Statistical Applications in Quality and

Productivity Management
 Chapter Goals

After completing this chapter, you should be

able to:
 Describe the concepts of Total Quality Management and
Six Sigma® Management
 Explain process variability and the theory of control
charts
 Construct and interpret p charts
 Construct and interpret X and R charts
 Obtain and explain measures of process capability
Chapter Overview
 Total Quality Management

 Primary focus is on process improvement

 Most variation in a process is due to the
system, not the individual
 Teamwork is integral to quality management
 Customer satisfaction is a primary goal
 Organization transformation is necessary
 It is important to remove fear
 Higher quality costs less
 Deming’s 14 Points
1. Create a constancy of purpose toward
improvement
 become more competitive, stay in business, and provide jobs

2. Adopt the new philosophy

 Better to improve now than to react to problems later

3. Stop depending on inspection to achieve

quality -- build in quality from the start
 Inspection to find defects at the end of production is too late

4. Stop awarding contracts on the basis of low

bids
 Better to build long-run purchaser/supplier relationships
 Deming’s 14 Points(continued)
5. Improve the system continuously to improve
quality and thus constantly reduce costs
6. Institute training on the job
 Workers and managers must know the difference between
common cause and special cause variation
7. Institute leadership
 Know the difference between leadership and supervision

8. Drive out fear so that everyone may work

effectively.
9. Break down barriers between departments so
that people can work as a team.
 Deming’s 14 Points (continued)
 10. Eliminate slogans and targets for the
workforce
 They can create adversarial relationships
 11. Eliminate quotas and management by
numerical goals
 12. Remove barriers to pride of workmanship
 13. Institute a vigorous program of education
and self-improvement
 14. Make the transformation everyone’s job
The Shewhart-Deming Cycle

Plan

The
Act Deming Do
Cycle
The key is a
continuous
cycle of
Study improvement
 Six Sigma® Management

A method for breaking a process into a series of

steps:
 The goal is to reduce defects and produce near
perfect results
 The Six Sigma® approach allows for a shift of as
much as 1.5 standard deviations, so is
essentially a ±4.5 standard deviation goal
 The mean of a normal distribution ±4.5 standard
deviations includes all but 3.4 out of a million
 The Six Sigma® DMAIC Model
DMAIC represents
 Define -- define the problem to be solved; list
costs, benefits, and impact to customer
 Measure – need consistent measurements for
each Critical-to-Quality characteristic
 Analyze – find the root causes of defects
 Improve – use experiments to determine
importance of each Critical-to-Quality variable
 Control – maintain gains that have been made
 Theory of Control Charts
 A process is a repeatable series of steps
leading to a specific goal
 Control Charts are used to monitor variation in
a measured value from a process
 Inherent variation refers to process variation
that exists naturally. This variation can be
reduced but not eliminated
 Theory of Control Charts
(continued)
 Control charts indicate when changes in data
are due to:
 Special or assignable causes
 Fluctuations not inherent to a process
 Represents problems to be corrected
 Data outside control limits or trend
 Chance or common causes
 Inherent random variations
 Consist of numerous small causes of random
variability
 Process Variation

Total Process Common Cause Special Cause

Variation = Variation + Variation

 Variation is natural; inherent in the world

around us
 No two products or service experiences
are exactly the same
 With a fine enough gauge, all things can
be seen to differ
 Total Process Variation

Total Process Common Cause Special Cause

Variation = Variation + Variation

Variation is often due to differences in:

 People
 Machines
 Materials
 Methods
 Measurement
 Environment
 Common Cause Variation

Total Process Common Cause Special Cause

Variation = Variation + Variation

Common cause variation

 naturally occurring and expected
 the result of normal variation in
materials, tools, machines, operators,
and the environment
 Special Cause Variation

Total Process Common Cause Special Cause

Variation = Variation + Variation

Special cause variation

 abnormal or unexpected variation
 has an assignable cause
 variation beyond what is considered
inherent to the process
 Control Limits
Forming the Upper control limit (UCL) and the
Lower control limit (LCL):
UCL = Process Average + 3 Standard Deviations
LCL = Process Average – 3 Standard Deviations
Control Chart Basics
Process Variability
 Using Control Charts
 Control Charts are used to check for process
control
H0: The process is in control
i.e., variation is only due to common causes

