Topic 4

Statistical Process
Control
• Readings:
1. Chapter 6 (Supplement): Heizer, J., Render, B. & Munson,
C. (2020). Operations Management, 13th Edition. Upper
saddle River, New Jersey: Pearson.
2. Chapter 10: Stevenson W.J., & Chee Chuong S.C. (2014).
Operations Management, Second Edition (Asia). Mc
Graw Hill Education
Learning Objectives
When you complete this chapter you should be able to:

1. Explain the purpose of different control charts

2. Apply 𝑥𝑥̅ − 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 and 𝑅𝑅 − 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 for variables
3. Apply 𝑝𝑝 − 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 and 𝑐𝑐 − 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 for attributes
4. Conduct a run test to evaluate a process
5. Apply process capability measures; 𝐶𝐶𝑝𝑝 and 𝐶𝐶𝑝𝑝𝑘𝑘

WRWH/MGT6240/Chapter 6S 2
 Quality is Conformance to
Requirements
Philip Crosby was famous for his
dozen books on management and
quality. His 1st book in 1979: “Quality
is Free”, by McGraw Hill.

 Philip Crosby (1979): Quality is conformance to carefully

thought-out requirements  4 fundamentals are:
1. Quality is conformance to requirements, not goodness
2. The basic aim of quality management is prevention, not appraisal
3. That zero defects is the performance standard, not some acceptable
level of defects like 6-sigma
4. The measurement of quality is the price of nonconformance
• Refuting argument that quality as an expense, a trade-off, something that you have
to spend money on
WRWH/MGT6240/Chapter 6S 3
 Quality Control
 A process that measures output relative to a
standard and takes corrective action when output
doesn’t meet standards
 Results acceptable  no action required
 Results unacceptable  take corrective actions

The purpose
 To assure that processes are performing in an acceptable
manner
 Accomplished by monitoring process output using
statistical techniques  SPC

WRWH/MGT6240/Chapter 6S 4
Phases of Quality Assurance
Inspection and
Inspection of lots corrective Quality built
before/after action during into the
production production process

Acceptance Process Continuous

sampling control improvement

The least The most

progressive progressive

Stevenson & Chuong, 2014, Figure 10.1

WRWH/MGT6240/Chapter 6S 5
Inspection
 How much or How often?
 Where or When?
 Centralized vs. On-site?
 Attributes or variables

Inputs Transformation Outputs

Acceptance Process Acceptance

sampling control sampling

Stevenson & Chuong, 2014, Figure 10.2

WRWH/MGT6240/Chapter 6S 6
Examples of Inspection Points
Type of Inspection Characteristics
business points
Fast Food Cashier Accuracy
Counter area Appearance, productivity
Eating area Cleanliness, no loitering
Building Appearance,
Kitchen Health regulations
Parking lot Safety, good lighting
Hotel/motel Accounting Accuracy, timeliness
Building Appearance and safety
Main desk Waiting times
Supermarket Cashiers Accuracy, courtesy
Deliveries Quality, quantity

Stevenson & Chuong, 2014, Figure 10.1

WRWH/MGT6240/Chapter 6S 7
 Statistical Control
 A process used to monitor standards by taking
measurements and corrective action as a product
or service is being produced
 The essence is to assure that the output of a process is random
so that future output will be random
 Process is in statistical control when variation only triggered
by random (natural) causes

 SPC measures process performance

 NOT the quality of a particular unit produced as being “good” or
“bad”; in or out of tolerance
 Applies statistical techniques to ensure that processes meet
the standards
 Control chart is important tool

WRWH/MGT6240/Chapter 6S 8
 Statistical Control: Terms
 Quality control
 Is concerned with the quality of conformance of a process
 Managers use SPC to evaluate if process output statistically
acceptable

 Statistical Process Control (SPC)

 Statistical evaluation of the output of a process during production
 A tool for quality control

 Quality of Conformance
 A product or service conforms to design specifications

WRWH/MGT6240/Chapter 6S 9
 Statistical Process Control
 Assumption: variability is inherent in every process
 Should expect it to occur thus need to
accordingly plan for it
 Could be due to
a) Natural or common causes
b) Special or assignable causes

 SPC provides a statistical signal when assignable

causes are present
 To quicken appropriate corrective actions
 To detect then eliminate assignable causes  process brought into
statistical control  performance predictable
 Use control charts to try to determine when the variation is
too far beyond the norm
 Assumed to be likely caused by something extraordinary
WRWH/MGT6240/Chapter 6S 10
Causes of Variation
1. Random Variations
 Also called natural or common causes
 Expected amount of variation
 Created by countless minor factors
 Affect virtually all production processes
 Natural variations form a pattern that can be described as
a distribution (output measurements)
 Assumed to form a normal distribution  could be
characterised by 2 parameters:
a) Mean (𝜇𝜇); a measure of central tendency
b) Standard deviation (SD); a measure of dispersion
 If the distribution of outputs falls within acceptable limits
 the process is said to be “in control”  variation is tolerated

