Do Scientists Really Need Ph.D.s? | Hacker News Reader

Do Scientists Really Need Ph.D.s?

nature.com

32 pointsbarry-cotter42 comments

Freeman Dyson on the topic:

http://michaelnielsen.org/blog/freeman-dyson-on-invention-an...

Brand: One of the things I got from Infinite in All Directions – it was a delight to me, and I’ve been quoting it ever since – is that you honor inventors as much as scientists.

Dyson: It’s as great a part of the human adventure to invent things as to understand them. John Randall wasn’t a great scientist, but he was a great inventor. There’s been lots more like him, and it’s a shame they don’t get Nobel Prizes.

Brand: Is it the scientists who are putting them down?

Dyson: Yes. There is this snobbism among scientists, especially the academic types.

Brand: Are there other kinds?

Dyson: There are scientists in industry who are a bit more broad minded. The academics look down on them, too.

Brand: Is that a weird British hangover?

Dyson: It’s even worse in Germany. Intellectual snobbery is a worldwide disease. It certainly was very bad in China and probably held back development there by 2,000 years.

Brand: How would you stop this intellectual snobbery?

Dyson: I would abolish the PhD system. The PhD system is the real root of the evil of academic snobbery. People who have PhDs consider themselves a priesthood, and inventors generally don’t have PhDs.

I'm handing in my PhD thesis tomorrow so this is a well timed question :-)

I have learnt a lot from my studies and feel that I have gained important skills in attacking problems and evaluating my solutions to them that will be useful inside and outside of academia. I couldn't have picked up these skills in the 15 years I spent in industry.

There may be many examples of people who can do great science without a PhD but for me the experiences of the last 4 years would be essential if I were to get a job in academia. As it is, I will be going back to programming -- I feel the academic job market is overly competitive for somebody my age.

No, they need knowledge, creativity, and courage to ask "why?" and "why not?" A PhD is a firm credential for a job in academia that may or may not indicate that you have these qualities. It is neither necessary nor sufficient to becoming a scientist.

The question isn't "Do people need the degree?" They're using PhD as a metonymy for the Western process of training researchers. What they're asking is "Is our current process of training researchers necessary?"

The current process involves some breadth requirement (exposure to most of your field), and doing a starter-project (the dissertation).

And willingness to spend years of their young adult life jumping through academia's hoops.

You can make that sort of snarky comment about pretty much any career. There are always hoops to jump through (though you might not see them till you get there).

Those two questions have the same answer. I have a PhD and my experience is that the process is broken because it's geared towards the wrong goals of (1) writing papers, (2) getting grants, and (3) the game of getting a tenure-track position and then tenure. If you happen to make a great discovery, that's nice, too. Most people I know get a PhD for career advancement, not to engage in the scientific and discovery process.

My experience was that the process was more "cargo cult science" and less science.

I think you just had the wrong expectations about what a PhD is. The idea isn't just to work on interesting problems for 5 years. You're also meant to learn how to function as an academic, which necessarily means learning how to write papers and (maybe) get grants. There's nothing "cargo cult" about that. After all, cynicism aside, one of the best ways to get a paper published is to actually have a good idea.

>Most people I know get a PhD for career advancement, not to engage in the scientific and discovery process.

If you're in your 20s you'd have to be a bit crazy to take 5 years out without thinking about your future career. If PhD programs didn't cater at all to people's career needs you'd have an even heavier bias towards wealthy students in grad school. That wouldn't be good for research.

It would appear so, but I did not. I have no regrets and enjoyed my time as student and professor. I self-funded my PhD through fellowships and work, which gave me freedom to work on any problem I wanted. That taught me self-reliance and the power of creativity, not just technical skill and process. That is something I don't think my counterparts who were supported by grants grasped very well. During my time as a professor, I tried to give my own PhD students the same freedom.

I agree that having good discoveries and novel algorithms should be a good way to get a paper published. I respectfully disagree that the process isn't "cargo cult", though. In practice, it tends to be more often than not. The research culture rewards quantity over quality, and most papers are rubbish. You will hear "Dr A has B publications" 100+ times more often than "Dr C has a few papers that have been cited hundreds of times." That's an indicator that our research culture is broken.

I guess it depends a lot on the field. I'm funded mostly by my department and there aren't any restrictions on what I can work on. I don't find the standard of published papers to be particularly awful, and my colleagues seem to be getting on fine.

I agree that the emphasis on publication volume is a pain, but it's not so easy to find good alternatives. In principle, it's all very well to say that one good publication should count for more than ten mediocre publications, but it gets a little tricky to make decisions that way in practice. It could well have the opposite of the desired effect by making the process even more cliquey that it currently is.

