The sample size that accounts for the statistical efficiency loss introduced by weighting; always less than or equal to the nominal sample size when weights are applied.