I am about to begin the data analysis part of my study, and I am confused that I need to do transformation before analysis or not. So, please, tell me what the situations in which we need to do the transformation. One more thing, please tell me how to do it.
Transformation aims to try to meet the hypothesis in a study made by the statistical tests or procedures. We do transformation when the assumptions aren’t met, and the standard errors are biased. All this will lead to incorrect conclusions, but transformation can improve it. If you wish to learn how to do the transformation, then, Tukey’s ladder could be one of the useful keywords. A great mathematician named John Tukey devised an ordered list of transformations.
Transformations could be done in the following situations: If the data in your study doesn’t fit in a normal distribution curve or if the variance of the data in your study is not homogeneous.
Transformation can be done by using log, square root or arcsine transformation. It is suggested to do the transformation in ms excel.