What Strategy Is Likely to Dominate Each Stage of the Negotiation?
Researchers and theorists have addressed this question with phase models and stage models (Brett, Weingart, & Olekalns, 2004; Putnam, 1990). Phase models focus on coherent chunks of similar strategic behaviors or tactics.
A new phase begins when the strategy changes (Holmes, 1992). Phase researchers then describe negotiation dynamics in terms of the order in which different phases dominate. In contrast, stage models focus on time. Based on negotiation length, they divide negotiations into halves, thirds, quarters, or even eighths (Gulliver, 1979) and then determine what strategy dominates each stage (Morley & Stephenson, 1977).Both stage and phase models are valid, but each provides somewhat different insight into negotiation dynamics. Because stages change with artificially imposed time periods, stage models reveal the interplay among multiple strategies at any given point in time. Because phases change when negotiators switch strategies, phase models clearly reveal the sequence of strategies but not the concurrent interplay of multiple strategies. Stage and phase models both suggest that negotiations typically begin with the parties using distributive strategy to position themselves and that doing so may be important for ultimately reaching an integrative outcome (Lytle et al., 1999; Morley & Stephenson, 1977; Olekalns et al., 1996; Olekalns et al., 2004). This statement, however, appears to be inconsistent with the culture and negotiation strategy literature just reviewed. In an effort to reconcile these inconsistencies, the following sections lay out the evidence for starting distributive and then propose a theoretical reconciliation with the culture and negotiation strategy research.
Strong evidence for the conclusion to “start distributive” comes from Olekalns et al. (2004) and Weingart et al. (2007). These two studies, which draw from the same data set, also provide the basis for our ultimate theoretical reconciliation.
These studies used data from a multiparty, multi-issue simulation with integrative potential. There were 36 four-party groups composed of zero to four members who primed to act cooperatively or individu- alistically. Speaking turns (a statement by one negotiator surrounded by the other negotiator’s statements) were coded into 32 categories of behavior. Correspondence analysis reduced the 32 categories to 6: integrative information about interests and priorities, value creation (e.g., multi-issue offers), distributive information about positions, value claiming (e.g., single-issue offers and substantiation), process (e.g., “this is going nowhere”), and closure (e.g., “do we have an agreement?”). Phase analysis (Holmes, 1992) revealed that 31 of 36 groups initiated negotiations with a distributive strategy phase, typically distributive information about positions (Olekalns et al., 2004). Of those groups, 29 ended with an integrative phase, typically value creation, and reached agreement. Five groups started with an integrative phase, typically integrative information about own interests or mutual interests, and four of them did not reach agreement!The researchers concluded that, at least in this multiparty situation, starting with integrative strategy allowed one or more of the counterparts to turn the negotiation distributive, breaking the trust associated with the opening integrative strategy (Olekalns et al., 2004). Once broken, trust was difficult to reestablish. In contrast, starting with positions (distributive) is expected in negotiations (Morley & Stephenson, 1977), and because expected behaviors do not violate norms, starting with positions has no negative trust implications. However, once positions are stated, the negotiators typically remain far from agreement. At this point, the parties must move to integrative strategy to search for agreement.
While Olekalns and colleagues’ (2004) study used phase analysis of a four-party, deal-making negotiation, Lytle et al. (1995) used a (four) stage analysis of a two-party dispute resolution negotiation.9 The two studies’ similar findings demonstrate the broad applicability of the Olekalns et al.
(2004) pattern. The first stage of the Lytle et al. (1995) negotiation was dominated by distributive strategy, the second by integrative strategy, the third by a mix of distributive and integrative strategy, and the fourth by integrative strategy. The researchers concluded that disputants began by stating and substantiating their positions. Having established their differences, they began to use integrative strategy to understand interests and develop settlement options. By the third quarter, parties seemed reasonably confident that a zone of agreement existed and again used distributive strategy to position themselves within that zone. To reach an agreement, they ended the negotiation with integrative strategy.Across real-world or simulated negotiations, deal-making or disputing negotiations, and qualitative or quantitative studies, the phase and stage research reveals a similar pattern of strategic dynamics: Opening with distributive strategy and moving into integrative strategy appears both normative and effective for generating agreements. Yet this start distributive conclusion is inconsistent with Adair and Brett’s (2005; see also Brett, 2007) conclusion that negotiators need to get integrative information sharing going (via a culturally normative strategy) during the first quarter of the negotiation. How can we reconcile these findings?
We suggest two factors that may contribute to reconciling what seems like irreconcilable differences between the dynamic studies and the cultural studies. The first has to do with categorizing negotiation behaviors into strategies. The second has to do with the difference between phase and stage analyses.
Regarding categorization, recall that the Olekalns et al. (2004) study reduced 32 behavioral codes into six strategic categories, which separated information sharing into integrative (about preferences and priorities) and distributive (about positions and substantiation). By separating integrative and distributive information sharing, Olekalns et al.
could see exactly what type of information exchange opened most negotiations (i.e., distributive). Recall, however, that Olekalns et al. used phase analysis, so distributive information strategy only prevailed until overcome with another strategy, for example, integrative information. Thus, Olekalns et al.’s negotiators may very well have moved on to integrative information sharing very quickly, at least by the end of the negotiations’ first quarter. However, phase modeling cannot reveal when negotiators moved, just that they did move. Insights about the timing of strategic movement in negotiations come from the stage models of Lytle et al. (1995) and Adair and Brett (2005). Thus, it is both possible and likely that successful negotiators start sharing distributive information about positions and then move toward integration, sharing interests, and priorities. It is also clear that, if negotiators are going to attain high joint gains, they need to make this transition rather rapidly as the negotiation progresses through time.Thus, the apparently divergent findings may represent a methodological artifact more than a substantive disagreement. This observation suggests a number of guidelines and opportunities for future research. First, careful coding of negotiation behaviors, which distinguishes information sharing about positions from information sharing about interests and priorities, seems essential. Second, using both phase and stage analyses within the same data set would shed the most light on negotiation dynamics (but would also require more time and resources).