BY: Statistics Fundamentals Team
Reviewed By: Minsa A (Senior Statistics Editor)
Educational reference only — does not reproduce copyrighted ANSI/ASQ Z1.4 standard tables. Consult the official standard for production use.

Acceptance Sampling Table (ANSI/ASQ Z1.4) — AQL Lookup Guide

A complete educational guide to ANSI/ASQ Z1.4 acceptance sampling by attributes. Understand AQL, lot size, inspection levels, sample-size code letters, Ac/Re numbers, and single versus double sampling — with worked examples and step-by-step lookup instructions.

Quick Reference Summary

ANSI/ASQ Z1.4 is the American National Standard for sampling by attributes. It works through a chain: lot size + inspection level → code letter → sampling plan table → sample size (n) + Ac + Re. The AQL (Acceptance Quality Limit) is the quality level at which the producer's risk is approximately 5% — it does not mean "the percentage of defects allowed." A lot is accepted when observed nonconformities ≤ Ac and rejected when observed nonconformities ≥ Re. Exact table values and switching rules must be obtained from the current official standard, available from ASQ.

Acceptance Sampling Lookup Walkthrough

Enter your inputs below to walk through the ANSI/ASQ Z1.4 lookup process step by step. This tool illustrates the lookup logic — it does not reproduce official standard tables. Obtain Ac/Re values from the current official standard for production use.

⚠ Code letter shown follows published ANSI/ASQ Z1.4 look-up ranges. Sample size and Ac/Re numbers must be read from the official standard tables.

What Is Acceptance Sampling?

Acceptance sampling is a statistical inspection method used to decide whether to accept or reject a group of products — called a lot or batch — based on examining a randomly drawn sample. Instead of testing every unit (100% inspection), you inspect a smaller, statistically representative sample and apply a decision rule.

The method is widely used in incoming inspection of purchased parts, final inspection before shipment, and supplier quality audits. It balances inspection cost against the risk of accepting a bad lot or rejecting a good one.

Key point: Acceptance sampling does not improve lot quality — it only provides a decision about the lot. The quality of the product itself is determined by the manufacturing process. This is why acceptance sampling works alongside — not instead of — statistical process control.

What Is ANSI/ASQ Z1.4?

ANSI/ASQ Z1.4 is an American National Standard published by the American Society for Quality (ASQ) that provides sampling procedures and tables for inspection by attributes. "Attributes" means each unit is classified as either conforming or nonconforming (pass/fail), rather than measured on a continuous scale.

The standard grew from military specifications (MIL-STD-105) used during World War II and has been revised and maintained by ASQ since then. It is harmonized with the international standard ISO 2859-1, though the two documents are not identical. Organizations using either for contractual compliance should verify which edition and standard their contract or regulation requires.

What it covers

Attribute inspection (pass/fail). Single, double, and multiple sampling plans. Normal, tightened, and reduced inspection. General and special inspection levels. AQL-indexed sampling tables.

What it does NOT cover

Variable inspection (measurements). Process capability. Product design requirements. The standard complements — it does not replace — statistical process control or 100% inspection for critical applications.

Where to obtain it

Purchase the current edition directly from ASQ or ANSI. Do not rely on photocopied or unverified reproductions, which may be outdated or incomplete.

What Does AQL Mean?

AQL stands for Acceptance Quality Limit. In ANSI/ASQ Z1.4, it is defined as the quality level that is the worst tolerable process average when a continuing series of lots is submitted for acceptance sampling. It is the quality level at which the sampling plan provides a high probability of acceptance — approximately 95% — meaning the producer's risk of having a good lot rejected is roughly 5%.

Common Misconception

"AQL means the maximum percentage of defects allowed in a lot." — This is not accurate. AQL refers to a quality level for a process average over a series of lots, not a guarantee about any individual lot. An accepted lot may contain more than the AQL percentage of nonconforming units; a rejected lot may contain fewer.

Accurate Statement

AQL is the process quality level at which the sampling plan is designed to accept lots approximately 95% of the time. Lots produced at the AQL quality level pass the plan most of the time, protecting the producer from repeated rejection of acceptable-quality production.

