A simple question about the signaling model of education

Let’s say, for purposes of argument, that education is 100% signaling, and furthermore let’s assume that the underlying traits of IQ, conscientiousness, and so on are not changing in the population over the relevant period of time.

Now consider a situation where income inequality is rising, at least in the early years of jobs.  Since employers cannot discern worker quality — other than by observing the signal that is — this should imply that getting an education is “more separating” than it used to be.

That in turn has to mean that an education is more rigorous than it used to be.  No, not “getting in” (employers could hire their own admissions officers), I mean getting through.  Finishing successfully is more of a mark of quality than it used to be, because finishing is harder.  Finishing is harder because there is more rigor.

Is this true?


Griggs v Duke Power

Griggs only applied to janitor jobs, but people certainly seem to misinterpret it.

Corporations have admissions officers, and then can access quality. They do this by administering IQ tests and pretending they are something else. But because they are busy, they only want to administer them to people they think will pass, and that is correlated with being admitted to schools that do IQ admissions screening. It is hard to discern people who were admitted but declined/quit/failed since people don't put that on their resume by convention. Since there is no rigor and finishing it trivial once admitted quiting/failing would also look kind of bad and someone who declined might be young and immature at first. But I suppose if you can find a Thiel fellow you should interview her.

I suppose people pay more unequally because they IQ matters, or conventions suggest it does. Low IQ people can certainly waste everyone's time and find novel ways to lose money.

Procter & Gamble paid a fair amount of money to jump through various federal hoops to validate their highly IQ-correlated hiring tests, and pretty much succeeded:


On the other hand, the marketing research company where I worked started off using one of the founders' Quantitative Methods in Marketing Research 302 exams. This was a brutal test -- I worked on it for like 3 hours, but the next day the CEO called me up to invite me back for a full day of interviews. In general, people who did well on it tended to extremely bright.

Eventually, we grew big enough that the EEOC noticed we were using a college test to hire people. At that point we could have ponied up a fair amount of money to have a consulting firm validate Professor Eskin's test, the way our big client P&G had successfully defended their tests, but our firm made the penny wise and pound foolish decision to dump our hiring test.

My (highly prejudiced) impression is the quality of our hires declined once we lost our test.

The real shame is that the federal government discarded the superb civil service exam it had designed in the 1970s when the outgoing Carter Administration in January 1981 folded in the Luevano discrimination case and abolished the federal PACE exam.

"Griggs only applied to janitor jobs, but people certainly seem to misinterpret it. "

Well it certainly seems that you've misinterpreted it.

I'm going to back up Sailer, and say that "People select by the criteria available to the people doing the selection, and in the absence of criteria they rely on intuition, and intuition merely selects for people you like", and this makes hiring fully dependent upon the first line of recruiters and interviewers.

We found (as did Microsoft, and many others) that putting people through a lot of interviews across the team was extremely expensive but tended to filter negatives more so than select positives, because positives are obvious.

IQ tests are very hard to fake. Personality tests easier, and cultural affinity (values) tests easier. So, they have to be painfully long in order to filter the liar problem, and again this is expensive. But that time constraint in and of itself is a filter.

Best indicator of hiring likelihood is the genetic, cultural, and class distance between the interviewers. Some of us organize recruiting organizations to take advantage of that fact.

You would have to say that education has always been 100% signalling. Otherwise the increasing inequality could reflect increasing reliance on the education signal (which is more likely given increased urbanisation).

Tyler, can you do something about the Marxist infiltration of our schools? I'm tired of see kids these days that can't do reading, writing, or arithmetic but can spout cultural marxism, feminism, and black lives matter like the best of em.

Where do you see these kids? I do alumni interviews for my alma mater and talk to 20-30 kids a year. Most of them have far better math skills than I did, and I have yet to meet one who could talk intelligently about Marxism.

Nowadays, there are CEO women, Blacks sitting on front seats at the bus and lynching is frown at. Hence "cultural marxism". https://giphy.com/gifs/season-16-the-simpsons-16x5-3orieUCTT4CsGXlRTi

Actually it's the lynching that is objected to as cultural marxism, and sadly it has not gone out of style

America is a systematically racist country. There will be no peace unless America makes amends.

In what way is it systematically racist? In systematically favoring one skin color over another in college admissions and hiring?

Concepts like "the Patriarchy" are obviously derived from Marxist ideas of class struggle. Its not hard to trace their pedigree. Doesn't mean they are wrong or useless, but cultural Marxism is a pretty accurate label for this kind of thinking. I can't understand why so many people disagree.

You do not know much about Marxism clearly. The idea that there are struggle between various human groups based on race, religion, or sex is as old as the world. The idea that all these distinctions do not matter, and that the struggles they underline are nocive diversions for the only struggle that matter, based on the only important distinction , the distinction in social classes based on one's role in the productive system, is the essential idea of Marxism. So quite the opposite of what you say.

