Cheating at Caltech

Seth Roberts

Caltech has a serious problem with undergraduates cheating on academic work, which Caltech administrators appear to be ignoring. A few years ago, one alumnus considered the problem so bad that he urged other alumni to stop donating. I attended Tech (that’s what we called it) for a year and a half in the 1970s. I didn’t think cheating was a problem then. Now it is.

A recent article in the Times Higher Education Supplement by Phil Baty praised Caltech’s “honor system”, which includes trusting students not to cheat on exams. A Caltech professor of biology named Markus Meister told Baty that “cheats simply cannot prosper in an environment that includes such small-group teaching and close collaboration with colleagues because they would rapidly be exposed.” That strikes me as naive. How convenient for Meister that there is no need to test his theory — it must be true (“cheats simply cannot prosper”).

A few years ago, a Caltech alumnus named Peter Seidel, after receiving a donation request, told his fellow alumni not to donate until the system was cleaned up. Here’s some of what he said:

I found out today that Dean of Students Jean-Paul Revel said the following to my dad on the phone while I was at Caltech (Not realizing that my dad is a former Caltech student and BOC [Board of Control] rep) “Peter has a real problem with cheating. The fact is that people cheat. Peter needs to get over it.”

I think it’s safe to say that the Caltech ‘Honor Code’ is obsolete. [= is no longer working — Seth]

There is a small and growing population of students at Caltech [who] are systematically cheating, and the Caltech administration is aware of it but refuses to do anything about it. I suspect the problem began when Caltech started advertising its ‘Honor Code’ to prospective high school students in the 90′s, which lead to self-selection of students who were willing to bend the rules.

In my personal experience, I caught students cheating red-handed while I was a student, and though I took my findings to the BOC, nothing ever came of it.

I also went to one of my professors (along with several of my classmates) and we explained that we were very concerned that there was a significant amount of cheating going on in his class. While he was very empathetic and gave us a significant amount of his time, ultimately he essentially said that his hands were tied because the school does not allow him to give proctored exams.

The Caltech exam system is set up in such a way that it is extremely easy to take extra time on an exam, open a book on a closed book exam, or search for the answers on the internet. Most exams are taken by students alone in their dorm room, with no one watching, at the time of their choosing, with the student timing themself and with both the coursebook and an internet connection in the room, with only the student’s integrity preventing them from using resources they are not allowed to use. For that matter, many quizzes and exams are turned in to unlocked boxes in empty hallways where it would be simple to take another students answered exam to copy or check answers against, and then return it when turning in one’s own exam. <

In my job in the financial industry I interview a number of Caltech seniors every year for potential jobs. And unfortunately, I have to try to answer the question ‘Is this person a cheater?’ as part of my interview process. I have seen examples of resumes where students flat out lied about their GPA.

But probably the most blatant example . . . is a student [he means graduate — Seth] that I recently interviewed [who] claimed, as his two ‘hobbies’, to be a member of the Caltech fencing team his freshman and sophomore years, and a member of the Caltech chess club all four years at Caltech. As it happened, when I was handed his resume, the coworker sitting to the left of me was a former Caltech grad student that coached the fencing team during those years, and the coworker sitting to the right of me was a former Caltech undergrad who was an avid member of the chess club as both an undergrad and an alum. Both of them also happened to be part of the group scheduled to interview this student, and received copies of his resume. I asked them what their opinion was of the candidate.

Neither of them had ever heard of him.

We decided to go ahead and give the candidate an interview, and give him a chance to explain, in case we were somehow misunderstanding the resume. The first person to interview him was the former fencing coach. The interview began normally, and then after a while they had the following exchange (I’m paraphrasing somewhat):

Former fencing coach: I see you have two years on the Caltech fencing team.

Candidate: That’s right.

Former fencing coach: Well, I was the coach at that time… and I don’t remember you.

Candidate: Well, it wasn’t actually my freshman and sophomore years; it was just my freshman year.

Former fencing coach: I was the coach both years.

Candidate: Well, I wasn’t really on the official team, I just took the PE class that taught fencing.

Former fencing coach: I taught that class.

Candidate: Well, I didn’t really take the whole class. I signed up for it, but I only went to the first week, and then I dropped it.

After the first interview, we decided we wouldn’t be making him an offer, but I decided to go in and talk to the candidate anyway. [In] the meantime, the coworker who was a Caltech chess club member asked another chess club friend of his if he knew the guy, and he didn’t. I told the candidate that we wouldn’t be offering him a job, but I wanted to talk to him about his resume. I told him I had heard about the previous interview, and that there were also a couple members of the Caltech chess club who did not know who he was. He responded ‘Well, it wasn’t a formal team, and not everyone went every time.’ I asked him what night of the week the club met, and he told me (confidently) ‘Saturday nights.’ (I knew that it was actually Friday nights.)

