UNIT 5 ACCEPTANCE SAMPLING PLANS

Structure
5.1 Introduction
Objectives
5.2 Inspection
5.3 Acceptance Sampling Plan
Advantages and Limitations of Acceptance Sampling
Types of Acceptance Sampling Plans
5.4 Implementation of Acceptance Sampling Plan for Attributes
5.5 Terms Used in Acceptance Sampling Plans
5.6 Producer’s Risk and Consumer’s Risk
5.7 Summary
5.8 Solutions/Answers

5.1 INTRODUCTION
In many situations, the product is so complex that all components/parts of the
product are not made by a manufacturer. In such cases, one or more
components of the product are purchased from an outside agent or supplier and
the manufacturer does not have direct control over the quality of such
components. Since the final product is produced by the manufacturer, he/she
faces problems such as:
 How to control the quality of components received from others?
 How to ensure that the lots produced do not contain an excessively large
proportion of defective products?
 Do the products meet the desired specifications?
Such problems belong to the category of product control.
Product control refers to control the products in such a way that these are free
from defects and conform to specifications.
You have learnt in Unit 1 that product control can be done by 100% inspection,
i.e., each and every unit produced or received from the outside suppliers is
inspected. This type of inspection has the advantage of ensuring that all
defective units are eliminated. However, it is very time consuming and costly.
Also, if the unit is destroyed under investigation, e.g., light bulb, cracker,
ammunition, etc., 100% inspection is not practicable.
In 1920, Harold F. Dodge and Harry G. Roming developed statistical methods
for product control known as acceptance sampling or sampling inspection as
an alternative of 100% inspection. Now-a-days, product control is achieved
through acceptance sampling.
In this unit, we introduce the acceptance sampling plans and the related
terminology. In Sec. 5.2, we explain the meaning of the term inspection as
used in industry. We discuss the concept of acceptance sampling plan, its
advantages and limitations and the types of acceptance sampling plans in Sec.
5.3. In Sec 5.4, we discuss the implementation of the acceptance sampling plan.
In Secs. 5.5 and 5.6, we explain the basic terminology related to acceptance

1
Product Control sampling plan, such as lot, probability of accepting a lot, acceptance quality
level (AQL), lot tolerance percent defective (LTPD), producer’s risk and
consumer’s risk. In the next unit, we shall discuss the rectifying sampling
plans.

Objectives
After studying this unit, you should be able to:
 distinguish between 100% inspection and sampling inspection;
 define acceptance sampling plan and discuss the procedure of its
implementation;
 compute the probability of accepting or rejecting a lot;
 define acceptance quality level (AQL) and lot tolerance percent defective
(LTPD) of the lot; and
 compute the producer’s risk and consumer’s risk for an acceptance
sampling plan.

5.2 INSPECTION
You have learnt briefly about 100% inspection and sampling inspection in
Sec. 1.3.1 of Unit 1. In this section, we discuss these concepts in some more
detail.
When a manufacturer produces a product or buys some parts of the product
from outside agents or suppliers, he/she would like to ensure that the final
product is as per specifications. For this purpose, he/she inspects the lot at
every strategic point. The method of checking, measuring or testing one or
more quality characteristics of the product or the parts and determining whether
it satisfies the required specifications or not, is called inspection.
Inspection is of two types:
i) Inspection of Variables
In this type of inspection, the quality characteristic (s) of an item/unit is (are)
measured and compared with the required specifications. For example, the
desired specification for the diameter of a ball bearing is 50 mm, length of the
refill of a ball pen is 14 cm, weight of a cricket ball is 162 gram, and so on.
ii) Inspection of Attributes
A Go-No-Go gauge is In inspection of attributes, actual measurements are not taken. Instead, the
an inspection tool which item/units are categorised as defective or non-defective on the basis of
is used to check an item
or a unit or a piece
Go-No-Go gauges. It means that if a unit fulfils all required quality
against its allowed characteristics, it is categorised as non-defective and if not, it is categorised as
tolerances. The name defective. In such type of inspection, the number of defective units (or defects)
Go-No-Go derives from is counted.
its use. It means that we
check the item and if the Methods of Inspection
item is acceptable (fulfil
the specifications), we There are two methods of inspection:
say Go and if
unacceptable, we say
i) 100% Inspection
No-Go. In this method of inspection, each and every item/unit of any given lot is
inspected. A decision regarding the quality of the entire lot is taken on the basis
of all inspected units of the lot. This procedure needs huge expenditure of time,
2
money, labour and resources. Also, if the product is such that it is completely Acceptance Sampling
destroyed under the process of inspection (e.g., a cracker, ammunition, etc.) Plans
then in such cases, 100% inspection is neither practicable nor economical.
As an alternative, we use sampling inspection.
ii) Sampling Inspection
In sampling inspection method, some items/units (called sample) are randomly
selected from a lot in such a way that the selected sample is a true
representative of the entire lot. Then each and every unit of the selected sample
is inspected. A decision regarding the quality of the entire lot is taken on the
basis of the information obtained from the sampled units.
The question now is: How do we take any decision about the quality of a lot on
the basis of sampling inspection? For this, we need to learn about the
acceptance sampling plan.

