Study guide

12.3 Attribute Control Charts: p, np, c, and u Charts

CSSC Certified Six Sigma Green Belt Exam

Track your progress as you read

Finished sections get a check mark.

1 Chapter 1: Six Sigma Overview & Exam Fundamentals

1.1 CSSC Certification Overview, Exam Structure & Body of Knowledge 1.2 History & Evolution of Six Sigma 1.3 Six Sigma Organizational Roles & Belts Hierarchy 1.4 The DMAIC Methodology Architecture vs. DMADV / DFSS 1.5 Other Process Improvement & Quality Methods, and When to Choose Six Sigma

2 Chapter 2: Quality Fundamentals & Defect Metrics

2.1 Critical to Quality (CTQ) Characteristics & VOC Translation 2.2 Defect Metrics: DPU, DPO, DPMO, & the 1.5-Sigma Shift 2.3 First Time Yield (FTY) & Rolled Throughput Yield (RTY) 2.4 Cost of Quality (COQ) & Cost of Poor Quality (COPQ)

3 Chapter 3: Lean Principles & Waste Elimination

3.1 Foundations of Lean & The Toyota Production System (TPS) 3.2 The Eight Wastes: CSSC's Seven Muda, TIMWOODS, and Muri/Mura 3.3 The 5S Visual Workplace Methodology 3.4 Value Stream Mapping (VSM), Cycle Time & Takt Time

4 Chapter 4: Project Selection & Team Management

4.1 Six Sigma Project Selection & Business Case Development 4.2 Project Charter Architecture 4.3 Stakeholder Management & Communication Planning 4.4 Team Dynamics & Tuckman's Team Development Stages 4.5 The 15-Point Project Viability Model & Process-Level Selection

5 Chapter 5: Define Phase & Process Mapping

5.1 High-Level Process Mapping with SIPOC 5.2 Flowcharts, Swimlane Diagrams & Process Value Analysis 5.3 Quality Function Deployment (QFD) & The House of Quality 5.4 Critical to Quality (CTQ) Flowdown & Operational Definitions 5.5 The CSSC Define Toolset: Ground Rules, In/Out of the Box, Is–Is Not & Tollgate Checklists

6 Chapter 6: Measure Phase & Measurement Systems

6.1 Data Types: Continuous (Variable) vs. Discrete (Attribute) Data 6.2 Data Collection Planning & Sampling Strategies 6.3 Measurement System Analysis (MSA) Foundations & Calibration 6.4 Variable & Attribute Gage R&R Studies

7 Chapter 7: Analyze Phase & Root Cause Analysis

7.1 Qualitative Root Cause Tools: 5 Whys & Cause-and-Effect (Fishbone) Diagrams 7.2 Failure Mode and Effects Analysis (FMEA) & RPN Calculation 7.3 Pareto Analysis & The 80/20 Rule 7.4 Graphical Data Exploration: Histograms, Box Plots, & Multi-Vari Charts 7.5 The Root Cause Verification Matrix

8 Chapter 8: Descriptive Statistics & Distributions

8.1 Measures of Central Tendency & Dispersion 8.2 The Normal Distribution & The Empirical Rule (68-95-99.7) 8.3 Standard Normal Distribution, Z-Scores & Probability Calculations 8.4 Common Non-Normal Distributions: Binomial, Poisson & Chi-Square

9 Chapter 9: Inferential Statistics & Hypothesis Testing

9.1 Fundamentals of Hypothesis Testing: Null vs. Alternative, Alpha & Beta Risks 9.2 One-Sample and Two-Sample t-Tests for Means 9.3 Paired t-Tests & Non-Parametric Alternatives 9.4 Analysis of Variance (One-Way & Two-Way ANOVA) & Chi-Square Contingency Tests 9.5 Sample Size Determination for Hypothesis Tests

