Sampling Methods in Quality Control and Product Inspection

Sampling is widely used in manufacturing quality control and product inspection. It allows manufacturers, quality teams, and third-party inspectors to evaluate a defined population of products or production units by examining a selected number of samples rather than every individual item.

However, the way samples are selected can significantly affect the reliability of inspection results. A sample that does not adequately represent the population may overlook defects, distort quality assessments, or lead to decisions that do not reflect the actual condition of a production lot.

Three commonly used approaches are random sampling, stratified sampling, and systematic sampling. Each has a different selection principle and is suitable for different inspection and quality control situations.

This guide explains how these methods work, where they can be applied, and what to consider when choosing a sampling approach for manufacturing quality control and product inspection.

What Are Sampling Methods in Quality Control and Product Inspection?

A sampling method is the procedure used to select a subset of items from a larger population for examination. The population may consist of finished products, incoming materials, production units, components, or items within a shipment.

Instead of examining every unit, the inspector or quality team selects samples according to a defined approach and evaluates them against established inspection criteria.

Sampling methods help determine which items are selected. They do not, by themselves, define how many items must be inspected or what constitutes an acceptable result. Those decisions depend on the inspection objective, population characteristics, risk level, and applicable sampling plan.

For example, a manufacturer may sample products throughout a production run to monitor process consistency. A third-party inspector may select finished goods from different cartons during a pre-shipment inspection to assess whether a shipment meets the buyer's requirements.

In both cases, the selection method matters because the sample needs to provide meaningful information about the population being evaluated.

Why Does Sampling Method Selection Matter?

The purpose of sampling is not simply to reduce the number of items inspected. It is to obtain useful evidence from a larger population while managing inspection time and resources.

An unsuitable sampling method can introduce selection bias or leave important parts of the population unexamined.

Consider a shipment containing products from several production batches. If samples are taken only from cartons near the entrance of a warehouse, products from other batches or storage locations may receive little or no attention. The inspection results could therefore fail to reflect the condition of the entire shipment.

A suitable sampling method helps to:

• Improve representativeness: Give relevant parts of the population an appropriate opportunity to be included.
• Reduce selection bias: Avoid relying only on convenient or easily accessible items.
• Support consistent decisions: Apply a repeatable selection procedure across inspections or production runs.
• Identify quality variation: Help reveal differences between batches, production lines, shifts, or product groups.
• Use inspection resources effectively: Focus examination on a manageable number of items without selecting samples arbitrarily.

The appropriate method depends on how the population is organized and what the inspection or quality control activity is intended to establish.

Random Sampling

Random sampling is a method in which individual items are selected by chance from a defined population. Each item has a known opportunity of being selected, and in simple random sampling, each item has an equal chance of selection.

The purpose is to reduce deliberate or unconscious selection bias and obtain a sample that can represent the overall population.

How Random Sampling Works

A typical random sampling procedure involves:

1. Define the population to be examined, such as all finished units in a production lot.
2. Assign an identification number to each item or use an existing numbering system.
3. Determine the required sample size based on the applicable inspection plan.
4. Use a random number generator or another suitable random selection procedure.
5. Locate and inspect the selected items.
6. Record the results against the relevant inspection criteria.

For example, if a production lot contains 2,000 units, each unit can be assigned a number from 1 to 2,000. A random number generator can then be used to select the required sample items.

The inspector should follow the selection results rather than replacing inconveniently located items with easier alternatives. If a selected item cannot be accessed, the reason and any replacement procedure should be documented.

Applications in Quality Control and Product Inspection

Random sampling is useful when the population is relatively homogeneous and there is no clear reason to divide it into separate groups before selection.

Common applications include:

• Selecting finished products from a single production lot for pre-shipment inspection.
• Checking incoming materials from a defined delivery.
• Selecting units from a continuous production run for routine quality monitoring.
• Evaluating whether a batch meets specified visual, dimensional, or functional requirements.

For example, during a pre-shipment inspection of 1,000 electric kettles, an inspector may randomly select units from the defined lot and check appearance, assembly, electrical functions, accessories, and packaging according to the agreed inspection checklist.

Advantages

• Reduces the influence of personal judgment during sample selection.
• Provides a clear and repeatable selection procedure.
• Works well for relatively uniform populations.
• Can support statistical analysis when the sampling design and assumptions are appropriate.

