'How could I normalise data including positive, negative and zero values?
I'm looking at whether changes in the range of different UK plant species differ between different species groups. The index of range change includes positive, negative and 0 values, as some have increased and decreased in range. However, the range index data isn't normally distributed, so I can't perform an ANOVA between groups, and log and boxcox transformations won't work on negative data. Any suggestions on possible transformations, or how to proceed? I've used a Kruskal-Wallis test, but additionally the variances between groups aren't homogenous, so I don't know whether this invalidates my test. Beginner here so any help is much appreciated!
Sources
This article follows the attribution requirements of Stack Overflow and is licensed under CC BY-SA 3.0.
Source: Stack Overflow
| Solution | Source |
|---|