- For Teachers
Is this correct?
The null hypothesis implies to assume that the ratios of the taxon frequencies between the different co-ocurrence categories are equal for all the pairs of distribution units.
You've got (1) ratio of the taxon frequencies , (2) co-ocurrence categories and (3) pairs of distribution units.
I can't see how either (2) or (3) comes into the null hypothesis? Just what statistical test are you performing? 'ratio' means a non-parametric like Binomial, yet you have too many categories, and too many for a Chi-square, so you seem to have a one-factor ANOVA going on with non-continuous data!
Can you tell me more about the data?
Firstly, though, I can mention that a null hypothesis doesn't 'imply' since the null hypothesis is specifically formulated to formally postulate 'no difference', or 'equal means' etc. In effect, with an experiment, you are trying to prove that all sheep are white. This is impossible, since you need to examine all sheep in the Universe, and still be left with doubt in case one is skulking on outer Andromeda somewhere. Hence, the null hypothesis is formulated specifically asserting 'all sheep are black'. You're experiment is then to show your results are so unusual you have found an exception, a 'white sheep', and so you can reject the null in favour of the alternative (experimental) hypothesis.