between precipitants (cause) and departures (effect). Precipitants refer to identifiable events that trigger changes in the negotiation process. The introduction of new ideas or information, changes to the structure or format of negotiations, and external events all have the potential to precipitate a departure in the negotiation process. Departures, then, are reactions to any of these precipitants. Negotiators may choose to incorporate or ignore new ideas, they may accept or reject suggested procedural changes, and they may react or fail to react to external events. When negotiators respond to any of these precipitants, the negotiation process will change. Thus, departures have consequences: as a result of departures, negotiations move either toward or away from agreement.
Our focus in these studies was on precipitants, which can be distinguished in terms of their sources, either external or internal. External events, as Druckman (2001) described, are events outside of the negotiation that nonetheless affect subsequent processes. They can include policy and leadership changes, as well as third-party interventions. Thus, for example, negotiations about ongoing support for a politically unstable country may change as a result of new political leadership in that country.
Internal events refer to actions that negotiators themselves take that can alter the negotiation process. The internal actions that change negotiation processes have received increasing attention and include such concepts as interruptions, out-of-keeping acts, and turns (Olekalns, Brett, and Weingart 2003; Kolb 2004; McGinn, Lingo, and Ciano 2004).1 These are actions that change the “flow” of a negotiation process (Donohue 2004). A common thread throughout these discussions is the idea that negotiators and external actors can directly or indirectly intervene in and change the negotiation process. For example, negotiators might suggest putting a seemingly tough issue to one side and returning to it at a later time — this is a procedural precipitant. Or a negotiator might reframe an issue in a manner that alters the terms of trade — this is a substantive precipitant. (See Druckman, Husbands, and Johnston 1991 for examples of substantive precipitants that reframed the Intermediate-range Nuclear Forces talks in the 1980s.)
Turning points/departures also differ in their likely consequences. Some departures have positive consequences (i.e., progress toward agreement), while the results of others are negative (i.e., movement away from agreements and impasses). Examples of positive consequences include parties drafting a treaty document, reaching an agreement on principles, or resolving issues that have blocked progress to this point in the talks. Negative consequences include parties breaking off the talks or threatening to do so, the surfacing of new divisions, or the introduction of new issues. (A turning points analysis addresses both short- and long-term consequences; see Druckman 2001.)
The experiments that we report in this article explored the paths from precipitants to departures to their consequences. In the first experiment, precipitants took the form of messages about external events, departures were the reactions of negotiators to those events, and consequences were the outcomes. In the second experiment, we asked bargainers to identify the internal events (precipitants) that altered the course of the talks, departures that occurred as a result of these events, and the outcomes. We examine these paths in terms of negotiator perceptions, which are the interpretive filters that intervene between the precipitant and the parties' reaction to or decision about that event. An understanding of these filters, we believe, can help answer the question: why do certain precipitants trigger departures while others do not (Donohue 2004)? We turn now to a discussion of these filters.
Social Cognition as an Interpretive Filter
Much of the earlier research on external and internal precipitants has assumed that negotiators will agree on how to interpret and react to them. Social cognition research, however, suggests that this may not be the case. This research has found that negotiators are responsive to social cues and that the interpretation of these cues influences their perceptions of both themselves and the other party. For example, not only do negotiators rate themselves more positively than they rate the other party, but these positive self-ratings are enhanced when negotiators believe that the other party felt happy rather than sad at the end of the negotiation (Kramer, Newton, and Pomeranke 1993; Thompson, Valley, and Kramer 1995). Michael Morris, Richard Larrick, and Steven Su (1999) demonstrated that negotiators who “haggle” are rated as more disagreeable than those who do not and that negotiators who “waffle” are rated as less emotionally stable than those who do not. Mara Olekalns and her colleagues (Olekalns et al. 2005) demonstrated that the same negotiation strategies are interpreted differently, depending on the context in which they occur. This research stream suggests that how negotiators interpret and respond to precipitants will be influenced by social cues.
Our broad argument is that how negotiators respond to precipitants will be determined by whether they assess the other party's behavior through a positive or a negative interpretive filter. We propose that negotiators develop their filters from two classes of inferential cues, person based and context based. Person-based inferences reflect negotiators' judgments about their opponent (impressions) and their perception of likely exposure to exploitation (other's trustworthiness). Context-based cues capture the environment in which the negotiation takes place — in the case of our experiments, whether external precipitants are positive or negative (Experiment One) or whether negotiators hold cooperative (communal) or individualistic (more competitive) goals (Experiment Two).
These inferential cues influence how the negotiators interpret their relationship. A positive filter orients negotiators to the interdependent aspects of their interests, leading them to engage in problem solving and in more integrative negotiating. A negative filter increases their sense of independence as well as their fear that the other party might exploit them (McKnight, Cummings, and Chervany 1998). We hypothesize that when either person- or context-based cues encourage parties to form positive interpretive filters, negotiators will be more responsive to positive precipitants, whereas when these cues create a negative interpretive filter, they will be more responsive to negative precipitants. Moreover, when person- and context-based cues are congruent, their impact will be amplified.
Impressions of the Other Party
How negotiators perceive the other party affects their reactions to both internal and external precipitants. To assess the nature of this relationship, we explored how perceptions of communality and power affect the identification of and reactions to precipitants. We used these two dimensions of social perception because of their pervasiveness in characterizing social relationships. Power and communality reflect, respectively, interests in status, hierarchy, and power, or interests in goodness, morality, and solidarity.
