"Hey! Teacher! Leave Them Kids Alone!"

This is the 23rd installment of 774: Weekly lessons from history about science, technology, and the miscellaneous.

Next Week:

Who Needs Stupid Books?

This Week:

“Hey! Teacher! Leave Them Kids Alone!”

Charles Goodhart, a British economist, wrote of Monetary Policy:

Any observed statistical regularity will tend to collapse once pressure is placed upon it for control purposes.

22 years later, an anthropologist from Cambridge made his statement more digestible:

When a measure becomes a target, it ceases to be a good measure.1

To think about “Goodhart’s Law,” consider a call center running tech support for Apple. In an effort to gauge how well the call center is performing, a manager might look at how many calls each customer service representative takes in a day. Over time, it becomes known that the manager is judging performance by measuring the number of calls completed in a day. How do the employees respond? They expedite calls and sacrifice helpfulness for speed in order to appear more productive, but in reality are performing worse. The measure, number of calls taken in a day, became a target, and when it became a target, it is no longer useful as an indicator of performance.

Measuring real performance is particularly difficult because everyone wants to signal that they are on track to perform well. Human beings want to game the system, constantly. In 1920, the U.S. Army began to test its soldiers on a variety of physical assessments. The many tests have changed over the years between exercises like pushups, burpees, man-carries, and squats, but one has remained a constant: running. Proportionate to how highly it’s weighted in evaluations, individual speed does not have a massive effect on most soldiers’ ability to do their jobs well. Nonetheless, if you want to rank very highly in army schools, you should be running 6:30 miles, minimum. This results in soldiers focusing excess time on a target that doesn’t actually improve job performance, but will improve their perceived performance on paper. There are certainly worse outcomes than soldiers spending extra time doing cardio, but some of the lengths that people will go to in order to meet a measure-turned-target can get costly.

Economist Bryan Caplan argues in his book, The Case Against Education, that higher education is a society-wide signaling business that provides an undersized benefit to individuals outside of income growth.2 Income growth is certainly nothing to scoff at, but it’s not the sole purpose of most universities. Universities, ostensibly, exist to teach students knowledge and skills, particularly those that prepare them well for the workforce. Caplan argues that very little of the college experience delivers any long term knowledge gains or employable skills. What they provide students is a signal.

Consider the college graduate who gets a degree in Finance, but ultimately ends up working in logistics. It’s likely that when they applied for the role, which primarily deals in scheduling and customer relations, the ability to calculate the time value of money was not on the list of job requirements. Much of what that Finance major spent four years studying is not utilized in his job. In spite of the inapplicability of this graduate’s degree to his work, he might not have gotten his good paying logistics job without it. If obtaining any degree is a requirement to get a particular job, but doesn’t necessitate a particular skill set acquired through college, what exactly is that degree signaling?


Intelligence tests, thanks to U.S. labor laws, cannot exist in a vacuum. It is illegal for U.S. employers to discriminate between potential employees on the basis of IQ.3 Employers incur hefty fines if they are found to conduct IQ testing that cannot be proven to be an absolute business necessity. While Bank of America cannot test your IQ if you want to work in their loan department, your local business school can absolutely make you participate in one of the largest tests of intelligence conducted in the nation: the SAT. So you go to business school, graduate, and apply to work at Bank of America, who can ensure you are up to snuff because anyone who got into your top tier university must have a high IQ. But Caplan points out that this explanation of IQ signaling isn’t enough to explain the premium set for higher education. If all an employer was searching for was high IQ, all you really need to do is staple a copy of your acceptance letter to your resumé. No one does this, so there must be other information that a degree signals.


In the section titled “What Does Education Signal?” Caplan opens with a quote

From the standpoint of most teachers, right up to the level of teachers of college undergraduates, the ideal student is well behaved, unaggressive, docile, patient, meticulous, and empathetic in the sense of intuiting the response to the teacher that is most likely to please the teacher.

-Richard Posner

Caplan argues that conformity is another item in the “package” of signals that a high school and college diploma send to employers. What sort of conformity though?Caplan says that

To be clear, employers aren’t looking for workers who conform in some abstract sense. As anthropologists point out, almost everyone conforms to something. Hippies strive to look, talk, and act like fellow hippies. This doesn’t make unkempt hair and tie-dye shirts any less repugnant to employers. Employers are looking for people who conform to folkways of today’s workplace- people who look, talk, and act like modern model workers.

One of the biggest conformist things people do today is become educated. If you’re a moderately well performing high school student, the default expectation is that you graduate and attend college. College is a $100,000 way of saying “I’m normal,” regardless of the truth in that statement.

