Operations

1st Case Analysis

ain the concept of a p-chart to a group to a group of bank tellers without a background in SPC, How would you do it?

P-chart is a kind of control chart which is used to identify defective units in the process. P-chart is appropriate when we are evaluating attribute data. Attribute data is used when measuring presence or absence of certain attribute, for instance, number of defective units, or percentage of incorrect forms. To use a p-chart, we should pick enough samples for the experiment and evaluate each sample whether it is OK or defective. Then we calculate the mean and set the upper bound and lower bound of the number of errors. With this result, we can identify whether the process is stable or not.

3. How large should each sample be for the experiment schuss and Kluck describe on page 7?

When the first-time accuracy numbers quoted at 99% was right, 300 samples were needed, because at least 3 errors should be found to chart. However, according to Exhibit 4 (page 12), the actual average accuracy during 30 weeks is about 94.27%. This is because p-bar is 0.0573. Therefore, by using this new accuracy, we can assume the size of sample that result at least 3 errors.

100 samples : 5.73 errors = x samples : 3 errors ▶ 300 / 5.73 = 52.356

So, we can decide that 53 samples are enough to chart rather than excessive 300 samples.

4. The first 12 weeks of the data in Exhibit 4 represent the diagnostic period for the Policy Extension Group. Create an appropriate control chart for the data. What are the 3-sigma controls limits for the process? In which of the subsequent weeks is the process out of control?

Results |

Solver - p-Charts |

| | | | | | |

| | | | | | |

p-bar : 0.0573

UCL(p)= + 3* = 0.0573 + 3 * 0.0573(1-0.0573)300 = 0.0976

LCL(p) = - 3* = 0.0573 – 3*0.0573(1-0.0573)300 = 0.0171

Therefore, according to a p-chart of the subsequent weeks, we can find that week 23 and week 24 is out of control.

5. Develop...