You do not need to transform your data with PERMANOVA or if the data has normal distributions :)

Thanks!

And what to use PERMANOVA or ANOSIM?

M

Dear Manuel

In general, the type of analysis used is completely dependent on the research objectives and research questions. It seems that before choosing the required type of analysis, the purpose of the research should be explained correctly. In other words, according to the extent of the concept of vegetation cover, determining the question "what component of the vegetation cover should be investigated?" It is essential.

For example, you can use different approaches such as species, functional or phylogenetic diversity, heterogeneity or plant composition, each of which requires its own analysis.

Regarding vegetation composition and its similarity between different treatments, ANOSIM analysis seems to be more useful. But if the goal is to investigate the effect of different treatments on plant composition, it is logical to use PerMANOVA analysis.

Also, most natural data such as vegetation cover data do not have normal distribution, and the use of mathematical functions such as square root or logarithm will provide better results.

See the following paper also:

Article Fire disturbance effects on plant taxonomic and functional β...

Thanks Reza.

Regarding vegetation composition, I do not clearly see why ANOSIM is used to check the similarity between treatments and PerMANOVA for investigating the effects of different treatments on plant composition. At the end, is it the same right? high similarity among treatments (ANOSIM) should be related to no effects of the treatments (PerMANOVA).

M

Dear Manuel

Regarding the difference between these two analyses, I should mention that in the ANOSIM test, the dissimilarity matrix is first generated for the community, and then comparisons are made based on the rank order of dissimilarity values. The existence of a significant difference in this test (ANOSIM) indicates that the differences between the groups is greater than the differences within the groups.

ANOSIM can only tell you if a difference exists, nothing more, and PERMANOVA can tell you the magnitude of difference. In addition, if you want to test significant interactions, you need the PERMANOVA.

Finally, it seems that the result of a single test does not provide complete information about the answers to the questions of this research and more additional analyses are needed. For example, using NMDS analysis (virtual comparison) followed by ANOSIM test (statistical comparison) provides far more complete and comprehensive results.

Thanks a lot. I highly appreciate your help