7.2 Check Sheets, Histograms & Pareto Analysis

7.2 Check Sheets, Histograms & Pareto Analysis

Data-driven decision-making lies at the core of modern quality assurance. Without structured data collection and effective visual analytics, quality teams risk wasting resources on superficial symptoms rather than root causes. On the ASQ CQIA examination, candidates must understand how to capture raw shop-floor data using Check Sheets, visualize continuous data distributions using Histograms, and prioritize corrective actions using Pareto Analysis.

---

1. Check Sheets: Structured Frontline Data Collection

A Check Sheet (not to be confused with a procedural checklist) is a structured, visual form designed for real-time, manual data collection directly at the location where work is executed (gemba). It enables frontline operators to record occurrences, defect types, or operational parameters quickly and consistently.

+-------------------------------------------------------------------------+
|                   REPRESENTATIVE DEFECT CHECK SHEET                     |
+-------------------------------------------------------------------------+
| Process: Final Paint Inspection    | Inspector: J. Doe   | Date: 09/01  |
+------------------------------------+---------------------+--------------+
| DEFECT CATEGORY     | TALLY OF OCCURRENCES               | TOTAL COUNT  |
+---------------------+------------------------------------+--------------+
| Surface Scratch     | |||| |||| |||| ||                  | 17           |
| Paint Sag / Run     | |||| ||||                          | 10           |
| Orange Peel         | ||||                               | 4            |
| Contamination / Dust| |||| |||| |||| |||| |||            | 23           |
| Thin Coverage       | ||                                 | 2            |
+---------------------+------------------------------------+--------------+
| TOTAL INSPECTED: 500 units | TOTAL DEFECTS FOUND: 56     | DPU: 0.112   |
+-------------------------------------------------------------------------+

Primary Formats of Check Sheets

1. Tabular / Defect Item Check Sheet: A matrix listing predefined defect classifications along one axis and time intervals, machines, or shifts along the other. Tally marks record defect frequencies.
2. Location Check Sheet (Measles Chart): A physical blueprint, engineering drawing, or 2D silhouette of the product where inspectors mark an "X" or dot at the exact physical location of each observed defect. It visually reveals spatial defect clustering (e.g., solder bridge defects concentrating in the lower-right corner of a printed circuit board, indicating a faulty wave solder nozzle).
3. Process Parameter Check Sheet: A structured log used to record quantitative machine variables at scheduled intervals (e.g., furnace temperature, hydraulic pressure, conveyor speed).
4. Checklist vs. Check Sheet: A checklist is a procedural error-proofing tool (a list of mandatory steps an operator checks off to ensure complete compliance, such as pre-flight pilot checks). A check sheet is a data collection instrument designed to gather empirical defect counts or measurements for subsequent statistical analysis.

---

2. Histograms: Visualizing Continuous Data Distributions

A Histogram is a graphical representation of the frequency distribution of a continuous numeric dataset. While a bar chart plots discrete, categorical groups, a histogram groups continuous variable data into adjacent, non-overlapping intervals called bins or class intervals.

Frequency
    ▲
 25 ┼                     ┌───┐
 20 ┼                 ┌───┤   ├───┐
 15 ┼             ┌───┤   │   │   ├───┐
 10 ┼         ┌───┤   │   │   │   │   ├───┐
  5 ┼     ┌───┤   │   │   │   │   │   │   ├───┐
  0 ┴─────┴───┴───┴───┴───┴───┴───┴───┴───┴───┴──► Measurement (mm)
         9.6 9.7 9.8 9.9 10.0 10.1 10.2 10.3 10.4

Construction Guidelines

• Number of Bins ($k$): Selecting too few bins oversimplifies the distribution and hides patterns; selecting too many creates a jagged, uninformative shape. Common rules of thumb include:
  • Square Root Rule: $k = \sqrt{n}$, where $n$ is the total sample size.
  • Sturges' Rule: $k = 1 + 3.322 \log_{10}(n)$.
  • General industry heuristic: 5 to 20 bins for sample sizes ranging from 50 to 500 data points.
• Bin Width ($w$): Calculated as the total data range divided by the number of bins:

$w = \frac{\text{Maximum Value} - \text{Minimum Value}}{k}$

---

3. Interpreting Histogram Distribution Shapes

The profile of a histogram reveals critical insights into underlying process behavior, tooling wear, material mixing, and measurement integrity.

