Bryan Caplan  

The Signaling Model of Education Standing on One Foot

Health Care: Let the Games Be... Bequest Question...
Look at what people learn in the classroom.  Look at what people do on the job.  How much of a connection do you see?

Comments and Sharing

COMMENTS (38 to date)
Zdeno writes:

Contra your implied point, I see a lot of the same behaviours - effort for effort's sake, conformity to bureaucratic and political imperatives over actual education/value creation, and a superstar economy in which 10% of agents are responsible for 90% of the value-added (or, answers for a given week's micro assignment).

Of course, the lesson I draw from this isn't that the signaling theory of education is flawed, but that it should be extended to the majority of workplace activity as well.



Ariel writes:

But, in certain instances, people do learn how to learn. That will help them learn on the job.

Robert Johnson writes:

Depends on your line of work. As an engineer I use my training daily. I've learned many new things on the job, but almost all of them built on concepts I learned in school.

And Ariel's comment about learning how to learn is spot on.

Mr. Econotarian writes:

For my industry, I expect someone hired out of college to be able to do at least rudimentary computer programming. This is taught in some high schools today, but anyone with a BS in computer science or computer engineering degree has written a lot of programs in college, and the more programs you write the better you get.

I also want them to understand practical computer networking and PC operation and administration - setting IP addresses, gateways, netmasks, as well as skills in Excel and Word. While some of this is taught in college, much of it is simply gained by using a PC outside of formal college classes.

We deal with a lot of radio frequency issues. The knowledge of field, matter, and waves as taught in college is necessary to understand why things are built the way they are (such as coaxial cable), but my industry is fairly applied so we generally don't figure out the equations of fields, although whoever designs the cables and connectors must do this. If I use the word "polarization" I expect the graduate to know what I am talking about. Basic circuit knowledge (such as Kirchhoff's circuit laws) need to be known as well.

I'd say that the computer programming skills learned in college are most transferable to my industry, but the engineering and physics learned in college simply provide background for skills learned on the job.

Most of the real skills in my industry are arcane and not taught in any college. Most of the particular knowledge comes from a study of standards documents (ITU, etc.)

Do I care how many novels they read in college? If they know European history? Nope. I'd prefer they have business courses (such as accounting) so they understand that they are in a business to make money!

JPIrving writes:

I see a pattern, the more math the more the skills transfer into the workplace. The bottom two thirds of history or polisci majors probably don't use their college skills managing the Starbucks though.

So why does the government subsidize humanities... ?

Milton Recht writes:

What does a college education signal?

Why are there no alternative mechanisms to a college education to present the same signal?

Is there a lower incidence of college education among certain groups who can benefit without a college education, such as children of very wealthy families?

It might be interesting to see why some successful young adult actors and sports players obtain college degrees and some do not.

Floccina writes:

"Look at what people learn in the classroom."

What did they learn in the classroom that could not have learned elsewhere faster and cheaper?

Elsewhere would include: (apprenticeship, books, radio, TV, asking smart people, Podcasts, youtube or other online video, blogs etc.) ?

Floccina writes:

If we had to pay for schooling ourselves and there was a separation of testing and educating would you go to school as currently structured or learn in a non conventional way and take the test?

BTW in my list above I forgot tutors.

Dan R writes:

I am currently going for an MBA, and I have my Master's in Econ and Finance.

I have taken many finance classes, but my most recent one is associated with my MBA and is taught by a CFO, as opposed to a PhD.

When asked whether or not all the NPV sensitivity analysis we do for class actually happens in the average capital budgeting decision process, he looked at the student and brusquely answered, "Nah."

Bob Smith writes:

I share your intuitions about education being pure signaling, but a standard implication of signaling models is that there there should be negative social returns to subsidizing education, but I don't think any study has ever found this. Sure, the evidence for positive social returns is weak, too, but this doesn't so much help the signaling model as it hurts the human capital model. Maybe the best interpretation is that we don't really have good data on the effects of education, but this view cautions humility.

Joe writes:

Why does college have to be technical/vocational training?

Jesse writes:

What exactly does Bryan do for a living again?

Simon K writes:

For my job, lots. But I'm an engineer, as are all my colleagues and all our customers. Even the sales folks are engineering grads. You have to understand the vocabulary, priorities and trends of the industry or you'd be totally stuck, and that requires at least familiarity with the academic background.

Apparently this is not true for economists ...

Dr. T writes:

What fields are we discussing? In my fields of science and medicine, the correlation between what was learned in the classroom and what is needed on the job is high. A physical therapist uses the lessons from her anatomy class every day. A clinical pharmacologist uses his classroom knowledge of drugs, how they are metabolized, and how they interact with other drugs every day. A physician uses his classroom knowledge of anatomy, physiology, pathology, medicine, history taking, and psychology every day.

