Chapter 9 Audit Sampling
Chapter 9 Audit Sampling
Chapter 9 Audit Sampling
AUDIT SAMPLING
The professional standards board require the auditor to obtain sufficient appropriate evidence to be able to
draw reasonable conclusions on which to audit opinion. In forming such an opinion, the auditor does not
normally examine all evidence available. Auditors usually draw conclusions about the account balance or
transaction class by examining only a sample of evidence.
“The application of audit procedures to less than 100% of the items within an account balance or class
transactions such that all sampling units have a chance selection”
Audit sampling is performed on the assumption that the sample selected for testing is representative of
the population. Thus, an inference or conclusion can be drawn about the characteristics of the population
based on the sample results.
Not all testing procedures performed by the auditors involve audit sampling. For example, the auditor
may decide that it would be more appropriate to examine the entire population (100% examination) of
items that make up an account balance since the population constitutes a small number of large items.
Likewise, the auditor may decide to apply audit procedures only to those items which have particular
significance (selective testing) (e.g. all items over a certain amount).
Regardless of the approach used, the auditor needs to be satisfied that sufficient appropriate evidence is
obtained to meet the objectives of the test
Risk Sampling
When performing audit procedures, the auditor is faced with an uncertainty of not detecting material
errors in an account balance or class of transactions. This uncertainty arises because of sampling and
non-sampling risks.
• Sampling Risk
Sampling risk refers to the possibility that the auditor’s conclusion, based on a sample may be
different from the conclusion reached if the entire population were subjected to the same audit
procedures.
This exists because the sample selected for testing may not be truly representative of a
population.
There are two types of sampling risk that could adversely affect the audit. The alpha risk and the
beta risk.
A. Alpha Risk is the risk the auditor will conclude,
a) In the case of tests of control, that internal control is not reliable when in fact it is
effective and can be relied upon (risk of under reliance); or
b) In the case of substantive test, that material misstatement exist in an account balance or
transaction class when in fact such misstatement does not exist (risk of incorrect
rejection)
This type of sampling risk results in auditor performing audit procedures more than what is
necessary, thus affecting audit efficiency
a) In the case of tests of control, that internal control is not reliable when in fact it is not
effective and cannot be relied upon (risk of overreliance); or
b) In the case of substantive test, that material misstatement does not exist in when in fact
material misstatement does exist (risk of incorrect acceptance)
This type of sampling risk results in an auditor performing audit procedures less that what is
necessary thereby affecting the auditor’s ability to detect material misstatements in the
financial misstatements in the financial statements.
Non-sampling risk
Non-sampling risk refers to the risk that the auditor may draw incorrect conclusions about the
account balance or class of transactions because of human errors such as, application of
inappropriate audit procedures, failure to recognize errors in the sample tested, and
misinterpretation of evidence obtained this includes all aspects of audit risk that are not due
to sampling.
The only way to eliminate sampling risk is to examine the entire population. Doing this, would
not be feasible because of time constraints. Accordingly, auditors do not normally attempt to
eliminate sampling risk. Instead, the auditors control sampling risk by
By increasing the sample size and carefully selecting the sample, the sample becomes more
representative of the population, thus decreasing the sampling risk.
Non-sampling risk, on the other hand, is something that cannot be eliminated even if the auditor
examines the entire population. This risk however, can be minimized by
Both statistical and non-statistical sampling methods are acceptable. Botha approaches will
require the use of auditor’s judgment in designing and selecting the sample, in performing audit
procedures and in evaluating the results. Most of all, both approaches cannot assure the sample
will be representative of the population.
The only difference between the two methods is that statistical sampling allows the auditor to
measure or quantify sampling risks by using mathematical formula. Thus, statistical sampling
helps the auditor to
However, the benefits cannot be obtained without additional costs of training audit staff,
designing plans and selecting items for examination.
❖ Attribute Sampling
➢ This is a sampling plan used to estimate numerical measurement of a population such as peso
value.
➢ It is generally used in performing substantive tests to estimate the amount of
misstatements in the financial statements.
