its often said that +-2 is the acceptable range, I have got a 3.24 value for skewness, does it mean that the distribution is non-normal? Do we have some references that say that 3.24 can be considered normal? thank you!
I found information from the link below will help you make a good decision.
https://imaging.mrc-cbu.cam.ac.uk/statswiki/FAQ/Simon
Kline writes in his book "principles and practices of structural equation modelling" on page 77 that if the data is 200+, skewness and kurtosis calculation is not required at all but in case of less sample size, he suggested skewness value as ≤ 3.0 and kurtosis as≤ 10.0. Based on his assumption, data seems non-normal if sample size is less than 200.
thank you for the response! The sample size is more than 200, so it's fine, and thus no need to report skewness and kurtosis!
Data normalization is an complex topic with many approaches (e.g. log, square root, Box-Cox transforms), but I believe presenting the raw data, while noting any transformations, show that researcher has found its understanding.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Graphs