My area of research is in social sciences (psychology) and my sample size is 300. According to K-S and S-W test, the data is non-normal, skewness for some variables (sub-scales) are 5 and the graphs also show skewed data. Some studies say a large sample does not make a difference to skewness, but I am confused.
Should I go according to sample size and disregard the skewness, or vice-versa?
What would be a better testing approach in this situation- parametric or non-parametric?
What is a large sample?
As for my opinion is concerned, There is no straight forward answer to this question regarding large sample. It depends on once expertise. Again regarding your first quire parametric tests are preferred over non-parametric, except in some situations. I think you should go with Parametric test.
Parametric tests make assumptions about the parameters of the population distribution from which your sample is drawn. Often, this assumption is that the population data are normally distributed. Since you expect some skewness and even some Kurtosis in the distribution of your sample data, non-parametric tests are required. Non-parametric tests are "distribution-free" and, as such, can be used for non-normal variables.
Thank you for your valuable responses. I think I will go for non-paramtric in this case.
- Badges
- Science topic
- Similar topics
- Counseling Psychology
- Confusion