Limitations

• Important subgroups may be underrepresented by chance.
• Selecting items randomly can be difficult when products are stored across many locations.
• It may not adequately reveal differences between production batches, shifts, or manufacturing lines.
• A random sample does not automatically guarantee that every type of defect will be detected.

Random sampling is therefore a useful general approach, but it may not be the best choice when the population contains distinct groups with potentially different quality characteristics.

Stratified Sampling

Stratified sampling divides a population into separate subgroups, known as strata, before selecting samples from each group.

The groups are formed according to relevant characteristics, such as production batch, manufacturing line, supplier, product model, size, color, or shift. Samples are then selected from each stratum, often using random sampling within the individual groups.

This method is particularly useful when the population contains identifiable subgroups that may differ in quality or production conditions.

How Stratified Sampling Works

A typical procedure includes:

1. Define the complete population.
2. Identify relevant characteristics that can be used to divide it into meaningful groups.
3. Divide the population into separate, non-overlapping strata.
4. Determine how many samples should be selected from each group.
5. Select the samples within each stratum, often randomly.
6. Inspect the samples and analyze the results by group as well as across the overall population.

There are two common approaches to allocating samples:

• Proportionate allocation: The number of samples selected from each stratum reflects that group's share of the total population.
• Disproportionate allocation: Different numbers of samples are selected from each stratum based on factors such as risk, importance, or the need to examine smaller groups more closely.

The appropriate allocation depends on the purpose of the inspection. Proportionate allocation can provide a balanced overview of the whole population, while disproportionate allocation can give additional attention to higher-risk or particularly important groups.

Example: Sampling Across Production Batches

Suppose a shipment contains 1,000 products from three production batches:

If the inspection plan requires 80 samples and proportionate allocation is used, the sample distribution would be:

This approach ensures that all three batches are represented in the sample.

If Batch C has a history of quality problems, the inspector or quality team may instead choose a disproportionate allocation to examine it more closely. In that case, the sampling results should be interpreted with the allocation method in mind. A sample distribution that differs from the population proportions should not be treated as a simple proportional estimate of overall quality without appropriate analysis.

Applications in Quality Control and Product Inspection

Stratified sampling is suitable when quality may vary across identifiable groups.

Examples include:

• Sampling products from different production lines or shifts.
• Checking different sizes, colors, or models within a product range.
• Selecting samples from multiple production batches during pre-shipment inspection.
• Evaluating incoming materials from different suppliers or delivery lots.
• Comparing defect rates across manufacturing locations or process stages.

For example, a buyer orders 5,000 garments in five colors and six sizes. If the inspection focuses only on the most common color and size, less common variants may not be adequately examined. Stratified sampling can help ensure that the selected sample includes the relevant product variations.

Advantages

• Ensures that defined subgroups are represented.
• Helps identify differences between batches, product variants, or production conditions.
• Allows focused examination of higher-risk groups.
• Can improve the precision of estimates when the strata are appropriately defined and differ meaningfully from one another.

Limitations

• Requires reliable information about the population and its subgroups.
• Grouping decisions can be difficult when production or shipment records are incomplete.
• Disproportionate allocation requires careful interpretation of the results.
• Creating and managing multiple strata may increase inspection planning and coordination work.

Stratified sampling is especially valuable when the overall population is not uniform and differences between groups are relevant to the inspection objective.

Systematic Sampling

Systematic sampling selects items at regular intervals from an ordered population. After choosing a starting point, the inspector or quality team selects every kth item until the required sample size is reached.

Unlike simple random sampling, systematic sampling follows a fixed interval. It can be easier to apply in production environments where products move through a line or are arranged in a sequence.

How Systematic Sampling Works

The sampling interval can be calculated using the following formula:

K = N / n

Where:

• (k) = sampling interval
• (N) = total population size
• (n) = required sample size

For example, if a production run contains 1,000 units and the required sample size is 50: K = 1,000/50 = 20

After selecting a starting point between 1 and 20, the inspector selects every 20th unit.

If the starting point is 7, the selected items would be numbered:

7, 27, 47, 67, 87, and so on.

The starting point is commonly chosen randomly to reduce the risk of consistently selecting items from a particular position in the sequence.

Applications in Quality Control and Product Inspection

Systematic sampling is useful when products are arranged in a continuous or clearly ordered sequence.

