Every PhD I know has stories from graduate school of THAT professor. The one who loses it, goes off the deep end and finally makes some decisions that are odd, weird, or just plain screwy. Stories of THAT professor are usually told over beers at a professional meeting and start with "Remember the time when..."

Well, yesterday I pulled a stunt that may put me into THAT category. I'm currently teaching a PhD Applied Econometrics class to 32 first and second year PhD students--yes, 32! (Due to a scheduling change we had to combine the first and second year classes this year so I have way more students than I can handle.) This class is the last in a a sequence of applied econometrics classes and my job is to take the students from classroom students to researchers. These are smart people taking the highest level class we have to offer, and frankly they can pretty much learn anything on their own at this point.

Anyway, as part of a project I am working on, I've been struggling for the past week with an econometric problem. It's something I should be able to figure out, and given enough time, could probably figure out, but given all of my other duties, I haven't yet figured it out. This is frustrating to me.

I don't mean frustrating as in 'gosh-dangit,' but frustrating as in @#$%^&*!!!

Losing sleep frustrated.

Questioning my training and abilities frustrating.

So yesterday morning I cracked. Rather then spending the morning thinking about class, I obsessed over my problem. Then I realized: "I have 32 captured students who know all the econometrics I have forgotten since graduate school." So...

Yesterday's class went like this: I began by announcing I was calling an audible. Instead of talking about simulation based approaches to discrete choice models, I am giving a problem to solve. I then spent 1/2 hour explaining (ranting about) the problem, my attempted solutions and where I thought the problems were.

After my insane hair-pulling rant, I told them their solution is due in two weeks. I don't care if you work individually, in small groups or as a class as a whole. You pick. Work together. Be smarter than someone else. Show me what you know. Talk to each other. If you work in a group everyone in your group will get the same grade.

In fact, I don't care about your grade.

I just want answers.

Then I left the room.

I have no idea what happened after that.

**Another in our WWWTF (What in the World Will I ever use This For?) series:**

**Math Concept**: Integration

Having spent multiple hours over the past week helping official oldest daughter of Env-Econ (OODEE) work through calculus problem after calculus problem in preparation for the pending AP exam, I thought it might be helpful to see how she might eventually move past solving math problems to applying the concepts. Right now I'm going to focus on integration, and hopefully later I will come back and talk about differentiation (seems backwards, but integration is fresh in my mind right now). In what follows, I assume a working knowledge of basic integral calculus--otherwise why would you care what you use this for?

As a reminder, a definite integral is defined as:

where f'(x) is the derivative of the function f(x) with respect to the variable x.

Graphically, the integral represents the area under the function y=f'(x) over the closed domain [a,b].

An example might help. Consider the simple linear function f'(x)=6-2X. What is the definite integral of f'(x) over the closed domain [1,3]?

What does this look like graphically?

Notice that for this case of a linear function (a line), the area under the function makes a triangle. A simple way to check your integration is to just find the area of the triangle on the graph.

The area of the triangle is .5(Base x Height) = .5(2 x 4) = 4. This is the same result we got for the intergral. For more complicated functions (non-linear), the area under the function isn't a nice neat geometric shape.

I know what you're thinking: Nice pictures and it's great fun to think about math, but WWWTF? (**W**hat in the **W**orld **W**ill I ever use **T**his **F**or?). Keep reading for an application to environmental economics.

**Application: **Measuring total willingness to pay for changes in environmental quality

Economists use demand functions to represent the relationship between how much someone is willing to pay for something and the quantity of the good consumed. It is generally believed (and accepted) that there is an inverse relationship between what someone is willing to pay for something and the quantity demanded. This is known as the law of demand.

For example, suppose the demand for trips to a local beach can be represented by the demand function:

where Q is the total number of trips demanded for by travelers (measured in thousands of trips) at a trip cost (price) of P. If the average trip cost (including gas and time costs) is $3.00 per trip, then the total number of trips demanded will be 1,289.

In addition to knowing how many trips will be taken at a particular price, economists are sometimes interested in knowing the total value of those trips beyond the price paid for the trip. The total value of the trips could be found by adding up the willingness to pay for each individual trip and then subtracting the price. We call this 'consumer surplus.'

For example, from the demand function, the willingness to pay for the first trip (the first trip is Q=0.001 because Q is measured in thousands) is $18.19. Subtracting the $3.00 price of a trip, the consumer surplus for the first trip is $15.19. The willingness to pay for the second trip (Q=0.002) is $16.71 with a consumer surplus of $13.71.