These criteria have been used frequently to characterize relationships across a wide range of settings (Moskowitz, Suh, and Desaulniers 1994; Dillard, Palmer, and Kinney 1995) and they can be seen as basic dimensions for understanding communication (Donnellon 1994). Moreover, research shows that whereas impressions of another person as “good” or “moral” can alleviate concerns about his or her intentions toward us, perceptions of the person as “powerful” can heighten such concerns (De Dreu, Giebels, and Van de Vliert 1998). The dimensions of communality and power are salient for negotiators and can shape their decisions to engage in deception (Olekalns and Smith 2007). Drawing on our more general argument, we expected that a positive interpretive filter would be created when negotiators assess the other party as high in communality, whereas a negative interpretive filter will be created when the other party is perceived as being powerful.
We propose that the relevance of impressions will vary as a function of whether precipitants are internal or external to the negotiation. This is because internal and external precipitants shift the attributional locus, which is the negotiators' sense of which sources they should attribute as being responsible for their counterpart's actions. In their analysis of social settings, Walter Mischel and Yuichi Shoda (1995) distinguished between strong and weak situations, which are defined by the extent to which they generate ambiguity about appropriate behavior. Extending this argument, we propose that external precipitants create less ambiguity, that is, a “stronger” situation than internal precipitants. By this, we mean that external precipitants encourage negotiators to attribute behavior to features of the situation rather than to the other party's motives (e.g., Kelley 1973; Hilton, Smith, and Kim 1995).
In our first experiment, external precipitants occurred in a context of situational clarity (“strong situation”), drawing negotiators' attention away from the other party's motives. In the absence of such clear cues, negotiators are more likely to focus on their impressions of the other party (e.g., Johnson-George and Swap 1982; McKnight, Cummings, and Chervany 1998).
In our second experiment, internal precipitants occur in the context of high situational ambiguity (“weak situation”), drawing negotiators' attention to the underlying motives of the other party. Although we did not directly test this proposition, our argument suggested that we would observe relationships between impressions and negotiation outcomes in Experiment Two but not in Experiment One.
Trustworthiness
Reactions to internal and external precipitants are also affected by existing assessments of the other party's trustworthiness. Defined as the willingness to make oneself vulnerable to another person (Lewicki and Wiethoff 2000), trustworthiness is central in determining the likelihood of exploitation by the other party. Without trust, negotiators are unlikely to engage in problem-solving behaviors, such as information exchange, that enable them to find mutually beneficial solutions (Pruitt 1981; Butler 1999). Importantly, their levels of trust can alter individuals' sensitivity to negative events: in the context of contracts, Sandra Robinson (1996) found that individuals who have high trust in their managers are less likely to identify contract violations than those who have low trust. This suggests that negotiators who perceive the other party as highly trustworthy are more likely to attune to positive precipitants and to discount negative internal and external precipitants. Consistent with this argument, Olekalns and Philip Smith (2005) found that the perceived trustworthiness of the other party affects a negotiator's identification of positive and negative precipitants, which in turn influences the level of trust at the end of negotiations. Paralleling our arguments about impressions, we expected high perceived trust to create a positive interpretive filter and low perceived trust to create a negative interpretive filter.
In this research, we distinguish between inferential cues that signal either competence-based or goodwill-based trust. Competence-based trust involves judgments about the other person's reliability and predictability, whereas goodwill-based trust involves judgments about the other person's morality, specifically, his or her benevolence and integrity (Lewicki and Wiethoff 2000; Das and Teng 2001). We expect that a negative interpretive filter sharpens negotiators' focus on the tangible deal-making aspects of the negotiation and prompts fears that the negotiation will break down because the other party lacks the skills to successfully conclude a deal, consequently orienting negotiators to competence-based trust. We expect that a positive interpretive filter orients negotiators to the less tangible relational aspects of negotiation. Negotiators are concerned about the other party's benevolence and integrity, increasing the salience of goodwill-based trust.
Finally, we hypothesized that the two experiments would differ in the relevance to negotiators of competence- and goodwill-based trust because these two different kinds of trust prompt different representations of the negotiation, specifically priming an in-group (Experiment One) or an out-group orientation (Experiment Two). According to social identity theory, when individuals identify with a group they attempt to define themselves in terms of “we” rather than “I” (Tajfel and Turner 1986). This shift from a personal to a collective orientation (Flynn 2005) is likely to be accompanied by a change in focus from competence-based to goodwill-based (affective) trust.
In Experiment One, individuals were asked to negotiate about weapons inspections. This negotiation highlighted the relational aspects of the negotiation because it emphasized interdependencies between the negotiators; that is, weapons threaten the security of both parties. The negotiation effectively placed them in a common fate situation, which we expected would increase the importance of affective-based trust. Conversely, Experiment Two clearly highlighted the economic, deal-making aspects of the negotiation. Because this experiment emphasized the negotiators' independence, we expected that it would increase the salience of competence-based trust. Consistent with this argument, we would expect that affective-based trust would predict outcomes in Experiment One while competence-based trust would predict outcomes in Experiment Two. We do not provide a direct test of this argument, but we found that affective-based trust predicted outcomes in Experiment One whereas competence-based trust predicted outcomes in Experiment Two, which would be consistent with the preceding discussion.
The Experiments
Earlier case-study analyses indicated that external precipitants led to progress on security issues while internal precipitants were instrumental in moving trade and political negotiations along (Druckman 2001). We preserved this connection in our experiments. We focused on external precipitants in the security scenario developed for Experiment One and studied internal precipitants by constructing an employment contract task for Experiment Two.
In Experiment One, we evaluated the impact of interrupting an arms negotiation with an external message intended to trigger a crisis and with another message intended to trigger a positive reassessment of the situation. These messages are outside events not controlled by the negotiators. In Experiment Two, we evaluated how congruent and incongruent goals held by negotiators affected the identification of internal precipitants in the context of a simulated employment contract negotiation. These precipitants were often substantive or procedural initiatives brought about by the negotiators during the process.