Typical applications include:

• Checking units at regular intervals during mass production.
• Monitoring products coming off an assembly line.
• Selecting items from a conveyor belt for routine quality checks.
• Sampling cartons or units arranged in a defined sequence during warehouse inspection.
• Conducting periodic checks across a long production run.

For example, a factory produces 3,000 plastic containers in one production run. A quality team may select a starting unit and then inspect every 30th container to monitor dimensions, appearance, or other defined characteristics throughout production.

During a third-party inspection, systematic selection may also be practical when products are arranged in an accessible sequence. However, the inspector should first confirm that the sequence does not create a selection pattern that could distort the results.

Advantages

• Easy to apply in continuous production or ordered populations.
• Distributes sample selection across a sequence.
• Reduces the need to generate a separate random number for every selected item.
• Can support routine monitoring when the production flow is stable.

Limitations

• Periodic patterns in the population may affect the results.
• A fixed interval may repeatedly select similar items if the sequence follows a repeating production cycle.
• It is less suitable when the population has no clear order.
• It may not provide adequate coverage of distinct batches or product groups unless these are considered separately.

Systematic sampling is convenient, but the sequence of the population should always be reviewed before using a fixed interval.

Comparison of the Three Sampling Methods

The three methods differ mainly in how samples are selected and the situations in which they are most useful.

No single method is suitable for every situation. The selection should reflect the population structure, the inspection objective, and the risks that need to be evaluated.

How to Choose the Right Sampling Method

Choosing a sampling method begins with understanding the population and the purpose of the inspection or quality control activity.

The following considerations can help guide the decision.

1. Understand the Population

Determine what is being sampled and how the population is organized.

Is it one production lot, several batches, different product models, or a continuous production run? If the population contains clearly identifiable groups, stratified sampling may be more appropriate than selecting samples from the entire population without distinction.

2. Identify Potential Quality Variation

Consider whether quality may differ between:

• Production batches.
• Manufacturing lines or shifts.
• Suppliers or component sources.
• Product models, sizes, or colors.
• Different stages of production.

Where these differences matter, the sampling approach should account for them.

3. Define the Inspection Objective

Different activities may require different sampling approaches.

For example:

• Production monitoring: Systematic sampling can help monitor products at regular intervals, while stratified sampling can compare different lines or shifts.
• Pre-shipment inspection: Random sampling can be suitable for a relatively uniform lot, while stratified sampling can help cover multiple batches or product variants.
• Incoming material inspection: Stratified sampling may be useful when a delivery contains materials from different suppliers or batches.
• Focused risk assessment: A targeted approach may be needed when previous inspection results or production records indicate specific areas of concern. Such selection should be distinguished from probability-based sampling.

4. Consider Practical Access and Traceability

The sampling method must be workable in the actual inspection environment.

Inspectors need to identify the population, locate selected items, and record where the samples came from. Clear carton numbers, batch records, product labels, and production records can make the selection process more reliable.

5. Follow the Applicable Sampling Plan

The sampling method and the sampling plan are related, but they are not the same thing.

The method determines how items are selected. The sampling plan establishes details such as sample size, inspection level, acceptance criteria, and the conditions for accepting or rejecting a lot.

These elements should be agreed upon before inspection begins.

Sampling Methods vs. AQL Sampling Plans

Random, stratified, and systematic sampling describe ways of selecting items from a population. AQL-based sampling plans address a different question: how many items should be inspected, and what acceptance and rejection criteria should be applied?

The ISO 2859-1 standard provides sampling procedures for inspection by attributes, including plans indexed by acceptance quality limit (AQL). It is commonly used as a reference for lot-by-lot inspection of products when results are classified into categories such as conforming and nonconforming items. ^[3]

The distinction can be summarized as follows:

For example, an inspector may use an applicable AQL sampling plan to determine the required sample size and acceptance criteria for a shipment. Random sampling can then be used to select the actual units for examination.

The sampling plan and selection method should be applied together in a consistent way. AQL should not be interpreted as a guaranteed defect rate or as a statement that a certain percentage of defective products will always be accepted.

Practical Example: Selecting Samples from a Mixed Production Lot

Consider a buyer who orders 2,400 storage boxes in three sizes from a factory. The products are manufactured in two production batches and packed into separate cartons.

The inspection objective is to assess the finished products before shipment.

A suitable approach may involve the following steps:

Step 1: Define the population

Confirm that the 2,400 units belong to the inspection lot and identify the relevant carton, batch, and product-size information.

