Cheating on Quizzes? My students wouldn't do that, right?

It recently came to our attention in a (face-to-face) class that two students were collaborating on online quizzes.  It came to light because a student claiming to have taken a quiz called and mentioned taking it about the same time as his study buddy.  Whatever happened, that particular quiz was not accessed by the student, but looking at other quizzes, the two students had increasingly similar results on their quizzes, and typically their access times overlapped, sometimes turning in the quiz at the same time.  On one occasion, they accessed the quiz from the same IP address (meaning they were likely at the same location).

So, out of curiosity, we looked at other students in the class and found five other examples where it appears the students may be collaborating.  This is a face-to-face class with supplements and testing online.

So, this brings up some questions.  

• Are students cheating in your class?  
• How can you tell?  
• And, how can you minimize the likelihood of cheating in online quizzes and tests.

Well, are students cheating in your class?  From Jared Stein's PowerPoint presentation on cheating¹, he cites figures that 75% of students admit to cheating and 95% of cheating is undetected.^2,3

I'll leave detection for another post.

Does this mean you have to avoid online quizzes?  They're very convenient and in some cases necessary, so how can you discourage cheating on your quizzes and tests?

• Make clear statements about what is permissible, and what is not.  Our syllabus boilerplate refers to academic honesty and the academic honesty web page specifically mentions collaboration on quizzes and tests.  Still, it would be helpful to make it completely clear by posting your policy in your syllabus addendum, and perhaps sending a message or posting an announcement at the beginning of the class.
• Have students sign an academic honesty statement, and/or provide a quiz that they take that has them indicate they know what is permitted and what is not.
• Make your quizzes more cheat-proof.  There are a number of approaches to this.  Here are some of them.
• Deliver a random subset of a larger pool of questions.  In the incident mentioned above, the questions and answers were randomized, but they were the same set for all.  The students learned quickly enough they could find the same question even though they were in different order.  If you randomly delivered 50 questions out of 100 (randomized subset), students would get substantially different quizzes, making this kind of collaboration more difficult.
• Include some essay questions.  Yes, it makes for more work for you, but it also means the student has to do his own work.  You could also use the randomized subset technique for these too, so they couldn't discuss their answers.
• In the same vein, move away from objective questions and assessments and move towards subjective responses.
• Time limit your quizzes so that they can be completed, but without a lot of extra time which can be spent coordinating the answers.  
• You may wish to allow open-book quizzes, since it is virtually impossible to stop in un-proctored quizzes
• Employ the 'nuclear' option of proctoring quizzes.  Tegrity provides a proctoring function to record the student's work on quizzes, and other options are available.  Note that in-person proctoring of distance students in other states may trigger rules that require CBC to register and in some cases pay exorbitant fees, and should be avoided.  There are on-line proctoring services that could be used if there is interest. Contact eLearning for more details (or maybe I can do a blog posting on this later).
• If there is interest, we could look into licensing Respondus Lock-Down Browser.  This won't prevent students from collaborating, but could prevent them from viewing other sites while taking a quiz.  This would work best in a proctored setting, though, because there's nothing to stop a student from having another computer next to the locked down one.
• Apply other strategies that minimize the impact of quizzes, such as a variety of assessments, including essays and papers or other original work.
• Require that students share drafts and work in progress.
• Make more quizzes formative, with lower point values (along with some of the other suggestions) in conjunction with other assessment strategies.
 

You can evaluate your course using the Measure of Course Cheatability, which includes many of these items and more with annotations (note that getting 100% is the 'holy grail' and Jared Stein, who co-developed this rubric, says he hasn't seen anyone get better than 80%.⁴  There is a Collaborate recording from the eLearning in WA group on Academic Integrity with Stephanie Delaney, Renee Carney and Connie Broughton.  Unfortunately the sound goes out multiple times during Stephanie's part, which is annoying and hampers the flow.  Nonetheless, there's good info there. 

Added note (12/17): turnitin, the cheating detecting system, has published a white paper on the top 10 types of plagiarism.  This also linked from an article in eCampus News, if you're interested in a summary.    

¹From PowerPoint notes in Jared Stein's presentation on cheating. He has a lot of references on the linked page and more links on his diigo site.
²Dick, M., Sheard, J., Bareiss, C., Carter, J., Joyce, D., Harding, T., & Laxer, C. (2003, June). Addressing student cheating: definitions and solutions. ACM SIGCSE Bulletin, 35(2), 172-184.³Kaczmarczyk, L. (2001). Accreditation and student assessment in distance education: Why we all need to pay attention. Proc. 6th Conf. on Innovation and Technology in Computer Science Education, Canterbury, UK, 113-116. 
⁴ Ibid. (Stein)