qwerty's p-value definition seemed different to me than woden's, so i went to look it up...
"A p-value close to zero signals that your null hypothesis is false, and typically that a difference is very likely to exist. Large p-values closer to 1 imply that there is no detectable difference for the sample size used."
so, woden's definition is correct for the way we are looking at this data - a 0 means the sets do not match, and a 1 means they do.
that said, qwerty's last bit about statistical significance requiring very low p-values is right on - the logic being you are seeking to prove that something is unlikely to happen normally. (like proving cancer rates are higher under power lines or something.) therefore, the closer the number is to 0, the more you are certain that your data is atypical.
Edited By igmo on 1083176216
Whoa....time out folks - My proposal
Guest wrote:Some of you need to take a basic statistics class.
"The observed significance level, or p-value, for a specific statistical test is the probability (assuming the null hypothesis is true) of observing a value of the test statistic that is at least as contradictory to the null hypothesis, and supportive of the alternative hypothesis, as the actual one computed from the sample data."
From Statistics 9th Ed. by McClave and Sincich, published by Prentice Hall, 2003.
Crudely restated: a p-value of .15 means that it is 15% likely that you will find evidence that supports the alternative if the null is true.
Just as our courts are set up to prove guilt (but never actually prove innocence) common statistical methods are designed around the idea of proving there is a difference between two populations, which is not the same as proving there is no difference between two populations.
In real world statistical work the extremes of the p-values are rarely reached, a statistician who saw a p-value of one would immediately think he had checked a set of data against itself. If a p-value of .9 was seen he would likely scowl at the researcher and tell him to quit wasting his time. This tends to come from the fact that there are just too many factors that can effect the outcome for all of them to be tightly controlled and that the data values will tend to range over a very wide set of numbers and any change in a population will probably be a small percentage of the overall range. A study for the American Lung Society can easily have 200 factors (i.e. 200 different things being tracked that might be effecting the data seen, which would be totally different from the actual measurements they would be taking: lung capacity, FEV, etc.).
Here our factors are limited enough and our ranges are small and possible differences large enough that we should be able to see extreme p-values.
Null Hypothesis: No Difference in ranges
Alternative Hypothesis: Difference in ranges
Factor: Version of Myth
p-values: ranging from .2 to .9 currently
thks. i tried that earlier, and even though it installed, myth crashed on load of any game or film...Brainbug wrote:try making a full disc image, that usually works even from scratched copies
and then you can even burn it afterwards
hmm when i tested it just now, it worked. irritating, since i bid on a copy of myth at ebay. boo
-
Pistol_Pete
- Posts: 183
- Joined: Sun Mar 21, 2004 3:45 pm
- Location: Massachusetts, USA.
bleh. Sorry to hear about your CD issue. I did a quick search of a site I know that sells some games cheap. I didn't find Myth II available there, but I did find Myth 3. You simply must click to see its price.
-
Pistol_Pete
- Posts: 183
- Joined: Sun Mar 21, 2004 3:45 pm
- Location: Massachusetts, USA.
Well, I won't many any pokes about Aussie currency, but if that price is in Austrailian Dollars - I sure hope it is - then it's about 57 bucks for Myth 3. Selling a few Myth 3 CD's might not let you retire, CIK, but it might help you fill up your gas tank a few times. Actually, I did lose my Myth 3 CD ages ago, and I was looking to replace it. What kind of offer are you willing to make me? 