Quality Assurance Sampling - Business Statistics

1. An ideal Operating Characteristic (OC) curve would be L-shaped: the probability of acceptance would be 1.0 for all lots with a defect rate less than or equal to a specified level, and 0.0 for all lots with a defect rate greater than that level. Why is such an OC curve not achievable with acceptance sampling?
   Answer: Because sampling provides incomplete information, there is always a risk of making an incorrect decision about the lot. The fundamental reason an ideal OC curve is impossible is that acceptance sampling relies on inspecting a sample, not the entire population (lot). A sample provides only partial information. Due to random variation, a good-quality lot might yield a sample with a high number of defects, leading to its rejection (producer's risk). Conversely, a poor-quality lot might yield a sample with few or no defects, leading to its acceptance (consumer's risk). This inherent uncertainty, known as sampling risk, prevents the perfect discrimination depicted by an ideal OC curve.
2. What is the most likely consequence of changing the acceptance number from c=2 to c=3 while keeping the sample size constant?
   Answer: The producer's risk decreases, while the consumer's risk increases. Increasing the acceptance number (c) from 2 to 3 means that a lot will now be accepted even if 3 defective items are found in the sample, whereas before it would have been rejected. This makes the plan more lenient. A more lenient plan increases the probability of acceptance for any given defect rate. This means that good lots (at AQL) are more likely to be accepted, decreasing the producer's risk (α). Conversely, bad lots (at LTPD) are also more likely to be accepted, which increases the consumer's risk (β).
3. A company shifts from sourcing a component from a single, reliable supplier to multiple new suppliers with unknown quality histories. The company wants to adjust its incoming quality assurance sampling to protect against a potential increase in poor-quality lots. Which adjustment is most appropriate?
   Answer: Decrease the acceptance number (c) and/or increase the sample size (n) to make the plan stricter. With suppliers of unknown quality, the risk of receiving poor-quality lots increases. To protect the company (the consumer), the sampling plan must be made stricter. A stricter plan has a lower probability of accepting bad lots (i.e., a lower consumer's risk, β). This is achieved by either decreasing the acceptance number (c), which tightens the criteria for acceptance, or increasing the sample size (n), which increases the plan's ability to detect defects. Both actions make the plan more rigorous.
4. If the incoming lots are typically of very high quality (very low percent defective), what is the primary advantage of choosing Plan 2?
   Answer: The total number of items inspected on average will be lower with Plan 2. When you encounter sampling plan comparisons, focus on the key operational differences between single and double sampling plans, especially how sample sizes vary with lot quality. The crucial insight here is understanding average sample number (ASN). In Plan 1, you always inspect exactly 200 items regardless of what you find. In Plan 2's double sampling approach, you start by inspecting 125 items. If you find 0 or 1 defects, you accept the lot immediately without needing the second sample. Only if you find 2 or more defects do you inspect the additional 125 items. For very high quality lots (very low defect rates), Plan 2 will frequently find 0 or 1 defects in the first sample of 125, allowing immediate acceptance. This means the average total inspection will be much closer to 125 items rather than the full 250, making option A correct. Option B is wrong because "psychological advantage" isn't a statistical benefit and doesn't relate to inspection efficiency. Option C is incorrect because smaller sample sizes (when Plan 2 stops early) actually provide less precise defect rate estimates, not more accurate ones. Option D contradicts the passage, which explicitly states both plans have nearly identical OC curves, meaning their consumer's risks are essentially the same. Study tip: Remember that double sampling plans' main advantage is inspection economy—they can dramatically reduce average inspection when lot quality is either very good (early acceptance) or very poor (early rejection). Always consider how the stopping rules interact with the expected lot quality.
5. Assuming the given lot is representative of the incoming process quality, what is the approximate Average Outgoing Quality (AOQ)?
   Answer: AOQ = (0.24)(0.04) * (2000-100)/2000 = 0.00912 The formula for Average Outgoing Quality (AOQ) is AOQ = $P_a$ p ((N-n)/N), where $P_a$ is the probability of acceptance, p is the incoming proportion defective, N is the lot size, and n is the sample size. The term (N-n)/N is a finite population correction factor. Plugging in the values: AOQ = 0.24 0.04 ( (2000-100) / 2000 ) = 0.0096 (1900/2000) = 0.0096 0.95 = 0.00912.
6. A pharmaceutical company uses double sampling for quality inspection. In the first sample of 25 tablets, 1 defective is found. The sampling plan specifies: if ≤0 defectives in first sample, accept; if ≥3 defectives, reject; otherwise take second sample of 25. Given that a second sample is needed, what decision rule typically applies for the combined samples of 50 tablets?
