Archives for August 2014

Yes, that’s right. Turnitin, the top originality checking software, is now available for use by all S&T instructors. Turnitin has been integrated into Blackboard, and I have to say, it’s a very clean integration. There are no extra accounts to manage, and no extra passwords to remember. As an instructor, you simply set up a Turnitin Assignment, students submit work to that assignment item, and you may then view Originality Reports based on those submissions via the Blackboard Grade Center. Easy-peasy!

Naturally, being a writing instructor myself, I wanted to test Turnitin, to run it through the gamut of possible tricks, just to see if I could in fact slip a plagiarized paper past Turnitin. So far, I’ve had very limited success on that front. Turnitin is a very capable program, and the developers are well aware of the popular tricks which less-than-honest students may use to try to fool the program. So, in that spirit, I copy-pasted together a short test paper on “Academic Dishonesty”, and then used every trick I knew of or could find on the darker corners of the internet so that you could read all about it..

Some popular strategies that Turnitin easily detects are:

• The “mix it up real good” hack – Taking material verbatim from several disparate sources and “remixing” the content to create something new. Rearranging sentences, adding transitions, substituting words, none of this will truly help the quick-and-dirty plagiarist. No matter how well the content is remixed, the Turnitin algorithm immediately flags suspiciously similar content in the originality report.

• The Google Translate hack – Running plagiarized work through the Google Translate function, for example translating from English to Spanish and back to English. Turnitin does a very good job of recognizing this type of altered content.

• The Cyrillic character hack – Using Cyrillic characters which look like English language characters in order to trick the word-recognition algorithm. (The Cyrillic letters “A, Ie, Dze“, look like the English letters “a, e, s.”). Turnitin recognizes this trick immediately, and generates the same originality report as if the Cyrillic characters were not even there.

• The DOC to IMG to PDF hack – Changing a Word document to an image file, and then changing that image file to a PDF. The idea behind this trick is to create what appears to be a legitimate submission, but one that cannot be effectively scanned. This trick does not work. Turnitin will not accept a submission that does not contain a machine-readable text layer.

Strategies that “sort of” work:

• The MS Word Macro hack – Employing a macro-enabled Word document to automatically change certain characters back to a “normal” format upon opening the file (e.g. automatically changing “~a” back to “a” as the file opens in Word). The idea behind this trick is to have a difference between the machine-readable state and the human-readable state of the file. This trick will result in a similarity match of 0% due to the unique nature of the extra characters in the individual words. However, because Turnitin will not allow macros to run, the instructor will see the original file, complete with wh~atever form~atting tricks ~an ~author h~as ~attempted to use to hide their ~ac~ademic misdeeds.

• The whited-out filler character hack – Replacing all spaces in the document with a character (such as “#”) that has been set to the color white. This is another attempt at the old switcheroo, where the document appears normal but becomes more difficult for the algorithm to “read” the underlying words. This hack sort of works, in that it will reduce the amount of similarity that Turnitin is able to detect, but again, an aware instructor will likely notice some oddball spacing in the document. Additionally, Turnitin has a “text only” viewing mode that will strip out any formatting, thus#exposing#this#type#of#hack.

• The PDF layer hack – This hack is a sort of variation on the MS Word Macro hack. The PDF layer hack uses a custom character map to spoof the machine-readable text layer of a PDF document with text that appears unique due to the actual text layer being a random hash of standard and non-standard characters. The image layer remains readable and appears normal. This technique will also return a similarity match of 0%, but is easily defeated by an aware instructor. Additionally, because the underlying text layer is different, attempts to print or highlight and copy a block of text from the PDF will expose the subterfuge.

• The synonym substitution hack – Replacing every word that is possible to replace with a synonym. There are sites which will automatically do this for you, as you supply a block of text (e.g. http://www.outsmart.it). This trick will return a similarity match of 0%, but is again easily detected by an aware instructor. Automatic synonym replacement utilities do not and cannot take into account language variables such as syntax, grammatical case, tense, and number when choosing synonyms. This leads to a characteristic type of oddly stilted prose with many unusual constructions and mistakes that native speakers are highly unlikely to make.

Strategies that absolutely work:

• The “actual work” hack – This one is the easiest, or the hardest, depending on your outlook towards the writing process. One sure-fire way to “beat” Turnitin is by reading from many sources, maintaining a bibliography of all sources read, using quotations properly, paraphrasing complex ideas from these sources with appropriate citations, and adding to them with newly synthesized ideas/observations of your own. This is a great method to defeat any type of originality checking software, and one of the oldest methods around…but it’s a lot of work, and it’s the method of last resort for the truly incorrigible!

Of course that last “strategy” was written with tongue firmly in cheek. Ultimately, Turnitin is a very capable program, but it is only a tool. Instructors must, as always, be aware and use their best discretion; the effectiveness of several of the above-mentioned tricks relies almost wholly on the instructor not closely reading the student paper, and instead just blindly trusting the originality report. An important thing to keep in mind about Turnitin is that just because something is flagged as being similar doesn’t mean it has been plagiarized. Proper citations and quotations will always be flagged as similar. On the other hand, just because something is 0% similar doesn’t mean it is original. For example, a 0% similarity match means, at the very least, that the paper lacks proper citations and includes no bibliography or works cited page. A 0% match should be a huge red flag for instructors in any case.

Questions about using Turnitin? Contact Educational Technology! We’re here to help YOU.