Bryan Caplan  

Pop Quiz

Baby Fever... California's Proposition 19...
Suppose you get your midterm back, and discover that you've scored considerably higher than you expected.  According to basic micro, how will you adjust your study effort in response to this pleasant surprise?

Comments and Sharing

COMMENTS (25 to date)
Zac Gochenour writes:

Before you took the exam, you were studying until MC=MB. You know your MC of study time but you just have some expectation of the MB. If you did much better than you expected, that means the MB is lower than you thought. So you should reduce your study time in response.

Jon Leonard writes:

The basic micro answer would be that I'll study less, since there's an opportunity cost to the studying, and I need less studying than my previous estimate. The new equilibrium will have a higher grade and less studying than the old equilibrium.

But this makes a lot of assumptions about the underlying circumstances. I could have been lucky on that test, and depending on my previous tests, I'll update my Bayesian estimate of studying vs. results appropriately. Or, it could have been a grading error, which would be handled much differently.

I might also take it as a sign that I have some sort of competitive advantage in the relevant subject, in which case I might study much harder: It might make a good career choice. (At least, a better choice than the subjects where I routinely do worse than expected on tests.)

david writes:

Depends on your priors, and how your midterm actually factors into your overall degree, and what you are actually maximizing here.

Plus: income effects, etc. Maybe your higher score makes it plausible that you will get a first instead of a second class, so you will study harder for other elements of the degree.

I think all of that qualifies as "basic micro"...

David Friedman writes:

As stated, the question isn't answerable--it depends both on the production function for grades and the utility function for grades.

To make that clear, imagine a student who will get a large benefit from receiving an A rather than a B--perhaps a parent has promised some reward for an A, or perhaps the students is applying to a graduate school whose standards are high enough that a B will almost certainly mean rejection. There might be almost no benefit to a B over a C. Doing well on the midterm signals the possibility of an A, so it's worth working to get it.

Describing a situation that goes the other way is left as an exercise for the reader.

Henry Milner writes:

At this higher grade level, the marginal benefit of a higher grade has decreased, assuming diminishing marginal returns (perhaps a dubious assumption, as others have pointed out). But perhaps the unexpectedly high grade means that my studying is more productive than I had thought, so my estimate of the marginal grade improvement due to studying has increased. It's not clear which effect will be greater. Also, I enjoy doing things I am good at (or at least better at than others), so the marginal cost of studying may decrease. The effect on my studying is ambiguous.

4 months after finishing my undergrad degree in econ (and now doing something mostly unrelated), this stuff seems even more useless than it did when I was learning it! In most cases, when intuition clashes with "basic micro," it's because basic micro has abstracted away the important details to make the model tractable, not because simple optimization theory has any unique insight into human behavior. I realize these kinds of exercises are just pedagogical tools, but I suspect that many economists actually believe in them on some level. Also, most people in introductory micro courses never take more advanced economics; if they're paying attention, they must either be lead dreadfully astray, or conclude that economists are a bunch of idiots.

kevin writes:

You would study more. You discovered that the price of grades is lower than expected, so naturally you buy more grades with your time-currency.

Daniel Kuehn writes:

It depends on whether an income or a substitution effect dominates. Kevin's story could be right - but realizing that the price of grades is lower than you thought, you might use your time elsewhere. It's ambiguous.

JPIrving writes:

It at least depends on if the the first driviatives of the utility function are linear or if it has some multinomial form (two or more peaks around grades or study levels). You could either continue to study hard and get high marks, or reduce studying.

Sam writes:

This comment thread is a lovely example of "ask two economists, get three answers".

Jeff writes:
This comment thread is a lovely example of "ask two economists, get three answers".
No, it shows that poorly posed questions don't have good answers.
Tracy W writes:

Henry Milner - the useful thing I find about these results from micro is that from then on you know that there's no clear theoretical result, so you know to be suspicious when someone makes an argument depending on the result being one way, without providing a good argument as to why the results would be that way.

So for example, I recall someone asserting that the rate of interest would be lower in a no-economic-growth world, because the demand for capital would be lower. I queried that, based on Econ 101 the result would be indeterminate. The person I was questioning didn't even understand the question, indicating that they hadn't worked out their economic theory of a no-economic-growth world.

My own answer to Bryan's question is that it's indeterminate. Eg in my final year of high school I was trying to get results on my bursary exams of over 400 across my top 5 subjects (I was taking 6), so as to get into the uni course I wanted without having to do an intermediate year. If I had gotten an unexpectedly good result in one of my maths papers I would have studied more in it, if I had gotten an unexpectedly good, or bad, result in history I would have put that down to the vagrancies of the marker and ignored the information.

Luís Fonseca writes:

It depends. It would function like a raise in wage: you could work less because of the income effect, or you could work more because it becomes costlier not to work. So, there, if the same time studying can give you a higher grade, you can study less if you're happy with that sort of grade, or you can study more because you are more efficient at it.

Steve writes:

You darn economists. Is the answer to anything ever something besides "it depends?"

Luís Fonseca writes:

Steve: When a question is this vague, no.

Sam writes:

@Jeff: Exactly: the "problem" with economists is that the questions we ask them are usually vague or ill-formed.