WRWH/MGT6240/Chapter 6S 11
Detecting Variations via Sampling
 If only natural causes of variation are present  the output of a
process forms a distribution that is stable over time and
predictable
 Trying to determine the shape of the sample output over time
 If it remains stable the process is said to be in control  no
corrective action necessary
Frequency

Prediction

Weight

Heizer et. al, 2020, Figure S6.1

WRWH/MGT6240/Chapter 6S 12
 Causes of Variation
2. Assignable Variations
 Also called special causes of variation
 Generally this is due to some changes in the process
 Variations that can be identified or traced to a specific
reason  can be eliminated
 Examples: Machine wear, misadjusted equipment, fatigue, skill,
defective materials, noise
 The objective is to discover when assignable causes are
present
 Eliminate the bad causes
 Incorporate the good causes

WRWH/MGT6240/Chapter 6S 13
Detecting Variations via Sampling
 If assignable causes are present  the process output is not
stable over time  future output is not predicable

?
?? ??
? ?
? ?
? ?
? ?
? ??
?? ?
Frequency

Prediction

Weight

Heizer et. al, 2020, Figure S6.1

WRWH/MGT6240/Chapter 6S 14
 Control Charts
 Graphical presentation of process data over time
 Constructed based on historical data
 Time-ordered plot of sample statistics obtained from an ongoing
process (e.g. sample means/ranges)

 To monitor process output to see if a variation is

due to random or assignable causes
 Its basis is sampling distribution that describes random
variability
 Sampling; samples of process output are taken  compute
sample statistics (sample mean, sample range)  plot charts to
observe
 Control limits define the range of acceptable variation within which
variations designated as randomness from mean

WRWH/MGT6240/Chapter 6S 15
Control Charts

Stevenson & Chuong, 2014, Figure 10.4

 Process monitoring
 Each sample statistic (red dot) on a control chart is compared to the extremes of
the sampling distribution (the control limits) to judge if it is within the acceptable
(random) range

WRWH/MGT6240/Chapter 6S 16
Control Charts
Setting Control Limits

WRWH/MGT6240/Chapter 6S 17
 Process Control
(a) In statistical control
and capable of
producing within control
Frequency limits

Lower control limit Upper control limit

(b) In statistical control
but not capable of
producing within control
limits

(c) Out of control

Size
(weight, length, speed, etc.)
Heizer et. al, 2020, Figure S6.2
WRWH/MGT6240/Chapter 6S 18
 Population and Sampling
Distributions
 Central limit theorem is the underlying statistical
foundation
 Regardless of the distribution of the population, the distribution of sample means
drawn from the population will tend to follow a normal curve as number of samples
increases

WRWH/MGT6240/Chapter 6S 19
Sampling Distributions
As the sample size increases,
the sampling distribution narrows
n = 100

n = 50

n = 25

Mean of process Heizer et. al, 2020, Figure S6.4

WRWH/MGT6240/Chapter 6S 20
Control Charts for Variables
 Process characteristics that can be measured or take any
real value
 Usually on continuous scale; length, width, weight, temperature, volume,
speed, strength; duration

 May be in whole or in fractional numbers

 Continuous random variables; infinite number of possibilities

 2 control charts for variables

a) x� − Chart tracks changes in the central tendency of a process
b) R − Chart tracks the process dispersion within a sample
 Indicates a gain or loss in uniformity of dispersion of a process

 Both charts must be used together; different perspectives

 Mean charts are sensitive to shifts in the process mean
 Range charts are sensitive to changes in process dispersion

WRWH/MGT6240/Chapter 6S 21
 Mean and Range Charts
(a)
These (Sampling mean is
sampling shifting upward, but
distributions range is consistent)
result in the
charts below
UCL
(x-chart detects
x-chart shift in central
tendency)
LCL
UCL
(R-chart does not
R-chart detect change in
mean)
LCL
Heizer et. al, 2020, Figure S6.5
WRWH/MGT6240/Chapter 6S 22
 Mean and Range Charts
(b)
These
(Sampling mean
sampling
is constant, but
distributions
dispersion is
result in the
increasing)
charts below
UCL
(x-chart indicates
x-chart no change in
central tendency)
LCL
UCL
(R-chart detects
R-chart increase in
dispersion)
LCL
Heizer et. al, 2020, Figure S6.5
WRWH/MGT6240/Chapter 6S 23
Setting Chart Limits:
Mean Control Chart
1. For x� − Chart when we know true σ
 Applicable when the process (population) standard deviation is known
based on past data or when a reasonable estimate is available

𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙 𝐿𝐿𝐿𝐿𝐿𝐿 = 𝑥𝑥̿ − 𝑧𝑧𝜎𝜎𝑥𝑥̅

𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙 𝑈𝑈𝑈𝑈𝑈𝑈 = 𝑥𝑥̿ + 𝑧𝑧𝜎𝜎𝑥𝑥̅
Where
x� = mean of the sample means or a target value set for the process
𝑧𝑧 = number of normal standard deviations; confidence level
𝜎𝜎
𝜎𝜎𝑥𝑥 = standard deviations of the sample means =
𝑛𝑛
𝜎𝜎 = standard deviations of the population (process)
𝑛𝑛 = sample size

WRWH/MGT6240/Chapter 6S 24
 Setting Control Limits: Example 1
The weights of boxes of Oat Flakes within a large production lot are sampled
each hour by randomly select and weigh nine (n = 9) boxes each hour.
Managers want to set control limits that include 99.73% of the sample means.
The population (process) standard deviation is known to be 1 ounce.

Average weight in
the first sample

Heizer et. al, 2020

WRWH/MGT6240/Chapter 6S 25
Setting Control Limits: Example 1

Average mean
of 12 samples

𝜎𝜎
𝑈𝑈𝑈𝑈𝑈𝑈 = 𝑥𝑥̿ + 𝑧𝑧𝜎𝜎𝑥𝑥̅ = 𝑥𝑥̿ + 𝑧𝑧
𝑛𝑛
𝜎𝜎
𝐿𝐿𝐿𝐿𝐿𝐿 = 𝑥𝑥̿ − 𝑧𝑧𝜎𝜎𝑥𝑥̅ = 𝑥𝑥̿ − 𝑧𝑧
𝑛𝑛
Heizer et. al, 2020
WRWH/MGT6240/Chapter 6S 26
Setting Control Limits: Example 1

Average mean
of 12 samples

Heizer et. al, 2020

WRWH/MGT6240/Chapter 6S 27
Setting Control Limits: Example 1

Control Chart
for samples Variation due
Out of to assignable
of 9 boxes control causes
17 = UCL

Variation due to
16 = Mean natural causes

15 = LCL

| | | | | | | | | | | |
Variation due
1 2 3 4 5 6 7 8 9 10 11 12
to assignable
Out of causes
Sample number control

Heizer et. al, 2020

WRWH/MGT6240/Chapter 6S 28
Setting Chart Limits:
Mean Control Chart
2. For x� − Chart when we do not know true σ
 Often the case in practice  calculate CLs based on the average range
values instead of SD

𝐿𝐿𝐿𝐿𝐿𝐿𝑥𝑥̅ = 𝑥𝑥̿ − 𝐴𝐴2 𝑅𝑅�

𝑈𝑈𝐶𝐶𝐶𝐶𝑥𝑥̅ = 𝑥𝑥̿ + 𝐴𝐴2 𝑅𝑅�
Where
∑𝑛𝑛𝑖𝑖=1 𝑅𝑅𝑖𝑖
𝑅𝑅� = = average range of the samples
𝑛𝑛
𝐴𝐴2 = control chart factor Table S6.1
x� = mean of the sample means

WRWH/MGT6240/Chapter 6S 29
Setting Chart Limits:
Mean Control Chart
Table for Control chart factor

WRWH/MGT6240/Chapter 6S 30
Setting Control Limits: Example 2

Heizer et. al, 2020

WRWH/MGT6240/Chapter 6S 31
 Setting Control Limits: Example 2
Overall process average = 12 ounces
Average range = 0.25 ounce
Sample size = 5
Number of sample = 10

WRWH/MGT6240/Chapter 6S 32
 Setting Control Limits: Example 2
Overall process average = 12 ounces
Average range = 0.25 ounce
Sample size = 5
Number of sample = 10
𝑈𝑈𝐶𝐶𝐶𝐶𝑥𝑥̅ = 𝑥𝑥̿ + 𝐴𝐴2 𝑅𝑅�
= 12 + (0.577)(0.25) UCL =
= 12 + 0.144 = 𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜
Mean = 12
𝐿𝐿𝐶𝐶𝐶𝐶𝑥𝑥̅ = 𝑥𝑥̿ − 𝐴𝐴2 𝑅𝑅�
= 12 − (0.577)(0.25) LCL =
= 12 − 0.144 = 𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜

Heizer et. al, 2020

WRWH/MGT6240/Chapter 6S 33
 Control Limits: Example 3
For salmon filets at Darden Restaurants
x Bar Chart
11.5 – UCL = 11.524
Sample Mean