The common themes in these experiments were the measures of inferential cues — perceived trustworthiness, communality, and power of the other party at the outset of talks — and our attempt to link these perceptions to the anticipated and actual outcomes of the negotiation. The different scenarios provide an opportunity to address issues of generality or robustness of findings: similar findings obtained from different tasks bolster the case for convergent validity (Cook 1985).2
Experiment One: External Precipitants
In this experiment, we evaluated how external precipitants can influence negotiators' assessment of their likely outcome as well as their actual outcome. More specifically, we asked whether the impact of either an external crisis or of a positive breakthrough in relations between the negotiating parties would be affected by how negotiators perceived the other party. We assessed two impacts: the negotiators' reactions to the messages and the outcomes of the negotiation. Reactions, which are indicators of departures, included a) reaching agreement, b) ending the talks immediately, c) changing opinions about what was possible, or d) no effect. Outcomes refer to the extent to which negotiators moved toward agreement on any or all of the issues discussed.
Our argument also suggests that person-based and context-based cues will interact. As shown in Figure One, we expected that negative person-based cues would increase the salience and impact of a negative external precipitant, whereas positive person-based cues would increase the salience and impact of a positive external precipitant. Thus, with regard to reactions, these were our hypotheses:
Hypothesis One: Crisis messages sent in the context of negative social cues (low trust, high power) will lead negotiators to change their thoughts about what is possible with regard to movement from their original positions.
Hypothesis Two: Breakthrough messages sent in the context of positive social cues (high trust, low power) will alter negotiators' evaluation of what is possible with regard to movement from their original positions.
With regard to outcomes, we hypothesized the following:
Hypothesis Three: Crisis messages sent in the context of negative social cues will result in less movement toward agreement than when the cues are positive.
Hypothesis Four: Breakthrough messages sent in the context of positive social cues will result in more movement toward agreement than when the cues are negative.
Taken together, these hypotheses suggest that social- or person-based cues moderate the relationship between reactions to messages (departures) and outcomes (consequences). The process of rethinking what is possible is triggered by the negotiator's interpretation of the message: negative social cues produce less movement toward agreements when a negative message is sent; positive social cues produce more movement toward agreements when a positive message is sent. This is illustrated in Figure One above.
Participants
Forty-two students in two honors classes on research methods at the University of Queensland participated in the experiment.3 Twenty-six were in a political science class and sixteen were in a similar class in the psychology department. They were similar in practically all demographic categories (age, nationality, educational attainment), the only difference being choice of field. We assigned the students randomly to the experimental condition and role. They were also paired randomly into negotiating dyads. A total of twenty-one dyads were formed, thirteen from the political science class and eight from the psychology class; eleven were placed in the positive message condition while ten were placed in the crisis message condition.
Scenario
We assigned participants to the roles of chief negotiators appointed by their foreign ministries to represent their countries, Anice or Izeria, in a bilateral negotiation concerning a number of security issues. We established the context for the talks by providing the negotiators with a chronology of key events from December 2003 to March 2006. We also told them that the presidents of their two countries considered these talks to be a top priority and that the United Nations secretary general would be closely monitoring each round. The initiation of talks was considered to be a welcome departure from the hostile atmosphere that had existed between these parties.
We then gave each negotiator background information about the issues (five distributive and one integrative) that were on the agenda for the talks. Two of the distributive issues concerned Anice's desire to inspect Izeria's presidential palace; at issue were the number of weapons inspectors and the period of inspection. The other three distributive issues concerned the deployment of an Anicean-led international military force in a border area that lies between Izeria and its neighbor, Kerejistan; specific questions were the number of troops, the period of deployment, and the amount of budget allocation.
The one integrative issue concerned ways to combat terrorism in the region. Both parties were encouraged to consider ways of cooperating on this issue that would be in their joint interest. The private background information provided to each role player described rationales for his or her party's positions on the inspection and deployment issues and incentives to cooperate on a joint plan for the terrorism issue. They were also informed that they would have an opportunity to consult with their foreign ministries during the break period of the thirty-minute negotiation.
Procedure
We made random assignments to the type of condition (positive or negative external precipitant), negotiating role, and pairing from the class lists. All participants in each class were given an overview of the exercise, which included a schedule of tasks and the amount of time allocated for each of them. Following the chronology of events, we asked them to complete a set of questions about their expectations for the negotiation and the opposing negotiator. We then allotted the students twenty-five minutes to study the background information about the three issues. This included scales of alternative outcomes for the distributive issues and a narrative about the integrative issue. At the conclusion of the preparation period, the negotiation began. Students in each class were divided into Anice–Izeria pairs and provided with a space for the talks. They were also given forms for recording the outcomes on each issue.
We divided the negotiation into two fifteen-minute periods, with a break for consultation in between. Students were free to decide on the order for discussion of the issues as well as whether they would be addressed sequentially or in combination. During the break, each dyad received a “Foreign Ministry Update,” which informed them of either a breakthrough on efforts to establish a regional security organization or a crisis in Izeria. They answered questions about the impact of the message before returning to the negotiation table. The second period continued for fifteen minutes or until an agreement was reached on all the issues. After they recorded the outcomes, we asked participants to complete sixteen postnegotiation questions and then debriefed them.
