Interaction Patterns in Crisis Negotiations: Persuasive Arguments and Cultural Differences

Interaction Patterns in Crisis Negotiations: Persuasive Arguments and Cultural Differences

Ellen Giebels University of Twente

Paul J. Taylor Lancaster University

This research examines cultural differences in negotiators’ responses to persuasive arguments in crisis (hostage) negotiations over time. Using a new method of examining cue–response patterns, the authors examined 25 crisis negotiations in which police negotiators interacted with perpetrators from low-context (LC) or high-context (HC) cultures. Compared with HC perpetrators, LC perpetrators were found to use more persuasive arguments, to reciprocate persuasive arguments in the second half of negotiations, and to respond to persuasive arguments in a compromising way. Further analyses found that LC perpetrators were more likely to communicate threats, especially in the first half of the negotiations, but that HC perpetrators were more likely to reciprocate them. The implications of these findings for our under- standing of intercultural interaction are discussed.

Keywords: hostage negotiation, cultural differences, influence tactics, proximity coefficient

On August 12, 2002, a Dutch volunteer named Arjan Erkel was kidnapped while working on the Médecins sans Frontières medical aid program in Dagestan, a republic of the Russian federation. The response of the authorities to this incident was to engage the perpetrators in dialogue and persuade them to release, and not physically harm, the volunteer. The subsequent negotiations in- volved multiple threats to the volunteer’s life and resulted in 607 days of captivity before the volunteer was released (Hargreaves & Cunningham, 2004). One factor that undoubtedly complicated the negotiations was the misconceptions that arose from the different cultural backgrounds of those involved. For example, in his review of the negotiations, Van Zwol (2005) shows not only how the Russian perpetrators were rather circuitous in their approach, but also how they anticipated indirect messages from the Dutch nego- tiators, who were in contrast more explicit and upfront in their communication. Understanding how such high-stakes negotiations are influenced by cultural differences can help authorities to man- age hostage incidents more effectively in an increasingly violent and globalizing world.

Police forces in the United States and Europe have reported a significant growth in the cultural diversity of the perpetrators of kidnappings and extortions (Giebels, 1999; Ostermann, 2002; Tay-

lor & Donohue, 2006). This trend reasserts the need for scholars to identify cultural differences in approaches to interaction, as well as understand how violent individuals from these cultures are likely to react to efforts to dissuade them from causing harm. Such an understanding has applications in a range of settings where factors such as high stakes and distrust can make individuals suspicious of behavior that diverges from their culturally driven expectations (Ember, 1988). In the crisis negotiation context, a more sophisti- cated understanding of cross-cultural communication will help police formulate culturally sensitive negotiation strategies and enhance their appreciation of why perpetrators react the way they do (Yang, Wu, & Huang, 2007). For example, understanding the different ways in which cultures respond to and use persuasive arguments will enable a police team to frame negotiation strategies in a way that works with, rather than against, a perpetrator’s interpersonal approach. In this article, we respond to these possi- bilities by examining the negotiation dynamics of 25 actual crisis negotiations perpetrated by individuals from different cultural backgrounds.

Culture is often defined as the characteristic profile of a society with respect to its values, norms, and institutions (Lytle, Brett, Barsness, Tinsley, & Janssens, 1995). It is an important determi- nant of people’s attitudes, self-construal, and behavior and hence their strategic choices in conflict situations (cf. Pruitt & Kim, 2004). Efforts to understand the impact of intercultural dynamics on negotiation have generally taken one of two perspectives. One line of research has sought to understand how the outcomes of conflicts are affected by cultural factors such as social identities and cultural stereotypes (Faure, 2002, 2003; Gelfand et al., 2001; Macduff, 2006; Tjosvold & Sun, 2000). This research has re- vealed, for example, how divergent framing of messages and poor understanding of cultural values can lead to nonoptimal outcomes and escalations of conflict (Hammer & Rogan, 2002). A second line of research has considered how negotiators respond to key negotiation behaviors, such as information sharing and pressure tactics (Adair & Brett, 2005; Adair, Okumura, & Brett, 2001;

Ellen Giebels, Department of Organizational Psychology and Human Resource Development, University of Twente, Enschede, The Netherlands; Paul J. Taylor, Department of Psychology, Lancaster University, Lan- caster, United Kingdom.

Part of this article was prepared while Ellen Giebels was at the Univer- sity of Groningen, Groningen, The Netherlands. This study was supported by British Council–NWO Partnership Programme in Science Grant PS844. We would like to thank Sigrid Noelanders and Alette Zijlstra for their help in coding the material.

Correspondence concerning this article should be addressed to Ellen Giebels, University of Twente, Faculty of Behavioral Sciences, Organiza- tional Psychology and HRD, P.O. Box 217, 7500 AE Enschede, The Netherlands. E-mail:

Journal of Applied Psychology © 2009 American Psychological Association 2009, Vol. 94, No. 1, 5–19 0021-9010/09/$12.00 DOI: 10.1037/a0012953


Tinsley, 1998, 2001). These studies examine the way in which negotiators use culturally normative negotiation behaviors and how their use is linked to the utilization or otherwise of integrative potential. By focusing on the actual building blocks of negotiation, these studies are able to make a significant contribution to our understanding of the communicative process that underlies and structures negotiation.

In this article, we take the second of these two approaches and build on the existing literature in two ways. First, we extend the focus of previous studies, which has typically been on the use of integrative bargaining strategies (e.g., information sharing), by emphasizing the role of distributive forcing strategies. Forcing strategies are those strategies of influence (French & Raven, 1959) by which a negotiator tries to satisfy individual goals and gain advantage over the other party (Deutsch, 1974; Lewicki, Saunders, & Minton, 1997). Forcing strategies may be particularly prevalent in high-stakes, crisis scenarios because they are not bound by the normative framework of cooperation and role obligation that de- fine day-to-day negotiations (Roger & Schumacher, 1983). More- over, crisis negotiations are typically win–lose in structure and usually contain little integrative potential. In line with this reason- ing, research by Donohue and Roberto (1993) shows that negoti- ators in crisis situations focus on power issues such as “who is in charge” and “how can I force the other,” rather than on substantive issues such as identifying win–win solutions.

In conceptualizing forcing strategies, research distinguishes be- tween two main categories of forcing behavior: persuasive argu- ments and threats. Persuasive arguments aim to convince the other party to comply with one’s own proposals by using task-related arguments and logic (Giebels, De Dreu, & Van de Vliert, 2003). In contrast, threats communicate the intention to punish the other party if it does not concede (Deutsch & Krauss, 1962). Research addressing the use of forcing behavior in negotiation traditionally focuses on the communication of threats (Deutsch & Krauss, 1962; Tedeschi, Schlenker, & Bonoma, 1973). More recently, research has also taken into account persuasive arguments (e.g., Adair & Brett, 2005; Adair et al., 2001), although little research has con- sidered both strategies simultaneously or differentiated persuasive arguments from the communication of threats. This is surprising, since there is evidence to suggest that both types of forcing behavior are different in their effects. For example, two experi- ments by Giebels et al. (2003) demonstrate that threats and per- suasive arguments are not correlated and that a negotiator’s social value orientation influences the use of threats but not persuasive arguments. Another study by Giebels, De Dreu, and Van de Vliert (1998) shows that negotiators with a power advantage communi- cate more threats than do negotiators with a power disadvantage, while the opposite is true for persuasive arguments. In this study, therefore, we differentiate persuasive arguments from the commu- nication of threats. We also focus our hypothesis development on persuasive arguments, since this strategy is the less examined of the two and one that is arguably important to cross-cultural crisis negotiations. Specifically, while the precise effects of persuasive arguments on negotiation have yet to be explored, many negotia- tion handbooks advise negotiators to develop and use arguments to support their claims, based on the notion that persuasion may change the opponent’s understanding of reality, enhance the prob- ability that one’s claim is seriously considered, and reduce the chance that the other person feels personally attacked (e.g., Fisher,

Ury, & Patton, 1991; Lewicki, Hiam, & Wise Olander, 1996; Neale & Bazerman, 1991; cf. Harinck, 2004). This is a concern since evidence suggests that the salience of such rational strategies may be dependent on culture (Drake, 1995; Fu & Yukl, 2000; Grice, 1975), which in turn opens up the possibility that they may be more or less effective as a strategy in different cultures. More- over, this dependence on culture may be particularly prevalent in negotiations with win–lose structures and a focus on individual gains (Adair & Brett, 2004), which are both factors that charac- terize crisis negotiations. Thus, given this evidence, it appears particularly important to better understand the role of persuasive arguments in cross-cultural negotiations.

Second, we focus our examination of cultural differences at the micro level of cue–response patterns. This focus represents a significant departure from previous research, which has typically examined cultural differences in terms of the frequency of behav- ior use across several time periods. A focus on the interrelation- ships among behaviors allows our analysis to consider how culture impacts on the way in which negotiators interpret and respond to a message or series of messages. Indeed, there is growing evidence to suggest that the order in which behaviors occur—the interrela- tionships among behaviors—has a significant impact on their meaning and influence on the negotiation (Adair, 2003; Adair & Brett, 2005; Giebels & Noelanders, 2004; Taylor, 2002a; Taylor & Donald, 2003, 2004). Interestingly, these impacts are not always the result of conditional responses; they are often the result of delayed responses to a key behavior or responses to a particular pattern of previous behaviors (Taylor & Donald, 2007). Thus, understanding the actual makeup of interaction is important for negotiation theory because it provides insight into the process by which various cues and responses come together to allow a conflict to begin, unfold, and be resolved. It is also important to an applied psychology of negotiation, since crisis negotiators can make best use of strategies that work to influence a perpetrator at any given moment, rather than in an aggregate fashion.

Cultural Dimensions and Negotiation Behaviors

The preceding section presented the foundation of our approach to examining negotiation behavior and, in particular, persuasive argu- ments. In this section, we consider the existing literature on cross- cultural interaction and draw hypotheses about the types of cue– response contingencies that will dominate the behavior of perpetrators from different cultural backgrounds. Because culture influences the way in which people communicate, and our study focuses on the communicative dynamics of crisis negotiations, we use Hall’s (1976) theory of low-context (LC)/high-context (HC) communication to build our hypotheses. According to Hall, LC communication involves the use of explicit and direct messages in which meanings are prin- cipally contained in the transmitted messages. In contrast, HC com- munication is characterized by messages in which information is more hidden and meaning is located in the social or physical context of the negotiation. Hall argues that people in a culture use both LC and HC communication but that one tends to be dominant (cf. Gudykunst & Matsumoto, 1996). Generally, LC communication is predominant in individualistic cultures while HC communication is found to be pre- dominant in collectivistic cultures (Gudykunst & Ting-Toomey, 1988; Hofstede, 2001; Triandis & Suh, 2002). For example, Triandis (1994, pp. 184–185) states that “while people in collectivistic cultures


pay more attention to context . . . in individualistic cultures people speak clearly . . . and it is possible to discuss different viewpoints objectively.” Ting-Toomey and Oetzel (2001, p. 31) explain this strong link, particularly in relation to conflict management, by arguing that for cultures that emphasize “I-identity” and self-initiative, the ideal way of resolving conflict is to talk and discuss it directly. In contrast, in cultures that emphasize values of relational harmony, the ideal way of conflict management is to talk around the point and not deal directly with the issues at stake. This means that the majority of individualistic, Western societies, especially in northern Europe and the United States, can be considered LC cultures, while the majority of collectivistic, non-Western societies, such as China and Russia, may be considered HC cultures (Hofstede, 2001, p. 212; see also Adair, 2003; Adair & Brett, 2005).

