tuner-automotiv
New member
What is the better format to transfer my Cd's to my ipod?
ACC or MP3?
Are they the same quality and size?
ACC or MP3?
Are they the same quality and size?
Just to make clear what statistical significance means.kornchild2002 said:Sure, the iTunes AAC encoder did provide SLIGHTLY higher results (a difference of 0.2 is not big) but these results were NOT statistically significant.
Thanks for explaining the statistical significance jargon. What I ment by slight difference was that, to the average person, when they see iTunes AAC getting ~4.5 with Lame getting ~4.3, they see it as being that iTunes AAC has a slight lead over Lame. However, as you pointed out and as the results explained, this is not true. The test could be conducted again and the results could be switched, they could be exactly the same, or they could be totally different.mnhnhyouh said:Just to make clear what statistical significance means.
It is based on the presence of variability in results.
If I have a coin, how do I tell that if is biased? I toss it three times. It comes up with three heads in a row. Does this mean it is biased?
Statistically heads will come up three times in a row 1 in every 8 lots of three tosses. So if I do 800 lots of 3 tosses, 100 of them will have 3 heads. That is with an unbiased coin!
Statistical significance is (mostly) based on there being a less than 1 in 20 chance that the result comes from no difference. So back to the coin analogy, 4 heads in a row is a 1 in 16 chance, so that would not be a significant result, but there is only a 1 in 32 chance of 5 heads in a row. So if we were testing coins for bias we would throw out all those that gave us 5 in a row, knowing that we would be getting it wrong less than 1 in 32 times.
Now back to the original 0.2 difference in the test above. Given that the results are not significant what we are saying that there is no difference. Calling it a slight difference is not true. When viewing non-siggnificant results we should realise that it is likely that if we repeated the test there is an equal chance that the results would be the other way around ...
h
If the difference is not significant- it is a slight difference and there's no reason to say there's "no difference". The difference is slight - as revealed by the data - just not significant - as according to the stat test. There may or may not be a "real" difference - it just wasn't detectable by the statistical test. We don't know how things would come out if we did the test again (we don't know if this is like flipping a coin or not), and in fact, the more comparisons you make, the bigger your sample size, and the more power/ability you have to detect (statistically speaking) a small difference.mnhnhyouh said:Just to make clear what statistical significance means.
...
Now back to the original 0.2 difference in the test above. Given that the results are not significant what we are saying that there is no difference. Calling it a slight difference is not true. When viewing non-siggnificant results we should realise that it is likely that if we repeated the test there is an equal chance that the results would be the other way around ...
h