normally distributed

[tarox]

Dear all

i have returns data for two indices ,two kind of mutual funds and free risk from Jan 2003 - Jul 2014.... i did the descriptive statistics and the result for Jarque Bera for all data less than 0.05 and i transformed the data to Ln but still the same problem but the other result like Skewness and Kurtosis
show better result than normal data.....my equations are

1- is it ok for my data to be not normally distributed ?
2- how could i make Two-sample t-tests for a difference in mean between mutual funds and indices in eviews ?

thank you

[Carlo Lazzaro]

Tarox:
- you can visually inspect your data distribution and check whether or not it is (or approaches) a normal one. However, as per the details you give, this would not seem the case.
- should your raw data be not normally distributed you can consider performing a bootstrap ttest on untransformed raw data.

Kind regards,
Carlo

[tarox]

thank you very much Mr carlo