Choosing an AQL involves balancing cost against risk. A very small AQL (e.g., 0.065%) requires tighter inspection and suits safety-critical components. A larger AQL (e.g., 6.5%) suits lower-risk cosmetic attributes where occasional nonconformities have minimal consequence. The AQL should be specified in your quality plan, customer contract, or applicable regulation — not chosen arbitrarily.

AQL vs LTPD

Two quality levels bracket a sampling plan from opposite sides. Understanding both helps you evaluate whether a plan actually protects your interests.

Feature AQL (Acceptance Quality Limit) LTPD (Lot Tolerance Percent Defective)
Full name Acceptance Quality Limit Lot Tolerance Percent Defective (also RQL)
Primary risk Producer's risk (~5% rejection of acceptable lots) Consumer's risk (~10% acceptance of bad lots)
Perspective Protects the supplier / producer Protects the buyer / consumer
Probability of acceptance ~95% at the AQL quality level ~10% at the LTPD quality level
Use in plan selection Index for selecting the sampling plan table Used in OC curve analysis; covered in ANSI/ASQ Z1.9 for variables

Lot Size in Acceptance Sampling

A lot (or batch) is the collection of units from which the inspection sample is drawn. The lot size is the total count of units in that collection. Defining the lot correctly matters: a lot should consist of units produced under essentially the same conditions, from the same production run, using the same materials and equipment.

Lot size is the first input in the ANSI/ASQ Z1.4 lookup. It partially determines the sample-size code letter. An important property: sample size grows much more slowly than lot size. A lot of 500 and a lot of 5,000 may require very different samples, but the ratio of sample to lot size shrinks dramatically as lot size grows. This reflects statistical theory: precision of proportion estimation depends mainly on absolute sample size, not on the fraction inspected.

Lot Size Range Example Code Letter (Gen. Level II) What changes vs lot size
2 – 8 A Very small lots; smallest sample sizes
51 – 90 E Code letter rises; sample grows but not proportionally
501 – 1,200 J Sample fraction inspected declines sharply
> 500,000 R Largest code letter; fraction inspected very small

Code letters are illustrative examples for General Inspection Level II. Verify exact ranges and codes in the current official ANSI/ASQ Z1.4 standard.

Inspection Levels in ANSI/ASQ Z1.4

The inspection level controls the relationship between lot size and sample size. A higher general inspection level yields a larger sample — more discriminating, but more costly. The inspection level is set by the responsible authority (quality plan, customer, or contract) and should reflect the risk associated with the characteristic being inspected.

General Level I

Smaller samples than Level II. Used when less discrimination is needed or the cost of inspection is high relative to the risk of nonconformity. Appropriate only when quality history justifies it.

General Level II

The default level in ANSI/ASQ Z1.4. Used unless specified otherwise. Provides the standard balance between sample size, cost, and statistical protection.

General Level III

Larger samples than Level II. Used when greater discrimination is needed — for example, when a characteristic carries high safety or regulatory significance.

Special Inspection Levels

Special inspection levels S-1 through S-4 allow very small samples. They are used when destructive testing is required, when inspection is very expensive per unit, or when slight discrimination is acceptable. Sample sizes under these levels are considerably smaller than general levels. S-4 gives the largest special-level sample; S-1 gives the smallest.

Important: The inspection level must be specified before inspection begins. Changing inspection levels during a production run — without following proper switching rules — undermines the statistical integrity of the sampling scheme. The choice of level should appear in your quality plan or be specified by the contracting authority.

Sample-Size Code Letters Explained

A sample-size code letter is an intermediate key — a letter (A through R, with some excluded) that links lot size and inspection level to the correct row in the sampling plan tables. The code letter is not the sample size itself.

Lookup Chain: Lot Size + Inspection Level Code Letter Sampling Plan Table n + Ac + Re

For a given sampling plan table (single, double, multiple) and AQL, each code letter corresponds to a specific sample size. Two different lot sizes may share the same code letter and therefore the same sampling plan.

