Introduction

Negotiation is a process by which a joint decision is made by two or more parties, who perceive that their interests conflict (Pruitt, 1981; Thompson & Hrebec, 1996). Driven by the increase in online transactions enabled by the internet (Moore et al., 1999; Katsh & Rifkin, 2001), a recent trend is for negotiations to occur over computer media, such as e-mail and instant messaging (Katsh & Rifkin, 2001; Dorado et al., 2002; Tyler, 2004). For example, in December 2007 over 6 million listings were added per day and over 2 thousand dollars worth of goods were traded every second on eBay (eBay, 2007). Many of these transactions take place between buyers and sellers with conflicting interests, and as a result there is a need for negotiations (Tyler, 2004). Online dispute resolution services, such as Square Trade, handle a large number of these negotiations in over 120 countries (Steiner, 2004). By allowing buyers and sellers to engage in 'direct negotiation' via computer media, this type of service can lower the cost of negotiation, increase its speed, make it more informal, and reduce the need for third parties (Katsh & Rifkin, 2001).

Unfortunately, using computer media for negotiation is not without its challenges. For a variety of reasons, this media can affect the negotiation process and thereby make it more difficult for negotiators to reach an agreement (Valley et al., 1998; Thompson & Nadler, 2002), which is the primary goal of negotiation (Fisher & Ury, 1991; Lax & Sebenius, 1991). With regard to the process, our focus in this paper is on flaming. Flaming is generally defined as the anti-normative hostile communication (e.g., the use of profanity) of emotions (Selfe & Meyer, 1991; Lea et al., 1992; Lee, 2005). This type of communication is important because (1) it can have disruptive effects on negotiation (Thompson et al., 1999), (2) its incidence is significantly higher when using computer-mediated channels (Sproull & Kiesler, 1986; Hiltz et al., 1989; Reinig et al., 1998), and (3) computer-mediated channels are increasingly being used for negotiations (McKersie & Fonstad, 1997; Moore et al., 1999).

The literature suggests that, though greater than in face-to-face communication, the incidence of flaming in computer-mediated communication is generally quite low (Lea et al., 1992). For example, few or no incidences were reported by Hiltz et al. (1989) and Lea & Spears (1991), while 3–4% incident rates were reported by Siegel et al. (1986) and Kiesler et al. (1985). However, these studies were based on group decision tasks (e.g., Hiltz et al., 1989; Lea & Spears, 1991). In contrast, the incidence of flaming during computer-mediated negotiation can be quite high. For example, in our pilot study flaming occurred in almost 40% of negotiating dyads. Additionally, in the study reported here, flaming occurred in more than 30% of negotiating dyads. This is in accord with Kayany (1998) who found flaming incidences of over 25% when conflict and tension existed among communicating parties (such as when groups discussed political issues).

Although there is general agreement that flaming can have disruptive effects on negotiation (Sproull & Kiesler, 1986; Reinig et al., 1998; Thompson et al., 1999), predictions concerning these effects are not consistent (Anderson & Thompson, 2004). For example, work by some researchers suggests that flaming can lead to conflicts, which significantly reduce the likelihood of negotiated agreements (Allred et al., 1997; O'Sullivan & Flanagin, 2003; Friedman et al., 2004). In contrast, other researchers' work suggests that flaming can increase the level of concession (van Kleef et al., 2004a), which is the most widely accepted determinant of negotiated agreement (Magneau & Pruitt, 1979; Pruitt, 1981; Pruitt & Carnevale, 1982).

These inconsistencies indicate the need for further investigation into the effects of flaming on negotiation (Postmes et al., 2001; O'Sullivan & Flanagin, 2003). We address this need by identifying different types of flaming that can account for their differential impacts on negotiation. In accord with suggestions by Druckman & Olekalns (2008), we distinguish between flames directed at the negotiation opponent (e.g., cursing the opponent) and those directed at the bargaining context (e.g., cursing bad luck). We also distinguish between the impacts of these types of flames on flame sender and flame recipient concession. We are thus able to offer novel predictions that relate different types of flaming to the negotiating behaviors of flame senders and recipients, and ultimately to the likelihood of negotiated agreements.

In contrast to much negotiation research in which participants interact with simulated opponents (Druckman & Olekalns, 2008), we employ a laboratory study that allows negotiators to interact with each other in a more natural context. This enables natural interactive dynamics, and makes our study more relevant to real-world negotiations (Druckman & Olekalns, 2008). We find that flaming has different effects on the negotiation behaviors of the flame sender and the flame recipient depending on whether the flame was directed at the negotiation opponent or at the negotiation context. For example, flames directed at the negotiation opponent slightly decreased the likelihood of reaching agreement, and when agreement was reached, resulted in outcomes significantly favoring the flame recipient rather than the flame sender. In contrast, flames directed at the context significantly increased the likelihood of agreement, although outcomes still favored the flame recipient over the flame sender. These results suggest that a flame sender is generally worse off than a flame recipient in negotiations. As described later, this is particularly important for computer-mediated negotiation, which promotes a sinister attribution bias (Thompson & Nadler, 2002) that can increase the likelihood of opponent-directed flaming.

Theory and research model

In this section, we first define five concepts that are the foundation of our research model. These concepts are: negotiation, concession, distributive negotiation, flaming, and computer mediation. We then develop our hypotheses that are based on these concepts and informed by prior theories.

Negotiation. Negotiation is a special case of group decision-making, with specific kinds of goals and relationships (Donohue et al., 1983). Negotiating parties perceive each other's goals to be incompatible with their own (Thompson & Hrebec, 1996), although they must cooperate to reach satisfactory outcomes (Morley & Stephenson, 1977). This results in a mixed-motive relationship 'in which the parties cooperate by competing for divergent ends' (Putnam & Roloff, 1992). For example, in negotiations over a fixed resource (distributive negotiations), each party competes to claim a larger portion of the resource for himself/herself. However, this competition is tempered by the realization that as one party claims more of the resource, the likelihood of the other party cooperating (i.e., agreeing to the resource distribution) decreases; and lack of agreement (or impasse) decreases the potential utility for both parties. Some authors distinguish between negotiation and bargaining, with negotiation being a special case of bargaining in which there is verbal communication and a goal to reach agreement (e.g., Morley & Stephenson, 1977). However, like Rangaswamy & Shell (1997), we will use the terms synonymously to mean negotiation aimed at reaching agreement.

Concession. Negotiation differs from the general notion of group decision-making in terms of concession. Concession is the act of changing one's offer in support of another party's interests such that there is a reduction of benefit to one's self in order to increase the likelihood of agreement (Jensen, 1963; Pruitt, 1981). For example, when negotiating the sale of a fixed resource, the seller offers a price that he/she is willing to accept as payment. The buyer may then make a counter offer lower than that of the seller. The seller may then make a counter offer that is lower than his/her first offer but higher than the buyer's offer. The difference between the seller's first and second offers represents a concession by the seller. In this way, concession is ritualized as an exchange of offers and counter offers (Morris, 1970; Zartman, 1977; Tutzauer, 1992), and is the most widely accepted determinant of negotiated agreement (Magneau & Pruitt, 1979; Pruitt, 1981; Pruitt & Carnevale, 1982).

Distributive negotiation. In general, negotiation may be concerned with a single issue, involving strategies for claiming resources (termed distributive negotiation), or with multiple issues, involving strategies for reaching win-win agreements (termed integrative negotiation) (Thompson & Nadler, 2002). Our focus is on a distributive negotiation, involving a single issue, for three reasons. First, this type of negotiation has received less attention in the extant literature (Kwon & Weingart, 2004), although there are many negotiations that are centered solely on claiming resources (Spangler, 2003). Second, even when a negotiation involves multiple issues, it tends to progress on an issue-by-issue basis (Morris et al., 2002), with a single issue (such as price) often dominating while other substantive issues are considered tangential (Kwon & Weingart, 2004). Finally, anger is frequently linked to distributive negotiation (Berkowitz, 1988; Pillutla & Murnighan, 1996; Thompson et al., 1999).

Flaming. Flaming is generally defined as the anti-normative hostile communication of emotions (Selfe & Meyer, 1991; Lea et al., 1992; Lee, 2005) that includes the use of profanity, insults, and other offensive or hurtful statements (Sproull & Kiesler, 1986; Kowk et al., 2002; Veazie, 2007). The anti-normative aspect of flaming is determined with respect to interactional norms present in interpersonal interactions. Such norms help individuals construct and interpret messages (Brown & Levinson, 1987; Martin, 1997; O'Sullivan & Flanagin, 2003). Communication that is contrary to the interactional norms of a specific population is considered anti-normative behavior for that population. Past research suggests that flaming may be directed at an individual's dyad opponent (e.g., calling the opponent a vulgar name) or at the negotiation context (e.g., cursing bad luck) (Reinig et al., 1998). We refer to the former as opponent-directed flaming and the latter as context-directed flaming.

Computer mediation. Compared to richer forms of communication, it is well established that there is an increased incidence of flaming in computer-mediated channels, which is thought to be due to perceptions of anonymity associated with reductions in the transfer of social cues (Sproull & Kiesler, 1986; Hiltz et al., 1989; Aiken & Waller, 2000; Kowk et al., 2002). Such reductions decrease individuals' self-awareness (Diener, 1980; Pinsonneault & Heppel, 1998) as well as their concern for social evaluation and fear of social sanctions or reprisals (Sproull & Kiesler, 1991; Pinsonneault & Heppel, 1998; Stuhlmacher et al., 2007). In turn, less self-awareness, less concern for social evaluation, and less fear of social sanctions promote a psychological state that weakens individuals' desires to conform to the positive expectations of others, thereby increasing anti-normative behavior such as flaming (Spears & Lea, 1992; Postmes & Spears, 1998). In addition to the evidence that negotiations are more hostile with computer mediation as compared to face-to-face (Stuhlmacher & Citera, 2005), there is also evidence that the relative synchronicity of computer-mediated channels may make a difference in flaming. For example, Pesendorfer & Koeszegi (2006) found more negative emotional communication among negotiators using chat as opposed to e-mail systems. They hypothesized that this was due to 'time pressure and the need for immediate reaction (being)...causes for spontaneous and unreflective emotional behavior' (p. 144).

