For starters, it's measuring non-fatal illnesses. If your regulatory regime is converting otherwise fatal accidents to non-fatal ones, then you're going to have higher rates despite being safer. (Historically, this has happened: the introduction of helmets in WW1 caused head injury rates to go up... because people who otherwise died now lived with a head injury instead).
Another point is that states have quite different workplace characteristics, so the baseline risks are going to be quite different. Working in a coal mine is going to be an inherently more dangerous job than being a paper-pusher in an office tower. Unsurprisingly, West Virginia (which has a lot more of the former than the latter) is in the above-average list and New York (which has the reverse) is in the below-average list.
Finally... there's no mention of the degree of difference at all.
This actually helps to mitigate your other point, because all types of industries can have non-fatal injuries from things like overexertion or slips on wet floors but fatal injuries are dominated by dangerous industries like mining, transportation and agriculture that differ by state.
The fatal injury rate between states is quite incomparable in particular because motor vehicle accidents are the leading cause of death among working-aged people, and the rate will be higher in states that are more spread out because workers in all industries there log more vehicle miles for work (and at higher average speeds).
> Finally... there's no mention of the degree of difference at all.
Texas can't be worse if it's better.
> States with Non-Fatal Workplace Illness and Injury Rates Above the National Average:
> California
> States with Non-Fatal Workplace Illness and Injury Rates Below the National Average:
> Texas
It's completely possible that Texas is less adversarial/bureaucratic or similar than California, but whatever they're doing is working.
Though it could be cultural differences. Government regulations are often half good best practices and half harmful backwards nonsense. A culture that only calls OSHA for not following best practices even after having it pointed out and not for not implementing backwards nonsense could pretty easily end up with a better safety record than one that prioritizes strict compliance over actual safety.