MAN5501 — Chapter 3 Study Guide

Quality and Performance

Exam 1 Prep | Summer 2026

📖 Review Guide Notes:

 ✅ Know how to calculate Capability Index (Cpk) and Capability

Ratio (Cp)

 ✅ Know how to create and interpret x̄-chart, R-chart, and p-chart

 ⚠️You will NOT be asked to create a c-chart

📋 Table of Contents

1. What is Quality?

2. Costs of Quality

3. Total Quality Management (TQM)

4. Six Sigma

5. Process Capability

6. 🔢 Capability Index (Cpk) & Capability Ratio (Cp)

7. Statistical Process Control (SPC)

8. Control Charts — Overview

9. 🔢 R-Chart (Range Chart)

10. 🔢 x̄-Chart (Mean Chart)

11. 🔢 p-Chart (Proportion Defective Chart)

12. Interpreting Control Charts

13. Key Terms Glossary

1. What is Quality?

Quality — The general level of satisfaction with a service or product (from

the customer's perspective).
To achieve quality, three things must be true:

1. The product must be designed to meet customer needs

2. The process must be capable of creating the product as designed

3. The process must be controlled to ensure it is working properly

💡 Modern quality management focuses primarily on #3 — controlling

processes — because it's ongoing and prevents defects before they happen.

Defect — Any instance when a process or product fails to satisfy its

customer (internal or external).

2. Costs of Quality

Costs of ACHIEVING Quality

Type Description Examples

Prevention Costs to prevent defects before Training, process design,

Costs they happen quality planning

Appraisal Costs to monitor and assess Inspection, testing, control

Costs process performance charts

Costs of NOT ACHIEVING Quality

Type Description Examples

Internal Defects found before the

Scrap, rework
Failure Costs product leaves the facility

Defects found after the

External Returns, warranty costs, lost
customer receives the
Failure Costs sales, lawsuits
product

Ethical Costs from deceptively Damage to stockholder value,

Failure Costs passing defective products customer trust, legal liability

⚠️The costs of NOT achieving quality are always higher than the costs of
achieving it. Spending money on prevention and appraisal saves far more in
failure costs — but the savings come later (delayed gratification).

3. Total Quality Management (TQM)

TQM — A philosophy that stresses three principles for achieving high levels
of process performance and quality:

The Three Pillars of TQM

Pillar Key Ideas

Customer Conformance to specs, value, fitness for use, support,

Satisfaction psychological impressions

Quality at the source; workers have the knowledge to fix

Employee
problems; problem-solving teams, quality circles, self-
Involvement
managed teams

Continuous Kaizen — always trying to improve; use of the Deming

Improvement Wheel (Plan-Do-Check-Act)

The Deming Wheel (PDCA Cycle)

1. Plan — Identify a problem and make a plan to fix it

2. Do — Test the plan (pilot program)

3. Check — Is the plan working?

4. Act — Implement, document, and move on to the next improvement

💡 The wheel shape is intentional — once you finish one improvement cycle,
you immediately start the next. It's a never-ending loop of continuous
improvement. Similar to Six Sigma's DMAIC.

W. Edwards Deming — American statistician who developed TQM

concepts; rejected by U.S. automakers, embraced by Japan → Japan's quality
revolution followed.

4. Six Sigma

Six Sigma — A comprehensive system for achieving business success by

minimizing defects and variability in processes.

The Core Idea

 All processes have variability (measured by standard deviation σ)

 The goal is to reduce σ until the process distribution is so tight that

the specifications are 6 standard deviations away from the mean
  At 6σ, the probability of a defect is extremely small (~3.4 defects per
million opportunities)

Why "Six Sigma"?

 Traditional target: 3σ → Cp = 1.0 (specs are 3 standard deviations from

the mean)

 Better target: 4σ → Cp = 1.33

 Even better: 5σ → Cp = 1.67

 Six Sigma target: 6σ → Cp = 2.0

💡 Six Sigma was developed by Motorola when they were struggling. Instead
of incremental improvement, they made radical changes — and it worked.