What are modern model workers like? They’re team players. They’re deferential to superiors, but not slavish. They’re congenial toward coworkers but put business first. They dress and groom conservatively. They say nothing remotely racist or sexist, and they stay a mile away from anything construable as sexual harassment. Perhaps most importantly, they know and do what’s expected, even when articulating social norms is difficult and embarrassing. Employers don’t have to tell a modern model worker what’s socially acceptable case by case.

Conformity tells your employer you’ll do what your boss asks, create a hospitable work environment, and won’t get the company sued. But conformity and intelligence aren’t enough to signal how good of a candidate you are. Most people know someone with a high level of intelligence who adheres to all obvious social norms, but would still rather skip class to play Call of Duty. They can be conformist and smart but remain lazy.


There’s a differential between the number of people who begin high school and college and those who leave with a diploma. Those who leave without a degree are not treated the same as those who receive them. People leave college early for all sorts of reasons, but Caplan states that finishing a degree signals conscientiousness to a company that’s hiring. College, for those who finish on time, is a four year endeavor. There’s little you can do to speed it up. Even an accelerated graduation track of 3 years is quite the time commitment. Completing a multi-year project such as a college degree signals to others your willingness to to stick with something difficult over a long period of time.

One could point out that college is by no means the only signal of conscientiousness. There are ample ways to show that you can put in hard work, day after day, for years. Successfully holding a job for a long time or undertaking a years long personal project both exhibit conscientiousness. Caplan’s point is that a degree (both high school and college) is a way to simultaneously signal intelligence, that you’re willing to conform to social norms, and that you’re conscientious.

The issue with this bundled signaling is that it used to be a measure. A measure of intelligence and conscientiousness (I doubt many college administrators would readily admit that they’re testing student’s conformity) has become a target. In becoming a target it has become wasteful. Very few student’s attend college to increase their knowledge and skill set, they do so to receive the higher pay checks waiting for them on the other side. Depending on one’s course load, professors teach material that interests them personally, students study it, are graded accordingly, and quickly wash it from their minds. Caplan quotes an old Soviet adage :

We pretend to teach, they pretend to learn.

Caplan walks through equations he built to determine whether this payoff is worth it. If there is $100,000 price tag on a B.S. from Penn State, it would do one well to predict the actual financial payoff that will likely result from the degree. Caplan argues that more students are accepting high prices that simply won’t pay down the road. If we are going to pour so much time, money, and human effort into education we could at least bear to do what colleges purport to do: educate and teach job skills that will actually make students better workers. The United States collectively spends $1.1 trillion on education annually. If even 1% of spending on education is wasteful, that’s $11 billion annually being squandered (Caplan’s lowball for percentage of education spending going to waste is 20%. His highball estimate is 80%). Caplan is in favor of slashing that spending that he calls “fat” and reallocating existing spending to actual job skills that translate to working. Less Shakespeare, more statistics.

Maybe you disagree with the claim that 20-80% of education spending is going to waste and you very well may disagree with taking away any funding at all from schools. The claim that centers of education have some inefficient practices is tough dispute. Not every student is going to be exuberant about his or her courses and you can’t remove a class simply because someone is bored. What can be done is to change some of the simple guard rails on the way education is practiced at all levels.


Wikipedia is disavowed by teachers as a source of information almost as soon as students are asked to do research. This is unproductive for two reasons. There have certainly been incorrect edits on countless Wikipedia pages, but unlike pretty much any other encyclopedia, Wikipedia is constantly and obsessively updated by dedicated users to be as accurate as possible. Consistent improvement and increase in knowledge on every Wikipedia page is not going to stop anytime soon. The same cannot be said of more traditional encyclopedias. The second reason to encourage Wikipedia use is that it rebuts the idea that there isn’t good information to be learned from strangers on the internet. That’s the foundation of this blog. Strangers online have useful things to say. If you want to use Python to build a neat little map of the concentration of AIRBNBs in New York City, you shouldn’t enroll in a costly program to teach you how to do data visualization. Go on github.com or stackoverflow.com and just look at how other people made something similar. Keep the pieces you like and ditch the rest. It’s much cheaper and easier than going through a certified course.

Open Book Tests

Open book tests should be utilized all of the time. School is one of the few places in life where you are encouraged to trust your poor, poor human memory with the truth. It’s incredibly difficult to remember even the most important information. The idea that you should try and rely on fleeting memory when you can quickly search for the correct answers is counterproductive. A test can at least imbue students with the importance of utilizing resources to answer difficult questions.


I would like to imagine that if Isaac Newton entered an 11th grade trigonometry course he would be horrified that students have access to world class, hand-held computers, but are rarely allowed to use them. Imagine what a young Albert Einstein could’ve accomplished with a TI-84. It’s insane to limit learning to the human mind’s pace when we have technology deliberately built to augment it.

These three steps won’t solve the problems Caplan argues plague the education system, but they would make better use of the time spent on education right away.