Overview
The Control phase protects the improvement after the project team steps away. Statistical process control (SPC) is its analytical core: understanding common and special cause variation, forming rational subgroups, selecting the right control chart for the data, calculating control limits and interpreting out-of-control signals with run rules. Sustaining improvements covers the control plan, updated procedures and work instructions, training, lessons learned and the formal handoff to the process owner, along with ongoing measurement to confirm benefits.
Lean tools for process control complete the section, chiefly total productive maintenance (TPM) with overall equipment effectiveness (OEE), and the visual factory, which makes the standard and any deviation visible at a glance. Candidates must choose charts correctly, read signals without overreacting, and know what a complete control plan contains.
How it's tested
Control questions focus on choosing the right chart from a data description, interpreting a chart against run rules, calculating simple control limits from constants, and deciding what belongs in a control plan or handoff. BEST and NEXT wording is common, and the right answer usually avoids tampering with a stable process while acting quickly on genuine special causes.
-
Choose the control chart
The stem describes the data, subgroup size and whether the sample size varies, and you pick X-bar R, X-bar s, I-MR, p, np, c or u.
-
Read the signal
A chart pattern is described, such as one point beyond a limit, a run on one side of the centerline or a steady trend, and you decide whether it is a special cause and what to do.
-
Complete the handoff
A project is closing and you identify the missing control plan element, such as a reaction plan, measurement method or owner, or the best way to sustain the gain.
An operator sees a point just inside the upper control limit after several normal readings and adjusts the machine. Asked whether this was correct, apply the rules: a single point within limits is not a signal on its own, and adjusting a stable process in response to common cause variation is tampering, which increases variation. The correct answer is to leave the process alone and continue monitoring.
Topics 3
Official topics from the ASQ outline, in outline order.
6.1 Statistical process control (SPC)Common and special cause variation, rational subgrouping, control chart selection, limits and run rules.
- What it means
- SPC uses control charts to separate common cause variation, which is inherent to the process, from special cause variation, which comes from an assignable source. Control limits are calculated from process data, typically at plus and minus 3 standard deviations of the plotted statistic, and describe what the process does, not what the customer wants.
- What to understand
- Select charts by data type and subgroup: X-bar R for small subgroups of continuous data, X-bar s for larger subgroups, I-MR for individual readings, p for proportion defective with variable sample size, np for number defective with constant size, c for defects in a constant area of opportunity, u for defects per unit with variable area. Rational subgroups should capture only common cause variation within the subgroup. Run rules flag non-random patterns such as a point beyond 3 sigma or a long run on one side of the centerline.
- How it shows up
- A process and its data are described and you pick the chart, or a pattern on the chart is described and you decide whether it is a special cause.
- Common confusion
- Control limits come from the process and specification limits come from the customer; candidates plot specification limits on a control chart and call the process in control because points are within spec.
6.2 Sustain improvementsControl plans, documentation, training, lessons learned and ongoing evaluation after handoff.
- What it means
- Sustaining improvements turns project changes into the normal way of working. The control plan lists each critical characteristic, its specification, measurement method, sample size and frequency, control method and the reaction plan when a signal or nonconformance occurs. Procedures, work instructions and training are updated, lessons learned are recorded, and the process owner takes over monitoring.
- What to understand
- The reaction plan is the element most often missing and most often tested: it says who does what when the control method detects a problem. Benefits should be verified over time against the baseline, and the project is formally closed with the sponsor and owner. The process FMEA and control plan should be updated together.
- How it shows up
- Questions show an incomplete control plan or a project that regressed after closure and ask what was missing, often the reaction plan, ownership or updated documentation.
- Common confusion
- Candidates think the Green Belt should keep monitoring the process indefinitely; the exam expects ownership to pass to the process owner with the tools to sustain it.
6.3 Lean tools for process controlTotal productive maintenance, overall equipment effectiveness and the visual factory.
- What it means
- Lean tools help keep the process in its improved state. Total productive maintenance (TPM) involves operators and maintenance staff in keeping equipment reliable through autonomous and planned maintenance. Overall equipment effectiveness (OEE) measures how well equipment is used. The visual factory uses signs, floor markings, shadow boards, andon lights and visual displays so the standard and any deviation are obvious.
