This video demonstrates how to use the Kolmogorov-Smirnov test (KS test) to evaluate the normality of a dependent variable using Microsoft Excel. The Kolmogorov-Smirnov test results are compared to the results from SPSS.
Very useful indeed. Thank you very much. So clear, direct, to the point (with a little bit of extra information (e.g. on absolute cell referencing), no irritating music in the background. This is EXACTLY how instructional videos should be made.
This video is really soo great. The only critical thing is the critical value!!! It is quite easy to pick the wrong table. KS is designed for having the true parameters, but we use sample estimates for mean and sd. This leads to the change in critical value of app. 25%. What you need is the extended KS test by Lilliefor. This KL test needs distribution dependent tables. So for testing e.g. for exponential distribution the critical values would differ. I prefer the simpler Jarque-Bera test, it needs skew abd kurtosis, which are also available in xls, and much easier to calculate.
Thank you for this video and all the others of statistics! Yours videos are very good! I would appreciate very much if you did a video explaining how to do Shapiro-Wilk test on Excel
I love you sir for such a wonderful tutorial on Kolmogorov-Smirnov Test, I am able now to analyze both in SPSS as well as in excel but if i ask that we why need to check Normality ?
Great video doc. I have a question for u, on the score column is that a value from standardized residual/unstandardized residual/predicted value? And please explain why.
Can somebody tell me what test to use to assess normality distribution for 50 to 100 respondents? Shapiro Wilk-Test was said to be for less than 50 respondents only.
If n "large" (greater than 45 that he shows), the K-S critical Value is 1.07/sqrt(N) for alpha 0.2, 1.22/sqrt(N) for alpha 0.1, 1.36/sqrt(N) for alpha 0.05, 1.52/sqrt(N) for alpha 0.02, 1.63/sqrt(N) for alpha 0.01.