Ok, I’ve pretty much figured this out. My problem was that I didn’t restrict the sample to those in the relevant group before estimating the model. When I do that I get the results I expected.

 

Sorry to bother you…

 

PS

 


From: gretl-users-bounces@lists.wfu.edu [mailto:gretl-users-bounces@lists.wfu.edu] On Behalf Of Summers, Peter
Sent: Friday, March 05, 2010 10:09 AM
To: Gretl list
Subject: [Gretl-users] weird bootstrap behavior

 

Folks,

 

I’ve just noticed some very strange behavior when I try to generate studentized bootstrap confidence intervals in a model with only a constant (this is problem 5.6 in Davidson/MacKinnon’s ETM). The model is earnings data regressed on a dummy variable for a particular income group, so it’s just estimating the mean.

 

Here are the results from running a studentized bootstrap:

For the coefficient on G3 (point estimate 27973.6):

 

  Studentized 95% confidence interval = 16009.7 to 17410.2

 

Based on 999 replications, using resampled residuals

 

The graph of the sampling distribution, on the other hand, is centered roughly on the point estimate (see ‘studentized.pdf’, attached).

 

When I do a regular bootstrap confidence interval, I get

For the coefficient on G3 (point estimate 27973.6):

 

  95% confidence interval = 42857 to 44843

 

Based on 999 replications, using resampled residuals

 

With the distribution in ‘bootstrap.pdf’.

 

Things seem to work ok in a model with all three group dummies and no constant.

 

PS

 

===============================

Dr. Peter Summers

Assistant Professor

Department of Economics

Texas Tech University

===============================