In general, I'm a bit bemused that people seem to hold academia to such absurdly high standards. Some people seem to go in expecting it to be a bullshit-free paradise of open and free enquiry. Well, it isn't -- academics are human too. But it's really not as bad as all that.

I partially disagree. The point of the PhD process is not to create knowledgeable, creative, courageous scientists. As you point out, they can (and do) come to be that type of person on their own.

The value of the PhD is to teach the budding scientist the process that surrounds modern science. It's an opportunity to learn what the social structures in the field are (societies and conferences), to meet fellow scientists and form a network of people to draw on for ideas and criticism, and to learn the vagaries of funding and formally writing up research. Having a PhD says "I know how to do all these things... I know who the important people in my field are... I know which conferences are the most important... I know how to get funding". All of these things are important things to know if you want to do science professionally, because you need the connections and the funding to put food on the table.

This isn't to say the PhD is some magical permission to do science. Lots of great science is done by amateurs, and more power to them. So scientists don't need a PhD to do science, but it is more than just a credential. It's a valuable training program on how to navigate all the social and bureaucratic things that surround the actual science.

The prerequisite seems to only be about a century old in western academia. Michael Faraday, James Maxwell, Betrand Russel, et al didn't have PhDs. Faraday actually had no university at all.

Edit: Ah, apparently doctorates weren't even broadly granted in the sciences until about a century ago:

http://en.wikipedia.org/wiki/Doctor_of_Philosophy#History

To make this clear, you may as well ask if surgeons need to go to med school. It's required by regulation and tradition. Same with lawyers. It's more an indication of the maturity of the field than the ability of the person, although it hopefully represents a relevant hurdle.

So if you can devise a quantum computer without a PhD then the lack of one won't hold you back.

Medics, lawyers, accountants etc. are professionally qualified, a PhD is an academic qualification. They need to be licensed to do their job and if they aren't they can't do it.

Not the same thing at all.

It's pretty similar in practice, though. A PhD is effectively a "license" to hold an academic position.

Perhaps we need a professional degree for research science. A JD is different from a Law Ph.D. and an MD is different from a Ph.D in Medicine--so it stands to reason that we could have a professional degree as a professional rather than academic qualification.

I don't really see any need for that distinction. The curriculum would presumably be about the same for both degrees.

I don't really see the point either - the reason why there are professional qualifications (apart from the monopolistic reasons) is to ensure that people who have responsible jobs meet some minimum level of competence.

If an academic screws up, nobody really gets harmed that badly. Contrast that with a medic, lawyers, PE etc.

If you don't want to be an academic the curriculum might not be different, but the dissertation requirement could probably be done away with. Some universities have changed their Ph.D. programs such that you can graduate with 3 published articles in a peer reviewed journal. That seems like a reasonable standard for a professional degree, as opposed to a full dissertation for an academic degree.

Yes but the big difference between a professional qualification and an academic one is that the former can be taken away from you if you screw up badly enough.

I think the key point of the article is "Given the increasing rigidity and length of the Western academic pipeline which now extends so far beyond the PhD that the average age for first-time principal investigators on grants from the US National Institutes of Health is 42 the BGI model may be worth serious consideration."

This isn't even just getting a PhD. It's the whole post-doc rigmarole after that too.

The question remains, I think, a valid one. Does our current postgraduate system work well, or could it be better? Sounds like BGI is trying a great experiment and I'll be interested to see the results.

I don't think it's the same question - the function of a PhD for science researchers is very different from the MD/residency for surgeons, which is actually very different from the function of a JD and the Bar for lawyers (lawyers don't go through anything like the apprenticeship a surgeon goes through to become licensed, and a PhD generally doesn't provide a license to do anything).

Also, keep in mind that many people believe that the Bar and AMA are actually large scale cartels that do more to harm the public than good. This isn't necessarily an argument that credentials and licensing are bad per se, it's an argument that the scarcity of services causes more damage than the dilution of talent (for instance, nurse practitioners might be able to provide a very high level of service in areas where MD's currently enjoy a monopoly). I'm one of the folks who are frustrated with the cartel-like behavior of professional associations in the US (and elsewhere), but there's obviously a wide range of opinions on the subject - some think they should be essentially abolished, others think they are fine in principle but need to be reigned in to ensure that "protecting the public" doesn't morph into "protecting the cartel."

It's controversial, of course. But you might be interested in reading Milton Friedman's essay on the licensing of Medical professionals (just google for it, you'll find it quickly). Regardless of whether you agree with it, you'll probably find it interesting.

Not exactly.

Look at computer scientists like Chuck Thacker, Robin Milner, Simon Peyton Jones never had a PhD and they did first class work.

It really isn't a question of which system is better. It is about supply and demand. If there is a huge demand for scientists (as there appears to be at BGI today) then the requrements for getting a good job will be reduced. With the scarcity of decent jobs in US Academia today the prolonged PhD/Postdoc cycle is just a symptom of the underlying lack of demand.

