Tag: Non-Statistical Sampling

Audit Sampling means applying audit procedures to less than 100% of the items within a population in such a way that each sampling unit has a chance of being selected. Under SA 530, Audit Sampling, the auditor uses sampling to obtain and evaluate audit evidence about selected characteristics of the population and to draw a reasonable conclusion about the entire population. The population may include invoices, transactions, account balances or other records. The auditor selects a sample based on the audit objective, assessed risks and characteristics of the population. Sampling may be statistical or non statistical. A properly designed sample should be representative of the population and should provide a reasonable basis for conclusions. Audit sampling helps the auditor obtain sufficient appropriate evidence while reducing the time and effort required compared with examining every item. The auditor should also evaluate sampling risk and the results of testing.

Designing an Audit Sample:

1. Determining the Objective of the Test

Before designing an audit sample, the auditor must clearly define the specific objective of the test to be performed, whether it is a test of controls, a substantive test of details, or both combined, as this determines the appropriate sampling approach and the characteristics of the population to be examined. Understanding the objective helps the auditor identify which assertions are being tested, such as completeness, existence, or accuracy, and ensures the sample selected is relevant to addressing the specific risk of material misstatement identified. A clearly defined objective forms the foundation for all subsequent sampling decisions throughout the process.

2. Defining the Population

The population refers to the entire set of data from which the auditor wishes to sample in order to reach a conclusion, and it must be appropriate, complete, and relevant to the specific audit objective being tested. The auditor must ensure the population is defined accurately, for instance, when testing for overstatement of accounts payable, the population might be the complete list of recorded payables rather than potential unrecorded liabilities. Errors in defining the population, such as excluding relevant items or including irrelevant ones, can lead to incorrect conclusions being drawn, even if the sampling methodology itself is technically sound and well-executed.

3. Determining the Sampling Unit and Stratification

The sampling unit refers to the individual items constituting the population, such as individual invoices, ledger entries, or account balances, which the auditor will select and examine. Stratification involves dividing the population into sub-groups with similar characteristics, such as separating high-value transactions from routine ones, allowing auditors to apply different levels of scrutiny to each stratum based on relative risk and materiality. This technique improves audit efficiency by enabling auditors to focus greater sampling effort on higher-risk or higher-value strata while applying lighter testing to lower-risk items, rather than treating the entire population as homogeneous throughout the sampling exercise.

4. Determining Sample Size

Sample size determination involves calculating how many items from the population need to be selected and tested to reduce sampling risk to an acceptably low level, considering factors such as the acceptable level of sampling risk, tolerable misstatement, expected misstatement, and the degree of variability within the population. Larger sample sizes reduce sampling risk but increase audit cost and time, requiring auditors to balance these competing considerations using professional judgment or statistical formulas. Higher assessed risk of material misstatement or lower tolerance for error typically necessitates a larger sample size to obtain sufficient appropriate evidence supporting the auditor’s conclusion.

5. Selecting the Sample Selection Method

Once sample size is determined, auditors must choose an appropriate method for selecting specific items from the population, such as random selection, systematic selection, haphazard selection, or monetary unit sampling, ensuring the method chosen supports the objective of obtaining a representative sample. Random and systematic selection methods are commonly used in statistical sampling to ensure every item has a known chance of selection, enhancing objectivity and reducing selection bias. The chosen method must align with the overall sampling approach, whether statistical or non-statistical, and should be applied consistently to maintain the integrity and defensibility of the sampling process.

Types of Sampling (Approaches to Sampling):

1. Statistical Sampling

Statistical sampling is an approach that uses random selection techniques and probability theory to select sample items, allowing the auditor to measure and quantify sampling risk mathematically. This method requires that every item in the population has a known, non-zero chance of being selected, enabling the auditor to project results from the sample to the entire population with a calculated level of confidence. Statistical sampling provides an objective, defensible basis for conclusions, as the risk of the sample not being representative can be explicitly measured using formulas. It is particularly useful for large, homogeneous populations where consistent, repeatable methodology is valuable, though it requires specialized statistical knowledge and audit software to design, execute, and evaluate results accurately and reliably.

