Can You Calculate a P-Value for Correlation Between Non-Parametric Data?

[sobored]

I have made a correlation between group A and B and the pearsson coefficient constant, r, is at 0.97. The values are expression levels and hence they are non-parametric. I wonder how can i calculate the p-value of this correlation between group A and B? I know how to calculate the p-value for the difference between group A and B, but not for the correlation between group A and B.


Thanks for any inputs!

 

[Moonbear]

Pearson's correlation coefficient is a parametric statistic. If your data require a non-parametric test (how did you determine this?), then you can't use Pearson's correlations, but should instead be using Spearman's correlation coefficient, which is the non-parametric counterpart to Pearson's. But, nobody will be able to help you beyond that without more information about the data set you're working with. What's your sample size? Do you have continuous or integer data, random or fixed variables, etc? Basically, without knowing your experimental design, nobody can recommend an appropriate statistical test.

 

[sobored]

Pearson's correlation coefficient is a parametric statistic. If your data require a non-parametric test (how did you determine this?), then you can't use Pearson's correlations, but should instead be using Spearman's correlation coefficient, which is the non-parametric counterpart to Pearson's. But, nobody will be able to help you beyond that without more information about the data set you're working with. What's your sample size? Do you have continuous or integer data, random or fixed variables, etc? Basically, without knowing your experimental design, nobody can recommend an appropriate statistical test.

Thanks for helping me out.

Both groups have 35 samples and the values are continues. Can i use student T test to predict this p-value, assumming that i use the parametric way?

 

[EnumaElish]

Thanks for helping me out.
Both groups have 35 samples and the values are continues. Can i use student T test to predict this p-value, assumming that i use the parametric way?

I believe you can; as explained in this link or http://faculty.vassar.edu/lowry/rsig.html. (Both from a Yahoo! search on significance test for correlation coefficient.)