Process Capability & Variation - Business Statistics

Two suppliers are being evaluated for process capability. Supplier X has $C_{pk} = 1.25$ with a process that drifts 0.5σ every month. Supplier Y has $C_{pk} = 1.10$ with a stable, well-controlled process. Considering long-term performance over 6 months, which supplier should be preferred?
Answer: Supplier Y because process stability compensates for the slightly lower initial capability measurement While Supplier X starts with higher $C_{pk}$ , the 0.5σ monthly drift means after 6 months, the process will have drifted 3σ from its original position, severely degrading actual capability. Supplier Y maintains consistent $C_{pk} = 1.10$ throughout the period. Choice A ignores the drift impact. Choice C treats the suppliers as equivalent, missing the stability issue. Choice D suggests insufficient information when the stability difference is the key factor. Process drift makes initially superior capability irrelevant over time.
An analyst reports that a process has a $C_p$ value of 1.8 and a $C_{pk}$ value of 0.7. The process is known to be stable and its output is normally distributed. What is the most logical first step for the process improvement team?
Answer: Investigate why the process mean is significantly off-center and take action to adjust it toward the specification target. A high $C_p$ value (1.8 is excellent) indicates that the process has very low inherent variation relative to the specification width. The process is potentially very capable. However, a low $C_{pk}$ value (0.7 is not capable) indicates that this potential is not being realized. The large discrepancy between $C_p$ and $C_{pk}$ is a direct indication that the process mean is far from the center of the specification limits. The most direct and often easiest improvement is to re-center the process mean.
A company manufactures pistons with a diameter specification of 75 mm ± 0.05 mm. An analysis of the manufacturing process, which is in statistical control, shows that the output is normally distributed with a mean of 75.01 mm and a standard deviation of 0.01 mm. In this context, the 'voice of the customer' is represented by what?
Answer: The piston diameter requirement of 75 mm with an allowable tolerance of ± 0.05 mm. The 'voice of the customer' refers to the requirements, specifications, and expectations of the customer. In this case, that is the specified target diameter and its tolerance (75 mm ± 0.05 mm). The 'voice of the process' is what the process is actually delivering, which is described by the distribution of its output (mean of 75.01 mm, std dev of 0.01 mm).
A process engineer is establishing an X-bar chart to monitor a process. A key distinction that the engineer must remember when constructing the chart is the difference between control limits and specification limits. Which statement most accurately describes this distinction?
Answer: Control limits are calculated from process data and represent the voice of the process, while specification limits are set by customers and represent the voice of the customer. This is a fundamental concept in SPC. Control limits are derived mathematically from the process's own data (mean and standard deviation), reflecting its natural, inherent variability (the 'voice of the process'). Specification limits are externally imposed requirements based on design or customer needs (the 'voice of the customer'). The two are independent, and there is no required mathematical relationship between them.
A manufacturing process has been monitored and is confirmed to be in statistical control. Its capability is assessed against specification limits of 100 ± 10. The process mean is 102, and its standard deviation is 2. A consultant notes that the process capability index ( $C_p$ ) is 1.67, while the process capability potential index ( $C_{pk}$ ) is 1.33. What is the most accurate interpretation of these indices?
Answer: The process has low inherent variation relative to the specifications, but its performance is slightly degraded because the process mean is not centered. A $C_p$ of 1.67 (which is (110-90)/(6*2) = 20/12) indicates that the process spread is much narrower than the specification width, meaning it has low inherent variation. A $C_{pk}$ of 1.33 (which is min[(110-102)/(3*2), (102-90)/(3*2)] = min[8/6, 12/6] = 1.33) is lower than $C_p$ , confirming that the process mean is off-center. However, a $C_{pk}$ of 1.33 is typically considered capable. Therefore, the process has low variation but its performance is slightly reduced by being off-center.
Two suppliers produce shafts with a target diameter of 25mm and specifications of 25 ± 0.5mm. Supplier A's process is centered at 25mm and has a standard deviation of 0.15mm. Supplier B's process is centered at 25.3mm but has a very small standard deviation of 0.04mm. All parts from both suppliers are within specification limits. From a modern quality perspective emphasizing being 'on-target with minimum variation', what is the primary concern with Supplier B's product?
