Reply With QuoteAh, would it be appropriate to use Spearman's rank correlation?
Hey there,
I have a problem in that I'm attempting to test a relationship between two variables, but each sample from one of those variables is most likely drawn from a different population. I think this violates several assumptions of standard analysis, including single population, statistical independance and normal distribution.
I'll explain my data. My data comes from a borehole into the earth's crust, down to a depth of some 400m. This is the first variable - depth (Y). The second variable is the chemical makeup of the rocks at a depth (X). The problem is that X most likely varies systematically with Y due to fluid moving through the rock. This means that any value of X is dependant on the value of X immediately below it. This means (and be aware my grasp of statistics is poor) that each value of X is not "statistically independent". In other words, each value of X is drawn from a different population.
Aswell as this, the samples (X) were taken at arbitrary spacings at any depth, so intuitively the average is meaningless, and the data has no reason to be normal (even if it appears to be so).
So, my question is, how do I test that the X varies systematically with Y?
Any help is appreciated!
Ah, would it be appropriate to use Spearman's rank correlation?
Posting Permissions
|
|