10 Chapter 10: Correlation, Regression, & Experimentation

10.1 Linear Correlation & Pearson's Correlation Coefficient (r) 10.2 Simple Linear Regression & Coefficient of Determination (R^2) 10.3 Introduction to Design of Experiments (DOE) & Factorial Designs 10.4 Analyzing Experimental Results: Main Effects & Interaction Plots

11 Chapter 11: Improve Phase & Solution Implementation

11.1 Brainstorming, Creative Thinking & Solution Selection Matrices 11.2 Piloting Solutions & Conducting Risk Assessments 11.3 Implementation Planning & Change Management (ADKAR) 11.4 Mistake-Proofing (Poka-Yoke) & Standard Work

12 Chapter 12: Control Phase, SPC, & Process Capability

12.1 Statistical Process Control (SPC) & Shewhart Control Rules 12.2 Variable Control Charts: I-MR, X-bar & R, X-bar & S 12.3 Attribute Control Charts: p, np, c, and u Charts 12.4 Process Capability & Performance Indices: Cp, Cpk, Pp, Ppk 12.5 Communicating Statistical Results to Business Audiences

12.3 Attribute Control Charts: p, np, c, and u Charts

Key Takeaways Explain This Topic AI

Attribute control charts track discrete qualitative data based on two distinct probability distributions: the Binomial distribution for defective units (pass/fail) and the Poisson distribution for defect counts (flaws per opportunity).

A defective (nonconforming unit) refers to an entire product or transaction that fails to meet specification, whereas a defect (nonconformity) refers to a specific individual flaw; a single defective unit can contain multiple defects.

The p-chart tracks the proportion of defective units and accommodates both constant and variable subgroup sizes (creating stepped control limits), while the np-chart tracks the integer count of defectives and STRICTLY requires a constant sample size.

The c-chart tracks total defect counts across a strictly constant area of opportunity (Poisson), while the u-chart tracks the average defect rate per unit and accommodates variable inspection areas or sample sizes.

Under the Poisson distribution, variance equals the mean (sigma^2 = c-bar), so the standard error for c-charts is simply the square root of c-bar without sample size denominators; for attribute charts, negative LCL calculations default to zero.

Last updated: September 2026

12.3 Attribute Control Charts: p, np, c, and u Charts

Quick Summary: Attribute control charts monitor discrete, qualitative quality characteristics such as pass/fail classifications, visual defects, clerical errors, or customer complaints. Selecting the correct attribute chart requires answering two sequential questions: (1) Data Classification: Are you evaluating defective units (Binomial distribution: Go/No-Go) or counting defects (Poisson distribution: flaws per unit)? (2) Subgroup Size: Is the sample size or area of opportunity constant or variable ? This yields the four standard charts: p p p , n p np n p , c c c , and u u u .

Defective Units vs. Defects: The Fundamental Distinction

The most critical conceptual hurdle in attribute SPC is distinguishing between a defective unit and a defect . Attribute Data Type

  • │
  • ┌────────────────────────────┴────────────────────────────┐
  • ▼ ▼
  • DEFECTIVE DEFECT
  • (Nonconforming Unit) (Nonconformity)
  • Whole unit is judged PASS or FAIL • Specific flaw, blemish, or error
  • Binary outcome: Yes or No • Countable quantity: 0, 1, 2, 3...
  • Governed by BINOMIAL Distribution • Governed by POISSON Distribution
  • Monitored by: p-Chart or np-Chart • Monitored by: c-Chart or u-Chart
  • 1. Defective (Nonconforming Unit) — Binomial Distribution

A unit is declared defective when it fails to meet one or more quality specifications, rendering the entire item unacceptable. It is a binary, qualitative assessment: an invoice is either correct or incorrect; a lightbulb either illuminates or does not; a medical claim is either paid or rejected. Regardless of whether an item has 1 flaw or 10 flaws, it counts as exactly one defective unit .

