I need help choosing stat method!

#1
Hi!

I am hopelssly lost and I need help finding the suitable statistict to describe some resluts I have. I hope someone here can help me!

I have collected numerical data at nine set locations and also while travelling between them. I would like to show that measurements at the locations differs from measurements during transfer between them. Thus I have 17 sets of data and I want to show that they can be divided in to two groups where one show significantly higher values than the other. The data sets are not normally distributed and of different sample sizes (between 330 to 630).

What would be a good (and preferrably easy!!) way to describe this statistically?

If possible I would also like to prove that values in the higher group are twice as high as the other. Is this possible in any way?

Im very grateful for any help I can get!
Thanks
Jenny
 
#2
Can you assume that data from the nine locations are homogeneous (similar mean, shape of empirical distribution)? Also, that data from the 8 transitory locations are homogeneous? If so, then these data can be pooled to form two groups, those that are from a set location and those that are from a transitory location. Since these data are not normal, you could then use an appropriate non-parameteric test for a difference in the medians of these two groups.

If the above assumptions cannot be made, and the data appeared normal, then you could use a nested ANOVA. Since these data do not appear normal you might need to go for a non-parameteric version of a nested ANOVA. However, none of these proceedures are simple.

~Matt