11.0 – = 10.959

10.5 – LCL = 10.394

| | | | | | | | |
1 3 5 7 9 11 13 15 17
Range Chart
0.8 –
Sample Range

UCL = 0.6943

0.4 –
= 0.2125
0.0 – | | | | | | | | | LCL = 0
1 3 5 7 9 11 13 15 17
Heizer et. al, 2020
WRWH/MGT6240/Chapter 6S 34
 Setting Chart Limits:
Range Control Chart
 A variables control chart
 Shows sample ranges over time
 Difference between smallest and largest values in a sample
 Monitors process variability (dispersion)
 Is used as a proxy for the standard deviation
 For small enough sample sizes
 The LCL for the range chart will be 0
 Suggesting that it is impossible to have an unusually small amount of
dispersion for samples that is small

 Independent from process mean

Although the process average is in control, process dispersion
may not be in control
 Example: A screw loose in a piece of equipment that fills cereal boxes  average of
the samples may remain the same but the variation within samples could be too large

WRWH/MGT6240/Chapter 6S 35
 Setting Chart Limits:
Range Control Chart
� − Chart
For R
𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙 𝐿𝐿𝐿𝐿𝐿𝐿 = 𝐷𝐷3 𝑅𝑅�
𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙 𝑈𝑈𝑈𝑈𝑈𝑈 = 𝐷𝐷4 𝑅𝑅�

Where
∑𝑛𝑛𝑖𝑖=1 𝑅𝑅𝑖𝑖
𝑅𝑅� = = average range of the samples
𝑛𝑛
𝐷𝐷3 = control chart factor Table S6.1
𝐷𝐷4 = control chart factor Table S6.1

WRWH/MGT6240/Chapter 6S 36
Setting Control Limits: Example 4

Heizer et. al, 2020

WRWH/MGT6240/Chapter 6S 37
 Setting Control Limits: Example 4
Average range = 8 minutes
Sample size = 4
Number of sample = 5

𝐿𝐿𝐿𝐿𝐿𝐿 = 𝐷𝐷3 𝑅𝑅�

= 0 8
= 𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚

𝑈𝑈𝑈𝑈𝑈𝑈 = 𝐷𝐷4 𝑅𝑅�

= 2.282 8
= 𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚

Heizer et. al, 2020

WRWH/MGT6240/Chapter 6S 38
 Setting Control Limits: Example 4
 𝑨𝑨 𝒔𝒔𝒔𝒔𝒔𝒔𝒔𝒔𝒔𝒔𝒔𝒔 𝒓𝒓𝒓𝒓𝒓𝒓𝒓𝒓𝒓𝒓 ≥ 𝟏𝟏𝟏𝟏. 𝟐𝟐𝟐𝟐𝟐𝟐
 Process variability had increased
 Process is producing too much variation UCL = 18.256
 need to investigate the cause to be removed

 𝑨𝑨 𝒔𝒔𝒔𝒔𝒔𝒔𝒔𝒔𝒔𝒔𝒔𝒔 𝒓𝒓𝒓𝒓𝒓𝒓𝒓𝒓𝒓𝒓 ≤ 𝟎𝟎 Mean = 8.0

 Process variability had decreased
 Process is producing less variation  better!
LCL = 0
 need to identify the cause to see if it is
i. Desirable or
ii. Undesirable; had caused negative impact
to others (reduced productivity)

Heizer et. al, 2020

WRWH/MGT6240/Chapter 6S 39
 Steps in Creating
Control Charts
1. Collect 20 - 25 samples from a stable process
 Compute the mean and range of each
 Often of n = 4 or n = 5 observations each

�
2. Compute the overall means (𝑥𝑥̅ and 𝑅𝑅)
 Set appropriate control limits; usually at 99.73% level
 Calculate the preliminary UCL and LCL

3. Graph the sample means and ranges on their control charts

 Determine whether any observations fall outside the acceptable limits

4. Investigate points or patterns that indicate the process is out

of control
 Try to assign causes for the variation  address the causes  resume
the process

5. Collect additional samples and

 If necessary, revalidate the control limits using the new data Heizer et. al, 2020
WRWH/MGT6240/Chapter 6S 40
Control Charts for Attributes
 Process characteristics that are categorical, counted
 Defective or non-defective; good or bad; yes or no; acceptable or
unacceptable; number of calls per day
1. p-Chart: Percent defective in a sample
 To monitor the proportion of defective outputs generated by a process
in one sample
 When the data consist of 2 categories
2. c-Chart: Number of defects in a sample
 To monitor the number of defects per output unit in a sample
 When can count number of occurrence during a given time period, but
not number of non-occurrence
 Useful to monitor processes having large potential for errors but actual
number of errors is relatively small