H1: The process is out of control

i.e., special cause variation exists

 If the process is found to be out of control,

steps should be taken to find and eliminate the
special causes of variation
 In-control Process

 A process is said to be in control when the

control chart does not indicate any out-of-control
condition
 Contains only common causes of variation
 If the common causes of variation is small, then
control chart can be used to monitor the process
 If the common causes of variation is too large, you
need to alter the process
 Process In Control

 Process in control: points are randomly

distributed around the center line and all points
are within the control limits
 Process Not in Control

Out of control conditions:

 One or more points outside control limits

 8 or more points in a row on one side of the
center line
 8 or more points moving in the same
direction
Process Not in Control
(continued)
 Out-of-control Processes
 When the control chart indicates an out-of-
control condition (a point outside the control
limits or exhibiting trend, for example)
 Contains both common causes of variation and
assignable causes of variation
 The assignable causes of variation must be identified
 If detrimental to the quality, assignable causes of variation
must be removed
 If increases quality, assignable causes must be incorporated
into the process design
Statistical Process Control Charts
Statistical
Process Control
Charts

p chart X chart and R

chart
Used for Used for
proportions measured
(attribute data) numeric data
 p Chart Example

You are the manager of a 500-room hotel.

You want to achieve the highest level of
service. For seven days, you collect data on
the readiness of 200 rooms. Is the process in
control?
 p Chart

 Control chart for proportions

 Is an attribute chart
 Shows proportion of nonconforming items
 Example -- Computer chips: Count the number of
defective chips and divide by total chips inspected
 Chip is either defective or not defective
 Finding a defective chip can be classified a
“success”
 p Chart (continued)

 Used with equal or unequal sample sizes

(subgroups) over time
 Unequal sizes should not differ by more than ±25%
from average sample sizes
 Easier to develop with equal sample sizes
 Should have np > 5 and n(1 - p) > 5
 Creating a p Chart
 Calculate subgroup proportions
 Graph subgroup proportions
 Compute average proportion
 Compute the upper and lower control limits
 Add centerline and control limits to graph
 p Chart-Example

Subgroup Sample Number of Sample

number size successes Proportion, ps
1 150 15 10.00
2 150 12 8.00
3 150 17 11.33
… … …
Average
subgroup
proportion = p
 Average of Subgroup Proportions
The average of subgroup proportions = p
It is a general formula (equal or unequal sample sizes):

X i
p  ik1
n
i1
i

where:
Xi = the number of nonconforming items in sample i
ni = total number of items sampled in k samples
 Computing Control Limits
 The upper and lower control limits for a p chart
are
UCL = Average Proportion + 3 Standard Deviations
LCL = Average Proportion – 3 Standard Deviations

 The standard deviation for the subgroup

proportions is
(p)(1 p)
n
 Computing Control Limits
(continued)

 The upper and lower control limits for the

p chart are

p(1  p)
UCL p  3
n
Proportions are
p(1  p) never negative, so
LCL p  3 if the calculated
n lower control limit
is negative, set
LCL = 0
 p Chart- Example

You are the manager of a 500-room hotel.

You want to achieve the highest level of
service. For seven days, you collect data on
the readiness of 200 rooms. Is the process in
control?
 p Chart Example:
Hotel Data
# Not
Day # Rooms Ready Proportion
1 200 16 0.080
2 200 7 0.035
3 200 21 0.105
4 200 17 0.085
5 200 25 0.125
6 200 19 0.095
7 200 16 0.080
 p Chart
Control Limits Solution
k

X i
16  7    16 121
p i 1
k
  .0864
200  200    200 1400
n
i 1
i

n i
200  200    200
n i 1
 200
k 7

p(1  p) .0864(1  .0864)

UCL p  3 .0864  3 .1460
n 200

p(1  p) .0864(1  .0864)

LCL p  3 .0864  3 .0268
n 200
 p Chart
Control Chart Solution

Individual points are distributed around _p without any pattern.

