Why more diverse groups make better decisions
From whether to have another cup of tea in the morning to choosing whom to employ - decisions are integral to every moment of our lives. But how do you ensure that you always make the best decision possible? A review of the literature out this week has revealed certain dos and don’ts - and one of the dos is to make your panel - or committee - as diverse as possible. Georgia Mills made the executive decision to speak to lead researcher Dan Bang, from University College London, and find out why diversity makes for a good decision…
Dan - We started by asking what characterises a good decision, and how do we know which action is the best among all the available options? Then what we did was to apply insight from psychology and mathematical models of decisionmaking to try to unpack this process into component parts, if you like.
Georgia - So you were looking at previous studies, and previous real world examples of decisionmaking where we knew, in hindsight, what the best decisions were, and then you analysed this to try and search for common problems, or common good things that people were doing?
Dan - Yes, exactly. So you identify a bad decision and then you try to reverse engineer what went wrong. This is where models of decisionmaking are very useful because we can ask what went wrong and at what stage of the process it went wrong.
Georgia - What did you find?
Dan - What is important for understanding the main findings of our review is a metaphor for decisionmaking that we often come back to and that involves hilly, misty landscape. You can think of each point in the landscape that is corresponding to the different actions that we can take at a given moment in time, and then the landscape contours indicates how good each action is. So what you have to do as a decisionmaker is to find the highest peak in this landscape but, because of the mist, this process is slow and taxing. And often we will wrongly believe that we have found the highest peak even when, in reality, we are standing on a small bump.
The benefit of group decisionmaking is that more people are involved in the search for the highest peak and, therefore, more likely to find it. However, if our review we uncovered several social dynamics and biases that stops groups from finding the highest peak. One problem that we identified was the lack of independent knowledge, and the reason why this was problematic is that then the benefit of combining information from multiple people is massively reduced.
Georgia - Oh, I see. So, if there’s one of you searching around this hilly landscape, you’re maybe going to find the highest peak but it’s a lot better if you have a lot
of people searching, but if everyone thinks in the same way it may as well just be one larger person searching round this hilly landscape?
Dan - Exactly. If we think in the same way we’re likely to start searching in exactly the same place. People have to be different in different ways, if you like. So what we do is that we distinguish between different kinds of diversity, we think of identity diversity which is individual differences in personal characteristics. Characteristics that we can all see, such as gender, age, cultural background. But then there’s also a more subtle form of diversity which is functional diversity and that is differences in how people think and solve problems.
If we have people from diverse backgrounds, then they’re not likely to share the background knowledge, and they’re also not likely to acquire new information in the same way. So in this way they’ll start up in different points in the landscape.
Georgia - With large groups making decisions you have more expertise, but then there’s this idea that you just spend all day talking, and then there's this kind of inertia of having too many people rather than one person just having a vision and saying “right, we're doing this and we’re getting this done”.
Dan - Absolutely. But I think when you make a group decision, what’s important is to acknowledge the tradeoff between speed and accuracy. Often, fast decisions are not the most accurate, whereas the most accurate decisions tend to be slower. In fact, what we uncovered in the review is that discussion is actually a very useful thing to do. What you often find in larger groups is that they then decide to vote on what they should do. This is not a very good idea because this throws away a lot of useful information. So if you want to combine votes from different people then you need to have some mark of reliability in place and you need to weight the words accordingly. You should assign more weight to an opinion that is based on better information.
Georgia - So not everyone in the group is equal?
Dan - No. We all think that expertise should play a role in group decision making. However, we are remarkably bad at recognising expertise. There’s this finding in the literature called Dunning-Kruger effect, that the worse you are at a task, the more delusional you are about your own ability. So, surprisingly, it acquires expertise to know the limits of your own knowledge. People are very good at something are not sufficiently confident, so they take other people’s opinion into account too much.
Georgia - Beware of the person with the loudest voice then?
Dan - Definitely.