Originally posted by Yume
Do you even know how many people were polled? That is your problem. I do know stats and there is a basic quota you should make to your sample size. This poll didn't make it. Statistivs are good and I like to use them, but the data you provided just doesn't work. There are basic requirements that people must achieve to have valid data. You have not even touched that argument thus you have failed to claim validity. I have said this many times and I gave you too many times to respond. It is you who need to listen or read in this case more closely. I recommend you look into your data more in depth before you use it. A big problem I have found with many people is they do not do an analysis on others data. They just accept it as true. That is not a very good idea. You should know how the research was done before you believe it.
Dear God, you really are this dense, aren't you?
If you "know statistics," answer my question. What's the highest level of mathematics instruction that you've had? To be fair, I'll list the pertinent math courses I've taken, so we can compare who knows best what they're talking about: Calculus I, Calculus II, Calculus III, Statistics, Differential Equations, and Linear Algebra, all of those taken and successfully completed with a cummulative 4.0 GPA within the last 5 years, with Calculus III and above taken at a first tier private university.
My response to your "argumenents" has been, for the first two times, and now this third one: REPRESENTATIVE SAMPLE. Go and look up the damn word, Yume, and then realize that your claims of a sample too small to accurately represent the American population are impotent when a small, REPRESENTATIVE sample is constructed and polled. If you actually bothered to look at the data you're criticizing before plopping down at your keyboard and chewing and regurgitating what you've already said, you'd realize that the household survey covered 106,418 households. That number is 90 times as large as the approximately 1000-1500 people needed to construct a representative sample of the US population with a margin of error of +/- 3%. If you know how to multiply numbers, you'll find that even with an error of 3% (which is unlikely, since the standard error for the mean and medium incomes is less than 1%), the averages DO NOT CHANGE enough for the overall trend to reverse. Would you like to know what margin of error is necessary for the trend to reverse? For such a thing to even be possible for a portion of the trend would require that the error be greater than the greatest magnitude of difference between two categories, and that would necessitate 11.3% error. To reverse the entire trend would require far more error. That is unheard of. If such were the case, US census data would never be used by government policy makers or researchers in academia and the corporate sector.
So, before for you start making an blithering idiot of yourself, take your own advice. Analyze the data before you criticize it.
|
|
Bookmarks