Sunday, January 01, 2006

Re: st: return of p-value for beta after regress command

At 04:10 PM 1/1/2006, Colin Cooke wrote: >I am interested in looking at the p-value for each beta after running >a series of regression commands. I have a program that performs >regression after each "leave one out" iteration in order to determine >how infuential outliers effect the beta coefficients and p-values for >my model. Unfortunately, I am unable to determine how to get stata to >return the p-value for each beta coefficient after a -regress- >command. -ereturn list- does not appear to have the beta p-values as >an option. Can anyone help? Thanks.

Here is a quick and dirty solution. You'll need to modify the calculation of p if the sample is too small to justify a z approximation, i.e. use the t distribution instead.

* Run the regression first, then run this mat b = e(b) mat v = e(V) local k = colsof(b) mat z = J(1, `k', 0) mat p = J(1, `k', 0)

forval j = 1/`k' { local z = b[1,`j'] / sqrt( v[`j', `j'] ) local p = 2* (1 - normal(abs(`z' )) ) mat z[1,`j'] = `z' mat p[1,`j'] = `p' } mat list z mat list p

------------------------------------------- Richard Williams, Notre Dame Dept of Sociology OFFICE: (574)631-6668, (574)631-6463 FAX: (574)288-4373 HOME: (574)289-5227 EMAIL: Richard.A.Williams.5@ND.Edu WWW (personal): WWW (department):

* * For searches and help try: * * *


Links to this post:

Create a Link

<< Home

This page is powered by Blogger. Isn't yours?