EViews.com

Hi,
My data is drawn from a randomised experiment, with the unit of randomisation being the village. My observations are taken at the household level. - the household is my cross-sectional unit. I wish to use standard errors that are robust to covariance in errors within the village. Any tips of how to do this in Eviews 8?

Any help would be greatly appreciated! Thanks.

Create an undated panel with village as the cross-section identifier. Then cluster robust standard errors are available in the least squares options dialog by selecting White-period.

Hello,
I have got a similiar problem. I also have data from a field experiment and I also want to use clustered robust standard errors. The standard errors are clustered and robust. The problem is, that I also need to weight the regressions to correct for oversampling. I did the weighting with the normal Options button in Equation Estimation. To get clustered robust standard errors I changed the workfile structure to undated panel, I selected an identification series and as coefficient covariance method I selected White period. The problem is that I can´t weight the regression anymore with this method. So do you have an idea how I can weight the regression and get clustered robust standard errors with one equation?
Thanks a lot in advance
Fran