5. Process Capability — Concepts

Process Capability — The ability of a process to meet the design

specifications for a product or service.

Key Terms

Term Definition

The target (bullseye) for a design

Nominal Value
specification

Allowable deviation above or below the

Tolerance
nominal value

Upper Specification Limit

Nominal value + tolerance
(USL)

Lower Specification Limit

Nominal value − tolerance
(LSL)

Measures how wide/spread the process

Standard Deviation (σ)
distribution is

Capable vs. Not Capable

Capable Process Not Capable Process

Distribution Narrow, centered on nominal Wide and/or off-center

 Capable Process Not Capable Process

value

Tails outside Very small (low defect Noticeable (high defect

specs probability) probability)

Reduce σ and/or re-center the

Action needed Continue monitoring
mean

💡 Two ways to make a process more capable: (1) reduce the standard
deviation (tighten the distribution), or (2) re-center the mean closer to the
nominal value. Widening the specs ("moving the goalposts") is cheating.

6. 🔢 Capability Index (Cpk) & Capability Ratio (Cp)

Capability Index — Cpk

Measures how centered the process is relative to the specifications. Takes

the minimum of the two sides.

$$C_{pk} = \min\left(\frac{\bar{\bar{x}} - LSL}{3\sigma},\ \frac{USL - \

bar{\bar{x}}}{3\sigma}\right)$$

 x̄̄ = process mean (average of sample means)

 σ = standard deviation of the process

 Take the minimum of the two values — this represents the worst-case
side

Capability Ratio — Cp

Measures the overall spread of the process relative to the specification width.
Does not account for centering.

$$C_p = \frac{USL - LSL}{6\sigma}$$

Sigma Level Benchmarks

Sigma Required Required

Level Cp Cpk

3σ
1.00 1.00
(traditional)

4σ 1.33 1.33
Sigma Required Required
Level Cp Cpk

5σ 1.67 1.67

6σ (Six
2.00 2.00
Sigma)

⚠️Cp measures potential capability (is the spread narrow enough?). Cpk
measures actual capability (is the process also centered?). A process can
have a good Cp but a poor Cpk if it's off-center.

✏️Worked Example — ICU Lab Turnaround Time

Given: Nominal = 25 min, Tolerance = ±5 min → USL = 30, LSL = 20 Data:

x̄̄ = 26.2 min, σ = 1.35 min Question: Is the process capable at the 4σ level
(Cpk ≥ 1.33)?

Step 1 — Calculate Cpk: $$C_{pk} = \min\left(\frac{26.2 - 20}{3(1.35)},\ \

frac{30 - 26.2}{3(1.35)}\right) = \min\left(\frac{6.2}{4.05},\ \frac{3.8}
{4.05}\right) = \min(1.53,\ 0.94) = \textbf{0.94}$$

Step 2 — Calculate Cp: $$C_p = \frac{30 - 20}{6(1.35)} = \frac{10}{8.1}

= \textbf{1.23}$$

Conclusion: Cpk = 0.94 < 1.33 → Process is NOT capable at 4σ level

 The process mean (26.2) is too close to the USL (30)

 Need to either re-center the mean closer to 25, or reduce σ further

7. Statistical Process Control (SPC)

SPC — The application of statistical techniques to determine whether a

process is delivering what the customer wants.

Two Approaches to Quality Monitoring

Method What It Monitors When

Acceptance Quality of product received After the fact (shipment

Sampling from a supplier already made)
Method What It Monitors When

Statistical Our own internal production Ongoing — prevents

Process Control process defects in real time

💡 Modern quality management focuses on SPC — doing it right the first time
— rather than acceptance sampling (catching problems after the fact).