2. Defect (Nonconformity) — Poisson Distribution

A defect is a single, specific failure, blemish, or departure from specification found on a unit. A unit can contain multiple defects and still remain functional. For example, a newly painted luxury car hood might contain 3 paint blisters, 1 scratch, and 2 dust nibs (a total of 6 defects), yet it remains a single unit. Defect counts represent rare events occurring across a continuous area of opportunity, modeled by the Poisson distribution .

Inspection Context Defective Unit (Binomial - p , n p p, np p , n p ) Defect / Nonconformity (Poisson - c , u c, u c , u )

Automobile Manufacturing A car that fails final inspection due to water leaks. Total number of water leaks, rattles, paint blemishes, and loose bolts across the vehicle.

Healthcare Records A patient admission file rejected due to missing data. Total number of clerical errors, misspelled names, and missing signatures across 50 patient files.

Textile Production A 100-yard bolt of fabric classified as scrap/reject. Total number of snags, loose threads, and dye blotches found per 100 square yards of fabric.

Printed Circuit Boards A motherboard that fails electrical continuity test. Total number of cold solder joints, bridge defects, and misaligned capacitors on the board.

The Attribute Control Chart Selection Decision Tree

To determine which of the four attribute charts to deploy, use the following two-axis selection matrix: Attribute Control Chart Selection

│

What type of attribute data is being collected?

  • │
  • ┌───────────────────────────────┴───────────────────────────────┐
  • ▼ ▼
  • DEFECTIVE UNITS DEFECTS
  • (Binomial: Pass/Fail) (Poisson: Count of Flaws)
  • │ │

Is sample size (n) constant? Is inspection unit constant?

  • │ │
  • ┌────┴────┐ ┌────┴────┐
  • ▼ ▼ ▼ ▼
  • YES NO YES NO
  • np-Chart p-Chart c-Chart u-Chart
  • (Number) (Fraction) (Count) (Rate/Unit)

*Note: p-charts can also be used for constant n. *Note: u-charts can also be used for constant n.

Chart Metric Tracked Statistical Distribution Sample Size / Area of Opportunity Mathematical Center Line (CL)

p p p Proportion / Fraction Defective Binomial Variable or Constant p ˉ = Total Defectives Total Units Inspected \bar{p} = \frac{\text{Total Defectives}}{\text{Total Units Inspected}} p ˉ ​ = Total Units Inspected Total Defectives ​

n p np n p Number of Defective Units Binomial Strictly Constant ( n n n ) n p ˉ = Total Defectives k n\bar{p} = \frac{\text{Total Defectives}}{k} n p ˉ ​ = k Total Defectives ​

c c c Total Count of Defects Poisson Strictly Constant Area ( n = 1 n=1 n = 1 ) c ˉ = Total Defects k \bar{c} = \frac{\text{Total Defects}}{k} c ˉ = k Total Defects ​

u u u Defects per Inspection Unit Poisson Variable or Constant Area u ˉ = Total Defects Total Inspection Units \bar{u} = \frac{\text{Total Defects}}{\text{Total Inspection Units}} u ˉ = Total Inspection Units Total Defects ​

The p p p -Chart (Proportion Defective)

The p p p -chart monitors the fraction or percentage of defective units in a sample. It is the most versatile attribute chart because it can handle variable subgroup sizes ( n i n_i n i ​ ) .

Formulas for p p p -Chart:

Subgroup Proportion ( p i p_i p i ​ ): p i = d i n i p_i = \frac{d_i}{n_i} p i ​ = n i ​ d i ​ ​ (where d i d_i d i ​ is number of defectives found in subgroup i i i of size n i n_i n i ​ )

Center Line ( p ˉ \bar{p} p ˉ ​ ): p ˉ = ∑ i = 1 k d i ∑ i = 1 k n i \bar{p} = \frac{\sum_{i=1}^k d_i}{\sum_{i=1}^k n_i} p ˉ ​ = ∑ i = 1 k ​ n i ​ ∑ i = 1 k ​ d i ​ ​

