Hello everyone,
I'm having hard time using Generalized Linear Models in hypothesis testing in community ecology. I'm trying to figure out whether a certain treatment has resulted in higher counts of individuals. In the attached file, you will see four treatment results of unequal sample sizes. I want to see whether "Treatment 1a" has attracted more individuals than "Treatment 1b" and, similarly, whether "Treatment 2a" has attracted more individuals than "Treatment 2b" (1a and 1b are compared - 2a and 2b are compared separately). Or I can simply say I want to test the hypothesis "Treatments 1b and 2b have attracted a higher mean number of individuals than their counterparts". Can anybody tell me how to do this hypothesis testing using R and Generalized Linear Models step by step? You can save your answers in the form of a script and just attach it to your answer. Thanks in advance.
If you want to compare just two treatments, then you can use a t-test, which obviously would have equal sample sizes. But if you want to compare 'counterparts', you should make one data table and use a GLM/ANOVA and post-hoc tests, which take into account all possible interactions. Differences in sample sizes won't be an issue in the GLM unless the differences are very big, in which case the variables (treatments) themselves can become correlated and variance homogeneity may be affected. You can also use non-parametric tests or PERMANOVA, which don't require such strict assumptions. But I don't have an R script though :)
You need special techniques for count data. See the links below:
https://www.google.com/search?q=regression+on+counts&ie=utf-8&oe=utf-8&client=firefox-b-1
https://www.google.com/search?client=firefox-b-1&ei=QbZVW9bCHpjCjgT31baQDQ&q=regression+on+counts+R&oq=regression+on+counts+R&gs_l=psy-ab.12..33i22i29i30k1l4.157574.159121.0.162235.2.2.0.0.0.0.123.215.1j1.2.0....0...1c.1.64.psy-ab..0.1.123....0.bGIsuHavNUI
For sure there are methods in R to handle this as the second link shows.
Best wishes, David Booth
Hello
Please refer to,
Michael J. Crawley(2008). Statistics: An Introduction using R. Wiley.
Regards,
Zuhair
I also did similar test We can have some discussions if you are OK. [email protected]
I agree with Andrew Paul McKenzie Pegman.
Thanks to him for his valuable answer with elaborate description.
The link for resource given by David Eugene Booth will be helpful Thanks to him.
The reference given by Z. A. Al-Hemyari will benefit Kaan Özgenc .
Thanks to Z. A. Al-Hemyari
and best wishes to Kaan Özgenc.
Thanks to Balima Larba Hubert fir his informative answer. Best wishes to Kaan Özgenc.
Generalized Linear Model (GLM) in R with Example - Guru99
https://www.guru99.com/r-generalized-linear-model.html
@ Kaan Özgencil ,
This is a problem of analysis of variance of one way classified data having unequal number of observations in the treatment groups.
Accordingly, go for performing the said analysis.
- Badges
- Science method
- Similar topics
- Biological Science
- Ecology