How, when and why do negotiators use reference points? A qualitative interview study with negotiation practitioners

Since the introduction of prospect theory (Kahneman and Tversky, 1979), the crucial role of RPs in human decision-making has been widely investigated. The term RP can be applied to many different evaluation standards; however, most research has focused on so-called outcome RPs, which serve as evaluation standards for potential or prospective outcomes that result from decisions among different alternatives (e.g. Tversky and Kahneman, 1981) ^[2].

In negotiation research, RPs are commonly conceptualized as “reference outcomes” (De Dreu et al., 1994, see also Gupta and Livne, 1988), which serve as evaluation standards to qualify potential outcomes resulting from negotiators’ own proposals or proposals by the other party. Importantly, RPs do not only serve as evaluation standards that differentiate between positive and negative negotiation outcomes (Kahneman, 1992). Building upon the extensive research on RPs, the present study conceptualizes RPs as reference outcomes that are used by negotiators as comparison standards to:

• gain orientation in the negotiation process on a cognitive level (e.g. categorize outcomes as realistic or unrealistic; Cohen, 2003);

• regulate their behavior on a motivational level (e.g. approach desired outcomes or avoid undesired outcomes; Larrick et al., 2009; Trötschel et al., 2013); and

• evaluate outcomes on an affective level (e.g. gains or losses; Bazerman et al., 1985).

Cognitive models in negotiation research focus on how negotiators process information to mentally construct their negotiation (Thompson, 1990, see also Carroll et al., 1988). Figuratively speaking, negotiators use RPs to develop a map of the negotiation territory by staking out realistic destinations based on current market prices, reference transactions or price expectations (Blount et al., 1996; Kahneman et al., 1986; Kristensen and Gärlinga, 1997), to ultimately gain orientation in the fog of negotiation (Wheeler, 2013). This line of research also focuses on negotiators’ perceptions about the other party, for instance estimations about opponents’ limits and the size of the bargaining range (Larrick and Wu, 2007; Prietula and Weingart, 2011). Based on this information, negotiators can develop a sense of (un)realistic or challenging (i.e. difficult to achieve) negotiation outcomes (see Cohen, 2003; Thompson, 2015). The aforementioned RP-based cognitive processes help negotiators to better understand the particular negotiation situation by using the processed information to assess potential outcomes as appropriate or inappropriate, realistic or unrealistic, ambitious or unambitious.

Pertinent motivational theories focus on the impact of goals and standards on negotiator behavior and outcomes (e.g. Brett et al., 1999; Thompson, 1990). Referring to the fog metaphor, this line of research is concerned with how negotiators use RPs to approach desired end destinations and make sure they are on the right track despite the foggy negotiation territory (Wheeler, 2013). Scholars typically differentiate between two concepts: first, the ideal outcomes negotiators hope to achieve (target points; Walton and McKersie, 1965), termed goals from here on (e.g. Huber and Neale, 1987). And second, the minimal outcomes they are willing to accept, which we hereafter refer to as limits (Pruitt and Lewis, 1975). The negotiation literature provides various terms for similar concepts that refer to negotiators’ bottom lines, such as reservation prices (Raiffa, 1982), resistance points (Walton and McKersie, 1965), minimal goals (Thompson, 1995), willingness to pay (WTP; Wertenbroch and Skiera, 2002) or walk-away points (Malhotra and Bazerman, 2007) ^[3].

Target points and limits represent important RPs, because they divide “the space of outcomes into a positive and negative region” (Heath et al., 1999, p. 84). Because goals shift more outcomes into the negative region, goals are typically more difficult to attain than limits (see Larrick et al., 2009). By contrast, limits are assumed to be more rigid than goals, as they define a minimum success level in negotiations (Walton and McKersie, 1965). Consequently, limits can be interpreted as thresholds, at which negotiators are (economically) indifferent between reaching an agreement with the other party and breaking off the negotiation (i.e. limits as disagreement point outcomes; Muthoo, 1999). Thus, it may not be surprising that various economists resort to limits to explain and predict transaction processes and outcomes (e.g. in search theory; Jäger et al., 2024, in consumer research; Wertenbroch and Skiera, 2002, in game theory; Muthoo, 1999).

Empirical findings illustrate that negotiators with high goals (or limits) are more demanding (Appelt et al., 2009; Galinsky et al., 2005), show a higher resistance to concede (Smith et al., 1982; Trötschel et al., 2013) and ultimately claim more resources at the negotiation table than negotiators with low or no goals and limits (Zetik and Stuhlmacher, 2002) ^[4]. However, high goals can also complicate negotiations, as they are known to induce more risky behavior (Larrick et al., 2009) and more frequently cause impasses (Bazerman et al., 1985).

Best alternatives in the event of an impasse in a given negotiation determine the minimally acceptable outcomes, beyond which parties are not willing to settle (Brett et al., 1996). Various scholars emphasize the immense importance of BATNAs in negotiations (e.g. Jang et al., 2018; Sebenius, 2017; Thompson, 2015), which are also referred to as outside options in the economic literature (e.g., Jäger et al., 2024; Sebenius, 2017). Fisher and Ury, the inventors of the BATNA concept, even describe the best alternative outcome as the ultimate “standard against which any proposed agreement should be measured” (1981, p. 104). In the pertinent literature, BATNAs are therefore regarded as a highly important RP in negotiations (Blount et al., 1996; Jang et al., 2018). Empirical research shows that negotiators with an attractive BATNA hold higher expectations (Pinkley et al., 1994), are more likely to move first (e.g. make the first offer; Magee et al., 2007), are more willing to walk away from the negotiation (Brett et al., 1996) and achieve better outcomes than negotiators with less attractive BATNAs (e.g. Gunia et al., 2013).

By using goals, limits or BATNAs as RPs, negotiators regulate their behavior during the negotiation process. These RPs direct negotiators’ attention toward relevant activities (e.g. focus on claiming value for one’s own party, finding more arguments that support the reference outcome), they increase the effort and persistence devoted to the negotiation and positively affect the use of strategies to achieve or exceed the reference outcome (Huber and Neale, 1987; Larrick et al., 2009, see also Locke and Latham, 2002). Consequently, RPs affect negotiator behavior, for example negotiators’ calibration of proposals and concessions (Blount et al., 1996).

Research on negotiators’ affective evaluations of negotiation outcomes relates to how negotiators use RPs to determine whether they are satisfied with their intermediate or final destination on their way through the negotiation fog (Wheeler, 2013). This line of research is concerned with how negotiators translate objective outcomes into subjective assessments, for instance, into perceived gains (i.e. outcomes better than the RP) or losses (i.e. outcomes worse than the RP; Larrick et al., 2009; Tversky and Kahneman, 1991). Because of loss aversion – people experience losses as more painful compared to the perceived attractiveness of equivalent gains (Larrick et al., 2009) – negotiators are more willing to risk potential agreements under loss frames by making higher demands and offering smaller concessions than negotiators under gain frames (Neale and Bazerman, 1992). In experimental research, the perception of gains or losses is typically a product of providing minimum or maximum outcomes as a salient RP (e.g. De Dreu and McCusker, 1997; Trötschel and Gollwitzer, 2007). The findings of these framing studies are clearly relevant in explaining negotiator behavior. However, maximum and minimum outcomes as RPs in the laboratory are rather artificial RPs that are probably not common in practice. Therefore, we will carefully study whether and how frequently they are mentioned by the key informants in our study.