Standard Error of Proportion ( σ p i \sigma_{p_i} σ p i ​ ​ ): σ p i = p ˉ ( 1 − p ˉ ) n i \sigma_{p_i} = \sqrt{\frac{\bar{p}(1 - \bar{p})}{n_i}} σ p i ​ ​ = n i ​ p ˉ ​ ( 1 − p ˉ ​ ) ​ ​

Control Limits: UCL p i = p ˉ + 3 p ˉ ( 1 − p ˉ ) n i \text{UCL}_{p_i} = \bar{p} + 3 \sqrt{\frac{\bar{p}(1 - \bar{p})}{n_i}} UCL p i ​ ​ = p ˉ ​ + 3 n i ​ p ˉ ​ ( 1 − p ˉ ​ ) ​ ​ LCL p i = max ⁡ ( 0 , p ˉ − 3 p ˉ ( 1 − p ˉ ) n i ) \text{LCL}_{p_i} = \max\left(0, \bar{p} - 3 \sqrt{\frac{\bar{p}(1 - \bar{p})}{n_i}}\right) LCL p i ​ ​ = max ​ 0 , p ˉ ​ − 3 n i ​ p ˉ ​ ( 1 − p ˉ ​ ) ​ ​ ​

Stepped (Variable) Control Limits

When sample sizes n i n_i n i ​ vary from day to day (e.g., daily call center volumes fluctuating between 800 and 1,500 calls), the denominator inside the square root changes for each subgroup. Consequently, the control limits expand on days with smaller sample sizes and contract on days with larger sample sizes, creating a distinctive "stair-step" appearance.

The n p np n p -Chart (Number of Defectives)

The n p np n p -chart tracks the raw integer count of defective units rather than a fraction. It is highly popular in factory environments because floor operators find integer counts (e.g., "6 bad parts") far easier to interpret than decimals (e.g., "0.012 fraction defective").

Strict Operational Requirement: The sample size n n n must be identical for every single subgroup . If n n n varies by even a single unit, an n p np n p -chart cannot be used; the team must switch to a p p p -chart.

Formulas for n p np n p -Chart:

Center Line ( C L n p CL_{np} C L n p ​ ): CL n p = n p ˉ = ∑ d i k \text{CL}_{np} = n\bar{p} = \frac{\sum d_i}{k} CL n p ​ = n p ˉ ​ = k ∑ d i ​ ​

Standard Error of Count ( σ n p \sigma_{np} σ n p ​ ): σ n p = n p ˉ ( 1 − p ˉ ) \sigma_{np} = \sqrt{n\bar{p}(1 - \bar{p})} σ n p ​ = n p ˉ ​ ( 1 − p ˉ ​ ) ​

Control Limits: UCL n p = n p ˉ + 3 n p ˉ ( 1 − p ˉ ) \text{UCL}_{np} = n\bar{p} + 3 \sqrt{n\bar{p}(1 - \bar{p})} UCL n p ​ = n p ˉ ​ + 3 n p ˉ ​ ( 1 − p ˉ ​ ) ​ LCL n p = max ⁡ ( 0 , n p ˉ − 3 n p ˉ ( 1 − p ˉ ) ) \text{LCL}_{np} = \max\left(0, n\bar{p} - 3 \sqrt{n\bar{p}(1 - \bar{p})}\right) LCL n p ​ = max ( 0 , n p ˉ ​ − 3 n p ˉ ​ ( 1 − p ˉ ​ ) ​ )

The c c c -Chart (Defect Counts for Constant Area)

The c c c -chart tracks the total count of nonconformities (defects) across a constant area of opportunity (such as one airplane wing, 100 printed pages, or 10 square meters of glass). Because it models Poisson events, variance equals the mean ( σ 2 = λ = c ˉ \sigma^2 = \lambda = \bar{c} σ 2 = λ = c ˉ ).