Use of Persuasive Arguments

Hall (1976) identified fundamental differences in the way peo- ple from different cultures communicate and labeled these differ- ences as “low context” or “high context,” referring to the impor- tance of the explicit content of the message versus the context. This distinction suggests that negotiators from LC cultures, in comparison to their HC counterparts, will focus on sending and receiving accurate messages and defining the interaction princi- pally in terms of message content (i.e., LC communication). In contrast, those from HC cultures will focus on sending and receiv- ing messages in which information is more hidden and meaning is located in the social or physical context of the negotiation (i.e., HC communication). These notions are consistent with previous re- search showing that LC U.S. negotiators are more likely than HC Japanese negotiators to exchange information directly (Adair et al., 2001; Graham, 1993).

However, it may also have consequences for different types of forcing behaviors. An important assumption that is characteristic of LC communication is what Grice (1975; see also Gudykunst & Matsumoto, 1996) labeled the quality maxim: One should state only that which is believed to be true with sufficient evidence. This maxim implies that LC communication is typically centered around logic and rationality, compared with HC communication (cf. Adair & Brett, 2004). In a similar vein, Triandis (1994) suggests that LC but not HC cultures focus on explicit logic and proof. This implication is consistent with research in a number of areas. For example, Ting-Toomey’s work on cross-cultural com- munication (e.g., Ting-Toomey, 1988) suggests that confronting the other party with rational arguments and factual evidence is more central to American than to Chinese conflict environments (see also Fu & Yukl, 2000). Similarly, research has shown that U.S. negotiators use more analytic statements than do Taiwanese negotiators in a simulated negotiation scenario (Drake, 1995), and that European Canadians rely on more arguments when justifying their choices than do Asian Canadians (Hoshino-Browne et al., 2005). Combined, this evidence leads us to hypothesize that

Hypothesis 1: Low-context perpetrators make more use of persuasive arguments than do high-context perpetrators.

Reciprocity of Persuasive Arguments

Related to the possibility of identifying differences in the use of persuasive arguments is the possibility of identifying differ-

ences in the degree to which the other party reciprocates per- suasive arguments. Consistent with previous research, we use the term reciprocation for patterns of behavior in which nego- tiators respond to each other’s strategies by matching them with the same behavior (Adair & Brett, 2005; Brett, Shapiro, & Lytle, 1998; Friedman, Anderson, Brett, Olekalns, Goates, & Lisco, 2004; Olekalns & Smith, 2000; Weingart, Prietula, Hy- der, & Genovese, 1999; Weingart, Thompson, Bazerman, & Caroll, 1990). The possibility of observing cross-cultural dif- ferences in reciprocation is significant, since previous research has shown that reciprocation of communications is common in dyadic interactions, including negotiations (e.g., Axelrod, 1984; Gouldner, 1960; Putnam & Jones, 1982). Yet, if logic and rationality are considered core elements in LC communication, then reciprocation of rational arguments should occur more frequently in negotiations with LC perpetrators compared with negotiations with HC perpetrators. This proposal is certainly consistent with more general research on message strategies and cross-cultural differences in preference and effectiveness (Hong, Muderrisoglu, & Zinkham, 1987; Tai, 2004). For exam- ple, two comparisons of LC and HC negotiators by Adair (Adair, 2003; Adair & Brett, 2005) found differences in the reciprocation of information exchange that matched the ele- ments that Hall (1976) associates with the two cultural groups. The Adair studies focus on information exchange because they consider an integrative negotiation task, where information sharing is likely central to good outcomes. However, parallel differences might be expected in win–lose negotiations, where influence strategies such as persuasive arguments may be cen- tral to success (cf. Adair & Brett, 2004).

Other evidence suggests that the propensity to reciprocate with rational arguments might be dependent on the negotiation phase. Research suggests that unlike in more normative contexts, where deadlines and ultimatums can lead to arousal in later stages of negotiations, crisis negotiations involve the most stress and arousal in the initial stages of interaction (Donohue, Ramesh, Kaufmann, & Smith, 1991; Taylor & Donald, 2007; Vecchi, Van Hasselt, & Romano, 2004). This occurs because of the intense emotional and physical reaction that perpetrators have to the overwhelming and threatening nature of law enforcement involvement (Vecchi et al., 2004). Such arousal is likely to inhibit a negotiator’s capacity to process information (De Dreu & Weingart, 2003). However, this capacity is arguably critical if he or she is to understand the other party’s reasoning, evaluate their arguments (i.e., identify weak spots in the other’s argumentation), and construct reasoned re- sponses to further his or her own goals. The inhibiting effect of arousal is therefore likely to limit the ability of negotiators to use and respond to persuasion during the early stages of negotiation. In contrast, the efforts of police negotiators to reduce tension and create a problem-solving environment (Donohue et al., 1991) will mean that later stages of negotiation involve less arousal and, consequently, are relatively more open to persuasive arguments playing a more central role. Thus, we hypothesize that

Hypothesis 2: The reciprocation of persuasive arguments will be higher in negotiations with low-context as opposed to high-context perpetrators, particularly in the second part of the negotiations.


Efficiency of Persuasive Arguments

Finally, persuasive arguments may be regarded as having sig- nificant parallels with influence tactics, which we define as delib- erate actions by one individual (the influence agent) toward an- other individual (the target) that are intended to alter his or her attitudes in a way that would not have otherwise occurred (Perloff, 1993). This suggests that the effectiveness of influence strategies may be measured as the extent to which they alter the other party’s behavior in the desired direction. If persuasive arguments are considered more central to interaction in LC rather than HC cultures, then this strategy is also likely to be more influential in LC rather than HC cultures. This expectation is consistent with the results of a scenario study by Fu and Yukl (2000), which found that LC U.S. managers perceive persuasive arguments as more effective in influencing people and resolving differences than do HC Chinese managers. It is also consistent with wider research on the commitment and consistency principle, defined as the human desire to achieve apparent consistency between previous beliefs or behavior and current attitudes or behavior (Festinger, 1957; Heine & Lehman, 1997). As such, the commitment and consistency principle appeals to logic and rationality and tends to be developed in “if–then” linear terms, both of which are associated with LC cultures (cf. Adair & Brett, 2004). Thus, while the consistency principle is robust in its influence on behavior, research (Cialdini, Wosinska, Barrett, Butner, & Gornik-Durose, 1999) has found that LC U.S. students are more likely to feel compelled to act in a way that is consistent with previous behavior than are HC Polish students. Similarly, Choi and Nisbett (2000) found that East Asians have a higher tolerance for contradictions and react less to cognitive inconsistencies than do their American counterparts. This collection of research suggests that influence exercised through the use of cognitive, rational strategies is likely to be more effective in negotiations with LC than HC perpetrators. Further- more, and in line with our previous reasoning, we propose that an important prerequisite for cognitive strategies to influence an interaction is a context in which individuals have the cognitive capacity to process them. This is less likely in the early stages of crisis negotiation than in the later stages. Accordingly, we hypoth- esize that

Hypothesis 3: Persuasive arguments by the police negotiator will be more effective in negotiations with low-context per- petrators rather than with high-context perpetrators, particu- larly in the second part of the negotiations.


Negotiation Sample

Data were transcripts produced from audiotapes of interactions from 25 crisis negotiations that began in the Netherlands or Bel- gium and were therefore available in Dutch or Belgian police files. These cases were selected on the basis that they took place over the last 10 years and that they included at least 10 min of conversation between the negotiator and perpetrator (M � 39 min). The tran- script for each case contained all of the conversation that was recorded during the incident. This was essentially the complete negotiation minus possible lead in time during which the police were still connecting the recording devices. Of the 25 cases, 15

involved kidnapping and 10 involved extortion. All 25 cases concerned the negotiation of instrumental issues. Specifically, in all 25 incidents the perpetrators demanded money or other valu- able items. All incidents were conducted by phone and ended with the perpetrators being arrested. None of the incidents involved simultaneous negotiation with a third party. All police negotiators and perpetrators were male.

Twelve of the negotiations were with hostage takers from soci- eties that can be regarded as LC: The Netherlands (3), Belgium (6), and the United Kingdom (3; cf. Brett, 2001; Hall, 1976; Hall & Hall, 1990; Hofstede, 2001; Triandis, 1994). Of these 12 cases, 6 were kidnappings and 6 were extortions. The remaining 13 nego- tiations concerned perpetrators from societies that can be regarded as HC: China (4), Kurdistan (1), Morocco (2), Suriname (1), Russia (2), Turkey (1), and Poland (2; cf. Adair, 2003; Adair & Brett, 2005; Brett, 2001; Fu & Yukl, 2000; Hall, 1976; Hall & Hall, 1990; Hofstede, 2001; Kozan & Ilter, 1994; Triandis, 1994). Of these 13 cases, 9 were kidnappings and 4 were extortions. In all cases, the police files showed that the non-Dutch and non-Belgian perpetrators were born and lived the largest part of their lives in their country of origin.

Our distinction, based on the notion that HC communication is typical for collectivist societies and LC communication is typical for individualistic societies, is consistent with Hofstede’s (2001) collectivism–individualism country index score. This index con- trasts the extent to which individuals are autonomous individuals versus embedded in groups and is validated against a large amount of country data from other sources (e.g., Schwartz, 1992; Smith, Trompenaars, & Dugan, 1995). On the index, the HC countries in our study are associated with index scores ranging from 20 to 60 (M � 37.8), while the LC countries are associated with index scores ranging from 75 to 89 (M � 79.8). Thus, there is no overlap in the index scores associated with cases in the LC and HC groups, supporting our assertion that they represent groups that differ on Hall’s (1976) HC/LC distinction. Furthermore, even for Poland, the country in the HC group with the highest index score of 60, research indicates that individual behavioral reactions are charac- terized by collectivistic rather than individualistic tendencies (Cialdini et al., 1999; see also Reykowski, 1998). Moreover, our categorization is supported by considerable previous research (e.g., Adair, 2003; Adair & Brett, 2005; Fu & Yukl, 2000; Kozan & Ilter, 1994; Vrij, Dragt, & Kippelaar, 1992).

