Like others, I have been seduced by the “what X are you” quizzes on FaceBook. I certainly understand why they’re compelling, but I’ve begun to worry about just why they’re so prevalent. And I’m a wee bit concerned.
People like to know things about themselves. Years ago, when we built an adaptive learning system (it would profile you versus me, and then even if we took the same course we’d be likely to have a different experience), we realized we’d need to profile learning a priori. That is, we’d ask an initial suite of questions, and that’d prime the system. (And we intended this profiling to be a game, not a set of quiz questions). Ultimately that initial model built by the questions would get refined by learner behavior in the system (and we also intended a suite of interventions ‘layered’ on top that would help improve learner characteristics that were malleable).
The underlying mission given us by my CEO was to help learners understand themselves as learners, and use that to their advantage. So, in addition to asking the questions, we’d share with them what we’d learned about them as learners. The notion was what we irreverently termed the ‘Cosmo quiz’, those quizzes that appeared in Cosmopolitan magazine about “how good a Y” you are, where one takes quizzes and then adds up the score.
Fast forward to now, and I began to wonder about these quizzes. They seem cute and harmless, but without seeing all the possible outcomes, it certainly seemed like it might not take that many questions to determine which one you’d qualify as. Yes, in good test design, you ask a question a number of times to disambiguate. But it occurred to me that you could use fewer questions (and the outcomes are always written intriguingly so you don’t necessarily mind which you become), and then wonder what are the other questions being used for. And the outcomes here don’t really matter!
So, it’d be real easy to insert demographic questions and use that information (presumably en masse) to start profiling markets. If you know other information about these people, you can start aggregating data and mining for information. One question I saw, for instance, asked you to pick which setting (desert, jungle, mountain, city), etc. Could that help recommend vacations to you?
When I researched these quizzes, rather than finding concerns about the question data, instead I found that much more detailed information about your account was allowed to be passed from Facebook to the quiz hoster. Which is worse! Even if not, I begin to worry that while they’re fun, what’s the motivation to keep creating new ones? What’s the business relationship? And I think it’s data.
Now, getting better data means you might get more targeted advertising. And that might be preferable than random (I’ve seen some pretty funny complaints about “what made them think this was for me”). But I don’t feel like giving them that much insight. So I’m not doing any more of those. I don’t think they really know what animal/movie character/color/fruit/power tool I am. If you want to know, ask me.