IASSC Lean Six Sigma Green Belt Practice Questions — 10 Free (2026)

How should a Green Belt describe the contribution of the American quality teachers who lectured in Japan after 1950?

• A. They supplied statistical method and the plan do check act cycle that Japanese firms then developed in their own way ✓
• B. They designed the Toyota Production System and handed it over complete
• C. They taught only accounting
• D. They had no influence at all on Japanese practice, since the ideas that spread through industry there were entirely home grown inventions
• E. They insisted that quality was the responsibility of the inspection department rather than of production management

Visiting teachers brought statistical thinking and an improvement cycle that Japanese managers absorbed, taught widely and applied far beyond anything attempted at home. What they did not bring was the flow and pull architecture, which grew from Toyota's own constraints and experiments. Treating them as the authors of the whole system overstates the debt, while denying any influence ignores how quickly those methods spread through Japanese industry. Their central argument was in fact the opposite of leaving quality to inspectors, since they placed responsibility with management and with the process itself. IASSC Lean Six Sigma BOK, 1.4.2

A Kruskal-Wallis output reports 6 degrees of freedom. How many groups were compared?

• A. 4
• B. 6
• C. 7 ✓
• D. 10
• E. 12

Degrees of freedom for this test are one fewer than the number of groups, so the count of groups is the reported figure plus one. Six therefore points to seven groups. Reading the figure as the group count itself, or as the sample size, are the two usual slips. Seven groups also warn you that any follow up work involves twenty one pairs, so the alpha level needs careful protection. IASSC Lean Six Sigma BOK, 3.5.2

A bias study finds that a gage reads 0.06 mm high on a certified master. The characteristic being measured has a process spread of 1.20 mm. Expressed against that spread, how large is the bias?

• A. 2.5 percent
• B. 5.0 percent ✓
• C. 12.0 percent
• D. 20.0 percent
• E. 50.0 percent

Bias only becomes meaningful once it is compared with something, and here the chosen yardstick is the spread of the process itself. Dividing 0.06 by 1.20 gives 0.05, which is 5.0 percent of the spread the team has to work inside. Comparing the same 0.06 against half the spread would give 10.0 percent, and comparing it against a tolerance instead of a process spread would give a different figure again, which is why the basis must always be stated. A reading of this size is small but systematic, and it will bend every conclusion in the same direction until it is corrected. IASSC Lean Six Sigma BOK, 2.3.2

After a successful improvement project the average range falls to half of what it was. What should now happen to the control limits?

• A. Leave them alone, because limits must never change once they are set
• B. Leave them alone, because the smaller range will show as a nice trend
• C. Replace them with the tolerance, because the process is now capable
• D. Work out fresh limits from the data taken since the change was made ✓
• E. Halve the limits by eye so that the records can be brought up to date

Control limits describe a particular process, so when the process genuinely changes the old limits describe something that no longer exists and will fail to detect any further drift. Once enough subgroups have been collected under the new conditions, usually twenty or more, the limits are recalculated from that data alone and the old and new periods are shown separately on the chart. What must never happen is recalculating because a point was awkward, which is a different thing entirely from recalculating after a known and deliberate change. Halving limits by eye and swapping in the tolerance both abandon the arithmetic that gives the chart its meaning. IASSC Lean Six Sigma BOK, 5.2.3

An organisation trains dozens of green belts, but almost none complete a project. Which explanation should be examined first?

• A. Whether belts were given released time and sponsored projects to run ✓
• B. Whether the training material covered the statistical tools thoroughly
• C. Whether the belts were selected from senior enough grades of the staff
• D. Whether the examination standard was set too high for the candidates
• E. Whether the deployment had appointed enough master black belts first

The usual cause of a stalled deployment is not the training but the conditions afterwards: people return to a full workload, no named project and no sponsor asking about progress, and the learning fades within a few months. Checking release and sponsorship first is therefore the quickest way to find the real problem. Course content, seniority and coaching capacity all matter, but each of them would produce weaker projects rather than no projects at all. This is the same lesson the earlier improvement movements learned when structures were copied without the support that made them work. IASSC Lean Six Sigma BOK, 1.1.6

A plant that performs at roughly three sigma reports its COPQ as 4 per cent of sales. What is the most likely explanation for a figure that low?