While this method would work, it would be tedious to calculate calculate consumer surplues for 1,289 trips. Instead, we can find the total consumer surplus by taking the integral of the demand function between the current price ($3.00) and the maximum price that will cause the number of trips taken to go to zero, called the choke price. This is the equivalent of adding up all the willingnesses to pay for each trip.

In the current example, the choke price is infinity since the exponential function is asymptotic to Q=0 as the price goes to infinity.

Graphically, this looks like:

(Notice that this graph is drawn with Q on the horizontal axis and P on the vertical axis. This is a flaw of economists as we draw demand functions on the wrong axes. I apologize for those who preceded me who established this convention and just note that the range of integration is now on the vertical axis.)

Mathemaically, consumer surplus is measured as:

Multiplying by 1,000, the total consumer surplus for the $1,298 trips is $2,678.

**What does this have to do with valuation of environmental quality?**

Consumer surplus gives us a measure of the value that trip takers place on trips above and beyond the cost of the trips. Suppose that the government decides to invest in a clean up project at the local beach. Once the project is complete, we would expect the demand for trips to increase. If the demand increases, that means that the total amount consumers are willing to pay for the beach trips has now increased. The difference between the change in consumer surplus from before to after the clean up is a measure of the value of the clean up project itself.

To see how this works, suppose after the clean up project, the demand for trips increase to:

Using the integral technique, the consumer surplus after the clean up project is $7,279 (assuming the price of trips stays the same at $3.00). So consumer surplus has increased by $4,601 due to the clean up project. This is a value we can now compare to the cost of the clean up project to begin to assess whether the project passes a benefit cost test.

And just to complete the circle, here is one last graph that shows you what the valuation of the environmental quality improvement looks like:

I had my second exam scheduled today and next week is Spring Break. From the inbox:

All classes are canceled Friday, March 7. Campus is closed until noon for employees.

This makes an entire week that the university has canceled my MWF 10-10:50 Stats II class. Each time the cancellation was a good decision. With most students living off campus and public transportation not running, it is just too risky to have thousands of people on slick roads.

That said, a week of class is a long time. A significant chunk of material that I (optimistically) hoped to cover will not be covered this semester (nonparametric statistics, time-series analysis). Several years ago the TR morning classes missed two weeks of class. A few falls ago the university basically canceled finals week due to snow.

There are significant educational costs that have been ignored as part of our adverse weather policy and there is too much snowy weather in Boone to ignore these costs. The only things I can think of is to have (a) some extra class days built in to the academic calendar or (b) have all classes with web content to mitigate the cancellations. When will this happen? Maybe never. In terms of (a) the academic calendar is already too jam packed with two sessions of summer school and a two week break for 12 month contract upper administrator vacations. Our "reading day" before final exams has routinely been moved to a Saturday. In terms of (b) the majority of faculty still don't have the technical skills to get a lecture or quiz online. And given the low morale in the UNC system, faculty are in no mood to have the legislature, UNC general administration or Appstate upper administration pile on any sort of extra work.

The university makes sure to tell faculty that if we are out of town for a conference then we need to have some sort of educational opportunity as a substitute (guest lecture, online quiz). There is no such thing when the administration decides to cancel class. How seriously should I, as a part-time administrator (i.e., department chair), take the upper administration's mandate that I monitor all classes and make sure that faculty don't simply cancel class when they are out of town on university approved travel?

Which brings me back to the awkward logistics of my class. When do I have the exam? Most of the students had already studied by the time class was canceled. The Monday after spring break doesn't sound appealing (half the class will still be on their way back from their Colorado/Utah ski vacation). The Wednesday after makes it two weeks since the last lecture on the material on the exam. Complicating that is the CEO lecture that my students will attend on Friday during class time (I'll be at a conference and business students are strongly encouraged to attend the CEO lecture).

Update: The roads were fine by 11 am. The decision to cancel Friday afternoon classes was nothing more than a "hey, it is spring break and classes won't be doing anything anyway" decision. Sad that the administration has the same attitude as the college students.

I've been department chair for almost five years now and I haven't much enjoyed a 40% decrease to our department budget and trying to allocate a one-time 1.3% (or something like that) raise. News like this is disheartening (but I'll add a caveat at the bottom):

In a memo on Feb. 28, [Art Pope, the state budget director] took university leaders to task, saying they’re asking for far too much money at a time when the state has competing priorities such as Medicaid and raises for K-12 teachers and state employees. He said the university system had basically ignored his office’s instructions in December to come forward with budget expansion requests of no more than 2 percent. ...

This year, the UNC system received $2.5 billion in state money for operations and another $64 million for building repairs and construction.

Pope said the board has requested an increase of $288 million, or 11.3 percent over the current year’s state budget for UNC. Those figures do not include any raises for employees.

