Panel Estimation

[rjdroozendaal]

I'm doing a panel estimation in Eviews 7.

A normal least squares estimation.
With as effect specification: cross-section fixed
And as weights GLS: cross-section weights (This increases the R squared value significantly)
The problem is I don't know what will be the right coefficient variance method with this panel estimation. I've read the user manual but it's still not clear to me.

What would be the most appropriate coefficient variance method?
- Ordinary
- White Cross-section
- White period
- White Diagonal

With white cross-section I observe the highest significance across coefficients, can somebody explain the thought behind this?
Thanks in advance.

With kind regards, Rudolf