Article: "The disposable academic: Why doing a PhD is often a waste of time"

<p>Is it a waste of time to take a Ph.D in Computer Science?</p>

<p>

</p>

<p>Depends on what you want to do. If you want to do research, then of course PhD is not a waste of time. Even in industry, certain jobs require PhD. A small algorithms company I’m consulting for hires BS graduates for “project management” or “systems engineering” positions, which basically means they act as the intermediary between the client and the engineers. The technical people who write the algorithms all have PhDs. The founder claimed his PhD helped him get to where he is today even though his thesis research isn’t related to the work of his company. He said the PhD helped me meet the right people and get the right experiences, and I’m sure it didn’t hurt his attempts to get funding either.</p>

<p>

Non-academic researcher. It took me close to ten years to finish my Ph.D. (working full time and travelling, leaves of absence for both of the 2Kids_) but this was a practical and necessary step for me. With a Ph.D., I am Principal Investigator on my own grants. Without one, I would have forever been a helpful co-investigator on colleague’s grants. Whether a Ph.D. is necessary (or useful!) really depends on your field.</p>

<p>

</p>

<p>It wasn’t in my case :-). </p>

<p>I agree with None’s (^) comments. To elaborate a bit, a Ph.D. “union card” can be useful (if not critical) when working in an interdisciplinary environment where the norm is to have a Ph.D. I find that within the CS community your skill set is valued more than your degree (once you have had the opportunity to demonstrate what you can do). However when dealing with specialists in other fields, they are more likely to treat you as an equal if you have a Ph.D. If you want to move into management, that is, managing staff who are primarily Ph.D.s, having the degree can often be a de facto requirement.</p>

<p>While aspects of my time in graduate school were very stressful, I learned much more in the Ph.D. program than I would have with a terminal master’s degree. Even while focusing on my Ph.D. research, I was in an environment where I was exposed to everyone else’s research, and so came out with a broad appreciation of CS research and technology (at that point in time). Once you start working, you have deliverables that force you to become much more focused on your particular area. (I did not go into academia.) But awareness of the broader CS research activities was, and continues to be, very useful.</p>

<p>On the question of how important is the school you go to, I have only one datum. I found it incredibly useful to be associated with a group, and with a department, whose lead professors were well-known and very active within the research community. We had many visitors, and some of the contacts I made while in graduate school are still important to me now. This is one instance in which the general strength of the department is important (and not just the strength of your advisor). Perhaps things have changed since I was in grad. school, but I didn’t really pick my advisor until after passing my comprehensive exams. Having a variety of reasonable choices was to my advantage.</p>

<p>My suggestion would be (if you are interested in a research career) to enter a Ph.D. program that allows you to earn a master’s during the course of your studies. This gives you something for your work in case you decide that a Ph.D. is not to your liking. It also increases how much you earn during any summer internships after you are awarded the master’s degree. I would also encourage you to look for summer internships. This will give you options in case you decide that academia is not what you want. Even if you go into academia, having contacts with companies can be a useful way to motivate your research, and you can also work for them as a consultant.</p>

<p>My opinions are based on over 20 years at a National Laboratory working in the area of scientific computing. However, I do interact with many faculty and with researchers and staff at computer vendors as part of my job, so have some knowledge of other career paths within CS. Depending on what area of CS you concentrate in, your mileage may vary. If you are a theorist (and not doing cryptography), my comments may not be relevant.</p>

<p>Is a M.D.-PH.D a waste of time? Should I just be a M.D.?</p>

<p>

</p>

<p>The ‘overqualification’ argument keeps being proposed over and over again, to which I have always said: if that’s really a problem for certain job openings, why not just leave the PhD off your resume completely? There’s no formal requirement that you must list every single qualification you’ve ever earned in your entire life. While you can’t add qualifications that you never earned, you’re perfectly free to omit ones that you did earn. If somebody asks you what you did during that timespan, you can simply say that you did attend a PhD program but then ultimately decided that it wasn’t for you - a perfectly reasonable response given that half of all entering PhD students never finish the program. There is no requirement to tell them that you did finish. </p>

<p>Now, whether you should have earned the PhD in the first place is a different question, and that question is the point of this thread. But, given that you did, there is no reason why you should be punished by employers for doing so.</p>

<p>

</p>

<p>Given that you seem to be talking about the humanities, the answer to your three questions is, sadly, probably no, no, and no. </p>