- What to understand
- OEE = availability x performance x quality; for example 0.90 x 0.95 x 0.98 = 0.838, or about 84%. Availability losses come from breakdowns and changeovers, performance losses from slow cycles and minor stops, and quality losses from defects and startup scrap. Visual controls should show status at a glance without needing to ask anyone.
- How it shows up
- A scenario describes frequent unplanned stops or slow running and asks which tool or metric fits, or provides OEE components and asks for the overall figure.
- Common confusion
- Candidates add OEE components or take their average; OEE multiplies them, so a modest loss in each produces a much lower overall figure.
High-yield concepts
The terms, rules and formulas this domain turns into questions most often.
-
Common vs. special cause variation
Common cause variation is the natural, ever-present noise of a stable process; special cause variation comes from a specific, assignable source such as a new material lot or a worn tool.
- Remember
- Special causes are found and removed locally; reducing common cause variation requires changing the process itself, usually by management.
- Tested as
- Underlies every chart-interpretation question and the response that follows.
- Trap
- Treating every high reading as a special cause and adjusting the process, which is tampering.
-
Control chart selection
Continuous data: X-bar R for small subgroups, X-bar s for larger subgroups, I-MR for individual values. Attribute data: p (proportion defective, variable n), np (number defective, constant n), c (defects, constant area), u (defects per unit, variable area).
- Remember
- Defectives use p or np; defects use c or u; constant sample size allows np or c, variable size needs p or u.
- Tested as
- The single most frequent Control question pattern.
- Trap
- Choosing np when the daily sample size varies, or c when the inspected area differs between samples.
-
X-bar and R chart limits
For the X-bar chart, limits are the grand mean plus or minus A2 times R-bar; for the R chart, limits are D3 x R-bar and D4 x R-bar, using constants for the subgroup size.
- Remember
- For subgroups of 5, A2 is 0.577, D3 is 0 and D4 is 2.114.
- Tested as
- Short limit calculations with a constants table supplied or allowed in the open-book exam.
- Trap
- Using 3 times the standard deviation of individual values for an X-bar chart, which makes the limits far too wide.
-
Control limits vs. specification limits
Control limits are calculated from process data and show expected variation; specification limits are set by the customer or design and show what is acceptable.
- Remember
- A process can be in control yet out of specification, or out of control yet within specification.
- Tested as
- Interpretation questions that mix both sets of limits.
- Trap
- Drawing specification limits on an X-bar chart, where they are not comparable because the chart plots averages.
-
Run rules
Common rules flag one point beyond 3 sigma, two of three consecutive points beyond 2 sigma on the same side, four of five beyond 1 sigma on the same side, a long run on one side of the centerline (typically 7 to 9 points depending on the rule set), and a steady trend up or down.
- Remember
- Using more rules catches shifts sooner but increases false alarms.
- Tested as
- Pattern descriptions ask whether a signal exists.
- Trap
- Ignoring a long run on one side because every point is inside the limits.
-
Rational subgrouping
Subgroups are formed so that variation within a subgroup reflects only common causes, usually by taking consecutive units produced under the same conditions.
- Remember
- Mixing shifts, machines or lots inside one subgroup inflates within-subgroup variation and widens limits, hiding special causes.
- Tested as
- Questions on why a chart fails to detect known shifts.
- Trap
- Sampling one unit from each of five machines into one subgroup.
-
Control plan elements
A control plan lists, for each critical characteristic, the specification, measurement method, sample size and frequency, control method, responsible person and reaction plan.
- Remember
- The reaction plan tells the operator exactly what to do when a signal or nonconformance appears.
- Tested as
- Identify-the-missing-element questions and project closure scenarios.
- Trap
- Treating a control chart alone as a control plan.
-
Overall equipment effectiveness (OEE)
OEE = availability x performance x quality, measuring the share of planned production time that is truly productive.
- Remember
- Multiply the three factors: 90%, 95% and 98% give about 84%, not 94%.
