Thanks for making it easier for us to understand, can you please share link to reference for those ranges as i have tried search Arbuckle, James L. (2017) but unable to find kurtosis -+10 range. I need to quote in my paper, please
what about multivariate kurtosis value? Your data is not normal in this example. What about CR value. That has to be between +-1.96 to asssess if data is normal
Thanks for the suggestion. I have reported Mardia's coefficient in the past as well. It was super popular in the 2005-2010 range but I see less of it now. It is a good test to report in the write up.
Thank for video, but i have question base on your model.... I see more 1 variable on cfa and then why you focus in skewness and kurtosis value each variablet? Should be focus on multivariate value because the model (cfa) is more one constructs Need more explaining for me :)
I don't know if I quite understand what you are asking but I'll give it a try. With SEM, it is always multiple variables. You are going to assess normality on the individual item level. Thus, you are looking at each item that makes up your construct to assess normality.
@@joelcollier9387 In fact there are 2 types of normality which are univariate normality and multivariate. In SEM case we need to assess mardais cofficient because it's the techic that provides by Amos to check the multivariate normality.
Thank you for the video. It is extremely useful. As per my understanding, there are two types of normality- univariate and multivariate for multivariate techniques like SEM. In this case, the univariate normality is being established as it is within the acceptable ranges of skewness and kurtosis. However, multivariate normality condition is violated. Can we ignore the multivariate normality based on the argument of large sample size and proceed with analysis using SEM or should be report it and proceed with bootstrapping technique for estimation? Thanks in advance.
My best advice if you are concerned about normality is to run the analysis using the bootstrap technique. You will get far less criticism than just assuming the sample size is big enough to be normally distributed.
Thank you so much for sharing this video. I have some question. What is the reference for the acceptable range of skewness and kurtosis level that is not considered as nonnornality?
![Tee Grizzley - Detroit (Feat. 42 Dugg) [Official Video]](http://i.ytimg.com/vi/aeKfTtt0x14/mqdefault.jpg)
Thanks for making it easier for us to understand, can you please share link to reference for those ranges as i have tried search Arbuckle, James L. (2017) but unable to find kurtosis -+10 range. I need to quote in my paper, please
what about multivariate kurtosis value? Your data is not normal in this example. What about CR value. That has to be between +-1.96 to asssess if data is normal
Great video, but you should also consider the total values (i.e. Mardia's coefficient) to assess non-normality.
Thanks for the suggestion. I have reported Mardia's coefficient in the past as well. It was super popular in the 2005-2010 range but I see less of it now. It is a good test to report in the write up.
Thank for video, but i have question base on your model.... I see more 1 variable on cfa and then why you focus in skewness and kurtosis value each variablet? Should be focus on multivariate value because the model (cfa) is more one constructs
Need more explaining for me :)
I don't know if I quite understand what you are asking but I'll give it a try. With SEM, it is always multiple variables. You are going to assess normality on the individual item level. Thus, you are looking at each item that makes up your construct to assess normality.
@@joelcollier9387
In fact there are 2 types of normality which are univariate normality and multivariate. In SEM case we need to assess mardais cofficient because it's the techic that provides by Amos to check the multivariate normality.
Thank you for the video. It is extremely useful.
As per my understanding, there are two types of normality- univariate and multivariate for multivariate techniques like SEM. In this case, the univariate normality is being established as it is within the acceptable ranges of skewness and kurtosis. However, multivariate normality condition is violated. Can we ignore the multivariate normality based on the argument of large sample size and proceed with analysis using SEM or should be report it and proceed with bootstrapping technique for estimation?
Thanks in advance.
My best advice if you are concerned about normality is to run the analysis using the bootstrap technique. You will get far less criticism than just assuming the sample size is big enough to be normally distributed.
@@joelcollier9387 Thanks a lot
Thank you so much for sharing this video.
I have some question. What is the reference for the acceptable range of skewness and kurtosis level that is not considered as nonnornality?
My go to reference on those ranges are: Arbuckle, James L. (2017) IBM SPSS AMOS 25 User's Guide, IBM Corporation.