Answer: Supplier B's parts are consistently off-target, which may cause systematic problems in assembly or product performance. Modern quality philosophies, like the Taguchi loss function, argue that any deviation from the target value incurs a 'loss', even if the part is within specification. Supplier B, while having low variation (which is good), produces parts that are consistently far from the target of 25mm. This systematic deviation can lead to issues like premature wear or improper fit when the parts are assembled. Supplier A, while having more variation, produces parts that are, on average, closer to the ideal target value.
A quality improvement team is analyzing a stable process. The process mean is perfectly centered between the specification limits. The team implements a change that successfully reduces the process standard deviation by 50%. How will this change affect the process capability indices $C_p$ and $C_{pk}$ ?
Answer: $C_p$ will double, and $C_{pk}$ will also double. The formula for $C_p$ is (USL - LSL) / (6σ). If the standard deviation (σ) is halved, the denominator is halved, causing the $C_p$ value to double. Because the process mean is perfectly centered, $C_p = C_{pk}$ . Therefore, if $C_p$ doubles, $C_{pk}$ must also double. The reduction in variation improves both potential and actual capability proportionally when the process is centered.
A process engineer claims that increasing the specification width by 20% while keeping the process unchanged will improve $C_p$ from 1.0 to 1.2. A colleague argues this approach is flawed because it doesn't address the root cause of capability issues. Which statement best evaluates these positions?
Answer: Both positions have merit: the calculation is correct, but the approach doesn't improve actual process performance The engineer's calculation is mathematically correct: $C_p = \frac{USL-LSL}{6\sigma}$ , so increasing specification width by 20% increases $C_p$ from 1.0 to 1.2. However, the colleague correctly identifies that this doesn't improve actual process performance - the process still produces the same variation and defect rate relative to customer needs. Choice A ignores the fundamental issue that capability should reflect process performance. Choice B incorrectly states the calculation is wrong. Choice D incorrectly claims $C_p$ is independent of specification width.
An initial capability study for a cutting process yields a $C_{pk}$ of 1.1, below the company's requirement of 1.33. After months of work to improve the process, the calculated $C_{pk}$ remains 1.1. A subsequent Gage R&R study reveals that the measurement system is responsible for 75% of the total observed variation. What is the most likely truth about the cutting process itself?
Answer: The cutting process was likely already capable, and the low $C_{pk}$ was an artifact of the poor measurement system. Total observed variation is the sum of the true process variation and the measurement system variation. If the measurement system accounts for a large portion (75%) of the observed variation, the true process variation is much smaller than what was measured. Since $C_{pk}$ is calculated using the standard deviation of the observed variation, the calculation was based on an inflated figure for variation. The true $C_{pk}$ , based on the smaller, actual process variation, would be significantly higher, meaning the process itself was likely capable all along.
Capability index $C_p$ is calculated using the within-subgroup (short-term) standard deviation, reflecting process potential. Performance index $P_p$ is calculated using the overall (long-term) standard deviation. If a process analysis yields $C_p = 1.5$ and $P_p = 0.8$ , what does this discrepancy most strongly suggest?
Answer: The process is unstable, with the mean shifting or drifting between subgroups over time, degrading long-term performance. When a process is perfectly stable, its short-term variation (within-subgroup) and long-term variation (overall) are nearly identical, and $C_p \approx P_p$ . When $C_p$ is significantly greater than $P_p$ , it indicates that the overall variation is much larger than the short-term variation. This inflation of long-term variation is caused by instabilities between subgroups, such as shifts or drifts in the process mean over time.
A chemical process is in statistical control with an average pH of 7.2 and a standard deviation of 0.15. What are the natural tolerance limits of this process, defined as the range expected to contain 99.73% of the output?
