26 free ASQ CSSBB Certified Six Sigma Black Belt practice questions from the CertStash pack of 211, each with the correct answer and a full explanation. The rule is the same for every exam: ten free on the product page, 26 free here, all 211 in the pack, and each step includes the one before it.
Get all 211 questions (US$39) · Download the first 10 as a PDF
Question 1
A ___________ is used primarily to track the stability of the average value of a metric of interest.
Show answer and explanation
Correct answer: B. Xbar-R Chart
An Xbar-R chart plots the average of each rational subgroup on the Xbar chart and the subgroup range on the R chart. The Xbar half is designed to show whether the process mean stays stable over time, which is exactly what tracking the average value of a metric requires. The paired range chart confirms that the within-subgroup variation is stable, so the limits on the Xbar chart are valid.
Why the other options are wrong
- A. An NP chart counts defective units in constant-size samples, so it tracks attribute counts, not an average.
- C. An I-MR chart plots individual readings and moving ranges, not subgroup averages.
- D. A C chart counts defects per unit area or item, which is attribute data rather than a mean.
Question 2
For her injection molding project a Belt needed to track the percentage of defectives of a particular sample set so she used a ___________ to display the data?
Show answer and explanation
Correct answer: D. P Chart
A P chart plots the proportion (or percentage) of defective units in each sample, and it can handle samples of different sizes because the limits are calculated from each sample's n. Tracking the percentage of defectives for injection molded parts is a binomial, pass or fail measure, so the P chart is the correct display.
Why the other options are wrong
- A. An Individuals chart is for continuous measurements taken one at a time, not for percent defective.
- B. A C chart counts the number of defects, not the proportion of defective units.
- C. An Xbar chart tracks subgroup averages of continuous data, not proportions.
Question 3
Which of these graphs demonstrates conditions which would be sufficient to enable OCAP for the process?
Show answer and explanation
Correct answer: A. Xbar Chart
An Out of Control Action Plan (OCAP) is triggered by a defined out-of-control signal, such as a point beyond a control limit or a nonrandom pattern judged against the center line. An Xbar chart has a center line and statistically calculated control limits, so it provides the conditions needed to tell operators when to act. A time series chart (run chart) only plots data over time without control limits, so it cannot by itself define the trigger that an OCAP depends on.
Why the other options are wrong
- B. A time series chart has no control limits, so it cannot signal when the reaction plan should be started.
- C. The Xbar chart does provide the control limits and rules an OCAP needs.
- D. Only the Xbar chart qualifies; the time series chart lacks control limits.
Question 4
Control Charts were developed by Dr. Shewhart to track data over time. To detect Special Cause variation the Control Charts use which of these?
Show answer and explanation
Correct answer: C. Center Line and Control Limits
Shewhart control charts separate common cause from special cause variation by plotting data in time order against a center line and upper and lower control limits, usually set at plus and minus three standard deviations. A point outside the limits, or a nonrandom pattern relative to the center line, signals a special cause that should be investigated.
Why the other options are wrong
- A. Data shift analysis is not the mechanism; shifts are detected by comparing points with the center line and limits.
- B. Control charts do not use separate outlier analysis methods to flag special causes.
- D. The center line and control limits are the correct mechanism, so this option is wrong.
Question 5
Common and ______________ Cause Variation are the focus of Statistical Process Control.
Show answer and explanation
Correct answer: C. Special
Statistical Process Control is built on the distinction between common cause variation, which is the natural, inherent noise of a stable process, and special cause variation, which comes from specific, assignable events outside the normal system. Control charts are used to detect special causes so they can be removed, while reducing common cause variation requires changing the process itself.
Why the other options are wrong
- A. Uncommon is not the standard SPC term for the second type of variation.
- B. Ordinary variation is another way of describing common cause, not its counterpart.
- D. Selective is not a category of variation in SPC.
Question 6
Special Cause Variation falls into which two categories?
Show answer and explanation
Correct answer: C. Assignable & Pattern
Special cause signals are grouped into two kinds. Assignable causes appear as individual points outside the control limits and can be traced to a specific event. Pattern causes appear as nonrandom behavior inside the limits, such as runs on one side of the center line, trends or cycles, which the Western Electric and Nelson rules are designed to detect.
Why the other options are wrong
- A. Natural and unnatural describe patterns loosely but are not the two categories of special cause.
- B. Short term and long term describe capability periods, not special cause types.
- D. Attribute and discrete describe data types, not categories of special cause variation.
Question 7
Range Charts are the technique used to determine if Special Causes are occurring within the subgroups of the ______________.
