Yes, my course was photographed at Harvard.

We have seen a slow drip-drip-drip in the press about the surreptitious use of digital photographs to measure lecture attendance at Harvard.    Given that my class was one of the ones where attendance was counted by this method, I think that gives me some prerogative to comment on the episode.   

First of all, this was SPU 26, a General Education offering in the Spring of 2014.   I will shoulder some of the responsibility as I was aware of the use of the camera from the very first lecture.   Rather than play a guessing game about whose lectures were photographed, I will simply “out” myself, so that my students from the Spring Semester will know.  

It’s worth going over my own timeline.   Before the start of the first lecture, I discovered a small video camera, the kind that’s used for security purposes, perched on a ledge above the chalkboard and pointed toward the seats in the lecture hall.   A cable from it was dangling loose, as if someone had intended for it to be plugged in.   Not knowing what it was intended for, I went into the lecture prep room and alerted the media folks who video-tape lectures that a camera was unplugged.  

It turned out that it wasn’t the Media Services people who owned the camera.   A woman came by and thanked me for sounding the alert that the cam was unplugged.  She then told me what it was for.  There was a study being done about lecture attendance.   The digital images from the camera were not to be viewed by humans, but rather a computer algorithm would distinguish bodies from backpacks or coats and do an anonymous count of how many students were in seats as a function of time both during a given lecture and over the course of the semester.   The students were not supposed to know about it, and it was something of an accident that I discovered it.  In addition, she told me that other lecture halls had spy-cams in them as part of a relatively large-scale study to gather data.  

At that moment, I didn’t have much time to contemplate any issues surrounding the anonymous lecture census, as I had to jump right into my first lecture.    Over the course of the semester, I would occasionally mull over questions surrounding its use.  On one hand, I convinced myself that it was morally equivalent to a traffic survey where a municipality puts a little counting strip across a road to count cars.   On the other hand, if the digital images were on a hard drive somewhere, could it be hacked into?   I even imagined a situation where, God forbid, a shooter opened up in a lecture hall and that spy-cam might catch a murderer.

I tried to put it out of my mind.   I thought that if the purpose of the darn thing was to count attendance, and I wasn’t supposed to know, perhaps I should pretend it wasn’t there. 

Over the course of the semester, unsurprisingly, attendance dwindled.  Why?  Although I cannot say for certain, I suspect it’s because my course was video taped and then posted up to the course website. Perhaps students over-schedule themselves as the semester wears on.   I don’t know.   If I were a student I imagine that I could easily talk myself into skipping a lecture when time was tight and the deadline for an important term paper was looming.   After all, I could catch up by watching the video online at a later date.   But, I imagine that this could become a slippery slope.   You miss one lecture, and watch the lecture video.   Why not again?   

Over the course of the semester, I did my own eye-ball measure of lecture attendance, and by the end, I estimated that maybe only 25% of the enrolled class attended lecture. As one notable exception, three Knight Journalism Fellows from MIT faithfully attended nearly every lecture from start to finish.  

By the accounts of a few students who did not attend lecture regularly, they felt that they would have done better if they had attended lecture.   In some cases, critical pieces of information in lecture were missed, either because people fast-forwarded over parts of the lecture video, they were multi-tasking while watching, or perhaps didn’t watch some lectures on video at all.      

All of this is anecdotal and speculative on my part and doesn’t represent a systematic attempt to ascertain a cause-and-effect relation between regular lecture attendance and engagement with course material, but it does not seem far-fetched.  

This brings us back to the study itself.   Ignoring questions of privacy for the moment, all the study measured was bodies in seats, it did not measure the relationship between engagement in a course and attendance.    I don’t know how one could perform the kind of cause-and-effect study that could establish this relationship.  It would be far more invasive than the anonymous attendance study.   You would have to know that person X wasn’t attending lecture, and that person Y was, and that person X did poorly on the final exam, while person Y aced it. 

Over the summer, I forgot about the spy-cam and went on with my research work.   In mid-late September, the people doing the study contacted me and showed me the data for my course.   The attendance numbers seemed to confirm my impression of attendance: it dropped significantly as the course wore on.   I even recall being surprised at how ‘dense’ the classroom appeared during the midterm when attendance spiked, and the study confirmed this as well.  

I gave permission to the people doing the study to use the data from the course, and even gave them permission to cite my course specifically.  The data in the course were not shown at a HILT conference on Sept. 16^th. 

After the Sept. 16^th HILT conference, a fair amount of discussion ensued in the local press (Crimson and Globe) about the study.   I have to say that I consulted friends to get their opinions about the study.   Some felt that my analogy to a traffic survey was flawed as the data did contain images of people.    On the other hand, I don’t know if the study could have been done if people had known about it.   I also wondered whether the knowledge of the existence of the data would tempt hackers.     

I also wondered about public surveillance, as there are spy-cameras all over the place.   I learned from Globe Columnist Hiawatha Bray about the extent of public surveillance, where police photograph license plates as a regular practice.   Soon there will be a fleet of low orbit satellites that will update a global map on a daily basis with a resolution of 10 inches.    People can track your cell phone.   I tried to come up with a method where I might avoid detection and the only scheme I could imagine was turning off my cell phone completely, and bicycling around at night with a ski-mask on.   If we are concerned about a study where there were attempts made to preserve anonymity, should we not be concerned about surveillance methods that are known, and perhaps on sketchier ethical grounds?

Whatever the opinions or conclusions are about the ethics of the study, I hope we can pivot to the question of student engagement in courses.   Yes, attendance drops over the course of the semester, what are the causes and effects?