Independent Variable
During the break period, between rounds, all negotiating dyads received a message from their foreign ministry. About half the dyads received the following breakthrough message:
The latest development at the United Nations indicates that a breakthrough has occurred. A United Nations commission, which includes members from both Anice and Izeria, released a report today calling for the establishment of a regional security organization. The impetus for the report came from the results of an investigation into Ranistan's weapons program. The investigation discovered that Ranistan, a neighbor of Izeria, has been engaged in a program for developing nuclear, biological, and chemical weapons. These weapons pose a serious threat to the region. Realizing that other nations in the region could develop similar programs, the commission recommends “the establishment of an organization that would monitor weapons programs and encourage cooperation with the long-term objective of achieving lasting peace in the region.” Both Anice and Izeria strongly support this recommendation.
Keep in mind that this development has implications for both countries. It may also contribute to proposed joint efforts in combating terrorism. Remember that you have a limited amount of time to try to resolve the issues being negotiated.
The other half of the negotiating dyads received the following crisis message:
Late-breaking news indicates that a crisis has occurred in Izeria. The Associated Press reports that the president of Izeria, Sadam Ismaeli, succumbed unexpectedly to a fatal heart attack. This event has thrown the government into chaos as they hurriedly prepare for a succession. The vice president will serve as president until elections can be held. Negotiations with Anice continue until arrangements have been made for a public funeral.
Keep in mind that this development has implications for both countries. Recall that the president of Izeria has shown great interest in these talks and has closely monitored their progress. His death may pose a challenge to finding solutions for the issues including efforts at cooperation in the fight against terrorism. Remember that you have a limited amount of time to try to resolve the issues being negotiated.
As expected, none of the dyads reached agreement prior to receiving the message. We assessed the impacts of the new information on negotiator perceptions immediately after the messages were read.
Questionnaire Measures
We measured perceived trustworthiness using items from the Trust Scale developed by Roy Lewicki, Mary Stevenson, and Barbara Bunker (1997). This scale measures calculus-based trust, which is based on an analysis of the cost and benefits of violating trust (reliability coefficient for our data: 0.49); knowledge-based trust, which is based on how predictable one perceives the other person to be (reliability coefficient: 0.62); and identity-based trust, which is based on the degree to which the two parties share common goals and values (reliability coefficient: 0.79), using a 1-point (“strongly disagree”) to 7-point (“strongly agree”) scale. High scores indicate high trust.4
To measure impressions, we asked negotiators to rate the other party using a set of bipolar adjectives. We started with a list of fourteen adjectives that we selected to measure communality and power, and then applied a principal components factor analysis with a varimax rotation, which yielded three dimensions: communality, power, and morality.
We used three pairs of adjectives to measure how participants perceived communality: “uncooperative or cooperative,”“inflexible or flexible,” and “task-oriented or relationship-oriented” (reliability coefficient: 0.83). We measured participants' sense of relative power using the adjective pairs “powerful or powerless” and “weak or strong” (reliability coefficient: 0.93). Finally, we assessed the participants' judgments of their counterparts' morality using two adjective pairs: “dishonest or honest” and “bad or good” (reliability coefficient: 0.62). Participants used a 1 to 7 scale to rate the other party on these bipolar adjectives. We calculated a composite rating across participants for each dimension by averaging across adjective sets.
We also assessed the impact of the external precipitant (independent variable). After receiving the message, negotiators indicated whether the message a) strongly encouraged them to reach agreement on this issue now, b) served to change their thoughts on what was possible in this negotiation, c) had no effect, or d) encouraged them to end discussion of this issue without an agreement now. Preliminary analyses revealed that c) and d) were rarely chosen. Consequently, for the analyses, we collapsed across b) to d), all of which were clearly differentiated from item a) (strongly encouraged to reach agreement).
Finally, we assessed negotiators' outcomes by asking them to indicate how far they moved from their initial position in those cases in which an agreement was reached and how far they had been willing to move if an agreement was not reached on the issue. Negotiators were also asked to summarize their joint plan on the terrorism issue.
Approach to Data Analysis
We used hierarchical linear modeling (HLM) to test our hypotheses. The use of HLM offers the advantage of enabling us to examine individual behaviors while controlling for dyadic membership (Bryk and Raudenbush 1992; Kenny, Kashy, and Bolger 1998). Because models with random slopes and intercepts cannot be estimated for dyadic data, our model allowed for a random intercept but fixed the slopes. Although we had no hypotheses about role (Izeria or Anice), we included this variable in all our models to allow for the possibility that it influenced the ratings, although we found that role did not predict any of our outcome variables.
We fit a model for each set of predictor variables (trust, impressions, impact). In two-level models, Level One predictor variables describe attributes of the individual and Level Two predictor variables describe attributes of the dyad. We entered ratings of trust or impressions into the equation as Level One predictor variables. The type of external precipitant (crisis, breakthrough) was entered as a Level Two predictor variable. In setting up the model, we specified interactions between our Level Two predictor and the Level One intercept, as well as between our Level Two predictor and trustworthiness or impressions.
Experiment One Results
Message Impact
These analyses examine the impact of the crisis (Hypothesis One) and breakthrough (Hypothesis Two) messages on negotiators' reactions. First, we evaluated the role of perceived power in mediating the relationship between the type of message and negotiators' reactions. We found that perceived power affected negotiators perceptions of what was possible.5 Negotiators who perceived the other party as powerful were more likely to report wanting to reach agreement immediately after receiving both messages.
Second, we evaluated the role of trust in moderating the relationship between the type of message and reactions to it. This analysis showed that competence (calculus-based) trust interacted with the external precipitant to shape perceptions of what was possible.6 As can be seen in Figure Two, when perceived trustworthiness was high (positive climate), the type of external precipitant did not affect what negotiators thought was possible. When perceived trustworthiness was low (negative climate), however, a crisis message was more likely to change thoughts on what was possible.
