02/01/2014 01:01 EST | Updated 04/02/2014 05:59 EDT

Attack Plagiarism but Defend Student Creativity

Hours before my wife's water broke -- our son was born in 2005 -- I learned of a case of mosaic plagiarism in my class. My nerves twanged. When the water broke, it poured; stress over a potential plagiarism confrontation whilst careening in a new mini-van toward the hospital can be eerie. Lawyers? Ugh.

Mom and baby were fine. Not so much Dad in the week that ensued.

Rates of plagiarism among University students vary greatly depending on study design, but it's safe to say that a majority of undergraduates report that they have used others' words, ideas, or creative work without suitable acknowledgement.

In 2005, I took a low-tech approach to detecting plagiarism: inserting sentences into Web search engines when the grammar struck me as oddly placed. I then tried to sniff out plagiarism. I was old school in the ways of plagiarism identification, since I knew too well the limitations of popular software detection tools that excel at pinpointing stark plagiarism, not the far more common mosaic variety.

This appeared to me back then as mosaic plagiarism -- content stitched together from the web. Another challenge: it was a group project. Knowing which student was responsible and that the culpability was not intentional as I originally presumed, was reminiscent of childhood. I recall thinking about sitting in a playroom with my mother to hear which of my elder brothers would squeal on the other about whom had fiddled with my toy cars first.

The young woman accused of plagiarism was sobbing in the tiny conference room with too many corners, and I am claustrophobic in cornered spots. I was nauseated.

Many years later, now ensconced in the Internet software world, I have learned just how skilled young students are in the ways of mosaic plagiarism. I have learned the importance of attacking mosaic plagiarism full force. I believe in the rule of law and the science of intellectual property law; one's words are one's own and are not to be stolen by others wittingly.

Yet encouraging creativity in young minds should be paramount for any teacher.

What is more problematic than mosaic plagiarism (in social sciences this can be mitigated somewhat through appropriate grade allocation, including the use of presentations and participation grades and the submission of early drafts of papers and assignments) is the solution proffered by some in the well-intentioned professoriate: rarified fact-based essays about obscure topics that cannot be researched on the Internet. There is a reason for this: these topics are idiotic.

Instead of asking students to opine on the theory of land and identity ("A man without land is nothing") in the late Mordecai Richler's The Apprenticeship of Duddy Kravitz, it might be fashionable now for an English professor to suggest, say, a topic about the list of physical tools (pliers?) that Duddy possessed to live on a deserted island. Hey, if you have read the book, surely you can recount the nuts and bolts?

I suggest to these professors that they ask themselves if the writers or historians or philosophers under scrutiny would have approved any such assignment. If you kill creativity, you kill the University. A student without expression is not a student.

The common approach of plagiarism-detection software relies on natural language processing -- in large databases -- to help sift through whether or not there is something that seems too akin to someone else's paper inside that database. With the inevitable proliferation of online education, led by Coursera and MOOCs (or massive open online courses) at colleges and Universities, a greater proportion of essays submitted online could make plagiarism far more of a problem.

The solution lies not in analyzing reams of papers through pricey text analytics software, or through Big Data software -- searching for inconsistencies in grammar and argumentation -- but rather in simple, random surveys. Nuance and satire and mosaic plagiarism cannot be easily detected in all but the most extreme cases, and the best plagiarists (many bent on getting good enough grades so they can attend graduate or professional school) can easily cheat Big Data algorithms.

Surveys to detect plagiarism? The trick is not to ask a student whether he or she has plagiarized a paper or assignment, but rather to trend and track cheating norms at the University "unit level" through anonymous short surveys about whether students report plagiarism is rising or falling in different fields of study.

Consider an anonymous online short survey that asks: "Has a roommate or friend recently committed plagiarism at your University?" And then: "What field are they in?" "Which class?" Students know who's cheating and where.

A class with higher rates of plagiarism should suffer a grade bell curve shift leftward; your student colleagues, the real victims of your lack of creativity, will then snitch -- with glee. Incentives, as all students must one day learn, work.