There are circumstances when anonymity associated with computer mediation can decrease flaming. Such circumstances can occur when social identity and in-group status are salient, such as when anonymous communication takes place among members of a group who have a history of working together. Computer-mediated communication can then increase normative behavior because individuals focus their attention on the social context (and associated norms) rather than themselves (Lea et al., 1992; Moore et al., 1999; Postmes et al., 2001). However, we are interested in negotiations between parties who are anonymous and are strangers (i.e., unfamiliar with each other). This focus is important because social identity and in-group status become less salient, and therefore would not serve to mitigate the incidence of flaming. In addition, being strangers can exacerbate reductions in the transfer of social cues brought about by anonymity (Moral-Toranzo et al., 2007), thus increasing the potential for flaming. This focus on strangers is reasonable, since many parties transacting online do not know their opponents, at least for initial encounters and associated agreements, and are thus both anonymous and strangers. This is especially true in environments, such as eBay, when online resolutions take place as a result of disputes arising between buyers and sellers (Ba & Pavlou, 2002; Global Investment Technology, 2003; Tyler, 2004).

As described below and depicted in Figure 1, a flaming message is proposed to influence the likelihood of reaching agreement via effects on concession behaviors of the message recipient as well as the message sender. We theorize that the effect of a flaming message on concession behavior of the recipient depends on whether he/she perceives that the message is directed at him/her (e.g., because the flame sender believes he/she is not being fair) or directed at the negotiation context (e.g., because the flame sender is frustrated due to bad luck). Similarly, we propose that the effect of a flaming message on concession behavior of the sender depends on whether the flame is directed at the recipient or the context. Details of the Figure 1 model and associated hypotheses are provided next.

Figure 1.

Research model.

Full figure and legend (20K)

Concession and agreement

Concession is often described in relation to the norm of reciprocity, which guides many negotiations. Reciprocity is an aspect of fairness (Fehr & Gachter, 2000), and dictates that favorable treatment provided by one party is to be met by the other party returning the favor (Gouldner, 1960). This sense of obligation has been credited with the notion of being human (Leakey & Lewin, 1978), with enabling the division of labor thereby facilitating the human adaptive ability (Tiger & Fox, 1971), and with facilitating 'sophisticated and coordinated systems of aid, gift giving, defense, and trade...' (Cialdini, 1993). Within negotiations, this norm prescribes that a concession by an individual (e.g., a seller lowering his/her price) should be met with a concession by his/her opponent (e.g., the buyer raising his/her price) (Whatley et al., 1999; Eisenberger et al., 2001). It follows then that concession by one dyad opponent ought to promote concession by the other dyad opponent that in turn promotes further concession, resulting in a cycle of concession between negotiating opponents. Since agreement occurs when the opponents' demands are equal, the likelihood of agreement increases as concessions are made (Pruitt, 1981). This leads to the following hypotheses.

H1a

Concession by a flame recipient increases the likelihood of agreement.

H1b

Concession by a flame sender increases the likelihood of agreement.

Opponent-directed flaming and concession

Flame recipient concession. A negotiator's concessions can be encouraged to the extent that he/she feels common bonds with his/her dyad opponent (Pruitt, 1981). Such bonds are promoted when negotiators behave in accordance with their common interactional norms (Postmes & Spears, 1998; O'Sullivan & Flanagin, 2003). A flaming message consists of negative emotional communication that is counter to such norms. This type of communication can thus cause a flame recipient to evaluate the negotiation and flame sender less favorably (van Kleef et al., 2004b), thereby reducing common bonds, such as trust, and associated concession (Pruitt, 1981; Morrow et al., 2004). When a recipient believes that a flame is being employed as part of an intimidation strategy (Jones & Pittman, 1982; Rothbart & Hallmark, 1988; Goldner & Donath, 2004), this may result in an even less favorable evaluation of the flame sender, which further reduces common bonds and associated concession (Pruitt, 1981; Jones & George, 1998; Morrow et al., 2004).

Furthermore, flaming is the hostile communication of emotions (Selfe & Meyer, 1991; Lea et al., 1992), and such communication is likely to evoke feelings of anger in the recipient because of emotional contagion, in which one individual unconsciously mimics emotions displayed by another (Hatfield et al., 1994; Thompson et al., 1999, Van Kleef et al. 2008). Anger reduces the flame recipient's motivation to resolve differences (Friedman et al., 2004) and increases the recipient's competitiveness (Kopelman et al., 2006; Van Kleef et al., 2008). Coupled with a reduction in common bonds, this anger further reduces trust and interpersonal rapport (Allred et al., 1997; Moore et al., 1999; Dunn & Schweitzer, 2005), thereby discouraging concession (Pruitt, 1981; Donohue & Ramesh, 1992; Lewicki & Wiethoff, 2000).

However, these 'powerful affective tendencies' can be overcome by strategic considerations (Steinel et al., 2008, p. 367). For example, a flame can be perceived as a signal (i.e., tracking information) that the flame sender has reached his/her limit and that an impasse is likely. Faced with this information, the flame recipient has been found to be more likely to concede so that agreement can be reached (e.g., Van Kleef et al., 2004a). Such strategic behavior has been termed mismatching (Pruitt & Carnevale, 1993), where the recipient concedes more when the sender displays anger. In subsequent research, Steinel et al. (2008) found that when anger was directed at the opponent him/herself (i.e., opponent-directed) this had little tracking information and, as a result, the opponent's strategic considerations did not overcome his/her affective tendencies, and he/she reduced concession behavior. Taken together, this leads to the following hypothesis.

H2a

An opponent-directed flaming message will decrease concession by the flame recipient.

Flame sender concession. Opponent-directed flaming can result from feelings of anger when the flame sender believes that his/her dyad opponent has violated fairness norms. For example, such a violation may be perceived when the dyad opponent employs a positional commitment strategy by not making concessions (Pruitt, 1981). Employing this strategy, the opponent holds firm to his/her position and does not reciprocate concessions. This lack of reciprocation violates a fairness norm, which dictates that favorable treatment provided by one negotiating party is to be returned by the other party (Gouldner, 1960). Such perceptions of unfairness promote feelings of anger (Bies, 1987; Pillutla & Murnighan, 1996; Scherer, 1999), which can result in opponent-directed flaming (Murnighan & Straub, 1995; Pillutla & Murnighan, 1996). Although there may be social norms restricting the venting of anger in the form of flaming (Allred et al., 1997; Van Kleef et al., 2004a), anger can reduce individuals' cognitive control (Rogers, 1980; Mathieson, 2007), and so individuals will express their anger as a form of disrespect and reproach (Felson, 1982).

Flames may also be directed at the opponent as part of an intimidation strategy. With such a strategy, communications that appear to result from anger are sent in order to instill fear in the recipient, thereby making the recipient more submissive and willing to make concessions (Jones & Pittman, 1982; Rothbart & Hallmark, 1988; Goldner & Donath, 2004). However, we limit our discussion and hypothesis to flaming that results from anger rather than from an intimidation strategy (this limitation is examined later in the Discussion section).

When flaming results from anger, the act of flaming can increase concession by the flame sender in the following ways. Opponent-directed flaming is aggressive behavior in the form of negative socio-emotional communication (Berkowitz, 1988; Colomb & Simutus, 1996; Parks & Floyd, 1996; Thompsen & Fougler, 1996). Catharsis theory holds that acting aggressively against a provoking individual (e.g., venting) will improve one's psychological state (Breuer & Freud, 1955). The theory suggests that it is better to let anger out by directing it against a provoking entity and thereby reducing arousal (Geen & Quanty, 1977), rather than keep it inside and have it build up to the point at which a more dangerous uncontrollable reaction may result (Bushman, 2002). Furthermore, studies have shown that people often engage in such aggressive behavior because they believe that it improves their mood (Geen & Quanty, 1977; Bushman et al., 2001; Bushman, 2002). In other words, aggression is undertaken in the service of affect-regulation goals (Bushman et al., 2001); for example, Alonzo & Aiken (2004) found that a reduction in anxiety was associated with the sending of flaming comments. Such decreases in arousal and improvements in mood that can result from flaming induce helpfulness and cooperation thereby encouraging concession (Baron, 1990; Barry & Oliver, 1996; Forgas, 1998).

Moreover, feelings of anger are likely to cause the flame sender to engage in self-regulation, because people are motivated to regulate their emotions in order to feel good (Freitas & Salovey, 2000; Tice et al., 2001) and in order to increase their likelihood of goal attainment (Baumeister et al., 1994; Freitas & Salovey, 2000). Self-regulation can take the form of reducing one's level of aspiration (Baumeister et al., 1994), which produces smaller demands and more concessions (Pruitt, 1981). Finally, when negotiators are involved with repeated negotiations (as they are in our study), they are concerned with maintaining a good relationship (Fisher & Ury, 1991). In such circumstances, negotiators' current interests and the quality of their relationship are frequently traded off against each other (Fisher & Ury, 1991; Lax & Sebenius, 1991), which can result in greater concessions in order to mitigate damage to the relationship resulting from the flame that was sent. Taken together, this discussion leads to the following hypothesis.

H2b

Opponent-directed flaming resulting from anger will increase concession by the flame sender.

Context-directed flaming and concession

The negotiation context includes all aspects of negotiation except for the negotiators themselves. Consistent with Catharsis theory (Bushman, 2002), an individual can become angry and behave aggressively toward aspects of the negotiation context (Berkowitz, 1988). For example, a survey of 1250 workers found that 80% had seen colleagues shouting abuses at their personal computers (Mori, 1999). In general, individuals become angry at the context when elements of the context, such as the computer or communication media, frustrate their ability to resolve problems or fulfill their needs (Berkowitz, 1988). Such anger can lead to aggressive behavior, such as flaming (Berkowitz, 1988; Anderson & Bushman, 2002; Friedman et al., 2004).

Flame recipient concession. Context-directed flaming provides a means for venting anger toward the circumstances surrounding the negotiation (such as bad luck or the slowness of the communication channel). Such communication may evoke feelings of sympathy in the flame recipient because it is frequently interpreted as a forceful plea for support (Heise & O'Brien, 1993; Galinsky et al., 2005). Since concession is a manifestation of support for another's interest (Pruitt, 1981; Kopelman et al., 2006), a flame recipient's concession is likely to increase with context-directed flaming. In addition, when a display of context-directed anger is received by an individual, this can be perceived as unpleasant by that individual (Kopelman et al., 2006), and can evoke a strong desire to end negotiation through, for example, increased concession (Forgas, 2001; Sinaceur & Neale, 2005; Kopelman et al., 2006). Taken together, this leads to the following hypothesis.

H3a

Context-directed flaming will increase flame recipient concession.

