From: Stan Hilliard
Date: 30 Oct 2002
Time: 23:54:35
1) Level of confidence: the level of confidence does not come from data, it is chosen. Example: 95% confidence.
2) Natural variability: Assuming that you are performing statistically designed experiments on the same system, it seems reasonable that the residual variability would not change much from experiment to experiment. This is assuming that the variables studied in the experiments are not affecting the residual variability.