Audit procedures carried out by means of sampling techniques require consideration of at least
the following basic steps:
The audit objective largely determines the audit procedures to be applied. Hence, before deciding
on the nature of the audit procedures to be performed, the auditor must define the specific
objective of the test.
For example, when auditing accounts receivable, the auditor’s objective could be to
determine whether accounts receivable balances exists as of the balance sheet date.
After defining the audit objective, the next step is to determine the specific audit procedure that
will be performed to satisfy the objective. This step also involves defining the population and the
characteristics to be tested.
Using the existence of accounts receivable as an objective, the audit procedure may
involve, sending confirmation letter to customers.
The population would be the customers’ account balances as of the balance sheet and the
characteristic to be tested would be the monetary amount of misstatement in an account balance.
once the auditor has decided to apply certain audit procedure to a sample items in a population,
the auditor must decide how many sampling units to include in the sample.
The auditor may decide to examine only 100 out of the total 5,000 customers’ accounts in
order to draw a conclusion whether the total accounts receivable is actually existing as of the
balance sheet date.
➢ When statistical sampling is used, the auditor determines the sample size using
statistically based formula.
➢ With non-statistical sampling, the sample size is determined but relying primarily on the
auditor’s professional judgment.
Another problem the auditor faces after determining the sample size is the method of selecting
the sample form the total population. A sample selection technique must be designed in such a
way that all items in the population will have an opportunity to be selected. Statistical sampling
requires the sample items be selected at a random so that each sampling unit has known chance of
being selected.
In selecting the customers to whom notification letters will be sent, the auditor may use a
computer software that produces random numbers that match with the customer numbering
system.
The auditor sends confirmation letter to the customers selected to determine the accuracy
of the recorded account balances.
Once the audit procedures have been performed on all sample items, the sample results must be
evaluated to determine whether sufficient evidence has been obtained to satisfy the objective.
The auditor will have to summarize the customers’ confirmation replies and decide
whether the account balance are materially misstated or whether additional audit procedures
need to be performed.
It is to be emphasized that steps (1) defining the objective of the test (2) determining audit
procedures, (5) applying the procedures, and (6) evaluation of results will be performed whether
the auditor uses sampling or not.
Hence, the difference between audit sampling and 100% examination is that audit involves:
the remaining sections of this chapter focus primarily on these three sampling-related steps when
performing tests of control and substantive tests.
Audit sampling for tests of control is generally appropriate when application of control leaves
evidence of performance. For those controls that leave no documentary evidence of performance,
non-sampling procedures such as inquiries and observation would be more appropriate.
➢ Determination of sample size
There are three factors affecting the determination of sample size for test of controls.
These are the:
Sampling risk is inherent in an audit sampling application. A sample drawn can only
be expected to be representative of the population. The size of the sample is affected
by the level of sampling risk the auditor is willing to accept.
There is an inverse relationship between the acceptable sampling risk and sample
size. The smaller the sampling risk the auditor is willing to accept, the larger the
sample size to be (and vice versa)
Tolerable deviation rate is the maximum rate of deviation the auditor is willing
accept, without modifying the planned degree of reliance on the internal control.
The tolerable deviation rate is inversely related to the sample size. Therefore, a
decrease in the tolerable deviation rate will cause the sample size to increase.
Establishing tolerable deviation rate and the acceptable sampling risks requires
professional judgment and involves consideration of
Consequently, if the auditor wants to place more reliance on the control (lower
assessment of control risk) the auditor should justify that assessment by accepting
lower rate of deviation and a smaller sampling risk.
Expected deviation rate is the rate of deviations the auditor expects to find in the
population before testing begins. The auditor can develop this exception based on the
prior year’s results or by examining few items in the population (pilot example)
The expected deviation rate has a direct effect on the sample size; the larger the
expected population deviation rate, the larger would be the sample size.
The expected population deviation rate should not exceed the tolerable deviation rate.