Step 2: Review the population structure

The shipment contains different sizes and two production batches. These groups may have different dimensions, assembly requirements, or production conditions.

Step 3: Select a sampling approach

Stratified sampling can help ensure that the relevant sizes and batches are represented. Random selection within each group can then reduce selection bias.

Step 4: Determine the sample size

Use the agreed inspection plan to establish the required sample quantity and acceptance criteria. The sample allocation should reflect the inspection objective and the population structure.

Step 5: Locate and inspect the selected units

Inspect the selected products against the agreed criteria, such as dimensions, material, surface finish, assembly, functionality, labeling, and packaging.

Step 6: Record and evaluate the results

Document defects and findings, including the affected product size, batch, and carton where applicable. Evaluate the results against the agreed inspection plan.

This example shows how a sampling method can be selected and applied within a broader product inspection process. The method helps determine which units are examined, while the inspection criteria and acceptance rules determine how the findings are evaluated.

Common Sampling Mistakes

Even when a suitable sampling method has been selected, errors in implementation can reduce the reliability of the results.

Common mistakes include:

• Selecting only convenient items: Choosing products that are easiest to access may introduce selection bias.
• Ignoring different batches or product variants: A sample drawn from only one group may not represent the entire population.
• Using a fixed interval without checking the sequence: Periodic production patterns may affect systematic sampling.
• Changing selected samples without documentation: Replacing selected items informally can undermine the sampling procedure.
• Confusing sample selection with acceptance criteria: Choosing a sampling method does not establish how many defects are acceptable.
• Failing to record sample origins: Without batch, carton, or production information, it may be difficult to trace and investigate findings.

A documented sampling procedure helps make the selection process more transparent and repeatable.

The Role of Third-Party Inspection in Product Sampling

In third-party product inspection, sample selection is an important part of obtaining reliable evidence about a production lot or shipment.

The inspector should first confirm the inspection scope, product quantity, lot definition, applicable sampling plan, and buyer's requirements. The selection method should then be applied consistently, with relevant product groups and storage conditions taken into account.

For example, during a pre-shipment inspection, random sampling may be appropriate for a uniform shipment, while stratified sampling can help cover different production batches, models, or colors. The selected products can then be checked against the agreed inspection criteria.

At TradeAider, product inspections are carried out according to the agreed service scope and inspection requirements. Depending on the order and inspection stage, services include Pre-Production Inspection (PPI), During Production Inspection (DUPRO), and Pre-Shipment Inspection (PSI). Inspection findings are documented through the reporting system so buyers can review the results and make decisions based on the agreed requirements.

Learn more about TradeAider's third-party quality inspection services.

Frequently Asked Questions

1. What is the difference between random and stratified sampling?

Random sampling selects items by chance from the entire defined population. Stratified sampling first divides the population into relevant subgroups and then selects samples from each group. Stratified sampling is useful when differences between groups may affect the inspection results.

2. When should systematic sampling be used?

Systematic sampling is useful when products are arranged in a clear sequence, such as items moving along a production line or units arranged in a defined order. Before using it, check whether the sequence contains repeating patterns that could affect sample representativeness.

3. Are random, stratified, and systematic sampling all used in product inspection?

Yes. All three methods can be used in manufacturing quality control and product inspection. The appropriate choice depends on the population structure, inspection objective, and applicable sampling requirements.

4. Is random sampling the same as AQL sampling?

No. Random sampling is a method for selecting items. AQL sampling plans establish sample sizes and acceptance or rejection criteria for lot inspection. A random selection procedure may be used to select the samples required by an AQL plan.

5. Can different sampling methods be combined?

Yes. Sampling approaches can be combined when the population or inspection objective requires it. For example, a shipment may first be divided into production batches using stratified sampling, followed by random selection within each batch. The combined procedure should be defined and documented before inspection.

Sources & References

[1] U.S. Environmental Protection Agency (EPA). Selecting a Sampling Design. https://www.epa.gov/quality/selecting-sampling-design

[2] NIST/SEMATECH. e-Handbook of Statistical Methods: Choosing a Sampling Scheme. https://www.itl.nist.gov/div898/handbook/ppc/section3/ppc332.htm

[3] International Organization for Standardization (ISO). ISO 2859-1: Sampling Procedures for Inspection by Attributes — Part 1: Sampling Schemes Indexed by Acceptance Quality Limit (AQL) for Lot-by-Lot Inspection. https://www.iso.org/standard/77582.html