How Valid Is Negotiation Research Based on Student Sample Groups? New Insights into a Long‐Standing Controversy

In general, professional negotiators are too busy to participate in research projects, especially when they perceive the research as time consuming, which is invariably the case in experimental negotiation research (Tomaskovic‐Devey, Leiter, and Thompson 1994). Because it is difficult to recruit professionals to participate in negotiation research projects (Moore and Murnighan 1999), most empirical negotiation research relies on student samples (Ramsay 2004; Agndal 2007). In this context, a recently published study shows that only 3 percent of empirical negotiation‐related studies are based on the experience of practicing managers (Buelens et al. 2008).

Although the use of students as subjects has been standard practice in negotiation research, much controversy has arisen about the general applicability of findings derived from research on these groups. Numerous researchers are skeptical that behaviors demonstrated by students in areas in which they have little or no experience are representative of the behaviors of trained professionals (Weingart, Prietula, and Hyder 1996; Pullins et al. 2000). Proponents, on the other hand, argue that negotiation is a common enough activity, undertaken by anyone who must interact with others to resolve conflicting interests (e.g., when buying a car or in negotiations between spouses and friends) that the negotiation behavior of students accurately reflects the behavior of professional negotiators (Moore and Murnighan 1999). A third group of researchers attempt to avoid the issue entirely with a compromise: instead of psychology students, for example, they use students in Master of Business Administration (MBA) programs who have usually already gained some practical professional experience (Buelens et al. 2008).

In spite of the debate's long‐running history, which includes arguments for and against the use of students versus professionals (cf. Wall and Blum 1991), negotiation researchers have so far reached no consensus on the matter because no conclusive evidence has been found to either prove or disprove the contention that negotiation research carried out on students is equal in general applicability to that carried out on professional negotiators. In fact, scholars conducting investigations into the applicability of student samples have either trained previously inexperienced students in order to find whether such “manipulated” negotiation experts arrive at better negotiating results than students in a control group (Thompson 1990a, 1990b) or have set up situations in which professionals with “real” negotiation experience faced off against each other.

In the latter case, the results were then compared with those obtained in studies with inexperienced student negotiators (Neale and Northcraft 1986). But can the results gained using trained students in experimental negotiation situations be compared to results obtained using professional negotiators and, if so, to what extent? Because using professional negotiators as subjects can be so difficult, this question is of central importance to researchers undertaking empirical negotiation research: should no significant differences be found between results obtained with “manipulated” students versus “real” experts, many questions in negotiation research could then possibly be answered using trained students.

Against this background, the present study seeks to complement previous research by analyzing the performance of both professional negotiators and trained student negotiators in a multi‐issue simulation game. More concretely, we measured the individual and joint gains of professional negotiation experts and “manipulated” negotiation experts in a variety of negotiation dyads. Moreover, we integrated control dyads in which untrained student negotiators without any negotiation experience face off against each other, as well as against “real” and “manipulated” negotiation experts.

Negotiation is a decision‐making process in which at least two parties decide jointly how to allocate scarce resources (Pruitt 1983). The structure of the negotiation is determined by the parties' interests. Distributive negotiations exist when the parties' interests are, on the whole, negatively correlated (Walton and McKersie 1965), but negotiations can be integrative when they contain a potential for joint gain (Follet 1940; Raiffa 1982; Thompson 1990c). In practice, most negotiation results are neither exclusively distributive nor exclusively integrative in character (Walton and McKersie 1965).

In the actual modern workplace, negotiation has become a frequent, everyday activity. A recent study indicates that management‐level professionals spend approximately 20 percent of their time negotiating in diverse business areas (Mestdagh and Buelens 2003). These negotiations usually involve repeated interactions with others within and between organizations. Therefore, professional negotiators are presumed to have had significant bargaining experience, both in general and in particular decision‐making situations (Neale and Northcraft 1986), giving them not only a knowledge of general bargaining principles but also familiarity with specific decision‐making tasks (Goldstein 1993). Thus, a major difference between student and professional negotiators would lie in the extent of bargaining experience that they have obtained (Mestdagh and Buelens 2003).

Research examining the impact of experience on negotiation performance has fallen broadly into two categories: investigations of “real” negotiation experience and examinations of “manipulated” negotiation experience. In the first category, the negotiation performance of research participants with a history of professional bargaining is compared to that of untrained and inexperienced student negotiators. In the second category, the influence of experience gained by “manipulation” through controlled multiple trials or via diverse types of negotiation training is examined. We will turn first to the above‐mentioned examinations of manipulated experience. The primary reason for this is that studies of manipulated experience — possibly because of the difficulties of obtaining the services of professional negotiators for experiments — account for the majority of studies that deal with negotiation experience.

