Here's the link to the scale of values page that Doggygirl mentioned, if anyone wants to check it out.
http://www.isu.org/vsite/vfile/page/fileurl/0,11040,4844-152085-169301-64123-0-file,00.pdf
Well, like Emma, I am supposed to be working, but...
I just ran some of the spin statistics through a least-squares linear regression analysis to test the hypothesis that the judges are using the GOEs to give a double reward to the technically most challenging spins. This is, are you more likely to get positive GOEs for a hard spin or for a particularly well-done easy spin?
I used the data from the top ten ladies in the free skate at worlds (sample size 40, since each skater did 4 spins). First I ran "level" versus "average GOE." Is there a statistical tendancy for higher level spins to get higher GOEs?
The coeficient of correlation was r = .36. Informally, this means that only about 13% (r squared) of the variation in GOEs was correlated with variation in level of difficulty. (I also tried logarithmic and exponential correlation, but the relation was even smaller for those tests.)
The rest, 87% of variance in GOEs, was due to "other factors" -- i.e., presumably to proper judgments about the quality of the element independent of the level. Indeed, the .36 correlation probably reflects nothing more sinister that the fact that the best spinners do the hardest spins
and do them well.
If instead you compare base value to GOE, the correlation is even lower: r = .14, so only two percent of the variance in COE is related to the base value (and presumably 98% to the judges' determination of quality).
So that shoots that particular conspiracy theory in the head.
BTW, in all but one of the spins (40 in all) done by the top ladies, they got a positive GOE. (The one exception was Ando's flying combination spin, which got a negative GOE of 4 one-hundredths of a point.) So this shows that the judges are not being very critical of weaknesses such as traveling. But at least they are ignoring these problems consistently across the board (IIRC Miki's combo spin was really attrocious).
Mathman