A Short Rant about "Dissertations"

Steven Horwitz

I just received the University of Michigan's (my alma mater) list of PhD candidates seeking employment in this job season.  There are 19.  Of them, 10 have a "dissertation" whose title begins with "Essays in/on...", 7 of them have a "dissertation" whose title begins with "Three Essays in/on" and the other 2 have an actual title (and maybe an actual dissertation too).

Yes, I understand the institutional pressures to get students out and yes I understand that articles are where the payoff is, but I can still lament the loss of sustained attention to a single argument that used to characterize a much higher proportion of Economics dissertations.  And a title such as "Essays on the Macroeconomics of Trade Flows" is singularly uninformative in helping me understand what this person might be writing about.

Yes, I sound like an old fart, but so be it.  "Three Essays in Search of a Staple" has won out in the "competitive" process, but that doesn't mean I have to like it.

Rant over.

Comments

[Steve Miller]

It gets even worse. We don't have to use punch cards to run regressions anymore!

[Peter G. Klein]

Steve, I see your point, though I personally favor the three-essays model over the grand treatise model. Isn't the solution to make the job-market paper, rather than the dissertation, the unit of analysis? Placement officers should ignore their students dissertation titles and simply circulate titles (and abstracts) of the "official" job-market paper.

[Steve Horwitz]

That would certainly be more helpful Peter.

And I'm not so much defending the "grand treatise" as I am something that's more integrated and systematic than what I often see, which is three essays only loosely linked by a theme. I think one can write a "Three Essays" diss that lands short of the "grand treatise" but still far ahead of what I've seen.

[liberty]

I agree with Steve - although I have no authority to speak on the subject. I have always felt that before focusing (as has become very usual these days) almost exclusively on writing academic papers for journals, possibly supplemented by the occasional (a) book for the layperson or (b) textbook, young academics should be forced to think more broadly about research and formulate a larger thesis for which they must accumulate vast evidence or proofs.

This would (in my opinion) tend to prevent economists from learning to write bite-sized mini-papers on narrow subjects simply in order to publish, and would encourage them to think more creatively, with a larger perspective that considers the big questions in economics - like methodology, history, philosophy etc that they will run into writing a long dissertation. At least, it would if taught right.

By the way - I think they still do this in other countries, like the UK.

[David Hoopes]

I always liked grandiose titles. I'd like to see students go out on a limb. Examples, "Financial Markets." "The Theory of the Firm." "Reworking Macroeconomic Application."

[Troy Camplin]

I'm with Steve on this. If you cannot sustain an argument for 250 pgs or so on an original contribution to the field, you have no business being granted a Ph.D.

[Steve Miller]

I can see how there's some greater good-type argument for a 400 page unpublishable monstrosity (exceptions prove the rule here), but practically speaking you may as well chain new assistant professors to 1000 lb anchors and throw them overboard: Here's a new job in a new town, with new preps, and by the way you need to forget your past 2-5 years of research and start writing new journal articles from scratch and have several published, in print, within the next five years.

[liberty]

Steve Miller: there is no reason, if the tide were to turn back to full-length dissertations, that professors should be expected to publish several papers in the first few years. I am more interested in good economics than prolific economists.

Nor, of course, need the papers be completely unrelated to the dissertation--I would expect a good dissertation should have at least a few good papers hidden in there that could be fairly easily extracted.

Again - I would look to the UK and elsewhere (and historically in the US) for examples of how this can work well.

[Roger Koppl]

That's a helluva an "if," Liberty!

[Jim Rose]

I presume the three essays approach is a response to market demand for more diverse skills and a more detailed signal of what a job candidate knows and has done.

employers want three smaples, not one, of the candidate grappling with trickly problems. Three essays may be a better way to reveal talents.

coursework components spread to PhD programmes in Australia because the PhD student was thought to be not tested enough and skilled in enough areas if they worked only on one topic.

[Peter G. Klein]

Right, the issue isn't the dissertation per se, but the way economics research is typically communicated. In a profession that values articles, but not books, why make writing a book asthe entrance requirement? As long as article-length treatments are required for hiring, tenure, and promotion, students should be trained to write articles. In academic fields where these rewards are based on books -- most of the humanities, for example -- students should be trained to write books. Of course I'm talking here about pragmatic considerations, not ideal states. (A book-length dissertation may be intrinsically superior, for some topics. But it is likely to be a huge professional handicap.)

[Jerry O'Driscoll]

The model began to change while I was going to graduate school. "Book" was king and anyone with a pretense to being a scholar would be expected to write one or more books. Your disseration would normally be the first.

Then, too, there is no reason a student can't carve out an article from his "book" dissertation. I did that in the midst of writing my dissertation.