<p>The core issue is that there is no truly reliable and agreed-upon way to measure ‘quality’ when it comes to humanities knowledge because no objective truth exists. In the sciences, math, and engineering, employers can check to see whether graduates understand a baseline of technical principles by asking questions for which right and wrong answers exist, and - in the case of engineering - can actually design a product that actually works or not. For example, if the graduates of an astrophysics program believe that the universe revolves around the Earth, then that program is clearly teaching its students poorly. </p>

<p>But in the case of the humanities, other than to check whether the graduates have actually completed the assigned reading, how would you know whether they were poorly taught? Who’s to say that whether the students’ interpretations of Shakespeare are ‘wrong’? As there is no reliable way to measure humanities knowledge, there is no way to ascertain whether somebody is being ‘mistaught’ humanities - or indeed whether that is even possible - then that means that universities can get away with providing a (supposedly) lower standard of humanities educational quality for decades, perhaps even forever, or at least until the current crop of university administrators reaches retirement after which any decline is no longer their problem. </p>

<p>The same is true - perhaps even more so - with the quality of humanities ‘research’. Keep in mind that we’re not talking about the quality of humanities works, for which one can indeed make the case that Shakespeare is clearly ‘better’ than the average playwright. We’re talking about humanities research, that is, research about those works. Again, as no reliable way to measure humanities research quality exists, university faculties can survive on reputational inertia for decades.</p>

<p>I don’t quite understand the goal of the article. It seems to lump all PhDs into one big boat and then attack everything at once. One has to look at who did the PhD, where they did it, what they did it in… all these things make a big difference. </p>

<p>Furthermore, the author makes several statements that don’t even make sense and only seem to highlight that they don’t really know much about what they’re writing. This damages their creditability and makes it seem more like they have an ax to grind. For example:</p>

<p>

What??? </p>

<p>A math thesis is typically short because one can just list an equation they developed and a few pages with a mathematical proof showing that it works. However, getting to that point of figuring out something nobody else has figured out before still takes years. The author implies such a short thesis can be completed very quickly, which is not true.</p>

<p>

Clearly they didn’t read last weekend’s New York Times which had a big piece on how a JD is largely worthless these days. (Yes I note the article was from December, but there have been many other article like the one in last weeks’ NYT) </p>

<p>I agree that good academic positions are often hard to come by, but not everyone who does a PhD wants to be an academic. Depending on what one studies there can be a lot of transferable skills that can be applied elsewhere. Yes one often has to take a momentary step back and work their way up, but there are plenty of options for career shifts out there that value the (particularly quantitative and analytical) skills a PhD can bring to the table. Given that many if not most PhD students actually get paid to go to grad school and come out without any grad school debt, it can be a very good deal depending on what path one takes.</p>

<p>

</p>

<p>I think they meant to say that the JD, however difficult the legal hiring market may be, is still a more marketable degree than is a PhD, especially one in the humanities.</p>

<p>

</p>

<p>Though it’s likely you’ll get questioned about the temporal gap in your r</p>

<p>sakky, while I agree with your general answers to the questions (will universities suffer from hiring mostly adjuncts? I don’t think they will, and I do think reputational inertia will kick in) I disagree with the idea that humanities students/programs cannot be evaluated in a meaningful way.</p>

<p>First of all, there are objective things that you have to learn in humanities programs, if we want to be reductionist. Factual questions about literature plots; dates and summaries of historical events; fact-based points that philosophers have made; and assessments of reading ability in the foreign languages of choice are pretty easy to come up with and those questions have right/wrong answers.</p>

<p>But more importantly - you would never evaluate an astrophysics doctoral student by asking them whether the universe revolves around the Earth. I know that was an extreme example but even beyond that, doctoral studies in the sciences don’t get evaluated by factual questions beyond coursework. I’m in the basic psychological sciences people don’t ask us to define operant conditioning or explain what a p-value us, anymore than they would ask a chemistry PhD to rattle off the chemical makeup of X substance or a math major to do a proof from memory (well, except in our qualifying examinations). That factual knowledge is <em>assumed</em>. If you don’t have it you are screwed from the beginning.</p>