Show answer and explanation
Correct answer: B. SPC Charts
On paired SPC charts such as the Xbar-R chart, the Range chart shows the spread within each subgroup. If the range chart shows points outside its limits or nonrandom patterns, special causes are acting within the subgroups, and the limits on the Xbar chart cannot be trusted until that is resolved.
Why the other options are wrong
- A. Histograms show the distribution of data but do not have subgroups monitored over time.
- C. NP charts are single attribute charts without a companion range chart.
- D. Pareto charts rank categories by frequency and have no subgroups or control limits.
Question 8
If the production is for higher volume and monitoring and the Mean and variability is to be monitored for four machines producing product and the characteristic to be monitored is Variable Data, which SPC Chart is best to be selected?
Show answer and explanation
Correct answer: A. Xbar-R Chart
High volume production of variable (continuous) data lends itself to rational subgrouping, for example taking one part from each of the four machines to form a subgroup. The Xbar-R chart then monitors both the mean (Xbar) and the variability (range) of the characteristic, which is what the scenario requires. Ranges work well for the small subgroup sizes typical of this kind of sampling.
Why the other options are wrong
- B. An I-MR chart is used when only one reading is available at a time, such as low volume or slow processes.
- C. An NP chart is for counts of defectives, not continuous measurements.
- D. A CUSUM chart is used to detect small sustained shifts in the mean and does not monitor variability the way Xbar-R does.
Question 9
When a Belt Poka-Yokes a defect out of the process entirely then she should track the activity with a robust SPC system on the characteristic of interest in the defect as an early warning system.
Show answer and explanation
Correct answer: B. False
When a defect has been mistake proofed out of the process entirely, the cause can no longer occur, so there is nothing left for an SPC chart on that characteristic to warn about. Poka-yoke is one of the strongest control methods precisely because it removes the need for ongoing monitoring of that failure mode, freeing SPC resources for characteristics that still vary.
Why the other options are wrong
- A. Monitoring a defect that the poka-yoke has made impossible adds cost without providing useful information.
Question 10
Following the completion of a LSS project the Belt not only creates a Control Plan he also develops a _____________ so those involved in the process know what to do when the critical metrics move out of spec.
Show answer and explanation
Correct answer: A. Response Plan
A Response Plan (often part of the Control Plan, also called a reaction plan or OCAP) tells process owners and operators exactly what to do when a critical metric moves outside its limits or specification: who is notified, what is checked, how the product is contained and how the process is corrected. It ensures the gains of the project are sustained after the Belt hands the process back.
Why the other options are wrong
- B. A call list only identifies people to contact; it does not describe the corrective actions to take.
- C. A chain of command shows reporting lines but not the actions required when a metric goes out of control.
- D. A defect analysis plan is not the document used to guide immediate reaction to out-of-spec metrics.
Question 11
If an experiment has 5 factors and no replicates for a 2-level Experimental Design with 16 experimental runs which statement is incorrect?
Show answer and explanation
Correct answer: C. The Main Effects for the 5 factors are not aliased or confounded but the 2-way interactions are confounded with the 3-way interactions
Five factors in 16 runs is a 2 to the 5 minus 1 half fraction, normally built with the generator E = ABCD, which gives Resolution V. In a Resolution V design, each main effect is aliased with a four-way interaction and each two-factor interaction is aliased with a three-way interaction. Claiming that the main effects are not aliased or confounded at all is therefore the incorrect statement.
Why the other options are wrong
- A. Sixteen runs is half of the 32 runs a full factorial for five factors needs, so the design is a half fraction.
- B. In this Resolution V half fraction each main effect is aliased only with a four-way interaction, so this is correct.
- D. Every column in a balanced two-level design has half its runs at the high level, so 8 of 16 runs is correct.
Question 12
The method of Steepest Ascent guides you toward a target inside the original inference space.
Show answer and explanation
Correct answer: B. False
The method of steepest ascent uses the fitted first-order model to move the experiment along the path of greatest improvement, taking steps beyond the region where the original experiment was run. Its purpose is to lead toward an optimum outside the original inference space, where a new experiment is then centered.
Why the other options are wrong
- A. Steepest ascent deliberately moves outside the original inference space, so the statement is false.
Question 13
Which of these items contribute to what is necessary for successful Kaizen events?
Show answer and explanation
Correct answer: D. All of these answers are correct
Successful Kaizen events need analysis tools to understand the process and the waste in it, management support to provide resources and remove barriers, and operator support because the people who do the work carry out and sustain the changes. Missing any of these weakens the event, so all of the listed items contribute.
Why the other options are wrong
- A. Analysis tools are needed, but they are only one of the success factors listed.
- B. Management support is needed, but operator support and analysis tools are also required.