Flame sender concession. Context-directed flaming can result when individuals become angry at circumstances surrounding the negotiation that frustrate their ability to resolve problems or fulfill their needs (Berkowitz, 1988; Klein et al., 2002). As with opponent-directed flaming described earlier, context-directed flaming provides a means for such anger to be vented, and it reduces an individual's level of arousal and improves their mood (Geen & Quanty, 1977; Bushman et al., 2001). Such decreases in arousal and improvements in mood that can result from flaming induce helpfulness and cooperation, and thereby encourage concession (Baron, 1990; Barry & Oliver, 1996; Forgas, 1998). This leads to the following hypothesis.

H3b

Context-directed flaming will increase flame sender concession.

Method

One hundred and forty-eight undergraduate students (40% female) in a large university participated in a study designed to test the hypotheses described above. Each student was randomly assigned to a dyad (resulting in 74 dyads) and the role of either seller or buyer. The identity of each dyad member was unknown to the other. Individuals participated in nine rounds of negotiating with the same opponent using computer-mediated communication. We next describe the task, how it is modeled in the tradition of experimental economics, and the procedure, followed by a discussion of how the constructs were operationalized.

Task

The task used in this study focuses on economic behavior, where a buyer and a seller set out to agree on a price in order to exchange something of value. This economic bargaining perspective allows us to examine the negotiation process as well as outcome (Lim & Benbasat, 1993). In contrast, for example, we do not take a game-theoretic perspective because that would provide more of an outcome than a process focus (Lim & Benbasat, 1993).

The task that is used in our study is similar to other experimental economic two-player games, such as zero-sum, ultimatum, mixed motives, and repeated games, as well as to what Binmore & Vulkan (1999) refer to as one-on-one bargaining. Specifically, our task is related to others in terms of the Bosman et al.'s (2006) five factors as follows. (1) Task uncertainty – our task is similar to the bribery game (Abbink & Hennig-Schmidt, 2006) in the sense that decisions are contingent on uncertain outcomes. (2) Public vs private decisions – our task is similar to the power-to-take game (Bosman et al., 2006) in the sense that each person makes decisions privately. (3) Number and type of players – our task is similar to the bilateral monopoly game (Rapport et al., 1995) in the sense that there is a seller with a single object that may be sold to only one buyer if there is agreement on a price. But our task is perhaps more similar to a hybrid of an ultimatum game (Güth et al., 1982; Alvard, 2004) with repeated trails and a single-stage bilateral bargaining game (Daniel et al., 1998) than to any other type of game in terms of a combination of the first, second, and third factors.

Our task significantly differs from other games in terms of the following two factors, which are particularly relevant for our study. (4) Rules of engagement – our task makes it possible for players to communicate directly with each other without restrictions on the content of their messages. In many other games, the content of players' messages is restricted to numbers or a few select words. For example, in the bargaining game used by Güth & van Damme (1998), players' messages were limited to three arrays of numbers. (5) Medium of communication – our task is modeled in terms of a text-messaging communication medium. It is important to note that although the communication medium can be critical, it has been ignored in many research studies (Valley et al., 1998; Bosman et al., 2006). With these differences, our task provides the means for us to not only classify flaming but also to link its occurrence to outcomes.

In our game, the negotiation object for each round was a different computer-displayed lottery ticket. Negotiations involved determining a mutually agreeable price for the lottery ticket to be exchanged from one negotiator to the other. Lottery ticket prices were measured in terms of points. Individuals were given $10 for participating and were told that after all rounds were over, one of the rounds would be chosen at random, and the individual in each dyad with the most points for that round would receive an extra $10. This reward scheme helped motivate individuals to negotiate in each round independent of past or future round outcomes. The points associated with each ticket were determined by two chance events. (1) Each ticket was displayed with two potential values (e.g., 15 and 30 points), each of which had a 50% chance of being the actual value. (2) Along side the ticket, a prize wheel was displayed (see Figure 2) that had the following differently colored portions: one (say 75%) representing the likelihood of winning if the actual value were 15, and one (say 95%) representing the likelihood of winning if the actual value were 30. With this example, the ticket holder had a 37.5% chance of winning 15 points (0.5 0.75), a 47.5% chance of winning 30 points (0.5 0.95), and a 15% chance of winning nothing (0.5 0.25+0.5 0.05). If the seller decided to keep the ticket and won the lottery (the prize wheel landed on the win area), he/she received points equal to the actual ticket value; if the seller lost, he/she got zero points. In either case, the buyer received zero points. If the seller sold the ticket, he/she received points equal to the selling price, and the buyer received points equal to his/her lottery winnings (determined as described above for the seller) less the amount he/she used to purchase the ticket.

Figure 2.

Example of prize wheel(s). (Since colors cannot be reproduced here, the colored prize wheel displayed to dyads is illustrated above as two different prize wheels.)

Full figure and legend (42K)

Sellers and buyers were situated in separate locations. Each seller and buyer used a personal computer to see the lottery tickets and prize wheels as well as to converse with each other, using their PCs for text-based instant messaging. Personal computers were connected such that each simultaneously displayed the same information to individuals in the same dyad.

Procedure

Prior to the start of the task, individuals received instructions about the game, watched a presentation of the game, and were allowed three practice rounds of negotiation with one of the authors. During these exercises, individuals were encouraged to ask questions if they were uncertain about the game procedures and objectives. In the practice rounds, they were told that the task began when a lottery ticket, a prize wheel, a timer, and instructions appeared on their computer screen. The buyers (sellers) were also told that sellers (buyers) were in a remote location and would see the same information at the same time. Further, they were told that either individual could initiate the negotiation round by contacting their dyad opponent via the PC. Finally, they were told that they had three and a half minutes to negotiate in each round. (This time limit was based on a pilot study in which dyads had no time limit. The maximum time taken by any dyad was about 3½ min) Before the time elapsed, the seller had to indicate via the PC that he/she was selling the ticket (if an agreement with the buyer was reached) or that he/she was keeping the ticket. If the seller did not make such an indication and the time elapsed, both the seller and buyer would receive no points.

Identification of flaming communication

Earlier, we defined flaming as the anti-normative hostile communication of emotions. Therefore, a three-step process was employed to identify flaming. First, transcripts of the negotiations were examined, in order to identify hostile communication. Two student raters blind to the research hypotheses worked independently and coded each negotiation message in terms of the degree to which it conformed to; Bales (1951) negative socio-emotional communication category 12: an individual shows antagonism, deflates other's status, and/or defends or asserts self. The resulting inter-rater reliability was 0.88.

Second, these students independently evaluated the category 12 messages regarding the degree to which they appeared to be anti-normative. The resulting inter-rater reliability was 0.95. These students then discussed and reconciled their different evaluations, and finally agreed on messages they believed were flames. This resulted in 25 dyads that had at least one round in which at least one flame occurred. For dyads having more than one flaming round, the round in which the first flame occurred was chosen.

Third, as described next, a different set of student raters blind to the research hypotheses employed multiple scales to evaluate the degree to which context and opponent-directed flaming occurred in these 25 rounds. This approach is reasonable since individuals are generally able to discern others' communicated emotional states (Ekman, 1993; Kopelman et al., 2006). For example, Turnage (2007) found that the degrees to which students perceived messages as flaming were relatively consistent and were not affected by student age or gender.

Preparation of transcripts

Data to test the hypotheses were gathered by having student raters not involved in the task examine the communication between dyad opponents and determine levels of context and opponent-directed flaming. Transcripts of dyad communication were prepared as follows to facilitate these determinations:

1. As described above, 25 dyads had at least one flaming message. These dyads represented 1/3 of the total dyads. When flaming occurred in more than one round for a dyad, only the round where the first flame occurred was selected. This resulted in 25 transcripts in which flaming occurred.
2. Transcripts for each flaming dyad were edited such that messages after the first flame in the flaming round were deleted. For example, if in round 4 there were 20 messages, and the first flame occurred on the 10th message, then messages 11–20 were deleted.
3. Transcripts of all rounds after the flaming round were deleted.
4. Transcripts of all rounds up to but not including the round prior to the flaming round were deleted. In this example, transcripts for rounds 1 and 2 were deleted; the round 3 transcript was included.
5. The lottery outcome (how many points were won by whom) for the round prior to the flaming round was included with the transcript. In this example, the outcome might indicate that the buyer purchased the ticket for 8 points, although he/she received zero points from the round 3 lottery.

This resulted in 25 transcripts, 17 in which the last communication (the flame) was from the buyer and 8 in which the last communication (the flame) was from the seller.

Operationalization of flaming constructs

Context-directed flaming and opponent-directed flaming were measured via questionnaires provided to student raters who were not involved in the identification of flaming communication. Two student raters examined the 25 transcripts and completed a three-item questionnaire, which asked them to evaluate the degree to which opponent-directed flaming occurred (see Table 1). Two other student raters examined the 25 transcripts and completed a three-item questionnaire, which asked them to evaluate the degree to which context-directed flaming occurred (see Table 1) The following are example flames and their average ratings by the student raters:

• 'actually let hurry up I wanna get the f^**k out'. This was in response to nine buyer and seller messages in which no offers were made. This flame had a context-directed rating of 6.3 and an opponent-directed rating of 3.6. (Asterisks replace actual letters sent by participants.)
• 'wtf bitch u went up????'. This was the buyer response to the seller initially offering 25 and then going up to 27. This flame had a context-directed rating of 1 and an opponent-directed rating of 6.

Table 1 - PLS construct indicators, factor loadings, average variances extracted, and composite scale reliabilities for reflective constructs.

Full table

Operationalization of unfairness and anger constructs

Although not part of the model illustrated in Figure 1, items relating to opponent unfairness and anger were collected and are addressed in the Discussion section below. Two student raters not involved with evaluating the degree of opponent or context-directed flaming examined the 25 transcripts and completed a two-item questionnaire that evaluated the degree to which unfairness appeared to exist (see Table 1). Similarly two other student raters completed a three-item questionnaire for anger (see Table 1). The example negotiations that included the flames described above had the following unfairness and anger ratings by the student raters:

• 'actually let hurry up I wanna get the f^**k out'. This negotiation had an opponent unfairness rating of 4.5 and an opponent-directed anger rating of 1.7. (Asterisks replace actual letters sent by participants.)
• 'wtf bitch u went up????'. This negotiation had an opponent unfairness rating of 7.0 and an opponent-directed anger rating of 6.3.

