FUQed: False Positives, False Negatives, and Authentic Assessment

Last year I asked an econometrics class for a write-up on the basic meaning of causal inference. At that point we had only covered simple regression with dummy variables. The answer I wanted was narrow. When treatment is randomly assigned, a simple comparison of the treated group with the comparison group gives the average treatment effect. What came back were fluent pages on compliers, defiers, instrumental variables, difference-in-differences, and synthetic control. I had put those words in white type on a white slide background, so they never appeared on the screen in class. If I accuse a student of cheating when that student actually learned the material and can explain it, I have a false positive. If I treat the paper as proof of learning when the student cannot state the basic point and the essay came from hidden slide text plus a chatbot, I have a false negative. Either way, I am FUQed. FUQed stands for a Fundamentally Unidentified Question.^[1]

John Paul Rollert’s Atlantic essay argues that this habit can harden into a society of cheats.^[2] Many universities answer by telling students to use AI responsibly. Rebecca Winthrop says that is not enough. We need responsible products. She puts it as leaving a six-year-old at the dining table with broccoli and brownies, saying do not eat the brownies, and walking away for eight hours. The child eats the brownies. Gary Becker’s economics of crime says the same thing in another language. When the benefit of the easy action outweighs the cost, people take the easy action. Teaching willpower is necessary. It is not sufficient. Our job is to build a bridge that can take harsh weather and bad drivers, not a bridge designed only for a good day.

We should not train students for the job market of 2022. The price of code is falling. Jevons’ paradox says cheaper tools invite more use, not less. When computers arrived, accountants had to learn the machine. The same pressure now sits on every major. Andrew Ng’s line is that a marketing student should leave as a marketing engineer, and a supply chain student as a supply chain engineer: a person who can see an end-to-end process and ship a working product, not only a dashboard. Domain knowledge still comes first. Without it, a student cannot tell when a model is making things up, and cannot do systems thinking about where a piece of content belongs. Faculty have that hard-earned expertise. Most students do not, not yet. Our job as an institution is still to validate, authenticate, and credential. The labor market pays for people who can test an idea, connect it, and finish a live problem.

Josh Hall, Milan Puskar Dean at West Virginia University, was one of my professors during my Ph.D. He used to tell me to make teaching into research. I have been tinkering with my teaching ever since, documenting what I try, and writing papers from it. This semester I dropped homework, midterms, finals, and the course book as the spine of the class. I am teaching the absolute theoretical minimum. Lev Landau used that standard in physics. Leonard Susskind’s Theoretical Minimum lectures on YouTube are the version I could follow with no physics background.

In statistics I teach the sample mean, the sample standard deviation, the t-statistic, and simple linear regression. I do not teach the z-statistic, because in applied work we do not know the population standard deviation. Hypothesis tests, p-values, and confidence intervals all use the t-statistic. A two-sample t-test is a simple linear regression with a dummy variable. An F-test is what you get when that regression expands.

In econometrics we start with the basics of R, then go straight to data frames. I teach data management with a short set of verbs: select, filter, mutate, group by, summarize, pivot longer, pivot wider, and join. For graphs I teach a trend line, a density plot, and a county-level and state-level map of the United States. We then move to linear regression and spend most of the remaining time on its properties, its assumptions, and its interpretation. I teach the for-loop. With that tool they can run Monte Carlo simulations, bootstrapping, and randomization inference. We then spend a long stretch on the philosophy of causal inference. Next come directed acyclic graphs, with confounders and colliders. Instrumental variables follow on those graphs. Canonical difference-in-differences comes next, then two-way fixed-effects models.

Every two or three lectures, students defend the minimum out loud. Both courses require oral and written examinations. I built a scalable online oral-exam system, MyMeritGuide.^[3] Professors upload their course material. The system questions each student on that material within a set time window, grades the oral, and flags cases that may raise an integrity concern. Faculty can read the transcript and override a grade. In my class the window is 45 to 60 seconds, and the questions come from my own reading packet, so students cannot wander off the assigned pages. Later we sit down face to face. I have taken 10 to 15 students to lunch for that conversation. You should never eat alone. Eating together builds enough trust that a hard oral is still a human check, not an accusation. We also run a peer-review loop. A student writes a narrative. Two or three peers respond. The author revises. Then the system questions them on their own pages. A chatbot can help with the draft. It cannot sit in for the defense. The product is under university intellectual-property disclosure. Professors can use it at no cost. It does not replace the professor. It points to where attention is needed in a large or online section. That takes the logistics of oral exams off the professor’s desk, so the job is teaching, not policing.

Once the minimum is part of how they work, we use agentic AI. In statistics I have them bring a CV, load it into an agentic system, and then learn hypothesis testing against their own work history. A student who works in a daycare can ask when arrivals spike and how an app should behave. A student who works as a barista can ask about wait times and coupons. In time-series forecasting I can now get undergraduates to ARIMA, then to Harry Markowitz’s portfolio math in a handful of R lines, then to Black-Scholes and to calibration for bets. Last year the final was a mathematical take on Scott Galloway’s The Algebra of Wealth: pull a long S&P 500 history, draw random 30-year spans, invest a dollar, run a Monte Carlo, and then tell me in person in which years you are unlikely to lose principal. There is no textbook chapter for that exact question. The floor underneath it is compound interest and a for-loop.

I do not have a campus-wide AI policy to announce. At Walker College of Business we read university rules against federal and state law and then decide what we can try. The line between using a design tool in slides and using a model to write a paper is fuzzy. I am less interested in policing that line than in whether we send out graduates who confuse familiarity with understanding. Edmund Phelps, in Mass Flourishing, treats dynamism as the setting in which ordinary people put new talent to work. AI will churn jobs. The live question is whether students use that churn to grow a new skill or only to finish the assignment. No model replaces the person who can find the right people, ask, and stay pleasantly persistent.

Our 2028 graduate is not the student who picked a lane and waited. That student can take a problem apart, build with cheap code, defend what they built, and finish the work with other people.