CAPA Trending & Data Analysis 2 — Questions and Answers
Question 1: A quality engineer notices that defect counts have been slowly increasing over 6 consecutive data points on a control chart. What does this pattern indicate?
- Random variation within control limits
- A run suggesting a non-random trend requiring investigation (Correct answer)
- An outlier caused by measurement error
- Normal process behavior that requires no action
Correct answer: A run suggesting a non-random trend requiring investigation
Six or more consecutive points moving in one direction constitute a run, signaling a non-random trend that warrants CAPA investigation.
Question 2: Which data transformation technique is most appropriate when CAPA trending data spans several orders of magnitude?
- Moving average
- Logarithmic transformation (Correct answer)
- Z-score normalization
- Cumulative sum
Correct answer: Logarithmic transformation
A logarithmic transformation compresses wide data ranges, making trends and patterns easier to detect visually and statistically.
Question 3: During CAPA effectiveness review, a team finds that repair rework rates dropped from 8% to 6% after corrective action. What additional analysis should confirm the improvement is statistically significant?
- Recalculate the process capability index Cpk
- Perform a hypothesis test such as a two-proportion z-test (Correct answer)
- Draw a new Pareto chart of defect categories
- Update the FMEA severity ratings
Correct answer: Perform a hypothesis test such as a two-proportion z-test
A two-proportion z-test compares before-and-after defect rates to determine whether the observed reduction is statistically significant rather than due to chance.
Question 4: A scatter plot of process temperature versus defect rate shows data points scattered with no discernible pattern. What does this suggest?
- Temperature is the primary root cause of defects
- There is a strong positive correlation requiring corrective action
- Temperature and defect rate are not linearly correlated (Correct answer)
- The measurement system is out of calibration
Correct answer: Temperature and defect rate are not linearly correlated
A random scatter with no pattern indicates little to no linear relationship between temperature and defect rate.
Question 5: When applying exponential smoothing to CAPA defect trend data, what does a smoothing constant (α) close to 1 indicate?
- Historical data is weighted heavily over recent data
- Recent data points receive nearly all the weight in forecasts (Correct answer)
- The trend is stable and requires no corrective action
- Seasonal variation has been removed from the data
Correct answer: Recent data points receive nearly all the weight in forecasts
An α close to 1 means the forecast reacts almost entirely to the most recent observation, giving minimal weight to older data.
Question 6: A CAPA team uses a box plot to compare defect distributions before and after corrective action. The post-action box plot shows a smaller interquartile range. What does this indicate?
- The median defect rate increased after the action
- Process variability decreased following the corrective action (Correct answer)
- Outliers were removed from the dataset artificially
- The sample size was too small to draw conclusions
Correct answer: Process variability decreased following the corrective action
A smaller interquartile range on the post-action box plot indicates the middle 50% of values is more tightly clustered, reflecting reduced process variability.
Question 7: Which metric best quantifies the rate of CAPA recurrence when evaluating long-term trend effectiveness?
- Mean time to detect (MTTD)
- Repeat CAPA rate over a defined time window (Correct answer)
- Process capability ratio Cp
- First-pass yield percentage
Correct answer: Repeat CAPA rate over a defined time window
The repeat CAPA rate—the percentage of CAPAs for the same root cause within a defined period—directly measures whether corrective actions are preventing recurrence.
CAPA Trending & Data Analysis 2
A quality engineer notices that defect counts have been slowly increasing over 6 consecutive data points on a control chart.
What does this pattern indicate?