Beyond separating potential outcomes into gains and losses, RPs are also known to affect negotiators’ subsequent valuation of these outcomes (see Tversky and Kahneman, 1992). They determine, for instance, the perception of success or failure and the experience of (dis-) satisfaction (Oliver et al., 1994; Thompson, 1995). These subjective feelings differ from the economic outcomes of a negotiation (Curhan et al., 2006; Thompson, 1990), and they matter because they affect post-agreement behavior (e.g. post-settlement compliance, Gelfand et al., 2006; willingness to negotiate again, Oliver et al., 1994) and the economic outcomes in subsequent negotiations (Curhan et al., 2010; Wiltermuth et al., 2023).

In summary, the cognitive function focuses on how negotiators understand the negotiation situation, while the motivational function relates to negotiators’ aspirations, and the affective function refers to the translation of (potential) negotiation outcomes into subjective experiences. The three functions are, however, related, and considering all three functions collectively provides a comprehensive understanding of RPs and their multifaceted impact on negotiation behavior and outcomes. Thus, by integrating these functions, our study aims to offer a holistic view of RPs, addressing the interconnected nature of cognition, motivation and affect in negotiation processes.

Despite the extensive research on how RPs impact negotiation behavior and outcomes, we argue that RPs are still an understudied issue on a conceptual level. Supporting this view, Koszegi and Rabin (2006) suggested that empirical work and theory development require a deeper understanding of what RPs are. It is currently open, for instance, which RP people choose in a given decision context (Barberis, 2013; Gupta and Livne, 1988). Thus, there are various open questions about how people use RPs in negotiations. In light of these questions, the present research aims to investigate whether the simplified and rather narrow perspective on RPs in negotiation research, which focuses on goals, limits and BATNAs, sufficiently reflects the RPs that negotiators choose in practice, or whether it neglects further potentially relevant and distinct types of RPs (e.g. status quo or previous contracts) ^[5]. This knowledge is important because other types of RPs not only alter negotiators’ perceptions, but they may also result in different decisions and behaviors (e.g. negotiators may act much more careful when using a concrete previous contract vs a rather arbitrary goal). Thus, we seek to examine if negotiators in fact use any further RPs and therefore pose the following research question (RQ):

Second, current pertinent theories and models do not explain the origin of RPs (Gimpel, 2007; Hyndman, 2011). In line with this notion, Barberis criticized “relatively little guidance on how the reference point is determined” (2013, p. 178, see also Baucells et al., 2011). Thus, it remains unclear, for instance, whether negotiators in organizational settings choose their RP freely or whether they focus on RPs imposed on them by others. Thus, we pose the following RQ:

Adding a temporal dimension to the conceptual perspective on RPs allows us to identify a further important question referring to how negotiators use RPs in the course of negotiations. In empirical (negotiation) research, RPs are usually treated as given and static or unchanging phenomena (Holmes et al., 2011; Prietula and Weingart, 2011) ^[6]. This perspective neglects initial evidence that RPs can change in the course of the negotiation (Kristensen and Gärling, 1997a; Thompson, 1990). Thompson (2015) even suggests that negotiators should alter their RP in the course of a negotiation. Her recommendation highlights the economic benefits of focusing on one’s own goal during the negotiation, while also emphasizing that taking the best alternative as RP after the negotiation may lead to a higher satisfaction with the agreement (see also Galinsky et al., 2002). Exploring whether and when negotiators change their RP in the course of the negotiation seems to be an appropriate starting point to further study the reasons why they do so or why they stick to a once-chosen RP. Thus, we pose the following RQ:

The previous considerations imply the question of when or under which conditions negotiators use different types of RPs in negotiations. Thompson’s (2015) recommendation described above indicates that the effective choice of a RP may depend on specific features of a negotiation, for instance, the particular temporal phase of the negotiation process. This is a rare exception in negotiation research, because it provides advice, which depends on the respective phase of the negotiation process, as suggested by Jang et al. (2018) in their first step toward developing a general theory of negotiation. Although other theoretical approaches on RPs often also refer to a certain phase of the negotiation (e.g. the setting of a goal prior to the negotiation; Malhotra and Bazerman, 2007), they are typically not integrated into a comprehensive RP approach that considers all (or even more than one) negotiation phases (i.e. from preparation to implementation; Jang et al., 2018). Importantly, the relevance of RPs may not only depend on temporal but also contextual factors, for instance, the particular negotiation context (i.e. business vs political negotiations), the negotiation content (i.e. the negotiation topics or issues), the negotiators’ roles (i.e. management vs union members or buyers vs sellers) or the relationship with the negotiation partner (i.e. one-shot vs continued negotiation in a long-standing relationship; see Borbély et al., 2017; Eliashberg et al., 1995). Negotiation theories are mostly silent about the contextual factors that influence negotiators’ choices of RPs. We therefore pose the following RQ:

The final RQ of our study focuses on why negotiators use RPs. In particular, we investigate whether the psychological functions associated with RPs in negotiations (i.e. cognitive orientation, motivational self-regulation and affective evaluation, see above) are validated and/or supplemented by the reports of the key informants in our sample:

The main objective of this study is to make theoretical contributions to the understanding of RPs in negotiations. We seek to narrow the theoretical gap that lies in the incomplete understanding of the full spectrum of RPs, their origins and their multifaceted roles throughout the entire negotiation process. First, we develop a comprehensive typology of RPs in negotiations beyond the traditional focus on goals, limits and BATNAs. With this typology, we aim to increase the conceptual clarity of RPs and provide a more robust framework for explaining and predicting negotiation behavior and outcomes. Second, by examining how RPs are actually used, we challenge the prevailing treatment of RPs as rather static phenomenon in psychological negotiation research and demonstrate how they can change throughout the negotiation process. Third, we aim to extend the understanding of the psychological functions of RPs in negotiations beyond their familiar roles in cognitive orientation, motivational self-regulation and affective evaluation. Thereby, we seek to enrich existing negotiation theories by integrating a more comprehensive view of RPs’ roles. Finally, we aim to contribute to the further specification of a general theory of negotiation as proposed by Jang et al. (2018). By emphasizing the central role of different types of RPs throughout the entire negotiation process – including the preparation, at-the-table negotiation and post-negotiation phases – we seek to offer a more holistic framework for understanding negotiation dynamics and deepen the academic discourse on negotiation.

Our research objectives suggest a qualitative design to identify, characterize and analyze relevant RPs in organizational settings (see Holmes et al., 2011). Qualitative studies are an appropriate way to further develop (negotiation) theory (see Bowles et al., 2019; Geiger, 2017; Mann et al., 2024; Reif and Brodbeck, 2021), in our case, to better understand RPs conceptually by exploring practitioners’ theories in use (Zeithaml et al., 2020, see also Diaz‐Moriana et al., 2024). We chose an inductive, open-ended approach based on structured and rule-governed procedures to ensure both a naturalistic representation of the material without biases from researchers’ preconceptions and to maintain maximum rigor (Lincoln and Guba, 1986). Rather than sticking to one methodology like the Gioia methodology in a standardized manner (Gioia et al., 2013), we decided to tailor the analysis to our research project to benefit from the combination of analytical approaches (“methodological bricolage”; Pratt et al., 2022, p. 7).

Our sample included 58 experienced practitioners who negotiated in various contexts and roles. Following the terminology of Eliashberg et al. (1995), we differentiate three negotiation contexts in our study:

1. business negotiations (28 interviews);

2. labor–union negotiations (18); and

3. political negotiations (12).

