Table 2.

Variation in estimated batch means for different sampling regimes. The true variation is 0.15 for all simulated data sets.

	Grab Random	Grab Mid Sloppy	Grab Mid Precise	In-Line
Simulation 1	0.28	0.21	0.20	0.15
Simulation 2 1	0.28	0.22	0.22	0.15
Simulation 3	0.30	0.27	0.27	0.15
Simulation 4	0.27	0.28	0.27	0.15
Real data (CP1)		0.29 2		0.09 3
Real data (CP2)		0.28 2		0.14 3

¹ Cyclic components have same size as observed data. ² Based on data from lab database. ³ Based on in-line NIR.