 A defective record
 A record isn’t exactly correct because it contains >=1 defect
 A bad record may have several defects
WRWH/MGT6240/Chapter 6S 41
 Setting Chart Limits:
p - Chart
1. When proportion of defective in the population is known
 Population will be a binomial distribution, but applying the Central Limit
Theorem allows us to assume a normal distribution for the sample statistics

𝐿𝐿𝐿𝐿𝐿𝐿 = 𝑝𝑝 − 𝑧𝑧𝜎𝜎𝑝𝑝 𝑝𝑝(1 − 𝑝𝑝)

𝑈𝑈𝑈𝑈𝑈𝑈 = 𝑝𝑝 + 𝑧𝑧𝜎𝜎𝑝𝑝 𝜎𝜎𝑝𝑝 =
𝑛𝑛
Where
𝑝𝑝 = average fraction percent of defective in the population
𝑧𝑧 = number of standard deviations; confidence level
𝜎𝜎𝑝𝑝 = standard deviations of the population (process)
𝑛𝑛 = sample size (number of observations in each sample)

WRWH/MGT6240/Chapter 6S 42
 Setting Chart Limits:
p - Chart
2. When proportion of defective in the population is unknown
 When the proportion of defective (p) in the population is unknown, it is
estimated from samples using 𝑝𝑝̅

𝐿𝐿𝐿𝐿𝐿𝐿 = 𝑝𝑝̅ − 𝑧𝑧�

𝜎𝜎𝑝𝑝 𝑝𝑝(1
̅ − 𝑝𝑝)̅
𝑈𝑈𝑈𝑈𝑈𝑈 = 𝑝𝑝̅ + 𝑧𝑧�
𝜎𝜎𝑝𝑝 𝜎𝜎𝑝𝑝 =
𝑛𝑛
Where
𝑝𝑝̅ = mean fraction percent of defective in the samples
𝑧𝑧 = number of standard deviations; confidence level
�𝑝𝑝 = standard deviations of the sampling population (process)
𝜎𝜎
𝑛𝑛 = sample size (number of observations in each sample)

WRWH/MGT6240/Chapter 6S 43
 Setting Control Limits: Example 5
Samples of the work of 20 clerks are gathered. This company fully examines
100 records entered by each clerk and counts the number of errors. They also
computes the fraction of defective in each sample and uses +/-3SD (99.73%) to
set control limits

Heizer et. al, 2020

WRWH/MGT6240/Chapter 6S 44
 Setting Control Limits: Example 5
𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑜𝑜𝑜𝑜 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 80
𝑝𝑝̅ = = = 0.04
𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑜𝑜𝑜𝑜 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 (100)(20)

(0.04)(1 − 0.04)
𝜎𝜎
�𝑝𝑝 = = 0.02
100

𝐿𝐿𝐿𝐿𝐿𝐿 = 𝑝𝑝̅ − 𝑧𝑧�

𝜎𝜎𝑝𝑝
= 0.04 −3(0.02) = 0
𝑈𝑈𝑈𝑈𝑈𝑈 = 𝑝𝑝̅ + 𝑧𝑧�
𝜎𝜎𝑝𝑝
= 0.04 + 3(0.02) = 0.10
Heizer et. al, 2020
WRWH/MGT6240/Chapter 6S 45
Setting Control Limits: Example 5
Possible assignable
.11 –
causes present
.10 – UCLp = 0.10
Fraction defective

.09 –
.08 –
.07 –
.06 –
.05 –
.04 – p = 0.04
.03 –
.02 –
.01 – LCLp = 0.00
| | | | | | | | | |
.00 –
2 4 6 8 10 12 14 16 18 20
Sample number

Heizer et. al, 2020

WRWH/MGT6240/Chapter 6S 46
Setting Chart Limits:
c - Chart
Population will be a Poisson distribution, but applying the
Central Limit Theorem allows us to assume a normal
distribution for the sample statistics

𝐿𝐿𝐿𝐿𝐿𝐿 = 𝑐𝑐̅ − 3 𝑐𝑐̅

𝑈𝑈𝑈𝑈𝑈𝑈 = 𝑐𝑐̅ + 3 𝑐𝑐̅
Where
𝑁𝑁𝑁𝑁. 𝑜𝑜𝑜𝑜 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑
𝑐𝑐 ̅ = mean number of defects per unit being sampled =
𝑁𝑁𝑁𝑁. 𝑜𝑜𝑜𝑜 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠
𝑐𝑐 ̅ = standard deviationof defects per unit being sampled