5.3 ACCEPTANCE SAMPLING PLAN

Let us first define acceptance sampling.
A sampling inspection in which a decision about acceptance or rejection of a
lot is based on one or more samples that have been inspected is known as
acceptance sampling.
In other words:
Acceptance sampling is a technique in which a small part or a fraction of the
units/items is selected randomly from a lot and the selected units are inspected
to decide whether the lot should be accepted or rejected on the basis of the
information provided by the sample inspection.
For example, suppose a manufacturer of cricket balls supplies them in lots of
500. A buyer wants to inspect 20 balls from each lot before accepting the lot
and takes a decision about acceptance or rejection of the lot on the basis of the
information provided by this sample. This is an example of acceptance
sampling because the decision about the lot is taken on the basis of a sample.
Consider another example. Suppose a mobile phone company packs the mobile
phones it produces in lots of 100 units. To check the quality of the lots, the
quality inspector of the company draws a random sample of size 10 from each
lot. He/she takes a decision about acceptance or rejection of the lot on the basis
of the information provided by this sample.
When a lot is rejected on the basis of the acceptance sampling, it does not mean
that all units of the lot are of no use or all units are defective. It means that all
units of the lot have been carefully examined and the defective units have been
either repaired or removed. Thus, the remaining units in the lot conform to a
particular quality level. The rejected units may be sold at a lower price or the
defects in the units may be removed so that it can be retained. For instance, the
clearance sales of, e.g., ready-made clothes or shoes, etc. are carried out to sell
those items which are not up to the specifications or have been rejected. This is
a loss to the producer or the manufacturer, which can be prevented if he/she
could maintain the desired level of quality. Therefore, manufacturers should be
very careful about the quality of goods, which they supply to the customers.
We can now answer the question: What is an acceptance sampling plan?

3
Product Control An acceptance sampling plan is a specific plan that clearly states the rules for
sampling and the associated criteria for acceptance or rejection of a lot.
Acceptance sampling plans can be used for the inspection of:
1. Manufactured units/items,
2. Components,
3. Raw materials,
4. Operations,
5. Materials in process,
6. Supplies in storage,
7. Maintenance operations,
8. Data or records, and
9. Administrative procedures.

Acceptance sampling has many advantages and some limitations, which we

now describe.

5.3.1 Advantages and Limitations of Acceptance Sampling

The main advantages of acceptance sampling are as follows:
i) It is less expensive in terms of money, time and labour in comparison to
100% inspection.
ii) For items, which cannot be used after single inspection, such as crackers,
bulbs, tube lights, food, etc., 100% inspection is not practicable.
Sampling inspection is the only way for inspecting such items.
iii) In acceptance sampling, a sample of a small number of items or units is
inspected and hence smaller inspection staff is required.
iv) In many cases, acceptance sampling provides better outgoing quality. In
general, it is seen and agreed that good 100% inspection removes only
85% to 90% of the defective items, whereas very good 100%
inspection removes only 99% of the defective items. However, due to
human error, it usually does not reach the 100% mark. In other words, we
can say that 100% inspection is not always reliable because it involves
too much routine work for the persons inspecting each and every item.
Due to this, defective items may also be labelled as satisfactory and may
also be accepted at times when these persons are distracted. Hence, an
appropriate sampling plan is preferable.
v) Due to quick inspection through the acceptance sampling, the scheduling
and delivery times are saved.
Acceptance sampling has some limitations which are given below:
i) Since, in acceptance sampling, the entire lot is accepted or rejected on the
basis of conclusions drawn from one or more samples, there is always
some risk of making wrong inference about the quality of the lot. These
risks are termed the producer’s risk and consumer’s risk. You will learn
more about these in Sec. 5.6.
ii) The success of acceptance sampling depends on the randomness of the
sample, quality characteristics to be tested, lot size, acceptance criteria,
4
 etc. Therefore, it is a specialised job requiring careful planning and Acceptance Sampling
execution and every one cannot undertake it. Plans

So far you have learnt about the acceptance sampling plan and its advantages
and limitations. It is also important for you to know about different types of
acceptance sampling plans used in industries.

