Weighing Up the Options: Escalation of Commitment

[Sponsored article] Assume you are the head of R&D at a high-tech firm that develops new accessories for cell phones. One year ago, $200 million was budgeted to develop a new cell phone battery instead of developing a new cell phone screen. To date, $100 million has been spent, and the project is 75% complete. Then the company hears that a competitor has just marketed a cell phone battery that is cheaper and has a longer battery life than yours. How do you allocate the remaining $100 million between the cell phone battery project (improving durability, reducing weight, and reducing pollution) and the cell phone screen project (improving brightness, increasing display resolution, and improving color display)? Now assume you are not the initiator of the project but just one of the team – how does this change your decision making?

Decision makers often exhibit a decision bias called “escalation of commitment”. This occurs when they receive negative feedback on a previous investment and they have to decide whether to invest further to recover the sunk costs or to terminate the investment to avoid further losses. Under these so-called “escalation situations”, decision makers tend to continue with or even increase their investment in their original failing course of action, particularly when they are responsible for making the initial investment. In other words, given the above scenario, the R&D head would continue to focus on the battery project.

Researchers Jessica Y. Y. Kwong and Kin Fai Ellick Wong set out to examine how escalation of commitment can be either be reduced and exaggerated by framing the information differently through “option partitioning”. As an example of this, we can either invest in a “general class”, which is a broad group such as “electronic products”, or we can divide the investment into “suboptions”, such as MP3 players, portable TV game consoles, or other specific types of electronic products.

Partition priming is key to understanding this research. A fundamental idea in decision making is that the attractiveness of any option – that is, its expected utility -- is a joint function of the probability that the option will materialize and the corresponding subjective value if that option indeed materializes. In the escalation context, decision makers often assess the probability of success for each option and their corresponding subjective values. According to the idea of partition priming, these probability estimates are not fixed but are malleable depending on how the options are partitioned.

The researchers came up with four hypotheses. First, people will escalate less to an escalation option when the alternative option is partitioned into multiple sub-options than when presented as a single class. Second, people will escalate more to an escalation option when that option is partitioned into multiple suboptions. Third, the effects of option partitioning on escalation of commitment are mediated by escalation utility. And four, the effects of option partitioning outlined in 1 and 2 are independent of and go beyond the effects of personal responsibility.

These hypotheses were tested with a series of four experimental studies involving groups of undergraduates faced with a variety of scenarios such as the cell phone battery case outlined above.  The findings demonstrate how commitment escalation is affected by option partitioning. Escalation tendency was suppressed when an alternative option was divided into subparts, whereas a reversed trend was observed when it was partitioned. The research also suggests a new direction for explaining why escalation of commitment occurs: as partitioning increases people’s tendency to stick to that option, it is possible that escalation occurs because decision makers are likely to partition the escalation option rather than the alternative option. Self-serving and self-justification motives lead people to focus on that option, presumably because they want to look for information consistent with previous belief.