Complete Inspection vs. Sampling

 Complete inspection = inspect every item → accurate but expensive

and slow

 Sampling = inspect a subset → cheaper and faster; basis of SPC

In Control vs. Out of Control

Condition Cause Action

Only common causes (random,

In Control Allow process to continue
unavoidable variation)

Out of Assignable causes present Stop the process, find and

Control (identifiable, fixable) eliminate the cause

Common Causes vs. Assignable Causes

Type Definition Examples

Random, inherent variability —

Common Normal machine vibration,
unavoidable with the current
Cause minor material variation
process

Machine wear, broken

Assignable Identifiable, non-random
component, worker not
Cause variation that can be eliminated
following procedure

8. Control Charts — Overview

Control Chart — A time-ordered diagram used to determine whether

observed variation signals an in-control or out-of-control condition.

Three Lines on Every Control Chart

 Center Line (CL) — The nominal/average value

  Upper Control Limit (UCL) — Upper boundary (typically 3σ above
center)

 Lower Control Limit (LCL) — Lower boundary (typically 3σ below

center)

⚠️UCL and LCL are NOT the same as USL and LSL (specification limits).
Control limits are based on the sampling distribution; specification limits are
based on customer requirements.

Types of Control Charts

Chart Measures Used For

R- Variability (range) of
Variables — used first to check variability
Chart the process

x̄- Variables — used after R-chart confirms

Process mean
Chart variability is in control

p- Proportion of defective
Attributes (pass/fail, defective/acceptable)
Chart items

c- Number of defects per

Not required for this exam
Chart item

Type I and Type II Errors

Erro
What Happens Consequence
r

Conclude process is OUT of control when it's

Type Unnecessary stoppage;
actually fine (random variation triggered the
I wasted time
alarm)

Type Conclude process is IN control when it's Defective products

II actually out of control continue to be made

💡 With 3σ control limits, the probability of a Type I error is 0.0026 (very small
but not zero).

9. 🔢 R-Chart (Range Chart)

Purpose: Monitor the variability of the process. Always construct the R-

chart first — if variability is out of control, the x̄-chart is meaningless.
Formulas

$$UCL_R = D_4 \bar{R}$$ $$LCL_R = D_3 \bar{R}$$ $$\text{Center Line}

= \bar{R} \text{ (average of all sample ranges)}$$

 R̄ = average of all sample ranges

 D₃, D₄ = constants from Table 3.1 (based on sample size n)

 Range (R) = largest value in sample − smallest value in sample

Table 3.1 — Control Chart Constants

Sample Size
A₂ D₃ D₄
(n)

1.88 3.26
2 0
0 7

1.02 2.57
3 0
3 5

0.72 2.28
4 0
9 2

0.57 2.11
5 0
7 5

0.48 2.00
6 0
3 4

0.41 0.07 1.92

7
9 6 4

0.37 0.13 1.86

8
3 6 4

0.33 0.18 1.81

9
7 4 6

0.30 0.22 1.77

10
8 3 7

⚠️The sample size (n) = number of items measured within each sample,
NOT the number of samples taken. This is a common mistake!
✏️Worked Example — Webster Chemical (Caulking Tubes)

Given: 5 samples of n = 8 tubes each; target = 8 oz

Avg Range
Sample
(x̄) (R)

1 8.040 0.76

2 8.160 0.43

3 7.940 0.32

4 8.050 0.41

5 7.980 0.33

Averag x̄̄ = R̄ =
es 8.034 0.45

From Table 3.1 (n = 8): D₄ = 1.864, D₃ = 0.136, A₂ = 0.373

R-Chart: $$UCL_R = 1.864 \times 0.45 = \textbf{0.839}$$ $$LCL_R = 0.136

\times 0.45 = \textbf{0.061}$$

All sample ranges (0.76, 0.43, 0.32, 0.41, 0.33) fall between 0.061 and 0.839
→ Variability is IN CONTROL ✅ → Proceed to x̄-chart.

10. 🔢 x̄-Chart (Mean Chart)

Purpose: Monitor whether the process mean is staying on target. Only

construct after confirming the R-chart is in control.