5.3.2 Types of Acceptance Sampling Plans

There are several types of acceptance sampling plans based on different
approaches that can influence the decision about a lot. These are categorised as
follows:
1. Acceptance sampling plans for attributes, and
2. Acceptance sampling plans for variables.
1. Acceptance Sampling Plans for Attributes
Acceptance sampling plans in which actual measurements of the quality
characteristics are not made, but the units/items are categorised as defective or
non-defective on the basis of Go-No-Go gauges are called acceptance
sampling plans for attributes. These plans are of two types:
A. Lot-By-Lot Acceptance Sampling Plans for Attributes
Lot-by-lot acceptance sampling plans for attributes are the most commonly
used sampling plans and, therefore, are simply called acceptance sampling
plans for attributes. Such plans are used whenever the units to be inspected can
be conveniently grouped into batches or lots.
In lot-by-lot acceptance sampling plan for attributes, generally, we use the
following plans:
i) Single sampling plans
In the single sampling plan, the decision about the acceptance or rejection of a
lot is based on a single sample that has been inspected. This is the simplest type
of sampling plan.
ii) Double sampling plans
In the double sampling plan, the decision about the acceptance or rejection of a
lot requires the evidence of two samples drawn from the lot. If the lot quality is
good (or bad), the lot is accepted (or rejected) on the basis of the first sample.
If the first sample shows an intermediate quality, the decision about the lot is
taken on the evidence of the first and second sample combined. This is more
complicated than the single sampling plan.
iii) Multiple sampling plans
A multiple sampling plan is an extension of the double sampling plan. This
sampling plan may require more than two samples to reach a decision about the
acceptance or rejection of a lot.
iv) Sequential sampling plans
In the sequential sampling plan, the sample size is not pre-fixed as in the case
of single, double and multiple sampling plans. The units/items are drawn from
the lot, one at a time and inspected. The decision about accepting or rejecting
of the lot or continuing with the inspection by taking one more unit from the
lot, is made on the basis of information available up to that stage.

5
Product Control The multiple and sequential sampling plans are beyond the scope of this
course. However, in Units 7 and 8, we shall discuss the single sampling plans
and double sampling plans in detail.
B. Continuous Production Acceptance Sampling Plans for Attributes
Continuous production acceptance sampling plans for attributes are used when
the units/items to be inspected cannot be grouped into lots or batches. Many
manufacturing operations do not create lots because in these operations, the
units are produced in a continuous process on a conveyor belt or other
straight-line systems.
Continuous production acceptance sampling plans for attributes are beyond the
scope of this course.
2. Acceptance Sampling Plans for Variables
The acceptance sampling plans in which actual measurements of the quality
characteristics are taken are called acceptance sampling plans for variables.
The acceptance sampling plans for variables are beyond the scope of this
course. So we now focus on the acceptance sampling plans for attributes and
describe the general procedure for implementing it.

5.4 IMPLEMENTATION OF ACCEPTANCE

SAMPLING PLAN FOR ATTRIBUTES
Suppose that lots of the same size, say, N, are received from the supplier or the
final assembly line and submitted for inspection one at a time. The procedure
for implementing the acceptance sampling plan to arrive at a decision about the
lot is described in the following steps:
Step 1: We draw a random sample of size n from the lot received from the
supplier or the final assembly.
Step 2: We inspect each and every unit of the sample and classify it as
defective or non-defective on the basis of certain criteria. At the end
of the inspection, we count the number of defective units found in the
sample.
Step 3: We compare the number of defective units found in the sample with
the acceptance criteria.
Step 4: If the acceptance criteria are satisfied, we accept the entire lot.
Otherwise, we reject the entire lot.
The steps described above are shown in Fig. 5.1.

6
 Acceptance Sampling
Plans

Fig. 5.1: Procedure for implementation of the acceptance sampling plan.

Let us explain these steps further with the help of an example.
Suppose a buyer of cricket balls wants to inspect a sample of 20 balls from
each lot to check their quality. The acceptance criterion is that if the sample
contains at most one defective ball, the lot would be accepted. Otherwise, it
would be rejected. The buyer draws 20 balls from each lot and inspects each
and every ball of the sample and classifies each ball as defective or non-
defective on the basis of certain defects. At the end of the inspection, he/she
counts the number of defective balls found in the sample and compares the
number of defective balls with the acceptance criterion. If the number of
defective balls in the sample is more than 1, he/she rejects the lot. If the
number of defective balls is zero or 1, he/she accepts the lot.
We shall discuss different types of sampling inspection plans in detail in the
remaining units. But, before studying these plans, you should be aware about
the related terminology. This is what we discuss in Sec. 5.5.

