January 2002

Vol. 4, No. 1

Quality Quiz

After working in a bakery for most of his life, Ikant Singh has taken a position with a musical supply manufacturer, with responsibilities for improving the quality of the metronomes that are produced on the assembly line. His bakery experience has taught Singh the importance of sampling in order to monitor product quality, and he brings the same concept to the metronome production process.

Singh recalls that in the bakery, it was necessary to select a certain number of samples periodically in order to assess the texture of baked goods'good or bad'and the same technique to evaluate whether other aspects of the products were acceptable or not. While some might argue that one can never be sure of quality without examining every single product that emerges from a process, Singh scoffed at that concept. 'If we'd tested every cupcake, there would be none left to sell! ' He brings the same approach to the metronome line, insisting that 50 metronome hands be measured on every shift, and that operators use a go/no-go evaluation for the external dimension of these hands.

From the ensuing data, Singh is able to create control charts to demonstrate the stability of the process. Ikant Singh brings harmony to the manufacturing process, and all is well.

One day, however, after nearly a year of sampling and charting data, he discovers inconsistency in the number of samples taken each shift by the operators. Some use sample sizes of 48, and others go as high as 53. Singh is outraged. 'How can we see if the process is in control or not, if the sample sizes are all different? ' he rants. 'Now we have to start over, and we don't know anything about the process! ' He is beside himself. His supervisor, however, assures him that in creating p-charts, any variation in the sample size that amounts to less than 25 percent of the average sample size does not need to be accounted for. It is only if the variation is greater than 25% that the validity of the data is in question. Singh insists, however, that the study is flawed and the samples must be the same size.

Your job is to determine whether Singh is correct or the supervisor, and the reason. Select from the following responses:

A) The supervisor is correct, based on Shewhart's Economic Control of Manufactured Product, page 236;

B) The supervisor's figure of 25% is generally accepted, based on a well-known simulation model that demonstrates the efficiency of this standard;

C) The 25% figure is generally accepted as a rule of thumb by statisticians;

D) Singh is correct; the entire sampling process is corrupted, and the operators should begin anew;

E) The supervisor cites a statistical operation, establishing the need to square 5 and move the decimal point two places to the left in order to determine appropriate variation in sample size;

F) Singh is correct about the need for less variation in the sampling process, but statisticians agree that only 10%, not 25%, is the allowable variation.

Click on the correct answer, and link to the solution.



Copyright 2002 PQ Systems.

Please direct questions or problems regarding this web site to the Webmaster.