Dear list,

I have read that while two-stage least squares and other instrumental variables estimators are consistent, they are not unbiased in small samples.  I wonder how large should the sample be in order to safely use two-stages least squares/instrumental variable estimation.  Is there a cutoff N?

Thanks
Carola


      

**********************************************************
             Political Methodology E-Mail List
   Editors: Melanie Goodrich, <[log in to unmask]>
            Delia Bailey, <[log in to unmask]>
**********************************************************
        Send messages to [log in to unmask]
  To join the list, cancel your subscription, or modify
           your subscription settings visit:

          http://polmeth.wustl.edu/polmeth.php

**********************************************************