Practical implication: If a code letter produces an arrow in the standard's table (directing you to a different sample size), follow the arrow and use the plan at the pointed-to row. This happens when the AQL column has no plan for the original code letter at the specified AQL.

Ac and Re: Acceptance and Rejection Numbers

The two numbers that determine lot disposition are the acceptance number (Ac) and rejection number (Re). These appear in the same row as the sample size in the official standard's tables.

Ac — Acceptance Number

The maximum number of nonconforming units or defects in the sample that still results in lot acceptance. If observed nonconformities ≤ Ac, accept the lot.

Re — Rejection Number

The minimum number of nonconforming units or defects that triggers rejection. If observed nonconformities ≥ Re, reject the lot.

Re is not always Ac + 1

In double and multiple sampling plans, there can be a gap between Ac and Re where a second (or further) sample is required before a final decision is made. In single sampling plans, Re = Ac + 1. Always read both values from the same table row in the official standard.

Observed nonconformities ≤ Ac → Accept the lot
Ac < Observed < Re → Second sample (double/multiple plans only)
Observed nonconformities ≥ Re → Reject the lot

Worked Example: Single Sampling Lookup

Illustrative Example — Values Are Hypothetical

This example uses hypothetical values to illustrate the lookup process. Do not use these values in production inspection. Obtain actual Ac/Re values from the current official ANSI/ASQ Z1.4 standard.

Scenario: A quality engineer receives a shipment of 800 electronic connectors. She applies normal, single sampling at General Inspection Level II, AQL = 1.0%.

Step-by-Step Lookup

Step Action Result
1 Determine lot size N = 800 units
2 Select inspection level General Level II (default)
3 Find code letter from Table I of the standard Code Letter J
4 Select AQL AQL = 1.0%
5 Look up Table II-A (single, normal) in the standard [Consult official standard]
6 Read sample size (n) — hypothetical for illustration n = 80 (hypothetical)
7 Read Ac/Re — hypothetical for illustration Ac = 2, Re = 3 (hypothetical)
8 Inspect 80 randomly selected connectors Found: 1 nonconforming unit
9 Apply decision rule 1 ≤ Ac (2) → Accept lot ✓

Interpretation

Acceptance does not mean the lot is defect-free. It means the sample evidence did not trigger rejection under this plan at this AQL. Some nonconforming units may still be present in the uninspected portion of the lot. The risk of accepting a lot with more than the AQL percentage of nonconformities is the consumer's risk, described by the OC curve of the plan.

Normal, Tightened, and Reduced Inspection

ANSI/ASQ Z1.4 provides three inspection modes that can switch based on the supplier's recent quality record. Switching rules are defined in the standard and must be followed exactly — the exact rule thresholds must come from the official document.

Inspection Mode Purpose Relative Sample Size Trigger
Normal Standard inspection mode Base Default start; used when no switch criterion is met
Tightened Increased protection; stricter acceptance Same or slightly larger; lower Ac Specified number of consecutive rejections under normal (see standard)
Reduced Reward for consistent quality; smaller sample Smaller Specified number of consecutive acceptances under normal (see standard)

Switching Rule Caution

The standard defines specific numerical thresholds for switching between modes. These must be read from the current official edition — they are not reproducible here. Applying switching rules incorrectly (e.g., moving to reduced inspection without meeting the qualifying conditions) invalidates the statistical properties of the sampling plan.

Single, Double, and Multiple Sampling Plans

ANSI/ASQ Z1.4 provides three types of sampling plans indexed to the same AQL and code letter. They differ in how many samples are drawn before a final decision is made.

Single Sampling Plan

One sample of fixed size n is drawn. Count nonconformities. Accept if ≤ Ac; reject if ≥ Re. Simple to administer and document. The sample size is larger than in equivalent double plans, but the procedure is the most straightforward.