Heizer et. al, 2020

WRWH/MGT6240/Chapter 6S 47
 Setting Control Limits: Example 6
Total number of complaints received by a cab company
within the past nine days is 54. Using 99.73% confidence
level, set the control limits.
54
𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑛𝑛𝑛𝑛. 𝑜𝑜𝑜𝑜 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 𝑝𝑝𝑝𝑝𝑝𝑝 𝑑𝑑𝑑𝑑𝑑𝑑 = 𝑐𝑐 ̅ = = 6 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐
9
14 – UCLc =
𝐿𝐿𝐿𝐿𝐿𝐿 = 𝑐𝑐̅ − 3 𝑐𝑐̅
Number defective

12 –
10 – =6−3 6
8 – =
6 – c= 6
𝑈𝑈𝑈𝑈𝑈𝑈 = 𝑐𝑐̅ + 3 𝑐𝑐̅
4 –
2 – LCLc =
=6+3 6
| | | | | | | | |
0 – 1 2 3 4 5 6 7 8 9 = 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐
Day Heizer et. al, 2020
WRWH/MGT6240/Chapter 6S 48
Use of p-charts vs c-charts
1. p-charts
 When observations can be placed into 2 categories
 Examples: Good or bad; Pass or fail; Operate or don’t operate
 To count defective records in a sample
 proportion of defective outputs in a sample

2. c-charts
 When the number of occurrences per unit of measure can
be counted; non-occurrences cannot be counted
 Scratches, chips, dents, or errors per item
 Cracks or faults per unit of distance
 Breaks or tears per unit of area
 Bacteria or pollutants per unit of volume
 Calls, complaints, failures per unit of time
To control number of defects per unit of outputs in a sample
 Helpful for monitoring processes having large potential errors but
actual number of errors is relatively small
WRWH/MGT6240/Chapter 6S 49
 Managerial Issues and
Control Charts
1. Select points in the processes that need SPC
 Critical parts? Parts with tendency to be out of control?

2. Determine sample sizes to take

3. Determine the appropriate charting technique
a) Variables; to monitor weights, dimensions
b) Attributes; more of “Yes-No” or “go-no go”; tend to be less costly to
implement

4. Set clear policies and procedures on SPC to be

implemented

WRWH/MGT6240/Chapter 6S 50
Which Control Charts?

WRWH/MGT6240/Chapter 6S 51
 Run Test
 A test for randomness
 Control charts test for points that are too extreme to be
considered random (s.t. SD)
 Even when all points are within the control limits, the process may
not be random
 Any sort of pattern in the data would also suggest a non-random
process
 Runs of 5 or 6 points above or below the target or centreline suggest
assignable causes may be present

 Conducted in addition to control charts

 Better able to detect abnormalities in process

 Run
 A sequence of observations with a certain characteristic followed
by another with different characteristic
WRWH/MGT6240/Chapter 6S 52
 Nonrandom Patterns in
Control Charts
 When a process is stable or in statistical control
 The output generated will exhibit random variability over a
time period
 Presence of patterns however indicates existence of
assignable causes of variation despite all points within
control limits
i. Trend:
ii. Cycles:
iii. Bias:
iv. Mean shift:
v. Too much dispersion:

Stevenson & Chuong, 2014

WRWH/MGT6240/Chapter 6S 53
 Patterns in Control Charts
 Observe pattern by counting runs
 2 useful tests:
i. Examine the number of runs
up and down
(Denoted by series of Us and Ds)
ii. Examine the number of runs
above and below the median
(Denoted by series of As and Bs)

Stevenson & Chuong, 2014

WRWH/MGT6240/Chapter 6S 54
Patterns in Control Charts

WRWH/MGT6240/Chapter 6S 55
 Nonrandom Variation
 Managers should have response plans to
investigate cause
 May be false alarm (Type I error)
 To resume the process but monitor for a while to confirm

 May be assignable variation

 A real indication for the presence of problem
 To be immediately addressed

 If it is a good result (below LCL on p-chart, c-chart,

R-chart)
 To change the process to achieve similar results continuously

WRWH/MGT6240/Chapter 6S 56
 Process Capability
 Process capability
 Is a measure of the relationship between the natural variation of
the process and the design specifications
Variability of a process can significantly impact quality
 terms to refer to process output
a) Tolerances or specifications; output
 Range of acceptable values by engineering design or
customer requirements for output to be acceptable
b) Process variability; process
 Natural or inherent (random) variability in a process
 Measured in terms of the process SD
c) Process capability (Control limits for sampling); output
 Inherent variability of process output relative to variation
allowed by design specification
 Statistical limits within which sample statistics allowed to vary
WRWH/MGT6240/Chapter 6S 57
 Process Capability
 Concerns with the ability of a process to meet design
specifications (engineering specification limits or standards)
 Process is “capable” when it is within the specifications for the
process output

