372 



BELL SYSTEM TECHNICAL JOURNAL 



existence. The practical situation, however, is that in the majority 

 of cases there are unknown causes of variability in the quality of a 

 product which do not belong to a constant system. This fact was dis- 

 covered very early in the development of control methods, and these 

 causes were called assignable. The question naturally arose as to 

 whether it was possible, in general, to find and eliminate causes of 

 variability which did not form a part of a constant system. Less than 

 ten years ago it seemed reasonable to assume that this could be done. 

 Today we have abundant evidence to justify this assumption. We 

 shall, therefore, adopt as our third postulate: 



Postulate 3 — Assignable causes of variation may be found and 

 eliminated. 



Hence, to secure control, the manufacturer must seek to find and 

 eliminate assignable causes. In practice, however, he has the difiiculty 

 of judging from an observed set of data, whether or not assignable 

 causes are present. A simple illustration will make this point clear. 



2. When Do Fluctuations Indicate Trouble? 



In many instances the quality of the product is measured by the 

 fraction non-conforming to engineering specifications or as we say the 

 fraction defective. Table 1 gives for a period of 12 months the ob- 



TABLE 1 



served fluctuations in this fraction for two kinds of product designated 

 here as Type A and Type B. For each month we have the sample 

 size n, the number defective Wi and the fraction p = ni/n. We can 



