Quality Management & Statistical Quality Control — tap each question, test yourself, then take the mini-quiz.
00:00
0 / 20 opened
12
Quality Management & Six Sigma
1
What does TQM actually mean — and why does "total" matter?
Total Quality Management means managing the entire organization so it excels on all dimensions that matter to customers. The two operational goals are: (1) carefully design the product or service, and (2) ensure your systems can consistently produce that design. The word "total" is key — quality isn't just the inspector's job; it's everyone's job from the CEO to the front-line worker.
🎓 Campus Example
Think about your campus dining hall. TQM wouldn't just mean checking food temp before serving (that's inspection). It would mean the menu designers, the cooks, the servers, the cleaning crew, and even the purchasing department all working together so every meal is consistently safe, tasty, and on time.
2
What are the four costs of quality — and which one actually saves you money?
The four categories: Prevention costs (training, process design — spending upfront to stop defects), Appraisal costs (inspection, testing — catching defects before the customer), Internal failure costs (scrap, rework — defects caught inside), External failure costs (warranty claims, returns, lawsuits — defects the customer finds). The rule of thumb: every $1 spent on prevention saves $10 in failure and appraisal costs.
🎓 Campus Example
Imagine a campus print shop making 500 graduation programs. Prevention = proofreading the template before printing. Appraisal = spot-checking the first few copies. Internal failure = reprinting 50 copies because of a typo caught before delivery. External failure = the Dean hands out programs with the wrong date — embarrassing and costly to fix.
3
What is Six Sigma, and what does "six sigma" literally mean statistically?
Six Sigma is a philosophy and set of methods for eliminating defects by reducing process variation. Statistically, "six sigma" means fitting 6 standard deviations between the process mean and each specification limit. The goal: no more than 3.4 defects per million opportunities (DPMO).
DPMO = (Number of Defects / (Opportunities per Unit × Total Units)) × 1,000,000
🎓 Campus Example
Your university sends 10,000 financial aid letters per year. Each letter has 4 things that could go wrong (wrong name, wrong amount, wrong date, wrong address). That's 40,000 opportunities. If 12 errors occur: DPMO = (12 / 40,000) × 1,000,000 = 300 DPMO — that's roughly 5-sigma quality. Not bad, but not Six Sigma yet!
4
Walk me through DMAIC — what happens at each step?
DMAIC is the 5-step roadmap for Six Sigma improvement projects:
D – Define: Who are the customers and what are their priorities? M – Measure: How is the process performing right now? Collect data. A – Analyze: What are the most likely causes of the defects? I – Improve: What changes will remove those causes? C – Control: How do we lock in the improvements so they stick?
🎓 Campus Example
The campus coffee shop gets complaints about long wait times. Define: Students want coffee in under 3 min. Measure: Time 200 orders — avg is 5.2 min. Analyze: Fishbone diagram shows the blender station is the bottleneck. Improve: Add a second blender and pre-batch smoothie ingredients. Control: Post a timer visible to baristas; review weekly data.
5
What's a fishbone (Ishikawa) diagram, and when would you use one?
A cause-and-effect diagram that looks like a fish skeleton. The "head" is the problem; the "bones" are categories of potential causes (commonly: Methods, Machines, Materials, Measurements, People, Environment). You use it in the Analyze phase of DMAIC to brainstorm root causes before jumping to solutions.
🎓 Campus Example
Problem: "Late pizza deliveries on Friday & Saturday nights." Bones might include: People (drivers don't know campus), Machines (ovens too small for peak demand), Methods (poor dispatching), Materials (running out of dough). Each bone gets sub-causes branching off it.
6
What is poka-yoke, and why does Shingo say SPC isn't enough?
Poka-yoke = mistake-proofing. Shingo argued that statistical methods detect defects after they happen, but poka-yoke prevents them by designing the process so errors are impossible or immediately obvious. Types include checklists, special tooling, and source inspection (checking conditions before the operation).
🎓 Campus Example
Online course registration won't let you enroll in two classes at the same time slot — that's poka-yoke. The system prevents the mistake rather than letting you register and discovering the conflict later. Another: your USB-C cable only fits one way, so you can't plug it in wrong.
7
What are ISO 9000 and ISO 14000 — and why should a company care?
ISO 9000 is a set of international standards for quality management systems — it defines best practices for consistently meeting customer requirements. ISO 14000 focuses on environmental management. Certification tells your customers (and suppliers) that you follow a documented, auditable quality system. Many global companies require ISO certification from their suppliers.
🎓 Campus Example
Think of ISO certification like accreditation for your university. Without accreditation, employers might not trust your degree. Similarly, without ISO 9000, a manufacturer in Mexico might not be able to sell parts to Toyota — it's a globally recognized stamp of credibility.
8
What Six Sigma analytical tools should I know beyond the fishbone?