In studies of manipulated experience, scholars have mostly analyzed the effect of diverse types of experience (i.e., process or task experience) on negotiation performance (Thompson 1990a; Goldstein 1993). Such studies sought to analyze negotiation performance either on the basis of individual or joint gains achieved by the negotiation parties or on the basis of their ability to apply effective negotiation skills. Moreover, scholars often take the effect of experience on a negotiator's accurate judgment of his opponent's interest as a measure of performance. This is due to the fact that many studies emphasize the key role of understanding the interests of the opponent in achieving better negotiation performance (Thompson and Hastie 1990; Steinel, Abele, and De Dreu 2007). In addition to studies that have analyzed the impact of experience on negotiation performance in general, a modest body of research has also identified the factors that can either intensify or detract from the impact of negotiation experience on final performance. Independent of the object of investigation, all these studies started by measuring the performance of students after they received some kind of bargaining training (Weingart, Prietula, and Hyder 1996; Tinsley, O'Connor, and Sullivan 2002) or as they accumulated experience through multiple trials (Thompson 1990a, 1990b).

Max Bazerman, Thomas Magliozzi, and Margaret Neale (1985) were among the first to analyze the influence of experience on the achievement of joint outcomes. Students were randomly assigned to be either buyers (retail stores) or sellers (manufacturers of refrigerators). Thereafter, they were instructed to negotiate a specific conflict of interest as often as they could with different partners within a certain amount of time (usually about twenty‐five minutes). The results indicated that negotiation contracts became more integrative as the bargainers gained experience with their negotiation tasks. At first, the negotiators generally arrived at distributive agreements, but later they tended to engage in more integrative behavior. According to Bazerman, Magliozzi, and Neale (1985), these results confirm the often faulty “fixed‐pie assumption” to which many negotiators fall prey when entering a negotiation (Bazerman and Neale 1983). In several subsequent studies of experimental bargaining markets, researchers also found that negotiation contracts become more integrative as bargainers complete more transactions of the same type (Neale, Huber, and Northcraft 1987; Moran and Ritov 2007; Steinel, Abele, and De Dreu 2007).

Although these results indicate that experience has a powerful impact on the degree to which negotiators reach integrative outcomes, they fail to explain whether experience in general leads to more effective bargaining results or whether its impact is limited to a specific negotiation task. The reason for this is that the improvement in negotiation performance might have been due to the fact that the negotiators gained knowledge of the best solution in a particular negotiation scenario rather than because they gained a comprehensive understanding of effective general negotiation principles (Thompson 1990a, 1990b; Gentner, Loewenstein, and Thompson 2003).

Against this background, scholars have also examined the effect of experience in different negotiation settings. In this context, Leigh Thompson (1990b) compared the performance of negotiators who engaged in the same task on two separate occasions with the performance of those who encountered a similarly structured but different task in the second negotiation. Both groups achieved a higher joint result in the second negotiation with no significant differences between the two groups in terms of their performance (Thompson 1990b). Additional experiments in which students conducted multiple bargaining sessions had similar results (Thompson 1990a, 1990b; Thompson and DeHarpport 1994; Bereby‐Meyer, Moran, and Unger‐Aviram 2004).

More integrative results were also found when initially naïve negotiators received bargaining training prior to the negotiations (Weingart, Prietula, and Hyder 1996). In contrast to a control group, untrained students provided with instructions of both integrative and distributive negotiation tactics achieved a higher level of Pareto efficiency and engaged more often in integrative behavior, although only 20 percent of the members of the group that had received this information achieved optimal solutions in a simple bargaining scenario (Weingart, Prietula, and Hyder 1996).

Gaining negotiation experience has also been shown to improve the individual achievements of negotiators, as experienced negotiators outperformed inexperienced negotiators in a series of experiments (Thompson 1990a; Murnighan et al. 1999; Tinsley, O'Connor, and Sullivan 2002). But Thompson (1990a) also found that this advantage vanished as inexperienced negotiators improved their performance via repeated experience. More concretely, she found that the difference between experienced and inexperienced negotiators was greatest when the inexperienced bargainer had no prior experience; it diminished significantly even when the inexperienced bargainer had had only a single previous bargaining experience (Thompson 1990a).

