I apologize in advance in case this is a silly questions but I could not find enlightening clues neither on Stata forum nor on internet. I have a panel data consisting of 59 cross-sectional units (countries) and 10 time-units (year) and I am trying to assess whether my residuals are normally distributed. This is the code I run:
Code:
. jb residuals Jarque-Bera normality test: 1.377 Chi(2) .5022 Jarque-Bera test for Ho: normality: . . xtsktest, reps(500) seed(123) (running _xtsktest_calculations on estimation sample) Tests for skewness and kurtosis Number of obs = 465 Replications = 479 (Replications based on 59 clusters in country) ------------------------------------------------------------------------------ | Observed Bootstrap Normal-based | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- Skewness_e | 2.276566 11.81486 0.19 0.847 -20.88013 25.43326 Kurtosis_e | 353.2939 232.8903 1.52 0.129 -103.1627 809.7504 Skewness_u | 355.1194 112.5327 3.16 0.002 134.5594 575.6794 Kurtosis_u | 3539.433 1691.43 2.09 0.036 224.2903 6854.576 ------------------------------------------------------------------------------ Note: One or more parameters could not be estimated in 21 bootstrap replicates; standard-error estimates include only complete replications. Joint test for Normality on e: chi2(2) = 2.34 Prob > chi2 = 0.3106 Joint test for Normality on u: chi2(2) = 14.34 Prob > chi2 = 0.0008
Thank you in advance for your time
Best regards
0 Response to Conflicting results - xtsktest and jb -
Post a Comment