What is the appropriate test for testing the difference between 3 variables?

[zeydabadi]

Hi all,

I have a group of n subjects. Each subject has 3 measurements x, y, and z.

1-What kind of statistical analysis I should perform to see if those three measurements are significantly different or not? (units are the same for all variables)

2- I can tell by visual inspection that mean(x)>mean(y)>mean(z). How can I statistically back up this observation?

Thank you!

 

[Karabiner]

What kind of statistical analysis I should perform to see if those three measurements are significantly different or not?

It depends on the scale of the measurements. And, possibly, on your sample size.

I can tell by visual inspection that mean(x)>mean(y)>mean(z). How can I statistically back up this observation?

So you have an interval scaled variable? You could perform
a repeated-measures ANOVA if sample size is large enough
(n > 20 or so).

With kind regards

K.

 

[zeydabadi]

Hi,
Sorry for not being precise.
The sample size n=25.
measurements are distances along x,y, and z-axis all in mm.

Does this make any difference in what you advised?

Thank you!

It depends on the scale of the measurements. And, possibly, on your sample size.


So you have an interval scaled variable? You could perform
a repeated-measures ANOVA if sample size is large enough
(n > 20 or so).

With kind regards

K.

 

[Karabiner]

I suppose that repeated-measures ANOVA can be used,
but you have to check its assumptions. Alternatively, as
a more robust method, Friedman test and (for pairwise
comparisons) Wilcoxon's signed rank test could be used.
Friedman and Wilcoxon do not test differences of means,
though.

With kind regards

K.