How many roofers or lumberjacks are killed by homicide in the course of doing their job?

Untrue at least for USA, only 28% are road related deaths.

https://www.bls.gov/iif/oshwc/cfar0020.pdf

Your source is only for the year 1995, and even there it was the second most common cause. It was the most common cause on average over the last ten years (traffic fatalities are split, you have to add them together): https://nleomf.org/facts-figures/causes-of-law-enforcement-d...

Even on your source they aren’t the majority

The sum of traffic accidents is 335+4+1+2+58+4+130=534 vs. 528 who were shot.

Edit: Even better: the chance of being shot is way lower for an LEO than the average US citizen: Its .000122 for the average person vs. .000066 for an LEO.

(2017 numbers: 1/32510000039773=.000122 for the average vs. 1/80000053=.000066 for LEOs, gun deaths from here https://time.com/5476998/risk-of-guns-america/ and LEO deaths from the source above)

Did you sum the numbers for all types of moving vehicle sources of hazard?

If you do the math, it does appear to exceed the number of incidences of being shot on the job.

Yes and even then it doesn’t

Additionally the claim wasn’t, ‘exceed the number of incidences of being shot on the job’ it was, ‘majority of [deaths] are traffic accidents’

Depending on your perspective, the claim is not entirely incorrect.

The most neutral reading of the presented data is that police are equally likely to be shot, have a moving-vehicle accident, and have a job related illness.

The remainder of job related hazards are extremely marginal in comparison to those three.

Guns, vehicles, and illness compose the near totality of incidents and those incidents are distributed nearly equally between the three hazards.

This is your source:

> SOURCE: US Department of Labor, Bureau of Labor Statistics, Census of Fatal Occupational Injuries, 1995.

Notably from 1995.

Doesn't your source say that only 47% of police deaths are homicide? That's less than half of the statistic used for this ranking. Am I reading that table correctly?

I don't have my own stats or sources; I don't particularly have a dog in this hunt. This seems like not a useful source at all in my opinion.

Those are very old stats.

https://www.cdc.gov/niosh/topics/leo/default.html