Some statistics are just a function of term definitions

A good question to ask of any headline-generating statistic: to what degree is this number a function of how terms are being defined?

n% of major US cities are now considered “unsafe”