<p>When evaluating PhD programs it’s more important to see how students <em>use</em> that information to come up with new knowledge. No, there’s no one right interpretation of Shakespeare; but some interpretations are better than others. Some correctly use the tenets and theories of British literature and others do not. And even in an “objective” field like quantitative psychology, there are several statistical techniques that one can use to analyze the same data but some are better than others. There are several psychological theories I can build the framework of my study on but some are better than others. Doctoral work is rarely evaluated on factual/objective knowledge once you get past a certain point and evaluations focus on how well you are able to USE the objective knowledge you’ve gained (mostly in undergrad, some in the MA/MS part of your degree) to create new knowledge and interpretations.</p>

<p>So in that sense, yes, you can evaluate humanities programs in much the same way as you would evaluate science programs. However, my agreement with you begins here - I don’t think that professors often evaluate their own graduate programs; there’s not a lot of control within academia for quality and what it means to have a PhD, what we’re expected to know, etc. In fact, a big part of the rankings is faculty citations and publications. That stuff means next to nothing if the professors never want to interact with the students and pass on their knowledge.</p>

<p>

</p>

<p>But my point is that you don’t really need to have top-ranked faculty with glittering research records and storied CV’s to teach those factual elements of humanities curricula to students. Indeed, it’s arguable whether you would even need any faculty at all. You could run purely computer-driven standardized tests to measure whether students learned the factual elements of humanities - i.e. what were the names of the characters in this book, what year was this particular piece of art created, how do you translate this paragraph into French - and you could then tell the students that the exams would be drawn from a given set of material, and then have them study that material on their own time. Hence, however precipitous the research reputation of the faculty might decline, it would have practically no impact on the ability of that faculty to actually teach the factual elements of a humanities discipline. </p>

<p>

</p>

<p>Actually, they do, in the sense that their dissertation work itself is supposed to be factual in that it actually produces a reliable and non-obvious explanation of a scientific phenomenon that had heretofore not previously been known. The work is then factual in the sense that somebody else in principle could actually utilize that explanation to make predictions, often times regarding the results of a repeat of the experiment. Or, in the case of mathematics, computer science, or more theoretical branches of the natural sciences (i.e. string theory), the dissertation is supposed to produce a new framework - or an extension of an existing framework - usually in the form of a mathematical proof that is logically self-consistent. </p>

<p>Now, to be clear, the above obviously doesn’t hold for everybody’s dissertation. Clearly, some science PhD students will produce explanations that somebody else had already said, but which they didn’t know about. Some math PhD graduates will have written dissertations that have subtle errors in their ostensibly self-consistent mathematical proof that are only revealed later. And of course some students will produce explanations that upon later repetition prove not to be reliable because of methodological errors or even outright fraud. But at least the ideal exists that you can produce ‘factual’ predictions or self-consistent logical statements that nobody can really deny. “E pur si muove” as the (apocryphal) Galileo legend reportedly goes. </p>

<p>But the humanities doesn’t really have that, because nobody can definitively “prove” that one interpretation is better or worse than another. </p>

<p>

</p>

<p>But, to quote Shakespeare, there’s the rub. Sure there may indeed be better interpretations of Shakespeare than others, but the question is who decides which are the better interpretations? You might have just written the greatest interpretation of Shakespeare in the history of the world, but if your dissertation committee members happen not to like it, then you’re not going to graduate. The converse is also true - you could have written a terrible dissertation, but successfully graduate anyway if your committee happens to like you. The same is also true when dealing with peer-reviewed journals within the humanities, as much of it seems to devolve to a cynical game of the ‘publication lottery’ of just hoping to luckily land three referees who approve your paper. {One prof I know made the trenchant comment that ‘successful’ research consists of 3 people agreeing.} </p>

<p>Now, again, to be clear, nobody is denying that politics plays a role within the sciences as well. However, at the end of the day, factual results do matter. Science is not a democracy, and the most unpopular member of the community can ultimately prevail if the objective results are on his side. Barry Marshall won a Nobel Prize for famously demonstrating that bacteria does in fact cause stomach ulcers, overturning the all received scientific wisdom. </p>

<p>But you can’t do that in the humanities because there are no “objective results”. If the community believes that your Shakespearean interpretation is wrong, then it is necessarily wrong, as the notion of ‘wrong’ is inherently a socially defined construct. </p>

<p>

</p>

<p>In which case you could publish the results of all of those techniques - perhaps in an addendum section on your website - to demonstrate the robustness of your results. </p>

