Transcript Chapter 17 - McGraw Hill Higher Education

A PowerPoint Presentation Package to Accompany
Applied Statistics in Business &
Economics, 4th edition
David P. Doane and Lori E. Seward
Prepared by Lloyd R. Jaisingh
McGraw-Hill/Irwin
Copyright © 2013 by The McGraw-Hill Companies, Inc. All rights reserved.
Chapter 17
Quality Management
Chapter Contents
17.1 Quality and Variation
17.2 Pioneers in Quality Management
17.3 Quality Improvement
17.4 Control Charts: Overview
17.5 Control Charts for a Mean
17.6 Control Charts for a Range
17.7 Other Control Charts
17.8 Patterns in Control Charts
17.9 Process Capability
17.10 Additional Quality Topics (Optional)
17-2
Chapter 17
Quality Management
Chapter Learning Objectives
LO17-1:
LO17-2:
LO17-3:
LO17-4:
LO17-5:
LO17-6:
LO17-7:
LO17-8:
Define quality and explain how it may be measured.
Distinguish between common cause variation and special cause variation.
Name key individuals and their contributions to the quality movement.
List common statistical tools used in quality improvement.
List steps toward continuous quality improvement and variance reduction.
Make and interpret common control charts (x-bar, R, and p).
Recognize abnormal patterns in control charts and their potential causes.
Assess the capability of a process.
17-3
Chapter 17
LO17-1
17.1 Quality and Variation
LO17-1: Define quality and explain how it may be measured.
What is Quality?
•
Quality may be measured as •
Quality includes these attributes:
- a physical metric
- Conformance to specifications.
- an aesthetic attribute
- a functional characteristic
- Performance in the intended use.
- As near-zero defects as possible.
- Reliability and durability.
- a personal attribute
- Serviceability when needed.
- an efficiency attribute
- Favorable customer perceptions.
17-4
LO17-2:
Chapter 17
LO17-2
17.1 Quality and Variation
Distinguish between common cause variation and special
cause variation.
Common Cause versus Special Cause
•
•
•
Common cause variation
(random “noise”) is
- normal and expected and
- present in any stable,
in-control process.
Special cause variation is due
to factors that are abnormal
and require investigation.
Special cause variation must be
eliminated in order for the
process to be in control.
•
Sources of variation
in processes include
- human abilities
- training
- motivation
- technology
- materials
- management
- organization
17-5
LO17-3:
Chapter 17
17.2 Pioneers in Quality Management
LO17-3
Name key individuals and their contributions to the quality
movement.
Brief History of Quality Control
Early 1900s
•
Quality took the form of improved inspection and improvement in the methods of mass
production.
1920 – just after WWII
•
Walter A. Shewhart – process control charts
Harold F. Dodge and Harry G. Romig – acceptance sampling from lots
1950s and 1960s
•
1970s
1980s
W. Edwards Deming and Joseph M. Juran
train Japanese manufacturers to become high-quality producers by applying American
quality control techniques.
• Genichi Taguchi and Kaoru Ishikawa, Japanese statisticians also train Japanese
manufacturers.
• North American firms had lost their initial leadership in quality control.
• Japanese devised and perfected new quality improvement methods.
•
•
•
Deming, Juran, and Armand Feigenbaum advise North American firms in quality
improvement and Japanese lean production methods.
Japanese push quality frontier forward under the teachings of Taguchi and the perfection of
the Kaizen philosophy of continuous improvement.
Europeans articulated the ISO 9000 standards.
17-6
Chapter 17
LO17-3
17.2 Pioneers in Quality Management
W. Edward Deming
17-7
Chapter 17
LO17-3
17.2 Pioneers in Quality Management
W. Edward Deming’s 14 Points
17-8
Chapter 17
LO17-4
17.3 Quality Improvement
LO17-4: List common statistical tools used in quality improvement.
Business Quality Philosophies (Total Quality Management (TQM))
•
•
Total quality management or TQM requires that all business activities
should be oriented toward
- meeting and exceeding customer needs
- empowering employees
- eliminating waste or rework
- ensuring the long-run viability of the enterprise
through continuous quality improvement
TQM includes elements such as statistics, benchmarking, process
redesign, team building, group communications, quality function
deployment, and cross-functional management.
17-9
Chapter 17
LO17-4
17.3 Quality Improvement
Business Process Redesign (BPR)
•
Business process redesign or BPR seeks radical redesign of processes
to achieve breakthrough improvement in performance measures.