What explains these improvements in negotiation performance? To gain greater insight into the experience–performance relationship in negotiations, researchers have examined the impact of experience on negotiators' abilities to apply effective skills in different negotiation situations (Thompson 1990a, 1990b). Two skills examined in particular were logrolling, in which the negotiators make trade‐offs regarding the issues such that each party realizes some of its preferences on important issues in exchange for a less preferred outcome in less important issues (Froman and Cohen 1970), and the identification of compatible interests.

Whereas experience was clearly shown to improve logrolling skills, usually no measurable effect was found on negotiators' ability to identify compatible issues (Thompson 1990a, 1990b; Thompson and DeHarpport 1994). In one bargaining experiment, the results showed that subjects were even less likely to recognize compatible issues the more trials they completed (Thompson 1992). One possible explanation might be that there is considerable discrepancy between truly compatible issues and how they are perceived; that is, negotiators may intuitively take a fixed‐sum approach in regarding their priorities rather than trying to find issues compatible with those of their opponents (Thompson 1992).

Similar results were found concerning the influence of experience on negotiators' judgment accuracy. Although experience has been shown to improve the negotiator's assessment of the other party's interests in general, a closer look has shown that this improvement is due only to an enhanced perception of the relative importance of issues to the other party (logrolling accuracy). There is no evidence, however, that experience improves the correct assessment of concrete negotiation positions, for example. In contrast to this, a series of experimental studies has shown that compatibility accuracy did not improve as the negotiators completed more rounds of negotiation (Thompson 1990a, 1990b, 1992; Thompson and DeHarpport 1994). In addition to this, Simone Moran and Ilana Ritov (2007) found that experience does not improve the negotiators' ability to assess the other party's values for specific offers. This result is noteworthy because an accurate assessment of the specific gains that one's opponent would realize from a particular offer has been shown to enhance integrative performance more than a mere assessment of the other party's general profit schedules (Moran and Ritov 2007).

Results from studies in which experience gained through multiple bargaining sessions was enhanced by immediate feedback were also inconclusive (Thompson 1992; Thompson and DeHarpport 1994). Negotiators provided with feedback were no more accurate in identifying both logrolling and compatible issues than were subjects in a control group. Such findings suggest that feedback‐enhanced experience is generally no more effective than experience alone (Thompson 1992). This is surprising because experimental investigations in the more general field of behavioral decision theory have found process feedback to be effective in improving decision‐making performance (Hammond, Summers, and Deane 1973). Presumably, feedback provides people with information that enables them to compare their present behavior with ideal strategies (Hogarth 1981).

The judgment accuracy of experienced negotiators also did not improve when additional information was provided about the other party's preferences. Although Keith Murnighan and colleagues (1999) found that experienced negotiators obtained superior outcomes to those of inexperienced negotiators, it must be noted that experienced negotiators performed better than inexperienced ones regardless of whether they had received preference information or not. The authors argue that this is because prior experience taught negotiators that they needed to gather as much information as possible during the bargaining process (Murnighan et al. 1999). In this study, experience was again manipulated via training of inexperienced student negotiators prior to the negotiations.

It seems that the positive influence of negotiation experience alone on different performance criteria cannot easily be enhanced. In contrast, Catherine Tinsley, Kathleen O'Connor, and Brandon Sullivan (2002) found less of a correlation between experience and performance when reputation became a factor in the negotiation. In a large‐scale experiment that investigated the impact of reputation on negotiation behavior, researchers found that the negotiation outcomes of experienced negotiators were no better than those of the inexperienced bargainers they were matched with when the novice negotiators had been informed that the experts had a reputation for distributive bargaining (Tinsley, O'Connor, and Sullivan 2002).

What is to be learned from these results? Prior experience — be it acquired through the completion of multiple bargaining trials in succession or through training — leads in general to a better negotiation performance. The studies described above suggest that experienced negotiators achieve better integrative outcomes and obtain more resources than their inexperienced counterparts, regardless of whether the negotiation settings change or not. Other results indicate, however, that the experience–performance relationship should not be overestimated. For example, most studies found that a successful transfer of bargaining skills to different tasks was not easily achieved. Improvements in bargaining behavior were mostly restricted to logrolling skills, even though the negotiation tasks used to simulate changing negotiation settings were unrelated to the level of diversity that negotiators encounter in real‐life situations. Moreover, even when experience improved the negotiators' assessment of the other parties' interests, this improvement was limited to a better judgment of logrolling issues.