I don't think your model is right -- the sign of the effect of rigor on wage differentials is ambiguous.
Here's my counterexample:
There are three worker qualities: 1, 3, and 4, equally represented in the population.
Workers are paid their estimated marginal product.
Before, only workers will quality 4 get education. Their wage is 4.
Workers without education get paid their average quality -- 2.
Wage difference is 2 (one third get 4, two-thirds get 2).
Now let education get less rigorous, so that workers of quality 3 and 4 get educated. They'll get a wage of 3.5.
The uneducated get a wage of 1.
Wage difference has gone up to 2.5 (two-thirds get 3.5 and one-third get 1).
Putting more people through education increased the wages of the incremental educated workers (quality 3), but lowered everyone else's wages (quality 1 is hurt because they no longer pool with quality 3, and quality 4 is hurt because now they pool with quality 3).

+ 1

Interesting analysis.

To add to Steven's analysis, consider a population in 6 buckets. The buckets increase in productivity by 1, so 1, 2, 3, 4, 5, 6. The number of people in each bucket is 1, 3, 6, 6, 3, 1. People are either educated or not educated and are paid the average productivity of their type.

Education bucket level - Educated wage - non-educated wage - difference
none - n/a - 3.50 - n/a
6 - 6.00 - 3.37 - 2.63
5 - 5.25 - 3.06 - 2.19
4 - 4.50 - 2.50 - 2.00
3 - 3.94 - 1.75 - 2.19
2 - 3.63 - 1.00 - 2.63
All - 3.50 - n/a - n/a

So as education becomes less rigorous, income inequality decreases until a certain point, and then starts increasing again. If income inequality is increasing, perhaps that's a sign that too many people get educated. Also note that as education becomes less and less rigorous, life for the uneducated gets worse and worse.

Someone more awake than me could probably throw some basic calculus at this and prove it more rigorously.

I don't know if I'm misunderstanding the post, but there is of course the selection bias: the smarter ones are much more likely to get more education, doesn't mean education adds value.

Not quite. Since the institutions of higher education are highly heterogeneous, it can also imply that getting into the right one takes a lot more conscientiousness and conformity. As indeed we observe in the growth of the crazy hoops of extracurriculars, diversity, charity work and admission-essay-writing that entrants to the Yales and Harvards of America are expected to jump through.

This is the correct answer.

Honest question here, if (higher) education is all signalling, then what exactly should the human race do to pass on information and knowledge from older to younger, every generation?

I submit the structure and socialization of a 'school' is actually how humans at least see the accumulated knowledge before them. Even kids that go and decide to drop out and become a billionaire, they at least experience a place where past knowledge is meant to be heard and seen.

How to pass on information to the next generation. Sing. Reminisce with elders. Recite epic poetry.

Admittedly the “Dance of the SQL Insert syntax” is not going to be very efficient but for professional knowledge you give the answer in your comment below; guilds and apprenticeships.

"if (higher) education is all signalling"

No one says its ALL signaling, obviously, the civil engineering curriculum is more than just signaling. The question the signaling theory answers is "why do all these jobs which didn't used to require college degrees and don't seem to have changed much now require them?"

Good question. Fortunately, Caplan notwithstanding, higher education is not all signaling.

A lot of it - I'd say almost all - is about learning interesting and/or useful stuff.

I had a friend in college, a very smart guy, who wanted to go to medical school, which in fact he did. But he took a lot of completely unrelated classes. The one I remember was a class in Greek mythology, because he would regale us at lunch with the latest tales of Hercules or whoever.

My point is that if you see people only as economic actors you will make many mistakes.

Is there a model where kids figure out what they are interested in? and they have to be introduced to something at a fairly high level to figure out that interest. That fits my professor teaching intuition pretty well. Students that like writing code will only know it when they have been pushed through a high level project. (except for a few geniuses.)
I tell my non-economics inclined students to take a micro class to see if they enjoy the thinking style.

The question I would then ask is what the failure rate is. If college works as a hard work signal, what percentage of people going to college drop out or fail?

What's the incentives for colleges, particularly in subjects that are rarely applied to make easy courses to get people passes? What's the disincentives for them to do that?

The failure rate could have little to do with IQ or academic performance. It could be due to the soaring cost and resultant debt.

The most commonly reported number is graduation rate, usually measured after 6 years.

"The 6-year graduation rate for first-time, full-time undergraduate students who began seeking a bachelor's degree at a 4-year degree-granting institution in fall 2009 was 59 percent. That is, 59 percent had completed a bachelor's degree by 2015 at the same institution where they started in 2009.

Six-year graduation rates for first-time, full-time students who began seeking a bachelor's degree in fall 2009 varied according to institutional selectivity. In particular, 6-year graduation rates were highest at institutions that were the most selective (i.e., had the lowest admissions acceptance rates) and were lowest at institutions that were the least selective (i.e., had open admissions policies). For example, at 4-year institutions with open admissions policies, 32 percent of students completed a bachelor's degree within 6 years. At 4-year institutions where the acceptance rate was less than 25 percent of applicants, the 6-year graduation rate was 88 percent. "


One could just as easily conclude the opposite given the premise, I think. Rather than education being more rigorous than it used to be, it is less rigorous; anyone who cannot pass it successfully is signaling that their underlying traits are especially low, lower than would have been the case in the past. The income inequality, then, is not fundamentally because the high end is getting paid more, but because the low end is getting paid less.