Draw n → Count defects d → d ≤ Ac: Accept | d ≥ Re: Reject

Double Sampling Plan

A first sample (n₁) is drawn. If the result is clearly good (d₁ ≤ Ac₁) or clearly bad (d₁ ≥ Re₁), the lot is decided immediately. If d₁ falls between Ac₁ and Re₁, a second sample (n₂) is drawn. The combined count (d₁ + d₂) is compared to the second-stage Ac₂ and Re₂. Average sample number is lower than single sampling when lots are very good or very bad.

Sample 1: n₁ → d₁ ≤ Ac₁: Accept | d₁ ≥ Re₁: Reject | Ac₁ < d₁ < Re₁: Draw sample 2
Sample 2: n₂ → (d₁+d₂) ≤ Ac₂: Accept | (d₁+d₂) ≥ Re₂: Reject

Multiple Sampling Plan

Extends the double-sampling concept to up to seven sequential small samples. Each stage has its own Ac and Re. The plan reaches a decision as soon as the cumulative count crosses a threshold, or after all stages. Average sample number is the lowest of the three plan types for clearly good or bad lots, but administration is the most complex. ANSI/ASQ Z1.4 tables support up to seven stages.

Feature Single Double Multiple
Number of samples 1 1 or 2 1–7
Relative avg sample size Largest Medium Smallest
Administrative complexity Lowest Moderate Highest
Common use Routine inspection Efficient batches High-cost testing

The Operating Characteristic (OC) Curve

The Operating Characteristic curve is the single most important tool for evaluating a sampling plan's performance. It plots the probability of accepting a lot (Pa) on the vertical axis against the lot's true fraction nonconforming on the horizontal axis.

0 0.25 0.50 0.75 1.00 0% AQL LTPD High % Prob. of Acceptance (Pa) Lot Fraction Nonconforming → ~95% Pa at AQL (Producer protection) ~10% Pa at LTPD (Consumer protection) Illustrative OC Curve — Acceptance Sampling Plan

Reading the curve: at the AQL quality level (left side), the probability of acceptance is high — approximately 95%. This means a supplier consistently producing at the AQL will rarely have lots rejected. At the LTPD quality level (right side), the probability of acceptance is low — approximately 10%. This means bad lots are usually rejected.

Why the OC Curve Matters

No sampling plan has a perfectly sharp cutoff. There is always a range of lot quality where the plan may accept or reject a lot by chance. The shape of the OC curve reveals how sharp or gradual this transition is — a steeper curve means better discrimination between good and bad quality. Larger sample sizes produce steeper OC curves and better discrimination, at higher inspection cost. The NIST/SEMATECH Engineering Statistics Handbook provides further background on OC curves.

Producer's Risk vs Consumer's Risk

Acceptance sampling creates two kinds of errors, each affecting a different party.

Concept Meaning Who bears it? Statistical term
Producer's Risk (α) Probability of rejecting a lot that is actually at or better than the AQL Supplier / producer Type I Error
Consumer's Risk (β) Probability of accepting a lot that is worse than the LTPD Buyer / consumer Type II Error

Both risks are inherent in sampling. They cannot both be minimized simultaneously without increasing sample size. The sampling plan's design (sample size, Ac, Re) balances these risks according to the protection specified by the AQL and LTPD. For a deeper understanding of Type I and Type II errors in statistical testing, see the Type I and Type II errors guide on Statistics Fundamentals.

Defects vs Defectives vs Nonconformities

These terms are used in specific, distinct ways in quality standards. Using them interchangeably leads to errors in applying sampling plans.

Defect

A single instance where a characteristic does not meet a requirement. One unit can contain multiple defects. Sampling plans that count defects per unit (c-charts) use a different mathematical model from those that classify units as defective or not.

Defective (Defective Unit)

A unit that contains one or more defects. A unit is defective as a whole — even if it contains five distinct defects, it counts as one defective unit when classifying by attributes.

Nonconformity / Nonconforming Unit

Modern quality standards prefer "nonconformity" and "nonconforming unit" over "defect" and "defective" to avoid legal implications the word "defective" can carry. ANSI/ASQ Z1.4 uses these preferred terms. The underlying mathematics is the same.