I'm a pathologist who used to be a chemist. Do I use all my classroom knowledge in my job? No. I don't use info from my differential equations, early 20th century American literature, or comparative vertebrate anatomy classes. But, I enjoyed all of them and benefited from what I learned. Job-relatedness should not be the sole selection criterion for deciding upon a class schedule.

The "college degree as signal" is more applicable for those majoring in business, sociology, humanities, education, etc. There are a few key courses and the rest are filler.

Bill writes:

Showing up. Following meaningless rules. Figuring out which written rules are actually rules and which are bs. Figuring out what the unwritten rules are. How and when to lie. Figuring out which of your colleagues is useful to you and how to get what you want from them. Managing jerks with formal power over you. Managing random bureaucratic/clerical jerks with informal power over you. Figuring out the locally approved buzzwords, rationales, taboos, attitudes, social norms and using them appropriately.

In short, figuring out what is *really* going on and what you need to do to attain your goals subject to those constraints. Someone has to bear the costs while you thrash around to learn this stuff --- given parental and gvt subsidies for university, why not parents and gvt?

I.e. Zdeno's answer except more dimensions.

SydB writes:

In the engineering and technology field I'm familiar with, the correlation is high. True, one learns a lot (in terms of breadth) in college while one's job assignment is by nature specific. But flexibility is important, and a top-notch corporation wants people who have experience and knowledge that extends beyond the narrow limits of what they do on a day to day basis. College gives kids time to demonstrate interest. And technology corporations want people stock piled with ideas--not code pigs (as they are called).

True, college is also signaling, but the four years of a technical degree are invaluable for creating domain appropriate depth in students that can be then be measured by interviewers.

While he's not said as such, I suspect Caplan believes an IQ test is all that is necessary to provide the appropriate signal. If that is the case, I suggest he and his research associates hire high school graduates--or even middle school graduates--with high IQs and then create the next generation of nobel laureate economists.

I think they'll have a high fail rate.

That said, I just finished reading American Higher Education by Lucas and come away with the opinion--one I had prior--that we still don't have a clue what higher education--and education in general--is for. We're just doing it because we've been doing it.

Mike Rappaport writes:

I am a law professor, so what I say does not really speak to the college experience, but for what it is worth: Much of what we teach is either used by lawyers or is done to give students an opportunity to decide what area of law might interest them. Thus, a law student learns about litigation, which helps them decide they don't want to litigate.

My guess is that we could make the law school experience more like practice (in a desirable way for learning purposes)by about 20 to 30 percent, but instead because of inefficiency -- law professors don't want to do it and there is not enough competitive pressure to do so.

I am not sure that really supports the signallig theory that much. I guess you could say there would be more competitive pressure if this 20 to 30 percent did not signal.

Against the grain writes:

As an engineer who recently became a teacher, it is completely amazing how socially skilled but technologically challenge a 16 year old can be. While challenging and even given there may be more efficient ways to educate. I can see a world of difference between a high school sophmore and a college graduate. I agree that signalling is part of the game, but growth and knowledge happen as well.

The law prevents employers from describing and officially testing for job-specific knowledge. If employers could describe and give their own tests, then prospective employees could qualify by acquiring knowledge in any way they wanted, including self-study.

This is why "experience" is so highly stressed. It is used to validate a person's knowledge without giving a specific test.

I hired programmers as part of managing a software group. The HR department told me that I could ask technical questions, but to never write them down in any "formal" way. They were worried that someone would claim I was giving a "test". Any test was illegal unless proven to be non-discriminatory in effect when applied to different races.

James Taranto of the Wall Street Journal: [edited]

Most professional jobs require basic intellectual aptitude. Since the 1970s the Court has developed a body of law that prevents employers from directly screening for aptitude.

In Griggs v. Duke Power Co. (1971) a black coal miner claimed discrimination because his employer required a high-school diploma and an intelligence test as prerequisites for promotion. The court ruled 8-0 in the miner's favor. "Good intent or absence of discriminatory intent does not redeem employment procedures or testing mechanisms that operate as 'built-in headwinds' for minority groups," Chief Justice Warren Burger wrote.

This became known as the "disparate impact" test, and it applies only in employment law. Colleges and universities may use aptitude tests. Elite institutions lean heavily on exams such as the SAT in deciding whom to admit.

For a prospective employee, a college degree is a very expensive way of showing that he has, in effect, passed an IQ test.

College is an Expensive IQ Test

More on the value of college.

