Dak Posted September 25, 2008 Share Posted September 25, 2008 basic question: 1/ on day one, 45% of people are in favour of x 2/ on day two, 50% of people are in favour of x assume a margin of error of [math]{\pm}[/math]0% and a certainty of 95% (i.e., 5% chance of each statement being false). so, what's the certainty of the conclusion 'therefore, support for x has grown between days 1 and 2'? I suppose that if there's a 5% chance of each precept being wrong, and two precepts (i.e., 2 chances for a precept to be wrong), then there'd be a 10% chance of... what? the conclusion being not neccesarily true? The certainty of the conclusion seems as if it'd be 10%, but that feels a tad wrong. e.g., precept 2 could be wrong, but in actual fact on day 2 support could have been 60%, thus the conclusion would still be correct. so, yeah, basically how do the certainty intervals combine in this case to give a certainty for the conclusion? (for bonus points: how would that work with confidence intervals instead of certainty?) Link to comment Share on other sites More sharing options...
NeonBlack Posted September 25, 2008 Share Posted September 25, 2008 I'm not an expert on statistics and probability, but I don't think this question is well-formed. For example, what does it mean for the statement "on day one, 45% of people are in favour of x" to be "false"? Link to comment Share on other sites More sharing options...
Dak Posted September 25, 2008 Author Share Posted September 25, 2008 my appologies. I meant if the statement was false, then z% of people would be in favour of x on day 1, where z is not 45. i've mostly seen statements of certainty (as opposed to confidence intervals) on public oppinion polls, hence why i worded my example like that. (think estimates of parent-population mean based on sample mean: you usually end up being able to say 'there is x % chance that the parent population mean is between y and z' iirc) Link to comment Share on other sites More sharing options...
Josy Posted September 29, 2008 Share Posted September 29, 2008 Since the proportion of the population that supports x is a continuous quantity, it's not simply a question of true or false; presumably you'd model the proportion of the parent population as a normal distribution with a mean of 45% (or 50% on Day 2) and a s.d. of something or other, then move forward from there. It's a while since I did any stats as well, so I can't really be much use on this. Link to comment Share on other sites More sharing options...
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now