Which ANOVA? ...very stuck!

I have been looking for days now which ANOVA would be best but I am not sure which one I can use to analyse my data! Any help would be massively appreciated!

I have carried out a study on frogs within 3 forest types. I have in total 10 frog species and I have recorded the number of individuals from each species I saw within each forest type. Below is my raw data. How would it be best to analyse this? I have already used Shannon index, Simpsons index of diversity and would now like to use an ANOVA? How can I go about choosing which one? I also have SPSS program! Many thanks in advance for any help!

Species Forest 1 Forest 2 Forest 3
A 3 12 14
B 2 0 0
C 0 0 2
D 0 0 3
E 2 4 2
F 0 1 5
G 2 2 1
H 0 2 4
I 0 0 2
J 1 0 0
Hi CONFUSEDstudent1991 and Karabiner,

im with Karabiner, what do you want or need to find out?

From my point of view you might want to tell, if there is a significant difference in frog species populations concerning the forest and the species itself. This would mean to go for a Two-Way Anova (Factor 1 = Species, Factor 2 = Foresttype, Count = Dependent Variable).

With two way Anova you may only find out, that there is a signifikant difference (or not!) either between the forests and or the species itself. To get to know exactly what kind of difference ther is IN BETWEEN the Speciesgroup and/or forestgroup you might want to run Tukeys Honest Significance Test.

I tried it with R and could only find a significant difference for species. I was not able to run Tukeys HSD. Maybe Karabiner can enlighten us here, when you gave us your scientific question.


Hi Karabiner and teste1234,
Thank you both for replying! My principle question is find out whether there is a difference in frog abundance between the 3 forest types! I don't think I have enough data to additionally question whether there is a statistical difference between the species and the number of individuals of that species. Excluding this i thought a one way ANOVA would be most appropriate? And then maybe use a Turkey HSD or Duncan Post Hoc test to analyse this. These are just my thoughts and I would welcome your suggestions if you can see I am likely to end up going wrong at any point?! I really know very little about statical analysis of data!
Many thanks once more!
I have been looking for days now which ANOVA would be best

For days!??

I have been looking for years, and I don't know which anova to use!

And absolutely not which one would be best!

I start to believe that there is no "best" method (except in theoretical cases where all conditions are carefully set up, then maybe there is "best test"). There are really bad test, bad test and relatively good test. But "best test"? No, I don't think it is possible to suggest something like that.

I guess that there are many ways to crawl around in the forest and look for frogs, but is there a "best way"? If I may send that question in return?

In addition, it is also true for me what you said in the last statement:

I really know very little about statical analysis of data!
Since it is a number I thought of a Poisson distribution for the dependent variable. Thus a linear model with forest and frogs as explanatory factors. That can be estimated with a generalised linear model. I believe that they have hat in spss. Still the number is very low so, I am sorry, but I guess that it will be difficult to get precise estimates.

Don't use the Tukey HSD test or Duncans test. Duncans test is from the 1950:ies when they had no clear view about multiple inference, so the error rate is larger than the promised 5%. Tukeys hsd test (is great otherwise) but it is based on the studentized range distribution, which is based on the normal distribution, and these data are not normal. Besides when comparing among only three forest it would have been better to use the Bonferroni method, or preferably Bonferroni-Holm. But, the sample size is so small anyway so I suggest to skip all the multiple comparisons anyway.

Besides, I start to wonder: Are you infering about forests in general or was it just three forest that were investigated? (I would guess so.) You could possibly have sampled from ten times three different forests. Now it seems like the conclusion can only be about these three forests (an not forests in general).

Also about the choice of frogs species. The usual thing would be to choose a fixed set of frog species in advance. Now, maybe you have been crawling around in the forests and noted the one you happened to discover. (There is no specie with only zeros.) I am not sure of what would be the statistical consequences be of such a procedure. To be frank, I believe that most would just ignore tricky little things like this, but still the problem can be there.

Let's see if the other ones have any solutions.


TS Contributor
My principle question is find out whether there is a difference in frog abundance between the 3 forest types!
Looks like you have n=3 subjects (not factor levels) which you
want to compare with regard to frog abundance (8 measurements,
which perhaps could be aggregated, if needed). Don't know whether
statistical testing makes much sense in case of comparing individuals
(inidividual forests, that is), but it could well be.

With kind regards



TS Contributor
An Analysis of Means (ANOM) for Poisson data might be an option. However, be aware that the null hypothesis for ANOM is different that that for ANOVA.

The ANOVA Ho is that there is no effect (i.e., mean1 = mean2 = mean3 = ... = mean_n)
The ANOM Ho is that there is no difference from the Group mean (i.e., mean1 = mean_group, mean2 = mean_group, ...)