You might remember that last summer I taught the very first fully-online course my department had ever offered, the standard Calculus 1 course but fully online and asynchronous (i.e. we never met as a class in the standard sense). Since my last blog post, that course was approved to run again next summer, and I’ll be teaching it again. Although I have a lot of things to do between now and then, I’m already starting to think about what I should do differently when it kicks off in May 2016.

One of those items to re-think is academic integrity in the course. I don’t think there was ever a problem with cheating or collaborating in my course last summer – but then again, how would I know? Maybe the students are just really good at circumventing security. And to be honest, I went into the course design process with the security issues of online education pretty clear in my mind, but I also wanted to have a learning community in the class predicated on trust rather than constant policing. So it was entirely possible for students to be academically dishonest in the course. My thought was: I would rather get burned on academic dishonesty than try to set up and enforce a police state.

But I’m coming to understand that in online teaching, there is middle ground: You can set up and enforce simple measures that promote academic integrity that do not transform the class into an Orwellian landscape. The greatest density of aha-moments on this score that I’ve experienced came this weekend when I read through this collection of readings from Faculty Focus on academic integrity in online learning, shared with me by my friend and colleague Matt Roberts. I encourage you to read this too, even if you aren’t teaching online.

Here are some of the bigger take-aways I got from this collection:

• It seems like to focus most of our attention regarding academic integrity on policing – catching and punishing cheaters – but in reality it’s a three-pronged effort: policing, yes, but also prevention and promoting ethics. On the policing side we still want to be vigilant and not take academic honesty for granted, but have reliable systems that alert us to academic dishonesty as well as systems for punishing students who violate the rules. But also, we want to have systems in place that will help prevent cheating from happening in the first place – for example changes to test design that disincentivize cheating, continuing education in the class about what academically honest and dishonest work looks like, and simply being clear and explicit about what academic integrity is and what behaviors violate it in our classes. And this ties into ethics – the gold standard would be to create a class of students who simply don’t want to cheat. We want a learning community that believes in fair play, real learning, and ethical behavior. It seems to me that a truly effective policy for academic integrity must draw significantly from all three areas.
• It was clear to me from multiple parts of this report that there is a correlation between academic work in a course that promotes deep learning and academic work in a course that promotes effective security. That is, assignments that engage students in authentic work at a high level are by and large the very assignments that are hardest to cheat on, or at least the easiest to detect if there is cheating. For example giving a student a project in which they have to collect their own data, do a regression analysis on it, and then perform calculus opertations on it and interpret the results in context is an assignment that, while not “cheat-proof”, is a lot harder to cheat on than a five-question quiz in which students take the derivatives of five functions, because the student owns that project whereas the student does not have ownership over quiz questions. The degree of ownership leads to deeper learning and greater security.
• A lot was made in this report about technological methods for providing security, for example the use of remote proctoring services (such as the services provided by Kryterion). If you’re unfamiliar, remote proctoring typically involves having the student purchase a device that plugs into their USB port and includes a webcam and microphone that is turned on while the student takes an exam. While the student is taking the exam, the video and audio is beamed back to the HQ of the company, where a team of human proctors is watching the student (and presumably more, like I’m imagining each proctor looking at a 5x5 grid of monitors with a unique student on each) take the exam in real time. If the student does something fishy – like stare at a patch of space behind the computer, or put his hands in his pockets – the proctors can call the student out and ask them to pan the camera around the room, or take their hands out of their pockets, or whatever. I don’t doubt that this setup provides a great deal of security to the testing environment, and it’s probably the closest thing we have right now to a live, proctored in-seat exam setting. Some online education providers use this system and swear by it. But for me? And for everyone else in the committee meeting where we discussed this report? Definitely not our style and definitely not the sort of community I want to have in my classes. I don’t think surveillance is really the answer to academic integrity, again at least for me.
• Also frequently mentioned in the readings was the use of biometrics like keystroke analytics to provide security measures. Again this seems overly surveillance-oriented and I’m not sure how effective keystroke logging would be in a math class anyway. However, there’s something to be said for simple and noninvasive biometrics such as having handwriting samples from students to compare against in later assignments. I wrote a while back at Wired Campus about student-created video content in the calculus class, for instance. I used student videos purely for pedagogical reasons, but the more I read from this report the more I realized that it provides pretty effective security as well because it has the student’s face, voice, and handwriting all in the same frame. So if a student turns in work later that’s handwritten, and it looks nothing like the handwriting in their video, that’s cause for concern. So next summer I think I am going to greatly expand the use of student-created video to provide a larger corpus of examples for students and of student biometric data for me.
• Another effective practice is assessing students in a variety of ways. Again this is just good assessment practice generally speaking. But also, one of the papers in the report made the point that when you have a variety of assessments – quizzes and papers and projects and programming assignments for instance rather than just homework and tests – then it simply becomes impractical for students to cheat by, say, paying someone to impersonate them. It’s too much work for the impersonator! And it’s too much money for the student to pay for an impersonator who’s willing to do the work.

Again the whole report is really interesting reading and I encourage you to check it out. Bruce Schneier reminds us constantly that there is no truly secure system anywhere and that’s certainly true in online teaching, but I think too that there are simple measures that we can take that promote integrity and deeper learning and better community, all at the same time.