Schneier on Security
A blog covering security and security technology.
« Internet Quarantines |
| Security Haiku »
November 16, 2010
Term Paper Writing for Hire
This recent essay (commentary here) reminded me of this older essay, both by people who write student term papers for hire.
There are several services that do automatic plagiarism detection -- basically, comparing phrases from the paper with general writings on the Internet and even caches of previously written papers -- but detecting this kind of custom plagiarism work is much harder.
I can think of three ways to deal with this:
- Require all writing to be done in person, and proctored. Obviously this won't work for larger pieces of writing like theses.
- Semantic analysis in an attempt to fingerprint writing styles. It's by no means perfect, but it is possible to detect if a piece of writing looks nothing like a student's normal writing style.
- In-person quizzes on the writing. If a professor sits down with the student and asks detailed questions about the writing, he can pretty quickly determine if the student understand what he claims to have written.
The real issue is proof. Most colleges and universities are unwilling to pursue this without solid proof -- the lawsuit risk is just too great -- and in these cases the only real proof is self-incrimination.
Fundamentally, this is a problem of misplaced economic incentives. As long as the academic credential is worth more to a student than the knowledge gained in getting that credential, there will be an incentive to cheat.
Related note: anyone remember my personal experience with plagiarism from 2005?
Posted on November 16, 2010 at 6:36 AM
• 79 Comments
To receive these entries once a month by e-mail, sign up for the Crypto-Gram Newsletter.
"As long as the academic credential is worth more to a student than the knowledge gained in getting that credential, there will be an incentive to cheat."
Maybe consider "to a society" as well.
Both my wife and I teach college. Every semester she receives dozens of art history, history of architecture, etc. term papers from her students.
When some knuckle-dragging Neanderthal all of a sudden begins to wax eloquent on paper she becomes suspicious.
So we plug suspect phrases into Google and "Bingo" the plagiarized source appears.
It becomes a game for us. Who can find evidence of cheating first. It's like shooting fish in a barrel.
There are plenty of people out there who get others to complete their job applications, write their resumes and complete applications for places at University or to get funding for research or work, then they put their name and signature on the documents. And can pay someone to tutor them to understand the content without needing to think much. How many famous speeches were actually written by the person to delivered them, acting script writing rote learning etc. So is any of the above morally equal to plagiarism?
I find it amusing that "Ed Dante" from the first article has written papers on ethics. (In the big list of topics he has written about.)
"anyone remember my persona experience with plagerism from 2005"
I remember it well, as I do other cases of plagiarism.
However I also remember me unintentionaly doing it (with "welcome to the goldfish bowl").
And this is an important question at what point does using a phrase or two that you might have heard in the past become plagiarism.
And don't get me wrong I don't begrudge people taking my ideas if they can use them or put work into improving them as long as they atleast tip their hat in my direction. That is yes you can take my work as long as you do something productive with it, but don't claim it as your own unless you have put real work into it.
As I know technical authors find they get accused by others of taking liberties with ideas, trademarks, copyright etc by those who sometimes you think are little more than patent trolls, which is why we see the disclaimers in the front of books.
But as Sir Issac Newton pointed out with his "standing on the shoulders of giants" a work can be both derivative and original. The more mature the field of endevor the less easy it is to make an original contribution that stands out sufficiently to be described as such.
I do not know where the balance point is between derived work and plagiarism is and I assume it is actually a moving target.
For instance various governments have tried prosecuting people for breaching official secrets simply because they had collected publicaly available information into collections of information, and the prosecution has argued infront of the judge that the collection is more than the sum of it's individual parts...
So again I guess the argument boils down to "what's in it for me" when an accusation is made.
I guess it's an argument that is going to run and run.
This happens more frequently than one would expect. When I was in first year at University (Glasgow) I had been asked to code somebody's programming assignment - I told the guy to take a run and jump. I think he eventually found someone to do it.
Thankfully, no. 3 in Bruce's list caught him and his mates out; that and exams are the protection against plagiarism.
"As long as the academic credential is worth more to a student than the knowledge gained in getting that credential, there will be an incentive to cheat."
The paper is all to frequently used as a filter to find 'those who know' from 'those who do not know'. Are we really teaching topics that are useless to students that they can learn nothing and still succeed in a job after graduation?
(if using Word)
* You might mandate that the student turns on revision tracking and a five minute auto-save interval.
* Require the students to submit their outline before they start.
* Mandate the use of a template that tracks the source of text changes (paste, keyboard, Insert file, etc.)
None of these would catch a transcribing plagiarizer.
Discussing this problem with a friend who teaches at a local community college, I suggested that he begin the semester with an introduction to the cheats that have been kicked out of school and the number of tools at the teacher's disposal.
I don't think the entire writing process needs to be proctored, but a couple of sessions should suffice.
