My independent variable is positively skewed. To compensate for this in my lavaan path model I used the MLR estimator. However, I am now trying to make a scatterplot in R using this variable. There is a lot of criticism towards transforming variables e.g., sqrt or log. So is there another way to account for this when graphing the data in R?
Erinn L Acland The general aversion to transforming data is unfounded. In some cases, it makes sense to transform data, but this depends on the type of data, certain characteristics of the data, the type of transformation, and the goal of the researcher. Without knowing what kind of data it is, it's impossible to know if a transformation makes sense.
"The way we measure data may not necessarily reflect the way we should use the data in a particular analysis" (McElreath, 2014).
See these links:
https://influentialpoints.com/Critiques/displaying_distributions_using_R.htm
https://www.researchgate.net/post/How_do_you_transform_a_non-normal_set_of_data_into_a_normal_distribution
https://stats.stackexchange.com/questions/185175/how-to-test-the-linearity-between-two-non-normal-distributed-variables
https://bookdown.org/ndphillips/YaRrr/regression-on-non-normal-data-with-glm.html