Outcomes
These analyses examine the impact of the crisis (Hypothesis Three) and breakthrough (Hypothesis Four) messages on negotiation outcomes. They focus on the way that outcomes are influenced by the type of message, its impact, and impressions (social cues). First, we evaluated the relationship between the impact of both types of messages and outcomes. We found that message impact affected outcomes.7 The negative relationship indicates that negotiators obtained poorer outcomes when the external precipitant changed their thoughts on what was possible, apparently diminishing their sense that agreement was possible.
Second, we examined the relationship between trust (positive or negative climates), type of message, and outcomes. We found that goodwill (identity-based) trust affected outcomes: more movement toward agreement occurred as the level of perceived trustworthiness increased.8 Our analysis also showed that outcomes are affected by an interaction between competence-based (calculus) trust and the type of external precipitant.9 As can be seen in Figure Three, when perceived trustworthiness was low, the type of external precipitant had no effect on outcomes. When perceived trustworthiness was high, however, a negative precipitant (crisis message) led to more movement toward agreement than did a positive precipitant (breakthrough message).10
Experiment One Discussion
The results provide support for Hypotheses One (crisis messages lead negotiators to reconsider the range of possible positions when social cues in the negotiation have been negative) and Hypothesis Three (crisis messages will result in less movement toward agreement when the negotiation was conducted in a negative social context than when the context was positive). We found that negotiators did indeed change their thoughts about what was possible after receiving a crisis message in the context of negative social cues, but they also moved fewer steps toward agreement in that environment.
Our results did not provide support for Hypotheses Two and Four, which predicted the impacts of the positive breakthrough message; thus, positive breakthrough messages had no impact. It appears that crises set in motion departures (rethinking) that have implications for agreements, but only in a negative low-trust climate. This relationship can be depicted as a path:
negative climate (low trust) + a crisis message → rethinking the possibilities → less progress toward agreements.
A crisis message can also lead to more frequent agreements when negotiators already trust one another. In this environment, the crisis did not lead to rethinking. This path takes the following form:
positive climate (high trust) + crisis message → no rethinking → more progress toward agreements.
Interestingly, negotiators who perceived the other party as being powerful were more eager to rush to agreement. This reaction did not, however, result in more actual movement toward agreement. In terms of the turning points framework, we found that negative (but not positive) precipitants produced departures with consequences for outcomes. The direction of the consequences depended, however, on the social cues, which defined the environment in positive or negative terms.
The findings from this experiment can help clarify the relationship between negotiation crises and turning points. Case analyses have shown that crises often precede departures in negotiation: escalations are typically followed by de-escalations (Druckman 2001) and crises provide an impetus for progress in difficult talks (Druckman 1986). These experimental findings suggest that reactions to crises may depend on the social context (Donohue 2004). Contexts that lead negotiators to rethink their strategies seem to discourage them from reaching agreements. In addition, the impetus for progress seems to occur only in trusting environments — it does not occur when negotiators mistrust one another. The rethinking process set in motion when there is mistrust seems to shift expectations for agreements lower. Crises can move negotiators in either direction, toward raised expectations for progress or lowered ones. Although the professional negotiators in most of the cases analyzed in the earlier studies seemed to revise their expectations in a positive direction, the opposite result is possible and did occur in a few cases. Analyses of cases that produce one or the other reaction to crises would be valuable next steps in research on turning points.
More generally, this study illuminates an important contribution of experiments such as these: they provide opportunities to collect data about perceptions and social cues. By shedding light on the factors underlying negotiators' reactions, these data expand our understanding of the processes that lead toward or away from agreements. They complement case analyses by asking negotiators to answer questions about their impressions. Real-world negotiators are rarely available in sufficient numbers to provide this sort of information. We implemented further probes along these lines in our next experiment on the impacts of internal precipitants.
Experiment Two: Internal Precipitants
In this experiment, we took a somewhat different direction. Following Olekalns and Smith (2005), we asked negotiators retrospectively to identify and describe salient events within the negotiation that they believe altered its course (internal precipitants). As was the case in Experiment One, we asked whether impressions of the other party affected negotiation outcomes. We added a further dimension to our analysis of turning points by asking whether the context within which the negotiation was conducted affected both the identification of and the consequences of turning points. To do this, we compared dyads in which both negotiators had the same goals (either cooperative or individualistic) to dyads in which each negotiator had discrepant goals from the other party (cooperative versus individualistic).
Consistent with past research, we expected negotiators with cooperative goals to obtain better outcomes. Past research has also shown that negotiators with individualistic goals express less trust in and form more negative impressions of the other party (Beersma and De Dreu 1999; Olekalns, Lau, and Smith 2002). Our interest, however, lies in the relationships between negotiators' goals, the impressions that they form of the other party, and their identification of internal precipitants. Extending our earlier argument, we expected that a negative negotiating climate would be established when negotiators had an individualistic goal and concluded that the other party was either not trustworthy, lacked “goodness,” or was more powerful. Conversely, a positive negotiating climate would be established when negotiators had a cooperative goal and concluded that the other party was either trustworthy, was “good,” or had low power. This leads us to hypothesize that:
Hypothesis Five: When the negotiating climate is positive (high trust, low power, cooperative negotiators), negotiators will report more cooperative turning points than when the climate is negative.
Hypothesis Six: When the negotiating climate is negative (low trust, high power, competitive negotiators), negotiators will report more competitive turning points than when the climate is positive.
Hypothesis Seven: When the negotiating climate is positive (high trust, low power, cooperative negotiators), negotiators will achieve better outcomes than when the climate is negative (low trust, high power, competitive negotiators).