Any improvement in the process must come from reduction
of common-cause variation, which is the responsibility of
management.
 Understanding Process Variability:
Red Bead Experiment
The experiment:
 From a box with 20% red beads and 80% white
beads, have “workers” scoop out 50 beads
 Tell the workers their job is to get white beads
 10 red beads out of 50 (20%) is the expected
value. Scold workers who get more than 10,
praise workers who get less than 10
 Some workers will get better over time, some
will get worse
 Morals of the
Red Bead Experiment
1. Variation is an inherent part of any process.
2. The system is primarily responsible for worker
performance.
3. Only management can change the system.
4. Some workers will always be above average,
and some will be below.
 R chart and X chart
 Used for measured numeric data from a
process
 Subgroups usually contain 3 to 6
observations each
 For the process to be in control, both the R
chart and the X-bar chart must be in control
 Example: Subgroups

 Process measurements: for example,

thickness of steel plates after rolling
 The R Chart
 Monitors variability in a process
 The characteristic of interest is measured
on a numerical scale
 Is a variables control chart
 Shows the sample range over time
 Range = difference between smallest and
largest values in the subgroup
 Steps to create an R chart

 Find the mean of the subgroup ranges (the

center line of the R chart)
 Compute the upper and lower control limits
for the R chart
 Use lines to show the center and control
limits on the R chart
 Plot the successive subgroup ranges as a
line chart
Average of Subgroup Ranges

Average of subgroup ranges:

R
 R i

k
where:
Ri = ith subgroup range
k = number of subgroups
 R Chart Control Limits

 The upper and lower control limits for an

R chart are

UCL D 4 ( R )
LCL D3 ( R )
where:
D4 and D3 are taken from the table
for subgroup size = n
R Chart Control Limits-
Constants
 R Chart Example

You are the manager of a 500-room hotel.

You want to analyze the time it takes to deliver
luggage to the room. For 7 days, you collect
data on 5 deliveries per day. Is the variation in
the process in control?
 R Chart Example:
Subgroup Data
Day Subgroup Subgroup Subgroup
Size Average Range
1 5 5.32 3.85
2 5 6.59 4.27
3 5 4.89 3.28
4 5 5.70 2.99
5 5 4.07 3.61
6 5 7.34 5.04
7 5 6.79 4.22
 R Chart Center and
Control Limits

R
 R i

3.85  4.27    4.22
3.894
k 7

UCL D 4 ( R ) (2.114 )(3.894) 8.232

LCL D3 ( R ) (0)(3.894) 0

D4 and D3 are from TABLE (Factors for

Constructing Variables Control Charts
Table (n = 5)
 R Chart
Control Chart Solution
 The X Chart

 Shows the means of successive subgroups over

time
 Monitors process average
 Must be preceded by examination of the R chart
to make sure that the variation in the process is
in control (Check R chart First)
 Steps to create an X chart

 Compute the mean of the subgroup means

(the center line of the X chart)
 Compute the upper and lower control limits
for the X chart
 Graph the subgroup means
 Add the center line and control limits to the
graph
 Average of Subgroup
Means

Average of subgroup means:

X
 X i

where:
Xi = ith subgroup average
k = number of subgroups
 Computing Control
Limits(continued)
 The upper and lower control limits for an X chart
are generally defined as
UCL = Process Average + 3 Standard Deviations
LCL = Process Average – 3 Standard Deviations

 so
UCL  X  A 2 ( R )

LCL  X  A 2 ( R )
 X Chart Example

You are the manager of a 500-room hotel.

You want to analyze the time it takes to deliver
luggage to the room. For seven days, you
collect data on five deliveries per day. Is the
process average in control?
 X Chart Example:
Subgroup Data
Day Subgroup Subgroup Subgroup
Size Average Range
1 5 5.32 3.85
2 5 6.59 4.27
3 5 4.89 3.28
4 5 5.70 2.99
5 5 4.07 3.61
6 5 7.34 5.04
7 5 6.79 4.22
 X Chart
Control Limits Solution