If prior to testing, the auditor anticipates that the actual population deviation rate is
higher than the tolerable deviation rate, the auditor generally omits testing of that
control procedures and either seeks to obtain assurance by testing other relevant
internal control policies and procedures, or assess control risk at a high level.
b) Systematic selection
Another method of selecting sample that is easier to use compared to random
number selection is the systematic selection. This method involves determining a
constant sampling interval and then selects the sample based on the size of that
interval.
To illustrate the use of systematic selection, assume that the auditor wants to
examine a sample of 100 invoices from the population of 20,000 invoices. The
sampling interval is computed by dividing the population size by the sample size.
Hence, the sampling interval would be;
When using the systematic selection, the auditor should determine that the
population is not structure in such a manner that the sampling interval
corresponds with a particular pattern in the population.
c) Haphazard selection
When using this method, the sample is selected without following an organized
structure technique. Haphazard selection is useful for non-statistical sampling,
but it is not used for statistical sampling because the auditor cannot measure the
probability of an item being selected when using this method.
In selecting the sample and applying the appropriate audit procedures, the auditor may
encounter the following situations.
Evaluation of results
When evaluating sample results, both the qualitative and the quantitative factors of
deviations should be considered. Here are some general guidelines that may be used
when evaluating sample results for test of controls.
For example, if the auditor found 4 deviations out of the 200 sample items, the
sample deviation rate is
4 / 200 = 2%
This sample deviation rate represents the auditor’s best estimate of the deviation rate
in the population.
2. Compare the sample deviation rate with the tolerable deviation rate and draw
an overall conclusion about the population.
The next step is to compare the sample deviations rate with the tolerable deviation
rate. This comparison may result to the following situations.
▪ The sample deviation rate is less than the tolerable deviation rate.
If the sample deviation rate is less than the tolerable rate, the auditor should
consider the allowance for sampling risk that is, the possibility that these
sample results could have occurred even if the actual population deviation rate is
higher than the tolerable rate.
Hence, before making a conclusion, the auditor should determine how close the
sample deviation rate is to the tolerable deviation rate. As the sample deviation
rate approaches the tolerable deviation rate the allowance for sampling risk
decreases.
b) If the sample deviation rate is barely lower than the tolerable deviation rate
(for example, 8% as compared to tolerable rate of 10%), there is a high
possibility that the actual deviation rate will exceed tolerable rate.
The auditor then compares the maximum deviation rate with the
tolerable rate to evaluate the results.
If the maximum deviation rate exceeds the tolerable rate, the auditor
can conclude that the sample results do not support the auditor’s
preliminary assessment of control risk and therefore the scope of
substantive tests should be increased.
Under this method, the auditor does not use fixed sample size. It is sometimes called
stop-or-go sampling because after testing the sample, the auditor makes a decision
of whether to stop or go with the sampling plan.
For example, if no deviations are found in the sample, the auditor may conclude that
the internal control procedure is reliable and therefore may stop sampling plan. Also,
if the auditor observes many deviations, the auditor may also terminate the sampling
plan and conclude that the planned degree of reliance on internal control is not
justified.
On the other hand, if one or few deviations are found, the auditor may decide to go
with the examination of another set of samples in an attempt to obtain more evidence
to support the planned assessed level of control risk.
Discovery Sampling
This form of attribute sampling is most appropriate when no deviations are
expected in the population and therefore even one deviation would cause concern.
This is normally used when the auditor suspects that an irregularity might have been
committed. Under this method, the auditor determines a sample size sufficient to
discover at least one deviation to confirm whether an irregularity has occurred.
As mentioned in the preceding chapter, substantive tests are concerned with the amounts reported
in the financial statements and are of two types: analytical procedures and test of details of
transactions and balances. Analytical procedures that involve comparison of the financial
statements with the auditor’s expectations do not involve sampling. Audit sampling is appropriate
when performing test of details to estimate the amount of misstatements in the financial
statements.
There is an inverse relationship between the acceptable sampling risk and the sample
size: the lower risk the auditor accepts, the larger the sample size must be.
• Tolerable misstatement
Tolerable misstatement is the maximum amount of misstatement the auditor will permit
in the population and still be willing to conclude that the account balance is fairly stated.