Key tools to remember: Flowchart (maps process steps), Check sheet (tally data at the source), Pareto chart (80/20 rule — focus on the vital few causes), Run chart (trend over time), Control chart (process stability — Ch. 13 goes deep), FMEA (Failure Mode and Effect Analysis — ranks risks by severity, likelihood, and detection), and DOE (Design of Experiments — tests multiple variables simultaneously).
🎓 Campus Example
Your student org runs a big concert. A Pareto chart of last year's complaints might show 60% came from just two things: sound quality and long entry lines. Fix those two and you've knocked out the majority of problems — that's the 80/20 principle in action.
9
What's the difference between "design quality" and "conformance quality"?
Design quality is the inherent value of the product in the marketplace — the features, aesthetics, and performance the designer intends. Conformance quality is the degree to which the product actually meets those design specs during production. You can have a brilliant design that's poorly manufactured (high design quality, low conformance) or a simple product made perfectly every time (lower design quality, high conformance).
🎓 Campus Example
A food truck designs a gourmet burger with wagyu beef and truffle aioli — that's high design quality. But if every third burger comes out overcooked and with the wrong toppings, conformance quality is low. Meanwhile, a basic vending machine snack has modest design quality, but it's identical every time — high conformance.
10
How do Crosby, Deming, and Juran differ in their approach to quality?
Crosby: Quality = conformance to requirements. "Do it right the first time." Rejects statistically acceptable defect levels — demands zero defects. Focuses on prevention.
Deming: Quality = a predictable degree of uniformity at low cost. Emphasizes reducing variation through statistical methods. Says management is responsible for 94% of quality problems (his 14 Points). Continuous improvement is everything.
Juran: Quality = fitness for use (satisfying customer needs). Management should own quality strategy. Recommends SPC but warns it can become a "tool-driven" approach if overemphasized.
🎓 Campus Example
Three professors grading essays — Crosby says "follow the rubric perfectly, no exceptions." Deming says "reduce variation in how we teach so students are better prepared — the system drives the results." Juran says "does the assignment actually prepare students for what they need?"
13
Statistical Quality Control & SPC
11
What's the difference between common (random) variation and assignable (special) variation?
Common variation is the natural, inherent randomness in any process — it's always present and predictable in the aggregate. Assignable variation is caused by specific, identifiable factors (a broken machine, a new untrained worker, a bad batch of materials). SPC's whole purpose is to detect assignable variation so you can investigate and fix the root cause.
🎓 Campus Example
Your morning commute to class varies between 12–18 minutes daily due to traffic lights and pedestrians — that's common variation. One day it takes 45 minutes because of a car accident blocking the road — that's assignable variation. SPC would flag the 45-minute day as abnormal.
12
What's the difference between "attributes" and "variables" in SPC?
Attributes = quality characteristics classified as conforming or not conforming (pass/fail, yes/no, good/defective). You count them. Variables = characteristics measured on a continuous scale (weight, length, temperature). You measure them with actual values.
This distinction determines which control chart you use: p-chart or c-chart for attributes, X̄ and R charts for variables.
🎓 Campus Example
Attribute: Checking if concert tickets were scanned correctly — each ticket either scanned or didn't (pass/fail). → p-chart. Variable: Measuring the temperature of coffee served at the campus café — 165°F, 172°F, 168°F → X̄ and R charts.
13
How does a p-chart work? When do you use it?
A p-chart tracks the proportion (fraction) defective in each sample over time. Use it when you're dealing with attribute data — items are either defective or not.
p̄ = Total defectives / Total items inspected
σp = √( p̄(1 − p̄) / n )
UCL = p̄ + 3σp
LCL = p̄ − 3σp (if negative, set to 0)
🎓 Campus Example
The campus bookstore ships online orders. Each day they check 200 orders and count how many had errors (wrong book, wrong edition, missing item). That fraction defective each day is plotted on a p-chart. If Monday's error rate suddenly jumps above the UCL, something assignable happened — maybe a new temp employee wasn't trained on the system.
14
How do X̄ and R charts work together?
The X̄ (X-bar) chart monitors the process mean — is the center of the process drifting? The R chart monitors the range (spread) within each sample — is variability increasing? You use them together because a process can shift its mean while staying consistent (X̄ out of control), or become erratic while the average looks fine (R out of control).
When σ is known: UCL/LCL for X̄ = X̿ ± zσ_x̄, where σ_x̄ = σ/√n When σ is unknown (new process): Use the A₂, D₃, D₄ factors from Exhibit 13.7.
The campus gym smoothie bar wants consistent 16 oz smoothies. Every hour they measure 5 smoothies. The X̄ chart shows if the average fill is drifting above or below 16 oz. The R chart shows if the barista is being inconsistent — some smoothies at 14 oz, others at 18 oz. Both need to be in control.
15
How do you read a control chart — what patterns signal trouble?
