Chapter 11: Audit Sampling Concepts

Selecting Items to Test and Sample

When designing audit procedures, auditors can select items for testing by:

• selecting all items in a population (100% examination),

• selecting specific items, and/or
• audit sampling.

Audit sampling is selective testing, used when:

1) Less than 100% of the items in the population are being tested, and
2) Each item (described as a sampling unit) in the population could be selected
as part of the sample.
Comparison of Different Selection Approaches

Selection
Approach Explanation
All items Testing 100% of the population is often not efficient. However, it may be
appropriate in some circumstances:
• The population is made up of a small number of large value items that
individually exceed performance materiality.
• When performing tests of details, if the risk of error is expected to be
high, 100% examination or the use of a large sample size may be
appropriate.
• If the auditor plans to use audit data analytics, it may be easier for the
auditor to efficiently test the entire population for certain attributes or
other processes.
Selecting specific This approach is appropriate for:
items • High-value or key items that could individually result in a material
misstatement
• All items over a specified value
• Any unusual or sensitive items or financial statement disclosures
• Any items that are highly susceptible to misstatement
• Items to test control activities (the auditor may use judgment to
select and examine specific items to determine if a particular control
activity is being performed)
Representative • This is appropriate for reaching a conclusion about an entire -
sample of items population (data set) by selecting and examining a representative
from the sample.
population
Stratification of the • Stratification separates the population into distinct groups of items
population with similar characteristics. For example, all large items can be
extracted and tested separately and then a representative sample
could be taken from the remaining items.
Representative Sampling and Its Risks

Sampling represents a key part of the auditor’s risk response. When an auditor
selects a sample from a population, the objective is to select a sample that is
representative.

• Population—a set of accounts or transactions from which the auditor wishes

to draw a conclusion.
• Representative sample—a sample with the same characteristics as those of
the population. (ie: sampled items are similar to items not sampled – we can
never know this for sure unless we tested everything, but that is the goal)

There is a potential risk that an auditor comes to an incorrect conclusion due to an

error in the sampling process. This is called sampling risk and non-sampling risk:

o Sampling risk is the risk that the auditor reaches an inappropriate

conclusion because the sample is not representative of the population.
o Non-sampling risk is the risk that the auditor makes an incorrect
conclusion due to the failure to recognize exceptions, and/or
inappropriate or ineffective audit procedures.

Auditors have two ways of controlling sampling risk:

1. Adjust sample size. (bigger sample size = lower sampling risk)

2. Use an appropriate method of selecting sample items in the population.

Increasing sample size reduces sampling risk, and vice versa. At one extreme, a
sample of all items of a population has zero sampling risk (which, in reality, is not
sampling). At the other extreme, a sample of one or two items has an extremely high
sampling risk.

There are two common causes of non-sampling risk:

1) Auditor not recognizing (or incorrectly interpreting) exceptions (could be due

to not properly understanding the test, boredom, exhaustion)
2) Inappropriate or ineffective audit procedures (or performing the procedures
incorrectly)

Example #1:

You are in the process of testing the existence and valuation of accounts receivable
(AR) during the audit of a large pharmaceutical distribution company. The company
relies on a few key customers who place significant orders throughout the year. You
obtain an aged AR listing from the client and note that just three outstanding
amounts make up over 85% of the total AR balance at year end. The value of the
remaining 15% of the population is below performance materiality.

Determine how you would select items for testing.

Selecting specific items will be the best way to test this population, since just three
significant items make up the majority of the population. In addition, after testing
these three items, the remaining untested balance falls below performance
materiality and is therefore unlikely to contain an error that would be material to the
users of the financial statements.

Example #2:

You are auditing the existence of inventory for a small, privately owned art gallery.
The individual art pieces held by the gallery are of such a high value that it never
holds more than 12 pieces at any one time.

Determine how you would select items for testing.

Selecting all items will be the best way to test this population, since each item in the
population is high value and there are very few items. The small size of the
population makes it feasible for us to test all items, and the high value of each of
the items makes it meaningful to do so.

Example #3:

You are responsible for testing the occurrence of expenses as part of the audit of a
construction company that records revenues based on a percentage of costs
incurred relative to budget. The client has provided you with a detailed listing of
construction costs recorded during the period, and you note that there are
thousands of line items.

Determine how you would select items for testing.

Audit sampling will be the most practical approach to testing this population given
the high volume of transactions. It is unlikely to be possible or practical to test all, or
even a significant proportion, of items in the listing.
Statistical versus Nonstatistical Sampling and Probabilistic versus Non-
probabilistic Sample

Audit sampling methods can be divided into two broad categories – Statistical and
Non-Statistical. Both involve 3 phases:

1. Planning—plan the sample and determine sample size;

2. Performance—select the sample and perform the tests; and
3. Evaluation—evaluate the results and conclude on the acceptability of the
population tested.

• Statistical sampling applies mathematical rules so that auditors can measure

sampling risk in planning the sample and evaluating the results.
• In Non-Statistical sampling, auditors use their professional judgment to select
sample items that will provide the most useful information, and to reach their
conclusions on the populations. (sometimes called “Judgmental sampling”,
meaning using professional judgment rather than statistical methods to select
sample items for audit tests.)