Operationalizations of concession and agreement

Concession was determined for the round in which the flame occurred. Concession is typically measured as the difference between the initial and final offers (e.g., Komorita & Kravitz, 1979; Smith et al., 1982; Mannix & Neale, 1993). However, this ignores the potential that the first offer might include a concession; for example, a seller may make his/her initial offer lower than he/she would otherwise because he/she likes the buyer (e.g., Kopelman et al., 2006) or because he/she wants to increase the likelihood that the buyer will reciprocate with a high counter offer (Latz, 2003). Therefore, we use the difference between an individual's last quote and the average of equivalent individuals' last quotes to measure concession.

This equivalency is determined based on two criteria: (1) whether the individual is a buyer or seller and (2) based on the bargaining round involved. The first criterion is important because buyers' first quotes will generally be lower than sellers' first quotes. The second criterion is important because this controlled for the different lottery ticket expected values employed in the study, since these values were the same for all dyads in the same round. For example, if the individual is a seller and the bargaining round is 4, then the last quotes of sellers in bargaining round 4 in dyads that did not flame were averaged; the individual's last quote was then subtracted from this average. Thus, positive values indicated more concession, since seller concession results in price decreases. This calculation was similar for an individual who was buyer in bargaining round 4, except that the average of buyers' last quotes in round 4 was subtracted from the individual's last quote so that positive values indicated more concession.

An agreement is defined as a lottery ticket being exchanged from seller to buyer. The agreement construct had a value of 1 when an exchange occurred and a value of 0 when it did not.

Construct variances

There was no manipulation of the independent constructs. Variances found in the flaming constructs were therefore due to naturally occurring individual differences within and across dyads. One might reasonably expect naturally occurring variances in opponent-directed flaming due to differences negotiation strategies attempted by opponents. For example, if an opponent maintained a tough bargaining strategy, this resulted in very little concession (Sinaceur & Tiedens, 2005). Such a strategy could be perceived as unfair and result in anger and associated flaming because concessions by an individual were not reciprocated by his/her opponent (Gouldner, 1960; Torestad, 1990; Fehr & Gachter, 2000).

In contrast, naturally occurring variance associated with context-directed flaming does not appear as straightforward. For example, since all dyads are using the same computer communication media, there is less reason to believe that anger from frustration with the media would exist due to differing perceptions of media richness. However, prior research has found considerable variance among such perceptions with individuals employing the same media because these perceptions are in part socially defined (Fulk et al., 1987; Hollingshead et al., 1993).

Control constructs

Two control constructs were employed with paths leading to both concession constructs and to agreement. The first control construct reflected the two types of lottery tickets that were the object of negotiation. One type differed from the other in terms of the expected value of winning. For example, the ticket offering 15 and 30 and is illustrated in Figure 2 had a higher expected value. A lower expected value ticket would have greater areas in the prize wheel where zero points would be received. The second construct reflects whether the buyer or the seller sent the flaming message; this controls for the fact that twice as many flames were sent by buyers than sellers.

Results

In order to reduce problems associated with common method bias (Podsakoff et al., 2003), multiple questionnaires were created such that questions relating to constructs and their antecedents were not in the same questionnaire. Two student raters, blind to the hypotheses, were assigned to each questionnaire; working independently, they responded to the questionnaire items relative to each of the 25 dyad transcripts. Inter-rater reliability was good, with 0.89 for the context and 0.91 for the opponent-directed questionnaire items. In addition, two other student raters, also blind to the hypotheses, were assigned to the unfairness questionnaire and two more students were assigned to the anger questionnaire. Working independently these students responded to questionnaire items relative to each of the 25 dyad transcripts. Inter-rater reliability was good, with 0.88 for the unfairness and 0.93 for the anger questionnaire items. Therefore, responses were averaged across the two respondents for each questionnaire, and resulted in 25 data points for each construct.

These data were analyzed with partial least squares (PLS). Given the relatively small sample size and the fact that the two flame constructs have multiple indicators, the use of PLS was preferred to covariance structure analysis approaches such as LISREL (which typically require larger samples) and to more traditional techniques such as 2SLS (which cannot accommodate multiple measures) (Wold, 1985; Chin, 1998; Hulland, 1999). PLS can be used when there are as few as five for each path leading to the construct that has the most incoming paths or the formative construct with the most indicators, whichever is greater (Falk & Miller, 1992).

The path coefficients obtained from a PLS analysis are standardized regression coefficients, while the loadings of items on individual constructs are factor loadings. Factor scores created using these loadings are equivalent to weighted composite indices. Thus, PLS results can be easily interpreted by considering them in the terms of regression and factor analysis. As such, PLS allows us to assess how well the set of items used to measure one construct (e.g., opponent-directed flame) served to predict a dependent construct (e.g., flame recipient concession). Before examining the relationships among our theoretical constructs, the measures employed to represent the constructs (the measurement model) must be determined to be valid and reliable.

To ensure that items reliably measure constructs they are designed to measure, each item should have a loading of at least 0.6 on its own construct (Bagozzi & Yi, 1988). As illustrated in Table 1, all items exceed this requirement. Composite scale reliabilities were calculated for the constructs based on item loadings, and as illustrated in Table 1, were all above the 0.70 threshold (Nunnally & Bernstein, 1994). Construct convergent and discriminant validities were determined by comparing the average variance extracted for each construct with the covariance among constructs. When the average variance extracted for a construct is larger than the construct's covariance with other constructs, this indicates that each construct is more highly related to its own measures than to other constructs, and that convergent and discriminant validities are supported. As presented in Table 2, this is the case for all constructs. Moreover, all average variances extracted were above the recommended level of 0.50 (Fornell & Larcker, 1981; Chin, 1998).

Table 2 - For reflective constructs: Average variance extracted by PLS constructs (diagonal elements) and shared variance between constructs (off-diagonal element).

Full table

Descriptive statistics are provided in Table 3, and PLS results are illustrated in Figure 3. PLS generates estimates of standardized regression coefficients for the paths in a model's structural component. In order to determine the significance of these paths, jackknifed standard error estimates for the paths were obtained using the blindfolding procedure with an omission distance of 11 (Sambamurthy & Chin, 1994). Figure 3 also shows R² values in parentheses for concessions and agreement, which is the proportion of variance of a construct explained by constructs having paths leading to it. This analysis leads to the following hypothesis evaluations.

Figure 3.

Research model: PLS results (all paths are significant at P<0.05).

Full figure and legend (23K)

Table 3 - Descriptive statistics for complete data set.

Full table

All hypotheses are supported at P<0.05. When a flaming message is sent that is directed at the opponent, the recipient reduces concessions (H2a) while the individual sending the flame increases concessions (H2b). When a flaming message is sent that is directed at the context, the recipient increases concessions (H3a) while the individual sending the flame increases concessions (H3b). In addition, concessions by both flame recipient and flame sender increase agreement (H1a and H1b). Finally, the model accounts for a reasonable amount of construct variance, explaining an average of 24% of construct variation.

H2b is predicated on the belief that opponent-directed flames were related to flame sender anger that is directed toward the opponent. Including the anger construct (described earlier) with a path to opponent-directed flame and rerunning the model resulted in a path coefficient of 0.48 (significant at P<0.05), explaining 23% of the variance in opponent-directed flame. This supports the contention that opponent-directed flaming messages were associated with anger toward the opponent. However, flame sender anger could be part of an intimidation strategy, rather than resulting from the flame sender's feelings of anger. As described earlier, feelings of anger can come from perceptions that the bargaining opponent is being unfair. An important aspect of fairness is whether the opponent adheres to the reciprocity norm, which, for example, requires favors by one individual to be met with favors by the other (Gouldner, 1960). Therefore, two unfairness questions described earlier were included in a construct with a path leading to anger and the model to rerun. This resulted in a coefficient of 0.51, with unfairness explaining 26% of anger variance. The significant paths from unfairness to anger and from anger to opponent-directed flame thus support the assumption surrounding H2b that these flames resulted from flame sender feelings of anger rather than from a flame sender strategy.

Discussion

Our analysis illustrated in Figure 3 indicates that when an individual sends a flame in anger directed at his/her opponent, the likelihood of agreement is slightly reduced, with a combined path coefficient of -0.03 (as determined by: 0.29 0.38-0.40 0.36). Although not specifically tested here, prior research indicates that the flame sender will significantly reduce concession when opponent-directed flaming is employed as an intimidation strategy (Thompson & Nadler, 2002; Kopelman et al., 2006). This would decrease the H2b path coefficient in our model, suggesting that such a strategy may be risky due to a significantly decreased likelihood that agreement will be reached.

Although opponent-directed flaming reduces the likelihood of reaching an agreement, it is not clear from Figure 3 whether the outcome when agreement is reached might be increased enough for the flame sender such that it overcomes this reduced likelihood and thereby makes opponent-directed flaming an attractive strategy (or at least does not make it a liability). To examine this issue, the data set was reduced from 25 to 11, representing all rounds in which the agreement was reached. As illustrated in Figure 4, when there was opponent-directed flaming and agreement was reached, concessions by flame recipients were significantly reduced and those by flame senders were significantly increased. This indicates that flame senders not only risk lower likelihoods of agreement, but when agreement is reached they are also likely to be worse off than the flame recipients.

Figure 4.

Impact of flame when there was agreement (all paths, except that noted as 'ns', are significant at P<0.05).

Full figure and legend (20K)

In contrast to opponent-directed flaming, when an individual sends a flame directed at the context, the likelihood of agreement is significantly increased, with a combined path coefficient of 0.22 (as determined by 0.24 0.38+0.37 0.36 in Figure 3). However, in accordance with opponent-directed flaming, when agreement is reached, context-directed flames result in outcomes that are disadvantageous to the flame sender (Figure 4). Although not tested here, one might use these results to speculate on the potential for using a context-directed flaming strategy. As theorized earlier, and supported by our analyses, context-directed flames can evoke feelings of sympathy in the flame recipient and thereby result in a significantly increased likelihood of agreement. This increased likelihood comes from increases in concession by both the flame sender as well as the recipient. Concession by the flame sender was theorized to be due to an improvement in the sender's mood that resulted from the venting of his/her anger. However, if the sender used context-directed flaming as a strategy, then venting would not be an issue and he/she might be able to reduce his/her degree of concession (i.e., reducing the H3b path coefficient in Figure 3) while maintaining a high level of concession from the flame recipient. In such circumstances, strategically communicated context-directed flames could well result in higher likelihoods of agreement without substantial concession losses to the flame sender.