Simon K writes:

Andrew, there's a huge gulf between the disparate impact test and "you can't ask technical questions" because you only have to prove business necessity, and since 1989 the test for business necessaity has been extremely low. You do not have to prove non-discriminatory effect as long as you can show business necessity. Its obvious that technical knowledge is a business necessity for software engineers in a way that its not obvious that a high school diploma a required for a coal miner. I think your HR department is excessively paranoid.

Johan writes:

I use lots of what I learned in university in my current job but then I am an academic mathematician so it does not say much.

My more important point is that Bob Smith is right. So what if the human capital model is wrong? That does not in itself prove that the signalling model is right.

I can think of numerous other models. Perhaps people like studying, certainly some people do. Perhaps people are mistaken about the benefits that college gives, certainly some people think you learn important things there even if they perhaps are wrong.

Some people study because of occupational licensing requirements, which in turn might be the result of rent-seeking from unversities.

The above are just some alternatives to the signalling model of higher education and none of them require that you actually learn anything valuable in university. (And of course you do learn some valuable things in university, at least for some jobs.)

fundamentalist writes:

There is very little correlation between what workers need to know on the job and what is taught in school. McKinsey and Co have an interesting paper on the fact that education that increases productivity takes place on the job, not in the classroom.

eccdogg writes:

Me personally quite a lot.

I studied Economics and Finance with quite a bit of Math and Statistics thrown in.

I use the knowledge and skills I learned every day and have in every job I have had since I graduated. I work on a trading floor as a quant/structurer.

My wife is an engineer, she uses her training everyday. Most of the people I know and work with use thier training every day, but I don't know many people who got Liberal Arts types of degrees so I am probably biased.

Jaap Weel writes:

I think there is one big exception to the college-as-signaling-and-nothing-but story: engineering.

Engineering degrees are the most lucrative degrees for a reason. (Check any listing of lucrative college majors, e.g. Engineering and math classes actually teach you something you might use on the job and that you wouldn't easily pick up. Engineers really do solve partial differential equations, perform finite element analyses, design asynchronous integrated circuits, and so on. While most computer programmers put together tedious data processing applications that require nothing but elementary programming mixed with generous helpings of patience and attention to detail, the top 5% or so of computer programming really does involve error correcting networking protocols, optimizing compilers, algorithms to find the shortest path, and so forth.

That does not take away that math, science, and engineering degrees also have tremendous signaling value. Wall Street is teeming with physicists, and none of them do physics. The fact that physicists happen to know a lot of applied math may explain why they become quants for some hedge fund, but not why they become management consultants for McKinsey. As far as I can tell, for McKinsey, a physics degree is an elaborate sort of IQ and personality test.

This leads to another interesting question: why is engineering the one profession directly and usefully taught at the undergrad level? My answer: there is no good reason why law or medicine cannot be taught to undergrads, but the bachelor's degree is used purely as a barrier to entry in those fields. In fact, other countries do teach law and medicine to undergrads, and usually come up with other barriers to entry to make up for it.

SydB writes:

Few more comments, some based on previous ones:

1. My technical experience includes significant interviewing, which involved candidates solving tough problems, going to the white board, writing stuff down. The problem were stated verbally, not written. But testing all the same. So I don't agree with the above comment that says testing is not possible. And my own two cents: a written test is probably not a good way to gauge competence--at least in a technical environment.

2. My intuition tells me that foreign universities are much more pure signaling than American ones. And the candidates produced by the foreign ones tend to be pretty bland creatively speaking. It is the US degree programs (e.g. Berkeley engineering) that seem to produce flexible candidates who can engage with real world problems. And I don't think this is just signaling. Something happens on the ground with the people receiving these educations.

Tracy W writes:

Rather high, but then I did degrees in engineering and economics. A lot of what I use from the engineering degree is not the technical knowledge but what I learnt about testing and quality management.
Although ironically, most of what I know about writing I learnt either in the Latin courses I took in high school or on the job, not in English classes.

Johan writes:

I realized that the models I tried to answer the wrong question before. The difficult question is not why people pursue higher education since employers clearly value it. The question is why do employers value higher education?

Two of my suggested models still make sense however. Number one, it is possible that employers are simply mistaken about the value of higher education. I know it will sound impossible to an economist but I consider it a possible but unlikely theory. Perhaps employers are discriminating against uneducated the way they used to discriminate against women.

The occupational license model also still works. It is simply illegal to hire a person to audit your corporations books if they haven't received certain education. This is also important for the legal and medical professions.

Brian B writes:

"Look at what people learn in the classroom. Look at what people do on the job. How much of a connection do you see?"