Probabilistic versus Non-probabilistic Sample Selection

• As noted, sample selection may be probabilistic or non-probabilistic.

o In probabilistic sample selection, the auditor picks items at random, and
each item has a set chance of being chosen.
o When using non-probabilistic sample selection, the auditor selects
sample items using professional judgment.

Common Sample Selection Methods

Types of Probabilistic Sample Selection

 • Random sample—a sample in which every possible combination of population
items has an equal chance of being selected (ie: using a random number
generator).
• Systematic selection—the auditor calculates an interval and selects the items
for the sample based on the size of the interval and a randomly selected
number between zero and the sample size. (ie: You have a listing of 100 items,
and need 10 samples, you select every 10th item in a listing)
• Probability-proportionate-to-size sampling—a modified form of physical
attribute sampling that focuses on the individual dollar (or unit of currency,
such as the euro) as the unit of interest, rather than a physical unit. This is
also called simply proportionate to-size sampling, monetary unit sampling, or
dollar unit sampling.

Types of Non-probabilistic Sample Selection

• Directed sample selection—each item in the sample is selected based on

some judgmental criteria (ie: based on risk)
• Block sample selection—auditors select the first item in a block and the
remainder of the block is chosen in sequence. (Ie: sampling blocks of items
that are sequential, such as choosing 100 samples from the third week of
March)
• Haphazard sample selection—items are chosen without regard to their size,
source, or other distinguishing characteristics. (ie: random, but not
statistically random)
• Stratified sampling—a method of sampling in which all the elements in the
total population are divided into two or more subpopulations that are tested
independently. (ie: specifically testing 2 large balances in a population, then
haphazardly sampling the rest)

An Overview of the Audit Sampling Process

• The main differences among tests of controls, substantive tests of

transactions, and tests of details of balances are in what the auditor wants to
measure.

Type of Test What it measures Type of Exceptions

Test of Controls The operating Deviations from Client’s
effectiveness of controls established Controls
 Test of Details The monetary Monetary misstatements
correctness of individual in populations of
transactions OR account transaction data or
balances account balances

What are Exceptions?

Control Testing results are typically a binary (Y/N) – ie: was the control operating
effectively?

Test of Details results are typically a dollar value (quantifying the

misstatement/error) – ie: The invoice differed from the GL listing by $75
Sampling for Tests of Controls

• Steps involved are:

o Determining the test objective (risk addressed and cycle being tested)
 o Define population characteristics (define attribute, population and
sampling unit)
o Determine the initial sample size (depends on tolerable exception rate –
how many errors can there be while still calling the control effective?)
o Select the sample and perform audit procedures
o Evaluate sample results
o Conclude on acceptability of the population (if sample finds control to be
ineffective – do we expand testing (ie: bigger sample size) or find other
ways to test the population? (ie: test of details))
o Communication with the Audit Committee or management

Nonstatistical Sampling for Substantive Tests

• Steps in the audit sampling for tests of details of balances:

o Determine whether audit sampling applies (ie: not sampling if testing the
entire population)
o Determine the test objective (what assertion are we providing assurance
on?)
o Define population characteristics: (misstatement conditions, population
and sampling unit)
o Determine initial sample size
o Select sample and perform audit procedures

• Tolerable misstatement is the application of performance materiality to a

particular sampling procedure.
o The planned sample size increases as the amount of misstatements
expected in the population approaches tolerable misstatement.
• Acceptable risk of incorrect acceptance (ARIA)— the risk that the auditor is
willing to take of accepting a balance as correct when the true misstatement
in the balance is greater than tolerable misstatement.
• Evaluate the Sample Results
o Quantitative evaluation (quantify the results – project the total
misstatements to the population)
o Request client to correct known misstatements
o Qualitative assessment (evaluate the nature and cause of the
misstatement – are any changes to the audit risk model needed? Does it
impact other areas of the FS?)
 • If misstatement is larger than tolerable misstatement, several course of action
to take:
o Take no action until tests of other audit areas are completed (must
evaluate whether FS statements as a whole are materially misstated)
o Perform expanded audit tests in specific areas
o Increase the sample size
o Adjust the account balance
o Refuse to give an unqualified opinion

Audit Data Analytics versus Sampling

• If the auditor plans to use a data analytics tool, it may be easier for the
auditor to efficiently test the entire population for certain attributes or
processes.
• Audit data analytics (ADAs) can be used for tests of controls, substantive
tests, and substantive analytical procedures.

• The auditor may use ADAs when:

o There is data that is relevant for the assertion or attribute that is being
testing (ADA is most suited for testing accuracy and cut-off assertions);
o The data is reliable (if it is from the client, then there are strong controls in
place or if it is from an external source, that source is reliable); and
o The data is accessible and has been scrubbed of errors so it is in a usable
format.

Adequate Documentation and Professional Judgment

• The auditor needs to retain adequate records of the procedures performed,

the methods used to select the sample and perform the tests, the results
found in the tests, and the conclusions reached.
• Documentation is needed for both statistical and nonstatistical sampling to
evaluate the combined results of all tests.