• A. The plant makes a cheap product, so each failure costs very little
• B. Three sigma performance yields very few defects in the first place
• C. The finance team has wrongly included prevention and appraisal
• D. COPQ remains between three and five per cent in every industry
• E. Only the costs that were easy to count have gone into the total ✓

Performance around three sigma normally carries a quality cost well into double figures as a share of sales, so a reported 4 per cent points to an incomplete count rather than to an unusually cheap failure. Almost always the visible items have been added up and the hidden ones left out, which is why the estimate should be treated as a floor. Three sigma is not a low defect rate, and including prevention and appraisal would push the figure up rather than down. No single band of COPQ applies across all industries. IASSC Lean Six Sigma BOK, 1.2.3

Two independent machines each produce a sample of parts and the team wants to compare their average diameters. Which test is appropriate?

• A. A one sample t test using the average from the two machines
• B. A paired t test, matching each part on one machine to a part on the other
• C. A chi square test of association between machine and diameter
• D. A two sample t test comparing the means of the two machines ✓
• E. A one way analysis of variance

Two separate groups of continuous data, with no natural link between an individual part on one machine and an individual part on the other, is the two sample situation. The paired form only applies when each reading in one group has a genuine partner in the other, such as the same unit measured twice. A one sample test would compare a single group against a fixed target rather than against another group. Analysis of variance would give the same verdict here but is designed for three or more groups, and a chi square test would need counts in categories rather than measured diameters. IASSC Lean Six Sigma BOK, 3.3.4

A team must decide whether the mean fill weight of a line has drifted away from the 500 gram target. They have weighed 15 bottles, the readings plot as roughly normal, and no reliable long term figure for process spread exists. Which test fits the question?

• A. A paired t-test, pairing each bottle weight with the 500 gram target
• B. A one sample t-test of the 15 readings against the 500 gram target ✓
• C. A two sample t-test, treating the target as a second set of readings
• D. A one way ANOVA, because the target and the sample form two groups
• E. A one sample z-test, because the 500 gram target is a known number

One sample tests compare a single set of measurements against a stated value, and here that value is the 500 gram target. The t form is the right one because the spread of the process has to be estimated from the same 15 bottles rather than being known in advance; a z-test would need that spread supplied independently, and knowing the target is not the same thing. Pairing needs two readings on each item, but each bottle was weighed once. A target is a constant, not a second sample or a second group, so neither the two sample test nor ANOVA applies. IASSC Lean Six Sigma BOK, 3.4.1

A report states that a 95 per cent confidence interval for the mean response is 118 to 126, and adds that 95 per cent of units will therefore measure between 118 and 126. What is wrong with that addition?

• A. Nothing is wrong, since both statements mean the same thing
• B. That claim needs a prediction interval ✓
• C. The confidence level would have to be raised to 99 per cent first
• D. Individual units cannot be predicted from the fitted model at all
• E. The interval should have been centred on the median instead

A confidence interval for a mean response is a statement about where the average sits, and averages vary far less than the individual units that make them up. Covering individual units requires the prediction interval, which adds the residual scatter and is always wider. Quoting the narrow interval as though it covered single units will lead to promises the process cannot keep. Raising the confidence level does not repair the error, because the wrong quantity is being described in the first place. IASSC Lean Six Sigma BOK, 4.2.3

A line reports 99 per cent yield at final inspection, yet the rolled throughput yield across its eight steps is only 84 per cent. What does the difference measure?

• A. The work repeated inside the line before final inspection was reached ✓
• B. The units that failed at final inspection and were then sent for scrap
• C. The measurement error carried by the gauges used at each of the steps
• D. The share of orders that were despatched later than the date promised
• E. The count of separate opportunities each unit carries through the line

Final inspection only asks whether a unit is acceptable when it arrives there, so anything repaired, retested or quietly redone on the way is invisible to it. Rolled throughput yield asks the harder question of how many units went through every step correctly the first time, and the gap between the two figures is the cost of that hidden rework. Here roughly fifteen units in every hundred were touched more than once, which is where capacity, overtime and delay are being consumed. Scrap at final inspection is already inside the 99 per cent figure, and neither gauge error nor delivery performance is what the comparison measures. IASSC Lean Six Sigma BOK, 2.4.3