06-21-2006, 12:41 AM
Quote:listing known original helmets and fragments, I noticed that much more right cheek guards seem to be found. A total of 16 finds contains 10 right pieces, 3 left pieces and 3 pairs.
I'm going to interpret this as 13 rights and 6 lefts in a total population of 19. I'm not a professional statistician, buy my Z-statistic calculator with these proportions yields a Z-statistic of [1.50, implying a confidence level just over 85% that these are statistically different -- not strong. At least, you wouldn't want to go to court with it. ]
[However, if we had 38 total finds, and there were 26 rights and 12 lefts (which is the same proportion), then the Z-stat would be 2.13, for a confidence level of well over 95% that they are statistically different. You would probably be able to convince a jury.]
We need to go find more cheek pieces!
rkmvca/Rich Klein
EDIT: Please strike the statistical results in [brackets] above. I believe I defined the populations incorrectly. With the correct populations, I believe the z (or t) statistic is now about 2.25, indicating a very strong probability that there is a real difference in the predominance of left vs right cheek guards. I'll check with a real pro to make sure I have done the calculations correctly.