I think, by the way, that this is one of the tragedies of the present push to send nearly everyone to college: those who honestly don't want to go, or who are idiosyncratic enough to pursue other interests, are lumped in with those who *can't* succeed in college, and (hypothetically) lowering standards mean the quality of the latter group will decline.

Good comment. I wonder about returning to an older model, with guilds and apprenticeships, but can that scale in a world of 7 billion+?

I think its not the 7 billion per se, its that today 98% of the population are not subsistence farmers. Will it work for 50% of the population rather than 2%?

Manufacturing (in the US) is down to 10% of the working population, and seems headed toward the 2% for agriculture. Do guilds and apprenticeships apply to service industry jobs? Do apprenticeships even make sense for jobs where mastery takes a few months (McJobs) rather than the traditional 7 years?


+1, I also fail to see the correlation between rising inequality and more rigorous job requirements. Perhaps TC meant fewer jobs that cannot be automated rather than rising inequality? In any event, I think perhaps the opposite is true: due to grade inflation, there is less signaling than before. So whether or not you get a good job depends on the old boys network, same as it ever was.

We've replaced "rigorous" with "time-consuming."

I'm not 100% on this theory being true, but it completely fits the evidence: What used to count as a high-school-level of mastery from 60 years ago is now a college-level of mastery. People from higher SES can easily afford to put more time into education. (This is true even if education is "free" to the user.) What we are measuring is how long your family can afford for you to not be working, and since a lot of the big-five are highly heritable this turns out to be a very good signal. It's also one that filters out the lower SES but in a "safe" way for the corporations.

You almost did a reversal there. Was it implied, or actually avoided?

Yes, you could start with the idea that education was signaling, but then with increased rigor, more skills transfer, completion becomes something else.

Otherwise you have the silly notion that learning to analyze a mass spectrometer output is just an arbitrary hurdle chemists picked up along the way. A way to make "completion" harder.

Obviously that depends on whether reading a mass spectrometer is something that will be at all useful in the rest of that person's life

Reading spectra is a reasonable expectation for chemists, and going into a lab that does without the skill would be a day one deficit.

Finishing is hard, but for many different reasons - academic, financial, and others. Anything can go wrong mid- semester, but students who are financially well supported are better able to weather these events than their less fortunate peers.

Of course, but not around here. Around here those things all reduce to "family ability" and ultimately the "deserving" rich.

I think this is an arch and spandrel thing. Skills transfer is the arch. The subsequent filtering effect is the spandrel. But for some reason economists want to see the filter standing alone.

Maybe they just need more grounded skills and knowledge. Is it a coincidence that education as signaling comes out of the wreckage of macroeconomics?

And that "some reason" is evidence and the scientific method

lol, no.

This is a rejection of measured skills transfer for some phlogiston instead.

You didn't say anything about the society not changing, so that could be another explanation.

Classic example would be automation, where the higher skill end gets paid more due to increased productivity, but the lower skill end doesn't find a job anymore because they are outcompeted in everything they are capable of doing.

In the best case, the production or operation of automated machines is slow enough or constrained by ressources so that there are still some jobs left, but since the companies will definitely let the machines work in the fields that would be most expensive otherwise, the lower skill end will still earn significantly less.

But if there is no such constraint, comparative advantage doesn't work anymore and they are completely screwed.

automation, where the higher skill end gets paid more due to increased productivity

A common mythology that has no basis in fact. Salary and wages are determined by supply and demand, just like every commodity.

'Salary and wages are determined by supply and demand'

Somebody has probably never seen ever nepotism nor the good ole boy network in action. Or at least recognized it.

To give something along the lines of analogy - the price of diamonds has not acted like a commodity since De Beers' extremely successful marketing created a market that demanded something it was unaware of wanting beforehand.

Not everything is simply reduced to straightforward supply and demand - as De Beers, a monopoly supplier for generations, demonstrates.

De Beers is a perfect example of supply and demand

No. DeBeers is not a perfect example of supply and demand. As a monopolist it sets the price, given a demand curve and cost structure.

...and there is a lot of demand for high-productivity workers

Re: Finishing successfully is more of a mark of quality than it used to be, because finishing is harder. Finishing is harder because there is more rigor.

Is it? There are common tests across universities which give estimates of the quality of the respective students. Then there are the graduation rates. Among the universities on average there is a general consensus about the graduation rates and students quality. However, there is a university whose graduation rate is 2.73 SD above the norms with respect to the student quality.

Deviation from the regression line GradRate = 0.124*Senior -86.03; #n=67; Rsq=0.4384; p=1.05e-09 *** (VVSig)

Quality University

zscore ExpectedRate GradRate SeniorScore

-2.39 53.87 31.0 1128 U1

-1.76 64.79 48.0 1216 U2

-1.67 58.96 43.0 1169 U3

Few Students Left Behind

zscore ExpectedRate GradRate SeniorScore

+1.79 45.94 63.0 1064 U8

+1.81 55.73 73.0 1143 U9

+2.73 42.96 69.0 1040 U10

This engineering professor ponders on that question. https://profbillanderson.wordpress.com/2013/05/22/engineering-failure-rates/

What is finishing successfully? What is failure? Prof. Anderson shows there's a blurry divide between them. Even more, "failure" may be just one step on the path to successful skill development and happy life for people.

