Few of the predictions about differential frame shift were confirmed. Nevertheless, some unexpected, but interesting, findings emerged. The following section discusses the results of the differential frame use over time together with the frame shift equivalence findings.
In the cooperative condition, a trend for competitively-oriented frames to converge over time was observed. In particular, use of the non-inclusive process frame was found to converge significantly over time. When two cooperators negotiate, high levels of cooperation are likely. One of the pitfalls of high levels of cooperation is that such an orientation can limit the extent to which issues are explored (Ben-Yoav & Pruitt, 1984). Some level of competition in cooperative settings can provide a catalyst to stimulate greater exploration of issues, which promotes more optimal outcomes (Ben-Yoav & Pruitt, 1984). Excessive competition, however, can become destructive and result in an impasse. In the current study, convergence in the use of this non-inclusive process frame was detected, but tests of frame shift equivalence revealed that joint use of the frame also diminished over time. Thus, although negotiators were using this frame more similarly by the second stage of the negotiation, they were also using this frame significantly less often. In cooperatively motivated dyads, process-related interventions were introduced occur earlier. The current finding is consistent with research that suggests that negotiators under low time pressure experience temporary impasses - which might correspond to non-inclusive process frames -- early in the negotiation (Harrinck & de Dreu, 2004). According to this research, these impasses foster a shift from lower levels to higher levels of integrative behaviour. In the current research, however, there was no support for a change in levels of integrative behaviour as indicated by the use of cooperatively-oriented frames over time.
In summary, within the cooperative condition, all frames, apart from the non-inclusive process frame, were used similarly by both negotiators across time. For all other frames, the level of frame use was maintained. For cooperatively-oriented frames in particular, there was limited support for the frame convergence findings of Pinkley and Northcraft (1994), who found that use of the cooperate frame converged over time. The difference may be, in part, attributable to variations in the operationalisation of cooperative frames.
In the Pinkley and Northcraft study (1994), convergence towards the cooperative end of the cooperate/win frame dimension was observed. Following Follett (1942), their interpretation was that, in part, negotiators had influenced each other during the negotiation to arrive at a more similar appraisal of the conflict. A key difference was that Pinkley and Northcraft (1994) evaluated changes in framing after the negotiation was complete, whereas the current study evaluated changes while the negotiation was still in process. In their study, negotiators could reflect on proceedings once an agreement had been reached, and thus the convergence of the cooperate frame may have been indicative of negotiators reaching some form of consensus, rather than a direct indication of the way that negotiators mutually influenced each other during negotiation proceedings. Thus, even though cooperatively motivated dyads in the current study did not converge in their use of cooperative frames, their cooperative framing might have converged once an agreement had been reached.
Within the competitive condition, power/rights-based justification frames were predicted to converge over time. Although competitively motivated negotiators have been found to interpret information in terms of power (Van Lange & Kuhlman, 1994), and that the tendency to frame information in this way is likely to mutually increase as negotiators strive to defend their positions within a context of spiralling conflict, there was no empirical evidence for the convergence of this frame. Likewise, no support for the pattern of frame shift equivalence emerged. The comparatively low incidence of these frames in the competitive motivation condition, however, could accommodate these observations.
As with the power/rights-based justification frames, no other frames showed convergence over time. Cooperative outcome frames showed increased joint use over time. This finding is consistent with negotiation stage models, which maintain that, following attempts to outline their arguments, negotiators shift to adopt more problem-solving behaviours (Putnam & Holmer, 1992). In these later phases, framing information in terms of the desired outcomes is more likely as negotiators shift their focus towards the generation of alternative solutions and the exchanging of offers (D'Zurilla, 1988; Putnam & Holmer, 1992).
Tests of frame shift equivalence in the competitive condition also revealed that joint use of the substantive issues frame decreased over time. This finding is also consistent with stage models of negotiation, which assert that framing information in terms of the concrete facts and details is likely to occur in the earlier stages of the negotiation (Putnam & Holmer, 1992). During this time, negotiators are utilising the more factual elements of the dispute to build their case, to position themselves, or to test their power (D'Zurilla, 1988; Putnam & Holmer, 1992). Interestingly, decreasing joint use of this frame showed some evidence of a relationship with more distributive outcomes. Although not the subject of extensive investigation, some research suggests that differences in the way that issues are defined is likely to inhibit agreement (Bacharach & Lawler, 1984). This account implies that disagreement over definitions of issues will encourage negotiators to either return to contest the facts of the case or choose to proceed through the issues without shared understandings. In continuing to negotiate without shared understandings, subsequent bargaining is likely to be conducted on a less stable foundation. Although not tested directly in the current research, this explanation is consistent with the finding that joint use of the substantive issues frame decreased over time and was related to more distributive outcomes. Conceivably the facts of the negotiation were not sufficiently established to enable integrative outcomes.
