The behavioural approach to organization and decision making argues that negative feedback regarding a choice deters someone from doing the same thing again. That is, decisions should be guided by the law-of-effect—a concept central to adaptive learning. On the other hand, escalation research shows that decision makers often continue or reinforce their commitment to a previous decision after receiving negative feedback. These contradictory patterns form what is known in organizational research as a learning–escalation paradox.

Providing a theoretical basis to resolve this paradox, a recent paper by Wong and Kwong explains that while the law-of-effect has traditionally been used to explain decision-making at the decision level, it can also be extended to a group of planned decisions—the strategy level.

The authors define an ‘escalation strategy’ as organizing decisions according to the escalation of commitment pattern, and a ‘law-of-effect strategy’ as organizing them consistent with the law-of-effect pattern. They show that although decisions made under an escalation strategy might seem to be inconsistent with the law-of-effect pattern when evaluated separately, when viewed holistically at the strategy level the escalation strategy follows the law-of-effect.

The study finds that at the start of a learning process, people will underestimate the inherent value of the law-of-effect strategy. Moreover, the chances of correcting this underestimate are small, since the likelihood of choosing the law-of-effect strategy decreases over time due to the associated negative consequences. As such, when decisions are derived from the strategy level, the escalation strategy is preferred. This, the authors explain, is because “people prefer options that are subjectively more favourable, thereby developing a pattern of behaviour that is contradictory to the law-of-effect at the individual level, but consistent at the strategy level”.

Designed to capture the various aspects of escalation commitment, the study demonstrates this learning pattern through the use of three simulations of different combinations. The simulations show that the escalation strategy, which organizes decisions in ways that resemble the escalation of commitment pattern, will be developed to be more favourable than the law-of-effect strategy, which organizes every associated decision to follow the law-of-effect pattern.

In addition to resolving the learning–escalation paradox, the study also offers a novel theoretical perspective of escalation, termed the adaptive learning approach by the authors. Existing literature has paid little attention to the origin of escalation biases. Traditional approaches assume that decision bias is rooted in a deficiency in human cognition or individual motives, such as not wanting to admit failure. Instead, the authors suggest that decision makers do not initially have an escalation bias; rather, “escalation and errors in decision making are developed naturally and may occur without distorted motivation or cognition”. The bias is acquired gradually as a result of an adaptive sampling of escalation strategy over law-of-effect strategy. Over time, people learn to repeat the escalation strategy, which further sustains their biased sampling.

When running their simulations, the authors point out that they made a number of assumptions. First, that everything was equal between the two strategies. In reality, other factors may come into play. Second, they assumed that decision makers have no personal preference for either strategy, when, in fact, some may initially favour one over the other. Third, the authors presume that there is no transaction cost if individuals cease using either strategy. Finally, the study assumed that decision makers are able to freely choose a particular strategy based on what they have learned about their reinforcement values. Again, this may not hold in organizational contexts where decision makers are not able to freely make their own decisions.