It's not clear that "more rigorous" and "finishing is harder" means a "better selector for underlying traits which correlate with on the job performance" for employers.

Imagine a more rigorous liberal arts department and a less rigorous engineering department. Which one will students interested in delivering economic value immediately out of school choose? Which one will students with a love of theory and distaste for "applying theory" choose?

Signal vs human capital theory of education... I think it is like the wave-corpuscle theory in physics. Despite having totally different or even opposite conclusions whether you suppose one or the other, you have to admit it is both at the same time. And I think you can find the same duality in every aspect of economics.

Newly-minted aeronautical engineer: look I can design an airplane that actually flies!

Economist: No, you have just completed an arbitrary course of studies, which filters for people who just happen to design airplanes. Thus signalling that they can design airplanes.

N.m.a.e: You guys are weird.

If aeronautical engineering is anything like software engineering which is the field I’m familiar with.

Old grizzled aeronautical engineer: Oh look, here’s another young whippersnapper I have to teach the thousand things not mentioned in his textbooks that you actually need to know to design a plane that can fly.

Of course, humans learn their whole lives.

But still a degree in a serious program is still skills acquisition by the shovel full.

Yes but is it an “arbitrary set of studies” or the skills actually needed to design an aeroplane? I can see possibly having an “Engineering Arts” curriculum that contains the basic skills any type of advanced technical professional should know but specific niches such as aeronautics would be more effectively taught on the job. A bachelors degree in the field really is just signalling.

I was a chemist hired to be a programmer, back when programmers were rare. I worked with an ex electron microscopist. So I know that signalling can be a path.

But to go from tbere to "a bachelors degree in the field really is just signalling" is just silly.

Not least because signallers like me had a lot of catch-up to do. When it is time to write a language parser you have to start with more basic books. You start with a skills deficit.

In the worst case you don't know in which skills you are deficient, which is why bachelor's degrees tend to survey the landscape. Transfer signallers need to take up that survey themselves.

The first programmer I ever hired was a BS/MS MIT/Stanford electrical engineer, who had worked in the application domain for 2 years between BS and MS. Only incidental programming experience, he started the first day with a C manual.

He proved to be an exceptional employee, and wrote software for the firm for 25+ years.

We hired in those days with a strong bias for smart, and what can this person do over their career rather than just next week. Lots of PhD physics people, for instance. STEM degrees were and I think still are a pretty good filter for smart and conscientious.

You are confusing necessary and sufficient conditions.

Just because the new engineer may not be able to design a plane doesn't mean the skills learned are not needed to design one.

Knowing arithmetic doesn't mean you can design an airplane, but I bet you can't design one without it.

Do you actually have an aerospace degree? Because I do, and I can tell you that not one person graduating with an aerospace degree can design an airplane. Not one.

I don't mean from a clean sheet, but I mean by examining home-build plans, choosing elements, running models, designing an rc scale model etc.

Ultralights, not jumbo jets.

I think a person with an aerospace degree would have a rather marginal advantage over an equally smart person without such a degree in designing an ultralight airplane. I mean if someone told me to design an ultralight airplane, my first step would be to Google it and I would go from there.

Do you know less about modeling a design than a smart person off the street?

"Economist: No, you have just completed an arbitrary course of studies, which filters for people who just happen to design airplanes. Thus signalling that they can design airplanes."

It's always funny when dumb people are sure they've found a flaw in the reasoning of smarter people.

'completed an arbitrary course of studies, which filters for people who just happen to design airplanes'

Actually, the course of study is far from arbitrary, based on the reality that designing a workable aircraft is not about signalling, but engineering.

"Let’s say, for purposes of argument, that education is 100% signaling .."

We are left with two choices. Either smart people say dumb things because they believe them, or smart people say dumb things because they think silly premises lead somewhere productive.

So, if we want to start thinking about what happens to projectiles when they're fired into the air, should we start by accounting for wind resistance, mascons, the rotation and curvature of the Earth, imperfections in the shape of the cannonball, etc? Or might it be smarter to reduce the situation down to the simplest situation we could think of (spherical cannonball, no friction, flat unmoving Earth) and then extend it gradually to deal with all those complexities when they're needed?

Or should we just say the projectile signalled its ability?

What were you saying about smart people saying dumb things? Or was it dumb people saying dumb things?

That was a good joke, playing as it did on albatross's error.

He lists a bunch of stuff he learned in a physics course. Knowledge transfer for the win.

Even if he did not have the introspection to see it.

As far as I can tell he has not taken any position on that question, so I don't see that he has made any error.

Isn't the signalling model becoming less true?
One obviously cannot and never could have become a successful doctor, lawyer, engineer (civil or computer), physicist etc. just through signalling.
Finance used to be the obvious exception but most firms are now looking for harder skills, esp. mathematics. And of course entrepreneurs who start their own firms don't need to signal to anyone.

A lot (~75%) of the recent college grads we hire, mostly engineers and a few physical scientists, have little self-motivation. They get good grades and are capable of doing great work, but they prefer to do nothing. They serially procrastinate and spend most of their time not working (chatting, messaging, shopping, Reddit, Facebook). If you give them a well-defined task and a deadline, they can get it done, but anything open-ended is a nightmare, they just freeze. This may be true for all of history, I have no idea. But even if a 4-year engineering degree is a good signal for capability, it's a crappy signal for "worker quality".