Goodness, was that 2005? Yes, I remember. Apparently I've been reading this blog longer than I thought :D
That reminds me of when I was in secondary school (thats grade 9-12 in the US iirc) a good 15 years ago (in the Netherlands btw). I had to write 10 bookreports (more like summaries) and each time I delivered one the teacher would say that he knew I was cheating and would find out where I got it from the internet. He continued the same story for the first 9 reports. Obviously he never discovered where I got them from. (I wrote them myself, I didn't hate reading at all.) Ironically he never complained about the 10th even though I was a tad more lazy on that one and used another report as template for structure.
Of course, that's nothing compared to essays and thesis work. I cannot imagine someone actually graduating on such a paper. I would've expected that cheaters would've been discovered via questioning and mandatory presentations.
An incredible collection of intellectual laziness on the part of the teachers.
If you ask questions which can be plagiarized -- then you're too damn stupid to be teaching. If the blue-book solution is too expensive -- then it's all just fake jumping through hoops anyhow. If your students can plagiarize a thesis that is acceptable -- oh my god, you're willing to accept anything as a thesis subject anyhow.
Plagiarism isn't the problem in academics. The problem is that most academics have been reduced to diploma mills anyhow.
Discussing this problem with a friend who teaches at a local community college...
Well, there's the problem. Why the hell do we have "community colleges"? The CCs and a big chunk of universities are simple glorified high-schools posing as institutions of higher-learning, eating up massive amounts of funding. Diploma mills essentially, selling diplomas.
So what if the papers are plagiarized? Shit, the students who do that half-decently are already by far the best students in those organizations.
That’s an amazing essay. I’m not surprised that people cheat; I believe cheating (and lying, stealing, etc) is the second oldest profession. What does surprise me is how much money people are willing (and able) to pay for his writing services. I certainly didn’t have that kind of disposable income while I was in school.
I remember an experience in an Engineering class at college where during a test, the proctor, for some unknowable reason, left the classroom for 20 minutes. Full-scale cheating ensued, with at about 50% of the class participating in it. The professor quickly picked up on what happened, and made us repeat the test.
I believe that you cannot teach an Ethics class at the college level and actually hope to make people more ethical. Morals and ethics are taught (and caught) from parents and other figures of authority from infancy to about 16 years old, not sitting in a classroom somewhere.
Widespread cheating is more of an indictment on parental mores and behavior than an indictment on the educational institution, but the school is the one responsible to try to level the playing field for those that do honest work.
>>So what if the papers are plagiarized?
I also teach a martial art and equate academic plagiarism with cheating in the dojo. As the instructor, it is MY reputation that can be sullied if I promote a student to a rank they haven't earned. Accredited institutions of learning (at ALL levels) are in the same position -- school and teacher reputations are diminished or ruined if we fail to catch cheaters.
BTW..."Back to School" discussion between Rodney Dangerfield and Sally Kellerman glossed over the personal accountability (both teacher and student) issue.
My experience in teaching (though this was not extensive) was that even if students didn't 'fess up, it was still the professor's judgement on whether to simply fail the student. While I'm sure some might contest such a policy, usually the threat of taking things further quiets them.
Yeah, it's not as 'just' as getting them sanctioned, but it tends to serve its purpose well enough. The risk is if schools follow along a standardization path to the extent that this sort of discretion is no longer allowed - if they technology's not there to prove plagiarism, we need to rely on good old fashioned human judgment.
This omits one of the economic incentives that predominates in these days of budget cuts: When the academic unit values the tuition dollar more than the integrity of the teaching process, cheating is inevitable.
The Chronicle article says what many of us in academia know - cheating happens and is often rewarded (or ignored) rather than punished. I have found plagiarism everywhere from freshman papers to submissions to the federal government by well-established researchers (where the rewards are directly monetary). Per @billorites, Google can find things like this just as quickly as the original plagiarists found their shortcut.
Sadly, the liability for these accusations lies with the instructor - who will find himself with little support from the academic unit. It is hard to make a case for academic integrity when it does not reach any higher than those directly in contact with the students.
@aikimark if not using word (latex ftw), setting up a version control server and mandating some work behind made under the supervision of the teacher could show some cheating patterns.
A good thing we had in our curriculum (not an US school tho) were oral presentations of papers. 15mn to explain what a paper contains to a class of students is short, but hard to fake.
I used to TA pre-calculus/Calculus. I remember being surprised at the extremely high cheating ratios on proctored quizzes. (The first one would often be over 40% with some evidence of cheating, and a small percentage with fairly definitive evidence of cheating) - the quizzes weren't identical, but were designed to look very similar from a distance.
I viewed my task for the first few weeks, other than teaching the subject, was to get students out of the habit of cheating.
As for thesis work, I actually was in a class that did proctored thesi. I hated it, in no small part because the sources I could use were very tightly constrained. I like your quiz suggestion, though.
This omits one of the economic incentives that predominates in these days of budget cuts: When the academic unit values the tuition dollar more than the integrity of the teaching process, cheating is inevitable.
The Chronicle article says what many of us in academia know - cheating happens and is often rewarded (or ignored) rather than punished. Tom Lehrer's "Lobachevsky" is a great example of humor taken from the long academic tradition of intellectual theft.
