Do you have a tutorial on how to check for multivariate normality using Q-Q plots with a Chi-Square distribution and residuals? I have seen mention of both a GUI method and a syntax method, but can't get it to work. I can save residuals when conducting a MANOVA, but the Q-Q plot with chi-square distribution doesn't work. SPSS says it can't make the graph due to some of the residuals having negative values…any suggestions?

I agree that both the tests employed in SPSS to check for violations of assumption of normality are really poor and I regret of having used them in the past. I then foolishly used non-parametric tests when, as you show in your other videos, the t-tests and ANOVAs are pretty robust to such violations. I found D'Agostino-Pearson test (implemented in MedCalc) to be a much better test for the purpose that experimenters usually have in mind. It takes into account the sample kurtosis and skewness.

Hi, if half of the demographic data with p value <0.05, but many outcome measures in pre-test with p value> 0.05, then i should use non-parametric or parametric test??????? THANKS 🙂

A very useful video for my dissertation, I was just wounding if you could let me know the references you have used to support the use of Shapiro-Wilk test for determining normality in data. Thank you

I like the way this guy explains but listening his dry mouth kills me!

The KS test is only valid with sample sizes of greater than 2,000

Sir, thanks a lot ,, that was really helpful 🙂

Thanks very clear and useful

Hi, great video! What reading suggests that the K-S test in no longer a valid test?

Christopher Walken!

Seriously speaking, very helpful and well laid out sir!

Which are the reasons about you said that Kolmogorov-Smirnov test is not good?

Thanks for the video. What N do you need to conduct these tests?

Do you have a tutorial on how to check for multivariate normality using Q-Q plots with a Chi-Square distribution and residuals? I have seen mention of both a GUI method and a syntax method, but can't get it to work. I can save residuals when conducting a MANOVA, but the Q-Q plot with chi-square distribution doesn't work. SPSS says it can't make the graph due to some of the residuals having negative values…any suggestions?

Distracted by the fact you kind of sound like Robin Williams..

I agree that both the tests employed in SPSS to check for violations of assumption of normality are really poor and I regret of having used them in the past. I then foolishly used non-parametric tests when, as you show in your other videos, the t-tests and ANOVAs are pretty robust to such violations.

I found D'Agostino-Pearson test (implemented in MedCalc) to be a much better test for the purpose that experimenters usually have in mind. It takes into account the sample kurtosis and skewness.

Hi, if half of the demographic data with p value <0.05, but many outcome measures in pre-test with p value> 0.05, then i should use non-parametric or parametric test???????

THANKS 🙂

Thanks.

A very useful video for my dissertation, I was just wounding if you could let me know the references you have used to support the use of Shapiro-Wilk test for determining normality in data. Thank you

This video was one of the most helpful tutorials during my dissertation process. Thank you for taking the time to help others!!!

hard to understand!