[Peter Boettke]

Steve,

We are Lavoie students and he required us to write books for our dissertation work, and to Steve Miller, these were not professional suicide but helped establish careers because Don was a relentless editor and insisted you write a publishable book. And it worked, a quick count would include besides Steve and myself the other students include David Prychitko, Roy Cordato, Emily Chamlee Wright, Howie Baejter, and you should also count Virgil Storr and Tom Rusticci. Each wrote thesis and then published the thesis as a book. Among more recent students I would simply point to Tony Evans and Chris Coyne who wrote theses and then published them as books. And Pete Leeson has already published 2 books as well as journal articles.

The norm in our profession is articles as Peter points out and the best strategy is to follow the norm, but you can buck the trend if you do it right.

[David Emanuel Andersson]

I know for a fact that in Sweden a book is counted as equivalent to four journal papers (adjusted according to the quality of the publisher, just like journal papers are counted differently depending on the prestige of the journal). In Taiwan (where I live), however, nobody knows how to evaluate books since the only books that anyone writes here are undergraduate textbooks. But then I would guess that Taiwan is very much influenced by the US, since about 90% of the professors here were educated in the US or Canada. An interesting observation is that when I was applying for a promotion here in Taiwan, I was explicitly recommended to use my shortest journal paper as my "best product," since committee members are apparently "too lazy" to read more than 15 pages (I was told this by three different professors).

Also, when I was a PhD student in Sweden in the nineties, most professors at my department preferred a book-length dissertation. On the other hand, I have been told that the US standard (3 papers) is becoming increasingly popular, at least in Scandinavia (I don't know about the rest of Europe).

[Jim Rose]

Peter Klein,

Your post on books versus articles as the method of professional communication in different professions is an excellent idea for an article on the labour market for job market candidates.

Where has this trend persisted and where has it not? Business orientated degrees versus soft humanities and history?

The focus of the test is the responsiveness of universities to the job market needs of their fee-paying students and later graduates. Who changed faster to the three essays format? Is the response faster in universities where students do not get so many scholarships and other fee discounts?

Fees have gone up in response to the rising higher education premium. Students are willing to pay more for better educations.

Do students enroll more in those PhD programmes that make them most marketable in their chosen professions by emphasising three essays that can easily be turned into journal articles or stress the writing of a PhD that can become a book? Whose market share is increasing?

Stigler’s survival test of efficiency applies.

[George Selgin]

Of course the three essay model "wins" as a means for getting assistant professors tenure. But it does so thanks to a general decline in what's expected from them--a change from expecting them to actually be experts in a subject to expecting them to have X pubs within Y years. There's no point in trying to reverse the trend in dissertation length and substance unless doing so is part of a general attempt to restore the older emphasis on scholarship, which would include insisting (horror of horrors) that every PhD be versed in economic history and history of thought; that is, that every one actually learn to appreciate some fat books.

Mr. Miller won't approve of such a program, which he likens to reviving punch cards. But what good is it to maximize the survival prospects of young econ PhDs if it means having fewer and fewer competent economists? What use is another economist of the sort who can't imagine how a hefty book can be anything but a "monstrosity"?

Perhaps Miller and others who share his perspective imagine that the assistant profs whose survival so concerns them will become real scholars once they make tenure. With all due respect: not likely. To be "trained" in the Stigler-Miller fashion is to be so trained for life, that is, to be good for running regressions and tossing out every now and then a superficial 25pp MS, armed with the usual bells and whistles, and with a half life of about 2 years (or until the next econometric fad takes hold). The majority of such PhDs become deadwood by 45.

[Steve Miller]

Prof. Selgin, you are mistaken -- I don't remember taking a normative stand on this issue at all. There have been consequences to low-cost regressions, and they're certainly not all "good" when it comes to developing scholars (econometric fads are just one consequence). My point is simple: as a positive strategy, it makes little sense to avoid the three essay approach. Pete noted some exceptions. But that's what they are -- exceptions. And what may have been true 20-30 years ago is much more true today. X pubs within Y years is the norm, and until that changes it doesn't make much sense to whine about the trend in dissertations. It is senior faculty who must convince their colleagues that 1 important book is as much evidence of scholarship as half a dozen (or ten, or whatever) journal articles. Right now we all know that a book essentially counts *less* than one mid-level journal hit in the eyes of most tenure committees. I like big fat books, I'm just not crazy enough to try to write one pre-tenure.

[Peter Boettke]

George,

Steve is right, he was not passing a normative judgement, but instead pointing to a reality. You need 6 quality journal publications in 6 years, and if you have 3 publishable essays you hit the ground running professionally.

Also, as Peter pointed out, our profession for good or bad is a journal culture, not a book culture. Political science is still a book culture, history even more so. But economics --- journal articles in SSCI ranked outlets is the name of the game.

No normative judgment, just a positive statement of what in fact is our shared professional reality.

Pete

[Steve Miller]

In retrospect I did use value-laden terms like "unpublishable mostrosity". However, my take is that most "book" dissertations are at first not publishable, at least a couple of years away from being final products. Heck it can easily take three years to publish a journal article. There are major exceptions, dissertations that quickly became important published works. But looking at the senior faculty I know (who wrote "traditional" dissertations), I see most simply abandoned their dissertation topic work on articles that would get them tenure. Some converted their epic dissertations into articles, but it was a process that took upwards of five years. I know one who got three very good journal hits out of his dissertation, but that took a decade! Maybe the tenure clock should be longer; I don't know.