- Tested as
- Short calculations and identifying which loss category a problem falls into.
- Trap
- Averaging or adding the three components.
-
Total productive maintenance (TPM)
A maintenance approach where operators perform routine care such as cleaning, inspection and lubrication, and maintenance staff focus on planned and preventive work to maximize equipment reliability.
- Remember
- TPM aims at zero unplanned breakdowns and shares ownership of equipment with operators.
- Tested as
- Scenarios with frequent breakdowns or slow running after an improvement.
- Trap
- Thinking TPM means the maintenance department alone takes on more work.
-
Visual factory
Visual controls such as floor markings, shadow boards, kanban signals, andon lights and performance boards make standards and abnormalities visible immediately.
- Remember
- A good visual control lets anyone see within seconds whether the process is normal or not.
- Tested as
- Choosing the tool that helps sustain a standard without constant supervision.
- Trap
- Posting dense reports on a wall and calling it visual management.
-
Tampering
Adjusting a stable process in response to common cause variation, which adds variation instead of reducing it.
- Remember
- Only adjust when a chart shows a special cause signal or when the process has a known, deliberate change.
- Tested as
- Operator-adjustment scenarios and the funnel experiment idea.
- Trap
- Adjusting after every reading that moves away from the target.
Side by side
Concepts candidates mix up, compared on the point that decides the answer.
p chart vs. np chart vs. c chart vs. u chart
| Concept | Meaning | Key difference | Common trap |
|---|---|---|---|
| p chart | Proportion of defective units per sample | Defectives; sample size can vary | Choosing np when sample size varies |
| np chart | Number of defective units per sample | Defectives; sample size must be constant | Using it for counts of defects |
| c chart | Number of defects per inspection unit | Defects; constant area of opportunity | Using it when unit sizes differ |
| u chart | Defects per unit | Defects; area of opportunity can vary | Using it for pass/fail defectives |
X-bar R vs. X-bar s vs. I-MR
| Concept | Meaning | Key difference | Common trap |
|---|---|---|---|
| X-bar R | Subgroup averages and ranges of continuous data | Small subgroups, commonly up to about 8 to 10 | Using it with one reading per sample |
| X-bar s | Subgroup averages and standard deviations | Larger subgroups, where the range wastes information | Thinking it is only for automated data |
| I-MR | Individual values and moving ranges between consecutive points | Subgroup size of one, such as batch results or slow processes | Forgetting it is sensitive to non-normal data |
Control limits vs. specification limits
| Concept | Meaning | Key difference | Common trap |
|---|---|---|---|
| Control limits | Calculated from process data, usually plus or minus 3 sigma of the plotted statistic | Voice of the process; show stability | Setting them by management target |
| Specification limits | Set by the customer or design for individual units | Voice of the customer; show acceptability | Plotting them on an averages chart |
Exam traps
Each trap is an answer that looks right. Learn the clue that should change your mind.
-
np chart with varying samples
- What candidates pick
- An np chart for daily defective counts when the number inspected changes each day.
- Why it looks right
- The data are counts of defectives, which np charts plot.
- Why it's wrong
- The np chart assumes a constant sample size; with variable sample sizes the p chart, which plots proportions with limits that vary by sample, is correct.
Clue to watch for A stated range of sample sizes, such as 80 to 140 per day.
-
In spec means in control
- What candidates pick
- Conclude the process is in control because all points fall within specification limits.
- Why it looks right
- No defective product is being made.
- Why it's wrong
- Control is judged against control limits and run rules; a process can produce good parts yet show special causes that will soon cause defects.
Clue to watch for The stem mentions specification limits but describes a run, trend or point beyond a control limit.
-
Adjusting on every reading
- What candidates pick
- Have operators adjust the setting whenever a value moves away from target.
- Why it looks right
- It seems responsive and quality-focused.
- Why it's wrong
- Adjusting a stable process for common cause variation is tampering and increases variation; act only on special-cause signals.
Clue to watch for The chart shows no rule violations, yet the option recommends adjustment.