When people cheat and get away with it, they are more likely to cheat in the future, Seidel believes — a very plausible idea. Given the disinterest of professors and administrators in the problem, the Caltech mascot should be a monkey with its hands over its eyes.

23 thoughts on “Cheating at Caltech”

  1. Everything should be examined the old-fashioned way unless there is something intrinsic to the material that requires, say, “continuous assessment” examining.
    So the candidates enter a room just before nine a.m., or 1 p.m., and have three hours to answer a paper that they’ve never seen before. They are equipped only with calculators of a specified type – no phones, laptops, books or whatever. Repeat five, six, seven or more times across the course of a week or two. The papers should be set and marked by academics who are not the people who taught the course.
    The Caltech system described here is absurd. Why on earth would anyone think it a good idea?
    Seth: It makes life easier for professors. Much easier.
    Reply
  3. interesting…I also don’t believe the honor code system works. But even without honor code systems, there is cheating occurring!
    For example, most Greek houses keep records of previous exams and finals and then new pledges use them. I went to a college where my affiliated housing unit also had this. I have heard stories of people back in the 70’s using their frat’s finals and I have heard stories of people in the 2000’s using their frat’s finals. Apparently the professors didn’t think it necessary to change the exams over the past ten years, or they cycle them in a very predictable fashion.
    Which just made me think of mark twain about never letting school interfere with my education. I think the greater lesson is that for many people, going to university is almost like a finishing program and cheaters do prosper (Steven A. Cohen for example). From what I saw, many go to college in order to fit into an upper class… and professions that used to not require a degree now do for reasons I don’t understand. David Brooks has spoken about this, nothing new. For what it’s worth, and though I saw both women and men cheat, women were more likely to get caught, confess, and take the penalty for cheating than men were. Caltech is a much higher proportion of men, which makes me wonder if its environ is anything like the Tour de France? Everyone’s doing it and you’re not competitive if you don’t cheat.
    Reply
  4. Seems likely that there is an element that’s not being mentioned in this story:
    “The stereotype, delicately put: first and second generation Chinese, Korean, and Indian Americans, as well as nationals from these countries, often fail to embody the sterling academic credentials they include with their applications, and do not live up to the expectations these universities have for top tier students.
    Less delicately put: They cheat. And when they don’t cheat, they game tests in a way utterly incomprehensible to the Western mind, leading to test scores with absolutely zero link to underlying ability. Or both. Or maybe it’s all cheating, and we just don’t know it. Either way, the resumes are functional fraud. “
    Reply
  7. Steve Johnson and “as” beat me to quoting Education Realist.
    A quick look at Caltech’s student demographics shows it as 39+% “asian/pacific islander/native Hawaiian”. So there’s a 4-in-10 chance that the cheaters are Asian.
    Reply
  8. To paraphrase Charlie Munger, if you’re a shopkeeper it’s immoral if you don’t make it difficult for your employees to steal from you. We know people are more likely to cheat when they feel the chances of getting caught are lower. The university is encouraging bad (but rational?) decision making by students. Better off not testing them than playing this game.
    Reply
  11. For some reason this didn’t post before, so here it is again…
    I graduated from Caltech in the 1980s. Cheating was rare then, as far as I know. At risk of sounding quaint, most students attended for knowledge, not prestige or financial gain. Cheating just wouldn’t be interesting to people I knew there.
    I was aware of one small cheating network: students who knew each other from a particular private high school openly shared answers on homework and take-home tests. That private high school is in a developing country notorious for corruption. It is not any of the countries mentioned in previous comments above, and I won’t say which, lest it spawn unproductive nationalist finger-pointing.
    Not to say the other commenters’ are wrong. It’s just that generalizations about cheaters’ ethnic origin are not actionable. Maybe a more useful observation is that the cheaters I saw at Caltech had, compared to the rest of us, bigger incentives to cheat (financial and prestige compared to home-country alternatives), and lower barriers (knew each other well, and trusted in each another’s cultural assumption that anything goes).
    In the ensuing 30 years, the U.S. in general has become more like those students. Incentives are greater, because college is more expensive, income disparities are greater, and so your lifetime return on investment from attending a top-tier school is now much higher than that from a second-tier school. Barriers are lower, because the Internet eases cheating, and because for a decade, high-profile corruption and fraud in the US has often gone unpunished or lightly punished (finance, banking, corporate governance), reinforcing a message that anything goes.
    Until those things are corrected, clearly a school must police itself to preserve its reputation. Sad but obvious.