Method 1 — When σ (process standard deviation) IS known

$$UCL_{\bar{x}} = \bar{\bar{x}} + z\sigma_{\bar{x}} \quad \

text{where} \quad \sigma_{\bar{x}} = \frac{\sigma}{\sqrt{n}}$$ $$LCL_{\
bar{x}} = \bar{\bar{x}} - z\sigma_{\bar{x}}$$

 z = 3 for 99.74% confidence (standard); z = 1.96 for 95% confidence

(5% Type I error)

Method 2 — When σ is NOT known (use this method — it's on the

exam)

$$UCL_{\bar{x}} = \bar{\bar{x}} + A_2\bar{R}$$ $$LCL_{\bar{x}} = \

bar{\bar{x}} - A_2\bar{R}$$
  x̄̄ = grand mean (average of all sample means)

 R̄ = average range (from R-chart)

 A₂ = constant from Table 3.1 (based on sample size n)

✏️Worked Example — Webster Chemical (continued)

x̄-Chart (Method 2, n = 8, A₂ = 0.373): $$UCL_{\bar{x}} = 8.034 +

0.373(0.45) = 8.034 + 0.168 = \textbf{8.202}$$ $$LCL_{\bar{x}} = 8.034 -
0.373(0.45) = 8.034 - 0.168 = \textbf{7.866}$$

All sample means (8.040, 8.160, 7.940, 8.050, 7.980) fall between 7.866 and
8.202 → Process mean is IN CONTROL ✅

✏️Worked Example — Sunny Dale Bank (Method 1, σ known)

Given: Mean = 5 min, σ = 1.5 min, n = 6, Type I error = 5% → z = 1.96

$$\sigma_{\bar{x}} = \frac{1.5}{\sqrt{6}} = 0.612$$ $$UCL_{\bar{x}} =

5 + 1.96(0.612) = \textbf{6.20 min}$$ $$LCL_{\bar{x}} = 5 - 1.96(0.612) =
\textbf{3.80 min}$$

New samples: 3.70 and 3.68 min — both below LCL → technically out of
control, but in a good way (process improved). Investigate what changed
and institutionalize the improvement.

💡 Why z = 1.96 for 5% Type I error? A two-sided test splits 5% into 2.5% on
each tail. Look up 0.975 in the z-table → z = 1.96.

11. 🔢 p-Chart (Proportion Defective Chart)

Purpose: Monitor the proportion of defective items in a sample. Used for

attributes (pass/fail, defective/acceptable).

Formulas

$$\bar{p} = \frac{\text{Total number of defectives}}{\text{Total number of

items inspected}}$$

$$\sigma_p = \sqrt{\frac{\bar{p}(1-\bar{p})}{n}}$$

$$UCL_p = \bar{p} + z\sigma_p$$ $$LCL_p = \bar{p} - z\sigma_p$$

  If LCL < 0, set LCL = 0 (proportion can't be negative)

 z = 3 is standard (3-sigma control limits)

✏️Worked Example — Webster Chemical (Leaking Caps)

Given: 20 samples of n = 144 tubes each; 72 total leaking tubes found

Step 1 — Calculate p̄ : $$\bar{p} = \frac{72}{20 \times 144} = \frac{72}

{2,880} = \textbf{0.025}$$

Step 2 — Calculate σp: $$\sigma_p = \sqrt{\frac{0.025(1-0.025)}{144}}

= \sqrt{\frac{0.025 \times 0.975}{144}} = \sqrt{0.0001693} = \
textbf{0.01301}$$

Step 3 — Calculate control limits (z = 3): $$UCL_p = 0.025 + 3(0.01301)

= 0.025 + 0.039 = \textbf{0.064}$$ $$LCL_p = 0.025 - 3(0.01301) = 0.025 -
0.039 = -0.014 \rightarrow \textbf{0 (set to zero)}$$

Conclusion: All 20 sample proportions fall between 0 and 0.064 → Process

is IN CONTROL ✅

12. Interpreting Control Charts

A Process is IN CONTROL if:

 All sample points fall between the UCL and LCL

 No unusual patterns are present

A Process is OUT OF CONTROL if ANY of the following occur:

Signal Description Action

Point
Stop — find
outside Any point falls above UCL or below LCL
assignable cause
limits

8+ consecutive points all above or all Stop — process has

Run
below the center line shifted

Consistent upward or downward Investigate —

Trend
movement across several points process is drifting

Sudden Abrupt shift in the level of readings Monitor closely

Signal Description Action

jump

💡 When a process goes out of control in a good direction (e.g., mean drops
below LCL for a service time), investigate what improved and institutionalize
that change.

