Not sure where your Black Belt preparation stands? These ten diagnostic questions are drawn at the official syllabus weights — Define 13% · Measure 21% · Analyze 27% · Improve 17% · Control 22% — so Analyze, the exam’s center of gravity, carries three of the ten. Our Black Belt bank is built at exactly those weights. Work the set cold, note which phase costs you points, and start your study plan there; the per-phase pages linked from the Black Belt program page carry more worked questions for whichever phase needs the work.
Worked questions
1. At Meridian Packaging, twelve improvement projects each report a local win, yet plant on-time delivery barely moves. Viewing the portfolio through Y = f(x) logic, how should the Master Black Belt respond?
A. Reconnect every project Y to the plant delivery outcome, then retire or rescope efforts whose vital Xs cannot influence that outcome materially
B. Celebrate the twelve local wins publicly and trust that plant delivery will follow once teams repeat the same approach next quarter, if nobody objects
C. Ask each team to widen its project Y so the reported gains look larger when leadership reviews the portfolio at year end
D. Replace the plant delivery measure with an average of the twelve project scores so progress becomes easier to demonstrate
Correct: A. Local wins that never move the business outcome signal broken linkage in Y = f(x); the Black Belt restores flow-down from the plant Y to each project's Xs and drops work that cannot affect it.
2. A bottling team brainstorms causes of underfilled bottles and fills every fishbone branch, then adjourns declaring the filler valve 'the cause' because it drew the most sticky notes. What correction must the Black Belt make?
A. The diagram needs a seventh category for software, because vote leaders on modern lines are almost always control-system faults, since software runs the line now
B. The team should vote again with managers present, since frontline voting alone lacks the authority to declare a root cause
C. The filler valve conclusion stands if the team reached consensus, because fishbone validity comes from group agreement
D. The diagram organizes hypotheses only; suspected causes still require data — stratification, checks, or trials — before any is treated as verified
Correct: D. A fishbone structures brainstorming into hypotheses grouped by category; popularity in a brainstorm is not evidence, so each candidate cause must be tested against process data.
3. Sample standard deviation from 25 parts is 2.0 using the n – 1 denominator. A trainee recomputes dividing by n and gets about 1.96, then claims the gage improved. What does the Black Belt explain?
A. The trainee is right, because the smaller denominator choice reveals the gage's true precision more accurately
B. Both figures describe the same data; dividing by n – 1 corrects the sample's tendency to understate the population spread
C. The n – 1 version is a mistake carried over from hand calculation, and modern practice divides by the sample size
D. The difference proves one of the 25 parts was measured incorrectly and the study should be repeated from scratch
Correct: B. s uses n – 1 (Bessel's correction) because a sample tends to sit closer to its own mean than to the population mean; 2.0 versus about 1.96 is the same data under two denominators, not a gage improvement.
4. A multi-vari study of shaft diameter samples parts across positions in a fixture, within production runs, and across shifts. How are these three sources classified?
A. Positions are temporal, runs are cyclical, and shifts are positional, following the order in which samples were collected
B. Fixture positions are positional, repeated runs are cyclical, and shift-to-shift change is temporal variation in the study
C. All three are temporal, because every source in a multi-vari study is ultimately a function of when parts were produced
D. Fixture positions are cyclical because parts repeat through the fixture, while runs and shifts are both classed as random
Correct: B. Multi-vari families: positional = within-piece or position-to-position, cyclical = piece-to-piece or run-to-run in sequence, temporal = longer time spans such as shifts.
5. A researcher, after seeing a two-sided p-value of 0.08, re-runs the analysis one-sided in the observed direction and reports p = 0.04 as a discovery. Why does the Black Belt reject that maneuver?
A. One-sided tests are never permissible in industrial work, so the original two-sided figure is the only reportable number in any setting
B. Direction must be committed before the data speak; choosing the favorable tail afterward doubles the real false-alarm rate
C. The maneuver fails because halving a p-value is arithmetically wrong; one-sided values are always one quarter of two-sided values
D. The rejection rests on sample size: post-hoc direction changes are acceptable provided the study enrolled at least 100 observations
Correct: B. A one-sided test spends its entire alpha in a direction fixed by the decision in advance; picking the tail that flatters the data after the fact breaks that contract and inflates error risk.
6. An ANOVA on inspection times assumes independence, but inspectors processed parts in long shared queues so neighboring measurements drift together. Which repair fits best?