On the whole, therefore, it is clear that cognitive biases remain intact even as performance improves with negotiation experience. Past research suggests that the improvement in negotiation performance is achieved intuitively, rather than through an enhanced explicit, conscious understanding of the underlying negotiation structure (Thompson 1990a, 1990b; Moran and Ritov 2007). This result has been confirmed by studies that showed that inexperienced student negotiators' bargaining results significantly improved after a single negotiation experience (Thompson 1990a; Steinel, Abele, and De Dreu 2007). Likewise, inexperienced students' performances improved after they were exposed to a single training unit (Weingart, Prietula, and Hyder 1996; Nadler, Thompson, and Van Boven 2003). Moran and Ritov (2007) explain this phenomenon by arguing that subjects experience an “Aha!” moment without developing a deeper understanding of what constitutes an effective strategy and the reasons for that strategy.

According to past research on learning, this is because negotiators often focus on information that has superficial rather than structural similarity to the decision‐making process in which they are involved (Loewenstein and Thompson 2000; Gentner, Loewenstein, and Thompson 2003). Furthermore, Margaret Neale and Max Bazerman (1991) have noted that high‐quality outcomes can be achieved in one of three ways:

1. negotiators may randomly stumble upon effective strategies,

2. negotiators may learn particular strategies that are effective for a given situation, or

3. negotiators may acquire a common core of knowledge that explains when and why particular strategies are effective in different situations.

The second and third methods refer to an important distinction in the decision‐making literature between the terms “experience” and “expertise” (Dawes and Corrigan 1974): experience is defined as a form of repeated feedback that can cause unconscious behavioral adaptations without requiring a deeper understanding, while expertise is defined as an ability to effectively transfer strategies to different circumstances (Neale and Northcraft 1990).

Altogether, the studies on “manipulated” experience thus demonstrate that having or gaining more negotiation experience can improve a negotiator's performance but will not guarantee that he or she becomes a negotiation expert (Thompson 1990a). Whether this finding is also applicable in general to “real‐world” negotiations taking place beyond university walls is a separate question.

As noted, far fewer scholars have examined the experience–performance relationship on the basis of a professional sample, that is, “real” negotiation experts. Roland Scholz, Andreas Fleischer, and Andreas Bentrup (1983) compared the bargaining performance of buyers for department stores with that of student negotiators. Whereas the professional cohort resolved conflicts more quickly, joint outcomes for the two groups did not differ. A study conducted by Margaret Neale and Gregory Northcraft (1986) produced a different finding, however. They compared data from negotiations conducted by professional corporate real estate negotiators with data from a previous study of untrained student negotiators (Bazerman, Magliozzi, and Neale 1985). Their results indicate that although both groups reached increasingly integrative solutions as they gained experience, the experts were significantly more successful in reaching integrative outcomes. The authors found no difference, however, between the two groups with regard to the impact of framing (i.e., the perceptual bias a negotiator takes when evaluating alternatives) and performance constraints (negotiators' given goals and limits). The results suggest that the patterns of negotiation behavior, as influenced by framing and performance constraints, were similar in the two groups (Neale and Northcraft 1986).

Researchers also observed similar patterns of decision‐making behavior with regard to the biasing potential of the “anchoring‐and‐adjustment” heuristic (Neale, Huber, and Northcraft 1987). According to this heuristic principle, negotiators become prejudiced by arbitrarily selected values, such as opening offers (“anchors”) and subsequent objective valuations may fail to dislodge these anchors, thus creating a decision bias (Tversky and Kahneman 1974). In one experiment, students and professional real estate agents were asked to make pricing decisions about real estate properties in a real‐world setting; both groups were significantly biased by listing prices that had been provided to them prior to the investigation (Neale, Huber, and Northcraft 1987).

At first glance, these findings paint an inconclusive picture of the influence of “real” experience on negotiation performance. Whereas professional negotiators tend to achieve more integrative outcomes than inexperienced negotiators, the decision‐making behavior of both groups appears to be similar. Consequently, these studies reveal that having real‐world experience also does not immunize negotiators against faulty mental representations and judgment errors.

The results of these studies are comparable to the results of research on manipulated experience. For example, Margaret Neale and Gregory Northcraft's (1986) results suggest that professionals become more integrative in their negotiation practice as they gain experience in the market. A similar result was obtained from the studies in which novice negotiators were asked to complete multiple negotiation agreements (Thompson 1990a, 1990b). In addition, the fairly high impact of cognitive biases on negotiation behaviors suggests that professionals and students alike often act more intuitively than intelligently: if “real” experience had taught them to strategically conceptualize how and when to apply certain negotiation principles, the professionals should have performed better in incorporating their past learning to their experimental scenarios.

