Online experiment collecting location data

[Abby Clements]

Hello! People in my lab are looking into running a study on the connection between traveling mental/semantic and physical space. To do that, they want to track people's movement or location at several time points, likely through their phone. Has anyone on here done something like this? If so, what is the easiest way to do it--have people download their old data from Google Maps or something similar, or have them install an app on their phone that passively collects it? How much is the typical compensation rate for this type of data?

Additionally, even with privacy measures in place (the raw data would never be made publicly available and would only be analyzed in certain ways, the participants could proactively stop data collection for certain intervals or retroactively request that it get deleted, etc.), there is concern that the IRB wouldn't approve the experiment, or that the cost or delay would be too great. If anyone has any experience with modifying their IRB protocol in this way, that would be super helpful.

Thank you!

[Joshua Hartshorne]

HI Abby,

You certainly could get people to give you a data dump from Google Maps. The only study I know of along these lines is this one that just came out in Nature Neuroscience. It looks like in that case, they maybe had a specific app. Though that wouldn't change the data security issues, though. 

Since the data are going to show where people live, I think you are right that there's no way to really guarantee anonymity in the event of a data breach. So I'd want to have really good procedures in place, including strictly limiting who has access, using two-factor authentication, etc. Do all analyses on a single secure server, never on your own computer, etc. And basically telling people that if they are doing super sensitive things (like constantly going to a known brothel or meeting with spies at some embassy), they probably don't want to participate in your study.

As far as IRB, if you can make a good case that you are doing a decent job of protecting identity, that subject are adequately informed, and that the research is valuable, then in principle they should approve it, perhaps with some modifications and requests. The IRB's job isn't to prevent any risky research is done, only to ensure that the risks are appropriate, that subjects are giving informed consent, etc. 

You might also talk to the folks at LookIt! or Databrary, who have thought a lot about data security. 

Josh