A. Apply a log transformation to the inspection times, because transformations are the standard cure for correlated observations
B. Redesign data collection with randomized, separated runs so observations stop sharing queue-level influences with neighbors
C. Raise the significance level to 0.10, because correlated data need a more forgiving threshold to reach fair decisions overall
D. Pool the four groups into one large group, because dependence only matters when an analysis keeps groups distinct in a table
Correct: B. Dependence is a design defect, not a distributional one; no transformation or threshold change restores independence. Randomizing and separating runs so observations no longer share influences is the genuine repair.
7. A regression of fill weight on nozzle pressure gives slope 2.4 grams per psi with standard error 0.6. What is the t statistic for testing whether the slope differs from zero, and what does it say?
A. t = 2.4 x 0.6 = 1.44, borderline evidence since the product sits only modestly above the benchmark of one
B. t = 2.4 – 0.6 = 1.8, inconclusive evidence because subtracting the error leaves less than two full units
C. t = 2.4 / 0.6 = 4.0, strong evidence that pressure genuinely moves fill weight rather than the slope being zero
D. t = 0.6 / 2.4 = 0.25, weak evidence because the standard error is much smaller than the slope estimate itself
Correct: C. The slope t is estimate over its standard error: 2.4/0.6 = 4.0, far from zero, so the pressure effect is well supported.
8. A study's response is 'defects per panel,' but panels on the line vary in size by product. What response flaw needs fixing before runs begin?
A. The flaw is that defects per panel is continuous, and experimental responses should be attribute data
B. The response should be replaced by the number of panels produced, which correlates with defects anyway
C. The response is not comparable across runs unless panel size is accounted for, since bigger panels naturally accumulate more defects
D. No flaw exists, because defect counts are self-normalizing across products of different sizes in practice, and the budget would follow that reading
Correct: C. Counts need an exposure basis: defects per unit area or a size adjustment makes runs on different panels comparable.
9. A team debates adding either an alarm when a bin of inserts runs empty or a sensor that stops the press when inserts are absent. Volume is high and staffing is lean. Which reasoning is strongest?
A. Choose the press-stopping sensor: with lean staffing, an alarm may go unanswered while defective cycles continue
B. The alarm is stronger because it keeps a human in the decision loop, which Lean requires for every control point
C. Both are equivalent in effect, so the decision should be made purely on which device is cheaper to purchase
D. Choose the alarm, because stopping the press for a parts shortage wastes more value than the defects it prevents
Correct: A. Device selection weighs response reliability: a control that acts itself outperforms a warning when nobody may be free to respond.
10. A quality manager proposes judging a process improvement solely by counting out-of-control points before and after. The after period shows fewer signals but the chart's limits were also recalculated mid-period. Why is the comparison compromised?
A. It is not compromised, because signal counts are absolute measures unaffected by where limits were drawn during either of the compared periods
B. Signal counting is invalid because improvements should be judged only by specification conformance, never by control chart behavior of any kind
C. The comparison fails only because before-and-after studies require equal calendar lengths, and limit changes have no bearing on signal counting
D. Changing limits mid-comparison changes the measuring instrument itself; fewer signals may reflect wider limits rather than a better process
Correct: D. Signals are defined relative to limits. Recalculating limits during the after period moves the goalposts, so the count difference mixes process change with instrument change.
A note on format. The real IASSC exam mixes multiple-choice with true/false questions, while PassProfessor’s bank trains you with four-option multiple-choice questions throughout — the same knowledge, in a format that forces you to discriminate between close alternatives. No worked set on this page reproduces the real exam’s item-type mix.
These questions are drawn from the same bank as the full program. Start with the 20 free sample questions (also presented readably on the free sample questions page), read the IASSC Black Belt exam guide, then see the full IASSC Lean Six Sigma Black Belt Exam Prep program — 1,500 original questions built at the official syllabus weights (Define 13% · Measure 21% · Analyze 27% · Improve 17% · Control 22%), with an explanation for every answer, for $59 with 6 months of access.
Disclaimer
PassProfessor is an independent exam-preparation provider. This program is exam preparation only: it is not the real exam, it provides no education credit, and completing it does not by itself qualify you for any license or certification. The real exam is run by its own exam body — you register for it, pay its fees, and sit it through that body, not through PassProfessor. All practice questions in this program are original; they are not actual exam questions.
