Lately, I’ve been reflecting on Michael Rozier’s post, Bad Data Always Leads to Bad Ethics, where he underscores the importance of quality data for shaping public policy in response to COVID-19. To repeat the adage: garbage in, garbage out.

In the context of testing and infection rates, Michael brings up a great question: if we use personal geolocation to intervene in our current pandemic, will it meaningfully improve our ability to ethically respond to this disease?

There is no shortage of accuracy issues to contend with when working with mobile data. When we try to assess the closeness of two individuals through phone positions, at the scale needed for contact tracing, we have to consider:

Do these individuals have their phones with them (or have phones at all)?

Are they separated by walls, such as in adjoining apartments?

Are they on different floors of a building, while having the same X and Y coordinates?

Can a contact tracing application account for these nuances?

The smartphone adoption rate is particularly salient for mobile contact tracing. The latest figures from the Pew Research Center put cell phone adoption at 96% of U.S. adults, but smartphone adoption at just 81 percent. Older adults (65+) and those with incomes less than $30,000 have lower rates of smartphone adoption, at 53 and 71 percent. These are groups which may be particularly vulnerable to the disease.

I offer one more accuracy consideration to this stack: personal location masking. This is the practice of obscuring your personal location data. You could do this by swapping out your actual location for another or by reducing the precision of your location so that people don’t know exactly where you are. The purpose of personal location masking is to protect your geoprivacy by controlling the degree to which you share your location with others.

Just a few examples of personal location masking include:

• providing an incorrect home address
• using VPN to change your IP address (and its associated location)
• disabling smartphone location services
• moving a map pin away from your actual location

Do you ever mask your location when you don’t want someone to have it?

What drives individuals to mask their locations?

One of the questions I’m interested in is: who is masking their locations, and, why? I explored this with a study of California adults in the pre-pandemic era. An interesting surprise: Californians engage in location-masking behavior across demographic groups. This means that people of different ages, ethnicities, education levels, incomes, genders, and urban/rural geographies hide their true locations from time to time.

The rate of respondents who admitted providing incorrect home address data within the survey is 15 percent. In addition, 28% of participants with smartphones routinely keep location services turned off.

It turns out that the most consistent predictor of personal location masking is a lack of trust in websites to protect their personal data. If respondents reported lower levels of trust in websites to protect their information, they were significantly more likely to mask their locations by one or more means.

Trust and mobile contact tracing

The idea of trust will continue to be important as we respond to the COVID-19 pandemic. Do we trust our governments and companies to protect our health and location data? Will our trust in the COVID-19 response make us more or less likely to share these data? The pandemic is also one of the first times we can see a direct link between our patterns of smartphone use and policies which impact us (i.e. my phone captured that I saw my friend, and now I am being asked to self-quarantine). How will this change our patterns of phone use and the physical distance we keep from our phones?

I’ll be discussing more about these issues of geographical ethics and biases in a webinar hosted by the Geographical Sciences Committee of the National Academies of Sciences, Engineering, and Medicine on June 11, 2020 (link forthcoming).