Formulas for c c c -Chart:

Center Line ( c ˉ \bar{c} c ˉ ): c ˉ = ∑ i = 1 k c i k \bar{c} = \frac{\sum_{i=1}^k c_i}{k} c ˉ = k ∑ i = 1 k ​ c i ​ ​

Standard Error ( σ c \sigma_c σ c ​ ): σ c = c ˉ \sigma_c = \sqrt{\bar{c}} σ c ​ = c ˉ ​

Control Limits: UCL c = c ˉ + 3 c ˉ \text{UCL}_c = \bar{c} + 3\sqrt{\bar{c}} UCL c ​ = c ˉ + 3 c ˉ ​ LCL c = max ⁡ ( 0 , c ˉ − 3 c ˉ ) \text{LCL}_c = \max\left(0, \bar{c} - 3\sqrt{\bar{c}}\right) LCL c ​ = max ( 0 , c ˉ − 3 c ˉ ​ )

Notice: No sample size n n n appears in the formula because the area of opportunity is standardized as a single unit ( n = 1 n = 1 n = 1 ).

The u u u -Chart (Defects per Unit for Variable Area)

When the area of opportunity varies across inspection periods (e.g., inspecting rolls of fabric of varying lengths, hospital wards with different patient-day counts, or varying batch sizes), the c c c -chart cannot be used. Instead, the u u u -chart monitors the average defect rate per unit : ν i = c i n i \nu_i = \frac{c_i}{n_i} ν i ​ = n i ​ c i ​ ​

(where c i c_i c i ​ is total defects observed across n i n_i n i ​ inspection units).

Formulas for u u u -Chart:

Center Line ( u ˉ \bar{u} u ˉ ): u ˉ = ∑ i = 1 k c i ∑ i = 1 k n i \bar{u} = \frac{\sum_{i=1}^k c_i}{\sum_{i=1}^k n_i} u ˉ = ∑ i = 1 k ​ n i ​ ∑ i = 1 k ​ c i ​ ​

Standard Error ( σ u i \sigma_{u_i} σ u i ​ ​ ): σ u i = u ˉ n i \sigma_{u_i} = \sqrt{\frac{\bar{u}}{n_i}} σ u i ​ ​ = n i ​ u ˉ ​ ​

Control Limits: UCL u i = u ˉ + 3 u ˉ n i \text{UCL}_{u_i} = \bar{u} + 3 \sqrt{\frac{\bar{u}}{n_i}} UCL u i ​ ​ = u ˉ + 3 n i ​ u ˉ ​ ​ LCL u i = max ⁡ ( 0 , u ˉ − 3 u ˉ n i ) \text{LCL}_{u_i} = \max\left(0, \bar{u} - 3 \sqrt{\frac{\bar{u}}{n_i}}\right) LCL u i ​ ​ = max ( 0 , u ˉ − 3 n i ​ u ˉ ​ ​ )

Like the p p p -chart, the u u u -chart displays stepped control limits when subgroup sizes n i n_i n i ​ vary.

  • Step-by-Step Worked Calculation Examples
  • Example 1: n p np n p -Chart Calculation in Manufacturing

Scenario: An injection molding line inspects a fixed sample of n = 400 n = 400 n = 400 medical syringe barrels every hour. Over k = 25 k = 25 k = 25 hours, inspectors identify a total of 300 300 300 defective barrels.

Calculate the overall proportion defective ( p ˉ \bar{p} p ˉ ​ ): p ˉ = 300 25 × 400 = 300 10 , 000 = 0.030 ( 3.0 % ) \bar{p} = \frac{300}{25 \times 400} = \frac{300}{10,000} = 0.030 \quad (3.0\%) p ˉ ​ = 25 × 400 300 ​ = 10 , 000 300 ​ = 0.030 ( 3.0% )