Work by Steinel et al. (2008) suggests that there might be a third contingency of interest, in addition to whether the flames are directed at the opponent or the context. In accord with our results, they found that when anger was directed at the opponent him/herself (e.g., 'This person makes me really angry'), the opponent reduced concessions. However, when anger was focused specifically on the opponent's offer behavior (e.g., 'This offer makes me really angry'), the opponent increased concessions. Steinel et al. (2008) suggested that this increase was due to the opponent using the expressed anger as tracking information, and rather than risk an impasse (no agreement), the opponent increased concession behavior. Thus, flames directed at the opponent's offer behavior may act in the same way as context-directed flames, but for a different reason.

To explore this possibility, we examined our 25 flames and found only three that appeared to be directed specifically at the opponent's offer behavior; these are depicted in Table 4, along with their context-directed and opponent-directed ratings. Our raters tended to interpret these flames as directed toward the opponent (mean of 4.33) rather than the context (mean of 3.22), suggesting that Steinel et al.'s distinction between opponent-directed and opponent's offer-directed may have been too subtle for our dyads. It may be that when anger is expressed in terms of a flame (anti-normative and abusive), such a distinction tends to be lost to negotiators. This seems to be supported by the flames in Table 4, which we ordered in terms of abusiveness. The first flame ('sucky') was the least abusive, and our raters found it to be more context than opponent-directed (4.00 vs 2.33, respectively). The last flame ('shove it') was the most abusive, and was rated as more opponent than context-directed (4.33 vs 3.22). The middle flame was rated as about the same in terms of context and opponent-directed (4.67 vs 5.00).

Table 4 - Flames with a focus on opponent's offer.

Full table

Limitations and future research

Contributions of this study should be viewed with recognition of its limitations, and implications of these limitations for future research. The context of the study may limit its external validity because our subjects were strangers to each other. However, this is less limiting than it might appear since there is an increasing incidence of such negotiations, at least for initial encounters and associated agreements. This is especially true with online dispute resolution, which often occurs between buyers and sellers who are strangers to each other (Global Investment Technology, 2003; Tyler, 2004).

However, flaming decreases when individuals who are not strangers negotiate via computer media (Thompson & Nadler, 2002; Clark, 2004). Thus, in such circumstances, our model may require modification. For example, when negotiators are friends they may be more altruistic and less likely to engage in negotiating behavior that angers each other (Donohue & Ramesh, 1992). In contrast, when negotiators are friends, perceptions that one or the other is being unfair may lead to greater anger than if they were strangers (Mannix et al., 1995). It is therefore of interest to expand the examination of flaming during negotiation to the population of individuals who are familiar with each other.

Time limits are common in negotiation because negotiators desire an end to the uncertainty and emotional demands of the process, particularly when the negotiators' constituents press for quick resolution (Carnevale et al., 1993; De Dreu, 2003). However, negotiations with little or no time limitations would be interesting to examine. For example, more negotiation time can lead negotiators to better understand each other, and thereby help them more accurately predict and interpret their opponent's actions (Donohue & Ramesh, 1992). This might reduce the effect of flaming since negotiators may try to reduce behaviors that can anger each other.

Our investigation was within the context of distributive negotiation concerning a single issue. It would be helpful if the associations we have identified among the antecedents, anger, and flaming were also tested within integrative negotiations over multiple issues. Since integrative negotiation focuses on win-win agreement (Thompson & Nadler, 2002), principles of fairness are less likely to be violated and thus opponent-associated anger and flaming may also be less likely. However, if more issues are under negotiation, the potential for task uncertainty increases along with the associated potential for increased context-associated anger.

We used individuals other than the study participants to determine the extent to which participants flamed, were angry, and felt their opponents were unfair. However, assessing these constructs from actively engaged participants would raise other concerns. For example, if protocol analysis were employed, there would be concern about potential reactive effects, as participants would be required to interrupt negotiation in order to describe their psychological states. In addition, if post hoc questionnaires were administered to participants, questions regarding the accuracy of participant perceptions would arise. However, it would be interesting to triangulate with these other data gathering strategies to see if they produce results that are in accordance with ours.

Our focus on anger does not include strategic motivation as an antecedent to flaming. An attempt to intimidate one's opponent through demonstrated anger can be strategically motivated, independent from feelings of anger (Berkowitz, 1993; Geen, 2001). For example, in a sporting event a player may make a verbal attack on his/her opponent not because of anger, but because he/she is attempting an intimidation strategy. Similarly, negotiators can display anger in the form of flaming (Colomb & Simutus, 1996; Parks & Floyd, 1996; Thompsen & Fougler, 1996) that is strategically motivated (Barry, 1999; Sinaceur & Tiedens, 2005; Kopelman et al., 2006). This intimidation strategy by an opponent can yield concessions (Sinaceur & Tiedens, 2005; Kopelman et al., 2006) because a negotiator perceives his/her opponent to be dominant (Berkowitz, 1993; Sinaceur & Tiedens, 2005).

Such intimidation can be successful in situations where fear is justified, such as where physical violence or other significant retribution is possible. It is therefore less important in computer-mediated negotiations among strangers and concerning typical goods and services. In addition, the likelihood of intimidation being successful increases when anger is demonstrated via body, facial, and vocal cues (Scherer, 1986; Ekman, 1993; Sinaceur & Tiedens, 2005). However, because these cues are absent in computer-mediated communication, intimidation is less likely to be successful, and negotiators will be less likely to attempt such a strategy (Bandura, 1983; Mischel, 1999; Mischel & Shoda, 1999). We therefore are less concerned here with strategically motivated flaming.

We did not examine the potential of gender to influence the affect of flaming on concession making. Past research indicates that gender can play a role in negotiation. For example, Stuhlmacher & Walters (1999) performed a meta-analysis and found a decrease in the levels of outcomes negotiated by women as opposed to men. It has been suggested that such gender differences in negotiation are due to women being generally less assertive as well as due to their expectations regarding their own and others' behaviors (Stuhlmacher et al., 2007). However, most prior research on gender influence during negotiation has been done in face-to-face contexts, and such differences may disappear during computer-mediated negotiations. This is because such mediation can reduce the communication of social cues, which may make women's behavior less 'consistent with a female gender role' (Stuhlmacher et al., 2007, p. 333). In fact, a meta-analysis by Stuhlmacher et al. found women to be significantly more assertive and hostile when they were not negotiating face-to-face. In addition, Turnage (2007) found no significant differences between women and men regarding their perceptions of computer-mediated communication as being flames. Taken together, this prior research suggests that computer mediation of negotiations might reduce an ability of gender to moderate flame sender and recipient concession. This hypothesis would be interesting to examine empirically.

Our study involved students as participants. While many studies have found students to be good surrogates for older working individuals when involved in groups (e.g., Bettenhausen, 1991) and with decisions involving trust (e.g., Malhotra, 2004), potentially interesting differences have been found relative to participants age and background. For example, Turnage (2007) found that while older and younger individuals (graduate and undergraduate students) identified the same messages as flames, the older individuals typically perceived flaming messages as being more intense (e.g., more hostile or insulting) than the younger individuals. In addition, Sally (1995) found that psychology students tended to cooperate more than economics students. Thus, it would be interesting to replicate our study using older individuals of different backgrounds.

Finally, our study design did not involve the manipulation of independent variables. This has at least two implications. First, our hypotheses and the correlational support given for them by our data provide a basis for further investigation of cause-and-effect relationships in future studies. For example, we found a negative relationship between opponent-directed flaming and concessions by the flame recipient. This finding supported our hypothesis that flame recipients decrease their concessions because the receipt of flames increases their anger and thereby leads to increased competitiveness. However, this negative relationship could also result from a hold firm strategy employed by recipients, in which they do not reciprocate concessions. In such cases, flame senders may feel that the recipients are being unfair (Johnson et al., forthcoming), and may send flaming messages in response.

Second, we could not plan for the existence of flaming; rather, flaming that was observed came naturally during negotiations. As a result, the sample sizes of our analyses were somewhat small, with 25 involved in our primary analysis and 11 involved in our examination of dyads that reached agreement. While, as noted earlier, such numbers exceed the minimums suggested for PLS use, they do raise concerns regarding the power of our analyses. Thus, for example, the non-significant path from context-directed flame-to-flame recipient concession illustrated in Figure 4 may well have been the result of low power rather than the lack of a relationship.

Practical implications

As described earlier, negotiation is a pervasive form of social interaction, which is increasingly being conducted over computer-mediated channels. This trend is being driven by the increase in online transactions enabled by the internet (Moore et al., 1999; Katsh & Rifkin, 2001), with negotiations increasingly employing e-mail and instant messaging (Katsh & Rifkin, 2001; Dorado et al., 2002; Tyler, 2004). However, the use of computer mediation will also increase the likelihood of flaming (Postmes & Spears, 1998; O'Sullivan & Flanagin, 2003; Reinig & Mejias, 2004), which we found can affect negotiators' concession and the likelihood of agreement in the following ways.

When flaming is directed at a negotiating opponent the likelihood of reaching agreement will decrease and, when agreement is reached, outcomes will be less favorable for the flame sender. It is therefore in the best interest of a negotiating individual not to send opponent-directed flames; this would suggest that employing an intimidation strategy is not likely to be beneficial. Interestingly, it may be beneficial for negotiators to provoke their opponents into such flaming, since the slight decrease in agreement likelihood may well be overcome by increased outcome benefits once agreement is reached. This, for example, would support a positional commitment or 'hold firm' strategy that is suggested by the level-of-aspiration theory (Siegel & Fouraker, 1960), which posits that a negotiator should make very small (or no) concessions because it pays to be tough (Yukl, 1974; Pruitt, 1981). This strategy would likely be seen as unfair by the negotiating opponent and thereby provoke opponent-directed flames.

Computer-mediated channels promote a sinister attribution bias, which is the tendency for individuals to be irrationally distrustful of others (Kramer, 1994; Thompson & Nadler, 2002). For example, Fortune & Brodt (2000) found that negotiators interacting via e-mail were more likely to suspect the other party of lying or otherwise deceiving them, relative to negotiators interacting face to face. This bias is also characterized by a tendency to make cynical attributions about the intentions of others (Kramer, 1994, 1999). Such attributions and distrust are more likely to result in a negotiator becoming angry with his/her opponent and engage in opponent-directed flaming. For example, if an opponent is having problems with the communication channel (e.g., he/she is a slow typist), a negotiator may misinterpret these problems as a sinister tactic (e.g., stalling) and become angry with his/her opponent. Therefore, compared to richer forms of communication, computer-mediated negotiation will result in more opponent-directed flaming, which will decrease the likelihood of agreement as well as result in less flame recipient concessions and more flame sender concessions when agreement is reached.