    It would sure help if employers would hire more intelligently. When we fetishize a Caltech diploma, and rank it over skills, initiative and enthusiasm, we sustain the huge incentive for students to get that degree by hook or crook.
    Reply
  12. Your comment didn’t post earlier because the first time you make a comment it is moderated — held for approval — and I hadn’t yet approved it.
    I suppose the more people who know that Caltech still has an honor system — that is, makes cheating considerably easier than other schools — the better.
    Reply
  13. The honor system appeared to work when I was there. Decades later, Caltech remains one of the hardest things I’ve ever done. As the school’s president put it to incoming students at our orientation, “we will challenge not just your mental limits, but your physical limits,” meaning sleep deprivation. The intensity was incredible. Exhausting, but rewired me to think better.
    But it does depend upon the student culture. If cheating takes root, an honor system can’t work. I would hate to see them lose a well-earned reputation by not putting a lid on cheating, if it is a growing problem.
    Reply
  14. Actually I just realized the honor system was likely working in the 1980s, deducible from two facts: most courses there were still being graded on a C curve, and individual student scores in applied math and electrical engineering tests often averaged below 70%, with wide variance.
    If cheating were rampant at that time, this seemingly could not have occurred. The bell-curve grading would have driven competition among cheaters, which would either have driven scores up, or driven variance down, or both. That doesn’t seem to have happened.
    I have no idea if any of that is still true today.
    Reply
  15. The president said there would be sleep deprivation? Counter-productive. Sleep is when long-term memories are formed. Many studies have shown that sleep deprivation reduces long-term learning.
    I imagine that anyone at Caltech paying attention, including faculty, realizes there is now lots of cheating. If they (the faculty) keep the honor system in the face of that, they deserve to lose their reputation. It is honest students who suffer. They get lower grades and are tainted by the cheaters.
    It’s interesting that honest students don’t band together and demand proctored tests.
    Reply
  17. Agreed, the celebration of sleep deprivation seems obviously counterproductive.
    I’m curious now to inquire with someone currently involved at the university about prevalence of cheating. Will let you know if I learn anything.
    Reply
  18. Some confirmation of a very recent rise in cheating, but also assertions that core courses are already responding with in-class exams:
    One of the commenters there recalled something I’d forgotten: traditionally, Caltech’s honor code exams are unlimited-time and open-book. This allows exams to ask harder and more interesting questions, which in turn helps solve a challenge in teaching there: huge variance in student aptitude.
    Most Techers scored between 790 and 800 and the SAT math section, but the narrow spread of scores does not imply a narrow spread of aptitude. Instead, it means the SAT was simply too easy for the top tier of students there. The right tail on that aptitude distribution is very long. Traditional in-class, closed-book tests might result in a similar spread to the SAT: half the class scores 100%, and is bored out of their minds, while the other half has a more normal spread.
    Harder questions, open books and no time limits permit a wider variance of outcomes, and more closely resemble real-world problem solving, in which there is typically no time limit. As long as students are honest.
    A second challenge to testing in class is that students can no longer spend 15 hours on a test. Potential problem, because not every student at this level can think both deeply and fast. In-class exams would dramatically reduce grade outcomes for honest students that think deeply but deliberately — a profile that may exactly describe students destined to make great contributions later.
    So I don’t have a solution, but begin to appreciate the scope of the challenge.
    Reply
  19. Cheating is a great indicator that something is wrong with the educational system in question. But, hey, it could be worse. Some teachers I work with suggest they’d be happy if their students attempted to cheat. At least that would show they cared, and they might even learn a thing or two.
    Reply
  20. Those are good points. I agree, a good reason for take-home exams is to pose interesting thought-provoking questions.
    Here’s another point: Should grades be based (partly) on how smart students are? Or would an ideal grading system be solely a measure of whether they have mastered the material put in front of them by the teacher? If the latter, it is not so obvious that you want to pose problems that separate geniuses from everyone else. If you want to measure learning, not intelligence, you are pushed toward simpler questions.
    In my own teaching I avoid tests (and grading) altogether. My goal is learning. I find that other sources of motivation can replace grading. I have no interest in ranking my students according to their intelligence. I wonder why other teachers want to do this — or are they mindlessly copying how they were taught?
    Reply
  21. ‘I find that other sources of motivation can replace grading. I have no interest in ranking my students according to their intelligence’
    That’s the honorable way, it seems to me. Installing a grading and status widget into a student’s head is crippling for life, and once it’s in there, it’s very hard to remedy. I speak from personal experience.
    Reply

Leave a Reply

Your email address will not be published. Required fields are marked *