Zero Result

[Enes Coten]

When I apply the Jarque bera test or other tests, the result is always zero. What is the reason?

[Maarten Buis]

Your variable is just not normal. Have you look at a graph of that variable to see what might be going on?

[Enes Coten]

I've been trying to learn this program for a few days. Here is my data, there is no dummy variable in my test.

[Nick Cox]

qnorm hata

would be immensely more interesting and useful.

[Enes Coten]

here

[Nick Cox]

So, the distribution is indeed a long way from normal. Your residuals vary from about -3 million to about 15 million. That is suggestive to me of a need to work on logarithmic scale, but more information about your outcome and predictors is needed for good advice to be possible. Other approaches might be preferable.

[Nick Cox]

My guess is that your outcome is fiyat or flat or apartment price. That is always positive but almost inevitably highly skewed and fitting a line or plane or hyperplane won’t work well, Logarithm of that will work better in plain regression. Or use a generalized linear model with logarithm link, or equivalently Poisson regression with vce(robust) even though prices are not counted,

[Enes Coten]

I got at least non-zero values logarithmic, thanks. I will look into your other advice.