We chose these three contexts because they represent the most common types of professional (as opposed to private) negotiations, in which parties try to reach agreement without external help of mediators or arbitrators. Business negotiations refer to negotiations between different companies, regularly in the form of buyer–seller negotiations. The informants came from different major industries (e.g. automotive, chemicals, engineering, food and retail), worked in different functions (e.g. management, sales, purchasing) and at different hierarchical levels within their organization (e.g. director, manager). Labor–union negotiations describe negotiations between collective bargaining parties of varying scope (i.e. collective agreements with national, regional or company-specific scope). Informants were lead negotiators either from employer associations or from unions in various industries. Political negotiations refer to negotiations by representatives of a political institution. Our informants were either members of a parliament or in an executive position in a city’s administration (e.g. mayor, office manager). They belonged to various political parties.

The 58 key informants were employed in 50 different German organizations or public institutions. Their mean job experience was 25.95 years (SD = 9.60). The informants’ average age was 50.05 (SD = 8.65). Eighty-three percent were male, which largely corresponds to the proportions of the relevant subpopulations in Germany. To qualify as a knowledgeable key informant, each practitioner had to possess sufficient professional experience (i.e. at least 12 months) and negotiation experience (i.e. minimum of 30 conducted negotiations). These requirements were checked via phone prior to the interview and met by all participants.

We recruited the key informants following the guidelines for “theoretical sampling” (Glaser and Strauss, 1967, p. 45). We proceeded with an iterative process of data collection and data analysis (e.g. Gioia et al., 2013). The ongoing analysis of already conducted interviews helped us to determine focus areas for later data collection (Corley and Gioia, 2004) and enabled us to identify where to search for “disconfirming cases” to receive different perspectives on the phenomenon we investigated (Solarino and Aguinis, 2021, p. 665). For instance, in political negotiations, we first conducted interviews with members of national and regional parliaments. Based on the information received, we then decided to broaden our political sample to negotiators from city administrations, as their disputes seemed to differ significantly from those in the legislature (e.g. agreements on real estate transactions versus draft legislation). Further information about the recruitment procedure and the sample is available in the supplementary material (see appendices S1, S2 and S3).

We conducted in-depth interviews with all 58 informants between 2020 and 2021. About 84.48% of the interviews took place either in-person (ten interviews) or remote via video conferencing calls such as Zoom or Microsoft Teams (39). Remote interviews helped us to increase participation, as many key informants were reluctant to meet in person during the COVID-19 pandemic. Nine informants preferred to be interviewed by phone. All interviews were audio-recorded and lasted on average 75.60 min (SD = 21.44). The recordings comprised in total 4,035 min (i.e. more than 67 h) and were transcribed verbatim.

In the two main parts of the interviews, we asked the key informants to explain their standard approach to negotiation and to report particular features of very successful or unsuccessful negotiation cases they had previously conducted [cf. the critical incident technique (CIT); Flanagan, 1954]. In general, the CIT is aimed at systematically collecting behaviors that differentiate between good and poor task performance. In our study, this method ensured our informants’ reports were both highly relevant and concrete. It further facilitated the researchers’ comprehension of what was being said and made it easier to ask follow-up questions when necessary (Diaz‐Moriana et al., 2024). Because we were interested in understanding RPs preferably comprehensively, we also asked the informants what they typically do at different stages of the negotiation (i.e. preparation, negotiation and implementation).

A semi-structured interview guideline was applied (see appendix S11 of the supplementary material). This format allowed us to let informants speak freely and capture the phenomenon under investigation through an unsupported, bottom-up approach (Solarino and Aguinis, 2021). We made sure that the guideline was focused on the RQs and did not contain leading questions. We revised it as the research progressed (Gioia et al., 2013).

Following the Gioia methodology (Gioia et al., 2013), our data analysis adopted an inductive approach in the initial stages, which minimizes researchers’ confirmation biases that can result from knowing the literature too early in great detail and merely focusing on testing prior hypotheses (see also Van Maanen, 1979). In the beginning, we subjected the transcripts to open coding (Corbin and Strauss, 1990), using the informants’ own terms to describe relevant RPs. The number of informant-based codes quickly became unmanageable after a few interviews and therefore required the formation of homogeneous (first-order) concepts based on the discovery of similarities and differences between codes (Gioia et al., 2013). In doing so, we sought to adopt or integrate the key terms of our informants into the concept labels whenever possible (for similar approaches, see Corley and Gioia, 2004; Nag et al., 2007). We proceeded in this manner until no further distinct concepts emerged from the material and theoretical saturation was reached (Glaser and Strauss, 1967; Reif and Brodbeck, 2021). Table S6 in the online supplementary material contains representative quotes for each concept and provides a more detailed perspective on the categorization process.

As the analysis progressed, our approach transitioned to an abductive form of research by consulting the pertinent literature extensively, which helped to identify a deeper structure in the material (Gioia et al., 2013). In this iterative second-order analysis, the comparison and abstraction of concepts led to (second-order) themes representing theoretically separable clusters of RPs (Nag et al., 2007). In a final step of the categorization, we grouped the themes into four aggregate dimensions, differentiated by the origin of the RPs.

This categorization was complemented by frequency analyses counting the occurrence of different RPs across all interviews and identifying variables that made RPs relevant in negotiations (Seale, 1999). These variables stemmed from general information about the key informants and memos that were written during the coding process (Corbin and Strauss, 1990). We extensively used pivot tables of a spreadsheet software to identify patterns in the data set (Schreier, 2012). This final step of the analysis resulted in the development of a framework indicating relevant RPs depending on the context and phase.

We took various measures to ensure the trustworthiness of our findings (e.g. Corley and Gioia, 2004; Lincoln and Guba, 1986). First, all interviews were audio-recorded and transcribed verbatim. Second, the data analysis was carried out on the basis of a qualitative research software (f4analyse; see O’Kane et al., 2021). Third, we used a coding handbook to guide the coding process (see Table S5 of the supplementary material). We constantly developed it further to ensure a high level of coding accuracy. Fourth, part of the material was coded by two authors in parallel. Discussing the differences between the two coders until consensus was reached helped us to sharpen both the definitions of concepts and the coding rules. Fifth, the entire material went through multiple coding runs. Sixth and finally, following Nag et al. (2007), we engaged two researchers unfamiliar with the study to conduct a two-step review of our categorization scheme. In a first step, we asked the “outsiders” to group all concepts into themes, to label each theme and explain their decisions. This first step aimed at receiving feedback on how the emerged concepts could alternatively be grouped as themes. The qualitative feedback led to modifications of the categorization scheme. In the second step, the level of agreement was assessed between insiders and outsiders with regard to the categorization (a modified Q-sort approach; Stephenson, 1953). Intercoder agreement was good (Fleiss’ Kappa = 0.69; Fleiss, 1971; Landis and Koch, 1977; for qualitative studies using similar procedures, see Bendersky and McGinn, 2010; Gioia et al., 1994; Nag et al., 2007). This approach bolstered our confidence in our findings, although interpretative research does not necessarily require intercoder assessments (e.g. Gioia et al., 2013).

The findings in this section are based on 609 items (i.e. specifications of relevant RPs) our key informants reported within 58 typical approaches to negotiation and 61 critical negotiation incidents (Flanagan, 1954). We translated informant quotes from German into English.

