It's also possible that people who felt most at risk were also most likely to turn off their phones or leave them at home.

It's hard to know what kind of biases are in this data without access to it.