SAMPLE SELECTION METHOD

There are a number of techniques that auditors can use to select a sample. However, only
probabilistic techniques are applicable when using statistical sampling. Auditors can also use
probabilistic methods even when using nonstatistical sampling. Probabilistic sample selection
methods include the following:

1. Simple random sample selection

2. Systematic sample selection
3. Probability proportional to size sample selection

Nonprobabilistic sample selection methods include:

1. Haphazard sample selection

2. Block sample selection

Statistical sampling requires a probabilistic sample to measure sampling risk. For

probabilistic samples, the auditor uses no judgment about which sample items are selected,
except in choosing which probabilistic selection method to use.

Simple Random Sample Selection In a simple random sample, every possible combination
of population items has an equal chance of being included in the sample. Auditors use simple
random sampling to sample populations when there is no need to emphasize one or more
types of population items.

When auditors obtain a simple random sample, they must use a method that ensures all
items in the population have an equal chance of selection. Auditors most often generate
random numbers by using one of three computer sample selection techniques: electronic
spreadsheets, random number generators, and generalized audit software.

Random numbers may be obtained with or without replacement. With replacement means
an element in the population can be included in the sample more than once. In selection
without replacement, an item can be included only once. Although both selection approaches
are consistent with sound statistical theory, auditors rarely use replacement sampling.

Systematic Sample Selection In systematic sample selection (also called systematic

sampling), the auditor calculates an interval and then selects the items for the sample based on
the size of the interval. The interval is determined by dividing the population size by the desired
sample size.
 The advantage of systematic selection is its ease of use. In most populations, a systematic
sample can be drawn quickly and the approach automatically puts the numbers in sequence,
making it easy to develop the appropriate documentation.

A concern with systematic selection is the possibility of bias. Because of the way systematic
selection is done, once the first item in the sample is selected, all other items are chosen
automatically. This causes no problem if the characteristic of interest, such as a possible control
deviation, is distributed randomly throughout the population, but this may not always be the
case.

Probability Proportional to Size and Stratified Sample Selection In many auditing

situations, it is advantageous to select samples that emphasize population items with larger
recorded amounts. There are two ways to obtain such samples:

1. Take a sample in which the probability of selecting any individual population item is
proportional to its recorded amount. This method is called sampling with probability
proportional to size (PPS), and it is evaluated using nonstatistical sampling or monetary
unit statistical sampling.
2. Divide the population into subpopulations, usually by dollar size, and take larger
samples from the subpopulations with larger sizes. This is called stratified sampling, and
it is evaluated using nonstatistical sampling or variables statistical sampling.

Nonprobabilistic sample selection methods are those that do not meet the technical
requirements for probabilistic sample selection. Because these methods are not based on
mathematical probabilities, the representativeness of the sample may be difficult to determine.

Haphazard Sample Selection Haphazard sample selection is the selection of items without
any conscious bias by the auditor. In such cases, the auditor selects population items without
regard to their size, source, or other distinguishing characteristics.

The most serious shortcoming of haphazard sample selection is the difficulty of remaining
completely unbiased in the selection. Because of the auditor’s training and unintentional bias,
certain population items are more likely than others to be included in the sample.

Block Sample Selection In block sample selection auditors select the first item in a block,
and the remainder of the block is chosen in sequence. It is ordinarily acceptable to use block
samples only if a reasonable number of blocks is used. If few blocks are used, the probability of
obtaining a nonrepresentative sample is too great, considering the possibility of employee
turnover, changes in the accounting system, and the seasonal nature of many businesses.
 Although haphazard and block sample selection appear to be less logical than other sample
selection methods, they are often useful in situations where the nature of the data makes it
more difficult to use a probabilistic method. For many nonstatistical sampling applications
involving tests of controls and substantive tests of transactions, auditors prefer to use a
probabilistic sample selection method to increase the likelihood of selecting a representative
sample.

SAMPLING FOR EXCEPTION RATES

Auditors use sampling for tests of controls and substantive tests of transactions to
determine whether controls are operating effectively and whether the rate of monetary errors
is below tolerable limits. To do this, auditors estimate the percent of items in a population
containing a characteristic or attribute of interest. This percent is called the occurrence rate or
exception rate.

Auditors are interested in the following types of exceptions in populations of accounting

data:

1. Deviations from the client’s established controls

2. Monetary misstatements in populations of transaction data
3. Monetary misstatements in populations of account balance details

Knowing the exception rate is particularly helpful for the first two types of exceptions,
which involve transactions. Therefore, auditors make extensive use of audit sampling that
measures the exception rate in doing tests of controls and substantive tests of transactions.
With the third type of exception, auditors usually need to estimate the total dollar amount of
the exceptions because they must decide whether the misstatements are material. When
auditors want to know the total amount of a misstatement, they use methods that measure
dollars, not the exception rate.

The exception rate in a sample is used to estimate the exception rate in the entire
population, meaning it is the auditor’s “best estimate” of the population exception rate. The
term exception should be understood to refer to both deviations from the client’s control
procedures and amounts that are not monetarily correct, whether because of an unintentional
accounting error or any other cause. The term deviation refers specifically to a departure from
prescribed controls.

Because the exception rate is based on a sample, there is a significant likelihood that the
sample exception rate differs from the actual population exception rate. This difference is
called the sampling error. The auditor is concerned with both the estimate of the sampling
error and the reliability of that estimate, called sampling risk. Assume the auditor determines a
3 percent sample exception rate, and a sampling error of 1 percent, with a sampling risk of 10
percent. The auditor can state that the interval estimate of the population exception rate is
between 2 percent and 4 percent (3 percent ± 1) with a 10 percent risk of being wrong (and a
90 percent chance of being right).

In using audit sampling for exception rates, the auditor wants to know the most the
exception rate is likely to be, rather than the width of the confidence interval. So, the auditor
focuses on the upper limit of the interval estimate, which is called the estimated or Computed
Upper Exception Rate (CUER) in tests of controls and substantive tests of transactions. Using
figures from the preceding example, an auditor might conclude that the CUER for missing
shipping documents is 4 percent at a 5 percent sampling risk, meaning the auditor concludes
that the exception rate in the population is no greater than 4 percent with a 5 percent risk of
the exception rate exceeding 4 percent. Once it is calculated, the auditor can consider CUER in
the context of specific audit objectives. If testing for missing shipping documents, for example,
the auditor must determine whether a 4 percent exception rate indicates an acceptable control
risk for the occurrence objective.