2. Non-Statistical (Judgmental) Sampling

Non-statistical sampling, also called judgmental sampling, relies on the auditor’s professional judgment to determine sample size and select specific items, without using mathematical probability techniques to measure sampling risk formally. Auditors use their knowledge of the client’s business, past experience, and understanding of risk areas to select items they believe are most relevant or representative for testing purposes. While this approach offers flexibility and can be efficient for smaller or less complex populations, it lacks the mathematical rigor of statistical sampling, making it harder to objectively quantify and justify the precision of conclusions drawn. It remains widely used, particularly for smaller audits or specific targeted testing procedures where formal statistical projection is unnecessary.

3. Random Sampling

Random sampling is a statistical selection technique in which every item in the population has an equal and known chance of being selected, typically implemented using random number generators or computer-assisted audit tools. This method eliminates selection bias, ensuring the sample is representative of the entire population and allowing valid statistical projections of results. Random sampling is considered highly objective and defensible, as the selection process is free from auditor influence or unconscious bias toward particular items. It is commonly used when testing large, homogeneous populations such as sales invoices or payment vouchers, where each transaction carries a similar level of risk, making equal probability of selection appropriate and statistically sound for reliable audit conclusions.

4. Systematic Sampling

Systematic sampling involves selecting sample items at uniform, fixed intervals throughout the population after choosing a random starting point, such as selecting every fiftieth invoice from a sequentially numbered population. This method is easier and faster to apply than pure random sampling while still providing reasonable representativeness across the population, provided the population is not arranged in a pattern that coincides with the sampling interval, which could introduce bias. Systematic sampling is particularly practical for populations with sequential numbering, such as cheque registers or invoice listings, as it simplifies the selection process while maintaining a degree of objectivity. Auditors must remain alert to any underlying patterns in the data that could distort representativeness when applying this technique.

5. Monetary Unit Sampling (Value–Weighted Sampling)

Monetary unit sampling, also known as value-weighted or dollar-unit sampling, is a statistical technique where the probability of selecting a particular transaction or item is proportional to its monetary value, treating each individual currency unit as the sampling unit rather than each physical transaction. This approach naturally directs greater audit attention toward higher-value items, which are typically of greater audit significance, while still providing valid statistical coverage of smaller items. Monetary unit sampling is particularly effective for detecting overstatement errors in populations like accounts receivable or inventory, as it inherently emphasizes materiality through its value-weighted selection mechanism, making it a widely favored technique for substantive testing of significant financial statement account balances in modern auditing practice.

Sample Size

Sample size refers to the number of items selected from a population for examination during an audit. It is an important part of audit sampling because the auditor must select enough items to obtain sufficient appropriate audit evidence and reach reasonable conclusions about the population. The appropriate sample size depends on factors such as the auditor’s assessment of audit risk, expected misstatement, tolerable misstatement, population characteristics and the desired level of assurance. A larger sample may be required when the risk of material misstatement is high or when greater assurance is needed. A smaller sample may be appropriate where risks are lower. The auditor should use professional judgement while determining sample size and consider the requirements of SA 530.

Selection of Items for Testing:

1. Random Selection

Random selection is a method where every item in the population has an equal and known probability of being chosen for testing, typically applied using random number generators, computer-assisted audit tools, or random number tables cross-referenced to a numbered population. This technique eliminates auditor bias in item selection and forms the basis for valid statistical sampling, as it allows sample results to be mathematically projected across the entire population with a measurable degree of confidence. Random selection is considered the most objective approach and is widely used for large, homogeneous populations such as sales transactions, payment vouchers, or inventory items, where consistent and unbiased coverage across the dataset is essential for reliable audit conclusions.

2. Systematic Selection

Systematic selection involves selecting items at fixed, uniform intervals across the population, calculated by dividing the total population size by the required sample size to determine the sampling interval, after which a random starting point is chosen within the first interval. This method is quicker and more practical to implement than pure random selection while still achieving broad coverage across the population. However, auditors must be cautious of any hidden patterns or cyclical characteristics in the population that might coincide with the chosen interval, potentially skewing representativeness. It is especially suited to sequentially organized data, such as numbered invoices, cheques, or journal entries, where structured, interval-based selection naturally aligns with the data’s inherent organization.

