Returning to free-lance work after a year-long (interim) assignment, I also took some time for reading. More particularly I have read ‘The Difference: How the Power of Diversity Creates Better Groups, Firms, Schools and Societies’, 2007 by Scott Page.
In the prologue Page claims his book will offer a logic of diversity and show that diversity improves performance and, under certain conditions, trumps ability. It takes the first part of the book to, quite convincingly, argue and illustrate that cognitive differences improve the group’s capacity to solve problems. Cognitive differences are the basis for different predictive models. Page shows that, what he calls “a diverse crowd of models” always outperforms a single expert, even if the individual models that are part of the crowd are not as good as the experts’ model. For this claim to hold, it is required that the crowd members have at least some relevant abilities to contribute towards solving the problem, and that the problem the group deals with, is rather complex. Interestingly, these conditions are relevant to most management teams, hiring committees, juries and boards of directors.
Page’s claim is based on a link between cognitive differences and different predictive models. Predictive models can be complex or very simple (“red cars go fast” is the outcome of a – not so accurate – predictive model that links a car’s color to its speed). Critics of his book say Page does not adequately treats the issue of communication within his “diverse crowd of models”. Indeed, he does note that different (cognitive) perspectives on the same reality can lead to quite some miscommunication, but most of the time he seems to abstract from such problems.
While diverse predictive models are beneficial for the accuracy of predictions and create “wise crowds”, a major complication arises when these crowds are to actually decide on problems. This is because of their diverse personal preferences. These preferences are dealt with in the third part of the book. Page argues that members of a group tend to make different choices because they pursue different goals (different fundamental preferences) or they think – on the basis of their individual predictive models – that certain goals are best achieved in specific, different ways (different instrumental preferences). Too much diversity of preferences may lead to circular outcomes of decision-making processes, misrepresentation or strategic voting and agenda manipulation. Agenda manipulation is more of a risk when groups have diverse fundamental preferences because there are fewer incentives to be open about the agendas.
So much for diversity you would say. Nice for predicting outcomes, bad for decision-making. But that would be too quick a conclusion. Diverse preferences create problems but also have benefits. The fact that diverse fundamental preferences are linked to diverse perspectives, a major ingredient of cognitive diversity, mitigates their negative influence. Diverse perspectives are very beneficial to problem solving and innovative capacities of groups and make the group into a “wise crowd”. Members of the group can use the same problem solving technique (f.e. rank cases based on a hypothesis about a proxy for a single variable and test cases for their variable of choice until they are convinced they have a solid top 5). At the same time their preferences determine their perspective, the variable they use to rank those cases. This will lead to each having a different top-5 pick. A benefit of their diverse perspectives is that the group will not likely stick to a sub-optimal solution: what is good for one is not good enough for another group member and they will continue to improve on solutions (provided they communicate well…..)
So what is there to be learned from all this? To sum up Pages conclusions:
- Cognitively diverse societies, cities and teams perform better than homogeneous ones – diversity produces benefits
- Particularly fundamental preference diversity creates problems – people do not get along and public goods tend to be underprovided
A – likely – combination of cognitive diversity and fundamental preference diversity in a group helps to produce better outcomes, but it also produce more conflict.
So theoretically diversity has benefits and downsides. Do these actually add-up in real life? What is the net result? What does the evidence tell?
To look at the empirical “evidence” you need to be able to link the diversity we can see with real cognitive diversity. When we talk about diversity we normally mean “identity diversity”. People differ in terms of age, cultural and ethnic background, race, gender. Cognitive diversity is related to, but certainly not equal to identity diversity. Identities influence experience and opportunities, which in turn influence cognitive toolboxes, but assuming any linear relation would overly simplify the connections. And, the connections depend on the problem.
Reviewing a vast collection of empirical studies about identity diversity, Page claims they support that, when diverse identities are connected to cognitive differences that are relevant to the problem, diversity tends to have net benefits. His review of the empirical evidence ranges from studies related to firms and organizations, to studies of societies that look at linkages between ethnic and linguistic diversity of countries and their economic growth and history. For example, one study referred shows that within the same company, branch offices with more gender diversity generated more revenues etc. Page refers to a considerable variety of studies. Most of his evidence fits his model, not such a surprise, but it is indeed impressive. Too much to summarize here. He also includes interesting side-tracks: he reasons that the European Union is a way to harness diversity to produce benefits. Interesting perspective indeed.
The book comes with a warning that it contains a lot of mathematics and I did struggle at times. But then again, I find it equally daunting to read about diversity’s dialectically constitutive relationship to the discursive structure… (see call for abstracts, Critical Management Studies Conference 2009).
I will stick to freelancing for organizations and networks. Organizations can be supported to take some risks: there is no need to always recruit look-a-likes into their management teams, committees and boards of directors. Teams may need support to draw on the differences and allow them to inspire creativity and better predictions. And at the same time you can unpack and discuss with them their diverse preferences, separating goals from means. Teams can prevent diverse preferences becoming a source of conflict and make them generate new perspectives on complex problems. In his book, Page sometimes goes into details that I find hard to digest and sometimes he cuts corners where I dwindle in doubt, but I like the way he carefully un-wraps diversity, very insightful.