In the following, we will first address the question of how negotiators use RPs in negotiations by presenting a typology of RPs in negotiations. Second, answering the “when”-question, we will introduce a framework illustrating important contextual and temporal factors that affect the situational relevance of RPs. Third and finally, we will turn to the question of why negotiators use RPs or the psychological functions that RPs serve in negotiations.

Typology of RPs. In our effort to answer RQ1 (“What are relevant types of RPs in negotiations?”), we found a total of 14 different types of RPs (i.e. themes) with 33 relevant subtypes (i.e. concepts). To examine how frequently these different types of RPs occurred in our sample, we conducted frequency analyses. The results of the frequency analyses are presented later, because we believe that starting with the presentation of the data structure makes it easier for readers to understand our findings and, to present this data structure, we first address RQ2.

With respect to RQ2 (“Where do negotiators’ RPs originate from?”), our analysis revealed four different origins of RPs in negotiations. We found that RPs stem from the focal negotiators themselves, the negotiators’ organization, the parties’ collaboration and the broader environment (e.g. the market, legal order). Figure 1 illustrates the emergent data structure, which provides an overview of the different types of RPs according to their origin. In the following, the different RPs are explained in detail using this structure.

RPs originating from the focal negotiator. The first group of RPs originated from the focal negotiator or negotiation team. We found four types of RPs in this category: first, goals described self-set objectives, that is, what negotiators wanted to achieve in a negotiation. The informants’ goals differed in their level of aspiration and ranged from realistic goals to maximal or optimal goals. The following quote is representative for an individual goal as the RP:

Second, limits defined the minimum outcomes negotiators needed to achieve. If these minimum outcomes could not be accomplished, limits became walk-away points that supported negotiators’ decisions whether to continue or abort the negotiation. One informant (BN-S06) described limits as:

The negotiators’ expectations emerged as a third type. They referred to negotiation outcomes, but also to the negotiation process. An example for the latter was reported by a key account manager who successfully acquired a new customer (BN-S04): “In the first moment, the first offer was the most difficult, at least not to get beyond this piss-off point in terms of price, but to stay in the race.” The informant emphasized the importance of a certain price point in a very early stage of the negotiation, beyond which his counterpart would have discontinued their talks. Therefore, his price offer needed to remain below this RP (i.e. the expected limit).

Fourth and finally, in representative negotiations (i.e. negotiating on behalf of others), RPs also derived from negotiators’ anticipations about the demands of their constituents. The following quote of a retail manager (BN-B10) is representative for this type of RP: “If I end up below this average margin, based on the purchase price, our private merchants do not go along with it.” Further verbatim evidence for these four and all other types of RPs is provided in Table S4 in the supplementary material.

RPs originating from the organization. Internal financial indicators were the first type that originates from negotiators’ own organizations. Financial indicators stemmed from the internal accounting and took different forms such as budget values, break-even points and average profit margins. These RPs represented what organizations can afford or what they defined as required outcomes for a particular negotiation. The following quotes are representative examples of two financial indicators (i.e. a profitability measure and a budget value):

Second, the informants reported to focus on what others wanted them to achieve in negotiations (i.e. goals assigned by the supervisor). The following quote from a purchasing manager (BN-B32) is representative for this RP:

A third RP was related to the negotiator’s mandate or what focal negotiators were allowed to decide on their own. A sales manager (BN-S08) referred to his mandate by stating:

Fourth and finally, the organizational status quo described another relevant RP of negotiators’ organizations. A transaction between two B2B companies provides an illustrative example for this type of RP. Our informant, the head of procurement (BN-B28), intended to no longer purchase raw materials from the company’s sole supplier. In his negotiation with an additional, new supplier, he mentioned that his “reference point was the purchase price of the existing supplier.” This RP referred to the status quo of his organization.

RPs originating from the parties’ collaboration. A third origin of RPs was in the parties’ interaction at the negotiation table and their (prior) general collaboration. Both, previous proposals made by the focal negotiator or by the other party, were a first type of RP. A quote from a mayor (PN-P46) provides an illustrative example: “This negotiation was successful because our counterparty’s offer was lower than the final outcome.” A few informants identified a relevant RP in the middle of the two parties’ initial offers (i.e. compromise point).

A second RP derived from previous agreements between the parties, be they no longer valid, expiring, binding or temporary agreements. These points mainly referred to previous conditions once agreed between the negotiation partners in the past and to currently valid conditions that represented the (bilateral) status quo of the parties’ collaboration. A sales manager from the food industry (BN-S05) outlined an unsuccessful negotiation incident by using the previous agreement as his RP: “We were not satisfied with the outcome because we had to accept losses. […] The price was okay, but we really lost significant volume.” Another quote by a negotiator from a city administration (PN-P50) illustrates that the binding contract serves as an RP during the successful implementation of an important agreement:

Third, the key informants considered the outcomes resulting from a non-agreement in the current negotiation as a relevant RP. This RP refers to the literature on BATNA (Fisher and Ury, 1981) and inside options (e.g. Muthoo, 1999). Here, the RP is defined by the outcomes that result from saying no to a proposed deal (i.e. declaring an impasse) without an alternative agreement with another partner being available (see Lax and Sebenius, 2006). The quote of a national sales director (BN-S11) provides a representative example:

RPs originating from the environment. The fourth and final origin of RPs was the organization’s environment, for example particular markets, the economy and the legal system. The negotiator’s best alternative agreement represented a first environmental RP. A quote from our key informant, a head of procurement (BN-B32) illustrates this RP:

As this quote demonstrates, this RP results from outside options (e.g. Sebenius, 2017), agreements with other negotiation partners, for instance, closing a deal with another supplier in the market. Best alternative agreements can be seen as the prototype of the BATNA concept introduced by Fisher and Ury (1981). They are conceptually distinct from non-agreement outcomes in that they differ in the availability of a negotiated agreement with another party, although both refer to the BATNA concept.

External benchmarks as a second environmental RP were typically invoked to answer the question of fair or appropriate outcomes. Our key informants mentioned market prices, economic indicators such as the inflation rate, an expertise or the result of a price check as relevant external benchmarks. The following quote of a procurement manager illustrates this RP:

The expected outcomes of pursuing legal action were also included in this category. This RP can also be seen as a sub-variant of the BATNA concept representing the outcome from going to court as an alternative to negotiating with the other party (Sebenius, 2017). One of our informants, who negotiated with the work council for a public transport company, reported that the chances of legally enforcing video surveillance were zero (LN-E35). This external benchmark not only served him as an important RP in his negotiations with the members of the work council, but he also used it to manage the expectations of his constituents.

Precedents formed the third type of RPs stemming from the environment. In contrast to agreements between the negotiation parties, a precedent referred to an agreement with another negotiation partner reached by the focal negotiator or another negotiator of the organization, and to a known agreement between the counterparty and a third party or between third parties not involved in the current negotiation. Thus, precedents allow for social comparisons with other negotiators in similar situations (see Novemsky and Schweitzer, 2004). A management representative (LN-E24) provides a representative quote for this RP in labor–union negotiations:

The distinctiveness of RPs and an alternative approach to classifying RPs in negotiations. During the coding, we noticed that, in some cases, more than one RP was reported within the same text segment of a transcript (i.e. co-occurrence; Namey et al., 2008). We therefore checked our data for co-occurrences and found no significant patterns in the data except for goals, limits and the corresponding RP concepts with an organizational origin (i.e. assigned goals and mandates). These concepts were involved in 79.07% of all co-occurrences. This finding suggests that the RPs in our sample are distinct from each other; however, the boundaries between negotiators’ goals and limits and organizational RPs (e.g. budget values) are somewhat blurred. Based on our analysis, we assume that this is because negotiators rely strongly on RPs from their organization to define their goals and limits, presumably driven by their desire to meet or exceed organizational expectations.

