This week’s thesis work has focused on client work and website work, and has brought up some interesting conundrums about consent, privacy, and self-run studies. I thought that I would share some of the things that I’m working through right now, because it’s one of those topics that you don’t think about when you are first starting thinking about a thesis topic or your experiment in general.

First, a little background on the experiment I am running. I have a self-run experiment, where the subject will download a client onto their computer and run the experiment, and the program will gather data on the interactions between the user and the program and then submit it to a web server application for later processing and analysis by myself for the purposes of proving my hypothesis. I ran a study in a similar way last year titled Selection and Movement Methods for Multiple Mobile Agents. (The study server is down, so you can run the program, but no additional data is collected.)

In order to make the subjects want to participate in the study, I would like to let the subjects enter their email in the program which will then be submitted along with the data, and have a random prize drawing for participants of an online retailer gift card. This brings up some concerns for me as an ethical researcher, because I don’t want to have a list of emails just sitting around after the prize is over. It’s a privacy concern that I’d rather not deal with, and it relies on a level of trust that I will not use it for some nefarious purpose.

I am technically competent enough to assign each user an anonymous identifier and keep it separate so that I can delete all the emails without losing any important data (the user can run the program more than once, and that could provide even more useful data about training times and such). At the same time, I know that anonymizing data is hard and would like to avoid any issues with that as well.

The main conundrum at the moment is this: to get consent to monitor the user, I want to have a click-through title page at the beginning of the program. I need to word this in a way that clearly says that I’m only monitoring them for the purposes of the study itself, and will only use their email for the drawing and will delete that data later. I also need to be careful that this page isn’t too scary so that a lot of people don’t decide right then that they want to bail. Adding to the fun is that I need to be somewhat vague about what I’m actually testing, so that I don’t spoil the results. The wording is.. delicate.

As if that isn’t enough to think about, there are two special bonus rounds which are complicating matters further! First, I am going to get an IRB waiver for this study, so the review board will be reviewing my “consent form” that I am presenting, and will need to give it the OK. There really is a large amount of oversight for any study which involves subjects, including making sure there isn’t gender bias and many other things, but that’s another discussion. The second bonus round involves fancy statistics which I think that the subjects themselves would be interested in. Personally I think it’s nifty to see a page of your stats at the end of the thing, and would want to let each user see how they did, possibly in comparison to the rest of the subjects, or to earlier runs themselves.

So some part of this week will be nailing down the exact wording of that page, which is only a tiny fraction of the program itself. Because I have a study website to avail myself of during this, I am probably going to include a sample piece of data that they can look at if they’re curious exactly what I’m collecting, as well as an explainer page that is somewhat reassuring. I’m hoping that I can finish the site and client soon, and start the study in the next few weeks, so that I can get a decent amount of data to crunch before February, because I can then wrap this section up and move on to the rest of the thesis of doom.