Hello,
I'm doing a research about microbiology of sands. I need to know if there is a correlation between Most Probable Numbers from sands and sea. But, in some beaches I get negative results in all samples, and I can't find a way to calculate the correlation.
Minitabs message:
P-Value = *
* NOTE * All values in column are identical.
If anyone can help me to understand what can I do or how it works, I'll be very grateful.
Thanks.
Thank you, Dr. Chiou
So, Can I can conclude that there is no correlation when I get an undefined p-value? Am I wrong with that assumption?
Paul is correct. The whole concept of "correlation" can not be applied to a constant (the missing variation in one of your groups makes it indistinguishable from a "constant").
Hi Andrei,
your problem sounds more like your variables are set wrong. Check the variables for being numeric and varying (having at least two different values per variable). Also check whether your variables are truly "numeric". Your variables might have the class "character" because of some wrong import.
Hello, thank you all for your answers,
I think my variables are well determined. Because I'm looking for densities of microorganisms. But there are points where I only get the same result (number, quantity, density) in all the samplings, that's my biggest problem when I try to make a correlation.
Dear Andrei,
The correlation is between two variables. If one of them is constant, the covariance of both variables is zero, and the variance of the constant variable is zero too. For this, the correlation coefficient would be 0/0 which is an indetermination or the correlation coefficient is not defined in such case.
- Badges
- Science topic