Thread: Non parametric multivariate analysis. maybe :(

1. Non parametric multivariate analysis. maybe :(

Hello to everyone,
i need some help with my statistical analysis.
I have two different pharmacokinetic parameters. They result from two different models applied on the same data. These data are baseline and follow up data of 20 subjects.
To sum up, for each subject, i have 4 measures:
- two (baseline and follow up) of the first parameter and
- the other two for the second parameter.
Distributions are not normal and not omogeneous (Kolomogorv-Smirnov and Levene test already done).
I would like to find a correlation between the trend of these two parameters. As to find a statisfical meaning if, for example, the first param increase in the follow ups and the second decreases.
I hope i have been clear enough
Thank you.

2. Re: Non parametric multivariate analysis. maybe :(

I would like to find a correlation between the trend of these two parameters.
So you could calculate the difference between baseline and follow-up for each parameter, and correlate the differences.

With kind regards

K.

3. Re: Non parametric multivariate analysis. maybe :(

The differences or the percentage variation?

Thanks a lot.

4. Re: Non parametric multivariate analysis. maybe :(

Well, how should I know? You did not describe the variables in detail.
Percentages make sense only for rational scales, but even then simple
differences could be preferable.

With kind regards

K.

5. Re: Non parametric multivariate analysis. maybe :(

Originally Posted by Karabiner
Well, how should I know? You did not describe the variables in detail.
Percentages make sense only for rational scales, but even then simple
differences could be preferable.

With kind regards

K.
You are right. I have applied to my MR data two different pharmacokinetic models and my data are the parameters of these models that measure the perfusion of the tissue under examination (ktrans, kep, i don't know if you are familiar with that). Furthermore i have two parameters independent of the models (SUV and ADC). I have the first and the follow up scan (after treatment) for each patient and for each scan all these parameters. I already know that SUV and ADC significantly varies in the follow up scan in response to the treatment. I would like to understand if it is the same for my pharmacokinetic parameters.
Thank you.
Kind regars,
M.

 Tweet