Our findings contradict previous research that found flaming reduces the potential for agreement (e.g., Allred et al., 1997; Friedman et al., 2004; van Kleef et al., 2004b). In our study, context-directed flaming increased rather than decreased this potential. Thus, a negotiator might increase the likelihood of agreement by engaging in context-directed flaming rather than relying solely on making concessions.

Concluding remarks

There is little research regarding the influence of flame senders on flame receivers when using computer-mediated communication (Yuuki et al., 2007). We found that whether flaming directed at the negotiating opponent or at the negotiation context has differential impacts on negotiating behaviors of both flame senders and recipients. And, these different behaviors affect the likelihood of agreement as well as the outcomes once agreements are reached. Given the increasing use of computer-mediated channels for negotiations, it is of interest to continue this line of research in order to increase our understanding of the flaming phenomenon.

References

1. Abbink K and Hennig-Schmidt H (2006) Neutral versus loaded instructions in a bribery experiment. Experimental Economics 9(3), 103–121. | Article |
2. Aiken M and Waller B (2000) Flaming among first-time group support system users. Information and Management 37, 95–100. | Article |
3. Allred KG, Mallozzi JS, Matsui F and Raia CP (1997) The influence of anger and compassion on negotiation performance. Organizational Behavior and Human Decision Processes 70(3), 175–188. | Article |
4. Alonzo M and Aiken M (2004) Flaming in electronic communication. Decision Support Systems 36(3), 205–338. | Article |
5. Alvard M (2004) The ultimatum game, fairness, and cooperation among big game hunters. In Foundations of Human Sociality: Ethnography and Experiments in 15 Small-scale Societies (HENRICH J, BOYD R, BOWLES S, GINTIS H, FEHR E and CAMERER C, Eds), pp 413–435, Oxford University Press, Oxford.
6. Anderson CA and Bushman BJ (2002) Human aggression. Annual Review of Psychology 53, 27–51. | Article |
7. Anderson C and Thompson LL (2004) Affect from the top down: how powerful individuals' positive affect shapes negotiations. Organizational Behavior and Human Decision Processes 95, 125–139. | Article |
8. Ba S and Pavlou P (2002) Evidence of the effect of trust building technology in electronic markets: price premium and buyer behavior. MIS Quarterly 26(3), 243–268. | Article |
9. Bagozzi RP and Yi Y (1988) On the evaluation of structural equation models. Journal of the Academy of Marketing Science 16, 74–94. | Article |
10. Bales RF (1951) Interaction Process Analysis: A Method for the Study of Small Groups. The University of Chicago Press, Chicago, IL.
11. Bandura A (1983) Psychological mechanism of aggression. In Aggression: Theoretical and Empirical Reviews (GEEN RG and DONNERSTEIN EI, Eds), pp 1–40, Academic Press, New York.
12. Baron RA (1990) Environmentally induced positive affect: its impact on self-efficacy, task performance, negotiation, and conflict. Journal of Applied Psychology 20(5), 368–384. | Article |
13. Barry B (1999) The tactical use of emotion in negotiation. In Research in Negotiation in Organizations (BIES RJ and LEWICKI RJ, Eds), Vol. 7, pp 93–121, JAI Press, Stamford, CT.
14. Barry B and Oliver R (1996) Affect in dyadic negotiation: a model and propositions. Organizational Behavior and Human Decision Processes 67(2), 127–143. | Article |
15. Baumeister RF, Heatherton TF and Tice DM (1994) Losing Control: How and Why People Fail at Self-Regulation. Academic Press, San Diego, CA.
16. Berkowitz L (1988) Frustrations, appraisals, and aversively stimulated aggression. Aggressive Behavior 10, 59–73.
17. Berkowitz L (1993) Aggression. McGraw Hill, New York.
18. Bettenhausen KL (1991) Five years of Group Research: what we have learned and what needs to be addressed. Journal of Management 17, 345–381. | Article |
19. Bies RJ (1987) The predicament of injustice: the management of moral outrage. Research in Organizational Behavior 9, 289–319.
20. Binmore K and Vulkan N (1999) Applying game theory to automated negotiation. Netnomics 1, 2–9. | Article |
21. Bosman R, Hennig-Schmidt H and van Winden F (2006) Exploring group decision making in a power-to-take experiment. Experimental Economics 9(1), 35–51. | Article |
22. Breuer J and Freud S (1955) Studies on hysteria. In The Standard Edition of the Complete Psychological Works of Sigmund Freud (JAMES T and STRACHEY A, Eds), Hogarth Press/Institute of Psycho-Analysis, London.
23. Brown P and Levinson SC (1987) Politeness: Some Universals in Language Usage. Cambridge University Press, Cambridge.
24. Bushman BJ (2002) Does venting anger feed or extinguish the flame? Catharsis, rumination, distraction, anger, and aggressive responding. Personality and Social Psychology Bulletin 28, 724–731. | Article |
25. Bushman BJ, Baumeister RF and Phillips CM (2001) Do people aggress to improve their Mood? Catharsis beliefs, affect regulation opportunity, and aggressive responding. Journal of Personality and Social Psychology 81, 17–32. | Article | ChemPort |
26. Carnevale PJ, O'Connor KM and McCusker C (1993) Time pressure in negotiation and mediation. In Time Pressure and Stress in Human Judgment and Decision Making (SVENSON O and MAULE AJ, Eds), pp 117–127, Plenum Press, New York.
27. Chin WW (1998) The partial least squares approach for structural equation modeling. In Modern Methods for Business Research (MARCOULIDES GA, Ed.), pp 295–336, Lawrence Erlbaum Associates, Mahwah, NJ.
28. Cialdini RB (1993) Influence: Science and Practice. Scott, Foresman and Company, Boston.
29. Clark S (2004) Resolving online disputes. Available online at www.abc.net.au/rn/science/buzz/stories/s1088966.htm (retrieved October 2005).
30. Colomb GG and Simutus JA (1996) Visible conversation and academic inquiry: CMC in a culturally diverse classroom. In. Computer-mediated Communication: Linguistic, Social and Cross-cultural Perspectives (HERRING S, Ed.), pp 203–222, John Benjamins, Amsterdam/Philadelphia.
31. Daniel T, Seale D and Rapoport A (1998) Strategic play and adaptive learning in the sealed bid bargaining mechanism. Journal of Mathematical Psychology 42, 133–166. | Article |
32. De Dreu CKW (2003) Time pressure and closing of the mind in negotiation. Organizational Behavior and Human Decision Processes 91, 280–295. | Article |
33. Diener E (1980) Deindividuation: the absence of self-awareness and self-regulation in group members. In The Psychology of Group Influence (PAULUS PB, Ed.), pp 209–242, Lawrence Erlbaum, Hillsdale, NJ.
34. Donohue WA, Diez ME and Stahle RB (1983) New directions in negotiation research. In Communication Yearbook (BOSTROM R, Ed.), Vol. 7, pp 249–279, Sage, Beverly Hills, CA.
35. Donohue WA and Ramesh C (1992) Negotiator–opponent relationships. In Communication and Negotiation (PUTNAM LL and ROLOFF ME, Eds), pp 209–232, Sage, Newbury Park, CA.
36. Dorado M, Medina F, Munduate L, Cisneros I and Euwema M (2002) Computer-mediated negotiation of an escalated conflict. Small Group Research 33, 509–524. | Article |
37. Druckman D and Olekalns M (2008) Emotions in negotiation. Group Decision and Negotiation 17, 1–11. | Article |
38. Dunn JR and Schweitzer ME (2005) Feeling and believing: the influence of emotion on trust. Journal of Personality and Social Psychology 88, 736–748. | Article |
39. eBay (2007) http://news.ebay.com/fastfacts_ebay_marketplace.cfm.
40. Eisenberger R, Armeli S, Rexwinkel B, Lynch PD and Rhoades L (2001) Reciprocation of perceived organizational support. Journal of Applied Psychology 86, 42–51. | Article | ChemPort |
41. Ekman P (1993) Facial expression and emotion. American Psychologist 48(4), 638–652. | Article |
42. Falk RF and Miller NB (1992) A Primer for Soft Modeling. University of Akron Press, Akron, OH.
43. Fehr E and Gachter S (2000) Fairness and retaliation: the economics of reciprocity. Journal of Economic Perspectives 14(3), 159–182.
44. Felson RB (1982) Impression management and the escalation of aggression and violence. Social Psychology Quarterly 45, 245–254. | Article | ChemPort |
45. Fisher R and Ury W (1991) Getting to Yes. Penguin Books, New York.
46. Forgas JP (1998) On feeling good and getting your way: mood effects on negotiator cognition and bargaining strategies. Journal of Personality and Social Psychology 74(3), 565–577. | Article | ChemPort |
47. Forgas JP (2001) On being moody but influential: the role of affect in social influence strategies. In Social Influence: Direct and Indirect Processes (FORGAS JP and WILLIAMS KD, Eds), pp 147–166, Psychology Press, New York.
48. Fornell C and Larcker DF (1981) Evaluating structural equation models with unobservable variables and measurement errors. Journal of Marketing Research 18(1), 39–50. | Article | ISI |
49. Fortune A and Brodt S (2000) Face to face or virtually, for the second time around: the influence of task, past experience, and media on trust and deception in negotiation. Manuscript under review, Fuqua School of Business, Duke University.
50. Freitas A and Salovey P (2000) Regulating emotion in the long and short term. Psychological Inquiry 11, 178–179.
51. Friedman R, Anderson C, Brett J, Olekalns M, Goates N and Lisco CC (2004) The positive and negative effects of anger on dispute resolution: evidence from electronically-mediated disputes. Journal of Applied Psychology 89, 369–376. | Article |
52. Fulk J, Steinfield CS, Schmitz J and Power JG (1987) A social information processing model of media use in organizations. Communication Research 14, 529–552. | Article |
53. Galinsky AD, Ku G and Wang CS (2005) Perspective-taking and self-other overlap: fostering social bonds and facilitating social coordination. Group Processes and Intergroup Relations 8(2), 109–124. | Article |
54. Geen RG (2001) Human Aggression. Open University Press, Philadelphia.
55. Geen RG and Quanty MB (1977) The catharsis of aggression: an evaluation of a hypothesis. In Advances in Experimental Social Psychology (BERKOWITZ L, Ed.), Vol. 10, pp 1–37, Academic Press, New York. | Article |
56. Global Investment Technology (2003) Harborside+ Offers a Large-Block Trading Facility but Building Awareness is the Paramount Challenge. Global Investment Technology, Investment Media Inc..
57. Goldner SA and Donath J (2004) Social Roles in Electronic Communities. Paper presented at the Internet Research 5.0, Brighton, England.
58. Gouldner A (1960) The norm of reciprocity: a preliminary statement. American Sociological Review 25(2), 176–177. | Article |
59. Güth W and van Damme E (1998) Information, strategic behavior and fairness in ultimatum bargaining: an experimental study. Journal of Mathematical Psychology 42, 227–247. | Article |
60. Güth W, Schmittberger R and Schwarze B (1982) An experimental analysis of ultimatum bargaining. Journal of Economic Behavior and Organization 3(4), 367–388. | Article |
61. Hatfield E, Cacioppo JT and Rapson RL (1994) Emotional Contagion. Cambridge University Press, Paris.
62. Heise D and O'Brien J (1993) Emotion expression in groups. In Handbook of Emotions (LEWIS M and HAVILAND JM, Eds), pp 489–497, Guilford, New York.
63. Hiltz SR, Turoff M and Johnson K (1989) Experiments in group decision making, 3: disinhibition, deindividuation, and group process in pen name and real name computer conferences. Decision Support Systems 5, 217–232. | Article |
64. Hollingshead AB, McGrath JE and O'Connor KM (1993) Group task performance and communication technology: a longitudinal study of computer-mediated versus face-to-face work groups. Small Group Research 24(3), 307–333. | Article |
65. Hulland JS (1999) Use of partial least squares (PLS) in strategic management research: a review of four recent studies. Strategic Management Journal 20, 195–204. | Article | ISI |
66. Jensen L (1963) Soviet-American bargaining behavior in the postwar disarmament negotiations. Journal of Conflict Resolution 7, 522–541. | Article |
67. Johnson N, Cooper R and Chin W (forthcoming) Anger and flaming in computer-mediated negotiation among strangers. Decision Support Systems.
68. Jones EE and Pittman TS (1982) Toward a general theory of strategic self-presentation. In Psychological Perspectives on the Self (ULS J, Ed.), Vol. 1, pp 231–262, Erlbaum, Hillsdale, NJ.
69. Jones GR and George JM (1998) The experience and evolution of trust: implications for cooperation and teamwork. Academy of Management Review 23, 531–546. | Article |
70. Katsh E and Rifkin J (2001) On-line Conflict Management. Resolving Conflicts in Cyberspace. Jossey Bass, San Francisco.
71. Kayany JM (1998) Contexts of uninhibited online behavior: flaming in social newsgroups on Usenet. Journal of the American Society for Information Science 49(12), 1135–1141. | Article |
72. Klein J, Moon Y and Picard RW (2002) This computer responds to user frustration: theory, design and results. Interacting with Computers 14, 119–140.
73. Kiesler S, Zubrow D, Moses AM and Geller V (1985) Affect in computer-mediated communication: an experiment in synchronous terminal-to-terminal discussion. Human Computer Interaction 1, 77–104. | Article |
74. Komorita SS and Kravitz DA (1979) The effects of alternatives in bargaining. Journal of Experimental Social Psychology 15, 147–159. | Article |
75. Kopelman S, Rosette AS and Thompson L (2006) The three faces of eve: strategic displays of positive, negative, and neutral emotions in negotiation. Organizational Behavior and Human Decision Processes 99, 81–101. | Article |
76. Kramer RM (1994) The sinister attribution error: paranoid cognition and collective distrust in organizations. Motivation and Emotion 18, 199–230. | Article |
77. Kramer RM (1999) Trust and distrust in organizations: emerging perspectives, enduring questions. Annual Review of Psychology 50, 569–598. | Article | PubMed | ISI | ChemPort |
78. Kowk W, Lee J, Huynh M and Pi S (2002) Role of GSS on collaborative problem-solving learning: a study on knowledge externalization. European Journal of Information Systems 11, 98–107. | Article |
79. Kwon S and Weingart LR (2004) Unilateral concessions from the other party: concession behavior, attributions, and negotiation judgments. Journal of Applied Psychology 89(2), 263–278. | Article |
