Revision Date: 2005-09-06
The calculations are based on repeated measurements of a part (or standard) by an operator. The mean and standard deviation for the data are determined and a 95% confidence interval is constructed by displaying a plus/minus value in the uncertainty box and based on the t-distribution (for small sample sizes). For example, a value of 2.97496E-03 implies that the true measurement for the part lies between: the observation - .00297496 and the observation + .00297496
If you place a known value for the part (or standard) in the reference box, a bias figure will be calculated by: Reference value - average observed value
This value will be placed in the bias box.
If you enter specifications into the specification boxes, the percentage bias in terms of the specification spread is calculated and displayed in the bias(% of Specs) box. If the bias is negative, the sign will be carried over to the percentage calculation.
If you enter process variation (normally 5.15 sigma for gage R&R but 6 sigma for capability analysis) in the Process Variation box, the percentage bias in terms of the process variation is calculated and displayed in the bias(% of Process) box. If the bias is negative, the sign will be carried over to the percentage calculation.
Gordon Constable, Ph.D. | |
