COMM 225 — Production and Operations Management
CHAPTER 10: Statistical Quality Control
Comprehensive Exam Preparation Notes
1. Introduction to Statistical Quality Control
Statistical quality control (SQC) refers to the use of statistical techniques and sampling in
monitoring and testing of quality of goods and services. It provides an economical way to
evaluate the quality of products and meet the expectations of customers. There is a spectrum
of quality philosophy in organizations — from the least progressive (relying heavily on
inspection) to the most progressive (quality built so deeply into the process that inspection and
process control are unnecessary). Most companies occupy a middle ground that involves
some inspection and a significant degree of process control.
KEY Statistical Quality Control (SQC): The use of statistical techniques and
TERM sampling to monitor and test the quality of goods and services.
KEY Inspection: An appraisal activity that compares the quality of a good or service to
TERM a standard. Cannot be completely eliminated but should be minimized through
process design.
Two major components of SQC:
• Acceptance Sampling: Inspection of previously produced items (before or after
production). Covered in the chapter supplement.
• Statistical Process Control (SPC): Monitoring quality during production to detect and
correct assignable causes of variation before defects are generated.
Three points where inspection can occur:
• Before production: Inspect raw materials and purchased parts. Avoid wasting
resources processing bad inputs.
• During production: Monitor the transformation process in real time — this is the
domain of SPC.
• After production: Final verification of conformance before passing goods to customers.
Statistical Process Control Planning Process
Effective SPC requires a structured planning process. The overall planning framework consists
of two major steps:
• Step 1: Define the important quality characteristics, and how each is measured.
• Step 2: For each characteristic:
◦ (a) Determine a quality control point.
, ◦ (b) Plan how inspection is to be done, how much to inspect, and whether centralized
or on-site.
◦ (c) Plan the corrective action.
Define the Quality Characteristics
The first step is to define, in sufficient detail, what is to be controlled. Only characteristics that
can be measured are candidates for control. A standard must exist against which the
measurement can be evaluated. Different characteristics may require different control
approaches.
Determine a Quality Control Point
Every inspection adds cost. Inspection should be restricted to points where it does the most
good. In manufacturing, the three most important inspection points are:
• At the beginning of the process (raw materials and purchased parts): Avoid investing
labour and resources in material that is already bad.
• At the end of the process (finished products): Customer satisfaction and company
image are at stake. Field repairs are far more costly than production-point corrections.
• At operations where a customer-relevant characteristic is first determined:
Especially important to inspect before costly, irreversible, or covering (e.g., painting)
operations.
In the service sector, inspection points include wherever personnel and customers interact
(e.g., service counter) and the parts of the facility customers see (e.g., dining area, lobby).
How Much to Inspect
The amount of inspection ranges from zero to 100%. The optimal amount minimizes the total
cost of (1) inspection and (2) passing defectives. As inspection increases, inspection cost
rises but the cost of passing defectives falls. The goal is to find the point where their sum is
minimized. Key considerations:
• Low-cost, high-reliability processes: Minimal inspection is sufficient (e.g., nails, paper
clips).
• High-stakes, safety-critical items: Intensive inspection required (e.g., vehicle
components).
• Human-intensive operations require more inspection than mechanical ones (higher
variability).
• Unstable processes or processes with recent trouble require more frequent inspection.
• Many small lots require more samples than a few large lots.
• For high-volume repetitive operations, automated (100%) inspection at critical points
may be cost-effective.
Centralized Versus On-Site Inspection
On-site inspection is appropriate when the test is simple enough (e.g., measuring dimensions
with calipers or micrometers) or when the item cannot be moved (e.g., hull of a ship).
Automated systems (vision systems, laser gauges) can be installed directly on machinery to
achieve 100% on-site inspection. Centralized (lab) inspection is needed for specialized tests:
medical tests, food/water sample analysis, hardness testing, viscosity measurements,
destructive tests, or dimensions requiring large coordinate measuring machines.
, Plan the Corrective Action
When a process is judged out of control for an important characteristic, the process must be
stopped and corrective action taken. This involves: (1) uncovering the cause (e.g., worn tool
bits, failure to follow procedures); (2) testing and reworking or scrapping any defective output;
and (3) monitoring the process output for a sufficient period after correction to confirm the
problem has been eliminated.
2. Statistical Process Control
Statistical process control (SPC) is the primary tool used during production to statistically
evaluate process output for quality. It uses samples drawn from the process to determine
whether the process is in control — producing output consistent with historical, random
variation — or out of control, exhibiting assignable variation that must be investigated and
corrected. The ultimate goal of SPC is to assure that future process output will be random
(i.e., exhibiting only natural variation).
Types of Variations
All processes generate some variability in their output. The central question in SPC is whether
that variability is random (inherent) or assignable (correctable). Two questions must be
addressed:
• Question 1: Are the variations random? If nonrandom (assignable) variations are
present, the process is unstable. Corrective action is needed.
• Question 2: Given a stable process, is its inherent variability within design
specifications? This is process capability analysis.
Type of Variation Also Called Description Can It Be
(Deming) Corrected?
Random (Chance) Common Inherent, natural variation No — only by
Variation Cause due to countless minor changing the
Variation factors, each individually process itself
negligible. Present in every
process.
Assignable Special Cause Variation from identifiable, Yes — identify the
(Nonrandom) Variation specific causes: worn tools, cause and eliminate
Variation equipment needing it
adjustment, defective
materials, operator error.
KEY Statistical Process Control (SPC): The use of statistical methods (primarily
TERM control charts) to distinguish random variation from assignable variation in a
process, enabling operators to maintain a stable process.