Nevertheless, the aforementioned types of RPs are genuinely distinct from one another, which can be illustrated by including a temporal dimension into the analysis in addition to the (social) origin (see Yockey and Kruml, 2009). This alternative approach allows us to position the various types of RPs in a two-dimensional space. Our analysis results in a graphical representation of six distinct main categories of RPs (see Figure 2). The first category includes (individual) goals and limits, assigned goals and mandates, expectations as well as anticipated demands of negotiators’ constituents. In line with previous research, we label this category “level of aspiration” (see Oliver et al., 1994; Walton and McKersie, 1965), because these RPs typically refer to future states negotiators (or others) want (them) to achieve or avoid. Because they are prospective (see Heath et al., 1999), negotiators experience the achievability of the reference outcomes defined by these RPs as rather uncertain before and during the negotiation. The RPs in this category are set somewhat arbitrarily either by the negotiator themselves, the negotiation team or the organization. Conceptually, these RPs are perhaps best described by basic questions (for a similar approach of contrasting concepts, see Belitski et al., 2023) such as “What outcome do I hope to achieve?” (goals; Walton and McKersie, 1965), “What outcome would be my bare minimum?” (limits; Raiffa, 1982), or “What outcomes do others expect from me?”.

The second category “projected outcome” consists of current agreements as RPs, which are characterized as prospective and uncertain since they estimate the economic value of a concluded agreement after its implementation. They are typically substantiated by (financial) forecasts and stem from the parties’ collaboration. This RP relates to the question of “What outcomes will be realized by implementing the agreement?” Projected outcomes as RP relate to previous research in the legal sciences (contract theory; Hart and Moore, 2008).

“BATNA” (Fisher and Ury, 1981) as the third category consists of two different concepts: based on our data, BATNAs are best alternative outcomes resulting from possible agreements with other negotiation partners or through other conflict resolution approaches (i.e. best outside options), as well as reference outcomes without alternative agreements at hand (i.e. best inside options). They originate from the environment and the parties’ collaboration. As BATNAs refer to the present or the future, they can involve a degree of uncertainty (BATNA risk; Bottom, 1998). BATNAs should not be considered as arbitrarily chosen, as their identification follows a relatively rigid rule: focus on the alternative course of action that provides the greatest value if no agreement in the current negotiation can be reached (e.g. the best alternative agreement with another party; see Sebenius, 2017). BATNAs refer to questions such as “What would be our outcome if we reach an agreement with someone else?” or “What will be the consequences if we do not reach agreement?”

The fourth category “status quo” comprises the current organizational state and valid agreements (e.g. existing contracts) with the focal negotiation partner. Their reference outcomes relate to current states, which is why, they offer a high degree of certainty and are given or fixed rather than set arbitrarily by the negotiator. The status quo provides comparison standards from the negotiator’s organization and the parties’ interaction. These RPs relate to questions like “At what current price do we purchase?” or “What does the current agreement with the negotiation partner look like?”

The fifth category includes RPs that define a certain “level of efficiency” based on standards from the organization, the collaboration with the negotiation partner or the environment. These RPs are retrospective or contemporary, certain and substantiated. They can be described by questions such as “What is the available budget we were given by the organization?,” “Where is the point at which we suffer no losses?,” “What did we agree to in the past?” or “What is a fair outcome given the current market conditions?”

(Previous) Proposals and precedents can be classified into the sixth and final category of “social comparison.” These RPs are retrospective, rather certain, und substantiated. They originate from the parties’ collaboration and the environment depending on who negotiators are comparing themselves with. While proposals enable comparisons with the focal negotiation partner (internal comparison), precedents can relate to standards from negotiators’ own previous agreements with other partners or agreements of other negotiators in similar contexts and roles (external comparison; Novemsky and Schweitzer, 2004). Questions like “What did I agree to with other partners?” or “What did others agree to?”

As illustrated, the various types of RPs that emerged in our study are conceptually distinct according to their different characteristics. And importantly, just because goals and limits can be operationalized by other RPs (but not vice versa) does not imply that RPs can be conceptually reduced solely to these two types of RPs (for further notions supporting this perspective; see Heath et al., 1999; Bazerman and Neale, 1993; Blount et al., 1996; Loewenstein et al., 1989).

Frequency of RPs in negotiations. To examine how often different types of RPs occurred in our data (e.g. Namey et al., 2008; Schreier, 2012), we determined the number of informants who mentioned a particular RP on the level of themes. In qualitative research, frequency enables readers to “gain a sense of how representative and widespread particular instances are” (Seale, 1999, p. 128) and can be seen as an indication of salience (Namey et al., 2008, see Table 1).

Returning to RQ1 (“What are relevant types of RPs in negotiations?”), our analyses revealed that (individual) goals were the most frequent RP in our study (42 out of 58 key informants), followed by limits (34). Given the strong presence of goals and limits in negotiation literature, research and training, this result is not surprising. However, our findings also indicate that these RPs might not be as omnipresent in practice as is currently reflected in extant research. In fact, 26 informants reported that they focused on both goals and limits, implying that more than half of the informants only relied on a goal or a limit (24 informants) or neither of them (eight informants) over the course of a negotiation. The third most frequent RP was defined by the parties’ previous agreement or current contract (33 informants). Together with the current state within the organization (i.e. nine informants), this result not only underpins the importance of the status quo in decision-making, but further demonstrates its high relevance as a previously rather neglected RP in negotiations.

We also identified RPs that might be less relevant than assumed in the pertinent literature. First, only 16 informants reported that they focus on their best alternative agreement in negotiations ^[8]. Prior to this study, we had not expected this result, as BATNAs can be considered a cornerstone in the negotiation literature since its introduction by Fisher and Ury (1981). Second, proposals and demands articulated in the focal negotiation were no prominent RPs in our study. Only 15 informants mentioned to focus on reference outcomes defined by previous proposals. This finding, however, illustrates that previous proposals can be adopted as RP (e.g. an ultimatum) and affect how the potential outcomes associated with counterproposals are perceived (see Kahneman, 1992; Kristensen and Gärling, 1997b). Third, market values (e.g. notations at stock markets) were reported as RPs by ten informants. This finding may point to contextual factors that determine the relevance of market values in real-world negotiations. We will return to that point in the next section. Fourth and finally, we found little evidence in our data that the maximum or minimum possible outcomes are used as RPs in real-world-negotiation.

Our analysis of the origin of RPs (RQ2) also provided first answers to how free negotiators are in their choice of their RPs. The number of informants in our sample who used RPs that were not freely chosen by themselves totaled 38 (for further details see appendix S7 of the supplementary material). Thus, RPs imposed by the organization or supervisors represent a common phenomenon in professional negotiations.

