Bruno (2005) found that in the case of small N, the bias-corrected
LSDV (LSDVC) emerged as the preferred estimator in dynamic panel data fixed effects models.
LSDV [R.sup.2] is the measure of fit of the model with binary variable, while "within" [R.sup.2] - the measure of goodness of fit of the model for [[??].sub.it], i.e.
In such a scenario it is more appropriate to use least square dummy variables (
LSDV) technique or the fixed effect model.
The resulting equations are estimated using the
LSDV method without including cross-sectional means and without correcting the finite sample bias of estimates.
Columns (4) and (5) in Table IV present the results from the pooled OLS and
LSDV estimation, respectively.
However, conventional serological assays could not distinguish SPPV, GTPV and
LSDV due to the close antigenic and virulence relationship (Balinsky et al., 2008).
This implementation of the fixed effects model is also known as the
LSDV regression.
Taylor [22] found the feasible GLS estimator more efficient than least squares dummy variable (
LSDV) estimator for all but the fewest degrees of freedom and its variance never more than 17% above Cramer Rao bound.
Moreover, the OLS with year and province dummies will represent the Least squares dummy variable (
LSDV) estimator i.e.
Also, since there are time invariant variables in the model, and the cross-section or panel groups are few, the Least Square Dummy Variable (
LSDV) estimator is more appropriate for this study.