I have found plagiarism everywhere from freshman papers to grant submissions to the federal government by well-established researchers (where the rewards are directly monetary). Per @billorites, Google can find things like this just as quickly as the original plagiarists found their shortcut.
Sadly, the liability for these accusations lies with the instructor or reviewer - who will quickly find himself with little support from the academic unit. It is hard to make a case for academic integrity among students when it is not practiced by the institution.
@kangaroo: Incredible laziness? Let's say a teacher has the ability to tell that a paper is original and well-written, and assigns a paper on some obscure topic. Cheating student pays a ghost-writer for a paper especially written on that topic, and turns it in. Remember, this is not a case of looking for a pre-written paper on a topic. Now what's the teacher to do?
The teacher doesn't control the students' finances or communications, so there's no hope of identifying cheaters there. The teacher may be able to discuss the paper with the student, one on one, but that isn't always practical (not all educational institutions leave enough time for that). The teacher can try to learn students' styles, but that's going to be chancy at best. Moreover, since cheating is a serious thing, an accusation of cheating is a serious thing, so the teacher really does have to have a solid case to proceed with.
If you'd like to show how moderate diligence can stop the problem of custom-written class papers, I'm sure lots of people would be fascinated.
Plagiarism has been a problem in academic (and other) circles since time immemorial. There were people who would write your papers for a fee when I was in college (mid-1960's), or even take your final exam for you (usually really large classes where the professor/proctor would not likely know you or the proxy). It all comes back to that still-too-true saying - those who have the gold, make the rules...
In Poland we have some implementation of method #3. When you're writing your thesis you're doing it in parts (usually chapters, but it depends on agreement) and your professor is discussing with you every part (also suggesting adjustments, pointing errors, etc.). It's quite easy to detect a plagiarism then, but reality is that many professors are corrupted. So not only method is important, but also the person who will be using it.
Actually, I think that saying should more properly be:
Those who have the gold, make the rules. Those who want the gold, break the rules.
Doh! I see that I am repeating your link in the article! Sorry about that! Honest (but stupid) mistake. -Spiff
How does it deal with false positives occurring purely by chance?
The problem I see with trying to catch programmers who cheat is that, along with the cheaters, it is conceivable that it will falsely accuse the best of the students - i.e. the ones who write really efficient code that actually resembles what might be put into production. Granted, there are lots of ways to approach any particular problem, once you limit yourself to approaches that are close to ideal, you eventually HAVE to write something similar to another person's ideal approach.
@Piotr - I've never been to Poland, but when I spent time in Albania, university students there told me that paying for grades and graduation was commonplace, and many professors doubled their incomes by taking cash.
In my college years in the US, I saw or heard of no comparable professor-level corruption (besides the occasional "casting couch" story).
Semantic analysis in an attempt to fingerprint writing styles. It's by no means perfect, but it is possible to detect if a piece of writing looks nothing like a student's normal writing style.
From the comments, it's quite clear that many professors pick up on it and have no way to prove it or enforce the rules. So an automatic semantic analysis would not help, and in fact seems likely to make things worse, since now you're adding rote profiling and machines to the mix.
I have not been too impressed by what semantic analysis I've seen. Most of it seems like it would catch me for varying my writing voice to the audience.
The folks who pay people like that end up trying to plagiarize during their job interviews. I don't hire people who do that.
>> The paper is all to frequently used as a filter to find 'those who know' from 'those who do not know'. Are we really teaching topics that are useless to students that they can learn nothing and still succeed in a job after graduation?
It's hard to say whether the topics are actually useless, or if the students just believe them to be useless. I remember having a student in a linear algebra class once who was failing, and insisted that I give him an (unearned) pass so that he could take a job where he was expected to have successfully completed the course. It was a software development job and (naturally) I refused to pass him, and instead took to trying to point out how important linear algebra is to actually being able to come up with a decent algorithm for a problem. I don't think that I actually managed to make my point, and I suspect that the student in question still hates me to this day.
Oral exams. My Eastern European mother was shocked and dismayed to learn that some of my North American exams were multiple choice.
In her day, the sole purpose of a written exam was to qualify you for the oral exam, which was held in a lecture hall with open attendance (other students, prospective students scoping out a prof).
Two benefits to this scheme:
1. students have to learn *something* or they're publicly humiliated
2. profs can't be lazy and ask the entire class the same question
Mind you the state paid for the schooling so there was an incentive to make people work. In our western user-pay system, the economic incentive is for factory schooling and cheap, lazy evaluations (more money for the school), and for more systemic tolerance of cheating (because the clients are, in effect, paying for a degree).
A tenured faculty member with years of experience could potentially implement these solutions, absolutely. Unfortunately, an increasing amount of most departments' teaching workloads are carried by adjuncts and visiting assistant professors who have little experience with student work or time to engage in one-on-one student quizzes or other involved plagiarism detection methods. They are often teaching classes for the first time and thus having to craft the lessons/lectures during the semester while participating in the job market, conducting research, writing articles/conference papers, and sometimes even working second jobs. Plus, class sizes are getting bigger and bigger, which means more and more students to grill, papers to grade, and writing styles to identify. Then, when they do catch a cheater, they have to rely for support on a department and college that sees them as an expendable contractor and the cheater (and the cheater's parents) as a long-term cash cow and/or potentially plaintiff in civil court.
