Page 1 of 1

Panel Data with Random Effect

Posted: Tue Jul 10, 2018 9:33 am
by CCL242
Good Day,

I can trying to detect and correct heteroskedasticity, multicollinearity and autocorrelation in Eviews9 on a data set of 13 countries for a period from 1995 to 2015. I've ran a Hasman test thus far which resulted in a prob. of 0.077 which I assumed that a random effect model is needed. However I have very ll title knowledge about statistics or reviews. Can someone Kindly confirm the step I should take to check for the above?

Additional info: the Durbin-Watson stat is 0.361766. How would I improve this stat? As Eviews won't allow me to att AR(1) to a random effect model