My own normative view is that in a just world hiring committees would be impressed that Mr. ABD is sending his dissertation out to be published as a book, and has plans to write another manuscript once that process is complete, further understanding that the process can take years. And wouldn't it be nice if book-writing senior faculty stuck up for these candidates? Wouldn't it also be nice if tenure committee members could read as well as they count? When you all start hiring and tenuring book-writers we the grad students and junior faculty will change our behavior.

[Troy Camplin]

If each class one took as a Ph.D. student required the production of a publishable paper, then a book-length dissertation that showed one could think on a much larger scale would be a good way of bringing everything together. I fear focusing on papers will in fact narrow thinking rather than expand it. Personally, I would bet this anti-book trend has far more with econometrics and their inability to develop a complex argument than anything else.

[Andrew Larson]

A good dissertation topic will naturally have several "spin-offs" for journal articles. If you can't do that, you've chosen a dead-end topic. We need to think in terms of long-term strategy when choosing a topic. Otherwise, why go through all that work for a single or maybe three papers? A good book-lenth dissertation should produce journal articles for a long time.

[liberty]

"I would bet this anti-book trend has far more with econometrics and their inability to develop a complex argument than anything else."

Not just economics - my feeling is that the same formalization of modeling that Austrians lament is responsible for the trend toward papers over books in the discipline - of which the 3-paper-dissertation is only a symptom.

Most economic students - starting around the 1940s I guess - were taught formal models and techniques and little else. How can one learn that and then write a book? Its not easy. It makes much more sense to learn those techniques and models and then use them to analyze a problem, or at most tweak a model itself, and then write it up in a journal article.

Back when economics was about concepts, history, and logical rigor of analysis it made much more sense to write a book-length treatise and ultimately to write publishable books. Most economists today couldn't scrape together enough material for a coherent book from their entire careers.

[liberty]

That should have began "not just econometrics"

[George Selgin]

I apologize to Mr. Miller for misunderstanding his position: actually it wasn't just his reference to big books as monstrosities that threw me; it was also the remark about punch cards, which after all appeared to suggest that the values the loss of which Steve bemoaned were, not merely old-fashioned, but disntinctly inferior to the new ones!

[Steve Miller]

Prof. Selgin, I can see how it was interpreted that way, but I was thinking more about how both changes represent reduced barriers to entry in the profession. Now that I say that, it occurs to me that other barriers have been raised in their place, though.

[George Selgin]

Steve Just to be sure we really are on the same page: I'm all _for_ barriers to entry into professions so long as the barriers serve to exclude the unqualified. I hope you agree that we don't want it to be too easy for people to qualify as airplane pilots or brain surgeons. Ditto economics. Punch cards are lousy barriers because having the stamina to deal with them says nothing about one's grasp of economics (or econometrics, for that matter). A required big fat dissertation, on the other hand, might weed out anyone not prepared to gain some real expertise in their chosen fields of study.

[Troy Camplin]

Here's a question in regards to barriers to entry: if, say, I wrote several articles on economics that were published in economics journals, could I, with my Ph.D. in the humanities, be hired by an economics department? If so, what would be the number that would remove the barrier? Would a single book be enough?

This might answer the question regarding the appropriate number of articles for a "dissertation," or what the dissertation should be an equivalent of (the number of articles one should be able to get out of it).

[George Selgin]

Troy: There are examples, including some prominent ones, of academic economists lacking econ PhD's. Gordon Tullock comes to mind (although he got an honorary PhD from Chicago in '94), as does David Gordon (of Sargent and Gordon). John R. Commons never earned a doctorate; and neither for that matter did Keynes! But the way the profession works, I wouldn't count on any definite publication threshold getting one over the no degree barrier.

[Steve Horwitz]

Kenneth Boulding is on that list too.

[Steve Miller]

Prof. Selgin,

I think grappling with an important topic at the level required for a book is an important skill for a scholar to develop. But to be honest I'm not especially convinced that it's a better first step (which is what a dissertation is typically, a scholar's first serious attempt at original research) than three publishable journal articles tied connected by theme and method. It's worked out very well for some, but my guess is not for most, and all too often has been similar to punch cards in its function as essentially a pure barrier to entry. Some have what it takes to write meaningful books right out of the gate, but I don't think that those people are the only worthy scholars. To me it's more about getting a research program established. The mainstream doesn't really value books, and I think that's a mistake. But why does that mean that step one to being a scholar is necessarily writing a book? Is it backwards to publish several articles in an area before making a more comprehensive contribution? I'm suggesting that it can work either way.

[Russnelson]

Dr. David D. Friedman has not a PhD in economics. He hasn't even taken a single economics course!