I saw the article on

I felt obligated to criticize this article promoting the method of using genders and races to promote workplace diversity. My original comments were submitted in a rush a few months ago. I put a little polished version here:

__________________

I am stunned by the publishing of this article by a prestigious academic magazine from ACM.

This article first defines diversity as "heterogeneous in meaningful ways, for example, in skill set, education, work experiences, perspectives on a problem, cultural orientation, and so forth". The

author did not (dare to) list genders and races in her definition of meaningful heterogeneous qualities. But for the rest of the article, all she was talking about is data about diversity defined by genders and skin colors, such as "19.2% of Ph.D.'s were awarded to female candidates" and "5.3% of BS degrees were awarded to African American candidates". I cannot find a single data source studying diverse skills, education, work experiences and so on. I cannot help but wondering: is the author assuming people with the same gender will automatically not be diverse? Is the author also assuming people in the same skin color can not be possibly diverse in their qualities?

If the author really believes in her definition about diversity in qualities of people and wants to fight stereotypes based on genders and colors, the presented data would be totally different and gender/color blind. For example, in the case of testing how innovative a team is, the meaningful heterogeneous qualities would be the different perspectives for a domain, the different depth of knowledge for the problem, the degree of risk averse in the team members, the different thinking patterns like top-down and bottom-up, the preferences between incremental and radical changes, and so on. All these qualities are far more meaningful than the superficial, stereotyped gender and race differences in the team members.

The author also seems lack the basic understanding of (or knowingly ignore) why stereotypes happen at all. I would point to the concept of associative memory used by human brains. When two concepts often come to a human brain, the connection between these two concepts will be enforced. Later, mentioning one concept will automatically trigger the onset of the other concept in our minds. Essentially, neurons (brain cells) send signals to each others. The more signals sent

between two neurons, the stronger the connection grows. Our brains re-wires physical structures based on the input. For example, if we often see a suspect wearing a red hat committing burglaries in the news. Our brains will automatically associate red hats with burglaries and remind us to be cautious when seeing a person wearing a red hat. These kind of associations by themselves are just our evolutionary intuitions, which are not necessarily good or bad. Definitely we should not feel ashamed by these intuitions at all. What matters is how we use our logic and reasoning to make decisions from intuitions. Rather than weakening the stereotyped associations based on genders or races, what the author suggests (using genders and colors to measure and increase diversity) would in fact further strengthen the exact associations/stereotypes the author claims to fight against. In the end, the author is promoting gender and race based discrimination, at least in my opinion.

Maybe the biggest flaw in the article is something the author already admits: all cited studies about the benefits of gender/race-based diversity were correlational. None of them were designed to prove causal relationship, which I doubt they could ever be unless we use meaningful qualifies to re-design the studies. Still, while the author says the results only "to a degree that demands attention", the actions promoted are very alarming to me: acting now using gender/races to improve diversity even when there is no controlled, conclusive studies showing causal relation between gender/races and job outcomes.

I believe social policies to a society are like medical drugs to a human body. They must pass rigid controlled studies before fully pushed for implementation. Having correlational, observational studies are far from rushing to adapt the policies (drugs). The consequences can be very severe.

I don’t know how this article could pass the peer review. I am very disappointed by the publishing of this article and also disappointed by ACM.