80. Latz M (2003) Offer/Concession strategy must reflect ultimate goal. Business Journal, February, p. 1.
81. Lax D and Sebenius J (1991) The power of alternatives or the limits to negotiation. In Negotiation Theory and Practice (BRESLIN JW and RUBIN JZ, Eds), pp 97–114, The Program on Negotiation at Harvard Law School, Cambridge.
82. Lea M, O'Shea T, Fung P and Spears R (1992) Flaming in computer-mediated communication: observations, explanations, implications. In Contexts of Computer-mediated Communication (LEA M, Ed.), pp 89–112, Harvester Wheatsheaf, London.
83. Lea M and Spears R (1991) Computer-mediated communication, de-individuation and group decision-making. International Journal of Man–Machine Studies 34(2), 283–301. | Article |
84. Leaky R and Lewin R (1978) People of the Lake. Anchor Press/Doubleday, New York.
85. Lee H. (2005) Behavioral strategies for dealing with flaming in an online forum. The Sociological Quarterly 46, 385–403. | Article |
86. Lewicki RJ and Wiethoff C (2000) Trust, trust development and trust repair. In The Handbook of Conflict Resolution: Theory and Practice (DEUTSCH M and COLEMAN PT, Eds), pp 86–107, Jossey-Bass Publishers, San Francisco.
87. Lim L and Benbasat I (1993) A theoretical perspective of negotiation support systems. Journal of Management Information System 9(3), 27–44.
88. Magneau JM and Pruitt DG (1979) The social psychology of bargaining: a theoretical synthesis. In Industrial Relations: A Social Psychological Approach (STEPHENSON GM and BROTHERSON CJ, Eds), John Wiley & Sons, New York.
89. Malhotra D (2004) Trust and reciprocity decisions: the differing perspectives of trustors and trusted parties. Organizational Behavior and Human Decision Making 94, 61–73. | Article |
90. Mannix E and Neale M (1993) Power imbalance and the pattern of exchange in dyadic negotiation. Group Decision and Negotiation 2, 119–133. | Article |
91. Mannix E, Neale MA and Northcraft GB (1995) Equity, equality, or need? The effects of organizational culture on the allocation of benefits and burdens. Organizational Behavior and Human Decision Processes 63, 276–286. | Article |
92. Martin J (1997) Miss Manners' Basic Training: Communication. Crown, New York.
93. Mathieson K (2007) Towards a design science of ethical decision support. Journal of Business Ethics 76, 269–292. | Article |
94. McKersie RB and Fonstad NO (1997) Teaching negotiation theory and skills over the internet. Negotiation Journal 13, 363–368.
95. Mischel W (1999) Introduction to Personality, 6th edn. Harcourt Brace, Orlando.
96. Mischel W and Shoda Y (1999) Integrating dispositions and processing dynamics within a unified theory of personality: the Cognitive Affective Personality System (CAPS). In Handbook of Personality: Theory and Research 2e (PERVIN LA and JOHN O, Eds), pp 197–218, Guilford, New York.
97. Moore D, Kurtzberg T, Thompson L and Morris M (1999) Long and short routes to success in electronically-mediated negotiations: group affiliations and good vibrations. Organizational Behavior and Human Decision Processes 77(1), 22–43. | Article |
98. Moral-Toranzo F, Canto-Ortiz J and Gómez-Jacinto L (2007) Anonymity effects in computer-mediated communication in the case of minority influence. Computers in Human Behavior 23(3), 1660–1674. | Article |
99. Mori (1999) Employees get 'It' out of their systems. www.mori.com/polls/1999/rage.shtml.
100. Morley IE and Stephenson GM (1977) The Social Psychology of Bargaining. George Allen and Unwin Ltd, London.
101. Morris CG (1970) Changes in group interaction during problem-solving. Journal of Social Psychology 81, 157–165.
102. Morris M, Nadler J, Kurtzberg T and Thompson L (2002) Schmooze or lose: social friction and lubrication in e-mail negotiations. Group Dynamics 6(1), 89–100. | Article |
103. Morrow JL, Hansen MH and Pearson AW (2004) The cognitive and affective antecedents of general trust within cooperative organizations. Journal of Managerial Issues 16(1), 48–64.
104. Murnighan JK and Straub PG (1995) An experimental investigation of ultimatums: common knowledge, fairness, expectations, and lowest acceptable offers. Journal of Economic Behavior and Organization 27, 345–364. | Article |
105. Nunnally J and Bernstein JC (1994) Psychometric Theory, 3rd edn. McGraw-Hill, New York.
106. O'Sullivan PB and Flanagin AJ (2003) Reconceptualizing "flaming" and other problematic messages. New Media & Society 5, 69–94. | Article |
107. Parks M and Floyd K (1996) Making friends in cyberspace. Journal of Communication 46(1), 80–97. | Article |
108. Pesendorfer E and Koeszegi ST (2006) Hot versus cool behavioural styles in electronic negotiations: the impact of communication mode. Group Decision and Negotiation 15(2), 141–155. | Article |
109. Pillutla MM and Murnighan JK (1996) Unfairness, anger, and spite: emotional rejections of ultimatum offers. Organizational Behavior and Human Decision Processes 68(3), 208–224. | Article |
110. Pinsonneault A and Heppel N (1998) Anonymity in group support systems research: a new conceptualization, measure and contingency. Journal of Management Information Systems 14(3), 89–108.
111. Podsakoff PM, MacKenzie SB, Lee JY and Podsakoff NP (2003) Common method bias in behavioral research: a critical review of the literature and recommended remedies. Journal of Applied Psychology 88(5), 879–903. | Article | PubMed |
112. Postmes T and Spears R (1998) Deindividuation and anti-normative behavior: a meta-analysis. Psychological Bulletin 123, 238–259. | Article |
113. Postmes T, Spears R, Sakhel K and De Groot D (2001) Social influence in computer-mediated groups: the effects of anonymity on social behavior. Personality and Social Psychology Bulletin 27, 1243–1254. | Article |
114. Pruitt D and Carnevale P (1993) Negotiation in Social Conflict. Open University Press, Buckingham.
115. Pruitt DG (1981) Negotiation Behavior. Academic Press, Harcourt Brace Jovanovich.
116. Pruitt DG and Carnevale PJ (1982) The development of integrative agreements in social conflict. In Living with Other People (DERLEGA VJ and GRZELAK J, Eds), pp 151–181, Academic Press, New York.
117. Putnam LL and Roloff ME (1992) Communication perspectives on negotiation. In Communication and Negotiation (PUTNAM LL and ROLOFF ME, Eds), pp 1–17, Sage, Newbury Park, CA.
118. Rangaswamy A and Shell RG (1997) Using computers to realize joint gains in negotiations: toward an electronic bargaining table. Management Science 43(8), 1147–1163.
119. Rapport A, Erev I and Zwick R (1995) An experimental study of buyer–seller negotiation with one-sided incomplete information. Management Science 41, 377–394.
120. Reinig B and Mejias R (2004) The effects of national culture and anonymity on flaming and criticalness in GSS-supported discussions. Small Group Research 35, 698–723. | Article |
121. Reinig BA, Briggs RO and Nunamaker JF (1998) Flaming in the electronic classroom. Journal of Management Information Systems 14(3), 45–60.
122. Rogers RW (1980) Expressions of aggresion: Agression-inhibiting effects of anonymity to authority and threatened retaliation. Personality and Social Psychology Bulletin 6, 315–320. | Article |
123. Rothbart M and Hallmark W (1988) Ingroup–outgroup differences in the perceived efficacy of coercion and conciliation in resolving social conflict. Journal of Personality and Social Psychology 55(2), 248–257. | Article | ChemPort |
124. Sally D (1995) Conversation and cooperation in social dilemmas: a meta analysis of experiments from 1958 to 1992. Rationality and Society 7(1), 58–92. | Article |
125. Sambamurthy V and Chin WW (1994) The effects of group attitudes toward alternative GDSS designs on the decision-making performance of computer-supported groups. Decision Sciences 25(2), 215. | Article |
126. Scherer KR (1986) Vocal affect expression: a review and a model for future research. Psychological Bulletin 99, 143–165. | Article | ChemPort |
127. Scherer KR (1999) Appraisal theory. In Handbook of Cognition and Emotion (DALGLEISH T and POWER M, Eds), pp 637–663, Wiley, Chichester London, New York.
128. Selfe CL and Meyer PR (1991) Testing claims for on-line conferences. Written Communication 8(2), 163–192. | Article |
129. Siegel J, Dubrovsky V, Kiesler S and McGuire TW (1986) Group processes in computer-mediated communication. Organizational Behavior and Human Decision Processes 37, 157–187. | Article |
130. Siegel S and Fouraker L (1960) Bargaining and Group Decision Making; Experiments in Bilateral Monopoly. McGraw-Hill, New York.
131. Sinaceur M and Neale M (2005) Not all threats are created equal: how implicitness and timing affect the effectiveness of threats in negotiations. Group Decision and Negotiation 14(1), 63–85. | Article |
132. Sinaceur M and Tiedens LZ (2005) Get mad and get more than even: when and why anger expression is effective in negotiations. Journal of Experimental Social Psychology 42, 314–322. | Article |
133. Smith DL, Pruitt DG and Carnevale PJ (1982) Matching and mismatching: the effect of own limit, other's toughness and time pressure on concession rate in negotiation. Journal of Personality and Social Psychology 42, 876–883. | Article |
134. Spangler B (2003) Distributive bargaining. In Beyond Intractability (BURGESS G and BURGESS H, Eds), Conflict Research Consortium, http://www.beyondintractability.org/essay/distributive_bargaining/.
135. Spears R and Lea M (1992) Paralanguage and social perception in computer-mediated communication. Journal of Organizational Computing 2, 321–341.
136. Sproull L and Kiesler S (1986) Reducing social context cues: electronic mail in organizational communication. Management Science 32(11), 1492–1512.
137. Sproull L and Kiesler S (1991) Connections: New Ways of Working in the Networked Organization. MIT Press, Cambridge, MA.
138. Steinel W, Van Kleef GA and Harinck F (2008) Are you Talking to Me?! Separating the people from the problem when expression emotions in negotiation. Journal of Experimental Social Psychology 44, 362–369. | Article |
139. Steiner I (2004) Industry profile: Steve Abernethy, Square Trade President and CEO. Available online at http://www.auctionbytes.com/cab/abu/y204/m05/abu0119/s04 (visited April 2007).
140. Stuhlmacher AF and Citera M (2005) Hostile behavior and profit in virtual negotiation: a meta-analysis. Journal of Business and Psychology 20(1), 69–93. | Article |
141. Stuhlmacher AF, Citera M and Willis T (2007) Gender differences in virtual negotiation: theory and research. Sex Roles 57(5–6), 329–339. | Article |
142. Stuhlmacher AF and Walters AE (1999) Gender differences in negotiation outcome: a meta-analysis. Personnel Psychology 52, 653–677. | Article |
143. Thompsen PA and Fougler DA (1996) Effects of pictographs and quoting on flaming in electronic mail. Computers in Human Behavior 12, 225–243. | Article |
144. Thompson L and Hrebec D (1996) Lose-lose agreements in interdependent decision making. Psychological Bulletin 120(3), 396–409. | Article |
145. Thompson L and Nadler J (2002) Negotiating via information technology: theory and application. Journal of Social Issues 58(1), 109–124. | Article |
146. Thompson LL, Nadler J and Kim PH (1999) Some like it hot: the case for the emotional negotiator. In Shared Cognition in Organizations: The Management of Knowledge (THOMPSON LL, LEVINE JM and DAVID MM, Eds), pp 139–161, Lawrence Erlbaum Associates, Mahwah, NJ.
147. Tice DM, Bratslavsky E and Baumeister RF (2001) Emotional distress regulation takes precedence over impulse control: if you feel bad, do it! Journal of Personality and Social Psychology 80(1), 53–67. | Article | ChemPort |
148. Tiger L and Fox R (1971) The Imperial Animal. Holt, Rinehart and Winston, New York.
149. Torestad B (1990) What is anger provoking: a psychophysical study of perceived causes of anger. Aggressive Behavior 16, 9–26. | Article |
150. Turnage AK (2007) Email flaming behaviors and organizational conflict. Journal of Computer-Mediated Communication 13(1), 43–59. | Article |
151. Tutzauer F (1992) The communication of offers in dyadic bargaining. In Communication and Negotiation (PUTNAM LL and ROLOFF ME, Eds), pp 67–82, Sage, Newbury Park, CA.
152. Tyler MC (2004) 115 and Counting: The State of ODR 2004. In Proceedings of the third Annual Forum on Online Dispute Resolution, Melbourne, Australia (available online at http://www.odr.info/unforum2004/ConleyTyler.htm).
153. Valley KL, Moag J and Bazerman MH (1998) A matter of trust': effects of communication on the efficiency and distribution of outcomes. Journal of Economic Behavior and Organization 34, 211–238. | Article |
154. Van Kleef GA, De Dreu CW and Manstead AR (2004a) The interpersonal effects of anger and happiness in negotiations. Journal of Personality and Social Psychology 86(1), 57–76. | Article |
155. Van Kleef GA, De Dreu CW and Manstead AR (2004b) The interpersonal effects of emotions in negotiations: a motivated information processing approach. Journal of Personality and Social Psychology 87(4), 510–528. | Article |
156. Van Kleef GA, van Dijk E, Steinel W, Harinck F and van Beest I (2008) Anger in social conflict: cross-situational comparisons and suggestions for the future. Group Decision and Negotiation 17, 13–30. | Article |
157. Veazie J (2007) Email etiquette. Health Care Biller 16(6), 7–9.
158. Whatley MA, Webster JM, Smith RH and Rhodes A (1999) The effect of a favor on public and private compliance: how internalized is the norm of reciprocity? Basic and Applied Social Psychology 21(3), 251–259. | Article |
159. Wold H (1985) Systems analysis by partial least squares. In Measuring the Unmeasurable (NIJKAMP P, LEITNER H and WRIGLEY N, Eds), Martinus Nijhoff, Dordrecht.
160. Yukl G (1974) Effects of the opponent's initial offer, concession magnitude, and concession frequency on bargaining behavior. Journal of Personality and Social Psychology 30(3), 323–335. | Article |
161. Yuuki K, Shogo K and Kanji A (2007) Effects of emotional cues transmitted in e-mail communication on the emotions experienced by senders and receivers. Computers in Human Behavior 23(4), 1894–1905. | Article |
162. Zartman IW (1977) Negotiation as a joint decision-making process. Journal of Conflict Resolution 21, 619–638. | Article |