"If you give them a well-defined task and a deadline, they can get it done, but anything open-ended is a nightmare, they just freeze."

That's precisely the job of project leader: define tasks, agree deadlines and enforce them.

What have you expected from being a boss? Lightning cigars with $100 bills?If they ever become "good workers", you're no longer needed.

If they had great motivation, they wouldn't go to college, they'd acquire the skills by looking up a curriculum and reading the relevant books and manuals at home. They also wouldn't work for you, instead they'd found a startup.

It turns out the optimum age for starting up is around 40, building on both formal and informal education accumulated by that age.

I have noticed that there are several people around who, although quite able, reject the interview/selection process and "slum it" by starting gardening rounds or cleaning rounds, or indeed just a "handyman" business.
Link this with recent work that suggests that someone working a double shift at an unskilled job for the same length of time as the study required for a profession and investing the second wage, can end up having earned as much as a professional throughout his working life.
And mix in how Internet searches can make a specialist in almost anything.
Now where does this leave formal education and employee selection in say ten or twenty years?

'that education is 100% signaling'

Clearly, anybody who can read this is just signalling, since education played zero role in learning that skill, right?

Don't be obtuse.

Reading is clearly about signalling - it even appears you agree, as clearly your learning to read and write has resulted in telling a stranger not to be obtuse, which would seem to be a fine example of something that is obviously not based on education - Straussian virtue signalling perhaps? Mood affiliation? Or whatever terms you prefer, of course.

Education =/= Learning
In fact, this topic is usually raised to accuse formal education of a lack in learning. You don't need to go to school to learn to read or write. There are many programmers who have not graduated university, and there are many comp sci grads who do not know how to program.

Maybe you just went to some really awful schools.

Do your Venn diagrams show disjoint sets for learning and education?

Did your knowledge of Venn diagrams flow to you through the phlogiston?

Venn diagrams are intuitive and I hope your amazing school did not waste your time by teaching them. Anyway, I was explaining to clock what the debate is even about, since he seems to have missed it. You seem to already understand the naysayers' thesis so maybe you can move to one of the subthreads where people are arguing for one side or the other?

I think there is something weird going on here, where reasonably bright people amuse themselves by saying obviously false things like "Education =/= Learning" or "Let’s say, for purposes of argument, that education is 100% signaling"

Just don't.

If you want to be smart say that there are situations where signalling ability is more important than having acquired useful skills.

And obviously there are a lot of places in this world where skills are expected, on day one.

I'm not sure your logic chain holds up, and you appear to be positing the question as "if education is mostly signalling, then this logic chain is true, and therefore this ludicrous thing must be true, and I don't think that ludicrous thing is true therefore the premise must be false."

It seems to me the logic chain is flawed, therefore falsifying the result doesn't falsify the premise (I have no particular view on education as signalling, just on your logic chain).

The specific point of weakness I suggest in your logic chain is your assertion that if income inequality is rising in the early years of jobs, this must be due to signalling, as an employer could not possibly know much about an employee in early years. Whereas I'd argue that after about 3 weeks in a job you know quite a bit about an employee (and can judge if they are going to be successful or not). Perhaps there have been changes in willingness to provide differential pay rises based on this early information, which in turn would lead to inequality.

Unless you're asserting that there are significant differences in _starting_ salaries, not just in early years. Even then, I'd suggest that in a winner takes all market, some people will get paid far more based on the quality of their signalling (presumably not all signalling is equal).

"Finishing successfully is more of a mark of quality than it used to be, because finishing is harder."

My impression is the opposite: that colleges are nicer these days about helping students get through to graduation.

One reason is that the USNWR rankings weight % graduating positively. The highest ranked colleges have very high graduation rates, and a few old-fashioned colleges like Caltech that are pretty callous toward undergrads get dinged for not having the almost 100% graduation rates that their test scores normally imply.

For example, consider the Rolling Stone article "A Rape on Campus" by Sabrina Rubin Erdely. The U. of Virginia provided assistant dean Nicole Eramo to listen patiently for many hours to coed Jackie Coakley's complete BS stories about why she deserved extensions and handholding to keep her from being flunked out of UVA. Dean Eramo didn't believe Coakley's lies about Haven Monahan and being gang raped on broken glass, but part of Dean Eramo's job was to sit there and listen to coeds make up excuses.

America has a lot more Dean Eramos these days to coddle marginal students like Coakley through college than in, say, the 1970s.

"My impression is the opposite: that colleges are nicer these days about helping students get through to graduation."
Americans are confusing. Some Mericans (Krugman for instance) say students get into university less prepared than it used to be the case. Mr. Cochran says Americans Whites are where they always have been where K-12 academic achievement is concerned. I, myself, doubt Pasteur or Darwin would qualify for Harvard or Yale in today's America. Where is the truth?

Also, the rise of Advanced Placement testing has made it easier to graduate on time or even early. It used to be around 1980 that it was hard to graduate from the U. of California colleges in 4 years because classes were overbooked. But UC made an effort to make that less of a problem, and the AP credit helps more people get through with room to spare.

