From Max Bazerman’s “Judgement In Managerial Decision Making”
Reason Analogically (from p. 232)
Analogical reasoning, or the process of abstracting common lessons from two or more situations, turns out to be a remarkably simple debiasing approach (D. Gentner, G. Loewenstein, & L. Thompson, 2003a; Loewenstein, Thompson, & Gentner, 1999; L. Thompson, D. Gentner, & J. Loewenstein, 2000). Research shows that people learn far more from cases, simulations, and real-world experiences when they are able to take away an abstract form of the learning message. In the context of learning to negotiate through simulations, Gentner, Loewenstein, & Thompson (2003b) found that greater debiasing occurred among participants when they took part in two exercises that had the same lesson and were asked how the two simulations were related than when they assessed the same two exercises and were asked to explain the lesson of each one. When people learn from one episode at a time, they too often focus on superficial characteristics of the situation and assume that the message applies only to the specific context of the decision (such as learning how to buy a house). By contrast, the process of abstracting similar lessons from two episodes (such as learning to overcome the mythical fixed pie of negotiation following a house purchase and a workplace negotiation) creates more generalizable insight.
By assessing participants’ performance on a third task, Gentner, Loewenstein, and Thompson (2003a) have demonstrated evidence of debiasing decisionmaking and negotiation behavior through this type of analogical reasoning. They have replicated their research conclusions across a number of studies, many involving executives and consultants. Thompson, Gentner, & Loewenstein (2000) claim that when people make a comparison, they focus on the similarities between examples, whose common structure becomes more transparent. Identifying the common structure—the principle shared by both examples—helps the learner form a schema that is less sensitive to the irrelevant surface or context features of the particular examples. Such an abstract principle is more likely to be transferred to new situations with different contexts than a principle that is not abstracted from its original context. These impressive findings on the effectiveness of analogical reasoning open up important new directions for debiasing research and offer guidance on how to use cases and simulations to maximize generalizable learning
Building on Thompson et al.’s analogical reasoning work, Idson, Chugh, Bereby-Meyer, Moran, Grosskopf, and Bazerman (2004) suggest that understanding differences, as well as similarities, across problems may also be a very useful means of transferring knowledge. Idson et al. (2004) show that training based on differences can reduce bias in the Acquiring a Company problem, which, as discussed earlier, had proven resistant to many other debiasing techniques. Using the five problems from Tor and Bazerman (2003), Idson et al. (2004) had study participants either (1) examine the two versions of the Monty Hall problem and the two versions of the Dividing a Pie problem as four separate problems, or (2) presented the problems in pairs.
All participants were then given multiple trials to solve the Acquiring a Company problem, with pay based on performance. They also gave the same Acquiring a Company problem to other study participants who were not trained on the Monty Hall problem and the Dividing a Pie problem. Idson et al. (2004) found that allowing study participants to view the Monty Hall and Dividing a Pie problems as pairs helped them understand the differences between the two versions of each problem and generalize the importance of focusing on the decisions of other parties and the rules of the game. These lessons, which were also the keys to solving the Acquiring a Company problem, indeed enabled participants to perform substantially better on this problem. This research offers evidence that examining differences between seemingly related problems may be a successful direction for improving decision making. What is the optimal level of abstraction that should occur to help people form analogies across problems? Moran, Bereby-Meyer, and Bazerman (2008) argue that teaching people more general negotiation principles (such as “Value can be created” or “It is important to understand how parties’ interests interrelate”) enables successful transfer to a broader range of new negotiation tasks than the focused analogies of Loewenstein et al. (2003). Moran et al. (2008) argue that learning general principles will improve not only the ability to positively transfer specifically learned principles but also the ability to discriminate their appropriateness—i.e., to determine when a principle should and should not be applied.
Moran et al. (2008) found that learners who previously received training in analogical reasoning for one specific negotiation strategy (namely, logrolling issues to create value) did not perform well when confronted with a diverse, face-to-face negotiation with a very different structure. Thus, logrolling may have limited generalizability to other value-creating processes. To test this idea, Moran et al. adapted Thompson et al.’s analogical reasoning training to teach negotiators broad thought processes for creating value in negotiations. Moran et al. (2008) compared specific training, wherein learners compare two cases that illustrate the same specific strategy instances (e.g., logrolling), with diverse training, wherein learners compare two cases that illustrate different value-creating strategies (e.g., one illustrates logrolling and the other compatibility). Training effectiveness was
assessed by looking at performance and outcomes in a negotiation simulation with potential for various value-creating strategies, some of which learners previously had learned and others which they had not
Moran et al. (2008) found that more diverse analogical training, wherein negotiators learn and compare several different value-creating strategies, fostered greater learning of underlying value-creating negotiation principles than did more specific analogical training. This method facilitated transfer to a very distinctive task and improved performance on a variety of value-creating strategies, including some that participants had never previously encountered. Improved performance was accompanied by a deeper understanding of the potential to create value. Thus, more diverse analogical training can be effective for attaining greater expertise, which fosters understanding of which particular strategies might be effective in different situations and why. At the same time, when training becomes too diverse, the applicability of the message may be lost. The optimal level of abstraction remains an interesting question for future research, as does the question of how analogical reasoning can be applied to improve individual decision making.