Appendix 2 – Unbiased Methods of Selection for Generalizable Sampling
Auditors must strictly adhere to using random selection to ensure the integrity of sample results and to defend the results during a clearance or fact validation process with the auditees. Random selection tools include random numbers tables, computer-generated numbers, and utilities to generate randomly selected samples. All these options can be used to ensure that sample selection is unbiased.
There is a tendency to consider haphazard or arbitrary selection (i.e., trying to create a true random sample by haphazardly choosing items) to be equally good or good enough to ensure unbiased selection. However, despite the honest intentions of auditors to select a sample that will fairly examine the population, haphazard selection methods are not appropriate for statistical sampling. Haphazard selections can introduce bias into a sample. Factors such as convenience, interest, or prominence do influence selection decisions when not controlled through an objective process.
There are several methods of objectively selecting a sample.
Simple Random Sampling
This involves collecting a sample in a way that ensures every member of the population has an equal chance of selection.
Systematic Random Sampling
After randomly selecting a starting point in the population between 1 and n, every nth unit is selected, where n equals the population size divided by the sample size.
Cluster Sampling
Cluster sampling is a generalizable sampling technique where the population is divided into multiple groups (clusters). Random clusters are selected with a simple random or systematic random sampling technique for data collection and data analysis.
Proportional or Non-proportional Stratified Sampling
The population is subdivided into homogenous groups, for example regions, size, or type of establishment. The strata can have equal sizes or may be based on a higher proportion in certain strata. The specific type of sample selection used depends on the nature of the population, sampling frame, and analytical requirements. For most circumstances in performance auditing, a simple random sample is sufficient. If obtaining audit observations requires a highly complex methodology, whether sampling or otherwise, it can add significantly to the effort required to communicate and defend the results.
LinkedIn
Twitter