5.5 TERMS USED IN ACCEPTANCE SAMPLING

PLANS
In this section, we define some terms used in acceptance sampling plan.
1. Lot
A lot is the collection of units or items from which a sample is taken and
inspected to determine its acceptability.
It is found that lot formation influences the effectiveness of the acceptance
sampling plan. Therefore, formation of the lot is important for the success
of an acceptance sampling plan. Some guidelines for the lot formation are
as follows:
i) Units in the lot should be homogeneous. It means that the lot should
consist of units produced by the same machine, same operators, using
the same raw materials and approximately during the same time period.
ii) Units in the lots should be packed so that the shipping costs and
handling risks are minimum. This also makes the selection of units in
the sample easy.
Product Control The decision about the acceptance or rejection of all units of the lot is taken
on the basis of the results of inspection of the sample. Therefore, the
sample should be such that it is a true representative of the lot.
2. Sentencing
The act of accepting or rejecting the entire lot is called sentencing the lot.
3. Lot Size
The number of units in a lot is called lot size. It is denoted by N.
4. Sample Size
The number of units inspected to sentence a lot is called sample size. It is
denoted by n.
5. Lot Quality
The proportion of defective units in a lot is called lot quality or proportion
defective. It is denoted by p and defined as
Number of defective units in a lot
p ... (1)
lot size
From the definition of p given by equation (1), you should note that for the
same lot size, as p increases, the lot quality decreases.
6. Acceptance Number
When a deal is finalised between a seller and a buyer, they decide on the
maximum number of allowable defective units in a sample. This number is
called acceptance number and is denoted by c. The rule for acceptance or
rejection of a lot is that if the number of defective units observed in the
sample is less than or equal to the acceptance number c, the lot will be
accepted. Otherwise, it will be rejected.
Let us explain the above terminology with the help of an example.
Example 1: Suppose a cricket ball manufacturing company supplies lots of
500 balls. To check the quality of the lots, a buyer draws a random sample of
size 20 balls from each lot and accepts the lot if the inspected sample contains
at most one defective ball. Otherwise, he/she rejects the lot. If the lot consists
of 10 defective balls, find N, n, p and c.
Solution: Since the lot contains 500 balls, the lot size N = 500.
Since the buyer draws a sample of size 20 balls from each lot to take the
decision about the lot, the sample size n = 20.
The buyer accepts the lot if the inspected sample contains at most one defective
ball. So the acceptance number c = 1.
The lot consists of 10 defective balls. So from equation (1), the lot quality is
given as
Number of defective units in a lot 10
p   0.02
lot size 500
Now that you have learnt about terms such as lot size, lot quality, sample size,
etc., you may like to know: How can we be confident that a lot of good
quality will be accepted and a lot of bad quality will be rejected? We can
answer such questions by calculating the probability of accepting a lot of
specified quality. This probability will be 1 if all units in the lot are accepted
8
and 0, if all units in the lot are rejected. The probability of accepting a lot is Acceptance Sampling
denoted by Pa (p) or in short as Pa. Plans

7. Probability of accepting a lot (Pa)

Suppose that the lots of the same size, say, N, are received from the
supplier or the final assembly line and submitted for inspection one at a
time. The lot is accepted if the number of defective units observed in the
sample is less than acceptance number (c). Otherwise, it is rejected.
Suppose a random sample of size n is drawn from each lot for inspection.
If X represents the number of defective units in the sample, the lot is
accepted if X ≤ c. It means that we accept the lot if X = 0 or 1 or 2, ... or c.
Therefore, the probability of accepting the lot is given by
Pa  p   Pa  P  X  c  P  X  1 or 2 or 3, ..., or c ... (2) You have studied in Unit 3
of MST-003 that if A and
Since X = 0, 1,..., c are mutually exclusive events, from the addition B are mutually exclusive
theorem of probability, we have events then
P  A or B  P  A   P  B
Pa  p   P  X  c  P  X  0  P  X  1  ...  P  X  c 
c
or Pa  p    P  X  x  ... (3)
x0

We can easily calculate this probability if we know the probability

distribution of X. Generally, in quality control, a random sample is drawn
without replacement. So the number of defective units (X) in the sample
follows a hypergeometric distribution. However, we know that when the lot
size (N) is large compared to the sample size, i.e, N  10n , the
hypergeometric distribution approximates a binomial distribution with
parameters n and p where p is the lot quality. It is easier to calculate the
probabilities with the help of the binomial distribution rather than with the
hypergeometric distribution.
Therefore, the probability of getting exactly x defective units in a sample of The notation
n
Cx
size n using the binomial distribution is given by can also be
n x repersened as
P  X  x   n C x p x 1  p  ... (4)
 n .
x
Hence, from equation (3), we have  

c c
nx
Pa  p   P  X  c   P  X  x    n Cx px 1  p  ... (5)
x 0 x 0

However, for rapid calculation, we can use Table I entitled Cumulative

Binomial Probability Distribution given at the end of this block.
Let us now compute Pa (p) for Example 1.
In this example, we have
N  500, n  20, c  1, p  0.02
Suppose X represents the number of defective balls in the sample. The
buyer accepts the lot if the number of defective balls (X) in the sample is at
most one, i.e., c = 1. Therefore, the probability of accepting the lot of
quality p = 0.02 is given by
Pa  p   P  X  c  P  X  1  P  X  0  P  X 1

