- 4/16/2014 11:29:27 AM
@Randeroid, I suppose we can call your thinking cynically appropriate! Perhaps that's good reason to make "data science" a team sport, with lots of collaboration among the business owners, analytics professionals, and other key stakeholders, or to set up an Analytics Center of Excellence or the like.
- by Randeroid, Statistician
- 4/15/2014 10:44:05 PM
RE: Is that too naive of me?
RESP: No, yet I would say that measuring 'on the job performance' is subjective. We like certifications and other credentials because they provide some assurance that the leader understands the analytics team's technical capabilities. It is an easy next step to think that a Ph.D. will understand this even better and the technical side is where everyone is failing--AIG, Moody's, Fitch, etc,
One advantage of a Ph.D. is that they are regarded as a specialist on that topic. Hardly anyone will pressure a Ph.D., in a quant degree, into making the results fit a pre-selected conclusion. Whereas, those without that heavy investment are fair game. For example, if an MD says their new drug warrants a billion dollar investment despite the results of the clincal trials, who is going to stand up to them? If a highly regarded engineer makes the case that their Pinto gas tank is safe regardless on the crash test results, who is going to be crazy enough to stick their neck out?
- 4/15/2014 5:26:25 PM
@BachInsights -- glad we could provide you a "fun" moment, and something to chew on here as well. Regarding your question/comment on the bachelor degrees -- I didn't speak to the panelist directly, and so only incorporated the scenario he described. If I suggested that this is the only project type -- prescribed, rigid, etc. -- that those with bachelor's degrees were allowed to work on, I didn't mean to as I don't believe that was the intent behind the panelist's comment. I would agree that, theoretically at least, "a university degree holder + 10 years SHOULD be better at analytics than a masters degree holder + 3 years prior to that degree." It would all depend on an individual's experience and exposure to analytics challenges, though.
Likewise, I agree that PhDs don't have a monopoly on great analytics leadership. I don't know whether the panelists were playing to the audience or merely speaking from their own corporate experience -- but it does give you pause to think that PhDs are so readily perceived as being so above and beyond their master's-holding colleagues at the companies represented (or at least in the panelists' little corners of the corporate universe).
- by ophdesdi, Blogger
- 4/15/2014 5:16:59 PM
Beth, you are spot on. When they make me the Education Overlord, I would require that all degrees, at all levels, require some amount of practical experience (call it an internship, co-op, practicum). Advanced Analytics just seems to be such an obvious candidate for this. Its all a bit like swimming...if you learn the physics of buoyancy and the chemistry of water molecules, and then get thrown into a pool, you will drown. If you learn only theoretical mathematics and the derivations of the distributions and every problem you are faced with in the classroom is solved with a piece of paper and a pencil you will drown when faced with a real dataset (that you have to figure out how to pull and load and clean and transform and analyze...). Universities HAVE to teach kids how to "swim".
- 4/15/2014 5:04:32 PM
@ophdesdi, thanks for sharing your thoughts with us here. Your point that universities need to more closely partner with the private sector to ensure that PhD candidates are getting a "sufficient amount of practical exposure to real business problems" is a great one. And I've talked to other professors would have said the same -- but those conversations have centered on master's programs, with the need being to ensure collaboration between the enterprise and the classroom so students get real-life, practicable experience in analytics. Would you agree that this needs to start at the master's level and isn't simply a requirement for PhD candidates?
- by ophdesdi, Blogger
- 4/15/2014 4:55:05 PM
Analytics and the emerging discipline of Data Science presents a unique opportunity for universities to rethink education generally and the role of the Ph.D. specifically. Historically, Ph.D.s were the degree for those interested in academics and research - not application. Application was for the MS (or MA) degree. But, given the nascency of Data Science, combined with the HUGE (dare I say "Big") opportunities for innovative research and problem solving in Big Data and Advanced Analytics, there seems to be a perfect storm for universities to combine the practicality of the application of a traditional MS program with the research orientation of a Ph.D. My opinion (as a professor) is that a traditional MS degree does not provide sufficient training in the mathematics, computer science, statistics required to be a thought leader in Advanced Analytics. But, a traditional Ph.D., with its "ivory tower" orientation is not appropriate either. What we need to do as universities is partner more closely with the private sector to ensure that Ph.D. candidates are getting a sufficient amount of practical exposure to real business problems, in addition to the theoretical research back in the classroom.