As a consequence, both real and manipulated experience seems to enhance negotiation performance without leading automatically to the development of true expertise. Rather, as the studies of both manipulated and real experience show, simply having negotiation experience is not enough when the issue is a fundamental improvement in negotiation performance (Loewenstein and Thompson 2000; Bereby‐Meyer, Moran, and Unger‐Aviram 2004).

Although these findings provide valuable insights into the impact of experience in negotiation in general, they do not adequately answer the question of whether student test groups are suitable for carrying out negotiation research and, if so, to what extent, because up to now, the studies discussed above have shown only that the performance of both students and professionals can be improved by conducting multiple trials or receiving some kind of negotiation training. At the same time, they show that both manipulated and real experience does not always lead, by themselves, to optimum negotiation behavior. These studies have not, however, provided data on the extent to which experimental results gained by a manipulation of experience correspond to the bargaining behavior of real negotiators because the negotiation performance of trained student negotiators has yet to be compared with that of real negotiators. As our literature overview shows, previous studies have mostly compared the negotiation performance of trained student negotiators with the performance of untrained student negotiators. Similarly, studies of experts have compared the gains achieved by professionals with those obtained by untrained student negotiators.

To gain further insight into the applicability of experimental findings from experiments in which the study sample comprises students, we investigated the extent to which trained students attained results comparable with those of professionals. For this purpose, we used a full experimental design in which we measured the negotiation performance of untrained students, trained students, and professionals in a variety of negotiation dyads. In the first scenario, students with manipulated experience faced off against each other as negotiation partners (MM). In the second, professionals with a background of real professional negotiation experience faced off against each other (PP). In the third scenario, professionals faced students with manipulated experience (PM).

The third scenario is most important for the purpose of our study, because a mere comparison between the samples MM and PP would be subject to the potential risk that the behavior of the negotiators among themselves would eventually reach a point of common equilibrium. Moreover, we integrated several control dyads in which a third group of negotiators, untrained students (N), faced off against each other, as well as against real and manipulated negotiation experts (NN, PN, MN). Based on the insights gained from previous studies — in particular, that students with manipulated experience achieve better results than untrained students and that professional negotiators and trained students resemble each other in their basic patterns of negotiation behavior — we propose the following hypotheses:

To test our hypotheses, we asked eighty‐two professionals, seventy‐eight trained students, and ninety untrained students to negotiate a buyer–seller transaction in the market of biocides.¹

To generate the professional sample, we used alumni databases from three German universities (University of Muenster, University of Hohenheim, and University of Tuebingen). We sent a personalized e‐mail message inviting fifteen‐hundred professionals with extensive negotiation experience to participate in a negotiation simulation. We focused on employees in sales and procurement departments who routinely participate in buyer–seller negotiations as well as professionals in management positions. As an incentive, we told prospective participants that they would receive a personal evaluation of their negotiation results that would illustrate their relative negotiation performance in comparison with that of the other subjects (cf. Appendix A).

We sent two reminder e‐mails to stimulate responses. Within five weeks, eighty‐two professionals from diverse companies indicated that they were willing to participate. From those, 49 percent hailed from upper management. The age of these professional negotiators ranged from 25 to 52 years, with a mean of 38.8 years. Participants reported that the mean amount of their business day spent in negotiation‐related activities was 39 percent. Additional information about the professional sample is provided in Table One.

The trained student sample comprised master's degree students, enrolled in the business program at the University of Stuttgart‐Hohenheim. All student participants had completed a comprehensive negotiation training course within the previous three terms (between April 2009 and July 2010). This training included an in‐depth introduction to the different streams of negotiation research, the analysis of different videotaped negotiation situations, and the role‐playing of two complex negotiation simulations (one purely distributive and the other mixed‐motive integrative), which enabled the students to experience how different negotiation approaches could improve their negotiation performance (Gentner, Loewenstein, and Thompson 2003; Nadler, Thompson, and Van Boven 2003). Nevertheless, prior to registration, we checked the level of negotiation experience by means of a questionnaire regarding negotiation approaches, strategies, and tactics. The participants in the trained student sample ranged in age from 20 to 30 years, with a mean age of 23.9 years.

We recruited the untrained sample from among new master's degree students enrolled in the business program of the same university as the group with manipulated experience. We used a questionnaire to ensure that the students had no practical experience in negotiating and that they had never taken part in a negotiation training before. They ranged in age from 22 to 30 years, with a mean age of 23.3 years.