 SPC concerns with keeping a process in control

 Natural variation of a process must be stable
 The natural variation of a process should be small enough to
produce products that meet the standards required

 A process in statistical control does not necessarily

meet the design specifications
 Outputs being produced by a stable process may not conform to
specifications

WRWH/MGT6240/Chapter 6S 58
 Process Capability
Lower Upper
Specification Specification

A. Process variability
matches specifications
Lower Upper
Specification Specification

B. Process variability
Lower Upper
well within specifications Specification Specification

C. Process variability
exceeds specifications
Stevenson & Chuong, 2014, Figure 10.15
WRWH/MGT6240/Chapter 6S 59
Measures of Process Capability
 2 popular measures for quantitatively determining if a
process is capable
a) Process Capability Ratio (Cp)
b) Process Capability Index (Cpk)
 A capable process
 Its values must fall within upper and lower specifications
 Usually within +/- 3 SDS from process mean
 Process variability is the key factor in process capability
 Measured in terms of the process SD
 To determine if the process is capable  compare +/-3SD (i.e. 6
sigma) of the process to the specifications for the process
 Does not look at how well the process is centered in the
specification range
WRWH/MGT6240/Chapter 6S 60
Process Capability Ratio
If the process is centered, use Cp
𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 𝑊𝑊𝑊𝑊𝑊𝑊𝑊𝑊𝑊
𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶 𝐶𝐶𝑝𝑝 =
𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝑊𝑊𝑊𝑊𝑊𝑊𝑊𝑊𝑊

𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈 𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 − 𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿

𝐶𝐶𝑝𝑝 =
6𝜎𝜎

If the process is not centered, use Cpk

𝑥𝑥̅ − 𝐿𝐿𝐿𝐿𝐿𝐿 𝑈𝑈𝑇𝑇𝑇𝑇 − 𝑥𝑥̅

𝐶𝐶𝑝𝑝𝑝𝑝 = 𝑀𝑀𝑀𝑀𝑀𝑀 𝑜𝑜𝑜𝑜
3𝜎𝜎 3𝜎𝜎

WRWH/MGT6240/Chapter 6S 61
Process Capability Ratio
𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈 𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 − 𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿
𝐶𝐶𝑝𝑝 =
6𝜎𝜎

 A ratio for determining if a process meets design

specifications
 Relates to the spread (dispersion) of the process output relative
to its tolerance
 A capable process must have a 𝑪𝑪𝒑𝒑 ≥ 𝟏𝟏. 𝟎𝟎
 Common target used Cp = 1.33 (4 sigma or 8SD) to allow for off-
center processes
 6 Sigma (12 SD) quality standard requires a Cp = 2.0
 Only 3.4 parts per million can be expected to be out of
specification (3.4 defective parts per million) = very close to 0
defects!
WRWH/MGT6240/Chapter 6S 62
Process Capability Ratio: Example 7
Insurance claims process
Process mean x = 210.0 minutes
Process standard deviation = 0.516 minutes
Design specification = 210 ± 3 minutes

𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈𝑈 𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 − 𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿

𝐶𝐶𝑝𝑝 =
6𝜎𝜎
(210 + 3) − (210 − 3) 213 − 207
𝐶𝐶𝑝𝑝 = = =
6(0.516) 3.096

Process is
capable
WRWH/MGT6240/Chapter 6S 63
Process Capability Index
 Measures the difference between desired and
actual dimensions of goods and services produced
 A proportion of variation (3𝜎𝜎) between the center of the process
and the nearest specification limit

 A capable process is not necessarily in the center

of the specification
 But it falls within the specification limit at both extremes

 A capable process must have a 𝑪𝑪𝒑𝒑𝒌𝒌 ≥ 𝟏𝟏. 𝟎𝟎

WRWH/MGT6240/Chapter 6S 64
 Process Capability Index
𝑼𝑼𝑼𝑼𝑼𝑼𝑼𝑼𝑼𝑼 𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳
𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺 − �
𝒙𝒙 �
𝒙𝒙 − 𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺
𝑪𝑪𝒑𝒑𝒑𝒑 = 𝑴𝑴𝑴𝑴𝑴𝑴𝑴𝑴𝑴𝑴𝑴𝑴𝑴𝑴 𝒐𝒐𝒐𝒐 𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳 , 𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳
𝟑𝟑𝝈𝝈 𝟑𝟑𝝈𝝈

 A capable process must have a 𝑪𝑪𝒑𝒑𝒌𝒌 ≥ 𝟏𝟏. 𝟎𝟎

a) When Cpk is 1.0 for both USL and LSL
 Process variation is centered  process is capable of producing
within +/- 3 SDs  < 2700 defects/million
b) When Cpk is 2.0
 Process is capable of producing < 3.4 defects/million

WRWH/MGT6240/Chapter 6S 65
Process Capability Ratio: Example 8
You are the process improvement manager and have developed a new machine to
cut insoles for the company’s top-of-the-line running shoes. You are excited
because the company’s goal is no more than 3.4 defects per million, and this
machine may be the innovation you need. The insoles cannot be more than +/-
0.001 of an inch from the required thickness of 0.250”. The standard deviation for
the process is 0.0005 inches. You want to know if you should replace the existing
machine, which has a Cpk of 1.0.