9
Product Control 1
  P X  x 
x 0

Since the lot size (N) is large compared to the sample size (n), i.e.,
N  10n, X approximately follows the binomial distribution with
parameters n and p where p is the lot quality. Therefore, the probability of
accepting the lot of quality p is given by
1 1
nx
Pa  p    P  X  x    n Cx p x 1  p 
x 0 x 0

For rapid calculation, we can use Table I for obtaining this probability.
From Table I, for n = 20, x = c = 1 and p = 0.02, we have
1
n nx
 Cx p x 1  p   0.9401
x 0

Therefore, the probability of accepting the lot of quality p = 0.02 is given

by
1
nx
Pa  p    n Cx p x 1  p   0.9401 .. (6)
x 0

It means that if there are several lots of the same quality p = 0.02, about
94.01% of these will be accepted and about 5.99% will be rejected.
We can also calculate this probability manually using the scientific
calculator for n = 20, p = 0.02 as follows:

n x
Pa  p   P  X  1  P  X  0   P  X  1
P  X  x   n Cx p x 1  p 
20 0 20  0 1 20 1
 C0  0.02  1  0.02   20 C1  0.02  1  0.02 
20 19
  0.98   20  0.02   0.98 

 0.6676  0.2725  0.9401

Now, we can see the effect of the lot quality on the probability of
accepting the lot.
Suppose the lot in Example 1 contains 20 defective balls instead of 10.
Then the lot quality is:
Number of defective units in the lot 20
p   0.04
lot size 500
We can calculate the probability of accepting the lot of quality p = 0.04
with the help of Table I as discussed above.
From Table I, for n = 20, x = c = 1 and p = 0.04, we have
Pa  p   P  X  1  0.8103

In this case, about 81.03% lots will be accepted and about 18.97% lots will
be rejected.
If we compare this probability with the probability obtained for p = 0.02,
we observe that as the lot quality decreases (from p = 0.02 to p = 0.04), the

10
 probability of accepting the lot also decreases (from Pa = 0.9401 to Acceptance Sampling
Pa = 0.8103). Plans

We can also see the effect of the acceptance number on the probability
of accepting the lot.
Suppose the buyer accepts the lot if the inspected sample contains at most
two defective balls, that is, c = 2. Then we can obtain the probability of
accepting the lot using Table I.
From Table I, for n = 20, x = c = 2 and p = 0.02, we have
Pa  p   P  X  c  P  X  2  0.9929

It means that if there are several lots of the same quality p = 0.02, then for
the acceptance number c = 2, about 99.29% lots will be accepted and about
0.71% lots will be rejected.
If we compare this probability with the probability obtained for p = 0.02
and c = 1, we observe that when the acceptance number increases (from c
= 1 to c = 2), the probability of accepting the lot also increases (form
Pa = 0.9401 to Pa = 0.9929).
It is time for you to pause and solve some exercises for practice.
E1) A shopkeeper purchases pens from a pen company in cartons (lots) that
usually contain one thousand pens. To check the quality of the pens, the
shopkeeper selects 25 pens at random from each carton and visually
inspects each selected pen for certain defects. The shopkeeper accepts the
lot if the inspected sample contains at most two defective pens.
Otherwise, he/she rejects the lot. If there are 40 defective pens in each
carton, find N, n, p, c and Pa.
E2) For a sampling plan with n = 5 and c = 0, find the probability of
accepting a lot that has 2% defective units by assuming that the number
of defective units in a sample follows a binomial distribution.

8. Acceptance Quality Level (AQL)

The seller and the buyer generally know that the supply of completely
defect-free lots is not possible. So they usually negotiate and arrive at an
agreement that the buyer will accept all lots which have at most a definite
quality level or definite percentage of defective units. This definite quality
level is known as acceptance quality level (AQL). It is denoted by p1.
Hence, AQL can be defined as follows:

Acceptance quality level is the quality level decided in a negotiation of the

seller and the buyer: If the proportion of defective units in a lot is less than
or equal to AQL, the buyer will have to definitely accept the lot. Otherwise,
the buyer may accept or reject the lot.
Suppose 100% inspection of all lots is carried out and the lot quality (p) for
each lot is computed. Then all lots with lot quality p  AQL will be
accepted and all lots with p > AQL may either be accepted or rejected.
In Example 1, if the manufacturing company and the buyer decide that the
buyer will accept all lots which have at most 2% defective units, then
acceptance quality level (AQL) of this deal is 2%.

