By Josh Hemann
A casual conversation with a friend in the biotech industry got around to the subject of “PhD bias”: If you are working with a group of PhDs, they’ll want to know that you have paid your dues just like them. And if the highest level of the academe is not part of your experience, you'll certainly have to work harder to prove yourself. Having been in such settings over the years, I know this bias exists, and I wish there was a catchy phrase to label it.
Over these years I have also picked up on a particular flavor of anti-PhD bias in business settings. I'll describe this bias by way of a mock conversation that I have essentially witnessed multiple times...
Retailer: So, what do you think about approach XYZ for solving our business problem?
PhD: Well, actually, I would not use approach XYZ. Your problem is just a specific instance of a more general problem that is better solved by an approach called something-really-complex-sounding.
Retailer: OK, but our business actually is really unique, and we have to move fast here, so we think approach XYZ will meet our needs...
PhD: Sure, but years of academic research has taught me how to abstract away specific issues to tackle more general problems, and trust me, there is nothing special about your business. Something-really-complex-sounding is a more elegant, general approach that I strongly advise you take...
At this point, the retailer is probably thinking that the PhD has not paid her dues: she has not worked the routine 70 hour weeks, dealt with weekend fire drills because of a bug in the Point-of-Sale (POS) systems or a coupon that went viral, or had to adapt to constantly changing direction from executives as weak sales are further eroded by new competition. Having been on that side of things, I can commiserate with the retailer because even if it is true that their problem is not that unique, no one wants to hear that domain knowledge from the industry they have worked in for decades is not super relevant to solving the problem at hand. Thus, in such settings, I find that PhDs sometimes have to work harder at proving themselves as practical people with "real world" experience.
Is there a similar notion of bias and paying one’s dues in Big Data settings? I suspect there is in certain crowds. Somehow, working with terabytes of data (not huge by today's standards but certainly big enough to cause trouble) in an RDBMS with point-and-click analytics doesn't seem to garner as much street cred as working with the exact same data stored in HBase and analyzed with Hadoop and R. Some of this is just due to the hype cycle we are presently in, but some of it is warranted: Data scientists working in the latter stack have to know more hardware and software details to get the job done.
We all have our tribes and intrinsic notions of what it takes to be a member. I think the Big Data era upon us will make those notions fuzzier and the tribes more fractured.