Calculate the Center Line for the n p np n p -chart: CL n p = n p ˉ = 400 × 0.030 = 12.0 defectives \text{CL}_{np} = n\bar{p} = 400 \times 0.030 = 12.0 \text{ defectives} CL n p ​ = n p ˉ ​ = 400 × 0.030 = 12.0 defectives

Calculate the standard error ( σ n p \sigma_{np} σ n p ​ ): σ n p = n p ˉ ( 1 − p ˉ ) = 400 × 0.030 × 0.970 = 11.64 ≈ 3.4117 \sigma_{np} = \sqrt{n\bar{p}(1 - \bar{p})} = \sqrt{400 \times 0.030 \times 0.970} = \sqrt{11.64} \approx 3.4117 σ n p ​ = n p ˉ ​ ( 1 − p ˉ ​ ) ​ = 400 × 0.030 × 0.970 ​ = 11.64 ​ ≈ 3.4117

Compute Control Limits: UCL n p = 12.0 + 3 ( 3.4117 ) = 12.0 + 10.235 = 22.235 ≈ 22.24 defectives \text{UCL}_{np} = 12.0 + 3(3.4117) = 12.0 + 10.235 = 22.235 \approx 22.24 \text{ defectives} UCL n p ​ = 12.0 + 3 ( 3.4117 ) = 12.0 + 10.235 = 22.235 ≈ 22.24 defectives LCL n p = max ⁡ ( 0 , 12.0 − 10.235 ) = 1.765 ≈ 1.76 defectives \text{LCL}_{np} = \max(0, 12.0 - 10.235) = 1.765 \approx 1.76 \text{ defectives} LCL n p ​ = max ( 0 , 12.0 − 10.235 ) = 1.765 ≈ 1.76 defectives

Example 2: c c c -Chart Calculation in Automotive Quality

Scenario: A luxury automotive assembly plant inspects finished vehicle paintwork for surface blemishes (clearcoat craters, dust inclusions, orange peel). Exactly 1 1 1 vehicle body ( n = 1 n = 1 n = 1 ) is inspected per audit period. Over k = 30 k = 30 k = 30 vehicles, auditors count a total of 270 270 270 defects.

Calculate the Center Line ( c ˉ \bar{c} c ˉ ): c ˉ = 270 30 = 9.0 defects per car \bar{c} = \frac{270}{30} = 9.0 \text{ defects per car} c ˉ = 30 270 ​ = 9.0 defects per car

Calculate the standard error ( σ c \sigma_c σ c ​ ): σ c = c ˉ = 9.0 = 3.0 \sigma_c = \sqrt{\bar{c}} = \sqrt{9.0} = 3.0 σ c ​ = c ˉ ​ = 9.0 ​ = 3.0

Compute Control Limits: UCL c = c ˉ + 3 c ˉ = 9.0 + 3 ( 3.0 ) = 9.0 + 9.0 = 18.0 defects \text{UCL}_c = \bar{c} + 3\sqrt{\bar{c}} = 9.0 + 3(3.0) = 9.0 + 9.0 = 18.0 \text{ defects} UCL c ​ = c ˉ + 3 c ˉ ​ = 9.0 + 3 ( 3.0 ) = 9.0 + 9.0 = 18.0 defects LCL c = max ⁡ ( 0 , 9.0 − 3 ( 3.0 ) ) = max ⁡ ( 0 , 0 ) = 0 defects \text{LCL}_c = \max(0, 9.0 - 3(3.0)) = \max(0, 0) = 0 \text{ defects} LCL c ​ = max ( 0 , 9.0 − 3 ( 3.0 )) = max ( 0 , 0 ) = 0 defects

Example 3: u u u -Chart Calculation with Variable Subgroup Sizes

Scenario: A software testing firm logs functional bugs per module. Because software modules vary in complexity, the size is measured in thousands of lines of code (KLOC). Across k = 20 k = 20 k = 20 modules totaling 100 100 100 KLOC, testers discover 400 400 400 bugs.