In 19 cases, the negotiation occurred in a language other than Dutch. To facilitate analysis, we had the audiotaped dialogue in these interactions translated into Dutch by professional translators who were associated with the police but blind to the research hypotheses. These translators were bilingual, native speakers of the language in which the negotiation was conducted. Because of the sensitive nature of this material, it was not possible to use back-translation to evaluate the reliability of the transcription. However, a number of factors serve to mitigate the possibility of mistaken translation. First, the translator facilitated interpretation of the transcripts by providing “advisory notes” in which she or he indicated culture-specific meanings of particular words or phrases. Second, in their role with the police, the translators had prepared a large amount of material for legal proceedings (e.g., court cases), such that they were experienced in providing translations with high levels of accuracy. Third, as described in the next section, the dialogue was examined at the level of speaking turn and not in


terms of the use of individual words. An analysis pitched at this level arguably minimizes the impact that single word ambiguities will have on the analysis (cf. Taylor, 2002b).

Coding the Transcripts

Using the transcriptions of the audiotapes of the 25 negotiations, two trained judges (unaware of the hypotheses) coded all the speaking turns as 1 of 11 codes. These codes, derived in previous research (Giebels & Noelanders, 2004), capture the use of major influence tactics that occur in crisis negotiations. Three of the codes (“being kind,” “being equal,” and “being credible”) are primarily connected with the sender and his or her relationship with the other party. For example, the code “being equal” is used to capture an utterance that aims to highlight something that the parties have in common (e.g., “You and I, we will work this thing out together”). The remaining eight codes (“compromise,” “direct pressure,” “emotional appeal,” “imposing a restriction,” “informa- tion sharing,” “legitimizing,” “persuasive arguments,” and “threats”) are primarily connected with the content of the message and the information conveyed to the other party. For example, the code “emotional appeal” is used to measure occasions when a negotiator plays on the emotions of the other party (e.g., “Please take good care of her”). In contrast, the code “legitimizing” is used to measure occasions when a negotiator refers to what has been agreed upon in society or with the other party (e.g., “I’m sorry but I have to conform to regular bank procedures”). Finally, note that the code “information sharing” is unique in that it does not nec- essarily involve an attempt at influence but captures behaviors such as negotiators discussing priorities, comparing or contrasting positions, or acknowledging the other’s message. A full descrip- tion of this coding scheme and its derivation is given in Giebels and Noelanders (2004).

The judges were a 29-year-old woman from the Netherlands holding an MA in psychology and a 31-year-old woman from Belgium holding an MA in criminology. Both were specially recruited to work on a crisis negotiation research project (see Giebels & Noelanders, 2004) because they had experience with content-coding of cross-cultural interactions. They were trained to use this coding scheme on unrelated negotiation material. This training material included interactions that were translated by the same type of professional translator who translated the 25 negoti- ation transcripts. This ensured that judges gained exposure not only to the dialogue of HC negotiators, but also to the advisory notes made by translators to aid the interpretation of the translation (the negotiation data also included these notes). Thus, the training sought to ensure that the judges could not only apply the codes properly but also incorporate into their coding the transcribers’ notes on how to read the messages by HC negotiators. For exam- ple, the Chinese statement “then we did ‘it’” was interpreted as “in that case, we will kill (him)” and therefore coded as an explicit threat. This training continued until interrater reliability, as mea- sured by Cohen’s kappa, reached .75. This required approximately 30 hr of training.

The judges then coded the 25 negotiations by assigning each speaking turn of the police negotiator and perpetrator to one of the coding categories. As this assignment was just a placement of each speaking turn into a coding category, we did not attempt to assess the degree to which a speaking turn reflected an influence tactic.

This approach, which is in line with previous research (e.g., Donohue & Roberto, 1993; Weingart et al., 1999), allowed the analysis to focus on the transitions across speaking turns. The interrater reliability of this coding was very good, with an average kappa of .85 and a range across the incidents of .70 to .921

(Bakeman & Gottman, 1997). In cases of disagreement, the raters discussed the particular speaking turn and decided jointly which code was the best to apply.

For the purposes of this research, we focus on the use of four categories of behavior. The focus of our analysis was on patterns in negotiators’ use of the behavior “persuasive arguments,” coded as message behaviors that used rational persuasion or logic to assert a point of view or idea. Examples of persuasive arguments include “You just said you have the money available, so why can’t you bring it to me right away?” (perpetrator) and “I don’t think it is a good idea to let her parents drive the car because they are really stressed” (police negotiator). To test our assumption about the difference between persuasive arguments and threats, and to en- able comparisons with previous research (e.g., Adair & Brett, 2005; De Dreu, Giebels, & Van de Vliert, 1998; Giebels, De Dreu, & Van de Vliert, 2000; Giebels et al., 2003; Van de Vliert, Nauta, Giebels, & Janssen, 1999), we also examined negotiators’ use of the behaviors “threats” and “information sharing.” We defined threats as message behaviors that intimidate, accuse, or suggest punishment for not acting a particular way. Their inclusion in the analysis made it possible to test our assumption that persuasive arguments and threats are used by and influence negotiators in different ways. We defined information sharing as message be- haviors that provide information about procedural, emotional, or substantive issues. Their inclusion in the analysis, prompted by the fact that information sharing is the most frequently observed behavior in cross-cultural normative interactions (Adair & Brett, 2005; Adair et al, 2001), enabled a comparison between crisis negotiations and normative interactions, and allowed us to gauge the extent to which variation in information sharing might provide an alternative explanation for any observed difference in the de- gree to which negotiators compromised. Examples of threat mes- sage behaviors from the data include “I will kill [the hostage] if I do not get the money soon” (perpetrator) and “When you are hurting [the hostage] I cannot help you anymore” (police negoti- ator). Examples of information sharing message behaviors from the data include “There is no one else here at the moment” (perpetrator) and “I have no preference for any of the places you mention, but the exact timing is more important to me” (police negotiator).

Finally, to provide an indication of police negotiator efficiency (see below), we examined negotiators’ use of “compromise,” de- fined as an act of concession or engagement in give-and-take behaviors. To be coded as a “compromise” behavior, a speech turn had to indicate that the negotiator was taking action that would benefit the other party or the hostages. This may be a concession in terms of getting something to the hostages (e.g., a message or medicine), agreeing on providing proof of life, improving the physical circumstances of the hostages, or even releasing a hostage. Similarly,

1 The individual Cohen’s kappa for the four behaviors that were the focus of our analysis were .78 for “persuasive arguments,” .72 for “threats,” .75 for “information sharing,” and .87 for “compromising.”


police negotiators may concede on issues of importance to the perpetrator, often in relation to money and the terms under which this money is provided. Examples of compromising behaviors from the data include “Ok, I will give ‘the victim’ his medicine in time” (perpetrator) or “Ok, we will have the money available before next Friday” (police negotiator).

Police negotiator efficiency. The efficiency of police negoti- ators’ behavior was assessed by establishing the extent to which they were successful in eliciting compromising behavior from the perpetrators. This measure recognizes that police negotiators pur- sue goals that are more immediate and recurrent than are captured by the final outcome of a negotiation. Arguably, the main goal for police negotiators is ensuring the physical and psychological safety of the hostage(s), which is clearly promoted by the types of perpetrator concession making mentioned above (cf. Giebels, No- elanders, & Vervaeke, 2005). Furthermore, an important goal for police negotiators is to turn the negotiation into a more normative interaction characterized by mutual concession making. It is also consistent with research examining the behavioral correlates of successful win–lose negotiations (De Dreu, 1995; Hornstein, 1965; Michener, Vaske, Schleifer, Plazewski, & Chapman, 1975).

The decision to focus on the occurrence of compromising be- havior rather than overall negotiation outcome is consistent with our focus on cue–response dynamics. If we had used overall negotiation outcome as a measure of efficiency, it would be difficult to identify which of the persuasion strategies led to compromising behavior. For example, in the early stages of an incident a police negotiator may be effective at influencing a perpetrator to release the hostages, but in the later stages when the perpetrator becomes suicidal, the police negotiator may use an influence approach that is ineffective at dissuading the perpetrator from taking his life. In an analysis that uses final outcome as the measure of effectiveness, all of the police negotiators’ attempts at influence would likely be associated with the final unsuccessful act of the perpetrator taking his life. Such an analysis would not simply overlook the influence strategies that contributed to the successful release of the hostages, but it would inappropriately associate these strategies with an unsuccessful outcome. All ne- gotiations involve periods of persuasion that move the interaction toward and away from success, and it is impossible to distinguish how each of these patterns contribute to the unfolding negotiation without using a measure that captures the “outcomes” associated with different parts of the same interaction. Such a measure is provided by compromise behavior.

Time periods. To test whether or not the impact of influence behaviors is systematically related to time, we had to divide each negotiation into a series of interactions. To do this, we partitioned each negotiation into two equal-size periods, an approach consis- tent with previous research (e.g., Olekalns & Smith, 2000). This allowed for a comparison of the early period of interaction, often considered to be dominated by crisis and extreme reactions, from the later, usually more normative period of interaction (Donohue et al., 1991).

Analyzing Cue–Response Sequences

To examine the interrelationships among the four negotiation behaviors, we constructed event sequences (Bakeman & Gottman, 1997) from the coded transcripts. Specifically, for each transcript,

the series of assigned codes were used to create a single sequence in which one code appeared on one line of a data file. This sequence represented the codes across the complete negotiation transcript. Because the coding was based at the level of utterance, the resulting file contained a sequence in which the codes alter- nated between representing the police negotiator’s utterances and the hostage taker’s utterances.

There are currently a number of methods available to research- ers interested in examining the extent to which particular cues (e.g., police negotiator behaviors) lead to particular responses (e.g., perpetrator responses). One common approach is to conduct a log-linear analysis that examines the likelihoods of different responses occurring after a cue (Olekalns & Smith, 2000). How- ever, as Giebels and Noelanders (2004) note, the predominant focus of these analyses is the immediate relationship between behaviors (i.e., whether a behavior at time t impacts on behavior at time t � 1). This focus has the potential of overlooking important relationships between cues and responses that are delayed for a range of interpersonal reasons, including cultural dynamics. Even in advanced analyses where delayed responses are considered (Adair, 2003; Adair & Brett, 2005), the method employed restricts the analysis so that only one type of delay or “lag” (e.g., t � 3) is considered. By not considering the full range of intervals in these analyses, one loses valuable information about the range and type of interrelationships among behaviors. To overcome this problem, we examine the interrelationships among behaviors by using prox- imity coefficients (Taylor, 2006). Rather than consider the imme- diate relations among behaviors, the proximity approach considers the relationships among all behaviors as degrees of proximity. This captures more of the complex interconnections among behaviors in an interaction. As recent publications have shown, this allows for sophisticated comparisons across speakers and between single cases (Taylor, 2006; Taylor & Donald, 2007).