To answer RQ3 (“Do negotiators change their RP in the course of a negotiation?”) and clarify whether negotiators use RPs statically or dynamically, we analyzed the RPs in the critical incidents across the different phases of the negotiation process and searched for potential RP changes from one phase to another. Although we started our data collection with a differentiation of three main negotiation phases (i.e. the preparation, the negotiation and the post-negotiation phase; Jang et al., 2018), working with our material led us to reconsider our approach with respect to the “middle phase” of the parties’ interaction. Based on a simultaneous review of the pertinent literature and our transcripts, we decided to analyze the material during the negotiation phase itself in a more detailed way (e.g. as a two-stage process in negotiations; Bacon and Blyton, 2007; Pruitt and Carnevale, 1993, see also the two-phase-model in group decision-making; Gersick, 1989). We, thus, synthesized two perspectives on the negotiation process (Jang et al., 2018; Morris and Keltner, 2000) and differentiated four phases in our analysis:

1. the preparation phase;

2. an early phase at the negotiation table consisting of parties’ opening moves and positioning;

3. a later phase at the negotiation table involving parties’ joint problem-solving and the endgame (i.e. where parties arrive [or do not arrive] at the agreement); and

4. the post-negotiation phase of evaluation and implementation.

With this approach, two patterns could be identified ^[9]. First, in 29.6% of the incidents, negotiators changed their RP during the negotiation, for instance, between the early phase and the later phase of the negotiation. And second, in 38.2% of the incidents, they chose different RPs in the post-negotiation phase than the RPs they had focused on during the negotiation. These results illustrate that our informants frequently changed their RP and suggest a dynamic usage of RPs over the negotiation process. An exemplary case is provided by a major (PN-P46) negotiating on behalf of a group of cities and municipalities (i.e. the constituents) with the ministry of a federal state about tax compensations due to the COVID-19 pandemic. In the preparation, he used an internal financial indicator to determine his limit: “We knew that we would lose [range] euro in taxes according to our estimate. […] Our limit was the compensation of the trade tax loss.” After aiming for his higher goal in the early phase of the negotiation, he used his limit in the later phase. With respect to the final agreement, he concluded: “(We) have received an average of 93% of what […] has been lost. So we have reached the ideal value except for seven per cent.” However, after the negotiation, when concerned about getting internal approval for the agreement, he used a different RP, the opening proposal of his counterpart: “This negotiation was successful because our counterparty’s offer was lower than the final outcome.” Further exemplary cases are presented in Table S8 of the supplementary material.

Although this dynamic perspective on RPs is reflected in different theoretical models of the economic bargaining literature (e.g. Hyndman, 2011; Shalev, 2002), the adjustment of RPs has been largely overlooked in empirical negotiation research (see Gimpel, 2007). This finding raises further questions such as “When do negotiators change their RPs?” or “Why do they change their RP?” In the next section, we will address the first question and then return to the second ^[10].

To answer RQ4 (“What temporal or contextual factors affect the relevance of RPs in negotiations?”), we integrated temporal and contextual factors into our analysis. We used a four-phase model of negotiation (see previous section) and focused on the negotiation context (e.g. a business negotiation) and the role of the negotiator (e.g. buyer) as contextual factors. We thereby arrived at a context- and process-oriented analysis that explains the relevance of different types of RPs in negotiations. In this step, frequencies provided a valuable source for identifying similarities and differences between subpopulations, negotiation contexts and phases (Namey et al., 2008; Schreier, 2012). Figure 3 maps the use of RPs in the different negotiation phases.

In the following, we will elaborate on the use of RPs for each of the three negotiation contexts (i.e. business, labor and political negotiation) and highlight the drivers of the relevance of certain RPs. For business negotiations, we found that negotiators frequently focused on goals and limits in the early phases of the negotiation process, independently of their roles. The relevance of goals and limits as RPs decreased as the negotiation progressed. In the third and fourth phase of the negotiation progress, goals were more often relevant for sellers than for buyers.

A further role-dependent pattern emerged from the data: sellers focused on previous or current agreements with their negotiation partners, presumably in the hope of improving their contracts. By contrast, the buyers’ perspective concentrated more on RPs describing what their organization required (e.g. internal financial indicators) or what markets offered (e.g. best alternative agreements). Notably, best alternatives were particularly relevant for buyers who procured standardized goods (i.e. commodities such as raw materials). They proved to be (much) less relevant in other buying roles and for sellers. Overall, the origin of business negotiators’ RPs differed depending on the role: While buyers focused on RPs originating from their organization or the environment, sellers primarily considered RPs stemming from the parties’ collaboration.

For labor–union negotiations, we found a pronounced focus on goals throughout the entire negotiation process. Best alternatives were neither relevant for unionists nor for employers. Moreover and again, we identified differences between roles. Unionists often also focused on their limits and precedents. Precedents referred to previous outcomes with other employer associations and to collective agreements in other industries. By contrast, employers were more focused on internal financial indicators linked to the economic feasibility of potential outcomes for the companies they represented. In addition, anticipated demands of their constituents and previous agreements were also relevant RPs for negotiators on the employers’ side.

In political negotiations, negotiators were considerably more likely to be focused on their goals compared to all other roles. Paralleling business negotiations, the relevance of goals, however, decreased as the negotiation progressed. In the later phase of the negotiation, politicians became increasingly attentive to their limits. After the negotiation, they focused more on the RPs originating from their interaction (e.g. proposals). Again, best alternatives were no relevant RPs in this context.

In the final analytic step, we went through the material again to find evidence to clarify why negotiators use RPs at all to answer RQ5 (“What are the main (psychological) functions of different types of RPs?”). Our analyses not only validated the three psychological functions of cognitive orientation, motivational self-regulation and affective evaluation associated with RPs in negotiation research, but they suggest one further function (e.g. using RPs to justify achieved outcomes within the own organization) and also provide a more differentiated picture of the psychological functions of RPs in negotiations.

Various statements of our informants showed that negotiators use RPs for reasons of justification. This function differs from the other three psychological functions of RPs in the pertinent literature, as it relates to inter-individual rather than intra-individual processes. By using RPs to justify negotiated outcomes within their own organization, our informants aimed at obtaining approval and at positively affecting their performance evaluations by others. A good example of justification is provided by a retail manager who had to achieve a minimum profit margin to meet the demands of his internal customers (BN-B10, see above). This psychological function even applies to managers who feel compelled to justify their negotiated outcomes to their own employees, as the following quote illustrates: “What are my own goals that I have defined for my buyers? […] Logically, as a supervisor, you cannot achieve worse results than you have set (for your employees)” (BN-B17).

Our analyses also suggest a refined understanding of why people use RPs in negotiations. First, our key informants used RPs to get a sense of what realistic, appropriate and ambitious outcomes might look like. This helped them to navigate the complex task of negotiation, for example by reviewing previous procurement prices or calculating an appropriate price based on its components (e.g. raw material plus production, packaging, transport and mark-up, BN-B47, see above). Our analysis of co-occurrences suggest that negotiators use RPs to facilitate the setting of adequate goals and limits – a psychological function we refer to as RP formation (see research on RP formation in the decision-making literature; e.g. Baucells et al., 2011). A representative quote for this function stems from a management representative who negotiated with a work council (LN-E40):

Second, various cases indicate that informants used RPs for monitoring the implementation of their agreements or contracts. RPs provided orientation even in the post-negotiation phase and helped negotiators to differentiate between compliant and non-compliant actions with respect to parties’ agreement. A city administrator (PN-P50) provided a good example for the use of RPs in the implementation phase (see above). A sales manager also used RPs in the post-negotiation phase:

Third, and finally, the informants used RPs also for self-evaluation. When asked to explain why they perceived a negotiation incident to be particularly successful or unsuccessful, they frequently evaluated their own performance based on comparisons against retrospective RPs such as previous agreements, precedents or the status quo and less often with prospective RPs like ambitious goals. By doing so, they directed their attention toward incremental achievements that allowed a (more) positive performance evaluation. The quote of a union negotiator (LN-U33) provides an example of how negotiators use RPs for self-evaluation: “If you check the (cross-union) database, this collective agreement is one of the highest this year.” A further example is the case of a major (PN-P46) negotiating about tax compensations (see above). Further empirical evidence for each of the psychological functions can be found in Table S9 of the supplementary material.

