# Audit Sampling Methods

## Why Sampling?

It is impracticable for auditors to examine all available information. Audit sampling allows the auditor to draw valid conclusions about a population by examining only a portion of it. The key requirement: the sample must be representative of the population.

## Types of Sample Selection Methods

### 1. Stratified Sampling

The population is divided into subgroups (strata) based on a common characteristic (e.g., value range, geography, type).
Each stratum is treated as a separate population; samples are drawn from each independently.
The number of strata and percentage to sample from each is determined by auditor judgment.
Use when: The population is heterogeneous and proportionate coverage of each subgroup is required.

### 2. Block Sampling

A block of contiguous (consecutive) items is selected from the population.
Limitation: Transactions within a continuous block share similar characteristics. Unusual or different-type transactions occurring outside the block may not be captured, making this an unreliable sole basis for opinion on the entire population.

### 3. Random Sampling

Every item in the population has an equal probability of selection. Core of statistical sampling.

## Statistical vs. Non-Statistical Sampling

Feature	Statistical Sampling	Non-Statistical Sampling
Selection method	Random (unbiased)	Auditor's personal judgment
Evaluation method	Probability theory	Subjective assessment
Personal bias	None	Possible
Projection reliability	High — projectable to population	Less reliable
Best suited for	Large, homogeneous populations	Targeted testing of specific risks

## Why Statistical Sampling is More Scientific

1. Uses mathematical laws of probability to determine appropriate sample size in varying circumstances.

2. No personal bias — selection is purely random.

3. Sample results can be reliably projected to the entire population.

4. Wide application: compliance testing, trade receivable confirmation, payroll checking, vouching of invoices and petty cash vouchers.

5. Always recommended for large organisations with high transaction volumes.

Worked example

### Example 1

Case 1 (Stratified Sampling): An auditor divided trade receivables into three groups — balances above ₹20 lakhs, balances between ₹10–20 lakhs, and balances below ₹10 lakhs — and selected different percentages of items from each group.

Identification: Stratified Sampling

Why: The auditor divided the population into strata based on balance size and drew proportionate samples from each. This ensures adequate coverage of high-value items while still testing lower-value balances, making the sample representative of the full heterogeneous population.

### Example 2

Case 2 (Block Sampling): An auditor selected 50 consecutive cheques to test whether they were signed by authorised signatories, rather than picking 50 individual cheques spread throughout the year.

Identification: Block Sampling

Why: The cheques are selected as a contiguous block from one period. Key limitation: Cheques within a consecutive block tend to share similar characteristics. Unusual signing patterns (or unauthorised signatories) occurring at other times of the year will not be detected, limiting reliability of the conclusion about the full year's population.

### Example 3

Case 3 (Statistical Sampling justification): NP Ltd. has a huge volume of similar transactions. Team member Mr. Q wants to select samples based on personal experience and knowledge instead of using statistical sampling.

Why statistical sampling is preferred here:

With a large homogeneous population, mathematical probability laws determine the correct sample size — Mr. Q's intuition cannot do this reliably.
Mr. Q's personal selection introduces bias; statistical sampling eliminates it.
Results from a statistically drawn sample can be projected to the entire population with a measurable confidence level.
For large organisations, statistical sampling is always recommended as it is unbiased and the selected samples are not prejudged.

⚠️ Common exam mistakes

Confusing Stratified Sampling with Cluster Sampling — in stratified sampling, items are drawn FROM EACH stratum; in cluster sampling, entire clusters are selected as units.
Treating Block Sampling as a reliable method for drawing population-level conclusions — blocks miss unusual transactions occurring outside the selected consecutive period.
Assuming non-statistical sampling is always inferior — it can be appropriate for targeted testing of specific high-risk items.
Forgetting that statistical sampling requires BOTH random selection AND use of probability theory to evaluate results — random selection alone does not make it statistical sampling.

Reference:

Now that you've read this — what's next?

Move from understanding → mastery in 3 clicks. Each option below picks up from this lesson's topic.

🔥 Drill 5 Qs on this → ✍️ Adaptive practice →