So I want to see if there is a difference between paired sample means and there is nestedness. 2 years with 3 groups (Stages) with 5 plots in each and there are 20 subplots that were sampled in these plots.
My question is 1) Is this the right non-parametric test for nested paired samples? 2) Why would the p-value = 1 in the output when I can see that there is a significant difference in the means? I thought you would fail to reject the null if the Mean falls within the Z-score distribution, which it does not do here-- do I have the wrong perspective on this?
Thank you to anyone who can help :)
I used the nestedRanksTest() from https://rdrr.io/cran/nestedRanksTest/
I set it up this way:
Ry<- nestedRanksTest(Richness ~ Year|Plot, data = Chrono, subset = Stage == "Y")
And got this output:
`Nested Ranks Test`
data: Richness by Year grouped by Plot Z = -0.3795, p-value = 1 alternative hypothesis: Z lies above bootstrapped null values null values:
0% 1% 5% 10% 25% 50% 75% 90% 95% -0.37950 -0.19600 -0.13502 -0.10450 -0.05600 -0.00100 0.05550 0.10555 0.13500 99% 100% 0.19350 0.33300
bootstrap iterations: 10000 group weights: 1 2 3 5 7 0.2 0.2 0.2 0.2 0.2
plot(Ry)
[Z-score distribution] (https://i.stack.imgur.com/5hAHZ.png)
