Need Appropriate Test for comparing a series of measurements from one bone to another

My research involves taking measurements at 9 sections at different locations on the shaft of bones. I am having trouble determining which statistical methods I can use to test for significant difference in the distribution and change of the measurements as they progress along the shaft of the bone from 1-9 from one bone to the next. This also includes comparisons to many many bones. Does anyone have ideas as to how I can do this? To clarify, I have 9 measurements for each bone, all measurements are taken at 10% intervals of the total bone length. I am looking, first of all, just to compare the Left and Right pairs from within single individuals (they're monkeys). Ex: The Left and Right femora (femurs) from individual 025. The test needs to focus on the distribution of measurements for each bone, and not the size of the measurements across bones. Some individuals were larger than others, so tests such as Wilcoxon Rank Sum won't test accurately for distribution difference, only if the measurements of one individual are significantly higher than the others'. I don't know if there is a way to do this, but any suggestions would be greatly appreciated! Thank you in advance!


Well-Known Member
Re: Need Appropriate Test for comparing a series of measurements from one bone to ano

You need two things for a significance test - a measure of how different the two shapes are, and an idea of the usual distribution of the differences. This will let you calculate whether the difference you see is significant. One possible way to measure the difference in shape is to take your two sets of 9 measurements and scale them so that the mean of each set is 100. Calculate the 9 differences, then find the standard deviation of the 9 differences. This is a sort of % difference. If the answer is low, then the shapes are very similar. If the answer is high, then they are not.
The problem now is - is the difference you have found low or high? and if it looks high, is it sufficiently high to be regarded as significantly high and so the shapes significantly different? One way is to take lots of bones that you know are the same shape - 1000 mature right femurs of the same species, and measure them. Find the difference between them paired thousands of ways to get a usual distribution for the measure. Sounds like a lot of work.
Then if the difference you get is in the top 5% of your distribution, you claim a significant difference.
Re: Need Appropriate Test for comparing a series of measurements from one bone to ano

A few examples of the data (or many data points) would make it easier to understand.