Answer: 6.75 and 7.65 The natural tolerance limits of a process are typically defined as the mean ± 3 standard deviations ( $\mu \pm 3\sigma$ ), which encompasses approximately 99.73% of the data in a normal distribution. Calculation: Lower Limit = 7.2 - (3 0.15) = 7.2 - 0.45 = 6.75. Upper Limit = 7.2 + (3 0.15) = 7.2 + 0.45 = 7.65. Specification limits are not needed as this calculation is based purely on the voice of the process.
If the company requires both $C_p ≥ 1.33$ and $C_{pk} ≥ 1.33$ for regulatory compliance, what is the most cost-effective process improvement strategy?
Answer: Adjust process mean to 500 mg while maintaining current standard deviation of 6 mg LSL = 475 mg, USL = 525 mg. Current $C_p = \frac{525-475}{6 × 6} = \frac{50}{36} = 1.39$ (meets requirement). Current $C_{pk} = \frac{\min(525-498, 498-475)}{3 × 6} = \frac{\min(27, 23)}{18} = \frac{23}{18} = 1.28$ (fails requirement). Since $C_p$ already exceeds 1.33, only centering is needed. Adjusting mean to 500 mg gives $C_{pk} = \frac{25}{18} = 1.39$ , meeting both requirements. Choice A unnecessarily reduces variation when $C_p$ is adequate. Choice C involves unnecessary cost. Choice D assumes impossibility when centering solves the problem.
A company produces shafts where the specification for a critical diameter is 30.00 ± 0.15 cm. A process capability study finds that the process mean is 30.00 cm and the standard deviation is 0.05 cm. The process is in statistical control. Which statement is the most accurate assessment of this process?
Answer: The process is centered and capable, with both $C_p$ and $C_{pk}$ equal to 1.0. When you encounter process capability questions, you need to evaluate two key metrics: $C_p$ (process capability) and $C_{pk}$ (process capability index that accounts for centering). These indices tell you whether a process can consistently produce parts within specifications. Let's calculate both indices for this process. The specification range is 30.00 ± 0.15 cm, so the upper specification limit (USL) is 30.15 cm and lower specification limit (LSL) is 29.85 cm. The total specification width is 0.30 cm. $C_p = \frac{\text{USL} - \text{LSL}}{6\sigma} = \frac{30.15 - 29.85}{6 \times 0.05} = \frac{0.30}{0.30} = 1.0$ Since the process mean (30.00 cm) equals the target specification (30.00 cm), the process is perfectly centered. For a centered process, $C_{pk} = C_p = 1.0$ . Values of 1.0 or higher indicate a capable process. Option A is wrong because it focuses on an irrelevant comparison between standard deviation and mean. Process capability compares variation to specification limits, not to the mean itself. Option B incorrectly states the process is off-center when the mean perfectly matches the target specification. Option C contains a major calculation error, claiming $C_p = 0.5$ when the correct value is 1.0, and incorrectly states the process isn't capable. Remember: $C_p$ measures inherent process capability, while $C_{pk}$ accounts for centering. Both should be ≥1.0 for a capable process, and they're equal only when the process is perfectly centered.
A manufacturing process for resistors is stable and highly capable, with a mean centered at the target of 500 ohms and a very low standard deviation. The specification limits are 490 ohms and 510 ohms. A calibration error causes the process mean to shift to 508 ohms, while the standard deviation remains unchanged. What is the most likely consequence for the product output?
Answer: The product output will see a significant increase in the proportion of resistors that exceed the upper specification limit of 510 ohms. Even though the new mean (508 ohms) is inside the specification limits, the process output follows a distribution around this mean. With the mean shifted so close to the upper specification limit (USL) of 510 ohms, a significant portion of the distribution's upper tail will now fall outside the USL, leading to a high number of defects on the high side. There will be virtually no defects on the low side.
Based on the capability indices provided in the passage, which statement best describes the comparison between these processes?