Calculate the average defect rate per unit ( u ˉ \bar{u} u ˉ ): u ˉ = 400 100 = 4.0 bugs per KLOC \bar{u} = \frac{400}{100} = 4.0 \text{ bugs per KLOC} u ˉ = 100 400 ​ = 4.0 bugs per KLOC

Calculate limits for a specific module measuring n i = 4 n_i = 4 n i ​ = 4 KLOC: σ u i = u ˉ n i = 4.0 4 = 1.0 = 1.0 \sigma_{u_i} = \sqrt{\frac{\bar{u}}{n_i}} = \sqrt{\frac{4.0}{4}} = \sqrt{1.0} = 1.0 σ u i ​ ​ = n i ​ u ˉ ​ ​ = 4 4.0 ​ ​ = 1.0 ​ = 1.0 UCL u = 4.0 + 3 ( 1.0 ) = 7.0 bugs/KLOC \text{UCL}_{u} = 4.0 + 3(1.0) = 7.0 \text{ bugs/KLOC} UCL u ​ = 4.0 + 3 ( 1.0 ) = 7.0 bugs/KLOC LCL u = 4.0 − 3 ( 1.0 ) = 1.0 bugs/KLOC \text{LCL}_{u} = 4.0 - 3(1.0) = 1.0 \text{ bugs/KLOC} LCL u ​ = 4.0 − 3 ( 1.0 ) = 1.0 bugs/KLOC

Calculate limits for a smaller module measuring n i = 1 n_i = 1 n i ​ = 1 KLOC: σ u i = 4.0 1 = 2.0 \sigma_{u_i} = \sqrt{\frac{4.0}{1}} = 2.0 σ u i ​ ​ = 1 4.0 ​ ​ = 2.0 UCL u = 4.0 + 3 ( 2.0 ) = 10.0 bugs/KLOC \text{UCL}_{u} = 4.0 + 3(2.0) = 10.0 \text{ bugs/KLOC} UCL u ​ = 4.0 + 3 ( 2.0 ) = 10.0 bugs/KLOC LCL u = max ⁡ ( 0 , 4.0 − 3 ( 2.0 ) ) = max ⁡ ( 0 , − 2.0 ) = 0 bugs/KLOC \text{LCL}_{u} = \max(0, 4.0 - 3(2.0)) = \max(0, -2.0) = 0 \text{ bugs/KLOC} LCL u ​ = max ( 0 , 4.0 − 3 ( 2.0 )) = max ( 0 , − 2.0 ) = 0 bugs/KLOC

Example 4: p p p -Chart Calculation with Variable Subgroup Sizes

Scenario: A commercial bank processes incoming mortgage applications. Over k = 20 k = 20 k = 20 business days, total applications received are 10 , 000 10,000 10 , 000 (averaging 500 per day), and 400 400 400 applications are rejected due to compliance defects.

Compute the process average proportion defective ( p ˉ \bar{p} p ˉ ​ ): p ˉ = 400 10 , 000 = 0.040 ( 4.0 % \bar{p} = \frac{400}{10,000} = 0.040 \quad (4.0\% p ˉ ​ = 10 , 000 400 ​ = 0.040 ( 4.0%

Control limits for a high-volume day ( n i = 1 , 000 n_i = 1,000 n i ​ = 1 , 000 applications): σ p = 0.040 ( 1 − 0.040 ) 1 , 000 = 0.0384 1 , 000 = 0.0000384 ≈ 0.006197 \sigma_{p} = \sqrt{\frac{0.040(1 - 0.040)}{1,000}} = \sqrt{\frac{0.0384}{1,000}} = \sqrt{0.0000384} \approx 0.006197 σ p ​ = 1 , 000 0.040 ( 1 − 0.040 ) ​ ​ = 1 , 000 0.0384 ​ ​ = 0.0000384 ​ ≈ 0.006197 UCL p = 0.040 + 3 ( 0.006197 ) = 0.040 + 0.01859 = 0.0586 ( 5.86 % ) \text{UCL}_p = 0.040 + 3(0.006197) = 0.040 + 0.01859 = 0.0586 \quad (5.86\%) UCL p ​ = 0.040 + 3 ( 0.006197 ) = 0.040 + 0.01859 = 0.0586 ( 5.86% ) LCL p = max ⁡ ( 0 , 0.040 − 0.01859 ) = 0.0214 ( 2.14 % ) \text{LCL}_p = \max(0, 0.040 - 0.01859) = 0.0214 \quad (2.14\%) LCL p ​ = max ( 0 , 0.040 − 0.01859 ) = 0.0214 ( 2.14% )

