Repository navigation
Session tracking and data security in "Survey Unlinking" #611
|
Hi, I'm using the "Survey Unlinking" feature for a survey that collects participants' email addresses, with the goal of anonymizing personally identifiable data. Since email addresses often include names, I want to store them separately. I understand that once the unlinking is performed, there’s no way for me as the researcher to reconnect the datasets or identify individuals. So, in the unlikely event of a security breach (e.g., a hacker gaining access to the database), could such internal information be used to reconnect participants’ emails with their survey responses? Thanks in advance for your help! |
Replies: 3 comments
|
there is this issue. You are right. |
|
Yes, the data remains linked and is just not shown linked to the researcher. It shouldn't be possible for the researcher to see the email address in "pretend/spy mode" as @dfeistauer though, unless the email address is echoed somewhere in the study. I am interested in developing a better solution to this, i.e. encrypting sensitive survey data with the public key of the researcher, so that the sensitive data and the contact information are never readable on the same computer (ie. the sensitive data would be encrypted on the server, so it can only be decrypted offline by the researcher, and the contact information would not be accessible to the researcher). However, this is a fairly major feature, so I don't see it happening very soon unless someone can put programmer fees for this missing feature as a line item in a grant. |
|
For some, it might be an option might be use push notifications instead of emails with the upcoming version of formr. |
Yes, the data remains linked and is just not shown linked to the researcher. It shouldn't be possible for the researcher to see the email address in "pretend/spy mode" as @dfeistauer though, unless the email address is echoed somewhere in the study.
I am interested in developing a better solution to this, i.e. encrypting sensitive survey data with the public key of the researcher, so that the sensitive data and the contact information are never readable on the same computer (ie. the sensitive data would be encrypted on the server, so it can only be decrypted offline by the researcher, and the contact information would not be accessible to the researcher). However, this is a fairly major f…