What is the accuracy of "plagiarism programs"?

[Shananigans]

A few days ago a friend of mine called me up and told me that he was about to be accused of plagiarism. Apparently his TA ran his paper through a computer program that checks for plagiarism and it came up that he had plagiarized.

I know my friend and I'm certain that if he had plagiarized he'd admit it and accept the consequences. The crazy thing is that my friend showed me the part of the story that he supposedly plagiarized and it is a PERSONAL story.

I am just wondering if anyone else has had an experience with being accused of plagiarism or having their paper run through these programs. In my previous way of thinking, I thought these programs were pretty accurate, but now I am wondering.

What has your friend's/sister's/brother's/dad's/uncle's/
mom's/cousin's experience been?

[Shananigans]

1. Doesn't anybody love me? Doesn't anybody care? :(

[Snow]

2. Any scientific paper publishing house will use them,

and they typically give you a reference with the questionable passages highlighted for comparison with the paper in question. There will be cautions such as "both authors may have quoted from a third author". But you get a look at the original & the supposed plagiarism side-by-side & can make a judgement. Usually what journals will do is send the first author a letter saying something like 'ooops, we're sure you made a mistake here, but it seems what you say in paragraph so-and-so is just like what Albert Einstein said in his 1937 paper from Physics Today. Here's the printout so you can see the comparison. We ask you either rephrase or put quotes and properly attribute Prof. Einstein." The TA is probably young and inclined to be heavy-handed athoritarian. His right, I suppose, but not the best way to handle it IMHO.

[Shananigans]

3. What if it is for a class, though...

At this college, you face expulsion for plagarism.

Even if the program does find a similar story, who's to say that two people didn't have similar experiences?

[Snow]

4. Not a clue - that depends on the class & school, I guess.

I always prefer a lighter-handed approach myself. I teach graduate level small classes, so I give 3-4 take-home open book exams per semester. One time a couple of the exams had answers that were remarkably similar. I wrote a comment on each something to the effect of "your answers look a lot like another students. while collaboration is a key to successful scientific investigation, I am tring to see what your level of understanding is currently in order to teach you better, and if you collaborate on the exam I am hard-pressed to evaluate you both independently." And it didn't happen again. But there are places with these very strict, 'zero-tolerance' policies in effect. To my way of thinking, if they're proposing extreme penalties, then the burden of proof should be on them, with the understanding that a less-than-solid case will be followed by a lawsuit. Lotta places have gotten really paranoid about the term paper sites on the web.