3. Monetary Unit Sampling

Monetary Unit Sampling selects items based on their monetary value rather than treating each transaction as a single sampling unit, meaning transactions with higher values have a proportionally greater chance of selection. This value-weighted approach directs audit attention naturally toward larger, more material transactions while still providing statistically valid coverage of the population as a whole. It is particularly effective for identifying overstatement errors in accounts like receivables or inventory, since larger balances are inherently more likely to contain material misstatements. This method combines the benefits of statistical rigor with a built-in emphasis on materiality, making it a popular and efficient choice for substantive testing of significant financial statement account balances.

4. Haphazard Selection

Haphazard selection involves the auditor choosing sample items without following any structured or systematic technique, attempting instead to select items without any conscious bias toward particular characteristics, values, or ease of access. While this method may seem representative in practice, it lacks the mathematical objectivity required for valid statistical sampling and cannot support formal statistical projections of results to the broader population. Haphazard selection is more appropriate for use within a non-statistical sampling approach, where the auditor relies primarily on professional judgment to reach conclusions. Auditors must exercise caution to genuinely avoid bias, such as unconsciously favoring easily accessible or clearly organized items over others within the population being tested.

5. Block Selection

Block selection involves choosing a contiguous group or “block” of items from within the population for testing, such as examining all transactions recorded during a specific week or all invoices within a particular sequential number range. While block selection is simple and convenient to apply, it is generally considered the least reliable method, as most populations are not structured in a way that a single block would be representative of the entire population’s characteristics. This method is typically used only for very limited, specific audit purposes, such as testing controls over a particular short period, and is rarely relied upon as the primary technique for drawing broader conclusions about an entire population.

Sample Selection Methods:

1. Random Number Selection

Random number selection uses random number generators, computerized tools, or random number tables to select sample items, ensuring every item in the population has an equal, known probability of being chosen. Each item in the population must first be assigned a unique reference number, allowing the auditor to match generated random numbers to specific items for testing. This method is fundamental to statistical sampling, as it provides the mathematical basis necessary for valid projection of sample results across the entire population with measurable confidence levels. It is widely regarded as the most objective and defensible selection technique, minimizing any risk of conscious or unconscious auditor bias influencing which items are examined.

2. Systematic Interval Selection

Systematic interval selection involves calculating a fixed sampling interval by dividing the population size by the desired sample size, then selecting items at that consistent interval throughout the population after establishing a random starting point. This approach is administratively simpler and faster than random number selection while still achieving reasonably broad and objective coverage across the dataset. The key risk with this method is the possibility of an underlying pattern within the population that coincides with the chosen interval, which could distort the representativeness of the sample selected. It works particularly well for sequentially numbered records such as invoices, cheques, or journal vouchers, where data is naturally organized in a continuous, ordered sequence.

3. Value-Weighted (Monetary Unit) Selection

Value-weighted selection, commonly known as monetary unit sampling, selects items with a probability proportional to their monetary value rather than giving each transaction an equal chance of selection, meaning higher-value items are more likely to be included in the sample. This method inherently directs greater audit scrutiny toward transactions with greater financial significance, making it particularly effective at identifying material overstatement errors within account balances such as receivables or inventory. It combines statistical validity with a natural emphasis on materiality, allowing auditors to efficiently allocate testing effort toward the transactions most likely to contain significant misstatements, while still providing adequate representative coverage of smaller-value items within the broader population.

4. Haphazard Selection

Haphazard selection involves the auditor choosing items from the population without following any structured, mathematical technique, attempting to select items without deliberately favoring or avoiding any particular characteristics. Although intended to mimic randomness, this method inherently lacks the objective, mathematical basis required for statistical sampling, as there is no guarantee that every item genuinely had an equal chance of selection. It is therefore more appropriate within a non-statistical sampling framework, where the auditor relies on professional judgment rather than formal statistical projection to reach conclusions. Auditors using this method must remain vigilant against unconscious bias, such as unintentionally gravitating toward items that are more accessible, better organized, or easier to locate within records.