Hi everyone,
I am going to run ANOVA analysis with the post-hoc (Tukey's HSD) to see the statistical differences between 4 different methods for each of my different variables. I have tens of different variables and each of them will be ran accordingly.
I have one main method, which is going to be my baseline for the statistical results. Then, I will compare those different variables with this baseline method to see if there are any other variables have giving the similar statistical results or not? Then, I would like to rank these different variables to find the best matching one with my baseline method. Do you know how I can rank these different variables by using "widely accepted" method? Is there any way I can use point ranking system or not?
Thanks.
If you are using anova and Tukey HSD, it makes sense to rank the methods by similarity in means. (Because that's what these tests look at). However, if you want to judge the similarity of methods, simply comparing their means may not be the best way. .... If the results can be paired --- like Method 1 result for Sample A and Method 2 result for Sample A ---- you might start by plotting the two Methods and see if they match well over the range of Samples. To quantify this, you might use something like RMSE or MAE, but if you can explain more of what you are trying to do --- and maybe include an example plot --- someone here might be able to give you better advice.
- Badges
- Science topic
- Similar topics
- Biological Science
- Zoology
- Ornithology