Hi,
I'm in the middle of doing my research report, and am a bit confused with normality test. Since I have very limited knowledge and experience in running SEM AMOS, is that necessary to run normality test when you are analysing data in AMOS? If so, how can I do that?
I have also read that survey items (in Likert Scale) does not require normality test. Does that same assumption apply? thank you.
Usual statistical analysis is based on the assumption that the sample has been drawn from a population that follow normal distribution.
Depending upon the objective, the method of analysis varies.
If the sample is sufficiently large, the normality test is not essential.
Yes you need. But there are some estimator which are more robust even if distributions are not normal. With Mplus, see MLR estimator:
MLR – maximum likelihood parameter estimates with standard
errors and a chi-square test statistic (when applicable) that are robust
to non-normality and non-independence of observations when used
with TYPE=COMPLEX. The MLR standard errors are computed
using a sandwich estimator. The MLR chi-square test statistic is
asymptotically equivalent to the Yuan-Bentler T2* test statistic.
Look at: https://www.statmodel.com/download/usersguide/Mplus%20Users%20Guide%20v6.pdf
Best regards
Thanks to Claude Houssemand for his answer.
He has given a more accurate answer.
Appreciation to him.
Factor in this: there is normality for individual variables and multivariate normality. I often have normal variables but higher multivariate deviations from normality. Also there is the issue, mentioned above, about robustness.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Structural Equation Modeling