- C. Operator support is essential, but it is not the only requirement.
Question 14
A Belt working in a supply chain environment has to make a decision to change suppliers of critical raw materials for a new product upgrade. The purchasing manager is depending on the Belt's effort requiring that the average cost of an internal critical raw material component be less than or equal to $4,200 in order to stay within budget. Using a sample of 35 first article components, a Mean of the new product upgrade price of $4,060, and a Standard Deviation of $98 was estimated.
Select the answer that best states the Practical Problem.
Show answer and explanation
Correct answer: A. If the average cost per component is $4,200 or less, then the purchase manager will introduce the new product upgrade with new components.
The practical problem restates the business need in plain terms: the purchasing manager will go ahead with the new product upgrade only if the average cost per component is $4,200 or less, which keeps the project within budget. The sample mean of $4,060 and standard deviation of $98 are data used later to test that requirement statistically.
Why the other options are wrong
- B. An average above $4,200 would break the budget, so it is the opposite of the requirement.
- C. $4,060 is the sample mean, not the decision requirement, and the requirement is not an exact value.
- D. $180 does not appear in the requirement, which is an average cost of $4,200 or less.
Question 15
The Control Limits width varies if the sample size varies for which type of chart?
Show answer and explanation
Correct answer: A. P Charts
P chart control limits are calculated as p-bar plus or minus three times the square root of p-bar(1 minus p-bar) divided by n. Because n is in the formula, the limits widen for small samples and narrow for large samples, producing the stepped limits seen when sample sizes vary.
Why the other options are wrong
- B. NP charts require a constant sample size, so their limits do not vary in that way.
- C. Xbar-R charts are normally built with a constant subgroup size and fixed limits.
- D. Time series (run) charts have no control limits at all.
Question 16
Which of these elements are not included in Implementation plans?
Show answer and explanation
Correct answer: C. Cost/Benefit ratios
An implementation plan describes how an improvement will be rolled out: the tasks (work breakdown structure), schedule, owners, risk management plans and planned audits that confirm the work was completed. Cost/benefit ratios belong to project selection and justification, which is done before the implementation plan is built, so they are not part of the plan itself.
Why the other options are wrong
- A. A work breakdown structure lists the tasks needed and is a core part of an implementation plan.
- B. Risk management plans are included to handle problems that may arise during rollout.
- D. Planned audits confirm that the implementation steps were completed and are part of the plan.
Question 17
The Lean toolbox includes all of these items except ________________.
Show answer and explanation
Correct answer: C. Design of Experiments
The Lean toolbox focuses on eliminating waste and improving flow: mistake proofing, visual factory, inventory management, 5S, Kanban, standard work and similar tools. Design of Experiments is a statistical method from the Six Sigma toolbox used to model how inputs affect outputs, so it is not a Lean tool.
Why the other options are wrong
- A. Mistake proofing (poka-yoke) is a core Lean tool.
- B. Visual factory is a Lean method for making process status visible.
- D. Inventory management, such as Kanban and supermarkets, is central to Lean.
Question 18
Fractional Factorial, ____________and Response Surface Method are types of planned experiments.
Show answer and explanation
Correct answer: D. Factorial Design
Planned (designed) experiments include full factorial designs, fractional factorial designs and response surface methods. Each sets input factors at chosen levels in a structured matrix so that main effects and interactions can be estimated efficiently, which is what makes them planned experiments. In each, the experimenter actively changes the inputs instead of only observing them.
Why the other options are wrong
- A. Multi-vari analysis studies passive, observed data rather than a planned experiment.
- B. Baldrige refers to a performance excellence framework, not a type of experiment.
- C. One factor at a time testing changes inputs individually and cannot detect interactions, so it is not a designed experiment type.
Question 19
When a Belt properly analyzes the results of an experiment he must examine the Residuals in expectation of finding all of the following except _________.
Show answer and explanation
Correct answer: D. Residuals will represent a Linear Regression
Residuals from a good model are random: some are higher and some lower, scattered around zero with no structure, and normally almost all lie within about 2 standard deviations of the mean. Residuals that follow a linear trend show that the model has missed a relationship, so a linear pattern is not something to expect.
Why the other options are wrong
- A. Random residuals naturally include some higher values.
- B. Random residuals naturally include some lower values.
- C. Normally distributed residuals are expected to fall mostly within 2 standard deviations.
Question 20
If a Six Sigma project was to reduce repair station inventory and the team found the inventory was creeping up over time which Lean tools should be considered in the Control Phase to reestablish and sustain the project success?
