Authors Cass Sunstein and Reid Hastie have put together an excellent and readable taxonomy of the “bugs” that can afflict group decision making and cause groups to blunder in their book Wiser.
In the first part of this blog I discussed their dissection of the causes of group decision making failure.
How can we make groups function better?
The book carries a number of actionable insights into how to improve group decision making for the better:
Inquisitive and self-silencing leaders
Men out-talk women two to one in jury deliberations, the same happens in companies. Leaders and other high-status member of groups will exert a lot of influence, they can do the group a favour by remaining silent and indicating a willingness to hear all of the uniquely held information. That helps overcome one of the main issues identified with groupthink – self silencing which leads to a failure to share critical information.
Priming critical thinking
Rather than prizing consensus, give social reward to new and competing information.
The role of roles
Information aggregation becomes more likely with roles, as each member knows that the others have something to contribute, a division of labour in examining a problem will make hidden profiles less likely.
Shift perspectives:
“If a new CEO came in, what would they do?”
Does devil’s advocacy work?
It has become commonly thought that devil’s advocacy or can help get past groupthink influences. The authors are torn on this point. Devil’s advocacy as an idea is trying to formalize the commitment to expressing differing viewpoints (which is a good thing). Those assuming the devil’s advocate role are able to avoid the social pressure to agree. However the authors believe evidence is split on this – there is a difference between authentic dissent and a formal requirement for an assigned devil’s advocacy, who may be arbitrarily assigned and simply “acting out a role”. For this to work the devil’s advocate has to actually mean what they are saying.
An upgrade to devil’s advocacy, which has a greater chance of working is to get an entire team to work on a contrarian viewpoint, with the aim of defeating the primary team’s plan to execute a mission. so called, “red-teaming”.
The Delphi Method
This is a formal approach to aggregating individual views. It proceeds in a number of rounds. First round votes (or estimates) are taken, in complete anonymity. The second round estimates must all fall within the 25th – 75th percentile of the first round estimates. This process is repeated until the group converges on a single estimate. It’s an averaging process, but one that allows for a single stubborn (or convicted) group member to influence the final outcome a lot more than a simple average.
Separate Identification and selection.
The qualities that make a good identification process, particularly diverse and divergent thinking, are very different to a selection process which needs to favour convergent thinking. All too often these steps can be compressed together, making it difficult to juggle the requirements for both divergent and convergent thinking.
Harnessing experts
I found this section particularly interesting and relevant!
A few tips on how best to use experts:
- Obtaining a statistical answer from a few of them, rather than relying on just one
- Limit experts to domain areas where there is evidence that expertise gives an edge (eg how to combine asset classes to give a diversified portfolio as opposed to whether the stockmarket, or interest rates, will go up or down tomorrow or next month)
- Look for track records from experts (weight individual expert views by the track record)
Specifically for investment advice this points toward asking the questions
“What other strategies were considered, but ultimately didn’t quite make it into this advice?”
“Is this advice the work of one, or many, experts?
was this advice produced in an environment of challenge among experts, how were the different expert inputs incorporated and weighted (process)?
I for one will certainly be taking away a few of these insights and trying to apply them more consistently.
What do you think is the most helpful?
One thought on “[Books] Wiser (part 2) – getting past groupthink to better outcomes”