Is it possible to do a variation of a root data transformation that is below
x^(1/2)? Are there rules against this - for example: do they need to be whole numbers? Can anyone provide literature on this?
Context of transformation: Transforming to exhibit a normal distribution before applying a specific statistical test.