The proximity approach is based on the notion that behaviors contribute to the same part of the interaction and have more in common—in terms of the speaker’s motivating concerns, strate- gies, and cognitions—when they occur close together within an interaction than when they occur far apart. The importance of this behavioral “inter-connectedness” (Auld & White, 1959, p. 100) has long been recognized in theory and research. For example, in the social interaction literature, proximity underlies the notion of channeling, whereby every exchange of messages is seen as nar- rowing down the probability that other categories of talk will occur (Argyle, 1969; Kelly, 1997; Putnam, 1985). Those behaviors that occur immediately prior to current behavior may be shown to have the most influence on the direction of interaction development (Gottman, Markman, & Notarius, 1977; Thomas, 1985; Weingart et al., 1999). But those farther back in history may also have their own unique “lagged” association with current behavior (Olekalns & Smith, 2000; Taylor & Donald, 2003). A police negotiator’s persuasive argument may lead to an immediate concession by the perpetrator, but it may also lead to a concession following a number of intermittent behaviors. The difference between these two circumstances, according to the proximity approach, is that the intermittent behaviors have a bearing on the perpetrator’s conces- sion behavior. The effect of these intermittent behaviors is what Watzlawick, Beavin, and Jackson (1968, p. 131) describe as “lim- itation.” Each intermittent message reduces the number of possible next moves and is ultimately most predictive of the perpetrator’s


response. This dynamic is reflected conceptually in the proximity approach by associating less immediate, lagged relationships with lower proximity (see Taylor, 2006; Taylor & Donald, 2007, for more details).

This notion of proximity among two behaviors is measured with a proximity coefficient (Taylor, 2006). This coefficient, which varies between .00 and 1.00, expresses the relationship between two types of behavior as a direct function of their relative place- ments in a sequence. The coefficient equals .00 if the behaviors occur only once at the first and last position of the entire sequence, and it equals 1.00 if one behavior immediately precedes the second behavior without exception. A coefficient score between these two extremes reflects differing amounts of proximity between the two behaviors. A greater value indicates a more proximal relationship (i.e., fewer intermediate behaviors), while a lower value indicates a less proximal relationship (i.e., more intermediate behaviors and a significantly lagged relationship). Thus, the impact of various negotiation behaviors may be assessed by examining the relative proximity of particular responses to a speaker’s cue. A signifi- cantly high proximity indicates a strong relationship between the occurrence of a particular behavioral cue and the desired response.

There are a number of ways to assess the significance of proximity as measured by the coefficients (Taylor & Donald, 2007). Since our hypotheses require comparisons of coefficients derived from negotiations with LC and HC perpetrators, we adopt Taylor’s (2006) suggestion of using randomization tests. A ran- domization test (sometimes called a permutation test) is a statisti- cal significance test that compares the resulting test statistic (e.g., an F value) not to a traditional theoretical distribution but to a distribution obtained from resampling of the available data (see Edgington, 1995; Good, 1994; Ninness, Rumph, Vasquez, & Brad- field, 2002). The advantage of this approach is that it avoids many of the data assumptions of traditional significance tests (Dunlap, Burke, & Smith-Crowe, 2003; Switzer, Paese, & Fritz, 1992) and so provides a reliable way to derive significance levels for the proximity coefficients.

A randomization test begins by making a traditional statistical comparison of the dependent variable across two or more condi- tions (e.g., an ANOVA comparison). Then, rather than compare the resulting statistic to a table of critical values, the test derives a sampling distribution from the available data and examines the extent to which the test statistic falls toward the tail of this distribution. As with other statistical tests, the more the resulting statistic falls into the tail of the sampling distribution, the less likely it is that the difference between conditions occurred by chance. The sampling distribution used during a randomization test is derived through repeated random sampling of the available data. Specifically, after making the initial comparison across conditions, a randomization test shuffles the observed values (e.g., proximity coefficients) randomly between the conditions, calculates (from the newly formed groups) the same test statistic as computed for the original comparison (e.g., F), and repeats this process a large number of times (in our case 10,000 times). The result of this repetition is a set of test statistic values that form a distribution of the values likely to be observed for the statistic. From this distri- bution, it is possible to estimate the likelihood of observing a value of the test statistic that equals or exceeds the value actually observed in the data (i.e., in what proportion of the randomized data sets did the derived test statistic exceed the statistic obtained

from the original data?). This likelihood provides the equivalent of a p value, which may be understood in the same way as p values derived through traditional parametric tests (e.g., ANOVA).

To further illustrate the basic idea of a randomization test, suppose we are comparing HC and LC negotiations on one cue– response contingency, the reciprocation of persuasive arguments. Further, suppose the observed mean coefficients are .85 for the HC group and .96 for the LC group, and that we want to test whether the difference between these sample means (i.e., .11) is large enough to reject the null hypothesis of no difference in proximity. Given nHC and nLC as the sample size of each group, the test proceeds as follows: (1) the difference between the two samples is measured with a familiar test statistic (e.g., F); (2) the observations of the HC and LC groups are pooled; (3) from these pooled values, nHC observations are sampled at random without replacement; (4) this new group of nHC observations is compared against the re- maining nLC observations with the same statistic as used in Step 1 and the result is recorded; (5) Steps 2–4 are repeated many times to produce a reliable distribution of the test statistic (this distribu- tion is an estimation of the distribution of the test statistic under conditions where the null hypothesis is true, that is, no difference between groups); and (6) a p value for the test is calculated as the proportion of values in this distribution that are equal to, or more than, the original observed test statistic.

In the analyses that follow, we derive proximity coefficients for all possible contingencies among police negotiator and perpetrator behaviors. These are derived separately for each of the 25 case sequences. We then use a 2 (HC and LC) by 2 (Time 1 and Time 2) factorial ANOVA and the associated F statistics as a means of measuring the differences between the proximity coefficients de- rived for the HC and LC cases. Specifically, after each resampling of the data, we calculate F for the main and interaction effects and form a distribution of F for each effect. The significance of the difference between groups is then determined by locating the observed F value in the distribution provided by the resampling of F, rather than the sampling distribution used in traditional ANOVA testing. Thus, on some occasions the absolute value of the F statistic may be very small but the chance of observing it unlikely (i.e., p � .05). On other occasions the absolute value may seem large but such differences are regularly observed within randomized data.


Frequency of Behavior

The 25 incidents contain a total of 6,980 coded speaking turns, of which 3,498 were spoken by police negotiators and 3,482 were spoken by perpetrators. Of these speaking turns, 2,045 (29%) could be typified as one of the four negotiation behaviors as spoken by the perpetrators, while 1,810 turns (25%) could be typified as one of the four negotiation behaviors as spoken by the police negotiators. So in total, our analyses include 54% of all coded behaviors. A series of chi-square tests indicate that there are no significant differences in the frequency of police negotiators’ use of behaviors across extortion and kidnapping cases (all �2 � 1.0, ns) and no significant differences in the frequency of perpe- trators’ use of behaviors across extortion and kidnapping cases (all �2 � 1.0, ns).


To test our predictions about the frequency of occurrence of per- suasive arguments across cultures, we calculate the frequency of occurrence of the four negotiation behaviors included in this study as a function of perpetrators’ culture and time period. Table 1 shows the average frequency of speaking turns (coded as one of the four behaviors in the negotiations) as a function of culture and time period. These frequencies have been standardized to take account of the fact that the LC negotiations contained more coded speaking turns than did the HC negotiations. Specifically, the frequencies for HC and LC negotiations have been adjusted to sum to an overall number of speaking turns of 3,490, which is the mean of the coded speaking turns for LC negotiations (3,616) and HC negotiations (3,364). Without this adjustment, the frequency of occurrence of behaviors in LC negotiations would be marginally higher than that in HC negotiations simply due to the length of the interaction, thereby obscuring the differences in behavior use that result from police negotiators’ and perpetrators’ choices.

The data in Table 1 enable a test of Hypothesis 1. Consistent with Hypothesis 1, LC perpetrators made significantly more use of persuasive arguments compared with HC perpetrators, �2(1, N � 323) � 5.59, p � .05. The analysis also revealed an effect of culture in the use of threats. Specifically, LC perpetrators commu- nicated significantly more threats compared with HC perpetrators, especially in the first half of the negotiations, �2(1, N � 652) � 4.39, p � .05. Finally, our analysis found no significant differences in the use of information sharing and compromising across LC and HC perpetrators. We also examined whether or not police negoti- ators’ behavior differed as a function of perpetrators’ cultural background. Only for compromising behavior did we find an effect, with police negotiators making significantly more use of compromising when interacting with an HC perpetrator compared with an LC perpetrator, �2(1, N � 43) � 5.47, p � .05.

To determine whether or not this difference in the use of compromising represented a difference in the magnitude of com- promises being made to HC and LC perpetrators (i.e., that police negotiators used regular small compromises with HC perpetrators and fewer larger compromises with LC perpetrators), we had two raters rank-order the 44 police negotiator compromises from most significant to least significant in each incident. These rank order- ings, which were made reliably across raters (rs � .91, p � .01),

were correlated with a dichotomous variable denoting whether the compromise occurred in a negotiation with an HC perpetrator (scored as 0) or LC perpetrator (scored as 1). Specifically, we separately compared each rater’s rank ordering of the compromise utterances to the dichotomous culture variable with a rank-biserial correlation (Cureton, 1956; Willson, 1976). There was no signif- icant relationship between the ranked magnitude of compromises and the perpetrators’ cultural group (r � .14, ns, for Rater 1; r � .03, ns, for Rater 2), suggesting that the difference in police negotiators’ compromise behavior across LC and HC groups may not be explained simply as a variation in type of compromise being made. A similar result was found when we correlated time (0 � Time 1, 1 � Time 2) against the rank ordering of compromises (r � .02, ns, for Rater 1; r � –.05, ns, for Rater 2), suggesting that the magnitude of police negotiators’ compromises also did not vary systematically over time.