Critical, Major, and Minor Defect Classifications

Organizations often classify defects by severity. A commonly used framework, though not universally mandated by ANSI/ASQ Z1.4 itself, distinguishes three levels:

Critical

A nonconformity that judgment and experience indicate is likely to result in unsafe conditions or failure in a critical end use. Often requires AQL = 0 or 100% inspection, depending on the quality plan.

Major

A nonconformity that is likely to result in product failure, significantly reduced usability, or substantial dissatisfaction. Typically assigned a low AQL (e.g., 0.65% or 1.0%).

Minor

A nonconformity that is unlikely to materially reduce the usability of the product or result in failure. Typically assigned a higher AQL (e.g., 2.5% or 4.0%).

Note: Defect classification definitions belong in your product quality plan, customer specification, or applicable regulation — not in the sampling standard itself. ANSI/ASQ Z1.4 provides the inspection framework; your organization defines what constitutes each class.

Acceptance Sampling vs 100% Inspection

Acceptance sampling is often compared with inspecting every unit. The choice depends on the application, risk, cost, and whether inspection is destructive.

Factor Acceptance Sampling 100% Inspection
Cost Lower — only a sample inspected Higher — every unit inspected
Time Faster Slower
Detection guarantee Probabilistic — some bad lots accepted Finds all defects (if inspection is perfect)
Sampling risk Present — quantified by OC curve Zero (assuming perfect inspectors)
Destructive testing Necessary — only sampling is feasible Not possible for destructive tests
Inspector fatigue Less likely — shorter inspection runs Risk of fatigue degrading detection rate

100% inspection by fallible human inspectors does not guarantee all defects are found. Automated 100% inspection avoids this limitation but at high capital cost. Neither method is universally superior — the choice depends on defect consequences, product volume, and economic constraints.

Acceptance Sampling vs Statistical Process Control

Acceptance sampling and statistical process control (SPC) serve fundamentally different purposes. Confusing them is one of the most common errors in quality management.

Acceptance Sampling

Focuses on the lot: accept or reject this batch? Looks backward at a finished batch. Does not address how to prevent the production of nonconforming units. Makes a binary lot decision. Explored through the binomial distribution.

Statistical Process Control

Focuses on the process: is it stable and capable? Monitors production in real time. Prevents nonconformities by detecting process shifts. Uses control charts and hypothesis testing logic.

A well-run quality system uses both. SPC maintains process control during production; acceptance sampling provides an independent check at delivery or receipt. Substituting one for the other leaves gaps in quality assurance.

Common Acceptance Sampling Mistakes

These errors appear repeatedly in practice and can invalidate the statistical protection a sampling plan is meant to provide.

Treating AQL as a guaranteed defect rate per lot

AQL is a process average concept, not a certificate for each lot. An individual accepted lot may contain more than the AQL fraction of nonconforming units.

Confusing the code letter with the sample size

The code letter (e.g., "J") is an index, not a count. The actual sample size must be read from the sampling plan table.

Ignoring switching rules

Failing to switch to tightened inspection after consecutive rejections removes one of the plan's key quality feedback mechanisms.

Using unverified or photocopied table values

Old editions, unofficial reproductions, and truncated table copies may have errors or be incomplete. Always use the current official edition.

Choosing the AQL arbitrarily

The AQL should reflect the application risk, regulatory requirements, or customer specification — not be selected because it results in a small or convenient sample size.

Assuming a passing lot is defect-free

Acceptance means the sample result did not trigger rejection — it does not certify the uninspected portion of the lot. Consumer's risk always remains.

