The paper is recently published but it is old data from 1995-2002. The data set is adequate but narrow, some sixty offices of one professional services firm with 2-19 employees per office. Old data and narrow data set means that it is questionable how robust the results are or the extent to which it can be extrapolated to larger companies, currently, in different industries.
So at best, the conclusions are suggestive rather than reliable.
The study doesn't appear to look at racial diversity, just gender diversity.
It found employees are happier when they work with people of the same sex. The slightly puzzling flip side? Single-sex workplaces aren’t nearly as productive as those where men and women earn their livings side by side.I don't find this puzzling at all (confirmation bias). Human systems are an aggregate of unexamined trade-offs and that there might be a trade-off between increased diversity and happiness is not unsurprising nor that there should be an association between diversity and productivity.
I think too often researchers ask shallow questions: Does diversity lead to greater productivity? That is certainly a valid question but an incomplete one. The evidence is still mixed - some studies indicate that increased diversity leads to greater productivity and some studies indicate that it reduces productivity. Likewise with happiness.
The more pertinent question, I believe, is "Under what circumstances does increased diversity lead to greater productivity?"
Real diversity (religion, occupation, age, education, socioeconomic status, etc.) has real costs to it. Communication is an obvious one. Language and speech are weighted down with unexpressed assumptions and expectations and if you have many different people from many different backgrounds, it takes a while for people to calibrate with one another and even when they have done so, the probability of miscommunication remains much higher than among people with a shared background. This impacts efficiency (amount produced in a given time period with given resources). Efficiency is also reduced because diversity requires a much greater expenditure of management and executive time in coaching and dealing with the consequences of miscommunication. So that's the cost of diversity, what's the value? The value is inverse of the cost. People with different assumptions have different perspectives and different interpretations. With all those differences, you have a wider variety of hypotheses and have to have a more rigorous vetting.
In some circumstances, the cost of diversity will exceed the benefits, on others the benefits will outweigh the costs. So what are those circumstances. Just a couple of examples.
If you are in a very cost competitive and/or a very fast moving but stable environment, it is likely that homogeneity will be of greater value than diversity. The fact that there are homogenous assumptions is not such a big deal because it a stable environment and there is benefit from the efficiencies of seamless communication and the rapidity that comes from sharing common assumptions.
On the other hand, if you are in an industry subject to rapidly changing technology, regulations, customer expectations, etc. and one that is plagued by significant unknowns and uncertainty, then you are likely to find that having a diverse employee base is likely to be more beneficial. You don't know what you don't know but if you have a diverse employee base who bring a wide range of knowledge, assumptions, and experiences, then it is likely at least someone can speak to the emerging issues and the discipline of precise definitions and conscious vetting of hypotheses will also likely bring greater benefit.
I don't think that the issue is whether you should have diversity or not, the question is to identify under what conditions is diversity most likely to make you more viable. You have to survive and you have to make money - how does diversity facilitate those goals.
By making diversity a goal unto itself, you ignore reality (revenue has to come in and it has to be greater than your costs) which is always a dangerous thing to do.
Diversity as a strategic condition for success then fits into the other finding which the researchers found puzzling enough to remark upon.
The study also found that employees felt good about working for a company that cared about equality between the sexes — as long as that caring didn’t translate into action.The first line is a classic example of revealed preference. We like something in the abstract but often have different views on it when it is an experienced reality. We all think there ought to be better public transportation for other people to use so that my car commute is faster.
“They liked the idea of diversity more than they liked actual diversity,” Ellison said.
And what is the likely problem with diversity? Probably friction and tension. It is unsettling to have to confront someone about their beliefs or to deal with someone challenging our own. And yet that friction is what generates better decisions and greater productivity.