11
Product Control Even though the seller and the buyer decide that the buyer would accept all
lots of AQL quality, this may not happen in practical situations. In Example
1, the probability of accepting the lot of the lot quality p = AQL = 0.02 was
Pa  p   P  X  1  0.9401

It means that if there are several lots of the same quality p = 0.02 = AQL,
about 94.01% of these would be accepted and about 5.99% would be
rejected. This is obviously a risk of the manufacturer because it was agreed
upon by both that all lots of quality 0.02 will be accepted whereas the buyer
is rejecting 5.99% of them. This risk is known as producer’s risk which is
explained in detail in Sec.5.6.
9. Lot Tolerance Percent Defective (LTPD)

In order to reduce the producer’s risk, the buyer agrees to tolerate a lot
quality worse than acceptance quality level (AQL) up to a certain limit but
not beyond it. This limiting value is known as the lot tolerance percent or
proportion defective (LTPD). It is also known as rejectable quality level
(RQL), unsatisfactory quality level or limiting quality level (LQL).

LTPD is also decided at the time when the acceptance quality level is
decided in the negotiation between the producer and the buyer. They make
an agreement that the buyer will definitely reject the lot of quality equal to
or greater than LTPD. Therefore, LTPD can be defined as follows:
Lot tolerance percent or proportion defective is the quality level decided
in negotiation between the producer and the buyer: If the proportion of
defective units in a lot is equal to or greater than this level, the buyer will
definitely reject the lot. Otherwise, the buyer may accept or reject the lot.
The lot tolerance percent defective (LTPD) of a sampling plan is the level
of quality at which the lot is routinely rejected by the sampling plan. LTPD
is greater than AQL.
In Example 1, if the manufacturing company and the buyer decide that the
buyer will accept all lots which have at most 2% defective units and reject
all lots which have 5% or more defective units, then for this plan the AQL
is 2% and the LTPD is 5%.
This level represents the dividing line between good and bad lots.
Therefore, lot tolerance percent defective is the quality level of a lot that
the buyer considers bad and he/she would like to reject all lots that have
this level of quality. LTPD is denoted by p 2.
In acceptance sampling, the acceptance or rejection of the entire lot depends on
the conclusions drawn from the sample. Thus, there is always a chance of
making a wrong decision. It means that a lot of good quality may be rejected
and a lot of poor quality may be accepted. This leads to two kinds of risks:
i) Producer’s risk, and
ii) Consumer’s risk.
We now discuss these risks in some details.

5.6 PRODUCER’S RISK AND CONSUMER’S RISK

In quality control, a producer may be defined as follows:

12
A person or a firm or an organisation that produces/manufactures goods or Acceptance Sampling
provides services for use or consumption of another person or firm or Plans
organisation is known as a producer.
In Example 1, the manufacturing company that produces the cricket balls, is
the producer. Let us explain what is meant by producer’s risk.
Producer’s Risk
It may happen in practice that a sampling inspection plan leads to the rejection
of a lot of satisfactory or good quality. This means that there is a possibility of
rejecting a lot having a quality level less than or equal to the acceptance quality
level (AQL) due to sampling inspection. If a lot of good quality is rejected, the
producer suffers loss. Therefore, the producer always faces the risk of a good
lot being rejected. Such a risk is known as producer’s risk and is defined as
follows:
The probability of rejecting a lot of acceptance quality level (AQL) is known as
the producer’s risk. It is denoted by Pp(p) or in short as Pp and given by
Pp  p   Pp  P  rejecting a lot of acceptance quality level   

This can be written as

Pp  p   1  P  accepting a lot of acceptance quality level

 1  Pa (p  p1 ) ... (7)
The producer’s risk is equivalent to the type I error in hypothesis testing
discussed in Unit 8 of MST-004 entitled Statistical Inference. Therefore, it is
also denoted by α.
For computing the producer’s risk, we have to compute the probability of
accepting a lot of quality p = AQL. Then we use equation (7) to compute the
producer’s risk.
Let us explain how to compute the producer’s risk with the help of an example.
Example 2: Suppose in Example 1, the manufacturing company and the buyer
agree that AQL = 0.02. Find the producer’s risk for this plan.
Solution: We know that the producer’s risk is
Pp  p   1  P  accepting a lot of acceptance quality level  1  Pa  p 

For computing the producer’s risk, we first compute the probability of

accepting a lot of quality (p) = AQL = 0.02. We have already calculated this
probability in Example 1. So here we use the result:
Pa  p   0.9401

Therefore, the producer’s risk for this plan is

Pp  p   1  Pa  p   1  0.9401  0.0599

It means that if there are several lots of the same quality p = 0.02 as AQL, out
of these lots, about 5.99% will be rejected. This is obviously a risk of the
manufacturer (producer) because it was agreed upon by both that all lots of
quality p =AQL = 0.02 will be accepted whereas the buyer is rejecting 5.99%
of the lots.
Let us now explain the consumer’s risk.
13
Product Control Consumer’s Risk
A person or a firm or an organisation that purchases goods for its own need
or consumption or for use in the production of other goods (not for resale to
another person or firm or organization directly or indirectly) is known as a
consumer.
In Example 1, since the buyer purchases the cricket balls from the
manufacturer, he/she is the consumer.
Just as the producer has a risk of a lot of good quality being rejected, a
consumer also has a risk of buying a lot of unsatisfactory quality. Such a risk is
known as the consumer’s risk. If p 2 is the maximum proportion of defective
(LTPD) in the lot, which the consumer is ready to tolerate, the consumer’s risk
may be defined as follows:
The probability of accepting a lot of unsatisfactory quality, i.e., LTPD is
known as consumer’s risk. It is denoted by Pc(p) or in short as Pc.
Thus,
Pc  p   P  accepting a lot of quality = LTPD    ... (8)

Note that Pc(p) is the same as Pa(p) for p = LTPD.