While the state’s economy is improving, an 11 percent increase is a fantasy, he said. Such a spending increase for UNC, Pope said, would require the governor and legislature “to make major reductions in other state agencies and programs, such as our courts, the ‘K-12’ public schools, and health care.” ...

From 2007-08 to 2012-13, appropriations per student have declined 7 percent while tuition receipts per student have jumped 47 percent, according to the university system’s budget proposal. Controlling for inflation, education spending per degree at UNC has declined by 18 percent, UNC said. ...Pope, too, seems to be casting his eye toward the university’s ability to pay its own bills.

He pointed out that the system had a cash balance of nearly $269 million by the end of the 2012 fiscal year and collected $228 million in overhead payments accompanying grants and contracts, mostly from the federal government.

“How much of the overhead receipts are being used for the repairs and renovations for the facilities used to generate the overhead receipts, as opposed to requesting $163 million in General Fund appropriations for repairs and renovations?” Pope asked in his memo.

Appstate is looking for something like $90 million, I think, for a new nursing building. That is a lot of money in the current budget environment.

After almost 25 years in the UNC system, I've gotten way tired of the mission creep. ECU wanted to move up the Carnegie ladder and when we did they immediately announced a goal of moving up the next rung. Universities want new PhD programs, engineering schools, dental schools and et cetera. I've always wondered: why don't we just try to do better the stuff we are currently doing?

From the Chronicle of Higher Education:

From a former PhD student of our department (who graduated last May):

...60 undergrads at Penn State have now read your recent article (assuming they listen to my instructions) on CV analysis and critiquing Hausman. I had several classes on non-market values and used it to motivate CV analysis and its use in valuation.

For those who have forgotten, click here for access to the best paper of 2013.

From time to time, we find online college syllabi among those sites referring us traffic, and some professors have told us that they use Retraction Watch in their classes. We’re pleased and humbled by that.

In a new paper published in the

Journal of College Science Teaching, three professors at Clayton State University in Morrow, Georgia, discuss why retractions are good case studies for teaching ethics and examining the scientific process in class. Stephen Burnett, Richard H. Singiser, and Caroline Clower write:In this article, we discuss our experience using articles that have been “abandoned” (where their results are no longer accepted due to new evidence) and/or retracted as methods for teaching students about scientific literature in general and specifically about scientific ethics. Being presented with a more accurate picture of primary literature can help students develop an improved understanding of how science is actually practiced and how scientific ideas change over time. By examining retracted articles in which ethical lapses have been uncovered, students are able to develop a more clear understanding of the types of ethical problems that can occur and improve their ability to recognize them.

Their conclusion:

From retracted articles it is easy to stimulate discussion on a variety of issues, even if the specific methods might be too advanced. We have found that the best way to help define ethics is to show examples of ethical failures. Students can clearly see what is wrong in most of the retracted papers, and when they cannot see why a paper has been retracted, it becomes an important teaching tool to help them understand ethical conduct.

I'll be considering this to help prep for the next time I teach senior seminar (while the journal says this article is free, I haven't overcome the challenge of obtaining the PDF *it took awhile but I was able to obtain the free PDF*).

These poll results came out before the State Department issued their favorable report:

A slight majority of Americans favor the controversial Keystone XL oil pipeline that President Obama is expected to approve or reject this year, finds a poll conducted for USA TODAY.

About 56% say they favor the northern leg of the billion-dollar, Canada-to-U.S. project and 41% oppose it, according to the poll of 801 U.S. adults completed last month by Stanford University and Resources for the Future (RFF), a non-partisan research group.

More men (60%) than women (53%) support the 1,179-mile pipeline extension, which would carry heavy tar sands from Alberta through Montana and South Dakota to Steele City, Neb. Support was consistent regardless of education level but much stronger among self-described conservatives than liberals, a slight majority of whom oppose it. ...

The pipeline, proposed by Calgary-based TransCanada, has become one of the most contentious issues of Obama's presidency. Environmentalists have waged a grass-roots war against it, saying tar sands' development would worsen global warming and its delivery could risk oil spills. The oil industry and other backers say it would increase jobs and reduce U.S. dependence on unreliable foreign sources of oil. ...

via www.usatoday.com

The poll of n = 801 has a margin of error of +/- 4%. As I now well know (since I'm teaching stats and I know everything that is in the book!), the margin of error is determined by the formula +/- z x SQRT[(.5 x (1 - .5))/n]. With a 95% confidence level, z = 1.96 and the margin of error is 3.5%. So, where does the 4% come from? Are they just rounding up from 3.5%?