In the determination of lead in a paint sample, it is known that the sampling variance is 10 ppm while the measurement variance is 4 ppm. Two different sampling schemes are under consideration:
Scheme a: Take five sample increments and blend them. Perform a duplicate analysis of the blended sample.
Scheme b: Take three sample increments, and perform a duplicate analysis on each.
Which sampling scheme, if any, should have the lower variance of the mean?