   Answer: Accept if total defectives ≤ 2, reject if total defectives ≥ 3 across both samples combined In double sampling, when the first sample result falls in the intermediate zone (1-2 defectives), a second sample is taken and the decision is based on the combined total from both samples. The acceptance number for combined samples (n=50) is typically 2, so accept if ≤2 total defectives. Choice B only considers the second sample. Choice C uses too restrictive a limit. Choice D ignores the first sample result.
7. Based on this information, what is the producer's risk (α) for this plan, and what does it signify?
   Answer: The producer's risk is 0.08, signifying an 8% chance of rejecting a lot that meets the acceptable quality standard. The producer's risk (α) is the probability of rejecting a lot when its quality is at the Acceptable Quality Level (AQL). It is a Type I error. The problem states that the AQL is 2% and the probability of acceptance ($P_a$) at this quality level is 0.92. The probability of rejection is therefore 1 - $P_a$. So, α = 1 - 0.92 = 0.08, or 8%. This means there is an 8% chance that a lot meeting the AQL will be incorrectly rejected by the sampling plan.
8. A quality control manager implements a new single sampling plan. To increase the plan's ability to discriminate between lots of high quality and lots of low quality, the manager decides to significantly increase the sample size, $n$, while adjusting the acceptance number, $c$, proportionally. What is the primary effect of this change on the Operating Characteristic (OC) curve?
   Answer: The OC curve becomes steeper, reducing both producer's and consumer's risks for a given AQL and LTPD. Increasing the sample size (n) makes the sampling plan more powerful at discriminating between different quality levels. This is represented by a steeper OC curve. A steeper curve means there is a more rapid drop in the probability of acceptance as the lot's actual defect rate increases. This simultaneously reduces the producer's risk (α) at the AQL and the consumer's risk (β) at the LTPD, as the plan is better able to distinguish good lots from bad lots.
9. A food processing company uses variables sampling to monitor the fill weight of cereal boxes. The sampling plan uses $n = 16$ boxes with specification limits of 368g ± 12g. If the sample mean is 370.5g and sample standard deviation is 3.2g, what is the approximate upper specification limit capability index $C_{pu}$?
   Answer: 0.98 CPU = (USL - x̄)/(3s) where USL = 368 + 12 = 380g, x̄ = 370.5g, s = 3.2g. CPU = (380 - 370.5)/(3 × 3.2) = 9.5/9.6 = 0.99 ≈ 0.98. Choice A uses wrong denominator (2.5s). Choice C doubles the correct answer. Choice D uses LSL instead of USL in calculation.
10. A electronics manufacturer implements skip-lot sampling where every 5th lot is inspected when the process is stable. The normal sampling plan uses $n = 80$ with $c = 3$. After finding 4 defectives in an inspected lot during skip-lot sampling, what is the most appropriate immediate action?
    Answer: Return to 100% lot inspection until process stability is reestablished through consecutive acceptable lots Finding 4 defectives when c=3 means rejecting the lot, which indicates the process may no longer be stable. Skip-lot sampling requires process stability, so finding a rejected lot necessitates returning to 100% inspection until stability is reconfirmed. Choice B moves in wrong direction. Choice C doesn't address the stability issue. Choice D doesn't provide sufficient verification of process control.
11. An automotive parts supplier implements sequential sampling for critical components. The sampling plan has acceptance line $y_a = -2.5 + 0.1n$ and rejection line $y_r = 2.5 + 0.1n$, where $y$ is the cumulative number of defectives and $n$ is the sample size. After inspecting 40 parts with 3 defectives found, what action should be taken?
    Answer: Continue sampling because the cumulative defectives fall between the acceptance and rejection lines At n=40: acceptance line ya = -2.5 + 0.1(40) = 1.5, rejection line yr = 2.5 + 0.1(40) = 6.5. With 3 defectives found, we have 1.5 < 3 < 6.5, so continue sampling. Choice B is wrong because 3 > 1.5. Choice C misinterprets the decision rule. Choice D misunderstands sequential sampling protocol.
12. An appliance manufacturer compares two sampling plans for the same AQL: Plan A ($n = 32$, $c = 1$) and Plan B ($n = 50$, $c = 2$). If the true lot quality is 3% defective, which statement about the probability of acceptance is most accurate?
    Answer: Plan A will have higher acceptance probability due to its smaller sample size reducing detection sensitivity For Plan A: E(defectives) = 32×0.03 = 0.96, accept if ≤1. For Plan B: E(defectives) = 50×0.03 = 1.5, accept if ≤2. Plan A has higher chance of ≤1 defectives than Plan B has of ≤2 defectives. Smaller samples are less likely to detect defectives. Choice A incorrectly assumes larger sample helps acceptance. Choice B has backwards logic. Choice C ignores the different OC curve shapes.
