test of normality on large sameples - also ks or shapiro-wilk?

#1
if you have a large sample size...say 1000...do you still need to run and report tests of normality? I always thought you did but in my text book it says not to do test on skew or kurtosis if sample is large.

I did, and also mentioned in my report that i did and that it was not violated but that the test on kolmogorov was violated but was prob due to large sample.

anyways do you still run tests of normality on large samples?

Also, which test is better for large samples - shapiro-wilk or kolmogoriv-smirnov??

I thought it was KS but girl in class is saying shapiro and so is a website i just looked at.

Any help appreciated

Thank you
 
#3
Thank you so much! That is what I thought.

May I also ask....what is a possible reason as to why my KS test was violated??? I know my tests of normality may have been easily violated due to large sample size but why would KS be violated???

Thank you!!!