CHAPTER 10: Statistical Quality Control
Comprehensive Exam Preparation Notes
1. Introduction to Statistical Quality Control
Statistical quality control (SQC) refers to the use of statistical techniques and sampling in
monitoring and testing of quality of goods and services. It provides an economical way to
evaluate the quality of products and meet the expectations of customers. There is a spectrum
of quality philosophy in organizations — from the least progressive (relying heavily on
inspection) to the most progressive (quality built so deeply into the process that inspection and
process control are unnecessary). Most companies occupy a middle ground that involves
some inspection and a significant degree of process control.
KEY Statistical Quality Control (SQC): The use of statistical techniques and
TERM sampling to monitor and test the quality of goods and services.
KEY Inspection: An appraisal activity that compares the quality of a good or service to
TERM a standard. Cannot be completely eliminated but should be minimized through
process design.
Two major components of SQC:
• Acceptance Sampling: Inspection of previously produced items (before or after
production). Covered in the chapter supplement.
• Statistical Process Control (SPC): Monitoring quality during production to detect and
correct assignable causes of variation before defects are generated.
Three points where inspection can occur:
• Before production: Inspect raw materials and purchased parts. Avoid wasting
resources processing bad inputs.
• During production: Monitor the transformation process in real time — this is the
domain of SPC.
• After production: Final verification of conformance before passing goods to customers.
Statistical Process Control Planning Process
Effective SPC requires a structured planning process. The overall planning framework consists
of two major steps:
• Step 1: Define the important quality characteristics, and how each is measured.
• Step 2: For each characteristic:
◦ (a) Determine a quality control point.
, ◦ (b) Plan how inspection is to be done, how much to inspect, and whether centralized
or on-site.
◦ (c) Plan the corrective action.
Define the Quality Characteristics
The first step is to define, in sufficient detail, what is to be controlled. Only characteristics that
can be measured are candidates for control. A standard must exist against which the
measurement can be evaluated. Different characteristics may require different control
approaches.
Determine a Quality Control Point
Every inspection adds cost. Inspection should be restricted to points where it does the most
good. In manufacturing, the three most important inspection points are:
• At the beginning of the process (raw materials and purchased parts): Avoid investing
labour and resources in material that is already bad.
• At the end of the process (finished products): Customer satisfaction and company
image are at stake. Field repairs are far more costly than production-point corrections.
• At operations where a customer-relevant characteristic is first determined:
Especially important to inspect before costly, irreversible, or covering (e.g., painting)
operations.
In the service sector, inspection points include wherever personnel and customers interact
(e.g., service counter) and the parts of the facility customers see (e.g., dining area, lobby).
How Much to Inspect
The amount of inspection ranges from zero to 100%. The optimal amount minimizes the total
cost of (1) inspection and (2) passing defectives. As inspection increases, inspection cost
rises but the cost of passing defectives falls. The goal is to find the point where their sum is
minimized. Key considerations:
• Low-cost, high-reliability processes: Minimal inspection is sufficient (e.g., nails, paper
clips).
• High-stakes, safety-critical items: Intensive inspection required (e.g., vehicle
components).
• Human-intensive operations require more inspection than mechanical ones (higher
variability).
• Unstable processes or processes with recent trouble require more frequent inspection.
• Many small lots require more samples than a few large lots.
• For high-volume repetitive operations, automated (100%) inspection at critical points
may be cost-effective.
Centralized Versus On-Site Inspection
On-site inspection is appropriate when the test is simple enough (e.g., measuring dimensions
with calipers or micrometers) or when the item cannot be moved (e.g., hull of a ship).
Automated systems (vision systems, laser gauges) can be installed directly on machinery to
achieve 100% on-site inspection. Centralized (lab) inspection is needed for specialized tests:
medical tests, food/water sample analysis, hardness testing, viscosity measurements,
destructive tests, or dimensions requiring large coordinate measuring machines.
, Plan the Corrective Action
When a process is judged out of control for an important characteristic, the process must be
stopped and corrective action taken. This involves: (1) uncovering the cause (e.g., worn tool
bits, failure to follow procedures); (2) testing and reworking or scrapping any defective output;
and (3) monitoring the process output for a sufficient period after correction to confirm the
problem has been eliminated.
2. Statistical Process Control
Statistical process control (SPC) is the primary tool used during production to statistically
evaluate process output for quality. It uses samples drawn from the process to determine
whether the process is in control — producing output consistent with historical, random
variation — or out of control, exhibiting assignable variation that must be investigated and
corrected. The ultimate goal of SPC is to assure that future process output will be random
(i.e., exhibiting only natural variation).
Types of Variations
All processes generate some variability in their output. The central question in SPC is whether
that variability is random (inherent) or assignable (correctable). Two questions must be
addressed:
• Question 1: Are the variations random? If nonrandom (assignable) variations are
present, the process is unstable. Corrective action is needed.
• Question 2: Given a stable process, is its inherent variability within design
specifications? This is process capability analysis.
Type of Variation Also Called Description Can It Be
(Deming) Corrected?
Random (Chance) Common Inherent, natural variation No — only by
Variation Cause due to countless minor changing the
Variation factors, each individually process itself
negligible. Present in every
process.
Assignable Special Cause Variation from identifiable, Yes — identify the
(Nonrandom) Variation specific causes: worn tools, cause and eliminate
Variation equipment needing it
adjustment, defective
materials, operator error.
KEY Statistical Process Control (SPC): The use of statistical methods (primarily
TERM control charts) to distinguish random variation from assignable variation in a
process, enabling operators to maintain a stable process.