It's not just a lack of demand for PI's, it's a strong demand for lab monkeys aka grad students, like yours truly. So long as (1) most research labs require many grad-student-level-trained hands for every professor-level-trained supervisor and (2) grad students continue to be deluded about their chance of securing a tenured position in the future, the situation will continue.

I worked as an applied physicist / engineer at a smallish engineering company directly out of college, where about 50% of the employees had PhD's. When I decided to quit to get my PhD, the VP in charge of my boss gave me a 30 minute lecture about how a PhD would stifle my creativity and that some of the best people they had didn't have grad degrees. I was a little skeptical of that claim, as the only technical person I could think of without a graduate degree was that VP. He was very impressive, being both extremely intelligent and clearly someone who just absorbed technical material (he had two bookshelves full of grad-level math texts and seemed to know something useful on almost every topic).

Looking back, I feel that I would be much more useful to that company after a few years of grad school, receiving a more thorough mathematical and physics background and experience leading my own research projects. Of course you could attribute some of that to just "experience", which I might have gotten just the same at that job.

I'm listening to David Gilmour's "On An Island" right now. He didn't concern himself with pursuing a Ph.D in music before starting his career. But his body of work is extraordinary nevertheless and society has benefited tremendously.

Asking questions and following the scientific method doesn't need anyone's permission either.

My favorite example is Mark Daly (of GWAS fame). He basically got a PhD for signaling purposes. He was already doing his excellent work beforehand, and got the PhD for the benefit of the institution (MIT/Broad). Or, at least, so goes the lore.

The mathematician Stieltjes never completed a PhD, but was first rate. He was given an honorary doctorate so that he could apply for a chair at Groningen university.

Bob Floyd never even tried to get a PhD. (He did collect letters addressed to "Dr. Floyd".)

Stanford made him a full professor even though he didn't spend "enough" time on the tenure track.

credentials are for grant oriented signaling.

This comment has a bit of truth in it. Institutions look at grant money obtained before anything else when evaluating faculty. And without credentials, the NSF et al have little else to go on for a new researcher.

The entire system of academia is horribly twisted mess of politics and inefficiency that creates competition at wrong the level (individual in their 20/30s rather than teams of people at whatever age).

Who got human genome mapped in three months - a team from private industry (it took academia 15 years and boatloads of cash)

Who got a car to autonomously drive through the desert - s team people motivated by a competition (academia hadn't managed to get it done before then)

The current academic system of well connected 'star' PI's who are 'geniuses' directing around masses of lesser 'grads/postdocs' is a scale of snobbery not seen anywhere but show business and politics. It should be feel disgusting to anyone who sincerely values science and human progress.

I don't really get the point of your examples. No doubt one could also point to cases where academic researchers discovered something that industry didn't. And businesses also tend to have a hierarchical organization.

Businessman live or die by their skill at picking talent. Bureacrats who fund science live and die by their skill at picked themselves. I'd trust the businessmen to get the job done.

Monetary competitions like the X prize would get you scientific discoveries faster and cheaper than the current system, simply because you would have a businessman driving progress rather than studious scientist-cum-team leader (PI).

Academics aren't particularly beuracratic in my experience -- most of them hate doing paperwork. In any case, your views on this are obviously based on considering what "ought" to happen given your stereotypes, rather than looking at the facts. A vast amount of research gets done in academia.

It depends on the field. As far as I can tell, some undergraduate degrees are traditionally so watered down that most graduates simply haven't had the opportunity to become fit for research. (Chemistry, I'm looking at you.) Other programs do better. (Physics. EE. CS at the more hardcore schools.)

I was a physics undergrad at a very hardcore school, and I can tell you that I wasn't even remotely close to being able to do physics research upon graduation. Even after taking two years of full-time grad classes, I considered myself (and my fellow student) somewhat unprepared to choose a research field, much less do independent research.

Part of this stems from the fact that US undergrad degrees are much broader than in Europe. My understanding is that European undergrads specialize in a topic day one, whereas US undergrads take at most half of their college classes in their major.

It's more that there is too much hand holding and babying of american students. European students (at the good universities) are expected to thoroughly and independently study for exams that will come two or more years later. American students are driven by weekly assignments and all sorts of midterms and quizzes. When it comes time to do independent research, the american student is thoroughly unprepared compared to the european.

In my experience (physics graduate student), American students and European students do not differ significantly based on work ethic, planning, or organization. Certainly, the variance within each group in those categories is much larger than the difference between those groups.

The significant difference was that Europeans had completed more math and physics course before they had to choose an advisor and commit to a research topic.

Or alternatively, most undergraduate degrees are so irrevelant to either research or industry that they are essentially make work.

Reply on news.ycombinator.com