Taken together, these hypotheses suggest that the negotiating climate influences the way negotiators interpret precipitants, which in turn can lead to more or to fewer agreements. The perception of precipitants moderates the relationship between climate and outcomes. These relationships are depicted in Figure Four.
Negotiating Climate, Identification of Internal Precipitants, and Negotiation Outcomes
Negotiating Climate, Identification of Internal Precipitants, and Negotiation Outcomes
Participants
Ninety-six first-year psychology undergraduate students at the University of Melbourne completed this experiment as part of their course requirement. The sample comprised sixty-eight females and twenty-eight males.
Method
Participants negotiated a simulated employment contract (Olekalns and Smith 2000). Written instructions assigned participants to the role of either an employer or an employee, described the task, and gave a payoff schedule that described the points awarded for each possible contract. We embedded in the instructions a social motive manipulation that encouraged participants to focus on maximizing either individual (individualistic motive) or joint (cooperative motive) outcomes. Participants were told that their goal was to get either the best possible outcome for themselves (individualistic motive) or the best possible outcome for both negotiators (cooperative motive). Of the forty-eight dyads, seventeen had cooperative goals, seventeen had individualistic goals, and fourteen had mixed goals. Each negotiation was conducted face-to-face and was videotaped for subsequent transcribing.
Negotiators were encouraged to reach agreement on eight issues. These issues fell into one of three categories: distributive, integrative, and neutral. Each negotiator had one neutral issue, that is, one low-priority issue that was worth no points. For recruiters, this was job assignment; for applicants, it was start date. Two issues (vacation, relocation expenses) were distributive. For these issues, the total number of points available for division between negotiators was held constant, and an increase in points for one negotiator (better outcome) necessarily meant a decrease in points for the other (worse outcome). The remaining four issues (salary, annual raise, location, additional benefits) were integrative. These issues allowed trade-offs and offered the possibility of value creation. By making trade-offs, negotiators could increase the number of points available for division, allowing them to simultaneously improve outcomes.
Questionnaire Measures
We used the same measures reported in Experiment One. Reliabilities for the three trust subscales were calculus-based (reliability coefficient: 0.45), knowledge-based (reliability coefficient: 0.64) and identity-based trust (reliability coefficient: 0.72). We again used a 1 (“strongly disagree”) to 7 (“strongly agree”) scale with high scores indicating high trust. Using the same adjectives described for Experiment One, we undertook a principal components factor analysis with varimax rotation, which yielded three factors. Six pairs of adjectives measured communality: “bad or good,”“insincere or sincere,”“unhelpful or helpful,”“divisive or cohesive,”“not well-intentioned or well-intentioned,”“gives control or takes control” (reliability coefficient: 0.83). We use the adjective pairs “powerful or powerless” and “weak or strong” to measure power (reliability coefficient: 0.58). Finally, we measured predictability using two adjective pairs: “unpredictable or predictable” and “unreliable or reliable” (reliability coefficient: 0.79). Participants used a 1 to 7 scale to rate the other party on these bipolar adjectives. We calculated a composite rating for each dimension by averaging across adjective sets and across participants.
Identification and Classification of Internal Precipitants
The verbal protocol procedures used in relationship research enable researchers to directly access negotiator cognitions and to track changes in them over time. We adapted this procedure to elicit internal precipitants from participants. Immediately after completing the simulation, participants were asked to view a videotape of their negotiation. We gave them a booklet and asked them to note which events in the talks they considered critical in producing change in their negotiations and why. These protocols formed the basis for our analysis. We then classified the events as broadly cooperative (i.e., they focused on interests or made positive attributions about the other party's behavior) or broadly competitive (they focused on offers or made negative attributions about the other party's behavior).
Approach to Data Analysis
We again used HLM to test our hypotheses. In two-level models, Level One predictor variables describe attributes of the individual (trust, impressions) and Level Two predictor variables describe attributes of the dyad (number of cooperators). To capture the number of cooperatively oriented negotiators, we created two dummy variables as our Level Two variables. The first dummy variable contrasted dyads with no cooperatively oriented negotiators against dyads that had at least one cooperatively oriented negotiator. The second dummy variable contrasted negotiators who had the same goals (either cooperative or individualistic) against dyads in which negotiators had mixed goals (one cooperative negotiator, one individualistic negotiator). In setting up the model, we specified interactions between our Level Two predictor and the Level One intercept, as well as between our Level Two predictor and trustworthiness or impressions.
Experiment Two Results
Internal Precipitants
This experiment's first hypothesis (Hypothesis Five) posited the relationship between the negotiating climate (impressions, trust) and the number of cooperative internal precipitants that negotiators identified. Overall, our analyses showed that the identification of precipitants was linked to impressions of the other party along the dimensions of power and predictability but was not linked to perceived trustworthiness. We found that dyads with at least one negotiator rated as cooperative reported more cooperative internal precipitants than dyads with no member who was rated as cooperative.11 In addition, dyad composition interacted with perceived power12 and perceived predictability13 to affect the number of cooperative precipitants identified by negotiators. As can be seen in Figure Five below, we identified the greatest number of cooperative precipitants in negotiations in which there was at least one cooperative negotiator and the other party was perceived as having low power. Figure Five also shows a cross-level interaction: when the group (dyad) had at least one cooperative negotiator and the participant perceived the other party as highly predictable, then he or she reported a high number of cooperative precipitants; when the group had no cooperative negotiators and the respondent perceived the other party as highly predictable, he or she reported the lowest number of cooperative precipitants.
