Pc - Consumer's risk
In reference to sampling plans,
consumer's risk, (Pc), is the probability
that a "bad" lot will be accepted. Lots are considered to
be bad if they fail a LASP, (Lot
acceptance sampling plan). This can be a crucial
measurement when implementing six sigma.
This means the chance of a product deemed
unsatisfactory to a customer
could be accepted by Inspection. Engineers try to keep the ‘Consumer's
Risk’ as small as possible, keeping in mind the requirements of users
in each specific instance.
The Consumer's Risk of a sampling plan
can be estimated as follows… (the
Dodge and Romig ‘Sampling Inspection Tables’ defines ‘Consumer's
Risk’ tighter than it is defined here)
- Second, determine the percent defective
that the consumer wants to reject. This should be interpreted as the
point where the quality is so poor that the consumer would turn it
down a large percentage of the time.
- Third, find this value on an OC curve
scale, and use the curve to ascertain the probability of acceptance.
This indicates the risk of receiving rejects under conditions
assumed in the second step providing the product is of such poor
quality it could be actually submitted.
It is possible to calculate the ‘Consumer's
Risk’ without plotting the OC curve, however, most are generally
interested in the entire area around the percent defective. Due to this,
it is a real plus to have a completed OC curve first.
A Consumer's Risk of 10% does not mean that the consumer would have a
10% probability of getting a reject product. It would indicate products
such as that would have a 10% probability of being accepted providing
they were actually submitted to Inspection. Keep in mind that if the
process is running close it’s capability and is close to the shoulder
of the curve, then product could be as bad as 9% defective would
probably not be produced.