heruamon said:
|
although I always expect you to have some important information to contribute to the arguments, the fact of the matter is that niether can be taken by themselves.
Polling is part of nearly every statistical research, and polls have been proven to be exceptionally accurate. However, they are never good indicators of anything without additional information, and can even be used incorrectly to prove things that aren't true.
Of course, one warranty company's failure report in no better. I don't think that they are reporting incorrectly, but as any good researcher knows, you can't always take the word of a person who has something invested in the stats. The other thing every good researcher knows, is that every graph and statistic must be taken along with all additional information. Here we have no other information. We don't know how long these warranties last, we don't know what types of errors were reported, we don't know if this is an internal or external report. We don't know anything, and for you to say definitively and with excessive verve, that one company is more reliable than another, using evidence that I've shown to be exceedingly unsupported on both sides, is ignorant.
To really drive home my point, these two examples make a better case when used together, than when used separately.
On the other hand, lots of professionally used statistics use nothing but phone polling. In fact, VGC uses a type of polling to find it's numbers. There are no professionally used statistics though, that gather information from only one company. Especially not one that has vested interest in the service, and is only one of the companies involved. THat's like reporting a president as having won an election when only polling people from one state.