Cross-Cultural Differences in the Organization of Influence

Table 2 shows mean proximity coefficients for the responses of HC and LC perpetrators to police negotiators’ cues, analyzed by time period to allow a comparison of proximities over time. As can be seen from Table 2, the overall value of the coefficients across the contingencies is high. This is the case because most of the behaviors occur regularly across the interaction and not in discrete periods of interaction, which would lead to “blocks” of high- and low-value coefficients. This patterning is consistent with the pre- vious finding (Donohue et al., 1991) that crisis negotiations in- volve a complex organization of cues and responses, with negoti- ators quick to move through issues and respond with diverse influence tactics. However, despite their absolute values, the rela- tionships between police cues and perpetrator responses vary con- siderably. In some cases, the behavior of the police negotiator almost always resulted in an immediate reciprocation by the per- petrator. For example, as indicated by the coefficient of .97, perpetrators from LC cultures often reciprocated police negotia- tors’ information sharing almost immediately. In contrast, other responses were rarely given to police negotiators’ cues, as is true, for example, of both LC and HC perpetrators’ reciprocation of

Table 1 Relative Frequency of Occurrence for the Four Negotiation Behaviors as a Function of Cultural Context and Time Period

Initiator and behavior

High-context culture Low-context culture

Time 1 Time 2 Overall Time 1 Time 2 Overall

Perpetrator Persuasive arguments 66.4� (64) 66.4� (64) 132.8� (128) 92.6� (96) 96.6� (100) 189.2� (196) Threats 155.6� (150) 143.2 (138) 298.8� (288) 201.6� (209) 150.6 (156) 352.3� (365) Information sharing 244.8 (236) 273.9 (264) 518.7 (500) 257.6 (267) 247.2 (256) 504.8 (523) Compromise 14.5 (14) 10.4 (10) 24.9 (24) 13.5 (14) 6.8 (7) 20.3 (21)

Police negotiator Persuasive arguments 94.4 (91) 66.4� (64) 160.8 (155) 88.7 (92) 95.6� (99) 184.3 (191) Threats 51.9 (50) 55.0 (53) 106.9 (103) 58.8 (61) 41.5 (43) 100.4 (104) Information sharing 291.5 (281) 295.7 (285) 587.2 (566) 277.8 (288) 269.4 (279) 547.2 (567) Compromise 15.6 (15) 14.5 (14) 30.1� (29) 8.7 (9) 5.8 (6) 14.5� (15)

Note. Figures in parentheses are unadjusted frequencies. � p � .05 for chi-square difference across high and low contexts.


compromising in the second half of the negotiation ( P � .53 and P � .56). These cue–response relationships are quite different during the first period of the same negotiations ( P � .92 and P � .94).

The proximity coefficients reveal a number of other character- istics of the data. For example, the variance of proximity coeffi- cients in the four panels of Table 2 provides an indication of the extent to which perpetrators varied the way in which they re- sponded. This variance differs significantly across the HC and LC perpetrators. At Time 1, the variation of proximity coefficients, measured with the coefficient of variation (Howell, 1997), was .18 for HC perpetrators and .12 for LC perpetrators (M � .15). At Time 2, the variation of proximity coefficients was .17 for HC perpetrators and .12 for LC perpetrators (M � .15). One possible explanation for this finding, consistent with Hall’s (1976) original proposal and subsequent findings of Adair and Brett (2005), is that perpetrators from HC cultures show greater flexibility in the way in which they respond to cues compared with perpetrators from LC cultures. Finally, note the asymmetry in most of the contingencies. For example, during Time 1 of negotiations with HC perpetrators, the response of information sharing following a cue of persuasive arguments (.93) is typically more immediate than the response of persuasive arguments following a cue of information sharing (.80). These asymmetries in coefficients provide clues about the overall organization of the interaction, since those cues that initiate more proximal responses may be thought of as dominant in terms of the relationship between the two forms of influence (cf. Gottman & Notarius, 2000). In the above example, therefore, the dominant relationship between persuasive arguments and information shar- ing may be understood as one where persuasive arguments more often lead to information sharing than information sharing leading to persuasive arguments.

Comparisons across the high- and low-context sections of Table 2 enable an analysis of our predictions relating to the cue–response differences between LC and HC perpetrators. In partial support of Hypothesis 2, results show that LC perpetrators are faster than HC perpetrators in their reciprocation of persuasive arguments, but only during the second half of the negotiation (i.e., an interaction effect, F � 3.04, p � .05, �2 � .09), with HC perpetrators tending

to reciprocate persuasive arguments more quickly in the first half of the negotiations. Thus, over time, HC perpetrators reduce the immediacy with which they reciprocate persuasive arguments, while LC perpetrators increase the immediacy with which they reciprocate persuasive arguments. We also found significant dif- ferences in the immediacy with which perpetrators reciprocated threats. Interestingly, the reciprocation of threats by perpetrators from HC cultures was significantly more immediate than it was by perpetrators from LC cultures (F � 2.66, p � .05, �2 � .08).

In line with previous research (Adair et al., 2001), the coeffi- cients reveal a significant main effect of culture for information sharing across LC and HC cultures. Specifically, perpetrators from LC cultures were more likely than those from HC cultures to directly reciprocate information sharing (F � 2.65, p � .05, �2 � .06).2 Furthermore, the reciprocity of compromising behavior de- creased over time from Time 1 to Time 2, regardless of the perpetrator’s cultural background (F � 12.81, p � .05, �2 � .65).

Our final hypothesis relates to the extent to which the negotia- tion behaviors were effective in bringing about cooperation (i.e., compromising) by the perpetrator. Hypothesis 3 predicted that persuasive arguments by the police negotiator would be more effective in negotiations with LC perpetrators than with HC per- petrators, particularly in the second half of the negotiations. The analysis revealed a main effect of culture for the response of compromising following police negotiators’ cues of persuasive arguments (F � 4.72, p � .01, �2 � .25). This response was significantly more immediate following persuasive arguments in LC cultures compared with HC cultures, but this effect was not mod- erated by time. Thus, we found only partial support for Hypothesis

2 Note that, in absolute terms, the difference between coefficients for the reciprocation of information sharing in the HC and LC negotiations is negligible (M � .96 compared with .97). The reason we find a significant main effect for this small absolute difference may be explained by the low within-group variance found for the coefficients, particularly for the LC negotiations. This is best illustrated by the 95% confidence interval (CI) for the difference between coefficients across culture, which ranged from –.024 to .002. However, since this 95% CI passes through zero, albeit only slightly, we urge a cautious approach to interpreting this finding.

Table 2 Mean Proximity Coefficients for Perpetrators’ Responses to Police Negotiators’ Cues as a Function of High- and Low-Context Negotiations and Time Period

Negotiation context and police negotiators’ cue

Perpetrators’ response

Time 1 Time 2

Persuasive arguments Threats

Information sharing Compromise

Persuasive arguments Threats

Information sharing Compromise

High context Persuasive arguments .95 .91 .93 .66 .88 .88 .92 .65 Threats .91 .93 .96 .84 .87 .93 .92 .65 Information sharing .80 .85 .96 .73 .89 .91 .96 .70 Compromise .66 .88 .90 .94 .94 .98 .93 .56

Low context Persuasive arguments .86 .94 .95 .88 .92 .93 .94 .84 Threats .92 .89 .98 .94 .93 .85 .96 .77 Information sharing .81 .87 .97 .79 .90 .88 .97 .88 Compromise .81 .95 .97 .92 .78 .85 .96 .53


3. No such effect was found for the communication of threats. The analysis, however, revealed a significant main effect of culture for the response of information sharing following communication of threats. Specifically, this strategy was more effective at eliciting information sharing in LC cultures compared with HC cultures (F � 3.17, p � .01, �2 � .09). Interestingly, there was also a significant main effect of time for this contingency, with commu- nication of threats leading to less immediate information sharing at Time 2 compared with Time 1 (F � 2.89, p � .01, �2 � .08).


On a daily basis, newspapers report on individuals who have been kidnapped in crisis areas around the world, ranging from Dagestan to Iraq. The growing professional reporting of such incidents suggests that they are heavily influenced by the cultural background of the perpetrators. In this study, we examined the communicative dynamics of 25 crisis negotiations with perpetra- tors from LC and HC cultures. Taking the opportunity to examine message behavior in these extreme, win–lose structured negotia- tions, we focused on forcing strategies and, in particular, persua- sive arguments. To date, negotiation research and theory has largely neglected this type of negotiation behavior. This is surpris- ing because research shows that negotiators usually devote a relatively large share of their time to discussing arguments, some- times up to 40% (Giebels et al., 2003). Moreover, particularly in crisis negotiations, persuasive strategies are expected to be an important building block to resolving the crisis at hand (cf. Vecchi et al., 2004).

Culturally Dependent Negotiation Behaviors

We found evidence to confirm our expectation that persuasive arguments are more central to negotiations with LC perpetrators compared with negotiations with HC perpetrators. Compared with their HC counterparts, LC perpetrators were more likely to use persuasive arguments and were faster to reciprocate police nego- tiators’ persuasive arguments in the second half of the negotiation, where stress was likely to have decreased. Furthermore, LC per- petrators were more immediate in responding to persuasive argu- ments in a compromising way.

These results identify culture as a factor associated with the relative frequency with which negotiators use persuasive argu- ments in distributive negotiations, in much the same way as previous research has shown information sharing to be a factor of culture in integrative negotiations (Adair, 2003; Drake, 1995). The results also suggest that persuasive arguments are more effective at bringing about compromise from LC perpetrators. This is consis- tent with Adair et al.’s (2001) suggestion that behaviors are most effective when they are “culturally consistent” with the commu- nication style characteristic of the other party. In our case, previous evidence suggests that rational arguments and factual evidence are characteristic of LC communication (Ting-Toomey, 1988), and it is in this context that persuasive arguments were found to be more effective.

We did not find complete support for our hypotheses. One interesting and unexpected finding to emerge from our analysis concerns the difference in reciprocation of persuasive arguments across cultural groups in the first half of the negotiations. Rather

than finding that LC perpetrators were more immediate in recip- rocating persuasive arguments, we found that, in the first half of negotiation, HC perpetrators were more immediate in their recip- rocation of persuasive arguments. A possible explanation for this finding may relate to differences in the manner with which HC and LC perpetrators process persuasive arguments. Given that persua- sive arguments are less central to an HC communication style, we speculate that the HC perpetrators may have processed the police negotiators’ persuasive arguments more superficially (and perhaps responded with counterarguments less well developed) than did their LC counterparts. Since the first half of crisis negotiations is associated with high stress and degraded cognition (Donohue et al., 1991), this superficial processing of persuasive arguments would potentially allow the HC perpetrators to give faster re- sponses to the police negotiators’ persuasive arguments. The idea that differences in processing may account for our findings has parallels with recent research on information sharing, which found that HC and LC negotiators process shared information in different ways during the first half of negotiations (Adair, Weingart, & Brett, 2007).

A second interesting finding concerns the extent to which LC perpetrators responded to persuasive arguments with compro- mises. We predicted that persuasive arguments would be particu- larly effective during the second part of a negotiation, but we found that they elicited compromises from LC perpetrators irre- spective of negotiation phase. This suggests that persuasive argu- ments elicit compromising behavior regardless of whether the other party is experiencing the cognitive deterioration associated with the crisis phase of interaction. The reason for this may be that logic and deductive thinking are generally highly valued in LC cultures (Gelfand & Dyer, 2000), and tactics that make use of this way of thinking are being rewarded.