<p>

</p>

<p>Which has led me to believe that psychology - like the other social sciences - is not (yet) truly a science, although it may become one in the future, and in fairness, I would say that psychology is probably the closest of the social sciences to becoming a true science. The issue of having multiple theories that somebody could ‘choose’ from in order to frame a paper is the mark of a discipline that has not yet progressed to a true science, and is a problem that plagues all of the social sciences. There should only be one clear framework to choose from. If multiple frameworks exist, then each should have their boundary conditions clearly demarcated so that everybody knows under what conditions is each framework appropriate to use, or at least, that those boundary conditions should be a topic of active study. For example, all physicists understand that you cannot use the laws of classical electromagnetism to explain the precise interactions of individual electrons and photons, as that is the realm of quantum electrodynamics. </p>

<p>

</p>

<p>But that is probably the crux of our dispute: what exactly is ‘new knowledge’ in the context of the humanities? By re-interpreting Shakespeare, have you really created new knowledge? Did Shakespeare himself intend for those interpretations to actually exist in his works? Or are you simply reading additional meanings into his work that don’t truly exist? </p>

<p>

[quote]
Though it’s likely you’ll get questioned about the temporal gap in your r</p>

<p>Sakky, your “re-interpretation of Shakespeare” demonstrates a naivet</p>

<p>

</p>

<p>It’s true that the same data will often get different interpretations if it’s complicated enough. However, given enough data, the number of ways to explain an experimental result should diminish to one. The opposite seems true for the humanities, where, given more data, the number of interpretations of a work seems to grow without bound.</p>

<p>^^^ Agreed. </p>

<p>But if there weren’t a difference, they would both be sciences.</p>

<p>Since no one answered my question, I will just comment on the subject at hand. I think that PH.Ds are needed even though the way to get a PH.D can be problematic. The U.S. needs researchers to advance the country at whatever cost necessary. We Americans do not want another country to beat us in the race for progress.</p>

<p>

[quote]
Sakky, your “re-interpretation of Shakespeare” demonstrates a naivet</p>

<p>

</p>

<p>That actually seems to drive at the heart of my concern: what does it even mean to make ‘progress’ in the context of humanities? To be clear, I am not talking about the process of creating new works of art or literature, for I certainly agree that that is progress. I’m talking about the process of conducting research on existing works. </p>

<p>To be clear, I am hardly a science triumphalist, for I believe that much of the sciences fail to make demonstrable progress as well. String theory is the notorious archetypal example of a ‘science’ that has produced relatively few validated predictions relative to the manpower and budget that it consumes - hundreds of PhD’s granted and scores of physicists promoted to tenure for string theory ‘advances’ that likely will never be testable upon actual data during our lifetimes. Hence, in this context, it is not clear to me that the United States would truly be worse off compared to other countries if they produced more string theorists than we did.</p>

<p>“the question is who decides which are the better interpretations?”</p>

<p>“hundreds of PhD’s granted and scores of physicists promoted to tenure for string theory ‘advances’ that likely will never be testable upon actual data during our lifetimes.”</p>

<p>String theory is just so ugly, but let me say that one of the reasons it is falling out of favor is the same reason theories in the humanities can be deemed as less true than other theories:</p>

<ol>
<li> Lack of internal consistency (philosophy still being part of the humanities).</li>
<li> Lack of evidence in the tangible world.</li>
</ol>

<p>In other words, is it valid and sound?</p>

<p>Universal skepticism and relativism (which are actually less common than you might believe–they are much more popular among undergraduates than PhDs) aside, these two determine who is considered smarter in the sciences and the humanities alike.</p>

<p>It’s easier to test for soundness in certain hard sciences because it’s easier to do experiments on rocks and bacteria. However, validity plays an important role in both and soundness is a factor in the evaluation of humanities papers.</p>

<p>You can’t say that Shakespeare was a feminist if the prevailing definitions of feminism are contradicted by major themes and specific assertions throughout his works, in addition to concrete evidence for facts about his life that suggest that he did not practice feminism. (That is just an example I pulled out of my butt for argument’s sake: I’m sure someone has said he was a feminist, and I’m sure an argument about the validity and soundness of his arguments were had.)</p>

<p>Very few humanities papers are purely questions of opinion. Even when judging different works, they must outline specific measurement criteria and then use that to judge the works. They can’t just spout out random crap like, “I liked this book because it was really good.”</p>

<p>And if they can, I would suggest that this demonstrates a verifiable decline in the humanities.</p>

<p>Having come from the pharma industry, I can not stress enough just how overrated a PhD in chemistry is for most people.</p>