X
 X i

5.32  6.59    6.79
5.813
k 7

R
 R i

3.85  4.27    4.22
3.894
k 7

UCL  X  A 2 ( R ) 5.813  (0.577)(3.894) 8.060

LCL  X  A 2 ( R ) 5.813  (0.577)(3.894) 3.566

A2 is from Table
(n = 5)
 X Chart
Control Chart Solution
X-bar and S Control Charts
 X-bar and S Control
Charts(continued)
3-sigma control limits for S (m is the number of subgroups)
when the subgroup sizes are the same use the followings :

x When using S
 X-bar and S Control Charts
(continued)
3-sigma control limits for S (m is the number of subgroups, n is a subgroup
size), when the subgroup sizes are different use the followings :

x When using S
 X-bar and S Control Charts-
(continued)
When the subgroup sizes are different ……
A3, B3 and B4 used to calculate control limits on the x-bar and s
control charts can be found using the average sample size . If the
subgroup sizes are not very different, this approach may be
satisfactory in some situations; it is particularly helpful if the charts
are to be used in a presentation to management. Since the average
sample size ni may not be an integer, you may round it in away that
a conservative control limit is achieved or use the most common
sample size to use A3, B3 and B4 .
 Example

Use S chart and X bar chart

to analyse the given process,
pertaining to the inside
diameter measurements for
car engine piston rings
Example (S Chart)
Example (X bar Chart)
 Control Charts in Minitab

Lets review the following tutorials:

https://www.youtube.com/watch?v=BsNh7li6Dfs

https://www.youtube.com/watch?v=-O9Q4Z-nmfI

https://www.youtube.com/watch?v=gBCrvyeSpMU
 Process Capability
 Process capability is the ability of a process to
consistently meet specified customer-driven
requirements
 Specification limits are set in response to
customers’ expectations
 The upper specification limit (USL) is the largest
value that can be obtained and still conform to
customers’ expectations
 The lower specification limit (LSL) is the
smallest value that is still conforming
 Estimating Process Capability

 Must first have an in-control process

 Estimate the percentage of product or service
within specification
 Assume the population of X values is
approximately normally distributed with mean
estimated by X and standard deviation
estimated by R / d2
 Estimating Process
Capability(continued)
 For a characteristic with a LSL and a USL, the
probability that the output product is within
specification is shown by P :
P(outcome will be within specifications)
 
 
 LSL  X USL  X 
P(LSL  X  USL)  P Z 
R R
 
 d2 d2 
 Where Z is a standardized normal random variable
 Estimating Process Capability
(continued)

 For a characteristic with only an USL (for example,

the size of your product must be smaller than USL)
P(outcome will be within specifications)
 
 
 USL  X 
P( X  USL)  P Z  
R
 
 d2 
 Where Z is a standardized normal random variable
 Estimating Process Capability-
(continued)

 For a characteristic with only a LSL (for example,

the size of your product must be larger than LSL)
P(outcome will be within specifications)
 
 
 LSL  X 
P(LSL  X)  P  Z
R
 
 d2 
 Where Z is a standardized normal random variable
Variable control charts-
Constants
 Process Capability
Example
You are the manager of a 500-room hotel.
You have instituted a policy that 99% of all
luggage deliveries must be completed within
ten minutes or less. For seven days, you
collect data on five deliveries per day. You
know from prior analysis that the process is
in control. Is the process capable?
 Process Capability:
Hotel Data
Day Subgroup Subgroup Subgroup
Size Average Range
1 5 5.32 3.85
2 5 6.59 4.27
3 5 4.89 3.28
4 5 5.70 2.99
5 5 4.07 3.61
6 5 7.34 5.04
7 5 6.79 4.22
 Process Capability:
Hotel Example Solution
n 5 X 5.813 R 3.894 d2 2.326

P(outcome will be within specifications)

 
 10  5.813 
P( X  10)  P Z  
 3.894 
 
 2.326 
P( Z  2.50) .9938
Therefore, we estimate that 99.38% of the luggage deliveries
will be made within the ten minutes or less specification. The
process is capable of meeting the 99% goal.
 Capability Indices

 A process capability index is an aggregate

measure of a process’s ability to meet
specification limits
 The larger the value, the more capable a
process is of meeting requirements
 Cp Index