Interestingly, our data provide some indirect support for this explanation. While the proximities between persuasive argument and compromise are on average higher for negotiations with LC than HC perpetrators, the frequency of occurrence of compromise behaviors is higher for negotiations with HC than LC perpetrators. This reversal across proximity and frequency has at least two implications. First, it serves to eliminate the possibility that the greater use of persuasive arguments by LC perpetrators is an artifact of the fact that these negotiations happened to contain more compromise behavior. Second, and more importantly, it suggests that there may be an alternative form of social influence that is more effective at eliciting compromise behavior from HC perpe- trators. This alternative form is likely to involve types of influence tactics other than persuasive arguments, which were not consid- ered in our analysis and which might be expected to revolve around issues of relationship or face (Hammer & Rogan, 2002).

As expected, the analysis of threat behavior revealed a different pattern. While LC perpetrators were more likely to communicate threats, especially in the first phase of the negotiations, HC per- petrators were faster to reciprocate them. These seemingly contra- dictory findings are arguably explained by the same mechanism. Threats refer to a confrontational and assertive way of handling conflict, which is consistent with LC communication and consid- ered more inappropriate in HC cultures (Fu & Yukl, 2000). Con- sequently, not only was the communication of threats higher for LC rather than HC perpetrators, but this strategy was also more effective at eliciting information sharing from LC perpetrators


compared with HC perpetrators. Moreover, HC perpetrators were quick to “punish” police negotiators who used them with counter- threats, particularly given that crisis negotiation centers on issues of “who is in charge” (Donohue & Roberto, 1993) and HC nego- tiators may be more concerned with establishing dominance (Adair & Brett, 2004). The confrontational nature of threats may also draw attention to the need to preserve face, something that is considered more important within HC rather than LC cultures (Ting-Toomey & Oetzel, 2001).

To allow for further comparison with the findings of previous research, we also examined information sharing. We found evi- dence to suggest that LC perpetrators were more immediate in their reciprocation of information sharing than were HC perpetra- tors. However, recall that the absolute value of this difference in our data, while statistically significant, was small in magnitude (see footnote 2). This makes it difficult and inappropriate to draw any firm conclusions in the absence of further evidence, although we do note that our finding is consistent with previous research by Adair and Brett (2005; see also Adair et al., 2001). Interestingly, in contrast to what might be expected, we also found that LC perpe- trators did not use information sharing to a greater extent than did HC perpetrators. An explanation for this finding is that mutual information sharing requires cooperative goals and trust (De Dreu et al., 1998) and therefore is less prominent in win–lose negotia- tions directed at the value-claiming aspects of negotiation (cf. Adair & Brett, 2004). Yet, this reasoning seems inconsistent with the relatively high occurrence rate of information sharing for both LC and HC perpetrators. Considering the fact that all police negotiators originated from LC cultures, and therefore negotiations with HC perpetrators were in fact intercultural negotiations, adap- tation processes may have occurred. Indeed, there is some evi- dence to suggest that HC communicators are more flexible in their behavior and therefore may have adapted more to LC police negotiators than did their LC counterparts (cf. Adair, 2003; Adair et al., 2001). This successful adaptation might also explain why police negotiators tended to compromise more when dealing with an HC perpetrator compared with an LC perpetrator. Furthermore, this raises the possibility that ingroup/outgroup processes associ- ated with cultural differences may have played a role in shaping behavior. It is possible to argue that the reciprocation of persuasive arguments is higher in LC negotiations because the opponent is perceived as an ingroup rather than outgroup member. This alter- native explanation is based on the assumption that people sharing similar behavioral norms are more likely to match each other’s behaviors, and that such matching is on the whole beneficial (Adair, 2003; Patterson, 1983; Taylor & Thomas, 2008). However, if this was the case, we would expect to find an overall pattern of higher proximity coefficients for the same culture interactions compared with the mixed HC and LC interactions. Since we did not find this pattern, we suspect that in crisis negotiations, cultural identities are less salient than are identities associated with the role of perpetrator versus police and that, consequently, both LC and HC perpetrators regard police negotiators as outgroup members.

Taken together, our findings support the general expectation that persuasive arguments are more central and effective in negotia- tions with LC opponents and should be distinguished from other negotiation behaviors that may be considered similar in terms of being distribution-focused (threats) or in terms of being consistent with LC instead of HC communication (information sharing). At

an epistemological level, our findings highlight the need for ne- gotiation researchers to complement their analyses of the impact of psychological variables (e.g., emotion: Van Kleef, De Dreu, & Manstead, 2004; trust: Kim, Ferrin, Cooper, & Dirks, 2004) with an understanding of how this impact emerges in actual behavior. A single outcome may emerge from myriad different interactions, such that any linking of independent variables (e.g., distrust) with particular outcomes does little to help us understand why it is that that variable had the impact on negotiation that it did. Such questions about why variables have an effect can be understood only by looking under the hood of the communicative process and learning how such variables impact on negotiators’ strategic behavior.

Along with these theoretical and methodological implications, our findings also have implications for practicing negotiators. Indeed, given the previously discussed similarities between instru- mental crisis negotiations and day-to-day business negotiations, our findings have relevance to a range of (cross-cultural) norma- tive negotiations. Consistent with the advice given in handbooks of negotiation (e.g., Fisher et al., 1991), our findings suggest that developing and communicating persuasive arguments to support claims is a beneficial approach, particularly with opponents from LC rather than HC cultures. Similarly, our analysis suggests that threats will often have negative consequences in a negotiation, particularly with opponents from HC cultures, who will often be quick to punish the threatening party. Thus, at a general level, our findings (a) point to the importance of negotiators’ developing culturally refined understandings of persuasion tactics and (b) highlight the need for negotiators to be alert to the responses of their opponents.

Limitations and Questions for Future Research

While our research extends previous work by examining the communicative dynamics of protracted real-life negotiations, it leaves unanswered a number of important questions. For example, in order to compensate for the small number of interactions avail- able for analysis, we divided interactions into only two time periods. This ensured that the proximity coefficients were calcu- lated on sufficiently long sequences of interaction. However, if a larger pool of interaction were available, such that it were possible to study a subset of longer negotiations, it is likely that separating interactions into more periods would reveal further, important variations in the cue–response dynamics across HC and LC cul- tures (Adair & Brett, 2005).

A second area of unaccounted-for variation lies in the differ- ences among the national cultures we assigned to LC and HC categories. By assigning individuals to one of the categories on the basis of what is known about the culture of their country of origin, we essentially inferred rather than measured cultural differences. There is potentially considerable within-country variation in inter- personal and cognitive style that may affect how negotiators re- spond to persuasive arguments. For example, research suggests that some of the different national cultures included in our study may vary on other cultural dimensions, such as uncertainty avoid- ance (Hofstede, 2001). Moreover, while our categorization of countries into HC and LC groups is consistent with research linking these communication styles with individualistic and collectivistic national cultures (e.g., Hall, 1976), we cannot rule


out the possibility that other factors associated with individualistic and collectivistic cultures may have been (partially) responsible for our results. Finally, other factors unrelated to culture per se may also explain differences between HC and LC individuals. These include issues associated with translation or interpretation of messages, or differences in the amount of stress or anxiety expe- rienced due to being in a home or foreign country. Such factors may cause considerable within-group variability, although the data, as well as previous research on other behavioral categories (e.g., Adair & Brett, 2005), support our theorizing and hypotheses on persuasive arguments.

A third set of questions stem from the fact that our analysis does not shed much light on the impact of different manifestations of the behavioral categories that we examined. For example, in light of recent evidence showing that some types of arguments are less effective than others at eliciting compromise (cf. Harinck, 2004), it will be useful to determine whether or not specific types of arguments are particularly effective at eliciting compromises from LC perpetrators but not from HC perpetrators. This analysis will undoubtedly benefit from a refined coding scheme that captures the different forms of persuasive arguments. Similarly, and as argued above, the type of arguments used by HC perpetrators may differ systematically from those made by LC perpetrators. Evi- dence of such a difference in type of argument should help to provide an explanation for the more immediate reciprocation ob- served for HC perpetrators at Time 1. Similarly, the type (but not magnitude) of compromise behaviors may have been different in the first half of the negotiations compared with in the second half. This difference might explain why perpetrators’ reciprocation of police compromise behavior occurred with much lower proximity in the second half than in the first half of the negotiations. How- ever, it may also be the case that, as the negotiations unfold over time, the range of issues on which compromising was possible became restricted. This possibility is in line with the general decline in compromise behavior that we observed between Time 1 and Time 2 (see Table 1). Thus, by using a content analysis–based approach to examining interaction, we inevitably lost our ability to uncover some of the rich conversational dynamics that occurred during the interaction between negotiators (Lombard, Snyder- Duch, & Bracken, 2002). To examine such conversational dynam- ics would likely require a different conceptual and methodological approach to the data (e.g., Sacks, Schegloff, & Jefferson, 1974), but the findings would complement and clarify those derived from our analyses (Chapman, 2001).

A fourth set of questions arise out of our conceptual approach to understanding behavioral relationships, as encapsulated in the proximity coefficient. Our approach was based on the assumption that behaviors occurring close together within an interaction are more related than those that occur further apart. For example, we assumed that persuasive arguments were more effective at eliciting compromises when the compromise behavior occurred quickly after the persuasive argument compared with when it occurred after a delay. While this assumption is the dominant position in both theory (e.g., Watzlawick et al., 1968) and research (e.g., Adair & Brett, 2005), it is possible that delays in responding might be expected in HC negotiations, where direct responses to the other party’s communication would not be normative and might even be seen as inappropriate. Against this possibility, our data suggest that this is not the case for all behaviors for which we might expect

avoidance of direct responding to be the case. For example, we found that, compared with LC perpetrators, HC perpetrators are faster at reciprocating threats despite threats being a very direct response to police negotiators’ cues. However, this does not rule out the possibility that there may be something unique to HC negotiators’ responses to persuasive arguments that leads to de- layed responding, and this possibility merits further examination.

Finally, the kidnappings and extortions included in our research can be typically seen as instrumental crisis situations. Instrumental crisis situations look much like business transactions in which the victims are used by the perpetrators as “bargaining chips.” These situations can be contrasted with the expressive crisis incidents that have dominated research in the area so far (e.g., domestic crises: Van Hasselt et al., 2005). In expressive situations, the communication is usually intense and emotional, and the police response may be regarded as adopting a crisis counseling role rather than a negotiation role per se. It is thus an open question as to whether similar parallels will be observed when the dynamics of expressive situations are examined. In these situations, we expect other negotiation strategies, such as cooperative statements (Vec- chi et al., 2004) or affective influence strategies (Adair & Brett, 2004), to play a more important role.

