Privacy Implications of Demographic Questions?

I’ve heard a few questions lately about the privacy questions of collecting student demographic data with clickers, largely as a result of the vendor i>clicker adding this feature to their system. The two specific concerns I’ve heard are the following:

1. Let’s say you ask your students a series of demographic questions (gender, major, class year) followed by a series of opinion questions, maybe about sensitive topics like race or politics. Your plan is to parse out the responses to the opinion questions using some of the demographic variables to shed some additional light on student perspectives. However, given the demographic questions you use, you inadvertently uniquely identify one or more of your students, say the only female, junior math major in your class. Even though your students responded to the questions anonymously, you can now see exactly how that female, junior math major responded to your opinion questions. Have you violated this student’s privacy?
2. You ask your students whether or not they have done drugs in the last week. You use your clickers in anonymous mode, but it’s clear that some of your students are drug users. Then one of your students commits a crime while under the influence a couple of weeks later. Might you or your institution be held liable for not acting on the information that there were drug users in your class?

This is a new area for me, in that I haven’t heard much discussion about how to handle these kinds of issues from faculty using clickers in the past. What’s your response to these scenarios? Are there other privacy issues that might arise from the use demographic clicker questions?

Image: “Privacy,” Rob Pongsajapan, Flickr (CC)