Well, that depends. Does anyone you know read, write, or communicate with other people at their jobs? It's obvious that college provides crucial training for technical jobs (as others have noted), but it also provides crucial training in soft skills that have wide applicability.

If you think not, then why do college students in philosophy, english, history, etc. on average score higher on the GRE verbal than physics, mathematics, and economics majors, despite the latter students having demonstrably higher IQs? Do you really want to argue that humanities students don't learn anything of value?

Now, perhaps you want to argue that these skills can be learned without college. I wouldn't disagree. But the reality is that, for most young adults, these skills are NOT learned without going to college. Most people are just not driven enough to do it on their own, so the opportunity costs of not going to college remain HUGE.

In addition, one shouldn't underestimate the role of peer selection. Where else can one go and be assured of interacting with lots of other people who value rational thought? The academic community is a powerful force for shaping the human mind into a powerful organ.

Really, this is what a liberal arts education is all about. Higher education is the grease that smooths the action of human society. Honestly, Bryan, my freshman students understand these points very well. Why don't you?

SydB writes:

Tracy W wrote: "Although ironically, most of what I know about writing I learnt either in the Latin courses I took in high school or on the job, not in English classes."

Very true. In areas like grammar, linguistics, rhetoric, argumentation, and style, our education system--primary, secondary, and advanced--is lacking. Science and engineering present the student with a toolkit. A similar kit could be provided in the area of writing, speaking, and communication in general. It isn't. This is a shame.

To Simon K (89963):
I did ask technical questions. The HR department was frightened that anyone would regard these as a "test".

You are unafraid. The HR people were afraid of going to court and showing that any test, and each and every part of that test, is a business necessity. This would have to be shown by prior evaluation of that test, not just a (white, biased, discriminatory) opinion of what is necessary. The HR department wanted to stay far away from any such lawsuit.

It is great that something is obvious to you and me. Courts are different, and technically ignorant.

The case about the coal minor was about promotion. You may improperly think that coal mining supervision or management is a profession for the stupid.

To SydB (90046):
You wrote: "I don't agree with the above comment that says testing is not possible."

I didn't say "not possible". But business is restricted to placing the role of testing onto individual managers, who apply a verbal tradition within each company. Testing for hiring cannot be written, refined, and standardized to effectively train the hiring managers.

Prospective employees have no choice but to train as best they can at schools that teach things unspecific to the needs of any particular company.

Why is a technical degree from MIT worth more than one from less prestigious schools? All schools have access to the best textbooks, and MIT does not employ all of the highly qualified instructors and professors. I make the same point for a liberal arts degree from Harvard compared to "lesser" schools.

I suggest the difference is the admission standards and selectivity at MIT and Harvard. It is the aptitude and dedication of the student that is most important, not the knowledge available at a particular school. A prospective employer relies on the selectivity of MIT to screen applicants. The employer knows that these smarter, more dedicated students have been given a more difficult and superior education, tailored to the brightest.

Smart students can get a good education from most schools, or even from personal study. But, employers rely first on the screening of the "best" schools. The best schools rely on intelligence testing and many other measures to select the best students going in. Schools do not have to prove to any court that what they measure is a "necessity" for admission.

CJ Smith writes:

@ Dan R:

If your finance professor is saying that you shouldn't do NPV of other financial analysis because "nobody bothers with that stuff" you ought to run, not walk, to your investment advisor and short your finance teacher's company to the maximum possible. He's an idiot if he thinks he and his buddies' "gut feelings" win out over fundamental financial analysis in the long run. Sure, he may be successful now, but consider both the long run and realize how much MORE successful he and his companies could be if they actually used the business tools that were developed to make them more competitive.

While with Coopers&Lybrand n/k/a PricewaterhouseCoopers, I consulted with any number of companies and governmental who thought their rose-colored opinions were better than a simple "best case, worst case, most likely case" sensitivity analysis. Most of the CEOs an CFOs of organizations like that are now either unemployed or working as bankruptcy consultants - because they've taken so many companies through bankruptcy.

@ Mike Rappaport:

I agree that in law school we didn't generally didn't learn what the clerk of court's direct line was, or how many copies of motion and supporting memorandum of law to submit. We also spent a lot of time studying areas of law that we don't now practice in. But I don't think the point was to teach the lawyer how to be a procedures wonk or a field specific attorney - it was to develop lawyerly skills that are applicable in pretty much all areas of legal related work such as abilities to:
1. "teach yourself" on the fly;
2. critically evaluate another's position, finding the merits and demerits of both sides of the argument, and then responding persuasively;
3. parse oral and written arguments in a logical and/or persuasive manner.