Statistical Quality Control (SQC)
•
Statistical quality control or SQC refers to a subset of quality
improvement techniques that rely on statistics. Listed below are some:
17-10
Chapter 17
LO17-5
17.3 Quality Improvement
LO17-5: List steps toward continuous quality improvement and
variance reduction.
Continuous Quality Improvement (CQI)
•
•
•
•
Quality improvement begins with measurement of a variable or an attribute.
For a variable, quality improvement means reducing variation from the target
specification.
For an attribute, quality improvement means decreasing the rate of
nonconformance.
Goal is to use statistical methods to eliminate sources of special cause
(nonrandom) variation.
17-11
Chapter 17
LO17-5
17.3 Quality Improvement
Continuous Quality Improvement (CQI)
•
Six Sigma steps to quality improvement – DMAIC (define, measure, analyze, improve,
control):
17-12
Chapter 17
LO17-6 17.5 Control Charts for a Mean
LO17-6: Make and interpret common control charts (x-bar, R, and p).
Control Limits: Known  and 
•
The process mean  is the centerline of the control chart.
•
The upper control limit (UCL) and lower control limit (LCL) are set at + 3 standard
errors from the centerline.
The Empirical rule says that 99.73% of the sample means will fall within “3-sigma”
limits.
•
17-13
Chapter 17
LO17-6 17.5 Control Charts for a Mean
Control Limits: Known  and 
•
•
Sample means will vary but should stay within the control limits and be
symmetrically distributed on either side of the centerline.
If a sample mean falls outside of these limits, then we suspect that the sample
may be from a different population from the one specified.
Empirical Control Limits: Un-known  and 
When  and  are unknown, we can estimate them with the sample
mean and sample standard deviation. The control limits will then be:
17-14
Chapter 17
LO17-6 17.5 Control Charts for a Mean
Empirical Control Limits
•
If using R, where we estimate the standard deviation with the range for the sample
subgroups, the formulas become
We use this table
to obtain d2.
17-15
Chapter 17
LO17-6 17.5 Control Charts for a Mean
Detecting Abnormal Patterns: Four Rules
• Rule 1. Single
point outside 3
sigma.
• Rule 2. Two of
three successive
points outside 2
sigma on same
side of centerline.
• Rule 3. Four of
five successive
points outside 1
sigma on same
side of centerline.
• Rule 4. Nine
successive points
on same side of
centerline.
17-16
Chapter 17
LO17-6
17.6 Control Charts for a Range
Control Limits for the Range
•
•
•
The x-bar chart of sample means only reveals information about the centrality of
the process.
To determine if the process is in control, we need to also examine the variation
around the mean – traditionally, the sample range – in the R chart.
The R chart has asymmetric control limits since the sample range is not a normally
distributed statistic.
17-17
Chapter 17
17.7 Other Control Charts
Attribute Data: p Charts
•
•
•
The p chart for attribute data plots the proportion of nonconforming items using the
sample proportion p (the defect rate).
The number of nonconforming items X in a sample of n items is a binomial random
variable.
The population nonconformance rate  can be found by using one of the following:
an assumed value of  (e.g., a target rate of nonconforming)
- an empirical estimate of  based on a large number of trials
- an estimate  from the samples being tested (if no other choice)
17-18
Chapter 17
17.7 Other Control Charts
Other Standard Control Charts
•
Other common types of control charts include:
- I charts (for individual numerical observations)
- MR charts (moving range for individual
obs.)
- s charts (for standard deviations)
- c charts (for Poisson events)
- np charts (for binomial totals)
- zone charts (using 6 regions based on s)
17-19
Chapter 17
LO17-7
17.8 Patterns in Control Charts
LO17-7: Recognize abnormal patterns in control charts and their
potential causes.
Abnormal Patterns
17-20
Chapter 17
17.9 Process Capability
LO17-8
LO17-8: Assess the capability of a process.
•
•
•
Customer requirements must be translated into and upper specification limit (USL)
and lower specification limit (LSL) of a quality metric.
These limits do not depend on the process.
Whether the process is capable of meeting these requirements depends on
- the magnitude of the process variation () and
- whether the process is correctly centered ()
Cp Index
•
•
•
Cpk Index
Managers typically require Cp > 1.33
Although easy to understand, Cp fails to
show whether the process is well-centered.
Cpk remedies this weakness by considering
the relationship between USL and LSL and
the process centerline.
17-21
Chapter 17
17.10 Additional Quality Topics
•
•
•
•
•
•
•
Acceptance Sampling
Supply-Chain Management
Quality and Design
Taguchi’s Robust Design
Six Sigma and Lean Six Sigma
ISO 9000
Malcolm Baldrige Award
17-22