Related Standards

Standard Scope Relationship to Z1.4
ISO 2859-1 Attribute sampling; international Closely harmonized but not identical; verify for contract use
MIL-STD-105E Military predecessor to Z1.4 Cancelled in 1995; Z1.4 is its civilian continuation
ANSI/ASQ Z1.9 Sampling by variables Companion standard; applies when measured values are available
ISO 2859-2/3 Attribute sampling; isolated/skip-lot Covers use cases outside Z1.4's AQL-indexed scheme

Glossary of Acceptance Sampling Terms

Term Definition Related Concept
Acceptance Sampling Statistical method to decide lot acceptance based on a sample ANSI/ASQ Z1.4, ISO 2859-1
AQL Acceptance Quality Limit — process average quality level for producer protection Producer's risk, OC curve
Lot A definite quantity of a product produced under essentially the same conditions Lot size, batch
Code Letter Index key (A–R) linking lot size and inspection level to the sampling plan Sample size, inspection level
Ac Acceptance number — maximum nonconformities to accept the lot Re, sampling plan
Re Rejection number — minimum nonconformities to reject the lot Ac, sampling plan
OC Curve Operating Characteristic curve — Pa vs fraction nonconforming for a plan AQL, LTPD, producer/consumer risk
LTPD Lot Tolerance Percent Defective — quality level at ~10% probability of acceptance Consumer's risk, OC curve
Producer's Risk Probability of rejecting a good lot; approximately α = 5% at AQL Type I error, AQL
Consumer's Risk Probability of accepting a bad lot; approximately β = 10% at LTPD Type II error, LTPD

Frequently Asked Questions

What inspection level should I use for routine incoming inspection?

General Inspection Level II is the default and is appropriate for most routine incoming inspection. Level I provides less discrimination and is only appropriate when quality history justifies it or inspection cost is prohibitive. Level III provides greater discrimination for high-risk characteristics. Your quality plan or customer contract specifies which level to use.

Can I use Z1.4 for destructive testing?

Yes. Special inspection levels S-1 through S-4 are designed for situations where testing is destructive or very expensive, permitting very small samples. The trade-off is reduced discriminating power — the OC curve is less steep, so the plan provides weaker separation between good and bad quality levels.

How do I choose the right AQL for my product?

AQL selection should be driven by the consequence of the defect type. Critical defects affecting safety typically use AQL = 0 or 100% inspection. Major defects affecting function might use AQL = 0.65% to 1.0%. Minor cosmetic defects might use AQL = 2.5% to 6.5%. Check your customer specifications, industry standards, and applicable regulations before setting AQL.

What happens if a lot is rejected?

A rejected lot can be: returned to the supplier, subjected to 100% screening (with nonconforming units removed or reworked), accepted under deviation/waiver if the user accepts the risk, or scrapped. The disposition decision is a business and quality decision — Z1.4 specifies the inspection method, not the corrective action. Rejections also trigger the switching rule evaluation for tightened inspection.

Is double sampling always better than single sampling?

Not always. Double sampling has a lower average sample number when lots are clearly good or clearly bad, saving inspection effort. But it requires more complex administration and recordkeeping, and when lots are marginal quality, the average sample number can exceed single sampling. The best choice depends on lot quality history, inspection complexity tolerance, and whether the sampling environment can support staging two draws from the same lot.

Does ANSI/ASQ Z1.4 apply to service industries?

Z1.4 was developed for manufactured goods but the underlying attribute sampling principles apply wherever units can be classified as conforming or nonconforming — including service outputs like forms, invoices, or processed transactions. The key requirements are that units are independent and that the lot is well-defined before inspection begins.

Where can I access the official ANSI/ASQ Z1.4 standard?

The current edition is available for purchase directly from ASQ (asq.org) and from ANSI (ansi.org). Do not rely on free internet copies, which may be old editions or incomplete reproductions.

Using This Guide in Practice

This page is educational only

This guide explains how ANSI/ASQ Z1.4 works conceptually. It does not reproduce the official tables, switching rules, or exact Ac/Re values needed for production inspection. All production sampling must use the current official standard, available from ASQ.

Statistical foundations

Attribute acceptance sampling is grounded in the binomial distribution and the Poisson distribution. The OC curve for a sampling plan with sample size n and acceptance number Ac is derived from these distributions. Statistics Fundamentals covers both distributions in full.

Complementary resources

The NIST/SEMATECH Engineering Statistics Handbook provides additional OC curve analysis. The ASQ quality resources library covers acceptance sampling applications across industries.