NYC is Paying $2 Million For Anti-Plagiarism Software After Firing Teachers (vice.com) 44
An anonymous reader shares a report: Earlier this month, more than 1,000 educators and students at City University of New York institutions petitioned their board of trustees to not renew its contract with the anti-plagiarism software company Turnitin. The board ultimately voted unanimously, with the student senate representative abstaining, to renew Turnitin's five-year contract for nearly $2 million. Five months earlier, CUNY had laid off nearly 3,000 adjunct faculty and part-time employees as a result of budget shortfalls. (The college system's chancellor has pushed back against that characterization). The protest against Turnitin is the latest high-profile effort in what has become a nationwide backlash in higher education against educational technology vendors. As schools moved online during the pandemic and confronted slimming budgets, they increasingly turned to a wide array of software companies for solutions. The ed tech industry has boomed, and the school experience has been transformed in ways that are sure to outlive the pandemic -- not necessarily for the better, many experts say.
So outsource them (Score:1)
Don't tell our teachers are better. I don't believe that for a second.
Re:So outsource them (Score:4, Funny)
Re: (Score:2)
"My teacher almost speaks English."
Yeah, I had an American teacher too once. You have my sympathy.
Re: (Score:3)
You don't need to outsource teaching to India to save money.
The way to save money is to use pre-recorded mass-produced lectures.
Instead of having a thousand teachers give more-or-less the same lecture, you have the best teacher record the lecture and then show it to a thousand classes.
When the schools shut down in March, my son's math teacher did this. Rather than record each class, the teacher gave them a list of Khan Academy videos to watch and a list of homework problems to complete after each one.
So th
Re: (Score:2)
"Instead of having a thousand teachers give more-or-less the same lecture, you have the best teacher record the lecture and then show it to a thousand classes."
Why would you even need a teacher doing the lecture? That wouldn't be the best out of your money. Contents, maybe, are the realm of a good teacher, but doing the act? Let a Di Caprio or a Zellweger staging the lecture, that, of course, would be so much better!
No. The act of teaching is not "giving a lecture".
Math is hard. (Score:5, Insightful)
$2m for a five-year contract, or $400,000/year. How many of those 3,000 faculty members could have stayed on for that kind of money? Four or five, in other words less than 0.2% of them. A few million dollars for a service contract is nothing compared to payroll. While I'm sure there is a very real emotional and human component to decisions that appear to replace live human beings with bits of dubious software, the facts shown on a balance sheet are that the two events aren't really related, or at least not interchangeable.
Re: (Score:1, Informative)
Adjuncts are typically paid about $3000 per course, far lower than regular professors. $400,000 per year could have kept 150+ of them employed at 10 courses / year (which a a very full load). It's also not a zero sum game as putting on those courses likely means more enrollment and more tuition to allow even more to be hired.
The adjuncts have a positive impact on the University and the students they teach. For students who do not cheat, Turnitin causes additional anxiety and stress. For those who do cheat
Re: Math is hard. (Score:1, Informative)
Yes, math is hard and it seems that you fail at it.
You claim $3,000 per course for an adjunct. Then somehow claim that $400,000 would pay for 150 adjuncts with a 10 course load.
Hmm. Let's go with 1 course for 150... $3,000 * 150 = $450,000 which blows well past the $400,000. And multiply by 10 courses instead of 1.
Yep, math is hard and you need to go back to school and learn it.
Re: (Score:1)
Your math is very off. First off, a good rule of thumb is that the cost of having an employee on the payroll is roughly twice their actually pay. At ten courses per year, for $3000 a course, that means each adjunct costs $60,000 a year. $400,000 divided by $60,000 equals 6 and two-thirds. 150? Try less than seven.
Re: (Score:2)
How many of those 3,000 faculty members could have stayed on for that kind of money? Four or five, in other words less than 0.2% of them.
You speak as if it has been established that Turnitin software has completely replaced 3,000 faculty members. It didn't.
I do have a lot of fun watching students try to defeat Turnitin-like software by submitting differently cropped images, different format/font/colored files or otherwise tweaked answers. It does not help against human plagiarism detection though.
Re: (Score:2)
If the students have to work hard to defeat the anti-plagiarism software, at some point they will decide that it is easier to just do original work instead.
Re: (Score:3)
If these adjuncts and part time workers are paid minimum, and teach 4 classes per semester, we are looked at more than a few, but less than 20 or 30 that could be paid with this money.
Beyond this the software is a good value. Not all students are
Re: (Score:2)
>The software did not exist to help her learn not to
>plagiarize when she writing for Nixon.
On the bright side, it makes her eminently qualified to work for Mr. Biden . . . :)
returning to topic . . . in the early aughts, penn state sent us emails about this wonderful new service for us, that would help with plagiarism.
I tried it out on term papers.
I was indeed impressed: not only a false positive rate of 100%, but *also* a false negative rate of 100%.
Yes, *every single* thing it flagged was fine. Sim
Re: (Score:2)
Re: (Score:2)
Less than 20.
But, yes, I *assume* they have improved.
What they considered a product could not *possibly* get any worse, save perhaps by adding new classes false positives to waste faculty time . . .
Re: Math is hard. (Score:2)
It's vice, what did you expect?