Acknowledgements

We thank Blake Ives as well as the anonymous reviewers for their comments on earlier versions of this paper. We also thank Richard Holowczak and Bruce Weber for their contributions to the development of the negotiation task that was used in our study. We are grateful to Rebecca Biel and Brian Woods, along with Leah Duque, Vlad Krotov, Elham Mousavidin, Hesam Panhi, and Doug Steel for their efforts in coding our data.

About the authors

Norman A. Johson is an assistant professor in the Decision and Information Sciences Department in the Bauer College of Business at the University of Houston. He earned his masters and doctorate degrees from Baruch College, City University of New York. His current research interests include construct development, computer-mediated negotiation, communication and decision-making under uncertainty, affect and trust.

Randolph B. Cooper is a Professor in the Decision and Information Sciences Department in the Bauer College of Business at the University of Houston. He received his bachelors, masters, and doctorate degrees from the University of California at Los Angeles. He has published in a variety of academic journals, including DataBase, Journal of Management Information Systems, Management Science, MIS Quarterly, and Omega. His current research interests include the diffusion of information technology innovations, management of systems design creativity, and the impact of information systems on negotiation.

Wynne W. Chin is a Professor in the Decision and Information Sciences Department in the Bauer College of Business at the University of Houston. He received his doctorate from the University of Michigan in Computers and Information Systems, M.S. in Chemical Engineering (biomedical option) from Northwestern University, MBA from the University of Michigan, and a Bachelors in Biophysics from U.C. Berkeley. He has published in journals such as MIS Quarterly, Information Systems Research, and Decision Sciences. Dr. Chin's substantive interests include modeling the individual IT adoption process, end-user satisfaction, and developing group process measures such as cohesion, satisfaction, and consensus to understand the impact of electronic meeting systems. More recently, he has begun working on cross-cultural analysis. His research is largely empirical and quantitative relying on lab and Monte Carlo experiments as well as surveys. Methodologically Dr. Chin focuses on construct development through the use of structural equation modeling (both covariance-based and partial least squares) as well as developing new causal modeling techniques for topics such as assessing interaction effects.