Ironically, one of the best checks against cheating out there is the term paper market itself. Despite this guy's claims, a recent NYT article found that purchased term papers are often of extremely poor quality, written by non-native speakers, and often themselves plagiarized. There is little incentive for the seller to craft a good paper since the buyer has no recompense if they aren't happy with the product. What are they gonna do, call the BBB?
Plagiarism of a publicly available work is one issue; paying someone to privately write your paper is another. While both can be subtle and sneaky, I feel that the latter is more difficult to catch.
Can you PROVE someone cheated through semantic analysis? How? Its utility may be more in identifying whether there MIGHT have been something fishy going on, and this could be confirmed upon employment of any of these other techniques (asking to see a draft, questioning the student, debating the topic with them, etc.). I'm not knowledgeable of the metrics used in these programs, but one obvious weak point could be if a student buys a research paper and then changes the words that they feel they "wouldn't use" with a "Find-Replace" strategy (or by simply going through it manually - still probably faster than writing the darn thing).
The best way seems to be to monitor the actual process of drafting and writing, through proctored exams or demanding to see outlines/drafts. As has been noted, an alternative, if the prof has time for it, is peppering students with questions - this proves that they at least learned the material.
And it's not just writing papers - I once met a kid who graduated from a certain respected computer science school (to keep it anonymous, let's call it SUNY Tony Rook) with a CS degree, without ever having written a single computer program.
"As long as the academic credential is worth more to a student than the knowledge gained in getting that credential, there will be an incentive to cheat."
This should be testable. The more valuable the credential the higher the cheating rate should be.
The lower rates of cheating should correlate with the lower payed BFA/Social Work type degrees while the highest rate should be in the law, finance, and business administration.
I think there was a study to that effect too.
So is there a conflict of values here between academia and the knowledge seeking fields and the 'real' world of business where outcomes only, are rewarded?
What a surprise. The business world rewards cheating. What else would you expect from organizations who argue their only ethical concern is return on shareholder investment.
I am told at Apple they call it "incorporating ideas from elsewhere and polishing them".
Of course you can't polish Apple's products in return; it will get you slapped with all kinds hot legal warnings.
Perhaps you could add another way to deal with plagiarism, as I mentioned in 2005, by publicly outing the plagiarist. You mention pro-active detection and prevention, but what you are doing with your blog is reactive and may be just as effective.
I'm an English instructor at a major state university, so I've had my share of things copied and pasted from Wikipedia.
As long as the student doesn't crib all of the writing that you've seen, it's pretty easy to tell when the language isn't theirs (even for an overworked adjunct like myself). So if you assign in-class writings, you're pretty sure to have a sample of the student's actual writing to compare to.
Most students are also not actually good at cheating -- the whole process, that is. If you have some samples of their actual writing alongside something you suspect is not theirs, you can call them into your office hours and ask them questions about the suspicious piece: most will admit that they got it from somewhere else, which alleviates the need for any investigation.
Now, I don't think there's a defense against a sufficiently well-executed custom paper, but that requires a few unlikely things: (1) the false work is in a significant way similar to any samples of the student's actual writing, (2) the student is calm and assured enough to stand up to questions about the paper (I'm not talking cement-room-bare-lightbulb here; just "can you explain the argument of this piece?") and (3) the administration/review structure does not value the instructor's opinion. This combination seems rare enough that a careful instructor should be fine.
Now, if you're teaching a 5/5 course load and have 120 papers to grade every few weeks...
Really what you've got is an outcome from signaling theory.
In the market place, the end result (i.e. degree) is worth more than the knowledge gained.
One way to look at it is that the relevance of what is taught is in question. If the students are simply interested in acquiring the signal as an indicator instead of a necessary skill, this is clearly self maximizing behaviour. The change would have to be in the relevance of the material.
It's ironic, beautifully ironic.
The idea that the degree is worth more than the cost in time, tuition, debt, and opportunity cost is itself in many cases going to be found to be erroneous, as witness all the talk in the blogosphere these days about the higher education bubble....
The university sucks in money, bloats itself like Jabba the Hutt, awards you a degree you did not really earn, for money you did not have, so you can be underemployed, incompetent, and indebted. All that cheating for practically naught.
"In-person quizzes on the writing. If a professor sits down with the student and asks detailed questions about the writing, he can pretty quickly determine if the student understand what he claims to have written."
I happen to be a professor, and I'm fairly comfortable asserting that this option is the best. I find that an oral exam (i.e. student presenting their work and answering questions in person on the spot) is the single best way to measure a student's mastery of the material. This is exactly what I do in my upper level classes (which are small enough to spend several hours for each student).
