Bryan Caplan  

Tyler's Telling Question About My Education Book

PRINT
David Autor Signals Wit and In... Timothy Taylor on College Loan...
On Monday, Tyler asked me a telling question about my education book: Why do I keep re-interpreting other people's research to show that it's consistent with or supportive of signaling, instead of simply doing a new study of my own?  Here are my reasons:

1. My comparative advantage is in synthesis, not normal science.

2. On my reading, there's already more than enough evidence to convince reasonable, open-minded people that signaling is a big deal.  Assembling all this evidence into a coherent whole would therefore be a great contribution worth years of my life.

3. Many of the best arguments for signaling simply aren't publishable in academic journals because they don't fit the format.  Take my argument about the career consequences of failing versus forgetting.  Laugh if you must, but I think this blog post is a greater intellectual contribution than 95% of articles in top journals.  Why?  Because it might actually change the mind of a reasonable, open-minded person on an important topic.  Only in a book can I weave a bunch of these arguments together to create a lasting contribution.

4. One more academic paper on signaling would have little value.  The best case scenario is that a top journal publishes my paper and people cite it as a curiosity.  More realistically, it gets published in a mid-level journal and changes no one's mind.  Not even my own. 

If I thought the journal publication process were a good measure of the importance and quality of my scholarship, my forecast about my article's fate would be a strong sign that I should abandon the project.  Of course, by that standard, I might as well have given up on research as soon as I got tenure.  My real view, as you'd expect, is that academic journal publication is mostly a socially wasteful signaling game.  Hits in top journals are a strong sign of IQ, diligence, and creativity within the narrow confines of disciplinary conformity, but at best a weak sign that you've discovered important new truths.

Final point: Tyler is one of my main inspirations for writing books rather the articles.  His books are much better than the journal articles he could have written instead.  Tyler should take my imitation as the sincerest form of flattery.


Comments and Sharing





COMMENTS (11 to date)
Ross Emmett writes:

Good argument, Bryan. You've just convinced me that I should focus on books, also, since my comp. adv. is also in synthesis!

Clay writes:

If you want to be more than a novelty, you need an actual solution rather than just a well researched and argued rant.

Everyone already knows that higher education is largely about signaling and credentials. But people feel that there are various other positives that are important to preserve.

The only solution I've heard you make, is cut higher ed funding, and expand trade schools. Is the solution to the ills of society really just more electrians and computer repair people? I don't see a case for that.

Ken B writes:

I disagree with Bryan on a great deal. But I loved The Myth of the Rational Voter (rampant confirmation bias perhaps). And I'm willing to be persuaded on other topics. That will only happen if Bryan writes books; I won't hunt down his journal articles.
The argument deserves to be made, and it deserves a committed capable advocate. If not Bryan, whom?


Dan Hill writes:

@Ken B - I second that


Ironic that in this post Bryan is also showing quite effectively that journal publication is also about signalling rather than advancing knowledge!

Scott writes:

I don't really know about economics but journal articles in engineering and computer science are often of real value.

There are real, techniques and discoveries that have been researched and described. This is especially true in CS.

The empirically harder sciences are where journals make sense.

I'm sure there is a lot of signaling going on as well. But when I read Claude Shannon or Bonnie Baker, it isn't to see how smart they are. It is because they have information that aides my work.

BillD writes:

There's lots of signaling and bias even in the "hard" sciences. A lot of it is invalid, but ends up being published in prestigious journals.

http://cafehayek.com/2012/04/fake-science-everywhere.html

Mark Little writes:

All very true. I'm looking forward to reading The Case Against Education.

Scott writes:

@BillD

I would say that any scientific field that involves human case studies is not empirically hard in the sense I meant.

I read a lot of IEEE and SIAM journals. I think you would have a hard time making a case that they have the signaling problems being discussed in Econ journals.

I mean there is some, more than I would like (repeated articles about the material that will replace silicon in transistors come to mind).

But they are also a good source for problem solving ideas. Especially SIAM Journals.

Scott writes:

Here is a really interesting approach to research.

A CS "Web Journal?" That demands reproducibility to make the grade. The kind of thing that other fields just really couldn't do.

Andy writes:

Papers in computer science fall into a spectrum. Theory papers are just math, and those are generally pretty solid, since the reviewers will be following along to make sure the proofs are reasonable. There are errors sometimes, but they are rare.

Experimental papers (my algorithm/technique is faster/more effective/easier) won't necessarily be hard to reproduce (after all, software is mostly deterministic, so if you run the same code on the same machine you'll get similar results). But they can still be misleading in terms of their broader implications. For example your new method might only be fast for the data you happened to test, or it might be impractical for real world data sets, or it might be better but only by so little that no one cares.

chris writes:

mr. caplan have you written anything about the existence of government being about signaling or do you not have that view?

It would be interesting if the entirety of government existed so people will think someone is looking out for them, someone will protect them, someone is in charge... *evil laugh*

Comments for this entry have been closed
Return to top