New process mean: 𝑥𝑥̅ = 0.250 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖

Process standard deviation: 𝜎𝜎 = 0.0005 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖
Upper Specification Limit: 𝑈𝑈𝑈𝑈𝑈𝑈 = 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖
Lower Specification Limit: 𝑈𝑈𝑈𝑈𝑈𝑈 = 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖
𝑼𝑼𝑼𝑼𝑼𝑼𝑼𝑼𝑼𝑼 𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳
�
𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺 − 𝒙𝒙 � − 𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺
𝒙𝒙
𝑪𝑪𝒑𝒑𝒑𝒑 = 𝑴𝑴𝑴𝑴𝑴𝑴𝑴𝑴𝑴𝑴𝑴𝑴𝑴𝑴 𝒐𝒐𝒐𝒐 𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳 , 𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳
𝟑𝟑𝝈𝝈 𝟑𝟑𝝈𝝈

WRWH/MGT6040/Chapter 6S 66
Process Capability Ratio: Example 8
New Cutting Machine
New process mean: 𝑥𝑥̅ = 0.250 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖
Process standard deviation: 𝜎𝜎 = 0.0005 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖
Upper Specification Limit: 𝑈𝑈𝑈𝑈𝑈𝑈 = 0.251 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖
Lower Specification Limit: 𝑈𝑈𝑈𝑈𝑈𝑈 = 0.249 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖
𝑼𝑼𝑼𝑼𝑼𝑼𝑼𝑼𝑼𝑼 𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳
�
𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺 − 𝒙𝒙 � − 𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺
𝒙𝒙
𝑪𝑪𝒑𝒑𝒑𝒑 = 𝑴𝑴𝑴𝑴𝑴𝑴𝑴𝑴𝑴𝑴𝑴𝑴𝑴𝑴 𝒐𝒐𝒐𝒐 𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳 , 𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳
𝟑𝟑𝝈𝝈 𝟑𝟑𝝈𝝈
0.251 − 0.250 0.250 − 0.249
Cpk = Minimum of ,
3(0.0005) 3(0.0005)

Cpk = Both results in

Process is
NOT capable
WRWH/MGT6240/Chapter 6S 67
Process Capability Ratio: Example 9

𝐶𝐶𝐶𝐶 > 1.33 𝑖𝑖𝑖𝑖 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑

𝐶𝐶𝐶𝐶 = 1.00 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑖𝑖𝑖𝑖 𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐
𝐶𝐶𝐶𝐶 < 1.00 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑖𝑖𝑖𝑖 𝑛𝑛𝑛𝑛𝑛𝑛 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐

Stevenson & Chuong, 2014

WRWH/MGT6240/Chapter 6S 68
 Control Limits vs
Specification Limits
 Control limits = process variability (stability)
 Designed to monitor output over time to ensure that the system
continues to produce consistent output
 Output produced outside of control limits sends a signal that
something extraordinary has occurred
 the firm should investigate

 Specification limits = process capability

 Fixed engineering measurements that define exactly what
determines acceptable output or not
 It is possible for output to exceed control limits and still be within
spec limits
It is possible for output that exceeds spec limits be appear to be
within statistical control
WRWH/MGT6240/Chapter 6S 69
 Limitations of
Process Indices
1. Process may not be stable
2. Process output may not be normally distributed
3. Process not centered but Cp is used

WRWH/MGT6240/Chapter 6S 70
Improving Process Capability
1. Simplify
2. Standardize
3. Mistake-proof
4. Upgrade equipment
5. Automate

WRWH/MGT6240/Chapter 6S 71
Automated Inspection
 Modern
technologies allow
virtually 100%
inspection at
minimal costs
 Not suitable for all
situations
Laser tracking device by Faro technologies enables quality control
personnel to measure and inspect parts and tools during production.
The portable tracker can measure objects from 262 feet away and
takes up to 1000 accurate readings per second

WRWH/MGT6240/Chapter 6S 72
Making Vaccine
Bottles
What are vision systems and
why are they useful OM tool?

Which of the quality control

tools (Chapter 6) could vial
Producers employ?

WRWH/MGT6240/Chapter10 73