Seems like a good trade-off (Score:5, Informative)
$2 million dollars for a five-year contract is $400,000 per year. If the burdened cost of one of those teachers is $80,000 per year -- which could well be an underestimate, considering that the average salary within CUNY is somewhere north of $64,000 -- then this software costs the same as five teachers. In hour equivalents, that is 10,000 hours of labor per year, and probably less.
The CUNY system consists of 26 institutions, and serves 275,000+ students. The cost of the software works out to roughly 130 seconds of teacher time per student, per year. If the software saves more than that -- across all of the courses a student takes -- then it wins the cost-benefit tradeoff.
Re:Seems like a good trade-off (Score:4, Informative)
What value does Turnitin actually bring?
As a teacher when I was forced to use it, it meant another bureacratic headache and more time assisting students to solve upload and account problems. It never caught anyone cheating which wasn't instantly clear to me upon reading their paper. To me the value was zero.
For students who do not cheat, Turnitin causes additional anxiety and stress. I've been on the receiving end up trying to help my students manage that stress and it sucks. For those who do cheat, its just means that cheating costs a little more by having to pay someone to ghost write, rather than just paying for a pre-written paper. Again I think for students the value is either zero or negative.
Administrators love Turnitin because it creates graphs and numbers which they'll then fashion into reports to justify their enormous salaries. In reality its another expensive case of garbage data in, unreliable analysis out.
There are real pathways to reduce academic dishonesty. An easily fooled piece of very expensive pattern matching software is not it. This is $2 million that would be better spent hiring more teachers and tutors to help students learn.
Re: (Score:2)
Re: (Score:2)
Teaching and testing does not need a uniform standard, it need individual tests.
The main problem with teaching in the west is exactly your attitude.
It makes "people who can pass a test" and not people who have learned anything.
Re: (Score:3)
It makes "people who can pass a test" and not people who have learned anything.
If you pass the test, there may still be gaps in your knowledge.
If you can't pass the test, then you didn't even learn the fundamentals.
Tests don't identify the brilliant, but they do filter out the clearly unqualified.
Re: (Score:2)
If you can't pass the test, then you didn't even learn the fundamentals. ... it only shows you are bad in tests.
Wrong
Re: (Score:3)
The purpose of the software is not to catch cheaters but to deter students from trying to cheat in the first place.
So just because you didn't see it catching anyone, doesn't mean it wasn't working.
Re: (Score:2)
The purpose of the software is not to catch cheaters but to deter students from trying to cheat in the first place.
How do you arrive at that conclusion? I don't see anything about their software deterring cheating on Turnitin's website. [turnitin.com] AFAICT, its purpose is all about identifying plagiarism after the fact.
So just because you didn't see it catching anyone, doesn't mean it wasn't working.
qparadox didn't say the software didn't catch anyone. He said the software didn't catch anyone he wouldn't have caught himself. Besides, if the software isn't catching anyone plagiarizing, either no one is plagiarizing (and the software is unneeded) or the software doesn't work.
Re: (Score:2)
It's almost impossible to be sure, since the absence of something can be due to several (un)related things.
LOL, impossible? There's nothing absent here, unless you believe ShanghaiBill arrived at his conclusion with no reasoning behind it. If he answers my question, I can be reasonably sure of what lead him to his conclusion.
Now, with that out of the way, an indirect effect of such a system is that some who would otherwise have cheated, or tried to, will not even make the attempt. Hence, being deterred.
These plagiarism detection systems have existed for years, and students are quite aware of them. Yet plagiarism still exists in shockingly high numbers, so I question the deterrent effect of such systems. I would imagine that by now we'd have some data along the lines of "In the three ye
Re: (Score:2)
I was the lead Teaching Assistant for a few semesters for a senior level course with roughly 650 students in it each semester. With each student writing five essays during the semester, most of the TAs had to thoughtfully grade nearly 300 essays every semester. While it was literally their job to do so and most of them were good about it, some weren’t, and one of the best ways to identify them and take corrective steps was by using Turnitin.
In one case, we had a TA who apparently didn’t even rea
Re: (Score:1)
You and above are missing 'adjunct' and 'part time' in TFS. Adjuncts who make a living at it typically have at most a few classes from several schools. If they get benefits it's usually in the form of a COBRA contribution. That being said, there are a large number of line items in the CUNY budget and juxtaposing these two is a little disingenuous.
2 separate issues (Score:2)
Regarding Turnitin, it's an INCREDIBLY USEFUL tool for teachers nowadays. Worth 2 million for the NYC school system for sure. It's total
Re: 2 separate issues (Score:2)
Bollocks.
Such as turnitin encourages teachers to disengage from their students. Meaning, as a teacher you have closer to zero idea of their learning.
Re: 2 separate issues (Score:4, Insightful)
The time when a teacher can spend 45 minutes on each paper they need to grade is 50 years gone. Your vision of an idylic countryside classroom where a teacher hovers over 6 students is far out of date. There aren't enough hours in the day to do it the old fashioned way. Computer tools are necessary.
Re: (Score:2)
If teachers stopped plagiarizing assignment and tests, then their students would, too.
How does one detect plagiarism without software? (Score:2)
It's not practical to rely on human memory.
Students cheat because they attend school to obtain credentials for employment. There is every economic reason to cheat and that pressure pisses all over any residual idealism.
Teachers required automated help. (Eventually AI may replace them as teaching becomes too demanding for humans to manage at scale.) It's not 1960 any more.
My opinion (Score:1)
Antiplagiarism software (Score:1)