In this study we consciously departed from the perspective of police negotiators and a focus on how the behavior of police negotiators influences the subsequent behavior of perpetrators. That is, from a practical point of view we are interested in strategies that work to influence a perpetrator and result in a peaceful resolution of the incident at hand. However, interesting additional information may be obtained by examining to what extent the behavior of perpetrators influences the behavior of police negotiators. For example, it might be that immediate retal- iation of threat behavior by HC perpetrators is responsible for the somewhat higher occurrence of compromising behavior of police negotiators in encounters with HC rather than LC perpetrators. As mentioned previously, our research also lacks insight into possible adaptation processes occurring in negotiations with HC perpetra- tors. Future research should aim at including police negotiators from HC cultures as well.

Practical Implications and Conclusion

An important implication of our findings is that police negoti- ators should be sensitive to the influence strategies that they use when interacting with perpetrators from LC and HC cultures. Our research suggests that persuasive arguments, usually found to be a central element of negotiations (Giebels et al., 2003), may be particularly effective when seeking to influence LC perpetrators. In contrast, threats appeared to be ineffective at gaining concilia- tion from perpetrators, especially from HC perpetrators, since threats often lead to counterthreats and escalation. Interestingly, for both persuasive arguments and threats, the evident differences were found to be greater during the second half of interaction. Thus, cultural differences in responding seem to be particularly acute when the initial crisis response has given way to a more normative form of interaction.

These findings are important considering our general observa- tion that police negotiators appear to be quite consistent in their approach. Our findings suggest that although the existing police approach has been relatively successful (Greenstone, 1995), there


are opportunities for further fine-tuning the use of message tactics to incorporate growing knowledge of cultural differences. This may be particularly important considering the significant growth in the cultural diversity of the perpetrators of kidnappings and extor- tions (Giebels & Noelanders, 2004; Ostermann, 2002; Taylor & Donohue, 2006). Our findings suggest that, at the microlevel of ongoing crisis negotiations, there is value in critically reviewing the interaction process and the pathways that may allow negotia- tors to break through undesirable interaction patterns.


Adair, W. L. (2003). Reciprocal information sharing and negotiation out- come in East-West negotiations. International Journal of Conflict Man- agement, 14, 273–296.

Adair, W. L., & Brett, J. M. (2004). Culture and negotiation processes. In M. J. Gelfand & J. M. Brett (Eds.), The handbook of negotiation and culture (pp. 158–176). Palo Alto, CA: Stanford University Press.

Adair, W. L., & Brett, J. M. (2005). The negotiation dance: Time, culture, and behavioral sequences in negotiation. Organization Science, 16, 33–51.

Adair, W. L., Okumura, T., & Brett, J. M. (2001). How negotiators get to yes: Predicting the constellation of strategies used across cultures to negotiate conflict. Journal of Applied Psychology, 86, 371–385.

Adair, W. L., Weingart, L., & Brett, J. (2007). The timing and function of offers in U.S. and Japanese negotiations. Journal of Applied Psychology, 92, 1056–1068.

Argyle, M. (1969). Social interaction. New York: Atherton Press. Auld, F., & White, A. M. (1959). Sequential dependencies in psychother-

apy. Journal of Abnormal and Social Psychology, 58, 100–104. Axelrod, R. M. (1984). The evolution of cooperation. New York: Basic

Books. Bakeman, R., & Gottman, J. M. (1997). Observing interaction: An intro-

duction to sequential analysis (2nd ed.). New York: Cambridge Univer- sity Press.

Brett, J. M. (2001). Negotiating globally: How to negotiate deals, resolve disputes, and make decisions across cultural boundaries. San Francisco: Jossey–Bass.

Brett, J. M., Shapiro, D. L., & Lytle, A. L. (1998). Breaking the bonds of reciprocity in negotiations. Academy of Management Journal, 41, 410–424.

Chapman, S. (2001). In defence of a code: Linguistic meaning and propositionality in verbal communication. Journal of Pragmatics, 33, 1553–1570.

Choi, I., & Nisbett, R. E. (2000). Cultural psychology of surprise: Holistic theories and recognition of contradiction. Journal of Personality and Social Psychology, 79, 890–905.

Cialdini, R. B., Wosinska, W., Barrett, D. W., Butner, J., & Gornik-Durose, M. (1999). Compliance with a request in two cultures: The differential influence of social proof and commitment/consistency on collectivists and individualists. Personality and Social Psychology Bulletin, 25, 1242–1253.

Cureton, E. E. (1956). Rank-biserial correlation. Psychometrika, 21, 287–290. De Dreu, C. K. W. (1995). Coercive power and concession making in

bilateral negotiation. Journal of Conflict Resolution, 39, 646–670. De Dreu, C. K. W., Giebels, E., & Van de Vliert, E. (1998). Social motives

and trust in integrative negotiation: The disruptive effects of punitive capability. Journal of Applied Psychology, 83, 408–422.

De Dreu, C. K. W., & Weingart, L. R. (2003). Task versus relationship conflict, team performance, and team member satisfaction: A meta- analysis. Journal of Applied Psychology, 88, 741–749.

Deutsch, M. (1974). Resolution of conflict. New Haven, CT: Yale Univer- sity Press.

Deutsch, M., & Krauss, R. M. (1962). Studies of interpersonal bargaining. Journal of Conflict Resolution, 6, 52–76.

Donohue, W. A., Ramesh, C., Kaufmann, G., & Smith, R. (1991). Crisis bargaining in intense conflict situations. International Journal of Group Tensions, 21, 133–145.

Donohue, W. A., & Roberto, A. J. (1993). Relational development as negotiated order in hostage negotiation. Human Communication Re- search, 20, 175–198.

Drake, L. E. (1995). Negotiation styles in intercultural communication. International Journal of Conflict Management, 6, 72–90.

Dunlap, W. P., Burke, M. J., & Smith-Crowe, K. (2003). Accurate tests of statistical significance for r-sub(WG) and average deviation interrater agreement indexes. Journal of Applied Psychology, 88, 356–362.

Edgington, E. S. (1995). Randomization tests. New York: Marcel Dekker. Ember, M. (1988). The human relations area files: Past and future. Behav-

ior Science Research, 22, 97–104. Faure, G. O. (2002). Negotiation: The cultural dimension. In V. Kreme-

nyuk (Ed.), International negotiations: Analysis, approaches, issues (pp. 392–415). San Francisco: Jossey–Bass.

Faure, G. O. (2003). Negotiating with terrorists: The hostage case. Inter- national Negotiation, 8, 469–494.

Festinger, L. (1957). A theory of cognitive dissonance. Stanford, CA: Stanford University Press.

Fisher, R., Ury, W., & Patton, B. (1991). Getting to yes: Negotiating agreement without giving in. New York: Penguin Books.

French, J. R. P., & Raven, B. H. (1959). The bases of social power. In D. Cartwright (Ed.), Studies in social power (pp. 150–167). Ann Arbor: University of Michigan.

Friedman, R., Anderson, C., Brett, J., Olekalns, M., Goates, N., & Lisco, C. C. (2004). The positive and negative effects of anger on dispute resolution: Evidence from electronically mediated disputes. Journal of Applied Psychology, 89, 369–376.

Fu, P. P., & Yukl, G. (2000). Perceived effectiveness of influence tactics in the United States and China. Leadership Quarterly, 11, 251–266.

Gelfand, M., & Dyer, N. (2000). A cultural perspective on negotiation: Progress, pitfalls, and prospects. Applied Psychology: An International Review, 49, 62–99.

Gelfand, M. J., Nishii, L. H., Holcombe, K. M., Dyer, N., Ohbuchi, K., & Fukuno, M. (2001). Cultural influences on cognitive representations of conflict: Interpretations of conflict episodes in the United States and Japan. Journal of Applied Psychology, 86, 1059–1074.

Giebels, E. (1999). A comparison of crisis negotiation across Europe. In O. Adang & E. Giebels (Eds.), To save lives: Proceedings of the first European conference on hostage negotiations (pp. 17–27). Gravenhage, The Netherlands: Elsevier.

Giebels, E., De Dreu, C. K. W., & Van de Vliert, E. (1998). The alternative negotiator as the invisible third at the table: The impact of potency information. International Journal of Conflict Management, 9, 5–21.

Giebels, E., De Dreu, C. K. W., & Van de Vliert, E. (2000). Interdepen- dence in negotiation: Effects of exit options and social motive on distributive and integrative negotiation. European Journal of Social Psychology, 30, 255–272.

Giebels, E., De Dreu, C. K. W., & Van de Vliert, E. (2003). No way out or swallow the bait of shared exit options in negotiations: The influence of social motives and interpersonal trust. Group Processes and Inter- group Relations, 6, 369–386.

Giebels, E., & Noelanders, S. (2004). Crisis negotiations: A multiparty perspective. Veenendaal, The Netherlands: Universal Press.

Giebels, E., Noelanders, S., & Vervaeke, G. (2005). The hostage experi- ence: Implications for negotiation strategies. Clinical Psychology & Psychotherapy: An International Review, 12, 241–253.

Good, P. (1994). Permutation tests: A practical guide to resampling methods for testing hypotheses. New York: Springer.

Gottman, J. M., Markman, H., & Notarius, C. (1977). The topography of marital conflict: A sequential analysis of verbal and non-verbal behavior. Journal of Marriage and the Family, 39, 461–477.


Gottman, J. M., & Notarius, C. I. (2000). Decade review: Observing marital interaction. Journal of Marriage and the Family, 62, 927–947.

Gouldner, A. W. (1960). The norm of reciprocity: A preliminary statement. American Sociological Review, 25, 161–178.

Graham, J. L. (1993). The Japanese negotiation style: Characteristics of a distinct approach. Negotiation Journal, 9, 123–140.

Greenstone, J. L. (1995). Tactics and negotiating techniques (TNT): The way of the past and the way of the future. In M. I. Kurke & E. M. Scrivner (Eds.), Police psychology in the 21st century (pp. 357–371). Hillsdale, NJ: Erlbaum.

Grice, H. P. (1975). Logic and conversation. In P. Cole & J. L. Morgan (Eds.), Syntax and semantics: Vol. 3. Speech acts (pp. 41–58). New York: Academic Press.

Gudykunst, W. B., & Matsumoto, Y. (1996). Cross-cultural variability of communication in personal relationships. In W. B. Gudykunst, S. Ting- Toomey, & T. Nishida (Eds.), Communication in personal relationships across cultures (pp. 57–77). Thousand Oaks, CA: Sage.

Gudykunst, W. B., & Ting-Toomey, S. (1988). Culture and interpersonal communication. Newbury Park, CA: Sage.

Hall, E. T. (1976). Beyond culture. Garden City, NY: Anchor Press. Hall, E. T., & Hall, M. R. (1990). Understanding cultural differences.

Yarmouth, ME: Intercultural Press. Hammer, M., & Rogan, R. (2002). Latino and Indochinese interpretative

frames in negotiating conflict with law enforcement: A focus group analysis. International Journal of Intercultural Relations, 26, 551–575.

