Adventures of a historian-in-training, comments about history and how it is done, and anything else related to history. This means essentially everything, for everything is history as soon as it slips into the past.

Tuesday, December 4, 2012

Is plagiarism-detecting software reliable?

I've had an unpleasant experience with my paper about the suspension of Al Cahill.  The professor has required us to submit our papers to a plagiarism-detecting software used by the university.  I am not going to name the particular software.  And I'm not naming it because I think it's useless.

It said it found a 3% match between my paper and the sorts of sources it checks on the web.  Four of the six that it "found" had absolutely nothing whatsoever to do with the subject of my paper, and in fact had to do with things that have occurred in the second decade of the twenty-first century, not the 1950s.  Two of them did have something to do with a part of my paper.  One of them was a source I was unaware of; it never showed up in my searches.  The other was tagged as being from a newspaper in a Florida city other than the ones in which the central events of the paper took place, that is to say, Jacksonville and the state capital, Tallahassee.  I got the particular source from another site, that of a well-known historical society. 

If this is the best the software can do -- and the professor told me that this was its function -- it does not seem to me to be very useful in detecting actual plagiarism.  The software keyed in on single words, not phrases or paragraphs.  Using a single word that might have appeared somewhere in another source does not constitute plagiarism.  It's called "vocabulary."

Certainly I can see the thinking behind the use of these programs.  But in the old days, before we became so dependent on computers and the internet for everything, a teacher knew the capabilities of his or her students, and could pretty well tell when a paper was written in a style and manner above those capabilities, and would call the student in for The Talk.  And the student would get an F.  Well, the teacher OUGHT to have known the capabilities of his or her students, but sometimes it did not work that way.

On his senior paper in high school, my brother was accused of plagiarism and given an F with no appeal.  The teacher would not budge, even when our mother talked to the teacher  The teacher did not bother to attempt to verify whether the charge was justified.  She was right, and that was that.  She just "knew" my brother "couldn't" have written such a paper.  It was on the Yellow Fever epidemic in Jacksonville over a hundred years ago now.  What the teacher did not know -- and did not bother to find out -- was that our aunt was at the time Director of Health Information for the State of Florida, and had gotten my brother all sorts of publications.  She read the paper, our mother read the paper, and I read the paper.  We looked at the sources.  He footnoted everything he should have footnoted.  Where he did not use footnoted quotations, he used his own words.  He did not plagiarize.

Should twenty-first-century plagiarism-detecting software protect the student from lazy teachers as well as allow a professor to detect cribbed papers?  Perhaps, but it should do a better job of detecting than focus in on one word -- such as "waiver," which I used in the sense of a waiver of immunity from prosecution -- and return a result of a story from a few days ago (fifty-four years after the events about which I wrote) concerning the President of the United States.  I told my professor I was surprised the software did not return references to a bunch of sports articles!

The other thing that makes me very angry is that I was forced -- by the use of this software being required -- to place my paper out there on someone else's server forever.  This is my work, and I like to control my work completely until it is published.

Before I will think well of these sorts of programs or services again, they need to make some pretty massive improvements.
.

No comments:

Post a Comment