Northeastern University criminologist James Alan Fox has looked at the murder rate in individual cities over time and made a discovery. “I found 2-1 odds that if it goes up by 20 percent or more one year, it will go down the following year,” he told me. Random fluctuations are inevitable. In a period of relatively low crime rates, there is probably also an irreducible minimum — which means that everything else being equal, they are more likely to rise than to fall.

But it would be imprudent to ignore what happened in 2015. The numbers cited by those who see a Ferguson effect are not conclusive and could turn around this year. Still, as Mac Donald told me, “a 17 percent change in a year up or down is huge.” A one-year increase is not a lasting upward trend — but any lasting upward trend begins with a one-year increase.

“It is not possible to speak to police officers, in my experience, without hearing of their concerns about the current political climate regarding policing and the hostility they are encountering on the street,” she attests. But diagnosing the causes of an increase in violent crime is not the same thing as devising a cure. The Ferguson effect, if it is real, is the result of factors that are not terribly amenable to control.