I blogged in July that fatal shootings of police for the year, though up from last year, were still at the 2005-2015 average. I regret to say that they seem to have risen substantially from that rate, though they’re still below the high points of 2007 and 2011.

The AP (Lisa Marie Pane) reports:

Ambushes in Dallas and Baton Rouge, Louisiana, and other shootings around the country led to a sharp increase in the number of police killed in the line of duty this year.

From Jan. 1 through Wednesday, 135 officers lost their lives. Some died in traffic accidents, but nearly half were shot to death. That’s a 56 percent increase in shooting deaths over the previous year.

Of the 64 who were fatally shot, 21 were killed in ambush attacks often fueled by anger over police use of force involving minorities.

In July, such fatal shootings were at an annual rate of 50 per year, a bit below the 2005-2015 average of 53 per year:

(From the National Law Enforcement Officers Memorial Fund; rows corresponding to causes of death other than shooting deleted)

This year’s 64 (which statistically might rise to 65 in the remaining days of the year) is now well above the average, though still below the 70 we saw in 2007, and 73 in 2011. The much greater sharpness of the increase from last year stems from the fact that last year’s total, 41, was fairly low by recent standards.

It’s possible, then, that the increase stems from “anger over police use of force involving minorities.” But it’s also possible that this is normal year-to-year variation, which can be very high — for instance, I know of no evidence that the still-higher 2007 and 2011 totals stemmed from such anger or that the dramatic declines in 2007-08 and 2011-12 (and the two-year plummet from 2011 to 2013) stemmed from any great decrease in such anger.

These murders are of course awful, and, as with shootings of unarmed citizens by the police, they can cause great indirect harms as well: For instance, as police officers feel more threatened, they may be more likely in close cases to shoot at people whom they perceive as threatening, and they may also be more likely to avoid threatening situations, even if that means cutting back on effective law enforcement. And, unlike in July, we now do see an increase over the medium-term average, not just an increase over an unusually low year. Still, there is so much yearly variation in the numbers that I’d be hesitant to draw confident conclusions based on just this year’s data, just as there was no reason to draw such conclusions in 2007 and 2011, or draw conclusions about dramatic improvements in underlying social conditions in 2008, 2013 and 2015.