We conducted an inductive study to address the problem that there is currently a lack of conceptual clarity on RPs (Holmes et al., 2011). The rudimentary conceptual understanding of RPs in negotiation research limits the potential of extant theories to explain and predict experiences and behavior in negotiations and restricts the usefulness of existing reference-based recommendations. Our analyses suggest a shift in the current perspective on RPs in empirical research, as RPs turned out to be a remarkably multifaceted and dynamic construct serving different psychological functions. In the following, we will first outline the implications of our findings on how and when negotiators use RPs in negotiations for the advancement of understanding RPs as concepts. We will then discuss how this knowledge contributes to further specifying a general theory of negotiation, as proposed by Jang et al. (2018).

We contribute to further specifying RPs as concepts and thus extend the pertinent nomological net of RPs in research (see Cronbach and Meehl, 1955) by proposing additional types of RPs people use in negotiations and identifying conditions that affect RPs to become salient in a given negotiation.

First, we investigated how negotiators use RPs and identified relevant types of RPs, for instance previous agreements or the organizational status quo, which have received little attention in empirical negotiation research (cf. Caputo, 2013). Thus, in terms of its nomological net, the RP construct appears to be much closer related to further concepts beyond goals, limits and best alternatives than previously assumed. Based on our data, we suggest that previous agreements or organizational status quos serve as very relevant, additional types of RPs. Importantly, they have different characteristics as compared to the typically discussed types (goals, limits and best alternatives). They therefore broaden and specify the understanding of the RP construct.

For instance, concerning contracts as a form of status quo: a contract refers to an existing state, which parties may choose as their RP. They are based on retrospective views of what is currently in the books (e.g. a transaction price) and thus involve a low degree of uncertainty. By contrast, goals and limits are subject to high uncertainty, because they refer to prospective outcomes that have yet to be achieved or avoided. Consequently, the status quo will often represent a rather low bar or some kind of “natural” bottom line in a negotiation, as people typically approach negotiations with the expectation of benefiting from a change in the status quo through an agreement (see Boothby et al., 2023; Malhotra and Bazerman, 2007). We argue that, compared to ambitious goals and arbitrarily set limits, negotiators who focus on the status quo are more likely to:

• experience (potential) outcomes as gains rather than as losses;

• regard their own RP as a rigid benchmark; and

• persistently resist attempts by the other party to manipulate their own RP because outcomes that worsen the current state might violate common principles of fairness (mutual gains approach; Fisher and Ury, 1981) and can hardly be justified within negotiators’ organization.

Thus, our findings show that understanding other types of RPs beyond goals, limits and best alternatives is essential as these RP may not only evoke different perceptions of outcomes but also cause negotiators to make other decisions, behave differently during the negotiation process or experience another level of (dis-) satisfaction.

Second, we broaden the understanding about potential conditions under which certain RP concepts become relevant in negotiations (i.e. answering the “when”-question). Our analyses provide clues for specifying these conditions based on features of the negotiation context. Negotiation as an interaction between two or more negotiators virtually always involves individual RPs such as goals. Further, RPs like budget or market values are to be expected when negotiations unfold significant financial effects and are thus subject to interaction processes within negotiators’ organization (e.g. leadership and delegation, financial planning and monitoring). We assume that negotiators’ accountability plays a key role in these processes because constituents and supervisors are likely to use internal financial indicators or external benchmarks to evaluate (potential) outcomes and negotiators’ performance.

Moreover, the negotiation context is also characterized by whether the parties have reached agreements with each other before and whether they are cooperating continuously or occasionally. These contextual features increase the relevance of previous agreements as salient RPs in a current negotiation because ongoing collaboration not only establishes a bilateral status quo but also influences the status quo of both organizations. Finally, specific features of the parties’ environment influence the emergence of certain types of RPs such as best alternatives or precedents in negotiations. Our analyses suggest that best alternative agreements as focal RPs are bound to conditions such as the mere availability of alternative negotiation partners – which is rarely the case for instance in labor–union negotiations (see Sebenius, 2017) – or the relationship between suppliers’ capacity and buyers’ demand in a market (see Hebisch et al., 2022). This understanding of RPs refers to previous considerations in the literature linking the salience of RPs to specific decision-making contexts. For example, Koszegi and Rabin (2006) argue that people’s expectations as RP are better predictors than the status quo as RP when these two types of RPs diverge (rather than coincide), which is the case when people typically do not expect to maintain their status quo in a given situation (e.g. consumer decisions).

Our contributions to completing the understanding of the RP construct are based on empirical investigation, to the best of our knowledge the first of its kind in the pertinent literature. With the present study, we contribute to narrowing a critical gap not only relevant to various theoretical approaches in negotiation research, such as prospect theory (Tversky and Kahneman, 1992), goal-setting theory (Locke and Latham, 2002), the expectancy disconfirmation paradigm (Oliver et al., 1994) or the self-regulation model of negotiation (Brett et al., 1999), they are also important to further develop and specify a general theory of negotiation as proposed by Jang et al. (2018).

The present research exemplifies key elements of the negotiation process (i.e. multiple phases, multiple parties, multilevel structure) and also points to the usefulness of broadening the perspective on RPs to refine the form and features of a general behavioral theory of negotiation (Jang et al., 2018).

Our work illustrates the crucial role of RPs in explaining people’s perceptions, decisions, behavior and affective evaluations in negotiations. Although we generally follow the reasoning of Jang et al. (2018) in placing an RP (i.e. the BATNA) at the center of their theoretical model, our analyses suggest that many other RPs besides BATNAs are frequently relevant during the negotiation, including goals, representations of the status quo, external benchmarks and other approaches to defining negotiators’ lower bound (e.g. mandates, break-even points, budgets; see Mann et al., 2022). Because less than one out of three key informants in our sample reported to focus on their best alternative agreement, identifying a strong alternative in real-world negotiations seems to be quite a challenge or at least a relatively rare behavior. Conceptualizing the BATNA as the dominant RP may help to structurally map parallel negotiations with alternative partners. However, Jang et al. (2018) approach may not adequately capture the variety of RPs negotiators actually use and therefore runs the risk of falling short in understanding, explaining and predicting negotiation behavior and outcomes, especially when strong alternatives are absent and thus turn out to be irrelevant in a given negotiation. We contribute to identifying the RPs that a general theory of negotiation should address and further illustrate, for example, that maximum and minimum possible outcomes do not to tend to be among the relevant RPs in negotiations.

Our findings indicate that negotiators use RPs for justifying their outcomes to relevant stakeholders (e.g. managers) within their own organization. This psychological function of RPs relates to previous research on procurement decisions (Puto, 1987) and underscores the importance of modeling the negotiation process as a two-level game (Putnam, 1988). We contribute to further specifying a general theory of negotiation by demonstrating the relevance of RPs at both levels, the level of the negotiation process with the other party (i.e. level one) and the level of the internal alignment within the negotiator’s own organization (i.e. level two; Jang et al., 2018).