However, it isn't feasible in the lower level undergraduate classes. There are just too many students to spend that much time grading each one.
I'm perplexed, though, that this sort of thing can go on at the graduate level. My PhD committee grilled me mercilessly on multiple occasions, and my adviser required me to give many many professional talks as I was working on my dissertation. There's just no way that I could have passed off the work of someone else as my own. Of course, I'm in a very technical field... perhaps things are different in the softer fields.
Of course, the problem with semantic analysis is that it can only tell X wrote pieces A, B and C; it doesn't tell you if X is sitting in front of you or not.
Imagine a recidivist essay-purchaser, who uses it almost all the time - or, at least, all the time for a specific course.
If an essay-writing service is organised in such a way that a client is generally "handed back" to the same writer they used the first time - which is simple enough for a series of essays in the same course - then your semantic analysis will happily report they all check out...
option 3 seems perfectly reasonable. Part of the assignment for the term paper can simply include an oral component where you have to defend what you have in the paper, and the professor/TAs/whoever will question you on it and try to poke holes in your thesis. This is standard for a PHD thesis, no reason you can't make a smaller scaled down version standard for undergrad term papers. It actually does have an educational component too. (Thinking on your feat, defending your ideas in the face of alternate viewpoints, etc.)
At least in central Europe it's quite common to have someone proofread & edit/correct a large work like a thesis before submission. Sounds pretty ambitious to me to try & determine a "fingerprint" of writing style in that case.
0: for grammar, spelling and writing style, and correct references/footnotes, _not_ the actual academic content.
I could see some combination of #1 and #3 working: require one proctored writing assignment, from which the instructor gets a sample of the student's writing. For future assignments, the instructor may choose to compare the new writing with that sample and, if suspicious, quiz the student on the finer points of the paper.
I agree with Mr. Dante (and many commenters) that this is mostly a question of what incentives and disincentives are present for schools (hence instructors) to catch plagiarists and hold them accountable. Don't forget that the PR generated by plagiarism cases rarely works in favor of the schools involved.
The economics of cheating is a great model--thanks! In addition to perceived value, there's costs and resources available.
The costs of cheating include the possibility of getting caught. I try to deter cheating by telling students about the uncertainty in cheating, e.g., when two students have identical wrong answers it is very conspicuous; if a student copies another's work, there's always the possibility of copying a wrong answer that will be conspicuous to the grader...
My favorite teacher in high school--Physics--was what he called "tough but fair". He treated his classes like college-level classes: exams 70%; quizzes and labs 15% each.
He said that once, he found two identical papers submitted to him. Deserving of a 100, he gave them 100--50 each. If four students wrote a paper "together" that was deserving of a 100, they'd get 25 each.
So split it up with the Google results. If Googling the phrase finds 4 hits and the paper is deserving of a 100, the student gets 20 for doing 1/5 of the work. :-)
The problem underlying academic cheating is that most educational institutions care not one whit whether students learn anything. Schools are run by bureaucrats and by power-hungry or money-grubbing administrators.
In public schools, the administrators are concerned about average scores on standard exams and dropout rates. They address the latter problem by never investigating why 16-, 17-, or 18-year-old students stop coming to class and pretend that the students' families moved. They address the former problem by teaching to the standardized multiple-choice tests.
In colleges and universities, it's all about money. They now accept students who would never have been considered smart enough to attend a community college back in the 1970s. They cram students into massive lecture halls, spoon feed them pap from high school-level textbooks, give multiple-choice tests with simplistic questions taken directly from lecture notes, and either give no writing assignments at all or ones that require little work and are graded by teams of grad student teaching assistants who receive no guidance.
The outcome of this indifferent and uncaring approach to education is obvious: college graduates who wasted four to six years getting a Bachelor of Arts degree but no real education. Cheating is rampant because students put no value on learning a subject; they value only the credit hours and the grade.
I was a medical student in the late 1970s, and I have taught medical students and residents since the mid-1980s. Each year, the average first-year medical student arrives with less background knowledge and a lesser ability to learn complex subjects than the year before. The decline is appalling and partly explains the large increase in the numbers of incompetent practicing physicians. (The other parts of the explanation are that medical schools also fail to educate, and the residency training process rewards memorization and pattern recognition over knowledge and a skillful approach to diagnosis and problem-solving.)
Our entire educational system needs to be redesigned to focus on learning rather than bureaucratic needs or ineffective teaching paradigms created by idiotic pseudoeducators. The new educational system needs to address the needs of all students, including the one in nine with IQs below 80 and the one in nine with IQs above 120. It needs to tailor education to students with different learning styles and be flexible enough to teach those who learn by handling and doing as well as those who learn by reading. It needs to allow students to choose education options that span from minimal formal training to trade schools to focused higher education classes to graduate schools. But, it is my belief that this will not happen, and that we will muddle along as we are for decades.
Since the article in question is about a paper that wasn't plagiarized in the usual sense, things like Googling for it won't work. However, there is a technique that is supposed to be pretty good.