My son, for example, saved me a huge amount of money by graduating from his expensive private college in 3 years due to passing a vast number of AP tests.

Wealth inequality plummeted following the 1929 financial crisis, but income inequality remained high. Was the persistence of the great depression attributable to the former or the latter, or a combination of the two in tandem?

Could it just mean that more people are enrolling, more skilled people are graduating with stickers and more unskilled are dropping out with debt? Rigor seems an unnecessary variable.

Unless that passes for a twisted definition for rigor.

About 65% of those who enroll at post-secondary institutions are eventually awarded a BA degrees. Some of the others stop with an associate's degree. North of 40% of each cohort now receives a BA degree, which is about double the share so receiving a generation ago.

40% of each age cohort would mean everyone down to about 105 IQ gets a BA level degree. I've seen data suggesting that average IQ of college graduates had declined about 10 points over the last 40 years or so, from about 112 to about 100, which would appear consistent.

I would infer that we are sending too many people to college (more than are capable of true college level work). The fact that they are getting degrees strongly suggests that college, in aggregate, is less rigorous that it was.

Let’s say, for purposes of argument, that education is 100% signaling,

Vocational degrees account for about 65% of all baccalaureate degrees issued and some of the academic degrees (e.g. those in chemistry, microbiology, computer science, and even economics) map satisfactorily to private sector jobs. Why would you construct a model making that assumption?

"Now consider a situation where income inequality is rising, at least in the early years of jobs. Since employers cannot discern worker quality — other than by observing the signal that is — this should imply that getting an education is “more separating” than it used to be."

Or could simply mean that the personal qualities signalized by high education are more valuable in the market than before.

"Now consider a situation where income inequality is rising, at least in the early years of jobs. Since employers cannot discern worker quality — other than by observing the signal that is — this should imply that getting an education is “more separating” than it used to be."

Only if you assume that the cause of rising income inequality is rising inequality in worker quality. I don't think that's the case.

Well it certainly costs more money nowadays.

Even if the rigor and the probability of finishing school stays the same, the risk increases because people have to bet more money now. So people with low IQ and conscientiousness are less likely to apply, and that makes the signal even more valuable and more "separating".

Please clarify the antecedent for "this." More substantially: I suspect getting through might be in some ways harder -- the content seems sometimes more challenging and I think the burden of taking on loans is more difficult to carry than before -- but I also think there are more programs to help students get through. One more thing, your model does not take into account computers and the internet. I can type far faster than I can write, for example, and I can find a lot of information by searching that before I would have had to dig up through several layers of searching at the library.

Plug for a colleague's recent paper on how students use major choice (e.g., STEM or not) as a signal, and how the local geography of education influences this. Takeaway: Major choice is a signal. This might be relevant because STEM majors are plausibly more challenging to complete?


My personal experience was at a small very selective university. Many, perhaps most, students came in with the declared intent to be STEM majors; they all took the core freshman physics, chemistry, and calculus.

Many of these students switched to non-STEM majors and freshmen or sophomores. Almost no one switched from non-STEM to STEM. The non-STEM majors were universally understood to be easier.

All the anecdotal information I've heard at other institutions is consistent with this; with rare exceptions, all the switching is from STEM to non-STEM.

The first year of majors like engineering are designed to weed students out. Recently this has come under scrutiny because black and female students tend to disappear into the humanities. Inside Higher Ed has been covering various "reforms" which scare the bejesus out of me as someone who crosses bridges and the like.

In general, anytime the problems and irrationalities of education are looked at, race lurks in the background.

It is true. You need a good SAT and HS GPA for a good undergrad. And to get a good grade degree (for an eventual good job) you need a good undergrad GPA and then a good score on another standardized test (LSAT GMAT MCAT GRE).

That's significantly harder than just getting an undergraduate degree.

Could be! One question I'd ask about the situation is this: What incentives are acting on the schools? If students complain vociferously enough (and pay enough, in the case of colleges), could it be that schools actually get easier, in order to extract more tuition from more students?

Now consider a situation where income inequality is rising,

Couldn't it rather imply that demand for IQ, conscientiousness, etc. is rising relatively. IE a person with computer skills can now produce more relative to a work without.

Another thing could be that separation is getting clearer now a higher percent of the people who could graduate are graduating.

I know I am just using two anecdotes. Here goes.

Bill Gates and Mark Zuckerberg both got into Harvard, and both dropped out to become billionaires.

In those cases, the value of Harvard was not a diploma. It was first the signalling of getting in. And second the opportunity to network. Graduating is not much of a signal. Just how many admissions get flunked out?

I'm far from sure that either Gates or Zuckerberg relied on either of those two things.

Not true. What rigor? This is the post-rigor age. There wasn't even rigor when I went, because everybody needs to participate in the education bubble. Besides, these places weed out people above 130 IQ, because the signal (for corporations) is primarily about willingness to conform. So you get a sort of goldilocks zone of IQ where people are smart enough to figure out what is wanted, but not smart enough to have much of a problem with what is wanted.

By income inequality, aren't we usually talking about the top 1% vs the median. But this can't be what you have in mind because people in the top 1% are being getting a job based on their education.

Maybe you mean the median versus the bottom 10%? But has inequality grown between this group? I thought by inequality we meant the top is doing better than everyone, and everyone else is doing about the same.