Extending the model of Jang et al. (2018), we consider RPs as negotiators’ permanent companions, not only affecting the parties’ interaction at the negotiation table, but also their perceptions and decisions before and after the negotiation. In the preparation, our key informants used RPs to gain orientation and define their goals and limits. As the success or failure of real-world negotiations often depends on successfully implementing what the parties agreed on (e.g. Mislin et al., 2011), negotiators are likely to use RPs also in the post-negotiation phase to monitor whether the implementation of the agreement delivers the (anticipated) value. We suggest that negative deviations from the current agreement (i.e. RP) represent an important reason for negotiators to initiate renegotiations with their partner (see Reif and Brodbeck, 2021). Further specifying Jang et al. (2018) approach, we presume that a general theory of negotiation may benefit from incorporating RP-based comparisons as a key determinant of negotiation behavior over the entire negotiation process, which also affect future negotiations with the same partner.

The findings of the present study offer practical implications for both negotiators and managers. On an individual level, our typology of RPs can be used by negotiators as a resource to identify appropriate or realistic outcomes for a given negotiation situation. It also provides alternative RP candidates negotiators may consider for calibrating their proposals or evaluating a specific offer from the other party. For instance, before making major concessions or applying a risky strategy (e.g. making an ultimatum), negotiators could explore whether this negotiation strategy would still make sense if they chose other RPs as comparison standards (Malhotra and Bazerman, 2007).

We assume that many practitioners are not aware of the different sub-variants of BATNAs that exist (see Sebenius, 2017). Thus, they might be unable to identify best alternatives in the absence of alternative agreements with other parties. By systematically extending the findings of our study, we differentiate five potential sources of BATNAs:

1. best alternative outcomes with others (e.g. customers can buy a particular product from another supplier);

2. outcomes obtained by collaborating with the other party without an agreement being reached in the current negotiation (e.g.; they no longer buy the specific product from the supplier, but keep purchasing other products);

3. outcomes resulting from internal actions (e.g. they could produce a good on their own instead of buying it from a supplier);

4. outcomes generated by other approaches of conflict resolution (e.g. they can take the supplier to court); or

5. outcomes that result from retaining the status quo (e.g. they can decide not to buy the product from anyone).

Practitioners may benefit from systematically screening these different types of alternatives to strengthen their position with an attractive BATNA before entering into a negotiation.

On an organizational level, our typology of RPs in combination with the psychological functions provide a preliminary framework for improving negotiation-related decision-making and more effectively exerting influence on others within their organization and beyond. To manage both organizational performance and risk, managers are presumably well-advised to support negotiators’ RP formation processes before and during important negotiations. If negotiators choose too aggressive RPs, organizations are more likely to suffer from riskier strategies and bargaining inefficiencies (e.g. unnecessary and hurtful impasses, White and Neale, 1994; damaging the relationship with the negotiation partner, Olekalns and Smith, 2018). Conversely, if RPs are too low, organizations tend to leave valuable resources on the table (e.g. Huber and Neale, 1987; Zetik and Stuhlmacher, 2002). Where necessary, this managerial support could also prevent negotiators from setting only a goal or a limit instead of formulating both RPs as recommended in various guidebooks (e.g. Malhotra and Bazerman, 2007; Thompson, 2015). By doing so, negotiators are able to use both the performance-enhancing effect of goals and the protective impact of limits.

Our findings can also help managers influence the subjective feelings of their negotiators in the organization. Ambitious but realistic RPs such as high goals reinforce negotiators’ motivational striving for economic outcomes, but they can be detrimental to their satisfaction if negotiated outcomes fall short of reference outcomes (Galinsky et al., 2002). We argue that managers can use the present study’s insights by deliberately altering the RP away from ambitious goals to RPs that focus on the negotiators’ achievements and allow for a fair assessment of final outcomes. In business negotiations, for example, managers could alternatively compare negotiated outcomes to external RPs such as market prices, outcomes of others in similar roles, or previous agreements with the partner. This understanding enables managers to use RPs effectively to meet both their responsibility for performance and the people in the organization (Drucker, 2013).

Like any study, the current study has limitations, which provide opportunities for future research. Because our analyses relied on the key informants’ reflections on past experiences, the qualitative data may not fully and accurately represent negotiators’ typical approaches to negotiations and the critical incidents that were reported (e.g. due to memory distortions and loss). Problems with retrospective sensemaking or social desirability cannot be entirely avoided, although the data collection was based on an exploratory bottom-up approach (e.g. Bowles et al., 2019). Future research could use quantitative methods (e.g. surveys, experiments) to not only triangulate (i.e. cross-check) and generalize our findings (Lincoln and Guba, 1986) but also to further explore the role and impact of previously understudied RPs (e.g. the status quo, precedents) in negotiations, for example to clarify whether it is more effective to focus on one particular RP over another. Experimental research would also make it possible to uncover unconscious RP-based judgment processes that can hardly be captured using interviews or surveys as data collection method (Breuer et al., 2020). Building upon the present study’s insights, we suggest to consider the specific features of the negotiation context (e.g. intra-organizational hierarchies) to allow for the emergence of alternative RPs, for instance in internal negotiations with superiors, which are linked to external negotiations with negotiators’ counterparts.

Our analyses are based on interviews with experienced practitioners frequently involved in business, labor–union and political negotiations. We were careful to collect data from a highly diverse sample to promote the generalizability of the findings. However, as negotiation settings can differ significantly (Borbély et al., 2017), our findings might be limited to the negotiation contexts and cultural settings of our sample. Future research could also examine other negotiation contexts like mergers and acquisitions, job negotiations or international and peace negotiations (Jang et al., 2018) and different cultural settings in which negotiations take place. Cultural differences in the use of RPs could be investigated, for example, between negotiators from diverse international societies (e.g. in view of dignity, face vs honor cultures; Aslani et al., 2013) and various types of organizational cultures (Hartnell et al., 2011) to better understand how they affect behavioral tendencies in negotiation (e.g. information exchange; Adair et al., 2001).

RPs are known to have a substantial impact on human experiences, judgments and their resulting behaviors. Despite their ubiquitous use in various decision-making tasks, such as negotiation, empirical research aimed at understanding how, when and why negotiators use RPs in negotiations is scarce. The present study contributes to the advancement of this conceptual knowledge. We offer a comprehensive typology of RPs in negotiations showing that negotiators use a wider variety of RPs than previously assumed. In particular, our findings suggest that BATNAs do not serve as a salient RP in different negotiation contexts, although BATNAs are often presented as a generally applicable concept in the relevant literature. We provide valuable insights into why RPs should be considered to be more dynamic than assumed in previous negotiation research and illustrate that negotiators use RPs for reasons beyond merely motivating them to achieve certain outcomes. Specifically, we put justification as a previously neglected (psychological) function of RPs in negotiations on the (research) agenda. We hope that this research will pave the way for developing comprehensive theory on the effective use of RPs in negotiations.

The authors would like to thank Caroline Heydenbluth and Marco Schauer for their valuable support in reviewing and optimizing the categorization scheme. The authors would also like to thank Colleen Dollst for her support in compiling the reference list and translations.

Funding: The research project was supported by a research grant from the Hans Böckler Foundation (Project Number 2019–112–30) awarded to Roman Trötschel.

The supplementary material for this article can be found online: https://osf.io/8afec/?view_only=15ee54a0ed2d465bac80c7e8750df976