Both student groups (trained and untrained negotiators) participated voluntarily in the experiment. As an incentive, we promised each student a certificate of participation and — if they requested — a report of their negotiation performance compared with those of the professionals. To protect confidentiality and anonymity, we generated a blind experimental procedure by assigning identification numbers (ID) to the students and the professionals. To limit the influence that knowledge of the opponent's negotiation experience might exert on participants, we gave them no information about the experience level of their counterparts. So, all students (untrained and trained) and all professionals believed that they were negotiating with students or professionals with a level of experience similar to their own. To prevent disclosure of the negotiators' identities during the negotiation process, we also urged all participants to take on the roles described in the business case, and they were explicitly told not to reveal their actual identities. The voluntary and blind structure of this experimental procedure was designed to avoid impression‐management efforts that could otherwise have been expected — in this experiment, students had no incentive to influence teachers or fellow students, as they might in many other experiments (Wood and Mitchell 1981).

All negotiators (students and professionals) were engaged in the same task. To facilitate participation of professionals from all over Germany, all negotiations were conducted via a specially developed online negotiation platform, which resembled a chat room and facilitated synchronous negotiation, avoiding the delays that plague negotiation experiments conducted via e‐mail. The platform also enabled an administrator to observe the negotiations and intervene if technical problems or other difficulties arose.

All online negotiations were conducted over the course of four days. In each case, they involved two people. During this period, the participants were asked to identify several slots of one hour each in which they would be available for the negotiation experiment. The negotiation dyads, required for testing our hypotheses — professional versus professional (PP), trained student versus trained student (MM), professional versus trained student (PM), novice student versus novice student (NN), professional versus novice student (PN), trained versus novice student (MN) — were then designed according to their time slot preferences and their type of experience (real, manipulated, or novice). Table Two illustrates the final distribution of the negotiation dyads.

One week before the start of the experiment, all participants received a password and their personal user identification via e‐mail. This enabled them to log into the platform at their reserved time slot. In the same e‐mail, the participants received detailed information about the negotiation situation. In order to make the scenario more realistic, we developed a complex simulation and supplied the participants with extensive background and role‐specific information. According to Stephen Weiss (2008), complex simulations help experimental participants to develop a deeper understanding of complex negotiation settings than standard role‐plays. The task of the participants was to negotiate a potential agreement between a supplier of additives for biocides and a manufacturer of biocides on three distributive issues: the price for 200 kg of additive, a cost‐sharing arrangement at the production plant, and a cost‐sharing arrangement for the transportation of the additive. Moreover, we included one integrative issue, which was the delivery date (cf. Appendix B). The task description for each party set the target of increasing profits. The participants were given one hour in which to reach an agreement.

As in the case of real buyer–seller negotiations, we did not hand out a payoff schedule, but the parties were able to calculate the monetary value of each offer on the basis of the information we provided. Both parties were thus able to determine which agreements would be best for them and which profits would result. We selected this procedure because there is no provision of payoff schedules in real negotiation situations (Moran and Ritov 2007). Overall, we distributed role assignments fairly evenly (supplier and manufacturer) across the field of novice and trained students, as well as of professionals. (See Table Two.)

Because our intention was to analyze the general applicability of results based on student samples with regard to the process dimension of experience, we developed a negotiation simulation in an industry — chemistry — that was not represented by the professional sample. Therefore,we can assume that the professionals had no more experience in the particular chosen negotiation task than did the students. This also ensured that differences in negotiation behavior and performance did not reflect knowledge of strategies that would be effective in that specific context. Rather, differences were more likely the result of process experience, because student subjects would naturally be less experienced at negotiating a deal.

Our main hypotheses (Hypotheses One and Three) concerned the results obtained by student negotiators who have no practical experience but have received some in‐depth training (manipulated experience). We investigated whether these results have general applicability to the negotiation performance of real negotiators. To test these hypotheses, we measured the single and joint negotiation outcomes of professional negotiation experts (P) and manipulated negotiation experts (trained students) (M) in three types of bargaining dyads (PP, MM, and PM). Moreover, we integrated control dyads in which untrained student negotiators faced off against each other and against real and manipulated negotiation experts (NN, PN, and MN). First, this full experimental study design allowed us to examine both differences in single and joint gains of all the types of negotiation parties (P, M, and N). Furthermore, it allowed us to check the effect of the manipulation of experience (Hypotheses Two and Four).

