Statalist The Stata Listserver

[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: partial F test

From   Richard Williams <>
Subject   Re: st: partial F test
Date   Wed, 18 Oct 2006 22:14:23 -0500

At 08:49 PM 10/18/2006, Diego Bellavia wrote:
Hi Statalist,

Today in the class we approached the partial F test in
linear regression. The examples have been done in JMP
(that I do not like), but I would like to use STATA
instead. The questions are:
Your references to creating a new dataset make me suspect that jmp (whatever that is) must do things somewhat differently than Stata does.

1) Is possible to build a "parent" model (only b0)
without creating another dataset ?
Just do

regress y

2) How to perform a partial F test in STATA ?
I think what you call a partial F test I call an incremental F test. Either that or I don't understand the question. Perhaps you want something like this:

regress y x1 x2 x3 x4
test x3 x4

i.e. test the hypothesis that the effects of x3 and x4 are both 0. You could also do something like

nestreg: regress y (x1 x2) (x3 x4)

3) Is possible to temporarily exclude some observation
(potential influencing points) and run again the model
without them and without the necessity to create a new
dataset ? I know, this question is not directly
related to the partial F test, but it is critical to
me, just to quickly assess the outliers in the model.
Let's assume you have somehow created a variable called outlier which equals 1 for cases you want to exclude, 0 for cases you want to keep. Use the if qualifier on regress.

regress y x1 x2 x3 x4 if outlier == 0

For an overview of using Stata for OLS regression, see

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:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index