13. A chemical company uses chain sampling for expensive destructive testing. The chain sampling plan CSP-1 with parameters $n = 10$, $c = 0$, and $i = 3$ is implemented. The last 4 lots had 0, 1, 0, and 0 defectives respectively, with the first 3 lots accepted. For the current lot with 1 defective found, what is the acceptance decision?
    Answer: Accept the lot because the cumulative defectives across the chain are within acceptable limits Chain Sampling Plan CSP-1 is designed for expensive destructive testing where you want to minimize sample sizes while maintaining quality control. Unlike standard acceptance sampling that evaluates each lot independently, chain sampling considers the cumulative performance across a sequence of consecutive lots. In CSP-1 with parameters $n=10$, $c=0$, and $i=3$, the decision rule works as follows: if the current lot has defectives exceeding $c=0$, you look back at the previous $i=3$ accepted lots. If those previous lots collectively had zero defectives, you can still accept the current lot despite it exceeding the individual lot criterion. Here's the sequence: the last 4 lots had 0, 1, 0, 0 defectives respectively, with the first 3 lots accepted. The current lot has 1 defective. Looking at the previous $i=3$ accepted lots, they had 0, 1, and 0 defectives. Wait—this seems problematic since the second lot had 1 defective but was accepted. This suggests those lots met the chain sampling criteria cumulatively. For the current lot with 1 defective, the chain principle allows acceptance because the cumulative defective pattern across the sequence remains within acceptable limits. Option A incorrectly assumes any lot with defectives breaks the chain. Option B misunderstands that $c=0$ applies only when the chain condition isn't met. Option C uses vague language about "sufficiently few" rather than the specific cumulative logic. Remember: Chain sampling trades individual lot strictness for cumulative quality assurance—focus on the collective performance across the specified number of previous lots, not just individual lot compliance.
14. A quality control plan is specified by a sample size (n) and an acceptance number (c). If the acceptance number $c$ is set to 0, what is the most significant consequence for the sampling plan?
    Answer: The consumer's risk (β) is minimized, offering the strongest protection against accepting poor-quality lots. When you encounter questions about acceptance sampling plans, focus on how the acceptance number $c$ affects the two key risks: producer's risk (rejecting good lots) and consumer's risk (accepting bad lots). Setting $c = 0$ creates the most stringent possible sampling plan—you reject the entire lot if even a single defective item appears in your sample. This dramatically reduces consumer's risk because it becomes extremely difficult for poor-quality lots to slip through. Even lots with small defect rates have a high probability of producing at least one defective item in the sample, leading to rejection. This gives consumers maximum protection against accepting substandard products. Option A correctly identifies this consumer protection benefit. Option B reverses the risks—while producer's risk does increase with $c = 0$, the question asks for the "most significant" consequence, which is the minimized consumer's risk. Option C misunderstands the Operating Characteristic (OC) curve's behavior. The OC curve doesn't become linear when $c = 0$; instead, it follows the exponential relationship $P_a = (1-p)^n$, where acceptance probability drops exponentially as defect rate increases. Option D is simply false—there's no statistical requirement for $c$ to exceed zero, and $c = 0$ plans are commonly used in practice. Remember that lower acceptance numbers always favor the consumer at the expense of the producer. When $c = 0$, you're implementing a "zero tolerance" policy that prioritizes consumer protection above all else.
15. A semiconductor manufacturer uses continuous sampling plan CSP-2 with parameters $f = 1/8$ (inspect every 8th unit) and $i = 4$ (required consecutive conforming units). Currently in the sampling phase, the last 6 inspected units showed the pattern: conforming, defective, conforming, conforming, defective, conforming. What is the required next action?
    Answer: Return to 100% inspection because defective units were found during the sampling phase In CSP-2, when a defective unit is found during the sampling phase, the plan immediately returns to 100% inspection. The plan only allows sampling to continue when i consecutive conforming units are found during 100% inspection. Finding any defective during sampling triggers return to 100% inspection. Choices A and D ignore this fundamental rule. Choice C describes a different sampling protocol.
16. In acceptance sampling, which situation presents the strongest justification for moving from a single sampling plan to a double or multiple sampling plan?
