21

Is it true that "Companies with diverse workforces outperform those with homogenous teams"? That's what TrendsWatching says anyway. And is it "diverse workforce" or "workforce comprising people from a variety of backgrounds"? I don't immediately see what racial diversity per ce may have to do with performance. TrendsWatching references this report

enter image description here

Sergey Zolotarev
  • 1,197
  • 2
  • 9
  • 15

2 Answers2

26

Does diversity lead to more productivity?

Some diversity enhances productivity some doesn't. For example, age diversity Ilmakunnas & Ilmakunnas 2011), cultural diversity or knowledge/experience (Navon 2010) diversity among employees is related to increase in productivity.

On the other hand educational (Ilmakunnas & Ilmakunnas 2011) or linguistic diversity (Dale-Olsenab Finseraas 2021) is shown to negatively impact productivity.

These are some examples. A workforce diversity can be measured along plethora of multiple dimensions so no answer can look at every single possible dimension diversity.

I don't immediately see what racial diversity per ce may have to do with performance.

Neither the poster, nor the report to which you link, talks about racial diversity. It talks about homogeneous vs heterogenous teams. A team composed of same racial group can be heterogenous (i.e. diverse) along some other dimension such as education or age. Underrepresented group is any group that is not proportionally represented in some area relative to their proportion in total population, so for example in some industries that can be women, or Hindu etc.

Team heterogeneity can be measured among myriads of different dimensions (some examples: age, sex, gender, sexual orientation, experience, skill set, education, culture, language, geographic, nationality, religion, ideology etc...).

Community
  • 1
1muflon1
  • 56,292
  • 4
  • 53
  • 108
  • 14
    Maybe also mention that measuring productivity is non-trivial, and differs a lot between fields. Linguistic diversity was found to decrease productivity in manufacturing in the linked study, but may have a different impact in more communication-oriented types of work. – blues Nov 20 '22 at 19:24
  • 4
    I'd guess diversely educated teams are better than uniformly lowly educated teams and worse than uniformly highly educated teams. Not sure which comparison was done in that study. And if a "linguistically diverse" team means one where the people couldn't speak each others language, a worse performance is also somewhat expected. – Paŭlo Ebermann Nov 20 '22 at 21:08
  • 2
    How does cultural/ethnic diversity increase productivity? – Sergey Zolotarev Nov 21 '22 at 01:27
  • 1
    @SergeyZolotarev Intuitively, people with different life experiences are likely to have different heuristics (mental shortcuts), familiarity with different things, different attitudes and approaches to problems… and that means – at least, in thinky work – people are less likely to think the same thoughts as each other. That means less redundant work is done. – wizzwizz4 Nov 21 '22 at 02:56
  • 1
    But the data doesn't say that. The data shows there's a correlation, but the data doesn't say why. Experimental evidence can help you distinguish between multiple hypotheses, but at a certain point, it's only intellectual curiosity that motivates such research. (Not that there's anything wrong with that, but it doesn't really get you research grants.) We know how to increase productivity: hire diverse teams, and support them. – wizzwizz4 Nov 21 '22 at 02:57
  • @wizzwizz4 no data do not show just correlation. Data can show there is causal relationship between two variables. 1. Experimental evidence is also data it is just data generated in a special way to make sure untreated group is by construction valid counterfactual for treatment group. 2. Even observational evidence can yield data that allow you to make causal inference. For example, quasi-experimental research techniques such as IV, RDD, DiD or Synthetic control, or natural experiment. 3. Even good old OLS results can be given causal interpretation as long as you can show all assumptions hold – 1muflon1 Nov 21 '22 at 09:57
  • 1
    @1muflon1 Data do just show correlation. If you assume the correlation isn't coincidence, there is some causal relationship – but other than that, you've just got hypothesis testing, statistical tools and good old-fashioned elbow grease to figure out what's going on. Those techniques you've described allow you to infer a causal relationship – but only with additional assumptions. (Remember the classic "ice cream causes shark attacks" problem – high correlation, but no causal relationship; there's a third factor in the causal graph.) – wizzwizz4 Nov 21 '22 at 11:37
  • @wizzwizz4 1. you don't need to assume correlation is not coincidence. I don't know of any model that would assume that. Not even randomized control trials have such assumption. I would like to see any peer reviewed source for such claim. What you need to assume is that control is good counterfactual for treatment which is not strong assumption with good research design. 2. the graphs you are referring to are not even examples of a third causal factor. They are all examples of stochastic trends. Nobody claims that correlation automatically equals – 1muflon1 Nov 21 '22 at 11:53
  • causality. However, under modern understanding of science, the way how we uncover causal relationships is to test hypothesized relationships against observations (i.e. data) and eliminating ones that can be rejected by the data. Randomized controlled trial just generates data as explained above. Even physics laboratory controlled experiments are just ways of generating data to confirm or disprove that some relationship exist. – 1muflon1 Nov 21 '22 at 11:55
  • I recommend you read this especially the second half https://www.jmp.com/en_nl/statistics-knowledge-portal/what-is-correlation/correlation-vs-causation.html – 1muflon1 Nov 21 '22 at 14:19
  • Let us continue this discussion in chat. – wizzwizz4 Nov 21 '22 at 16:00
  • The OPs question specifically asks about racial diversity improving productivity. Cultural diversity is not the same as racial diversity. – carrizal Nov 22 '22 at 15:52
0

According to Nancy Adler and Tony Byers, there is a big difference between living an inclusive diversity and just having a diverse workforce.

In general, if the team members are well intentioned, there is clear commitment by management, everyone is welcome to share their ideas, and there is a good balance of everyone's interests, diverse teams bring so much more to the table, especially in creative jobs. (Repetitive and highly standardized jobs may have a slight advantage in mono-cultural teams because of less communication "friction").

During each project, teams will go through different phases. During the divergence phase, teams explore new ideas and innovative solutions. Diverse teams can develop more ideas, more diverse ideas, and they can include the perspectives of more different customer groups (access to new markets). More out-of-the-box thinking occurs and leads to more innovation.

During the ensuing convergence phase, teams will have to decide which ideas will make it into the final course of action. If not managed well, this phase may lead to some friction, misinterpretations, more disagreement, miscommunication and high stress levels. Diverse teams therefore should be allowed more time to find common ground initially.

If left unmanaged, the benefits of diverse teams can drain quickly. When team members feel their ideas are being ignored or the treatment and job opportunities are unfair, this will lead to resignation, disengagement, and a low identification with the organization. For the organization, this leads to low talent retention and missed opportunities. The daily humiliation sags like a heavy load, and it can get much worse. This is not an argument against diverse teams - it is an argument for the value of good leadership.

If managed well, diverse teams will outperform mono cultures.

naughtilus
  • 27
  • 1
  • 7
    This does not contain any hard evidence. And reading through your post, I mostly see a comparison between well managed and badly managed teams with the implicit assumption that a diverse team will be better managed – Manziel Nov 21 '22 at 12:48
  • Not my assumption. I agree that diversity and quality of management are independent variables. The research done by Carol Kovach / reported in Nancy Adler's International Dimensions of Organizational Behavior indicates that cross-cultural teams that are led poorly and have low cross-cultural competencies have the lowest performance, followed by mono-cultural teams. Cross-cultural teams with effective management and high levels of cross-cultural competencies have the highest performance. – naughtilus Nov 21 '22 at 23:10
  • Not totally independent, though. While diversity itself does not magically improve management, bad management and a lack of commitment to diversity & inclusion can impair an organization's ability to acquire and retain diverse talent. – naughtilus Nov 22 '22 at 00:17