Find a key paragraph. Take the first line or two. Present them to the student. Request he finish the paragraph.
I have never tried it, but it is supposedly quite reliable at catching even competent people working off a paper that isn't theirs.
I agree with you completely. I haven't been teaching as long as you have, but I have seen the same trend. Ultimately, it all comes down to money, of course. As a society, we have repeatedly cut education funding while demanding higher graduation rates. The result is that the educators (pressured by the administration) deliver the higher graduation rates in the most cost-effective manner---lowering standards.
We all build from the knowledge of others even those who produce unprecedented highly accomplished works. If we imbibe something and spit it out in another form that is real learning. If we copy paste it is plagiarism. In between there are many shades of gray. Where is the line exactly?
I find it interesting that so many readers are dealing with this as an opportunity to indict our education system - surely indictable! - rather than as a security/verification problem. While "Dante" wants us to view him as a symbol of a failed system, Bruce perhaps would urge us to think more specifically about the problem; indeed, he has done so in his post.
I think the use of Word time-stamps is interesting, though perhaps harder to implement and ultimately defeatable.
One approach that occurs to me is to use assignments for which this kind of purchase is difficult or impossible. These include: in-class writing, short-term take-homes where the turnaround is such to make a purchase more difficult, specific text-based assignments in which the author MUST have the text at hand. They allow us to create a kind of stylistic fingerprint to compare with a later, more sophisticated essay, that emerges. In combination, we might be able to start building a case.
But administrative support is crucial, and all too rare.
I'm not sure that anyone has mentioned this, but sometimes the paper is just for the grade! it is a subject that is a must do but is completely unrelated to your major. Under these circumstances, and given the high course work load that I had (EE and Biomed) I would have been very tempted to just buy some BS arts and humanities papers. I certainly didn't want to do the course and my attention was definitely elsewhere.
I realize it is intellectually dishonest, but so is the process of loading students with unrelated subjects that are cheap for the college to run. I've heard from several prof's that engineering courses, especially those with some practical aspects, are 10 times as expensive (per credit hour) as a humanities course. This cost differential is largely responsible for the "humanization" engineering courses.
Given that Colleges want to push the cheapest solution, is it any wonder that students seek their own "no work" solution?
We have NOT cut education funding. We spend more and more on public schools and colleges for less and less education (and more and more waste and bureaucracy). In public school districts salaries paid to administrative staff approach those paid to classroom teachers. At universities, despite increased funds, the ratio of tenure-track professors to students gets lower and lower. Administrators are sucking down the extra money and are hiring part-time faculty or poorly paid, nontenure-track instructors to handle the massive influx of students. Lecture halls, classrooms, labs (and dorms) are packed. I had medical students who told me that their undergrad science classes had over 300 students. Labs were essentially worthless. Not one of my students had dissected anything more complex than a frog before medical school. That's pathetic.
On final thought:
What's the difference between "buying term papers" and employing a private tutor, who maybe helps a little to much?
All my kids have private tutors, who come for over an hour each day to help them with assignments, extra topics. I know I've read a few assignments that were definitely more tutor than student! But I rationalize the "dishonesty" by telling myself that they learned something through the process.
Is my predicament just the rich mans variant of the same problem?
About a month ago, a close friend of mine who is in college was accused of plagiarism. She went to see her Professor and was shown a paper that had a passage surprisingly similar to hers. Tracking down the full text of that other paper showed that it was citing her paper with complete and correct attribution.
Plagiarism is supposed to require passing off another's work as your own. Yet accusations of plagiarism almost never are preceded by any attempt to determine who authored the identical section.
Regarding #2, while it wasn't actually a plagiarism issue, in high school a friend of mine forgot to get a test signed by a parent and so asked me to sign her mother's name (so it wouldn't look like her own handwriting). I did, and it actually became a fairly common thing. One day the other student actually remembed to get her mother's signature and was asked to stay after class and informed by the teacher that "the signature was not her mother's." The student agreed, said she had just forgotten and didn't think it that important but would get her mother's signature that night. I signed the test and the teacher was satisfied.
A 'comprehensive' test writing package would provide all papers, outlines, and edit versions (possibly from an automated system).
#3 is actually an excellent option as even if the paper was not written by the student, if the student becomes sufficiently familiar with the subject to answer questions, they will have learned the material anyway.
That in itself has some interesting security implications. If your security system sufficiently converts enemies to friends, then then system works even if it is somewhat porous ("Salt" the sleeper agent is another example).
andrej is on to something:
The real test of a thesis is not just the written paper, but also an oral defense, and/or a presentation of the research in that thesis to a public and possibly hostile audience. A school that is unwilling to put students through that has no business graduating PhD's.
As far as whether these papers actually matter, I'll just say that if you are unwilling to do the work, don't take the course. If you're unwilling to take the courses that you need to get the degree, don't get the degree. And you'd be surprised what helps you out in the real world - as a software developer, my knowledge of English writing style is at least as useful as my knowledge of C++ syntax.
