Stratified Sampling
Definition
A sampling design that divides the population into homogeneous subgroups (strata) and samples each stratum separately. Allows the auditor to apply higher sampling intensity to high-risk strata and lower intensity to low-risk strata, improving efficiency without sacrificing coverage.
- Field
- Fraud auditing, statistical sampling
- Method
- Divide population into homogeneous strata, sample each separately
- Benefit
- Higher intensity on high-risk strata, lower on low-risk
- Outcome
- Improved efficiency without sacrificing coverage
Common questions
How does an auditor decide what counts as a stratum?+
Strata are usually defined by a risk-relevant attribute such as transaction size, business unit, or vendor category, chosen so that items within a stratum are more similar to each other than to items in other strata. This lets the auditor target scarce testing effort at the segments most likely to contain irregularities.
Is stratified sampling always better than simple random sampling in a fraud audit?+
Not automatically. It only improves efficiency when the chosen strata actually correlate with risk or error rate. If the population is fairly homogeneous or the strata are poorly chosen, the extra design complexity adds cost without improving detection over a simple random sample.
Related terms
- Attribute Sampling
- A statistical sampling method that tests whether each selected item either has or lacks a specified attribute, for example whether a change...
- Cluster Sampling
- A design in which the population is divided into naturally occurring groups called clusters, a random sample of clusters is selected, and...
- Confidence Level
- An explicit label attached to an attribution assessment indicating how strongly the available evidence supports the conclusion. Standard tiers are low, medium,...
- Monetary Unit Sampling (MUS)
- A probability-proportional-to-size method that treats each currency unit in the population as a sampling unit. Larger transactions have a higher probability of...
- Risk-Directed Selection
- A judgmental sampling approach in which items are chosen because they exhibit specific risk indicators: unusual amounts, unusual payees, bypass of normal...
- Sampling Bias
- A systematic distortion in an estimate caused by a selection process that does not give all units a known probability of inclusion....
- Sampling Error
- The difference between a sample statistic and the true population value, arising from the randomness of selection. It can be estimated from...
- Sampling Frame
- The list or description of all units from which the sample can be drawn. Any unit not on the frame cannot be...
- Simple Random Sampling
- A design in which every possible subset of a given size has an equal probability of being selected. It requires a complete...
- Tolerable Deviation Rate (TDR)
- The maximum rate of control deviations the auditor is willing to accept before concluding that a control cannot be relied upon. Setting...
Explained in these topics
- Sampling Strategies and Representative DataA design in which the population is divided into mutually exclusive subgroups called strata, and units are sampled independently from each stratum. It reduces...
- Sampling Techniques in Fraud Audits