+-------------------------------------------------------------------------+
|                   HISTOGRAM DISTRIBUTION TAXONOMY                       |
+-------------------------------------------------------------------------+
|                                                                         |
|   1. BELL-SHAPED (Normal)       2. BIMODAL (Double Peak)                |
|             ┌─┐                           ┌─┐       ┌─┐                 |
|           ┌─┤ ├─┐                       ┌─┤ ├─┐   ┌─┤ ├─┐               |
|         ┌─┤ │ │ ├─┐                     │ │ │ ├───┤ │ │ │               |
|       ──┴─┴─┴─┴─┴─┴──                 ──┴─┴─┴─┴───┴─┴─┴─┴──             |
|       Stable, random variation        Two mixed streams / shifts        |
|                                                                         |
|   3. SKEWED RIGHT (Positive)    4. TRUNCATED (Cliff / Censored)         |
|         ┌─┐                           ┌─┐                               |
|       ┌─┤ ├───┐                     ┌─┤ ├─┐                             |
|       │ │ │   ├───┐                 │ │ │ ├─┐                           |
|       ┴─┴─┴───┴───┴──               ┴─┴─┴─┴─┴──────────────             |
|       Natural zero boundary         100% sorting / inspection culling   |
|                                                                         |
+-------------------------------------------------------------------------+

Detailed Diagnostic Profiles

---

4. Pareto Analysis and the 80/20 Rule

In the late 19th century, Italian economist Vilfredo Pareto observed that approximately 80% of the land and wealth in Italy was owned by 20% of the population. In the 1940s, quality luminary Joseph M. Juran recognized this mathematical phenomenon as a universal quality principle, coining the terms the "Vital Few and Useful Many" (originally the "vital few and trivial many").

The Pareto Principle in Quality Management

• In quality improvement, the Pareto Principle (the 80/20 Rule) states that roughly 80% of process problems, defects, or failure costs stem from approximately 20% of the underlying causes.
• Improvement teams cannot fix every defect simultaneously. Pareto analysis provides the empirical justification to focus limited organizational resources, budget, and engineering time on the vital few causes that will yield the greatest return on investment.

---

5. Construction and Anatomy of a Pareto Chart

A Pareto Chart is a specialized dual-axis hybrid chart combining sorted descending vertical bars with a cumulative percentage polygon line curve (an ogive).

Defect Count                                                  Cumulative %
    ▲                                                              ▲
100 ┼───┌────────┐                                            ─── 100%
 80 ┼───│        │───┌────────┐                  .- - - - •´  ───  80%
 60 ┼───│        │   │        │            . - •´             ───  60%
 40 ┼───│   45   │   │   25   │      . - •´                   ───  40%
 20 ┼───│ (45%)  │   │ (25%)  │  •-´ ┌────────┐ ┌────────┐   ───  20%
  0 ┴───┴────────┴───┴────────┴──┴───┴────────┴─┴────────┴───┴───   0%
         Dust /      Surface     Paint Sag     Orange      Other
         Contam.     Scratch                   Peel
        [  VITAL FEW (70%)  ]   [      USEFUL MANY (30%)     ]

Structural Components of a Pareto Chart

1. Horizontal Axis ($X$-Axis): Defect categories arranged in strict descending order of frequency or cost, starting with the largest category on the far left. An "Other" or "Miscellaneous" category is always placed on the far right, even if it is slightly larger than the smallest individual category.
2. Primary Vertical Axis (Left $Y$-Axis): Displays the direct measure of impact—defect count, frequency, labor hours, or monetary scrap cost ($).
3. Secondary Vertical Axis (Right $Y$-Axis): Displays the cumulative percentage scale, spanning strictly from 0% to 100%.
4. Cumulative Percentage Line (Ogive): Plotted points representing the running cumulative sum of percentages, beginning at the top of the first bar and terminating at exactly 100% above the final category.

Count-Based vs. Cost-Weighted Pareto Analysis

[!IMPORTANT] A frequent mistake in quality analysis is prioritizing defects strictly by raw frequency. A minor surface scratch occurring 100 times may cost $10 in polish rework ($1,000 total), whereas a cracked engine block occurring only 5 times may cost $1,500 per unit in total scrap ($7,500 total). Whenever cost data is available, construct a Cost-Weighted Pareto Chart to ensure financial priorities match operational focus.