Correlation of two variables through others

chomi

New Member
#1
Hi,
I have two data sets. In both sets there are 3 of the same variables (density, MOE and frequency). In addition to that, each set has one more variable (tension strength and bending strength). Since i can not measure both strengths on one piece, I would like to estimate the correlation (or any relationship) between the two.
Does anyone have any idea?
Thanks!
 

Miner

TS Contributor
#2
Can you provide a little more context to your scenario? MOE = Modulus of Elasticity? What does frequency mean in this context? How were the samples collected? Are they all from the same batch of material? One group per batch? One/two piece(s) per batch?
 

chomi

New Member
#3
Yes MOE is the Modulus. Frequency is the 1st Eigenfrequency of the specimens. The 1st set of data has 1000 pieces of wood, the 2nd 1000 different pieces of wood. A single piece is only in one group/batch/data set.
 

Miner

TS Contributor
#4
I think you are out of luck. The variability between each 1000 piece sample is likely to be much greater than the variability within each sample. This would overshadow any potential correlation that you are seeking.

If the material were relatively homogeneous (e.g., steel, aluminum, etc.), you can subdivide each piece to test torsion strength on one portion and bending strength on the other. I suspect that wood is much more variable than steel, but this variation may still be less than the variation between the 1000 piece samples.
 

chomi

New Member
#5
Thank you for the reply. I will try to divide both samples in bands according to the frequency values and try it with that. And a correlation zone instead of a single value would be a good enough result. Lets say a value between 0.5 and 0.8 instead of just 0.67 is already a result.