Journal · Evidence

Cognitive diversity: what the evidence supports, and what it does not

We build on this idea, so we have a reason to overstate it. Here is the honest version, including the peer-reviewed argument that the best-known formal result is wrong.

Evidence · 6 min read

Does cognitive diversity improve group decisions? There is reasonable support for diverse problem-solving approaches outperforming homogeneous ones on complex problems, and real disputes about the strength of the formal results, the size of the effects, and the coordination costs. It is a defensible working hypothesis, not a settled theorem.

The most cited formal argument is the Hong and Page result, popularly rendered as "diversity trumps ability". In 2014 a peer-reviewed critique in the Notices of the American Mathematical Society argued the theorem was mathematically incorrect, trivial, or irrelevant to the computational experiment supporting it, and that the experiment was constructed in a way that ensured the desired result. Responses have contested the critique in turn. The debate is live.

The frequently cited magazine study

A 2017 article reporting that cognitively diverse executive teams solved problems faster is widely quoted as research. It is an editorially reviewed magazine piece, not a peer-reviewed paper, based on an analysis of around 150 teams using a commercial assessment instrument. That does not make it false. It does mean it should not be presented as established science, and we removed it from our own method page for that reason.

We build on this idea. That is precisely why we should be the ones telling you where it is contested.

What is better supported

Structured dissent beats consensus-seeking on decision quality, with the satisfaction cost noted (Schweiger et al., 1986).
Trainable reasoning methods improve accuracy, measurably, in forecasting tournaments.
People cannot self-report their own biases, which is a strong argument for bringing in reasoning you did not generate.

The costs nobody mentions

Diverse groups are slower, less comfortable, and need active facilitation to convert difference into a decision rather than into a stalemate. Left unmanaged, cognitive diversity produces a longer meeting and the same conclusion.

Five Peers editorial note. Sources named in the text.

A working hypothesis we are honest about, applied through methods that are not.

Compose my board