saicho said:
you have misconception about a lot of things. |
It's time you took a look in the mirror. I've taken several levels of statistics and I told you that the specific numbers don't matter.
You want to know the numbers because you think there is a linear relationship that you could understand. You think there is a number like 39 out of 100, out of a thousand. But in reality, it could be any series of numbers. 35 out of 90, 70 out of 92, 180 out of 181, 731 out of 1000. They can potentially all lead to 39%. Because as I explained, there are several factors which change the numbers: several factors that are carefully calculated. You are not going to get a linear relationship like 39 out of 100, or 390 out of 1000. And that is what you don't understand/refuse to understand.
Secondly, you might reply to this saying, "I never cared about the specifics, I just want some proof and a look at the methodology". Well, like I said previously, you aren't going to get it. A company isn't going to give you, an anonymous forum member, the secret recipe that made them 20th stat company in the US. Of course, you can take the lack of evidence as evidence that the methodology is flawed. I'm not going to tell you that is wrong to do, even though it is. You can do what you want. But try to keep in mind that it would be a big mistake for a company to release baseless numbers. Sure, maybe they made a mistake, but don't count yourself so highly that you would be the one to figure it out, and never would the error be something so mundane as not having a large enough sample population.
But hey, don't listen to me...listen to kasz, he is far more advanced with stats and research than I, and I defer to his wisdom even though I may take argument with what he says at times. The way I look at it, when I have an argument with kasz the arguments I have are the footholds I'm stepping into. With each argument I go higher and higher until I overcome the obstacle of understanding. That is how you should treat this entire discussion.
