Dear Emily

all depends on your data sets, what type of variable you have. It is likely that the test you propose (in nonparametric or parametric test) can answer your question.

Hi,

First you have to check your data set follows which distribution, normal or binomial or passion?  Whether your data form a normal curive, then you have to go for t-test or otherwise, you have to do non-parametric test.

For further information, kindly go through the following link:

http://www.sciencedirect.com/science/article/pii/S235198941400064X

What data do you have? Is this birds from 1974 compared to birds in 2014, or birds in one part of a nature preserve relative to another part? How many replicates do you have? Are the data quantitative or qualitative, or some of each? The more you can tell us the better our answers will become.

Dear Dr. Timothy,

I am comparing birds in a zone with high anthropogenic noise level (>60dB) with birds in a zone low anthropogenic noise (

Ok, so now I understand why you suggested the paired t-test and U test. At least one approach would be to take the difference between high noise and low noise and then test to see if this value is zero using Mann Whitney U.

How did you get to the multi-response permutation test? Are you just counting the numbers of individuals, or are there other data (nest counts, blood work, )? Is this for a single bird species, or several?

So you have three sites, and you are taking paired samples (high noise, low noise) from each site. You are doing this every month for the next 6 months, for a total of 18 paired data points for each species of bird in the study?