2019 brainfood biasindiversity 883x432 1

Bias in diversity

28 October 2019

The article at a glance

Diversity on key corporate committees is supposed to produce better decisions. But a study at Cambridge Judge Business School finds that diverse …

Diversity on key corporate committees is supposed to produce better decisions. But a study at Cambridge Judge Business School finds that diverse views rarely ‘average out’ as assumed, but instead lead to systematic biases toward certain types of errors.

Geometric seamless pattern.

Diversity is usually applauded as something that provides a broader and more balanced perspective, so companies often seek a variety of backgrounds on top committees that make key decisions such as project selection. The assumption is that diverse views will “average out” to arrive at a sound conclusion through the “wisdom of the crowds”.

Yet such a common assumption is usually wrong, concludes a study at Cambridge Judge Business School published in the journal Manufacturing and Service Operations Management.

Instead, the research finds that diversity can lead to “systematic biases” toward certain types of errors in project selection. That’s because the “diversity” on such committees often reflects preference diversity (committee members’ own different perspectives about a proposed project’s market value) rather than interpretive diversity (the different emphasis that committee members place on the project’s likelihood of success based on information that arises during the project execution).

“This distinction is crucial,” concludes the study co-authored by Nektarios Oraiopoulos, University Lecturer in Operations Management at Cambridge Judge, and Stelios Kavadias, Margaret Thatcher Professor of Enterprise Studies in Innovation & Growth at Cambridge Judge.

Dr Nektarios (Aris) Oraiopoulos
Dr Nektarios (Aris) Oraiopoulos

“Higher preference diversity always leads to higher likelihood of making the wrong decision. Higher interpretive diversity instead can become a powerful lever improving the project decision-making process,” says the study, based on earlier fieldwork in the pharmaceutical industry, entitled “Is Diversity (Un)Biased? Project Selection Decisions in Executive Committees”.

The study finds, further, that committees with higher preference diversity are more likely to reject good projects that would have succeeded and also more likely to accept projects that are bound to fail. This conclusion was reached by examining the two fundamentals errors of decision-making: the probability of selecting a project that eventually fails, and the probability of terminating one that would have been successful.

Interpretive diversity reflects different “thought worlds” among committee members, and this allows them to filter new information about a project in truly different ways. Preference diversity reflects “differences in the perceived opportunity costs and outside options” faced by each committee member – and this introduces those committee members’ “own biases and objectives”, which might challenge the decision-making process.

Professor Stylianos Kavadias
Professor Stylianos Kavadias

The key implication of the findings is that managers need to identify and reduce preference diversity, which can be achieved through greater transparency in business unit pipelines including the resources required for a particular project. This can lead to sounder decisions on project selection, a vital part of many businesses such as pharmaceuticals in which firms must often decide whether to continue or terminate risky and novel projects.

A 2011 study found that nearly 30 per cent of pharmaceutical compounds are terminated at an early stage for strategic reasons, such as low expected financial returns, rather than due to efficacy or safety issues.

A recent report by the Boston Consulting Group, cited in the study, finds that members of executive committees have a natural tendency to prioritise the interests of their own business units – and that the risky nature of research and development initiatives allow those members’ interests to remain undetected, particularly when decisions result in project termination.

“The study shows that diverse perspectives about a project are rarely averaged out, but rather that diversity leads to systemic biases in decision making,” says co-author Nektarios Oraiopoulos. “Managers therefore need to better understand the sources of diversity, to determine whether such diversity reflects already formed individual preferences or new information that arises during the project.”