Answer: Process A has better inherent capability but Process B performs better relative to specifications due to better centering $C_p$ measures inherent capability (spread relative to specification width), while $C_{pk}$ accounts for process centering. Process A has higher $C_p$ (1.2 > 1.0), indicating better inherent capability, but lower $C_{pk}$ (0.9 < 1.0), showing it's poorly centered. Process B is perfectly centered ( $C_p = C_{pk}$ ) and meets minimum capability. Choice B incorrectly states B has better inherent capability. Choice C incorrectly averages the indices. Choice D ignores that A's $C_{pk} < 1.0$ .
A stable bottling process has a $C_p$ of 1.4 and a $C_{pk}$ of 1.4. A maintenance error causes the filler head to recalibrate, shifting the process mean closer to the upper specification limit without changing the process variation. What will be the most likely effect on $C_p$ and $C_{pk}$ ?
Answer: $C_p$ will remain the same, but $C_{pk}$ will decrease. $C_p$ measures the potential capability based on the spread of the process relative to the specification width. It is calculated as (USL - LSL) / 6σ. Since the specification limits and the standard deviation (variation) are unchanged, $C_p$ will remain the same. $C_{pk}$ measures the actual capability and accounts for the process mean's location. As the mean shifts away from the center, the distance to the nearest specification limit decreases, causing the $C_{pk}$ value to decrease.
An X-bar chart monitoring a production process shows the last seven consecutive points trending downwards, though all points remain within the control limits. The parts produced during this time were all measured and found to be within engineering specification limits. What is the most appropriate action for the process manager?
Answer: Initiate an investigation for a special cause of variation, as the downward trend is a signal of process instability. In statistical process control, a run of seven or more points in a consistent trend (up or down) is a rule for identifying an out-of-control condition, signaling the presence of a special cause of variation. Even if the points are within control limits and the output is within specification limits, the process is considered unstable. Process capability cannot be meaningfully assessed until the process is brought back into statistical control.
A process has specification limits of 50 grams (LSL) and 60 grams (USL). The process is stable with a mean of 56 grams and a standard deviation of 1.5 grams. Which calculation correctly determines the process capability index, $C_{pk}$ ?
Answer: $C_{pk} = \min[(60 - 56) / (3 \times 1.5), (56 - 50) / (3 \times 1.5)] = 0.89$ The formula for $C_{pk}$ is the minimum of two values: (USL - Mean) / (3σ) and (Mean - LSL) / (3σ). Here, USL=60, LSL=50, Mean=56, and σ=1.5. The calculation is $\min[(60 - 56) / (3 \times 1.5), (56 - 50) / (3 \times 1.5)] = \min[4 / 4.5, 6 / 4.5] = \min[0.89, 1.33]$ . The minimum value is 0.89. Choice A calculates $C_p$ . Choice D incorrectly uses 6σ in the denominator for $C_{pk}$ . Choice C uses an incorrect formula.
A process improvement team reduced common cause variation by 25% while keeping the process centered. If the original $C_p$ was 0.8, what process characteristic changed most significantly as a result of this improvement?
Answer: Both $C_p$ and $C_{pk}$ increased to approximately 1.07 due to reduced process variability with maintained centering Reducing common cause variation by 25% means the new standard deviation is 75% of the original. Since $C_p = \frac{USL-LSL}{6\sigma}$ , when σ decreases by 25%, $C_p$ increases by a factor of $\frac{1}{0.75} = 1.33$ . New $C_p = 0.8 × 1.33 ≈ 1.07$ . Since the process remained centered and $C_{pk}$ also depends on σ in the denominator, $C_{pk}$ increases by the same factor. Choice A incorrectly suggests the mean changed. Choice B misinterprets the relationship. Choice C incorrectly states $C_{pk}$ remained unchanged.
A call center manager is reviewing performance data. Which of the following observations is most likely attributable to a special cause of variation?
Answer: An unprecedented spike in call handle time on a single day, traced back to a network outage that affected all workstations. Special cause variation is due to specific, identifiable events that are not inherent to the process. A network outage is a distinct, non-routine event that would cause a significant, temporary deviation from normal performance. The other options describe common cause variation: random differences (A), predictable patterns like fatigue (B), and seasonal/cyclical effects (D) are all part of the inherent, expected variation of the process.

Start this quiz