1. What is the purpose of statistical quality control?
2. Is it important how Mary Witcher selected the two files that she “spot checked” in Alicia’s filing system?

It seems to make sense that if a firm is going to implement a program of quality control, then it would want to check its products carefully to see that each lived up to whatever standard of quality had been established. This careful scrutiny of every unit of the product is ideal, but in some cases, it is impossible. A firm may be producing so many units of a product that it is not cost efficient to check each item.

The textbook points out that a firm using destructive testing cannot test each item. Take an arsenal, for instance. If the firm tests each round of 50-caliber machine gun ammunition manufactured at the arsenal, then there will be nothing left to ship out to the armed forces.

One way to avoid checking each item is through the use of statistical quality control. In this method, quality control personnel take periodic samples from the production process and test them. The theory dictates that if the samples are chosen in a manner that is sufficiently scientific, then statistical sampling is just as efficient as inspecting the entire population of units of a product.

An example of sampling is illustrated by television surveys of American viewers. A few thousand monitors on TV sets scattered across the nation enable rating services to claim that millions of Americans were watching a particular show at a particular time. Some of the TV sampling activity has come under criticism, and Congressional investigation committees have called in representatives of the rating services that measure America’s viewing habits. Audience surveys are a very serious business when you consider that low-rated networks can lose millions of dollars in advertising revenue. Thus, the sampling techniques of the rating services are very important.

If you are a skeptic about sampling, then you will enjoy the anecdote about the 1948 presidential election. All the learned pollsters, with their scientific sampling methods, picked Thomas Dewey to win the election. As it turned out, Harry Truman won the election. The question remains whether statistical quality control is as safe as 100 percent inspection. Some years ago at a quality control (QC) conference, an experienced QC practitioner defended sampling by telling this story:

One of the QC veteran’s earliest jobs in the field was in a brewery. For several hours at a stretch, he was to look through a magnifying glass at empty beer bottles coming down an assembly line. The bottles themselves were big, and under a magnifying glass they were gigantic. He checked every bottle that went by—every bottle; 100 percent inspection. As a prank, a colleague placed a large wooden clothespin in one of the empty bottles. When the bottle with the pin inside went through the magnifying-glass inspection location, the QC inspector missed it. The point the QC veteran was making was that with 100 percent inspection, concentration and efficiency can still be lost.