    Answer: When the inspection process is destructive and the cost per item inspected is very high. When you encounter acceptance sampling questions, focus on understanding when the costs and benefits of different sampling plans make economic sense. The key insight is that multiple sampling plans allow you to make decisions with smaller initial samples, which is crucial when testing is expensive. Answer A is correct because destructive testing with high per-item costs creates the strongest economic justification for double or multiple sampling. In destructive testing, each inspected item is destroyed (like crash-testing cars or testing medicine potency), making every test expensive. Multiple sampling plans start with smaller initial samples, allowing you to accept good lots quickly or reject obviously bad lots without destroying many items. Only when the initial small sample is inconclusive do you take additional samples. This minimizes the expected number of items destroyed while maintaining statistical confidence. Answer B is incorrect because high rejection costs for producers don't change the fundamental sampling economics - they just shift who bears the cost, not the efficiency of the sampling method itself. Answer C is wrong because consistently very good or very bad quality actually favors single sampling plans, since you'll reach clear accept/reject decisions quickly regardless of plan type. Answer D misunderstands the relationship between lot size and sampling - small lots typically make single sampling more practical, not less, since you need fewer samples relative to the lot size. Remember this pattern: Multiple sampling plans are most valuable when the cost per individual inspection is high, especially in destructive testing scenarios. The ability to make decisions with fewer total inspections on average drives the economic benefit.
17. Based on the passage above, which sampling plan parameter is most critical for meeting the specified rejection requirement for poor quality lots?
    Answer: The ratio of sample size to acceptance number must provide steep OC curve slope The requirement specifies rejecting lots with ≥6% defective with 90% probability, which means P(accept) ≤ 0.10 at p=0.06. This requires good discrimination (steep OC curve) between acceptable (2%) and rejectable (6%) quality levels. The n/c ratio determines curve steepness. Simply increasing n or decreasing c alone may not achieve the required discrimination. Inspection levels relate to sampling standards, not discrimination.
18. What is the primary purpose of calculating the Average Outgoing Quality (AOQ) in such a program?
    Answer: To predict the long-run average percentage of defective items in the product stream after the inspection process. Average Outgoing Quality (AOQ) is a measure of the quality of the product that the consumer receives after the acceptance sampling and rectification process has been applied. It represents the long-run expected proportion of defective items in the final product stream, considering that accepted lots contain some defects and rejected lots are theoretically made perfect. Its purpose is to understand the final quality being shipped.
19. A company implements an acceptance sampling plan where rejected lots are returned to the supplier. How does this practice differ from a rectifying inspection plan, and what is its main implication?
    Answer: It makes the calculation of Average Outgoing Quality (AOQ) invalid as a measure of final quality. The concept of Average Outgoing Quality (AOQ) is predicated on the assumption of rectifying inspection, where rejected lots are 100% inspected and made perfect before being re-introduced into the product stream. If lots are simply returned to the supplier, they are removed from the outgoing flow entirely. The outgoing quality is then simply the quality of the lots that were accepted. Therefore, the AOQ calculation, which blends the quality of accepted lots with the (theoretically perfect) quality of rectified lots, is no longer a valid model for the final product quality.
20. A quality engineer observes that for a specific sampling plan, the Average Outgoing Quality Limit (AOQL) is 1.8%. Which of the following is the correct interpretation of this value?
    Answer: Regardless of the incoming quality of lots, the long-run average quality sent to the customer will not exceed 1.8% defective. When you encounter questions about Average Outgoing Quality Limit (AOQL), you're dealing with acceptance sampling theory and the concept of quality protection for consumers. AOQL represents a fundamental guarantee about the worst-case average quality that will reach customers. The AOQL of 1.8% means that no matter how poor the incoming quality might be from the supplier, the long-run average quality of products actually delivered to customers cannot exceed 1.8% defective. This protection comes from the sampling plan's rejection and screening process: when lots are rejected, they're typically 100% inspected and defective items removed before shipping. This screening action, combined with the acceptance of good lots, ensures the outgoing quality stays within the AOQL limit. Option A is incorrect because AOQL doesn't specify what the supplier's process quality should be—it's about outgoing quality regardless of incoming quality. Option B misinterprets AOQL as a fixed value rather than an upper limit; the actual average outgoing quality will typically be below the AOQL. Option C confuses AOQL with the rejection threshold concept—lots aren't necessarily rejected just for exceeding the AOQL percentage, as rejection depends on the specific sampling plan parameters. Option D correctly captures that AOQL provides a quality ceiling guarantee to customers, functioning as protection regardless of supplier performance variability. Remember this key distinction: AOQL is about consumer protection and outgoing quality limits, not about incoming quality requirements or exact quality predictions. It's the "worst-case" average quality guarantee your sampling plan provides.

Start this quiz