The finding is also consistent with Pinkley and Northcraft's (1994) assertion that, when negotiators start to adopt substantive issues frames, especially in competitive environments, the likelihood of more integrative agreements increases. Interpreting the current finding in light of previous research, this result suggests that the timing of the use of the substantive issues frame is important in obtaining integrative outcomes. That is, the facts of the negotiation need to be sufficiently established before proceeding to the next stage of the negotiation. To investigate further the issue of the importance of shared understandings of the facts, the content of these substantive issues frames could be examined. As Gray and Donnellon suggest (1989), each fact is ascribed meaning according to the negotiator. The meanings that negotiators ascribe to the issues in dispute are subject to change during the course of the negotiation. In this sense, similar use of substantive issues frames across the negotiation between negotiators might not always reflect a sharing of understandings.
There were two competing predictions derived about the behaviour of mixed motivation dyads. These predictions were based on previous research that suggests that either the cooperator in mixed dyads will assimilate his or her behaviours towards the less contextually responsive competitor (Kelley & Stahelski, 1970; Olekalns et al., 1996), or that both negotiators will shift away from their motivation-consistent behaviours to a set of mutually shared behaviours (Olekalns & Smith, 1999). Potentially, both of these could have occurred concurrently.
To test the first prediction, patterns of convergence were examined to determine whether cooperators had converged towards competitors in their use of frames. This test failed to find evidence of frame convergence. The analysis also revealed that negotiator frame use did not diverge over time.
The second prediction was tested by investigating whether the pattern of frame use in the first half of the negotiation differed from that in the second half. The focus was upon how negotiators shifted together and, because frame maintenance had been established for all frames in this condition, any shifts in joint frame use over time could be interpreted as simultaneous shifts in the use of frames.
An analysis of patterns of joint frame use across time showed significant shifts in use for five of the eight frames: an increase in joint use was found for the cooperative outcome and inclusive process frames, and decreases in use were found for the competitive outcome, cooperative aspiration and the power/rights-based justification frames. Joint use of competitive aspiration, non-inclusive process and substantive issues frames remained similar throughout the negotiation. In effect, negotiators shifted from a more competitive to a more cooperative framing of offers, curbed their use of the less collaborative power/rights-based justification frames and increased the inclusiveness of their process-oriented framing. Interpreting this more broadly, negotiators shifted across the course of the negotiation to mutually reframe information in a more cooperative way. Essentially, as negotiators became closer to reaching an agreement, they adopted frames that were more facilitative of an agreement. Moreover, a decrease in joint use of competitive outcome frames was found to predict more integrative outcomes. These two findings are compatible with negotiation phase models (Putnam & Holmer, 1992) in that over the course of the negotiation, as joint use of cooperatively-oriented frames increased and competitively-oriented frames decreased, they became more likely to secure an integrative or high joint gain agreement.
These findings together provide greater support for the second of the alternative hypotheses, which states that mixed motivation negotiators will shift in their use of frames during the course of the negotiation. Despite the pattern in frame use changing over time and notwithstanding significant shifts over time in the use of some frames, negotiators did not wholly abandon the frames they used in the first stage of the negotiation for another set. Moreover, although the pattern in frame use may have changed over time, certain types of frames were predominant in use across the two stages (e.g., outcome-oriented frames) and certain types that were used to a lesser extent (e.g., process-oriented frames). This finding is consistent with previous research that demonstrates that negotiators employ positional statements that are akin to the use of outcome-oriented frames (Donnellon & Gray, 1989) more frequently than they refer to the negotiation process (Carroll and Payne, 1991).
Although convergence and divergence was not prevalent within conditions, in the mixed motivation condition, frame shift equivalence was a relatively common phenomenon. The most likely explanation for the phenomenon of greater frame shift activity in the mixed motivation condition is the contrasting motivational orientations of negotiators in these dyads.