In case you haven't seen this (there's no meaningful date on the news article), apparently several people involved in your original plagiarism post, including the department head Tauseef-ur-Rehman, have been caught and forced to resign in a corruption and nepotism scandal a mile long, involving land deals, faked theses, phony journals, etc.:
Crikey. At some point, isn't it easier just to do the work...?
@RobertT: One reason you take humanities courses is to wind up as a more fully educated person. College should not merely be a way to get high-level vocational training.
Moreover, the technical courses I've taken were generally not demanding in the ability to write good English prose (academic papers usually don't qualify as good English prose). That was taught and evaluated in English and humanities classes. It's a skill that will serve you well if you bother to pick it up, rather than considering it useless.
I think the economic analysis here is, alas, flawed by an apples-to-oranges comparison. As long as prospective employers use degrees as a sorting mechanism for resumes, the credential is always going to be worth more than the actual knowledge in the short term. Perhaps in the medium or longer term some piece of knowledge (or the ability to marshal knowledge effectively) will be more valuable than the credential, but there's a lot of uncertainty in that, and students are not known for long time horizons.
If colleges and universities wanted to get rid of most of this practice, they'd: a) support faculty members who recognized cheating and discipline those who collaborated with it, and b) start up their own clandestine paper-writing operations to poison the well.
@Bruce @Jason "... forced to resign in a corruption and nepotism scandal..."
I guess they didn't really mean it when they apologized.
IMO, you need to have adequate controls in place to make getting good grades by cheating difficult.
If the class involves writing, make that a significant part of the grade, but use hard tests to catch the cheaters.
If cheating means that you're going to pay for A papers in order to get a C- in the class, you've established a pretty significant economic disincentive for cheating.
You are correct in that the total amount of money spent on Universities has increased over the decades. I should have been much more specific as I meant the money spent on faculty (per student), which has certainly decreased.
In fact, during the current budget crisis, my University has had a total budget decrease of around 20% over the past three years. During this same time the president of the university got a 30% pay increase, we spent over 100 million on the football stadium, and we hired several new coaches for another multi-million dollar cost. We were told that "this money comes from a different pot, so faculty shouldn't complain," however to me it is very clear where the priorities fall by how well each "pot" gets funded. And, lest anyone say, "athletics supports itself," let me point out that our students pay over $1,000 each per year to finance the athletics program, so it certainly isn't self-supporting. That portion of their tuition is only for the varsity athletes, the funding to intramural athletics, the gyms for normal students, etc. are taken from a separate "student activity fee" which is on the order of $200 per student.
Maybe detecting isn't the solution, maybe the solution is to reduce demand and decrease supply/increase cost of purchased essays. So schools could:
* Sue paper-houses for undermining the integrity of their grading system. Even if they lose, this would probably increase the price of papers due to increased risk of being the the buisiness
* Charge people caught signing into websites to submit papers that aren't theirs with criminal impersonation (my university did this to close down companies whose business model consisted entirely of providing people to take your tests for you)
* Have stings where they advertise fake essay services and punish the students who contact them. This would increase the risk to students contacting services
* Persue students who are caught. Schools fear legal trouble, but if a student sues them, they can then supeoana their credit card, email, etc. and probably get proof.
"Granted, there are lots of ways to approach any particular problem, once you limit yourself to approaches that are close to ideal, you eventually HAVE to write something similar to another person's ideal approach."
The MOSS paper discusses false positives
Couldn't agree more. After running a couple of businesses (which in turn consulted for many others), I learned very early that paper qualifications are useless (or worse, people who try to collect a bunch usually have something to hide thereby), HR departments that look at them are just doing the CYA dance, and that even if the guy didn't cheat in school, it's likely he didn't learn much of importance. As in how to learn -- which by golly, cheating might teach someone, by practice at using available resources efficiently and quickly.
I looked always for people with work ethic, hard to fake, and inquisitiveness -- people who live to learn, who do at home as a hobby what they want me to hire them to do for a living.
This automatically excludes people who think a nice looking piece of paper means they've got a ticket on the sinecure gravy train for life.
Funny thing, once I'd done some great work as an engineer in that first job -- no one else ever asked about my paper qualifications and I never had to actually look for work again -- word of mouth is powerful.
As someone who now trades stocks for a living, which includes intense study of productivity, it seems that right now, with all this unemployment, that only those people got laid off -- productivity is at an all time high per man-hour.
Since other successful businessmen probably have some of the same criteria as me, I'd bet it's just those cheaters (or other forms of loser) that they accidentally hired are the ones who lost their jobs in the main (I'm sure some innocents did too, don't take me wrong). I know that now that I'm trying to hire, what I get is just those people, and no way. I quit looking through the normal channels, there's no point and the signal to noise ratio is one over infinity at this point.
Myself -- I got hired before I was able to finish college, and at higher pay than my classmates got when I hired some of them in turn when they graduated. No substitute for real talent and work ethic (I was fortunate to have smart parents who'd already taught me to the level a BS would have given me anyway). No way to say that that doesn't sound a little like boasting, but it's the case.
