Acceptance Sampling Details

Acceptance sampling is a quality control process used when deciding to accept or reject a batch of products or services received from a supplier. When testing each product or service is expensive, time-consuming, or damaging, acceptance sampling offers a statistical, more feasible alternative for determining acceptability, saving time, energy, and cost.

In acceptance sampling, the consumer must decide its Acceptance Quality Level, or AQL, and its Rejection Quality Level, or RQL. The AQL is also known as the Alpha or Producer’s Risk, and is the highest defect rate that a consumer finds acceptable. The RQL, equivalent to the Beta or Consumer’s Risk, represents what the sampling plan will reject.

Acceptance sampling may be performed after the batch of products has been delivered or while they're still in the production plant. Either way, the method is not intended as a measure of quality but rather as a guide for the consumer when deciding to accept the products or not. Generally, if the number of defects is lower than the previously determined acceptable number, the lot will be accepted.

Acceptance Sampling Real World Example

A food manufacturing business has just received its order of 10,000 pieces of vacuum bags. It would be impractical for them to check each item for quality as this would take an unreasonable amount of time and energy. On the other hand, they cannot simply accept the products, not knowing how many of them are in good, usable condition.

As a solution, they decide to use the acceptance sampling method. They start by determining the sample size, or the number of vacuum bags that would be inspected, and the acceptance number, which is the acceptable number of defective products in the batch.

If the AQL is 1.5% and the RQL is 5.0%, alpha would be 0.5 and beta would be 0.1. Based on these figures, They should inspect 209 vacuum bags from the batch of 10,000 received. If the number of defective pieces is 6 or lower, they should accept the shipment. If the number of defective pieces is 7 or higher, they should reject the entire lot.

Significance of Acceptance Sampling

As mentioned, inspecting a full batch of products or services requires a huge amount of time and labor, constituting what is considered as a “deadweight cost.” It can even destroy the products in the testing process. Acceptance sampling offers a more realistic alternative that still gets the job done.

Humans as they are, inspectors are likely to grow bored and tired while checking an entire batch one by one, causing lapses in judgment as to the products’ acceptability. Again, this can be costly to the consumer as defective products are classified as acceptable and vice-versa. With the statistical option offered by acceptance sampling, every aspect of the process benefits from automation, especially finances and human resources.

Still, acceptance sampling is not a perfect process, and the possibility of making inaccurate conclusions remains. The method's success is also affected by such issues as sampling randomness, batch size, the testing attributes, and the overall acceptance criteria. Moreover, since accepting or rejecting an entire lot of products or services depends on the samples' results, there is a chance that those samples are not truly representative of the entire batch.

Two Types of Acceptance Sampling

The goal of acceptance sampling is consistent, but it can come in any of two types or forms. One is based on the product or service's straightforward attributes, and the other is based on variables.

In attribute sampling, data comes in the form of attributes, and the question solely rests on whether the attributes are present or not, regardless of the magnitude of the error. For example, in a fabric shipment, the presence of runs in any fabric roll would mean rejection of that roll. If there are no runs, no other factors will be considered, and the product would be accepted.

In variable sampling, data comes in variables, and the results are scaled according to how they meet predetermined requirements. For example, samples from a batch of body lotions will be assessed according to their toxicity levels. The products will be chemically tested, and if they meet the consumer’s specifications, they will be accepted.

History of Acceptance Sampling

The beginnings of industrial acceptance sampling go back to the early 1940s when the U.S. military used it to test bullets during the Second World War. Harold Dodge, a Bell Laboratories employee and then-Secretary of War consultant, developed the concept as part of the company’s quality assurance system.

Bullets had to be tested before the supplier accepted them but to test them meant to destroy them, and it was too risky to accept the bullets without testing their quality. To solve the problem, Dodge created a random sampling plan that would be used as a guideline, defining a set of criteria, such as sample size and a maximum number of tolerable defects.

Acceptance sampling rose to popularity during the war and was adopted by other industries in the years that followed. Then in 1969, Dodge pointed out the difference between acceptance sampling, which is no more than an immediate spot-checking of samples to assess lot acceptability and quality control, which is a broader, and independent part of the manufacturing process.