For the examination of single gains, we measured each party's individual gain by adding their negotiated monetary values for each issue. Because joint performance is an indicator of the integrativeness of agreements, we then summed up the parties' profits per dyad to create a measure of joint gain. To ensure a meaningful basis of comparison with regard to the distribution of roles, the standardized profits were used in each case for analysis.²

Hypothesis One concerned the single‐party gains realized by the professional negotiators versus the trained students. Our hypothesis that the experts would not receive higher scores than the students whose experience was manipulated was tested against the null hypothesis that the professionals would outperform their student counterparts. We analyzed all single gains for negotiators who bargained in the dyad constellation PM. The results indicate that real‐world experts and trained student experts tended to claim roughly the same single gains. Interestingly, with a mean single gain of = 0.09 euros (standard deviation [SD] = 0.67), the trained students actually achieved slightly higher results than the professionals, whose average single gain was mean = 0.07 (SD = 0.76). We used the analysis of variance (ANOVA) test for significance, however, and those results indicate that the difference between the groups was not significant.(See probability factors in Appendix Three.) Consequently, we see Hypothesis One as confirmed.

We also found evidence to support Hypothesis Two, which assumed that both trained students and real experts would achieve greater single gains than novice student negotiators. With an average achieved single gain of mean = 0.13 (SD = 1.06), professionals significantly outperformed novice negotiators whose average achieved single gain was mean = −0.46 (SD = 0.33) Our ANOVA analysis revealed this to be a significant difference. Trained students scored an average achieved single gain of mean = −0.06 (SD = 1.25) and thereby also significantly outperformed naïve ones whose average achieved single gain was mean = −0.68 (SD = 1.5).

According to Hypothesis Three, dyads in which both parties' experience was manipulated (MM) would not achieve lower joint gains than dyads in which only professionals (PP) negotiated against each other. With an average joint gain of mean = 0.33 (SD = 0.60) for the trained student dyads and mean = 0.38 (SD = 0.55) for the professional dyads, we found no significant difference. Furthermore, our results show that the difference was also not significant between the purely professional dyads (PP), with an average joint gain of mean = 0.38 (SD = 0.55), and those mixed dyads comprising trained students and professional negotiators, with an average joint gain of mean = 0.14 (SD = 0.68). This confirms Hypothesis Three.

Finally, we analyzed results relevant to Hypothesis Four. To test this hypothesis, we investigated whether both trained students and real experts would outperform novice student negotiators in terms of average joint gains. We first analyzed whether the average joint gains achieved in the negotiation dyads PP and MM differed from the average joint gains achieved in the negotiation dyad NN, which was mean = −0.03 (SD = 0.96). We found that both the differences between PP and NN and between MM and NN were significant.

In a second step, we also analyzed whether the mixed dyads PN (with mean = 0.09, SD = 0.81) and MN (with mean = −0.41, SD = 1.49) achieved lower joint gains than the pure expert dyads (PP and MM). We found that the mixed professional negotiator/novice student dyad (PN) achieved significantly lower joint gains than did the pure expert dyads (PP) and the pure trained student dyads (MM). The same was true for the mixed dyads of trained and untrained student negotiators (MN) when compared to the pure trained student (MM) dyads and pure expert dyads PP. Because dyads in which either both parties were novice negotiators or at least one party was a novice negotiator and the other was either a manipulated or a real expert always achieved significant lower joint gains than dyads in which both parties comprised either manipulated or real experience, we may also confirm Hypothesis Four.

We investigated whether student subjects who had received some kind of extensive bargaining training — whom we classify as “manipulated negotiation experts”— outperformed novice student negotiators in a simulated negotiation and whether these results are generally comparable to those of real negotiators. In doing so, we aimed to contribute further insights to the long‐standing controversy about whether experimental sample groups comprising students can legitimately and effectively be used for negotiation research.

Research carried out up to now has compared either the performance of novice students with that of students whose experience was manipulated (i.e., they were given some specific kind of negotiation training) or has compared the performance of untrained students with that of professionals. Against this background, we developed a full experimental study design in which we analyzed the individual and joint gains of trained and untrained student negotiators, as well as that of professionals who negotiate regularly in the workplace. The findings of our study go beyond the already established effect of experience on negotiation performance by also comparing manipulated student negotiators with real experts.

Generally, our data confirmed our hypotheses that students with some background of negotiation experience would outperform untrained student negotiators but are not outperformed by professional negotiators in terms of achieved single gains. Moreover, both the average joint gains of trained students and of professionals were significantly higher than those of novice student negotiators, although the average joint gains of trained students did not lag significantly behind the integrative level of agreements reached by professionals. Although these results appear striking at first glance, they are actually in line with previous research: negotiation performance seems to improve with repeated negotiation experience, whether that experience is gained in the same negotiation setting or in different situations (Thompson 1990a, 1990b).