John Thacker writes:
If you did much better than you expected, that means the MB is lower than you thought.

Not necessarily. If you've taken a test without studying before, and then studying this time gave you a larger jump in your grade than you expected, then the MB is higher than you thought.

If you're a supply-side economist, I suspect there is one obvious "correct" answer from "basic micro". It's the same one that some people think always and everywhere result from lower tax rates and some naive assumptions about MB and MC.

If you have a slightly more sophisticated view of the world, you realize that the outcome is ambiguous. In fact, I had a Dean many years advise me to grade the first quiz of the semester harshly to induce more effort from the students subsequently.

This real world behavior might come as a surprise to some, but as the comments above illustrate, it's not terribly hard to model using "basic micro".

Extra credit: If people in a society become more productive, would you expect them to work more hours or fewer hours?

Salem writes:

I disagree that the problem is vague or ill-formed. The question is a perfectly valid real-world scenario, with answers that can be determined empirically. The problem is with abstractions like "utility functions", not with reality, which is ticking along just fine.

I suspect that there will be a range of responses, but the median response will be to work harder because of the psychic gain. Has there been a study done on this?

Dan Weber writes:

Apparently I do well in the subject, which implies that further study will push me further along the lines of comparative advantage.

I expect the math prodigy to study math several hours a night. He doesn't need to in order to survive, but being the best mathematician and a poor geologist is worth more than being an average mathematician and an average geologist.

David Williamson writes:

Short answer: only if I am convinced that the grade is largely a function of skill gained from studying.

Long version:
A few questions must be addressed to reassess the benefits of study:
1a) Was the exam multiple choice, short answer, or longer essay? 1b) Can the student assume the same format for the final?
2) Was the student's raw score higher than expected, or did the professor include a very generous curve?

I'll assume 1b is answered with yes.

Chance of repetition with similar skill matters. If I knew my higher grade was a fluke (multi-choice), I would realize that for the final, p(A) increases with study. If I knew I was being too hard on myself with remembering what really wasn't terribly difficult material (short answer), because the vocabulary wasn't terribly difficult, I will study as much, or less if it makes it easier to think. If I knew that I could successfully BS my way through an argument to say things the prof liked without doing the reading (essay), then I would almost certainly study less, because study matters little compared to skills like writing well and sucking up to the prof.

If I knew that scoring a 58 on the exam would give me an A, I'm really not going to worry about studying, though there is a chance that getting the A on the midterm would change my tastes for going for an A on the final. How this manifests itself will be hard to judge; if all I need to study is a little more to keep my A, then absolutely study more. If it takes much studying to achieve a significant improvement, study the same. Finally if the test is arbitrary, study less.

In a way, you could suggest that the student will work as hard studying for the test as the teacher does making the test. Multiple choice tests are hard to write, and students will study more because they can improve more easily by studying. Short answer questions are a little easier, and students know they will either know it or they won't. An essay exam can be prepared 10 minutes before class, or right when the prof walks in by writing the question on the board, and students knowing there is little chance that any given topic will be on the exam and that it is most arbitrarily graded by a lazy teacher will study less.

Mandy writes:

I think this depends on which effect dominants: the income effect or the substitution effect.

An similar example is , say, I find a babysitting job. I am expected to be paid $12 per hour. At the end of the day the baby's mother pays me actually $15 hourly because she finds I am a wonderful babysitter. Suppose I think in the future if I provide services with similar quality (it will be different if the extra money will be gone if my performance falls), I will get $15 per hour. Will I work longer hours in the future?

Thus, the question is translated into the labor supply change according to the change of wage---we have learned that from the textbook.

saltmanSPIFF writes:

Instead of maximizing single test results in discrete time, my goal is to maximize my course grade. After a decade and a half of test-taking, I think I have rational expectations with respect to grades. I treat the pleasantly high score as an outlier. Ricardian equivalence holds; I don't change my study habits.

Bob Murphy writes:

For some reason, I have a hunch that Bryan was thinking like this:

"I bet you guys *thought* the obvious answer was you should study less next time, but actually that violates basic micro. What you've found is that the price of getting one more percentage point on your exam is lower than you thought, so you should 'buy' more of it. If it becomes cheaper to raise your GPA, you should engage in more of that activity."

But like I said, I think this is wrong (assuming that is in fact where Bryan was coming from), for the reasons you guys (and gals) have spelled out.

Dr. Friedman, if you're still reading, this is exactly the problem I had with your analysis of housing prices. :)

azmyth writes:

Unlike most goods, there is a clear maximum satiation point for grades. Students should aim for the lowest possible percentage that will still give them an A, usually 90%. If a student scores considerably higher, they can reduce their effort and still get the A.

If a student thinks they are close to failing, and get a C, they might increase studying in an effort to get an A or B instead of dropping out entirely.

Kevin S writes:

The answer to this question depends on whether the "expected" grade is in line with the desired grade. One can expect a C and receive a B while desiring an A, in which case he may study more if getting an A is worth the opportunity cost of the extra time it takes. On the other hand, one can expect a B and receive an A while only desiring a C, in which case he should study less, seeing that if he only desires a C, then the time it takes to achieve any higher than this could be better spend doing other things.

Comments for this entry have been closed
Return to top