The consumer’s risk is denoted by β because it is equivalent to the type II error
described in Unit 8 of MST-004.
Let us explain how to compute the consumer’s risk with the help of an
example.
Example 3: Suppose in Example 1, the manufacturing company and the buyer
agree that LTPD = 0.05. Find the consumer’s risk for this plan.
Solution: We know that the consumer’s risk is given by
Pc  p   P  accepting a lot of quality = LTPD   Pa (p  0.05)

It is given that
N  500, n  20, c  1 and LTPD  0.05
We first calculate the probability of accepting the lot of quality p = 0.05 as
follows:
Pa  p   P  X  c  P  X  1

We can calculate this probability as we have discussed in Example 1.

From Table I, for n = 20, x = c = 1 and p = LTPD = 0.05, we have
Pa  p   P  X  1  0.7358

Therefore the consumer’s risk is given by

Pc  p   Pa  p  0.05   0.7358

It means that if there are several lots of the same quality p = 0.05, out of these
about 73.58% of them will be accepted by the buyer even though this quality is
unsatisfactory. This is obviously the buyer’s risk.
We take up an example to further explain the terms AQL, LTPD, producer’s
risk and consumer’s risk.

14
Example 4: A mobile manufacturing company has decided to purchase the Acceptance Sampling
mobile batteries from a battery manufacturing company. Both manufacturing Plans
companies have decided that the batteries are to be supplied in lots of 1000
batteries each. The lot will be accepted up to quality level p = 0.05 and rejected
at more than quality level p = 0.20. Acceptance sampling plan is based on a
sample of size 25 drawn from each lot and the lot is accepted if inspected
sample contains at most one defective battery. Otherwise, the lot is rejected.
Identify which company is the producer and which one is the consumer in the
plan. Calculate the producer’s risk and the consumer’s risk.
Solution: Here the mobile manufacturing company purchases the mobile
batteries from a battery manufacturing company. So it is a consumer. The
battery manufacturing company supplies batteries to the mobile manufacturing
company. So it is a producer.
It is given that
N  1000, AQL  0.05, LTPD  0.20, n  25 and c  1
We know that the producer’s risk is defined as
Pp  P  rejecting a lot of acceptance quality level

 1  P  accepting a lot of quality = AQL = 0.05  1  Pa  p  0.05  ...(i)

To calculate the producer’s risk, we have to calculate the probability of

accepting a lot of quality p = 0.05.
Suppose X denotes the number of defective mobiles in the sample. The mobile
company accepts the lot if the number of defective mobiles (X) in the sample is
at most one. Therefore, the probability of accepting the lot of quality p is given
by
Pa  p   P  X  c  P  X  1

Since the lot size (N) is large compared to the sample size (n), i.e., N  10n, we
can use the binomial distribution and can use Table I for obtaining the
probability P  X  1 .

From Table I, for n = 25, x = c = 1 and p = 0.05, we have

P  X  1  0.6424

Therefore, the probability of accepting the lot of quality p = 0.05 is given by

Pa (p  0.05)  0.6424
From (i), we calculate the producer’s risk as follows:
Pp  1  Pa  p  0.05   1  0.6424  0.3576

Similarly, we know that the consumer’s risk is defined as

Pc  P  accepting a lot of quality p  LTPD  0.20  Pa (p  0.20)

From Table I, for n = 25, x = c = 1and p = 0.2, we have

Pc  Pa  p  0.20   0.0274

You may like to solve some exercises for practice.

15
Product Control E3) Identify the consumer and producer in the exercise E1. If the shopkeeper
and the pen company have decided that AQL = 0.03 and LTPD = 0.10,
calculate the producer’s risk and the consumer’s risk.
E4) If in the exercise E2, the acceptance quality level (AQL) and the Lot
tolerance percent defective (LTPD) are 1% and 5%, respectively,
calculate the producer’s risk and the consumer’s risk for this plan.
We end this unit by giving a summary of what we have covered in it.