Steps for Using a Control Chart

1. Take a random sample; calculate the performance measure (range,

mean, or proportion)

2. Plot on the control chart

3. If outside limits or unusual pattern → look for assignable cause

4. Eliminate the cause (or incorporate it if it's an improvement)

5. Reconstruct the chart with new data if needed

6. Repeat continuously

13. Key Terms Glossary

Term Definition

General level of customer satisfaction with a product or

Quality
service

Defect Any instance where a process fails to satisfy its customer

Prevention
Costs to prevent defects before they happen
Costs

Appraisal
Costs to monitor and assess process performance
Costs

Internal Costs from defects found before the product leaves the
Failure Costs facility (scrap, rework)

External Costs from defects found after the customer receives the
Failure Costs product

TQM Total Quality Management — philosophy of customer

satisfaction, employee involvement, continuous
Term Definition

improvement

Kaizen Japanese term for continuous, incremental improvement

Deming Wheel Plan-Do-Check-Act cycle for continuous improvement

Program to minimize defects by reducing process variability

Six Sigma
to 6σ from the mean

Nominal Value The target value for a design specification

Tolerance Allowable deviation above or below the nominal value

Upper/Lower Specification Limits — based on customer

USL / LSL
requirements

Process
Ability of a process to meet design specifications
Capability

Capability Index — measures centering; takes the minimum

Cpk
of the two sides

Capability Ratio — measures spread relative to spec width;

Cp
does not account for centering

Statistical Process Control — statistical techniques to

SPC
monitor process quality

Common
Random, unavoidable variation inherent in the process
Cause

Assignable
Identifiable, non-random variation that can be eliminated
Cause

Process variation due to common causes only — allow to

In Control
continue

Out of Control Assignable causes present — stop and fix

Upper/Lower Control Limits — based on sampling

UCL / LCL
distribution (≠ spec limits)

Control chart for process variability (range); always

R-Chart
construct first
Term Definition

Control chart for process mean; construct after R-chart

x̄-Chart
confirms variability is in control

p-Chart Control chart for proportion of defective items (attributes)

R̄ Average range across all samples

x̄̄ Grand mean — average of all sample means

p̄ Average proportion defective across all samples

Concluding out of control when actually in control (false

Type I Error
alarm)

Concluding in control when actually out of control (missed

Type II Error
signal)

International quality documentation standard; required by

ISO 9001
many supply chain partners

Malcolm
U.S. award for quality excellence; rigorous application
Baldrige
process helps companies improve
Award

📐 Quick Formula Reference

Formula Use

Cpk = min[(x̄̄ − LSL)/3σ, (USL Capability index — accounts for

− x̄̄)/3σ] centering

Cp = (USL − LSL) / 6σ Capability ratio — overall spread

UCL_R = D₄R̄ R-chart upper limit

LCL_R = D₃R̄ R-chart lower limit

UCL_x̄ = x̄̄ + A₂R̄ x̄-chart upper limit (σ unknown)

LCL_x̄ = x̄̄ − A₂R̄ x̄-chart lower limit (σ unknown)

UCL_x̄ = x̄̄ + zσ/√n x̄-chart upper limit (σ known)

LCL_x̄ = x̄̄ − zσ/√n x̄-chart lower limit (σ known)

Formula Use

p̄ = total defectives / total

Average proportion defective
inspected

σp = √[p̄(1−p̄)/n] Standard deviation of p-chart

UCL_p = p̄ + zσp p-chart upper limit

LCL_p = p̄ − zσp (min 0) p-chart lower limit

📌 Up next: Chapter 4 — Lean Systems & Kanban (no calculations required)

Good luck studying! 🎓