(please note, this is largely intended for humour purposes)
In research the topic of personal information is a touchy one. On one hand, a study may be predicated on some very important or sensitive personal information from its participants. Sexual preference, political views, involvement in criminal organizations and the like are hot button issues and ones which drastically increase both the vulnerability of those involved in the study, and the risk to them.
Vulnerability is generally defined as being how socially vulnerable the participants are. Its rated as Low, Medium, and High, with Low being your average man or woman on the street, medium being those who may have sensitive or damaging information but are probably OK, and high vulnerability being people like minors, the mentally disabled, and the criminally insane.
Risk is the actual risk that is posed to the participants by their involvement in the study. Low risk studies look at things that the participants are already doing, or might do, in their day to day lives. Studies about reading habits, preferred meals, and the like are low risk. Medium risk are things that participants don't generally do that might get them in trouble in certain circles or certain circumstances, but if they're low vulnerability are probably OK. High risk studies are things like drug deals and the number of accidents that occur by way of juggling chainsaws.
Now, one can reduce the risk to participants by making them as anonymous as possible. As such, I have created the Absurdity Test. The Absurdity Test is a basic metric of exactly how anonymous your survey is. It is a simple question you have to ask yourself. If the answer to this question is "No", then chances are your survey is about as anonymous as it gets. Here is the question:
"Would I know the difference between a dozen 15 year old mentally disabled illegal immigrant serial killers taking this survey, and a dozen geriatric patients from the retirement home down the road?"
Its absurd, certainly. But if there is no way to tell the difference, then the information you're taking in is absolutely anonymous. Congratulations. Now you just need to figure out how worthwhile that data is...but at least you can assure your research ethics review board that there is no way that the participants could be harmed by taking part in the survey.