So I don't think you question makes sense.

I'm stuck on the getting in vs. getting through distinction. There are probably at least two or three orders of magnitude more students that have the potential to not fail out of Harvard and Stanford than are allowed admission. It's the getting in that's the hard part. And that getting in part seems to get harder every year. Sure, private firms could invest in their own gatekeeping process, but why bother competing with academia when students, faculty, and parents are already oversubscribed in their willingness to participate in these existing institutions?

As part of my misspent youth, I taught at a college. Being an adjunct professor meant you often taught introductory classes to some very disinterested students who weren't majoring in the topic. (Or on occasion some were openly hostile to the topic because of religious, ethical, ethnic, the gender of the teacher, meeting time, or - well it is a potentially very long list of objections.)

Invariably you would get questions like; why do I have to take this course, why does this course matter, why, why, why,.... My response was often something like the following.

The administration has decided that for you to be considered educated, you must have a foundation in some basic skills. Those foundational skills once acquired, improve your ability to select a path for your life that will, hopefully, be rewarding for you and society. And, as my Sainted Irish mother was fond of saying; an education is an easy burden to carry through life.

A college degree does not mean that you are smart. Instead, it shows that you can start and finish tasks on time, even when you didn't care for the project. For some of you, I apologize in advance; this class may be such an event. Furthur, I regret to inform you, such burdens will be a part of your college experience and life.

How will this course have any value in your future life? Well, what are the skills that potential prospective employers are looking for? They would like you to have the basic skills (a high level of competency ideally) to do the work. But they also desire certain character traits. Do you have a strong work ethic? Can you complete tasks on time? Can you do what you're assigned to do, and finish it, even if you may not feel terribly motivated or perhaps hungover? Can you work with others? Can you take direction from people you don't like? That, of course, will not be a problem in this class, but it sometimes happens in other venues.

A few of you, perhaps one or two, will be so inspired by this course that the trajectory of your life will change: a professor can dream. Most of you will complete the course, check the box, and move on to things you find more interesting. In the meantime, I will help you acquire some foundational skills that will help you in other classes. That is assuming that the experience doesn't encourage you to do something more productive with your life and you drop out of college.

This is what college is. Acquiring knowledge and skills that will give you a shared base of knowledge, a shared patios, the proper buzzwords, to communicate ideas with others who have shared the path. A college degree will demonstrate your ability to overcome all the silly, wasteful, expensive hurdles that the college administration and faculty have placed in your path because; well I could say it was because it teaches how to learn, how to dig inside yourself and find your true gifts, but the truth is they are just a collection of cruel sadist who enjoy torturing gullible people who have the funds to support their intellectual perversions. They are a kinky group always looking for some mystical ivory tower. They are in desperate need of funding for their quest, so they thank you for your contributions and offer me up as a token of their appreciation. Make of it what you will.

So if we can begin, and to answer the burning question that you all have, you must know the following for the final exam.

An interesting thought, and increased rigor is the explanation, but not because the courses are more difficult, but because the students are sorting themselves by major differently to signal their quality more efficiently. So, fewer humanities majors and more STEM majors, ceteris paribus. Clearly some majors are more rigorous (at least in terms of the raw IQ it takes to finish) than others. 30 years ago there were lots of bright students in the English and History departments, today I suspect there are fewer. And if they can't do Physics or Engineering, then they can at least take Biology and Economics. The pressure to avoid being the barrista is real.

Tyler's conjecture is: school is now harder so the equilibrium is more separating (high ability workers get educated with a higher probability than low ability workers) and as a result, there's more inequality.

I'm surprised no one mentioned the fact that Spence's original paper had different costs of education for high and low ability workers. This was also to intend that for the lower ability type education would require a higher effort.

I believe Tyler's conjecture would be correct if the skills to send the signal are correlated with actual productivity. This is subtly different from Spence's requirement; sending the signal is correlated with being a high ability worker.

The subtle difference is more evident in the case of graduate school, where is not clear that the ability to get through is correlated with being a high ability researcher. If those two things were correlated, at least in the early stages making grad school harder would result in (at least) better young PhDs.

However, in the long term and in other settings things could be different. The fact that few econ PhDs produce enough good research to get tenure reflects that making grad school harder could led to nowhere.

Thus I believe the link to inequality is tenuous and contingent on many things.

I believe Tyler's conjecture would be correct if the skills to send the signal are correlated with actual productivity. This is subtly different from Spence's requirement; sending the signal is correlated with being a high ability worker.

If we assume Tyler is really smart, then this is the crux, and more than subtle. It actually destroys the starting assumption. It is the purpose of the essay.

(For those fields where it is true. This whole idea of normalizing "education" to a uniform quantity is suspect as well.)

This whole thing leads to asking: rising tuitions, academic hurdle-jumping, and inequality are efficient and therefore we shouldn't complain?

There would seem to be that risk.

Though if we are glib at one end about high value academic signalling and high value technical skills, we might be too pessimistic about non-academic and non-technical roles.

What would be an example of a role for which academic background and technical skills don't matter, and the applicant need only "present" with the right spirit? Possibly sales, where you present and sell yourself?

Are there too few of those, and is Linvega right to worry about changes in work requirements?

