Lean Six Sigma Black Belt Certification Lean Six Sigma Black Belt Measure Phase: Data Analysis 1 — Questions and Answers
Question 1: A Black Belt plots a histogram of process output data and observes two distinct peaks. What does this bimodal distribution most likely indicate?
- The process is normally distributed with high variation
- Two different process conditions or input streams are mixed in the data (Correct answer)
- The measurement system has excessive bias
- The sample size is too small to draw conclusions
Correct answer: Two different process conditions or input streams are mixed in the data
A bimodal distribution typically signals that the data is drawn from two distinct subpopulations or process conditions (e.g., two shifts, two machines, two operators). Stratifying the data by the suspected factor is the correct next step.
Question 2: A Black Belt calculates a Cp of 1.45 and a Cpk of 0.72 for a critical dimension. What does this combination most likely indicate?
- The process spread is acceptable but the process is significantly off-center (Correct answer)
- Both process spread and centering are within acceptable limits
- The process is centered but the spread is too wide
- The measurement system is inadequate for this characteristic
Correct answer: The process spread is acceptable but the process is significantly off-center
Cp measures potential capability (spread relative to spec width) while Cpk accounts for centering. A high Cp with a low Cpk indicates the process has adequate spread but is poorly centered — the mean is shifted significantly toward one specification limit.
Question 3: During the Measure phase, a Black Belt needs to detect a 1.5-sigma shift in the process mean with 90% power using a two-sided t-test. Which action will most directly increase the statistical power of the test?
- Increase the significance level from 0.05 to 0.10
- Reduce the sample size to decrease variability
- Switch from a two-sided to a one-sided hypothesis test only
- Increase the sample size (Correct answer)
Correct answer: Increase the sample size
Statistical power increases with larger sample sizes because larger samples reduce the standard error, making it easier to detect true differences. While loosening alpha also increases power, increasing sample size is the most direct and statistically sound approach.
Question 4: A Black Belt finds that 68% of the total variation in a Gage R&R study is attributable to reproducibility (operator-to-operator variation). What is the most appropriate corrective action?
- Replace the measurement instrument with a higher-resolution device
- Standardize the measurement procedure and provide operator training (Correct answer)
- Increase the number of parts used in the Gage R&R study
- Accept the measurement system since reproducibility is within 70%
Correct answer: Standardize the measurement procedure and provide operator training
Reproducibility variation reflects differences between operators using the same gauge on the same parts. High reproducibility error points to inconsistent technique or unclear measurement procedures, so standardizing the method and training operators addresses the root cause directly.
Question 5: A Black Belt uses a box plot to compare defect counts across four production lines. Line C shows several data points plotted as individual dots beyond the box plot whiskers. What do these points represent?
- Measurement errors that should be removed from the dataset
- Data points that fall outside 1.5 times the interquartile range and are flagged as potential outliers (Correct answer)
- The maximum and minimum values of the dataset
- Values that exceed the upper specification limit
Correct answer: Data points that fall outside 1.5 times the interquartile range and are flagged as potential outliers
In a standard box plot, whiskers extend to 1.5 times the IQR from the quartiles. Individual points plotted beyond the whiskers are flagged as potential outliers — they warrant investigation but should not be automatically discarded without cause.
Question 6: A Black Belt is building a data collection plan and must decide how many subgroups to collect for an X-bar and R control chart. The process cycles approximately every 2 minutes. Which sampling strategy best balances rational subgrouping principles with practical constraints?
- Collect one large random sample at the end of the shift to maximize sample size
- Sample consecutively produced units within short time windows to form subgroups, spaced across the shift (Correct answer)
- Use a single subgroup of 30 units taken at the start of production
- Randomly select units from the entire shift output without regard to time order
Correct answer: Sample consecutively produced units within short time windows to form subgroups, spaced across the shift
Rational subgrouping requires that units within a subgroup be produced under as similar conditions as possible (minimizing within-subgroup variation), while subgroups are spaced over time to capture between-subgroup variation. Consecutive units within short windows, sampled periodically across the shift, best satisfies this principle.
Lean Six Sigma Black Belt Certification Lean Six Sigma Black Belt Measure Phase: Data Analysis 1
A Black Belt plots a histogram of process output data and observes two distinct peaks.
What does this bimodal distribution most likely indicate?