Assumptions for comparing means and which test to use

[cjhey]

I am attempting to compare the means of soil moisture measured from two adjacent soil profiles at both 10cm and 30cm. Soil measurements were measured discretely six times over the course of two months at the same date and time for each profile. However, when looking to perform tests to compare the means, I come across two problems: 1) is it correct to assume that the samples aren't independent because they are both influenced by precipitation in the same way as a function of time? 2) The sample sizes are small (n=6 for each depth in each profile), and they aren't normally distributed. Is there an appropriate test I can use to compare the means based on these qualifications?

 

[rogojel]

hi,
I would go for a two-sample test, pairing each observation from one profile with the corresponding observation from the other. This could either be a t-test or a non-parametric test like Wilcoxon. However, your power will be probably low due to the small sample size.

regards