RE: Poll data relating to racism by party
August 22, 2012 at 9:48 am
(This post was last modified: August 22, 2012 at 10:23 am by jonb.)
(August 22, 2012 at 8:44 am)CliveStaples Wrote:(August 21, 2012 at 10:57 pm)Shell B Wrote: Your brain is sexy.
CS, will you ever post something that is not leading?
The sample sizes involved determine the width of the confidence intervals--i.e., the margin of error. For example, with the interracial marriage question, since we're talking about propotions, that means the margin of error is about .03 at a confidence level of .90. The margin of error is much smaller for some of these questions--the p-values for some are lower than .01.
Tiberius either doesn't know about how confidence level and width of confidence intervals work, or didn't bother to find out what the sample sizes were before he stated that "none of the numbers are really comparable".
Apparently, you were impressed by his answer. You must know a lot about statistics.
No, even if what you were saying was true, the questions would have to be exact, a term like 'black pushy people' can easily be misinterpreted by someone glancing at a question not seeing the significance of black, and giving an answer, and 0.1 would only need 1 in a thousand to do that.
Clive: You seem to constantly make the same mistake, you think as you know the equation you have the answer 1+1=2 but if the terms are erroneous, it does not matter about the maths. For instance one herd of cows plus one herd of cows makes one herd of cows, true the herd might be larger, but you put the two together and you get one.