Now, one thing that's worked for me -- I hire college students while they're still in school, if they meet the other criteria. I teach them what the schools cannot (those who can't do...), and pay them beer and pizza money at first, with the agreement that they'll work for me for a couple years after they graduate for something in the good six figures. Sadly, I usually lose them after that, as I've taught them enough (and picked the right guys) that now they are all CEO's themselves.
You can't cheat your way into that, or not for very long...
I was fortunate to attend a university which was largely free of the grade-grubbing cynicism which underlies some of the market for custom papers and the factory model system which enables it. At Oxford and Cambridge, teaching in all subjects is done via weekly tutorials where students read their paper and then discuss it for an hour with the professor and at most one or two other students. It would be hard for a cheater to hide in the intense scrutiny of a tutorial. Furthermore, grades for a permanent record did not come from the tutorials but were marked from proctored final exams administered anonymously over a hellish few weeks after three years of study. It was in the student's interest to work hard and to participate fully, and the esteem of our fellow students and professors was far more motivating than a simple grade. The system of debate and challenge as well as the separate final evaluation process worked against political correctness and parroting professorial views.
Of course, the tutorial system was hugely labor intensive and would be difficult to replicate at a huge state university, but there are elements that could be incorporated into American education- perhaps a successive small portion of the class could participate in a weekly tutorial with the professor, and the final grade could come entirely from a proctored final exam at the end of the course.
> as a security/verification problem.
This, IME, is actually the most common method, right here:
> Semantic analysis in an attempt to fingerprint writing
> styles. It's by no means perfect, but it is possible to
> detect if a piece of writing looks nothing like a
> student's normal writing style.
Although this process is not automated.
Most teachers are actually not bad at recognizing non-original work, provided they have a reasonable problem space. If you are teaching two classes of 25 students, odds are pretty good that you'd recognize rather quickly when something is not original if your class has a decent set of deliverables.
Of course, most colleges and universities have many more than 25 undergrads in a section, and it's also typical for non-professor staff to be the ones that grade most of the work. So the capability of fingerprinting just isn't there. This is, unfortunately, not solvable without changing the way the class is taught, which (in many institutions) is a problem with the greater educational system preventing the actual audit countermeasure from working.
It's just impossible to audit creative work unless the teacher has a pretty implicit grasp of the student.
One thing that is being missed here is that outside of Academia and the fields of professional writing there is nothing immoral about plagiarism. In every day life there is nothing immoral about quoting songs, movies, books, or anything else without attribution. It is called culture, we live in it every day.
So, yes, these kids are cheating, plagiarism in an academic setting is bad. But in general, that isn't true.
I have been testing plagiarism detection systems since 2004 - most find something, but generally not all plagiarisms. In my current test, the best only pick up on 70% of the plagiarisms.
Ghostwritten papers are not found. Period. Semantic analysis - or at least the systems I have tested - does not work.
"But as Sir Issac Newton pointed out with his "standing on the shoulders of giants" a work can be both derivative and original."
Actually, Newton stole this quote from someone else, it has been traced to Bernard of Chartre, around 1130: "Pigmaei gigantum humeris
impositi plusquam ipsi
It is just not possible in our oversized classes to to real teaching with people disinterested in learning.
Sounds a lot like marketing, advertising, or EnRon and mortgage executive training to me - -
With regard to Newton's comment about the "shoulders of giant's" there is quite a bit more behind it than many people realise.
First of it may well have been plagiarism and the reason for this is Newton was actually being quite rude about a certain third party.
Newton wanted unfettered access to another persons research that was a work in progress and they quite rightl told Newton to go away and wait, and if he could not wait to do his own research. Newton took umbridge and tried to get others to put preasure on the person.
So the comment was almost certainly ment in at best a sarcastic way and was by no means ment as an "original thought" (a bit like calling somebody "a real winner").
What amazes me is the way quite a few historians know this yet "natural scientists" still think it was Newton showing humility to others...
As once remarked in the UK about football "it's a funny old game".
Read nearly all comments. Think it's overall a typical engineers aproch to the problem of processing thousends of persons from cattelwagon to crematorium (no discussion board should do without Goodwin's law).
Plagiarism in course and final paper is no problem, more some sort of solution. Esseys are not suited for assesing the knowledge level of a pupil.
Readily available text souces make grades on writing assingnements useless.
Get over it.
A bad essey is a probability to learn and not something to harass a student over. Stop grading everything and plagiarism in college papers will disaper.
Bravo! I marvel at how you inserted at least 10 errors in that tiny, tiny piece of text to prove your point about writing!
You do make a very strong case. It is, however, the opposite of the case you were attempting to construct.
"Fundamentally, this is a problem of misplaced economic incentives."
You forget the universities also have a problem of misplaced incentives. With sky-high tuitions, Universities have a clear economic incentive *not* to police themselves and catch cheaters. Academic integrity is mostly a buzz-word these days.
Schneier.com is a personal website. Opinions expressed are not necessarily those of BT.