Number of Cooperative Precipitants as a Function of Number of Cooperators and Perceived Power or Perceived Predictability
Number of Cooperative Precipitants as a Function of Number of Cooperators and Perceived Power or Perceived Predictability
The second hypothesis (Hypothesis Six) focused on the predictors of competitive internal precipitants. We found that dyads with at least one cooperative negotiator reported fewer competitive internal precipitants than dyads with no cooperative negotiators.14 In addition, dyad composition interacted with perceived power15 to affect the number of competitive precipitants identified by negotiators. As can be seen in Figure Six below, perceived low-power dyad composition did not affect the number of competitive precipitants that we identified. When the other party has high perceived power, however, more competitive precipitants emerge when both parties have the same goals than when they have different goals.
Number of Competitive Precipitants as a Function of Dyad Composition and Perceived Power
Number of Competitive Precipitants as a Function of Dyad Composition and Perceived Power
Outcomes
The final hypothesis (Hypothesis Seven) examined the relationship between the negotiating climate (trust, impressions) and outcomes. Overall, our analyses showed that outcomes were linked to the perceived trustworthiness of the other party but not to impressions (positive or negative). We found that negotiators obtained better outcomes when there was at least one cooperative negotiator in the dyad.16 We also found that outcomes were affected by an interaction between trust and the number of cooperators in a dyad,17 as well as an interaction between trust and negotiator goal congruence.18 As can be seen in Figure Seven, negotiators obtained the highest outcomes when competence (calculus-based) trust was low and there was at least one cooperator in the dyad. Figure Seven also illustrates that negotiators obtained their poorest outcomes when competence-based trust was high and both parties had the same goals (either cooperative or individualistic).
Experiment Two Discussion
The analyses from Experiment Two indicate that the identification of cooperative and competitive precipitants was influenced by the negotiating climate, which support Hypotheses Five and Six. When dyads had at least one cooperatively oriented negotiator, their members reported more cooperative precipitants and fewer competitive precipitants. This basic pattern was amplified by the impressions that negotiators formed of the other party: negotiators reported more cooperative precipitants when the other party was perceived as low in power or high in predictability. Conversely, negotiators reported fewer cooperative turning points in a negative negotiating climate (no cooperatively oriented negotiators and high predictability). Overall, these findings support our hypothesis that the more benign a negotiating climate, the more likely it is to be punctuated by cooperative precipitants.
Interestingly, we found that whereas the identification of cooperative precipitants was influenced by the number of cooperatively oriented negotiators in a dyad, the identification of competitive precipitants was influenced by whether negotiators had the same or different goals, which suggests that the basis for a positive climate is created when negotiators have the same goals whereas the basis for a negative climate is created when negotiators have different goals. The studies of cross-cultural negotiations that have demonstrated that individuals in same-culture negotiations are better able to maximize joint gains than negotiators in mixed-culture dyads support this assumption (Adair, Okumura, and Brett 2001).
Our results also suggest that when negotiators have the same goals, low perceived power reinforces the positive climate and reduces competitive precipitants, whereas high perceived power disrupts the positive climate and increases competitive precipitants. A different dynamic appears to operate when negotiators start with a negative climate (mixed goals). Under these less benign conditions, low perceived power increases the number of competitive precipitants whereas high perceived power reduces the number of competitive precipitants. These patterns suggest that competitive precipitants reflect an internal power struggle. How this struggle plays out depends on whether negotiators have the same or different goals.
Finally, we found that both the number of cooperatively oriented negotiators in a dyad and the levels of trust influenced outcomes. Consistent with Hypothesis Seven, we found that dyads with at least one cooperatively oriented negotiator obtained higher profits than all-individualistic dyads. The perceived trustworthiness of the other party affected outcomes in dyads with at least one cooperative negotiator but not in dyads in which both parties were more individualistic.
Contrary to our expectations, however, we found that cooperatively oriented negotiators (and those in same-goal dyads) were more likely to obtain better outcomes when they believed the other party to be untrustworthy than when they believed the other party was trustworthy. Consistent with Dean Pruitt's (1981) dual-concern model, these findings suggest that a benign environment may result in premature closure. The additional tension created by a lack of trust seems to generate the necessary persistence to sustain a longer negotiation, resulting in better outcomes.
General Discussion
The results obtained from both experiments show that perceptions of trust and power — referred to in this study as social cues — serve to filter interpretations of events that precipitate turning points. These interpretations also influence negotiating outcomes. In Experiment One, the precipitants consisted of external events that were injected into the negotiating process. We defined these events as either a crisis (negative precipitant) or a breakthrough (positive precipitant). In Experiment Two, the precipitants were created by the negotiators themselves within the process.
We found that competence-based trust interacted with context to affect outcomes in both experiments. When negotiators expressed low trust in the other party, outcomes were unaffected by the context within which negotiations occurred. When trust was high, however, negotiators obtained better outcomes in a negative context (crisis message inExperiment One, mixed goals in Experiment Two) than in a positive context. We conclude that outcomes are improved when trust offsets concerns about the other party's intentions whereas they are not improved when high trust combines with an already positive environment. In the latter situation, negotiators may lack the necessary motivation to push for improved outcomes. This suggests the possibility of an upper threshold for positive information that, when reached, produces no further improvement; put another way, the relationship between positive cues and outcomes may be curvilinear, rising to a point and then flattening out. These findings also help explain the variation obtained among the cases analyzed in the earlier comparative study of turning points (see Druckman 2001).
