A METHOD OF SAMPLING INSPECTION 615 



to the finished product. The distribution of inspection activities 

 throughout any process must be so ordered that the net cost of pro- 

 duction will be consistent with the quality demanded by the customer. 

 To determine whether an inspection should be made, or how much 

 should be made at any one of the formative stages, is a major problem 

 involving questions of both quality and economy. 



One hundred per cent inspection is often uneconomical at a point 

 in the production process where inspection is clearly warranted, par- 

 ticularly when preceded or followed by other inspections, inasmuch as 

 the cost of more inspection at that point may not be reflected by a 

 corresponding increase in the value of the finished product. In 

 special cases, for example where inspection is destructive, 100 per cent 

 inspection may be totally impracticable. Sampling inspections are 

 often best from the standpoint of both the producer and the consumer 

 when the value of quality and the cost of quality are weighed in the 

 balance. 



To arrive at an answer to the question, How Much Inspection, it is 

 first essential to define clearly just what the inspection is intended to 

 accomplish and to weigh all of the important factors both preceding 

 and following the inspection in question which have a direct influence 

 on the quality of the finished product and which as a whole determine 

 how large a part this inspection step must play in controlling quality. 

 Should it serve as an agency for making sure that the product at this 

 stage conforms 100 per cent with the requirements for the features 

 inspected? If so, 100 per cent inspection is required. Or should it 

 serve to make reasonably certain that the quality passing to the next 

 stage is such that no extraordinary effort would have to be expended 

 on defective material? If so, sampling inspection may be employed. 

 Ahead of all else, decisions are needed as to the specific requirements 

 that must be satisfied by the inspection plan itself. This part of the 

 problem is a practical one — one which must be approached in the light 

 of experience, knowledge of conditions and the statistics of past per- 

 formance. Once the basic requirements of the plan are agreed upon, 

 probability theory can assist in formulating the details which will 

 accomplish the desired results. It is important to hold in mind that 

 statistical methods are aids to engineering judgment and not a sub- 

 stitute for it. 



An attempt has been made in Fig. 1 to show schematically some of 

 the outstanding general considerations which must be taken into 

 account in establishing a proper setting for any problem that seeks 

 to determine the economical amount of inspection. Inspections vary 

 widely in purpose, type and character. While their broad purpose is 