Show answer and explanation
Correct answer: A. Review the Visual Factory to assure inventory in excess of desired visible
If repair station inventory is creeping back up, the Control Phase response is to reinforce the visual factory so that any inventory above the desired level is immediately visible, for example with marked maximum levels and labeled locations. This lets operators and supervisors see and correct the excess before it grows, which sustains the project gains.
Why the other options are wrong
- B. Better lighting does not define or reveal the desired inventory level.
- C. Supplier delivery data analysis is an Analyze activity and does not by itself control inventory at the stations.
- D. Rewording instructions into active verbs does not address the inventory creep.
Question 21
When doing Hypothesis Testing on Non-normal data Belts will use a ________________ to compare more than two sample proportions to each other.
Show answer and explanation
Correct answer: D. Contingency Table
To compare more than two sample proportions, the counts are arranged in a contingency table and analyzed with a chi-square test of independence. The test compares observed counts with the counts expected if the proportions were equal, which works for attribute (non-normal) data with any number of groups.
Why the other options are wrong
- A. Z score tables apply to normal data and to comparing one or two proportions, not several.
- B. Sakami Table is not a statistical method.
- C. Mean-to-mode analysis is not a test for comparing proportions.
Question 22
Which of these might contribute to similar distributions having Unequal Variance?
Show answer and explanation
Correct answer: D. All of the above
Two distributions that look alike can still have unequal variances when one has heavier or extreme tails, contains outliers, or has multiple modes from mixed sources. Each of these spreads the data out and inflates the variance, so all of the listed items can contribute.
Why the other options are wrong
- A. Extreme tails contribute, but outliers and multiple modes do as well.
- B. Outliers contribute, but they are not the only cause listed.
- C. Multiple modes contribute, but tails and outliers also inflate variance.
Question 23
A Full Factorial experiment using a 3 level 3 factor approach has been proposed to test the viability of an extrusion machine experiment.
How many treatment combinations will this approach involve?
Show answer and explanation
Correct answer: C. 27
The number of treatment combinations in a full factorial is the number of levels raised to the power of the number of factors. With 3 levels and 3 factors that is 3 to the power 3, or 27 treatment combinations. A three-level design is used when curvature must be estimated for each factor.
Why the other options are wrong
- A. Six is 3 plus 3 or 3 times 2, which is not how full factorial combinations are counted.
- B. Nine is 3 to the power 2, which would be only two factors at three levels.
- D. Fifty-four would be 27 combinations replicated twice.
Question 24
It would be more likely than not for a Belt conducting a Regression Analysis to find that the _______________________.
Show answer and explanation
Correct answer: A. r2 value is smaller than the absolute value of r
The correlation coefficient r lies between -1 and +1, and the coefficient of determination is r squared. Squaring any number with an absolute value below 1 makes it smaller, so r squared is normally smaller than the absolute value of r; they are equal only when r is 0 or exactly plus or minus 1.
Why the other options are wrong
- B. The correlation coefficient is r, not r squared.
- C. The coefficient of determination is r squared, so it cannot be less than itself.
- D. The correlation coefficient is r, not r divided by 2.
Question 25
Fractional Factorial Designs are used to analyze factors to model the output as a function of inputs if Hypothesis Testing in the Analyze Phase was inadequate to sufficiently narrow the factors that significantly impact the output(s).
Show answer and explanation
Correct answer: A. True
If hypothesis testing in the Analyze Phase has not narrowed the list of inputs enough, a fractional factorial design is used to screen the remaining factors efficiently and to build a model of the output as a function of the inputs. This identifies the vital few factors to carry forward into further experimentation or control.
Why the other options are wrong
- B. Fractional factorials are used for exactly this screening and modeling purpose, so the statement is true.
Question 26
Situations where standardized work needs to be incorporated include all of these except _____________________.
Show answer and explanation
Correct answer: A. Machines continually operating to reduce the labor cost per piece
Standardized work is needed wherever inconsistency causes waste: no defined method for keeping inventory at the right level, incomplete changeover instructions, or the same assembly taking different cycle times. Keeping machines running continually just to lower the labor cost per piece is overproduction, one of the Lean wastes, and it is not a situation that calls for standardized work.
Why the other options are wrong
- B. Without a defined system, inventory levels drift, so standard work is needed to set the method.
- C. Incomplete changeover instructions lead to inconsistent setups, a classic reason for standard work.
- D. Varying cycle times for the same assembly show the method is not standardized.
That was 26 of 211.
The full ASQ CSSBB Certified Six Sigma Black Belt pack has all 211 questions, each with the answer, the explanation and why the other options are wrong, plus a questions-only copy for timed runs. US$39, paid once, with free monthly updates and a pass-or-your-money-back guarantee.
Not ready yet? Get the first 105 free as one PDF by email. Get the Half Pack →