5.7 SUMMARY
1. The technique of controlling the quality of products in such a way that
these are free from defects and conform to their specifications, is called
product control.
2. The method of checking, measuring, testing one or more quality
characteristics of a product (unit) to determine whether it satisfies the
required specifications or not is called inspection.
3. If quality characteristic of a unit is measured on a continuous scale, the
inspection is called inspection by variables.
4. If the actual measurements of the quality characteristic of a unit are not
taken, but the unit is categorised as defective or non-defective on the
basis of Go-No-Go gauges, the inspection is called inspection by
attributes.
5. A lot is the collection of units or items from which a sample is taken and
inspected to determine its acceptability.
6. If each and every unit of a lot is inspected, the inspection is known as
100% inspection.
7. If some items or units are randomly selected from a lot in such a way that
a selected sample is a true representative of the entire lot and each and
every unit of the selected sample is inspected, the inspection is known as
sample inspection.
8. Acceptance sampling is a technique in which a small part or a fraction of
units is selected randomly from a lot and the selected units are inspected
to decide whether the lot should be accepted or rejected on the basis of
the information supplied by the sample inspection.
9. The proportion of defective units in a lot is called lot quality or
proportion defective, which is denoted by p and defined as follows:
Number of defective units in a lot
p
lot size
10. The acceptance quality level (AQL) is the quality level decided
mutually by the manufacturer and the buyer. If the proportion of defective
units in a lot is less than or equal to AQL, the buyer has to accept the lot.
Otherwise, the buyer may accept or reject the lot.
11. The lot tolerance percent or proportion defective (LTPD) is the
quality level decided mutually by the manufacturer and the buyer. If the
proportion of defects in a lot is greater than or equal to this level, the
buyer will definitely reject the lot. Otherwise, the buyer may accept or
reject the lot.
16
12. The probability of rejecting a lot of acceptance quality level (AQL) is Acceptance Sampling
known as producer’s risk Pp(p). Plans

13. The probability of accepting a lot of unsatisfactory quality (LTPD) is

known as consumer’s risk Pc(p).

5.8 SOLUTIONS/ANSWERS
E1) The carton contains 1000 pens and so the lot size N = 1000.
Since the shopkeeper draws a sample of size 25 pens, the sample size
n = 25.
The shopkeeper accepts the lot if the inspected sample contains at most
two defective pens. Otherwise, he/she rejects the lot.
 The acceptance number c = 2
The lot contains 40 defective pens
Number of defective pens in the lot 40
 p   0.04
lot size 1000
If X represents the number of defective pens in the sample, then the
shopkeeper accepts the lot if X ≤ c = 2. Therefore, the probability of
accepting the lot is given by
Pa  P  X  2  P  X  0  P  X  1  P  X  2

Since the lot size is large as compared to the sample size ( N  10n ), X
approximately follows the binomial distribution with parameters n and
p. Therefore, we can obtain this probability by using Table I.
From Table I, for n = 25, x = c = 2, and p = 0.04, we have
Pa  P  X  c  2  0.9235

E2) It is given that

n = 5, c = 0 and p = 2% = 0.02
Suppose X represents the number of defective units in the sample, the
probability of accepting the lot is given by
Pa  P  X  c  P  X  0

Since X follows the binomial distribution, we can use Table I to obtain

this probability.
From Table I, for n = 5, x = c = 0, and p = 0.02, we have
P  X  0  0.9039

Hence, the probability of accepting the lot is given by

Pa  P  X  0  0.9039

E3) The shopkeeper purchases pens from a pen company and, therefore,
he/she is a consumer. The pen company is a producer since it sells pens.
We have
n = 25, c = 2, AQL = 0.03 and LTPD = 0.10

17
Product Control The producer’s risk is given by
Pp    P  rejecting a lot of acceptance quality level

 1  P  accepting a lot of AQL = 0.03

 1  Pa  p  0.03

` From Table I, for n = 25, x = c = 2 and p = AQL = 0.03, we have

Pa  p  0.03  0.9620

 Pp  1  Pa  p  0.03  1  0.9620  0.0380

The consumer’s risk is given by

Pc  P  accepting a lot of LTPD quality

 Pa  p  0.10 

From Table I, for n = 25, x = c = 2 and p = LTPD = 0.10, we have

Pa  p  0.10   0.5371

 Pc  Pa  p  0.10   0.5371

E4) It is given that

n = 5, c = 0, AQL = 1% = 0.01 and LTPD = 5% = 0.05
The producer’s risk is
Pp    P  rejecting a lot of acceptance quality level

 1  Pa  p  0.01

`From Table I, for n = 5, x = c = 0 and p = AQL = 0.01, we have

Pa  p  0.01  0.9510

 Pp  1  Pa  p  0.01  1  0.9510  0.0490

The consumer’s risk is

Pc  P  accepting a lot of LTPD quality

 Pa  p  0.05

From Table I, for n = 5, x = c = 0 and p = LTPD = 0.05, we have

Pa  p  0.05  0.7738

 Pc  Pa  p  0.05  0.7738

18