Understanding Negotiations

[The following text is from “So How Much Money Will I Make?“]

Project Lead : Ron Caneel

In the Spring and Fall of 2004, we examined whether conversational dynamics occurring within the first five minutes of a salary negotiation can predict negotiated outcomes. In a simulated employment negotiation, activity level, mirroring, prosodic stress, and conversational turn taking predict more than 20% of the variance in negotiated outcomes. The conversational dynamics associated with success among high status parties are different from the conversational dynamics associated with success among low status parties

For this study, we used results of a negotiation sessions which was conducted with first-year Sloan students. For a class assignment they had to carry out a version of the case study “New recruit”. One person plays the role of the vice-president while the other the role of the middle manager, how looks for a better job. The two parties have to come to an agreement on eight different dimensions. Each person gets two scores at the end: an individual score (claiming value) and a dyadic score (creating value). These values were taken to distinguish between experts and novice. Assuming that a certain skill set helps the each party to generate a high outcome.

As hypothesized, four conversational dynamics, occurring within the first five minutes of a negotiation, are highly predictive of negotiated outcomes. In fact, whereas the average effect size in past thin slices research is r = .39 (Ambady & Rosenthal, 1992), our effect size was r = .54 for Middle Manager Points and r = .52 for Vice President Points. Moreover, the strength of this effect is comparable to the predictive power of negotiator aspiration levels, commonly known in the negotiation field to be a strong determinant of outcomes (cf. Barry & Friedman, 1998; for reviews, see Hamner, 1980, and Pruitt, 1981).

When operationalizing the stress feature, we could not be certain whether stress would result from purposeful emphasis or unintentional jitter in the voice as a result of physiological discomfort. Our results suggest that the latter is true. In our study, vocal stress during the first five minutes of a negotiation was a liability—particularly for Middle Managers.

We modeled our measure of mirroring after the mimicry behavior described by Chartrand and Bargh (1999), but our measure pertained to speech patterns rather than body language. Nevertheless, like in Chartrand and Bargh’s experiment, mirroring in our study appeared to have a positive effect on negotiation. It is noteworthy that mirroring in our study was predictive of individual outcome only among low-status parties. Chartrand and Bargh found that mimicry tended to occur more among individuals with dispositionally greater perceptual activity directed at others. Since low-power parties tend to pay more attention to high-power parties (Keltner & Robinson, 1997), it is not surprising that low-power mimicry of high-power parties would be more common and thus more normative than the other way around. Furthermore, if mirroring is a consequence of perspective-taking, then the fact that mirroring by Vice Presidents predicted Joint Points would suggest that perspective-taking is beneficial for integrative bargaining, a controversial issue in the negotiation field (Drolet, Larrick, & Morris, 1998).

As we predicted, activity level was positively associated with negotiation outcomes, but this effect was apparent only among the Vice Presidents. Higher activity among Middle Managers correlated positively with points earned by the Vice President. Future research will be necessary to explore this effect and the mechanism behind it.


  • Thin Slices of Negotiation, Curhan, J.; Pentland, A.; Caneel, R.; Eagle, N.; and Martin, M. Proceedings of the Annual Meeting of the Academy of Management, Honolulu, Hawaii, Aug 5-10, 2005
  • A Negotiation Advisor, Pentland, A., Curhan, J., Khilnani, R., Martin, M., Eagle, N., Caneel, R., and Madan, A. Poster at Seventheenth Annual ACM Symposium on User Interface Software and Technology, Santa Fe, New Mexico, October 24-27, 2004. Also Media Laboratory Vision And Modeling Technical Report 577, MIT.
  • Social Dynamics: Signals and Behavior, Pentland, A. MIT Media Laboratory Technical Note 579, appears in IEEE Int’l Conf. Developmental Learning (ICDL) 2004, Oct. 20-22, San Diego, CA.
  • Social Signaling in Decision Making, Caneel, R. Masters Thesis in Media Arts and Sciences, MIT. June 2005.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s