CSSBB common mistakes
CSSBB · Understand the traps and how to avoid them.
On a control-chart question, declaring the process out of control only when a point exceeds the specification limits, or drawing specification limits on the chart.
Candidates carry over everyday quality thinking, where 'out of spec' equals 'bad,' and have not internalized that control limits are computed from process variation (typically mean ± 3 sigma), not from customer requirements.
What to do instead
Before answering, state the rule: control limits come from the process data, specification limits come from the customer, and the two are never mixed on an SPC chart. Drill the standard run rules (point beyond limits, runs, trends) separately from capability against specs.
Computing Cp from the specification width and process standard deviation, finding it above 1.33, and concluding the process is capable without checking Cpk or the process mean.
Cp is the simpler formula and is memorized first; the centering concept behind Cpk is treated as an afterthought rather than as the reason Cpk exists.
What to do instead
Always calculate and compare both indices. Check that Cpk ≤ Cp and interpret Cpk as the capability that reflects off-center performance. If the mean is shifted, Cp alone is misleading.
Accepting a gauge R&R result of about 28% GRR as 'good' because it is below 30%, without asking whether the measurement is used for process control or final product acceptance.
The 10%/30% thresholds are remembered as a single pass/fail cutoff, and the distinction between repeatability (equipment) and reproducibility (appraiser) is blurred.
What to do instead
Memorize the AIAG guidance explicitly: under 10% is generally acceptable, 10%–30% may be acceptable depending on the application and cost, and over 30% is unacceptable. Identify whether the variation is repeatability or reproducibility before recommending action.
Describing the champion as the person who owns the process after the project ends, or saying the Black Belt is the process owner.
The roles overlap in real organizations, and exam questions use titles that are not always used consistently in the candidate's workplace.
What to do instead
Use a one-line role map: champion/sponsor removes barriers and secures resources; Black Belt leads the project; Green Belt supports within their area; process owner owns the ongoing process after handoff. Match each duty to the role in the question.
Selecting a two-sample t-test for before-and-after measurements taken on the same units, or using a z-test with unknown sigma and a small sample.
Candidates match the test to the word 'two' or to the data type alone instead of to the structure of the data and the assumptions.
What to do instead
Classify the data first: paired or independent, normal or non-normal, continuous or count. Use the paired t-test when the same units are measured twice; use the two-sample t-test for independent groups; use z only when sigma is known or n is large.
Treating a Resolution IV design as though main effects are aliased with two-factor interactions, or ignoring the need to randomize the run order.
Resolution levels are memorized as labels without the alias structure behind them, and randomization is treated as a formality rather than as protection against nuisance variables.
What to do instead
Memorize the alias pattern: Resolution III confounds main effects with two-factor interactions; Resolution IV clears main effects but confounds two-factor interactions with each other; Resolution V clears both main effects and two-factor interactions. Always randomize unless blocking is deliberately used.
Describing the control plan as simply a list of inspections, or saying mistake-proofing works only if the operator remembers to check.
The control plan is often taught as a document to complete rather than as a system of process controls, monitoring, and reaction plans; poka-yoke is confused with awareness-based inspection.
What to do instead
Define the control plan as the documented system that identifies the critical inputs/outputs, the monitoring method, the frequency, and the reaction plan when a signal occurs. Define poka-yoke as a device or design that prevents or detects the error without relying on operator attention.
On a deployment question, treating a balanced scorecard, a dashboard, and Hoshin Kanri as interchangeable names for the same thing.
All three involve goals and visual tracking, so the terms blend together, and the candidate has not separated strategy deployment from performance reporting.
What to do instead
Keep the distinctions sharp: Hoshin Kanri is strategy deployment with catchball and annual objectives cascaded through the organization; the balanced scorecard organizes measures into perspectives; a dashboard is a visual display of metrics. Match the definition in the question to the correct term.
Classifying warranty claims as an appraisal cost or placing training costs under internal failure when categorizing the cost of poor quality.
The four COQ categories are memorized as a list without a clear decision rule, and 'quality cost' is assumed to mean only failure costs.
What to do instead
Use the timing rule: prevention before the process, appraisal during inspection/test, internal failure before delivery, external failure after delivery. Warranty and returns are external failure; training and design reviews are prevention; inspection and testing are appraisal; scrap and rework are internal failure.
Writing a Define-phase problem statement that embeds a cause or a solution, such as 'operators are careless, so we need more training,' or that lacks a measurable baseline.
Natural narrative style pushes candidates toward explaining why the problem exists rather than quantifying the gap, and the problem statement is treated as background rather than as a measurable charter element.
What to do instead
Write the problem statement as the measurable gap between current and desired performance, including metric, baseline, time frame, and location, with no cause and no solution. Move suspected causes to the analyze phase and proposed solutions to the improve phase.