 Cp index is a simple process capability index

that relates the allowable spread of the spec
limits (difference between USL, and LSL) to
the measure of the actual, or natural, variation
of the process, represented by 6*sigma,
where sigma is the estimated process
standard deviation
 Cp Index(continued)
While Cp relates the spread of the process relative to
the specification width, it does not address how well
the process average, is centered to the target value.
Cp is often referred to as process “potential”
 Cp Index (continued)

USL  LSL specification spread

Cp  
6(R / d2 ) process spread

 Cp > 1 implies a process has the potential of having

more than 99.73% of outcomes within specifications
 Cp > 2 implies a process has the potential of meeting
the expectations set forth in six sigma management
Cp Index-(continued)
 Cp Index- (continued)
In general :
 Cp<1 means the process variation exceeds specification, and a significant number
of defects are being made.
 Cp=1 means that the process is just meeting specifications. A minimum of .3%
defects will be made and more if the process is not centered.
 Cp>1 means that the process variation is less than the specification, however,
defects might be made if the process is not centered on the target value ( the below
figure, the right hand side curve is an example of CP>1 and centered, but
sometimes the curve is not centered)
 Example:

If a coating process is sampled with USL = 0.10 mm,

and LSL = 0.05 mm, and the standard deviation is 0.01
mm, then calculate and interpret Cp ?
 CPL and CPU
 To measure capability in terms of actual process
performance:
X  LSL
CPL 
3(R / d2 )

USL  X
CPU 
3(R / d2 )

 CPU and CPL are used for one-sided specification

limits
 Use CPU when a characteristic only has a USL
 Use CPL when a characteristic only has an LSL
 CPL and CPU (continued)

 CPL > 1 implies that the process mean is more than

3 standard deviation away from the lower
specification limit.

 CPU>1 implies that the process mean is more than 3

standard deviation away from the upper specification
limit
 Process Capability-Example
You are the manager of a 500-room hotel. You
have instituted a policy that all luggage
deliveries must be completed within ten
minutes or less. For seven days, you collect
data on five deliveries per day. You know
from prior analysis that the process is in
control. Compute an appropriate capability
index for the delivery process and interpret
that.
 X Chart Example-
Subgroup Data
Day Subgroup Subgroup Subgroup
Size Average Range
1 5 5.32 3.85
2 5 6.59 4.27
3 5 4.89 3.28
4 5 5.70 2.99
5 5 4.07 3.61
6 5 7.34 5.04
7 5 6.79 4.22
 Process Capability-
Hotel Example Solution
n 5 X 5.813 R 3.894 d2 2.326

USL  X 10  5.813
CPU   .833672
3(R / d2 ) 3(3.894 / 2.326)

Since there is only the upper specification limit, we

need to only compute CPU. The capability index for
the luggage delivery process is .8337, which is less
than 1. The upper specification limit is less than 3
standard deviation above the mean.
 Cpk Index
 The most commonly used capability index is the
Cpk index
 Measures actual process performance for
characteristics with two-sided specification limits
Cpk = min(CPL, CPU)

 Cpk = 1 indicates that the process average is 3

standard deviation away from the closest specification
limit.
 Cpk ranges
Larger Cpk indicates greater capability of meeting
the requirements.

Cpk > 2 indicates compliance with six sigma management.

 Example

A bagging operation is designed to place 50 pounds of sand into each bag.

The specifications for the operation are a minimum of 49.5 pounds and a
maximum of 50.5 pounds. So, the lower specification limit (LSL) is 49.5. The
upper specification limit (USL) is 50.5 pounds. The operation is being
monitored using an Xbar-R chart with a subgroup size of 4. Each hour, four
consecutive bags are weighed. The standard deviation (from the range
chart data) is 0.212. The average bag weight is 50.05.
Calculate CP and CPK and interpret that :
 Example-solution

CP= (50.5-49.5)/ 6*0.212 =0.786

 Chapter Summary
 Reviewed the philosophy of quality management
 Deming’s 14 points
 Discussed Six Sigma® Management
 Reduce defects to no more than 3.4 per million
 Uses DMAIC model for process improvement
 Discussed the theory of control charts
 Common cause variation vs. special cause variation
 Constructed and interpreted p charts
 Constructed and interpreted X and R charts
 Obtained and interpreted process capability measures