To summarize Tyler's hypothetical: income inequality between the college educated and the non-college educated is large and/or rising, and the only criteria employers can use for selection is the presence or absence of a college degree. It's quite a leap to go from that to a college education must therefore be more rigorous in my mind.

Let's say that in years past, X percentage of people capable of getting a college degree actually got one. And in those years past there was less incentive to get a college education because of low income inequality. Now with today's incentive structure, a higher number of degree capable people than X apply, along with a number of non degree capable people. Assuming wise admission policy and a consistent level of rigor, the graduation rate should rise because a higher quality of applicant is getting in.

Is the graduation rate today higher than in the past? That would go against the increased rigor idea. Is the rate the same? That might support the idea, or other factors may be at play. If the graduation rate is going down while the quality of applicant is going up, that would make a strong case for increased rigor.

In my mind there are too many factors at play to draw a link between income separation and college academic rigor.

It’s a loop.

I used to teach the LSAT. One of the things I told me students was that there are two tests, not one.
1) the LSAT itself tests intelligence.
2) the lsat PROCESS tests your ability to understand the relative importance of the LSAT in the admissions process; your ability to understand the admissions in your future career, and your ability to make good choices. Although test 2 was originally less important, it has grown exponentially in importance

In Tyler’s example the starting assumption is signaling. There are still some smart folks without the signal but not as many.

Fast forward a few years and the gap grows. Not only do no college folks get dinged on the signal but they get dinged in their failure to recognize the importance of the signal. And so on.

Can't make any sense out of this. How does changing "early income inequality" correlate to educational "rigor"? I started to count the number of other factors involved - including changing government laws/regulations, a changing economy and a changing work force which simply (and obviously) can't be ignored in any reasonable model of "income equity" and gave up. His "model" is risible. Here's an alternative: rising costs of government regulated worker (health) benefits has forced employers to arbitrarily pick winners and losers.

Yes it is more rigorous and harder...once you consider the educational signal as a whole rather than each component of the education. When a master's give the signal of a bachelors from the 50s, you now have to cover the course work of both a 4 year degree and the extra 2 years for the masters plus a thesis for many programs. Even if a bachelor's degree is easier than it was in the 50s, a bachelor's + masters seems clearly more difficult.

I think this question incorrectly considers part of the spectrum of educational signaling rather than the signal as a whole. High school and college can be easier if they are a lower signal than they once were and the total signal range has expanded.

What about party schools? I don't think they provide a valuable signal and I don't think they are rigorous.

Perhaps finishing is harder because there are more distractions. Easier availability of psychoactive chemicals (including the classic alcohol). TV, video games, etc. Cars, for gosh sake. So much else that is interesting and easier than learning what the courses are supposed to be teaching you. Those who make it through show they are more goal-directed and more able to resist distractions.

Am I the only one that can't make out what the hell Tyler is talking about here?

High school is much, much more difficult for students at every point on the intellectual scale.

College, on the other hand, is significantly easier because we're shoving everyone through. 30-40% of the kids we send to college can't read or write at an 8th grade level, and a non-trivial chunk of them graduate. We've increased the number of college graduates, yet all tests of adult and college graduate literacy have stayed constant.

The largest public university system in the country is dumping the remedial math requirement, meaning they will give credit for taking middle school math classes.

This isn't debatable.

Top students today are as good or better than the past. However, many schools are driven to fill seats and lower standards. Those are the students who don't graduate and are often stuck with loans they cannot repay.

Contrary to Tyler, I don't think finishing college is harder overall. But for some students who are admitted with weak skills, finishing is a real burden.

On one level, giving students the opportunity to prove themselves regardless of background is a plus for the few who can complete the journey. But for too many, you have given them nothing but an additional financial burden.

First, maybe going to college and graduating is easier so that not graduating from college becomes even more of a negative mark than before.

More importantly, the connection between inequality and signaling is dubious. Think of college as like a credit rating agency. Colleges rate students for hiring worthiness. Credit rating agencies rate companies for credit worthiness. During some periods, credit spreads widen so that inequality in borrowing rates between low-risk and high-risk borrowers increases. Is that because credit rating agencies have become more scrutinizing? Not usually. It just means that economic conditions have changed such that lenders have become less willing to lend to high-risk borrowers. Widening income inequality or increases in the college education premium may just indicate changes in employers' demand for different types of workers.

absolutely not. The signalling is about getting in, not getting through. (This is a pet peeve of mine--i'd rather find someone who perhaps was a late bloomer and finished strong at a public or lesser school than someone who got into an elite school.)

To give an answer you need to take into account the other side of the market i.e. the schools. What is their objective function and the market structure?
Take MBAs for instance, these programs are 100% signaling. The best way to see this is the fact that firms start recruiting students even before students have had their first class. In this model, business schools are screening machines. They serve as signaling devices that find the best businessmen/women from the pool of individuals for whom it makes sense (can) to pay the screening cost.
Also, in these programs, it is much harder to get in than to get through. I believe the rate of non-completion is very low, and schools colluded so that no school will reveal GPAs of students to prospective employers.

So, in this case, finishing is not much harder despite the fact MBAs' are mainly a signal.

Comments for this post are closed