Hargreaves, S., & Cunningham, A. (2004, June 12). The politics of terror. Lancet, 363, 1999–2000.

Harinck, F. (2004). Persuasive arguments and beating around the bush in negotiations. Group Processes and Intergroup Relations, 7, 5–17.

Heine, S. J., & Lehman, D. R. (1997). The cultural construction of self-enhancement: An examination of group-serving biases. Journal of Personality and Social Psychology, 25, 915–925.

Hofstede, G. (2001). Culture’s consequences: Comparing values, behav- iors, institutions and organizations across nations (2nd ed.). London: Sage.

Hong, J. W., Muderrisoglu, A., & Zinkham, G. M. (1987). Cultural differences and advertising expression: A comparative content analysis of Japanese and US magazine advertising. Journal of Advertising, 16, 55–62.

Hornstein, H. A. (1965). Effects of different magnitudes of threat upon interpersonal bargaining. Journal of Experimental Social Psychology, 1, 282–293.

Hoshino-Browne, E., Zanna, A. S., Spencer, S. J., Zanna, M. P., Kitayama, S., & Lackenbauer, S. (2005). On the cultural guises of cognitive dissonance: The case of Easterners and Westerners. Journal of Person- ality and Social Psychology, 89, 294–310.

Howell, D. C. (1997). Statistical methods for psychology (4th ed., pp. 48–49). Belmont, CA: Wadsworth.

Kelly, H. H. (1997). Expanding the analysis of social orientations by reference to the sequential-temporal structure of situations. European Journal of Social Psychology, 27, 373–404.

Kim, P., Ferrin, D., Cooper, C., & Dirks, K. (2004). Removing the shadow of suspicion: The effects of apology versus denial for repairing competence- versus integrity-based trust violations. Journal of Applied Psychology, 89, 104–118.

Kozan, M. K., & Ilter, S. S. (1994). Third-party roles played by Turkish managers in subordinates’ conflicts. Journal of Organizational Behav- ior, 15, 453–466.

Lewicki, R. J., Hiam, A., & Wise Olander, K. (1996). Think before you speak: A complete guide to strategic negotiations. New York: Wiley.

Lewicki, R. J., Saunders, D. M., & Minton, J. W. (1997). Essentials of negotiation. Boston: Irwin/McGraw–Hill.

Lombard, M., Snyder-Duch, J., & Bracken, C. C. (2002). Content analysis

in mass communication: Assessment and reporting of intercoder reli- ability. Human Communication Research, 28, 587–604.

Lytle, A. L., Brett, J. M., Barsness, Z., Tinsley, C. H., & Janssens, M. (1995). A paradigm for quantitative cross-cultural research in organizational behav- ior. In B. M. Staw & L. L. Cummings (Eds.), Research in organizational behavior (Vol. 17, pp. 167–214). Greenwich, CT: JAI Press.

Macduff, I. (2006). Your pace or mine? Culture, time, and negotiation. Negotiation Journal, 22, 31–45.

Michener, H. A., Vaske, J. J., Schleifer, S. L., Plazewski, J. G., & Chapman, L. J. (1975). Factors affecting concession rate and threat usage in bilateral conflict. Sociometry, 38, 62–80.

Neale, M. A., & Bazerman, M. H. (1991). Cognition and rationality in negotiation. New York: Free Press.

Ninness, C., Rumph, R., Vasquez, E., & Bradfield, A. (2002). Multivariate randomization tests for small-N behavioral research: A Web-based ap- plication. Behavior and Social Issues, 12, 64–74.

Olekalns, M., & Smith, P. L. (2000). Understanding optimal outcomes: The role of strategy sequences in competitive negotiations. Human Commu- nication Research, 26, 527–557.

Ostermann, B. M. (2002). Cultural differences make negotiations different: Intercultural hostage negotiations. Journal of Police Crisis Negotiation, 2, 11–20.

Patterson, M. L. (1983). Nonverbal behavior: A functional perspective. New York: Springer–Verlag.

Perloff, R. M. (1993). The dynamics of persuasion. Hillsdale, NJ: Erlbaum. Pruitt, D. G., & Kim, S. H. (2004). Social conflict: Escalation, stalemate,

and settlement (3rd ed.). New York: McGraw–Hill. Putnam, L. L. (1985). Bargaining as task and process: Multiple functions

of interaction sequences. In R. L. Street & J. N. Cappella (Eds.), Sequence and pattern in communicative behavior (pp. 225–242). Lon- don: Edward Arnold.

Putnam, L. L., & Jones, T. S. (1982). The role of communication in bargaining. Human Communication Research, 8, 262–280.

Reykowski, J. (1998). Belief systems and collective action: Changes in Poland from the psychological perspective. Applied Psychology: An International Review, 47, 89–108.

Roger, D. B., & Schumacher, A. (1983). Effects of individual differences in dyadic conversational strategies. Journal of Personality and Social Psychology, 45, 247–255.

Sacks, H., Schegloff, E. A., & Jefferson, G. (1974). A simplest systematic for the organization of turn-taking for conversation. Language, 50, 696–735.

Schwartz, S. H. (1992). Universals in the content and structure of values: Theoretical advances and empirical tests in 20 countries. Advances in Experimental Social Psychology, 25, 1–65.

Smith, P. B., Trompenaars, F., & Dugan, S. (1995). The Rotter Locus of Control Scale in 43 countries: A test of cultural relativity. International Journal of Psychology, 30, 377–400.

Switzer, F. S., Paese, P. W., & Fritz, D. (1992). Bootstrap estimates of standard errors in validity generalization. Journal of Applied Psychol- ogy, 77, 123–129.

Tai, H. C. S. (2004). The relationship of cultural values and message strategies in service advertising. Marketing Intelligence & Planning, 22, 438–454.

Taylor, P. J. (2002a). A cylindrical model of communication behavior in crisis negotiations. Human Communication Research, 28, 7–48.

Taylor, P. J. (2002b). A partial order scalogram analysis of communication behavior in crisis negotiation with the prediction of outcome. Interna- tional Journal of Conflict Management, 13, 4–37.

Taylor, P. J. (2006). Proximity coefficients as a measure of interrelation- ships in sequences of behavior. Behavioral Research Methods, 38, 42–50.

Taylor, P. J., & Donald, I. J. (2003). Foundations and evidence for an


interaction-based approach to conflict. International Journal of Conflict Management, 14, 213–232.

Taylor, P. J., & Donald, I. J. (2004). The structure of communication behavior in simulated and actual crisis negotiations. Human Communi- cation Research, 30, 443–478.

Taylor, P. J., & Donald, I. J. (2007). Testing the relationship between local cue-response patterns and global dimensions of communication behav- ior. British Journal of Social Psychology, 46, 273–298.

Taylor, P. J., & Donohue, W. A. (2006). Hostage negotiation opens up. In A. Schneider & C. Honeymoon (Eds.), The negotiator’s fieldbook (pp. 667–674). New York: American Bar Association Press.

Taylor, P. J., & Thomas, S. (2008). Linguistic style matching and negoti- ation outcome. Negotiation and Conflict Management Research, 1, 263–281.

Tedeschi, J. T., Schlenker, B. R., & Bonoma, T. V. (1973). Conflict, power and games: The experimental study of interpersonal relations. Chicago: Aldine.

Thomas, A. P. (1985). Conversational routines: A Markov chain analysis. Language and Communication, 5, 287–296.

Ting-Toomey, S. (1988). Intercultural conflict styles: A face negotiation theory. In Y. Y. Kim & W. Gudykunst (Eds.), Theories in intercultural communication (pp. 213–235). Newbury Park, CA: Sage.

Ting-Toomey, S., & Oetzel, J. G. (2001). Managing intercultural conflict effectively. Thousand Oaks, CA: Sage.

Tinsley, C. (1998). Models of conflict resolution in Japanese, German, and American cultures. Journal of Applied Psychology, 83, 316–323.

Tinsley, C. (2001). How negotiators get to yes: Predicting the constellation of strategies used across cultures to negotiate conflict. Journal of Ap- plied Psychology, 86, 583–593.

Tjosvold, D., & Sun, H. (2000). Social face in conflict: Effects of affronts to person and position in China. Group Dynamics: Theory, Research, and Practice, 4, 259–271.

Triandis, H. C. (1994). Culture and social behavior. New York: McGraw– Hill.

Triandis, H. C., & Suh, E. M. (2002). Cultural influences on personality. Annual Review of Psychology, 53, 133–160.

Van de Vliert, E., Nauta, A., Giebels, E., & Janssen, O. (1999). Construc- tive conflict at work. Journal of Organizational Behavior, 20, 475–491.

Van Hasselt, V. B., Flood, J. J., Romano, S. J., Vecchi, G. M., de Fabrique, N., & Dalfonzo, V. A. (2005). Hostage-taking in the context of domestic violence: Some case examples. Journal of Family Violence, 20, 21–27.

Van Kleef, G. A., De Dreu, C. K. W., & Manstead, A. S. R. (2004). The interpersonal effects of anger and happiness in negotiations. Journal of Personality and Social Psychology, 86, 57–76.

Van Zwol, C. (2005). Gijzelaar van de Kaukasus: De ontvoering van Arjan Erkel [Hostage of the Kaukasus: The kidnapping of Arjan Erkel]. Am- sterdam: Prometheus/NRC Handelsblad.

Vecchi, G. M., Van Hasselt, V. B., & Romano, S. J. (2004). Crisis (hostage) negotiation: Current strategies and issues in high-risk conflict resolution. Aggression and Violent Behavior, 10, 533–551.

Vrij, A., Dragt, A., & Kippelaar, L. (1992). Interviews with ethnic inter- viewees: Non-verbal communication errors in impression formation. Journal of Community & Applied Social Psychology, 2, 199–208.

Watzlawick, P., Beavin, J. H., & Jackson, D. D. (1968). Pragmatics of human communication. London: Faber and Faber.

Weingart, L. R., Prietula, M. J., Hyder, E., & Genovese, C. (1999). Knowledge and the sequential processes of negotiation: A Markov chain analysis of response-in-kind. Journal of Experimental Social Psychol- ogy, 35, 366–393.

Weingart, L. R., Thompson, L., Bazerman, M. H., & Caroll, J. S. (1990). Tactical behavior and negotiation outcomes. International Journal of Conflict Management, 1, 7–31.

Willson, V. L. (1976). Critical values of the rank-biserial correlation coefficient. Educational and Psychological Measurement, 36, 297–300.

Yang, S.-L., Wu, B., & Huang, S.-L. (2007). Kidnapping in Taiwan: The significance of geographic proximity, improvisation, and fluidity. Inter- national Journal of Offender Therapy and Comparative Criminology, 51, 324–339.

Received February 15, 2007 Revision received April 29, 2008

Accepted May 7, 2008 �


Comments are closed.