Workers Comp Statistics Come With Many Caveats
The Workers Comp statistics production is one area that I have noted a sharp increase in over the last few months. There are so many statistics out there that make no sense; have (very important) an underlying unreported statistic, or relate the obvious together.
For example, I can make the proper statistical prediction that assaults in a certain area can be directly related to the sale of ice cream and I can prove it. The unmentioned underlying statistic is the time of year. In the summer as more people are out and about, there are more assaults.
The same can be said for summer ice cream sales. Both statistics are increasing during the summer and are not related in any manner. The key is that if one wants to associate two variables, with enough work and number massaging, the results can almost always be proven.
Some days, I want to bang my head on the desk when I read some of the Workers Comp variables that are being related to one another. I am not saying any of the studies are incorrect, they just have many underlying variables that actually cause the relationship or tell us what we know to be a fact.
There are also studies that prove what we knew all along. One example from NCCI is here. Actually, for the most part NCCI, WCIRB, and the rating agencies for Worker Comp are usually accurate, timely, and informative.
They also seem to be very adept at not relating unrelated variables. WCRI deserves a mention as a non-rating bureau source of information.
For instance, I saw one today that said injured employees take less illicit drugs when drawing Workers Comp benefits than other types of insurance programs. Do you think it might be that the person is being supplied with legal drugs, so why smoke a joint or snort a line and possibly end up in jail? Could it be the injured employee knows they are under a bit more scrutiny and want to be clean if drug tested upon a return to work?
A non-WC example was released by the Census Bureau today. They indicated the percentage of the population that is insured for health has actually slightly increased overall. Of course, that is true due to two possibly hidden variables:
- The huge aging baby boomer population means more people are on Medicare.
- Due to the very bad economy, more people are on Medicaid.
I did not read the entire study, but the conclusions are obvious before the study was released. In my humble opinion, the best studies are ones that actually have a well-founded surprise relationship between two variables. The rating agencies crank out interesting studies all the time.
©J&L Risk Management Inc Copyright Notice