An interesting difference between the two experiments is the role played by identity-based trust. Consistent with our earlier argument, although affective trust predicted outcomes in Experiment One, it did not predict outcomes in Experiment Two. One interpretation of this finding is that the two experiments primed different identity orientations (Flynn 2005). In Experiment One, the manipulated crisis and breakthrough messages were not under the negotiators' control. They presented negotiators with a common experience that may have primed a shared (relational) identity and oriented them to the reciprocal aspects of the interaction, a sense that they were both “in the same boat.” Moreover, the crisis message was likely to be perceived as a threat, further strengthening in-group identification. Thus, affective trust influenced outcomes and better outcomes were obtained when negotiators recognized their shared identities, resulting perhaps from a shared lack of control over the external events. In Experiment Two, the more transactional context likely primed independent rather than shared identities (Flynn 2005). As a result, negotiators perceived themselves to be independent and focused on the competence of the other party rather than on a shared identity.
Also consistent with our earlier discussion, the two experiments differed in the role played by impressions of power and predictability. Impressions interacted with negotiators' goals to influence the identification of precipitants in Experiment Two: more positive precipitants were identified when at least one of the negotiators had a cooperative orientation and the other party was perceived as having low power or as being highly predictable. The same social cues, however, did not influence the impact of external messages in Experiment One. These findings suggest that impressions may be influenced by attributions about the origin of precipitants. When events are attributed to the context, as was the case in Experiment One, their interpretation is not filtered through impressions of the other negotiator. When events are attributed to the person, however, as was the case in Experiment Two, negotiators use social cues to interpret the other party's actions. This pattern of findings suggests an interesting refinement in our model of precipitants, identifying attributional processes as playing a key role in moderating the impact of social cues. The key here is thedistinction between external and internal attributions. Our impressions of the other negotiator as trusting or powerful influence the way we interpret events that precipitate turning points.
Impressions derived from social cues play an important role in the negotiating process. The earlier case studies showed that progress will not occur in the kinds of security talks represented by the scenario used in Experiment One without external events or interventions. Those studies also showed that internal actions based on negotiators taking control over substance and procedure are needed for progress in the kinds of commercial transactions reflected in the scenario used for Experiment Two. We have learned that the influence of both events and actions may depend on the way they are interpreted by the negotiators, based on their perceptions of trust, power, and orientations, which, in turn, define the climate within which the negotiation process unfolds.
Taken together, the findings from both experiments enhance our understanding of the way precipitants operate during the negotiation process. The crisis and breakthrough messages manipulated in Experiment One provoked negotiators to question how they should coordinate in light of the new development. These questions would seem to raise concerns about competence-based trust, shown to influence negotiators' reactions to the messages. Indeed, low trust led negotiators to change their minds about the possibilities of reaching agreement when crises occurred. Perhaps it led to a sense of urgency as negotiators reduced their expectations about possible outcomes.
The competitive and cooperative precipitants identified by the negotiators in Experiment Two were under their control. These were actions taken to move the talks forward and depended for their effectiveness on coordination (which reflects competence-based trust) and perceptions of power. Both of these social cues influenced the dependent variables: competence-based trust influenced outcomes while perceived power influenced the number of cooperative (more when power was low) and competitive (more when power was high) precipitants identified.
Interestingly, competence-based trust played a role in both experiments. That role seems to be related to the need to coordinate in light of sudden events (Experiment One) or in balancing cooperative with competitive moves (Experiment Two). A difference between the experiments, however, is the role played by affective trust in relation to a common experience and by power in conjunction with initiated actions. These may be fundamentally different processes that are moderated by different social cues.19 The type of cue elicited (trust, power), however, may depend on whether the negotiator is reacting to events or initiating an action. This is an issue that warrants further research on turning points.
NOTES
The research reported in this article was partially supported by a Discovery Grant from the Australian Research Council and by the department of psychology at the University of Queensland, Brisbane, Australia. We thank Charles Liu and Paul Jensen for assistance in data collection, transcription, and analysis. A version of this manuscript was presented at the 2008 International Association of Conflict Management Conference, Chicago.
Out-of-keeping acts are behaviors that are inconsistent with previous strategies used in the negotiation or with the overall tone of the negotiation, for example, a competitive behavior introduced after a prolonged period of cooperation (McGinn, Lingo, and Ciano 2004). Turns are behaviors intended to change the direction of the negotiation, for example, by redirecting an attempt to exert power toward problem-solving (Kolb 2004).
By convergent validity, we mean that constructs that should in theory be related to each other are, in fact, observed to be related to each other. Thus, similar findings about inferential cues obtained from the two experiments provide evidence for convergent validity.
Honors classes are taken by fourth-year students selected to be in the honors program at universities in Australia. This is a year of study between undergraduate and graduate school. Undergraduate degrees are three-year programs in Australian universities.
For details on the meaning and computation of this reliability coefficient, including upper limits, see Cronbach (1951).
t(23) = −4,24, p < .001.
t(23) = 2.69, p < .05.
t(27) = −2.92, p < .01.
t(33) = 2.28, p < .05.
t(33) = −3.52, p < .05.
In this experiment, we conceptualized trust as an attribution formed rapidly on the basis of initial impressions (see McKnight, Cummings, and Chervany 1998). We did not gather explicit information about the factors that triggered those impressions.
t(45) = −3.04, p < .005.
t(84) = 2.20, p < .05.
t(84) = −2.88, p < .01.
t(45) = −2.72, p = .01.
t(84) = 1.99, p < .05.
t(45) = −2.1, p < .05.
t(86) = 2.74, p < .01.
t(86) = 1.06, p < .05.
This is similar to the two parts of the boundary role conflict. It has been argued that responding to the preferences of one's own side, in the role of representative, and to the demands made by the other, in the role of bargainer, involve fundamentally different processes (Druckman 1977).