The confirmation of our second hypothesis, in which we predicted that both trained students and professionals would outperform novice student negotiators in terms of achieved single gains, further confirms these findings. Previous research has also pointed out, however, that the effects of (repeated) negotiation experience tend to be limited, and the results of numerous studies on experience in negotiation have shown that even experienced negotiators often act more on the basis of intuition than on the basis of an actual expertise that might tell them which negotiation behavior will be most effective in a given situation.

More specifically, our finding that trained student negotiators are not outperformed by real experts, in terms of achieved single gains, confirms the results of prior studies that found significant differences in this area only when one of the negotiation partners had no previous negotiation experience at all (Thompson 1990a). As far as the joint gain is concerned, our results also support the results of Scholz, Fleischer, and Bentrup (1983), who found no significant differences when comparing professionals with novice student negotiators.

Our results have important implications for empirical negotiation research. They indicate that sampling with student groups is applicable in general — as soon as these groups have received some kind of specific negotiation training.

Nevertheless, the general applicability of samples taken with trained students has certain restrictions. This is because some caveats pertain to our comparison between a background of manipulated experience and a background of real experience, first and foremost to the composition of our trained student sample: in contrast to the vast majority of empirical negotiation studies (Thompson 1990a, 1990b; Nadler, Thompson, and Van Boven 2003; Bereby‐Meyer, Moran, and Unger‐Aviram 2004), we did not work with students of social psychology, but exclusively with graduate business students. In view of the inherent topical relationship between their studies and the context of negotiating buyer–seller deals, it is reasonable to assume that the manipulation of experience among business students leads to a greater improvement in negotiation performance than it might in the case of students of social psychology. In addition, the training we provided in this study must be regarded as a specific feature of our trained student sampling group: it was extensive and lasted for several weeks. Of interest, in this context, is the question of whether a shorter period of training or less intensive training would have produced comparable results. In addition, we ensured that all sample groups (P, M, and N) participated voluntarily in the experiment and had the same motivation (we gave all participants the possibility of receiving an individual negotiation performance evaluation). We felt that it was important to ensure that all participants receive the same form of motivation, because it has been found that motivation can have an important influence on negotiation behavior (McClintock 1972; Weingart, Prietula, and Hyder 1996).

Finally, it is important to note that all negotiations were conducted online. Opinions in the broad field of negotiation research are mixed regarding the impact of electronically conducted negotiations on negotiation behavior and outcomes (Croson 1999; Valley 2000; Voeth and Herbst 2005). Therefore, we cannot exclude the possibility that interactional influences associated with our study design played a role in our results.

The question of generalized applicability of samples taken with trained student groups thus also becomes a question of the conditions under which such student groups can be used for empirical negotiation research. In this regard, the above — mentioned characteristics of our study have provided concrete guidelines for further research. Moreover it should be remembered that our findings are related only to the outcomes of negotiations. Further research is thus needed to compare the actual negotiation behavior of trained and novice students with that of professionals. It will also be of equal importance to ask in greater detail to what extent the results remain comparable when trained students negotiate against professionals in a task environment that is familiar to the latter. It must therefore be remembered that our study reflects only the generalized applicability of negotiation process experience.

Although our results offer useful insights for empirical negotiation research, they should not stand in the way of the practice of working with professional groups. According to James Wall and Michael Blum (1991), professionals should especially be used in negotiating simulations in environments with which they are familiar, as a result of their profession. This might make it possible to investigate not only process experience, which has always occupied center stage, but task experience as well. In addition, the often‐heard argument that it is too difficult to work with real professionals should no longer be accepted as an excuse. In developing an online platform and providing personal evaluations to all professionals, we have demonstrated a practical way of motivating professionals to participate in negotiation experiments.

The main purpose of our research was to examine the general applicability of results based on student samples. Therefore, we distinguished between novice student negotiators, trained student negotiators, and expert professional negotiators. The results of our study confirm our hypotheses that students with some kind of “manipulated” negotiation experience outperform untrained student negotiators but that they are not significantly outperformed by professional negotiators. From this, we conclude that many questions in the field of negotiation research can be effectively tested by using trained students as experimental subjects.

This study complements the results of previous studies that have examined the impact of experience on negotiation behavior and its outcomes. We end by noting that the investigation of experience in negotiation is far from complete. The question of sample validity for the generation of further insights into negotiation research is too important to be left unresolved.