421
Process Capability: Engineering and Statistical Issues
2 N = 1000
1 Observalions are dwe the bst class
Midpoinl |
Couit |
0.9 |
0 |
1.0 |
6 |
1.1 |
49 |
1.1 |
Ml |
U |
211 |
1.4 |
207 |
15 |
158 |
1.6 |
106 |
1.7 |
72 |
15 |
29 |
19 |
X) |
2.0 |
4 |
2.0 |
2 |
22 |
2 |
21 |
2 |
1
30
60
90
120
150
210
Figurę 5. Histogram of Samples from a Stable Process with Cp = 1.33.
should be reported. The interested reader should consult Kushler and Hurley (1992) and Franklin and Wasserman (1992).
The width of the confidence interval on the process capability index can be controlled by sample size decisions. The larger the sample size, the narrower the width of the confidence interval. In the Monte Carlo study above, sample sizes of n = 30 were used. A 90% approximate interval on Cp can be
found by using the lower and upper 5"* percentiles of the calculated indices.