st: to check the significance of the coefficient

[Kit Baum <baum@bc.edu>]

Keynes said

I run pooled OLS by using stata and I want to check the significance
of the coefficient.

I'm confuse about the P-value appear in the result.

Do I have to divide that P-value by 2 before I check the significance
of the coefficient?

Moreover, how can I check for autocorrelation in this pooled OLS model?


The p-value given is for the null hypothesis that the coefficient  
equals zero, i.e. that of a two-tailed t-test.

If this 'pooled OLS' indicates that you are using panel data, I would  
first of all worry about the possibility of unobserved heterogeneity.  
Have you tried estimating the model with fixed effects (xtreg, fe)?  
If the F-test for individual effects is significant then these pooled  
OLS results are not consistent.

In the case of fixed effects, you could use xtgls and allow for  
various forms of AR(1). You could also estimate the model with Mark  
Schaffer's xtivreg2 (sic; it estimates OLS fixed effects, not just  
IV) and use HAC ("Newey-West")  standard errors, which would correct  
for heteroskedasticity and autocorrelation of unknown form.

Kit

Kit Baum, Boston College Economics
http://ideas.repec.org/e/pba1.html
An Introduction to Modern Econometrics Using Stata:
http://www.stata-press.com/books/imeus.html


*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/