This morning I saw the following headline on a Wall Street Journal blog:
That’s an interesting assertion, so I read the study to see how the authors had arrived at that conclusion:
in comparing successful versus unsuccessful companies, the overall median proportion of female executives is 7.1% and 3.1%, respectively, demonstrating the value that having more females can potentially bring to a management team
This is a questionable paragraph. For the sake of the argument, let’s suppose the study said this:
“In comparing successful versus unsuccessful companies, the overall median proportion of Porsches in the parking lot is 7.1% and 3.1% respectively, demonstrating the value that having more Porsches can potentially bring to a management team.”
Even though Dilbert’s boss might use this as an excuse to ask the execs to buy him a German car, Porsches are most likely a consequence of success. Therefore, the presence of Porsches in the parking lot doesn’t demonstrate that they add value. All it does is show a correlation.
Note that I’m not suggesting that this is the case with women executives! The Porsches in the parking lot are a caricature example to point out the flaw in the conclusion. Another paragraph makes the bias more evident:
The thesis for this paper claimed that having a higher proportion of female executives at venture-backed start-ups improves the ultimate success of the company
If you read the paper, you will find that the data doesn’t support that assertion. It is true that companies with more women are more successful. It doesn’t follow that the women improve the ultimate success of the company. The honest truth is that we don’t know what makes the companies more successful, and the authors blithely jump to their desired conclusion. Let’s look at other plausible explanations consistent with the results:
- In baseball, it is known that having the best players is strongly correlated with success. It would be reasonable to suggest that the same could apply to venture-backed startups. What if some companies are more likely to hire men rather than women, even if those men are not as qualified for their jobs? Those companies would be at a statistical disadvantage with respect to those who hire the best person regardless of gender. In this case, the title of the study could be “Discrimination Makes Companies More Likely to Fail.”
- It’s also true that most companies are started by men. From the study:
1.3% of privately held companies have a female founder, 6.5% have a female CEO, and 20% have one or more female C-level executives. The most common positions held by female executives were within Sales & Marketing roles, accounting for 27% of the total population sample.
Now, let’s look at a typical company started by men. Let’s assume that it’s relatively likely that the company will fail before hiring a female executive. In this case, we are looking at survivor bias: the companies that hired female executives were already more successful than most. Let’s also suppose that Sales & Marketing executives are only needed by companies that are relatively advanced in terms of product/market fit, and that many companies fail before getting to that stage. This would add even more survivor bias.
Note that I’m not making any assertions about whether the above explanations are true or false. I’m simply showing that they would be consistent with the data analyzed by the study, and that they have not been ruled out.
The sad thing about bad science and journalism of this sort is that it does not help to create a more egalitarian society, or to right a wrong. Instead of being honest and wondering what’s missing from companies that hire fewer women, it irresponsibly asserts that hiring women executives will make a company more successful.
I humbly ask the authors of the study to revise their hypothesis, and to try to get to the bottom of the issue. For example, how successful are companies at the stage in which women are hired as executives? Is it possible that successful companies are more attractive to female executives than unsuccessful ones? Perhaps successful companies have a higher interest in hiring women as executives?
If you have any other questions or hypotheses, I’m interested in hearing them.