Lokasenna
Posts: 9297
Joined: 3/3/2012 From: Iowan in MD/DC Status: offline
|
quote:
ORIGINAL: PaxMondo quote:
ORIGINAL: Canoerebel There is some margin for people to interpret data differently but probably not as much as feared. If an instruction were handed down by a governor or the head of a state health department to "cook the books" to paint a rosier picture, or a bleaker picture, there would be leaks, the press would get ahold of the story, and the instigator would be hounded or demolished. It's far more likely that variations in reporting/tabulating are good faith differences in how jurisdictions or entities or individuals do things. As we've noted before, there is merit in the statistics, especially as long as a given jurisdiction is consistent in reporting. Thus, if Belgium reports 200 deaths on April 15 and 25 today, it means something, even if Belgium counts differently than Denmark. I'm not suggesting that anyone is purposely cooking books or anything like that. I'm just saying this is a group that loves numbers, and just cautioning you that the basis for the numbers being released is NOT consistent and this is known. If you understand that statement, then you understand why I am urging caution. The data is being released because of the need to provide something to the public that they can understand. It isn't being done to intentionally mislead, its just that at this point even the Wharton graduates need another couple of months to get all the data onto a consistent basis, and then it will take another month to get the required 3-4 peer comparables to confirm. My guess on this is Aug, but it might slip another month or two. Until then, we use other, more consistent data to infer what we need to know. You can't explain this to the masses, it isn't that it's rocket science, it actually is considerably more advanced mathematics than that. As long as one isn't trying to mix numbers from different models at this early stage, and knows the limitations of the numbers being cited (such as how deaths are counted, in this particular example), then one can make reliable observations over time. If you're only looking at one model, know its limitations (and its changes/updates and how that affects the numbers, if applicable), and you are using it as a barometer for "are things getting better or are they getting worse", then you're OK to make general observations using that. But not much more than that, at least not with any reasonable degree of certainty. Basically anything else falls under conjecture, at which point we're entering into the realm of "opinions are like... and everybody has one."
|