I don’t see that this article has reached its conclusions. With the intent being a general conclusion about the state of the industry for the variables in question, vs. Google’s state, possibly narrowed to a particular part of Google (org, team, etc.) that they could focus on with some depth. Google is not every company, and Google is also made up of multiple parts.
The only thing this has done is signal that the particular reviewers and particular team/org/whatever structural component of Google has some biases, which is Google’s problem (and everyone else’s who looks at this and says “wow, that seems like my team”). I’m not saying that this research is bad, unfounded or wrong, but that you need a larger cross-section of the biases from multiple companies in order to form a proper conclusion. We need more data, more studies, and the methodology to deal with all of that.