-
Ignoring a run inside limits
- What candidates pick
- No action because every point is inside the control limits.
- Why it looks right
- Points beyond limits are the most familiar signal.
- Why it's wrong
- A long run on one side of the centerline or a steady trend is a non-random pattern indicating a shift, even without a point beyond the limits.
Clue to watch for A description such as nine consecutive points above the centerline or seven steadily increasing points.
-
Green Belt keeps monitoring
- What candidates pick
- The Green Belt continues to track the chart indefinitely after the project ends.
- Why it looks right
- It shows commitment to the result.
- Why it's wrong
- Sustainment depends on the process owner owning the control plan and reaction plan; the project should close with a formal handoff and benefit verification.
Clue to watch for The question concerns project closure or long-term sustainment.
-
Averaging OEE components
- What candidates pick
- Overall OEE equals the average of availability, performance and quality.
- Why it looks right
- Averaging feels like a fair summary.
- Why it's wrong
- OEE multiplies the three factors, so 90%, 95% and 98% give about 84%, not 94%.
Clue to watch for Three percentages labeled availability, performance and quality.
Answer framework
For Control phase questions about charts, signals or handoffs, apply these steps.
- Classify the data as continuous or attribute, and for attributes decide whether you are counting defects or defectives.
- Check subgroup size and whether the sample size or area of opportunity is constant, then choose the matching chart.
- Interpret against control limits and run rules, never specification limits, and decide whether a special cause signal exists.
- Act only on genuine signals by investigating and removing the special cause; leave a stable process alone to avoid tampering and added variation.
- For sustainment, confirm the control plan names the characteristic, method, frequency, owner and reaction plan, and that training and documents are updated before handoff.
Practice question
Want more like this? The CSSGB practice test has questions for every domain.
Checklist
0 of 12
Tick each one you can do without notes. Your ticks stay in this browser.
Study strategy
Start with SPC fundamentals: common versus special cause variation, rational subgrouping, and the difference between control and specification limits. Then build a one-page chart-selection guide keyed on data type, defects versus defectives and constant versus variable sample size, and drill it until you can choose a chart in seconds. Add limit calculations with the constants table and practice applying run rules to described patterns.
Next study sustaining improvements, focusing on the control plan's columns and especially the reaction plan, plus documentation, training and the handoff to the process owner. Finish with TPM, OEE calculations and visual factory examples, which are quick points once the multiplication rule for OEE is clear.
Candidates commonly underestimate this section because it has the fewest questions, but chart selection and run rules are precise and reward preparation. After this section, take a full-length CSSGB mock exam to tie the six phases together, since many Control questions assume the CTQs, measurement system and improved settings from earlier phases.
FAQ
How many Control questions are on the CSSGB?
ASQ lists 11 of the 100 scored questions for this section, the smallest share of the six sections.
How do I choose between p, np, c and u charts?
Use p or np for defectives (units that pass or fail) and c or u for defects (flaws counted on a unit). Choose np or c when the sample size or area is constant, and p or u when it varies.
What is the difference between control limits and specification limits?
Control limits are calculated from process data and show the expected range of common cause variation; specification limits come from the customer and define acceptable product. They answer different questions and should not be mixed on the same chart.
What is the most important part of a control plan?
Every element matters, but the reaction plan is the part most often missing: it states who must do what when the control method detects a signal or nonconformance.
How is OEE calculated?
Multiply availability, performance and quality. For example, 0.90 x 0.95 x 0.98 gives an OEE of about 0.84, or 84%.
How long should I study Control?
Plan roughly 10–15 hours within a 120-hour plan, mostly on chart selection, limit calculations and run rules.
How does Control connect to the other phases?
The control plan monitors the CTQs defined in Define using the measurement system validated in Measure, and protects the root-cause fixes and settings delivered in Analyze and Improve.
Other Quality Management certifications
Where the same subject shows up in other exams. Cards marked "Same subject" open the matching domain.
Test your Domain 6 knowledge
You have covered chart selection, control limits, run rules, control plans and OEE. Check whether you can choose the right chart and response quickly under exam conditions.