Control limits for a low-volume day ( n i = 400 n_i = 400 n i ​ = 400 applications): σ p = 0.040 × 0.960 400 = 0.000096 ≈ 0.009798 \sigma_{p} = \sqrt{\frac{0.040 \times 0.960}{400}} = \sqrt{0.000096} \approx 0.009798 σ p ​ = 400 0.040 × 0.960 ​ ​ = 0.000096 ​ ≈ 0.009798 UCL p = 0.040 + 3 ( 0.009798 ) = 0.040 + 0.02939 = 0.0694 ( 6.94 % ) \text{UCL}_p = 0.040 + 3(0.009798) = 0.040 + 0.02939 = 0.0694 \quad (6.94\%) UCL p ​ = 0.040 + 3 ( 0.009798 ) = 0.040 + 0.02939 = 0.0694 ( 6.94% ) LCL p = max ⁡ ( 0 , 0.040 − 0.02939 ) = 0.0106 ( 1.06 % ) \text{LCL}_p = \max(0, 0.040 - 0.02939) = 0.0106 \quad (1.06\%) LCL p ​ = max ( 0 , 0.040 − 0.02939 ) = 0.0106 ( 1.06% )

Notice that on the lower-volume day ( n = 400 n=400 n = 400 ), the control limits widen considerably ( [ 1.06 % , 6.94 % ] [1.06\%, 6.94\%] [ 1.06% , 6.94% ] vs [ 2.14 % , 5.86 % ] [2.14\%, 5.86\%] [ 2.14% , 5.86% ] ), demonstrating why stepped control limits naturally expand and contract based on sample size.

Critical CSSC Exam Traps

Trap 1: Choosing an n p np n p -Chart with Variable Subgroup Sizes — This is one of the most common trick questions on the exam. If the scenario mentions that sample size varies from day to day (e.g., "inspecting all incoming invoices each day"), you cannot use an n p np n p -chart. You must choose a p p p -chart.

Trap 2: Choosing a c c c -Chart for Variable Inspection Areas — If a quality auditor counts surface scratches on steel coils, but the square footage of the inspected coils varies, a c c c -chart is mathematically invalid. The u u u -chart must be used to normalize defects per unit of area.

Trap 3: Confusing Binomial and Poisson Formulas — Memorize that Binomial charts ( p , n p p, np p , n p ) always include the term ( 1 − p ˉ ) (1 - \bar{p}) ( 1 − p ˉ ​ ) under the square root. Poisson charts ( c , u c, u c , u ) never contain ( 1 − p ˉ ) (1 - \bar{p}) ( 1 − p ˉ ​ ) ; their variance is based strictly on c ˉ \bar{c} c ˉ or u ˉ / n \bar{u} / n u ˉ / n .

Trap 4: Negative Lower Control Limits — When calculating LCL for attribute charts where the average is small, the mathematical formula often results in a negative number (e.g., 1.5 − 3 ( 1.0 ) = − 1.5 1.5 - 3(1.0) = -1.5 1.5 − 3 ( 1.0 ) = − 1.5 ). In physical reality, you cannot have negative defectives or negative defects. On the exam, you must report LCL = 0 \text{LCL} = 0 LCL = 0 .

Loading diagram... Attribute Control Chart Selection Decision Matrix