Beyond just "point outside the limits," watch for these red flags:
Single point beyond UCL or LCL → investigate immediately Run of 5+ points all above or below the center line → sustained shift Two points near a control limit → early warning of drift Trend (6+ points steadily increasing or decreasing) → progressive change Erratic behavior (wild swings) → multiple assignable causes Sudden level change → something changed abruptly in the process
🎓 Campus Example
You're tracking the wait time at the campus health center. If you see 7 consecutive days all above the center line, that's not random — maybe the flu season started and demand spiked, or a nurse went on leave. The pattern tells you to investigate even though no single day hit the UCL.
16
What is a c-chart, and how is it different from a p-chart?
Both are for attribute data, but they count different things. A p-chart tracks the fraction of units that are defective (defective units / total units). A c-chart tracks the count of defects per unit — a single unit can have multiple defects.
c-chart: UCL = c̄ + 3√c̄ | LCL = c̄ − 3√c̄
🎓 Campus Example
p-chart scenario: Checking if each student ID card printed correctly — pass or fail. c-chart scenario: Counting the number of typos on each page of the campus newspaper — one page might have 0 typos, another might have 3. You're counting defects per unit (per page), not just pass/fail.
17
What is Process Capability — and what does Cpk tell you?
Process capability asks: "Can our process meet the customer's specifications?" — a different question from "Is the process in statistical control?" (that's what control charts answer). Cpk measures how well the output fits within the specification limits, accounting for whether the process is centered.
Cpk = min[ (X̿ − LSL) / 3σ , (USL − X̿) / 3σ ]
Cpk > 1.0 → Process is capable
Cpk = 1.0 → Barely capable (3σ quality, 99.74%)
Cpk ≥ 1.33 → Good standard
Cpk < 1.0 → Not capable — producing defects
🎓 Campus Example
An airline specifies that seat pitch (legroom) must be between 30 and 34 inches. The manufacturing process produces seats with a mean of 31.5" and σ = 0.4". Cpk = min[(31.5−30)/(3×0.4), (34−31.5)/(3×0.4)] = min[1.25, 2.08] = 1.25. Capable, but the lower spec is the tighter side — the process mean is a bit closer to the minimum legroom.
18
"In control" vs. "capable" — what's the critical difference?
"In control" means the process is stable and predictable — only common variation is present (no assignable causes). Checked via control charts.
"Capable" means the process output fits within the customer's specification limits. Checked via Cpk.
A process can be in control but not capable (stable, but too much natural spread to meet specs). Or it could be capable but out of control (meeting specs now, but something is shifting — trouble is coming). You need BOTH before production.
🎓 Campus Example
Your campus shuttle consistently arrives within ±8 minutes of schedule (in control — predictable). But students need it within ±3 minutes to make class. It's stable but not capable of meeting that tighter expectation. Solution: either change the schedule (adjust the spec) or speed up the route (improve the process).
19
When do I use which chart? (Quick decision rule)
Use this quick decision tree:
Step 1: Is the quality characteristic measured on a continuous scale (weight, time, length)? → Variable data → use X̄ and R charts
Step 2: Is it pass/fail, good/bad? → Attribute data → then ask:
• Are you counting the fraction of defective units? → p-chart
• Are you counting the number of defects per unit? → c-chart
🎓 Campus Example
Staffing the campus rec center: tracking how many minutes each lifeguard takes to respond to a drill → X̄ & R. Tracking whether each fire extinguisher passes inspection → p-chart. Counting safety violations observed per shift → c-chart.
20
Big picture: what's the correct workflow — control chart THEN capability?
Yes — order matters! The three-step workflow:
1. Is the process in control? (Use control charts — remove assignable variation first) 2. Is the process capable? (Calculate Cpk — does it meet customer specs?) 3. Start production (only when both answers are yes)
If you check capability on an out-of-control process, the Cpk number is meaningless because the process is unstable and unpredictable.
🎓 Campus Example
Before launching a new campus meal delivery app, you'd first run a pilot to make sure order fulfillment is consistent (in control — no wild swings in delivery time). Then you'd check: can we deliver within the 30-minute promise 99.7% of the time? (capable). Only then do you go live.
🎯 Check Your Understanding
5 questions — answer, then submit to see your score
Q1.A campus coffee shop tracks whether each cup passes a taste test (pass/fail). Which chart should they use?
A X̄ and R chart
B p-chart
C c-chart
D Pareto chart
Q2.A Cpk of 0.85 means the process is:
A Highly capable with room to spare
B Exactly meeting specifications
C Not capable — producing defects outside spec limits
D In statistical control
Q3.In DMAIC, the "Analyze" phase focuses on:
A Determining the most likely causes of defects
B Identifying customers and their priorities
C Implementing solutions to remove causes
D Locking in improvements so they stick
Q4.You see 6 consecutive points above the center line on a control chart, but none exceed the UCL. What should you do?
A Nothing — all points are within limits
B Tighten the control limits
C Recalculate the center line
D Investigate — a run of 5+ points signals a sustained shift
Q5.Which quality cost category does "training employees on new procedures" fall into?