So yes, I think even law school taught job relevant related skills.

eccdogg writes:

"Very true. In areas like grammar, linguistics, rhetoric, argumentation, and style, our education system--primary, secondary, and advanced--is lacking. Science and engineering present the student with a toolkit. A similar kit could be provided in the area of writing, speaking, and communication in general. It isn't. This is a shame."

I count myself as very lucky to have taken just such classes in college. Technical writing was a required course and I lucked into taking a very good communication course. Both taught the practical aspects of writing and presenting and communicating with others in a way that insured that your message was received.

My mother is a highschool english teacher, and I often discuss with her why literature has such an important role in highschool english. It seems that the skills of critical reading and clear writing could be learned without such a narrow focus of subject matter and would seem applicable to a wider number of students.

Syd writes:

Andrew_M_Garland: I suspect you can create an interviewer training program in your workplace for all managers. We did. This can be as formal--and written--as you want.

Why is a degree from Berkeley, Cal Tech, or MIT worth more? True, signaling and sorting is taking place. And Justice Scalia says about the Ivy's: "You can't make a sow's ear out of silk purse." So he only hires from the Ivy's. Yet the best person he's had on his staff came from Ohio or such. He didn't hired him. Inherited him. So what you say is true. The Ivy's are signaling yet there are great people elsewhere. But--and here's the issue for me--at top notch universities the student is exposed to a social environment with top notch researchers and top notch students. That creates an environment that cannot be duplicated at a lower run school. It's about competition, first and foremost. Ideas floating about. I will never forget my professors nor the time the son of a nobel laureate (classmate) did some things on a blackboard that amazed me. And pushed me to try harder.


More on writing programs: I'm also quite disappointed that our education system doesn't focus on the nuts and bolts of story and narrative. It's a primary component of thinking--in addition to analogy and logic--yet it's not in the curriculum. It stinks.

Dr. T writes:

Jaap Weel writes: "This leads to another interesting question: why is engineering the one profession directly and usefully taught at the undergrad level? My answer: there is no good reason why law or medicine cannot be taught to undergrads..."

Mr. Weel is bit engineeringocentric. Chemists and physicists and medical technologists with B.S. degrees jump right into good jobs.

Medicine cannot be taught well at the undergraduate level because the basics (chemistry, organic chemistry, biology, etc.) need to be taught first, and because youngsters need to mature.

I went to a medical school where some students had only two or three years of college. Some of those students had skipped grades and graduated high school at 16. I had two 18-year-old and seven 19-year-old classmates. The Doogie Houser model doesn't work. These students were smart enough for medical education, but they were too immature. They also had no life experiences beyond being students living with their families. When it came time to see patients in the second year, most of these students did poorly. They didn't know how to question adults. They couldn't handle sexual history questioning. (This was in 1980 in Brooklyn, and we were trying to find the causes of health problems in the gay community. We took extremely detailed histories.) My point is that maturation (not signaling) is needed before medical training. A similar point could be made for law.

Scott Gustafson writes:

Might I suggest that you're asking the wrong questions. Look at what people use on the job and in their own personal lives. Look at what you're teaching in the classroom. If what you're teaching isn't useful to people, why do they bother with your class?

I didn't comment earlier on this topic because I was busy grading a microeconomics exam on production & cost, perfect competition and monopolistic competition. All ten short answer questions covered things that a small business owner must know and understand in order to be successful.

I know what I'm teaching. What are you teaching?

Vlad writes:

As a working programmer of a few years only now beginning his Computer Science degree, I can safely say that a degree is a waste of time compared to experience.

A small percentage of programming does require special knowledge that is taught in colleges, such as algorithms. But such topics are covered by 3-4 courses in college at most, and I'm convinced they can be learned on the job. Mathy stuff are even rarer (despite this, the first year of a CS degree is almost the same as that of a math degree).

One thing I find curious is that professional literature and academic literature are very different from each other. They don't even discuss the same topics. Perhaps we can measure the connection between formal education and job skills on this basis - the more overlap in literature, the more connection there is.

What books do engineers, doctors, nurses, lab technicians, therapists, architects and so on read on their job? Are they the same books studied in college? Do they discuss the same topics, use the same language?

The White Detroiter writes:

I'm a certified public accountant and I majored in accounting in college so there is a relatively strong connection between what I studied in school and what I do for a living.

That being said, a lot of what I learned in college is either irrelevant to my job, redundant, or outdated. The length of time required to earn degrees is rediculous. You need 150 credit hours and at least a bachelor's degree to become a CPA. There already is a challenging CPA Exam in place so much of this forced schooling is wasteful.

Daniel M writes:

I quoted you in my recent opinion column on the subject.

Comments for this entry have been closed
Return to top