This is determined in the planning stage of audit and it is related to the auditor’s
preliminary estimate of materiality.
There is an inverse relationship between the tolerable misstatement and the sample size.
A smaller measure of tolerable misstatement will cause the sample size to increase.
• Expected misstatement
Expected misstatement is the amount of misstatement that the auditor believes exists I the
population. Like the expected amount of misstatements may be determined based on the
results of prior year’s substantive tests or pilot sample. As the expected misstatement
draws near the tolerable misstatement, the auditor needs more precise information from
the sample.
There is a direct relationship between the expected misstatement and the sample size. An
increase in the amount of misstatement that the auditor expects to be present in the
population will cause the sample size to increase.
• Stratified sampling
In addition, the auditor may divide or stratify the population into meaningful groups in
order to decrease the effect of variance within the population. When selecting stratified
sample, the sample size should be determined for each stratum and selected from the
stratum. For example, the customers’ accounts may be stratified as follows:
a) It decrease the effect of variance in the population and as a result, decreases the
sample size, and
b) It allows the auditor to give more emphasis to those items with higher monetary
value.
Under value weighted selection, each peso is treated as one sampling unit. That is why,
value weighted sampling also called monetary unit sampling unit. For example, if the
auditor plans to select one customer form a total accounts receivable balance of
P1,000,000, a customer’s account balance of P100,000 will have a 10%
(P100,000/P1,000,000) probability being selected.
This method of sample selection is similar to stratified sampling in that large monetary
values are given greater representation in the sample and it is appropriate when the
auditor anticipates overstatement errors because the greater the overstatement in an
account, the greater the probability that the account will be selected and the
overstatements detected.
a) Ratio estimation; or
b) Difference estimation
To illustrate assume that the auditor sent out confirmation requests to verify the validity
of the accounts receivable. The following data are given:
In terms of Amount of
In terms of number of misstatements
Value customers found
Population P10,000,000 200 ?
Sample size P1,000,000 24 P48,000
The only difference between the two methods is the ratio estimation uses the book values
of the population size and sample size to project the misstatements, while difference
estimation uses the number of customers to project the misstatements to the population.
Hence, the use of ratio estimation is appropriate when the amount of misstatements found
is approximately proportional to the client’s book amount.
After projecting the sample misstatements, to the population, the auditor will have to
compare the projected misstatement with the tolerable misstatement.
If the projected misstatement is greater than the tolerable misstatement, the auditor
will conclude that the account balance is materially misstated. In this case, the
auditor may:
However, if the projected misstatement is less than the tolerable misstatement, the
auditor should consider allowance for sampling risk. The auditor should recognize
that the sampling risk increases as the projected misstatement approaches the
tolerable misstatement.
In some instances, the auditor may encounter anomalous errors. These errors or
misstatement that arise from an isolated event that has not recurred other than
specifically identifiable occasions and are therefore not representative of errors in the
population. Such errors should be excluded when projecting sample errors to the
population. However, the effect of such errors must be considered together with
projected errors in order to determine the combined effect of the errors on the
account balance or transaction class.
A summary of the essential audit sampling steps is presented on the next page.
Determine the Determine the appropriate audit Determine the appropriated audit
procedures to be procedures to be performed to satisfy procedures to be performs to satisfy
performed the objective. the objective.
Define the population and the Define the population and its
conditions that constitute a deviation. characteristics.
Consider the effects of the following Consider the effects of the following
Determine the
factors in determining the sample factors in determining the sample
sample size.
size: size:
• Acceptable sampling risk • Acceptable sampling risk
(Inverse) (Inverse)
• Tolerable deviation rate • Tolerable misstatement
(Inverse) (Inverse)
• Expected misstatement and
• Expected population population variation
deviation rate (Direct)
(Direct)
Use any of the following techniques
Select the sample Use any of the following techniques:
and stratify the population:
• Random number selection Random number selection
• Systematic selection Systematic selection
• Haphazard selection (applies • Haphazard selection
only in non-statistical (applies only in non-
sampling) statistical
sampling)
• Value weighted selection
Apply the audit Apply the audit procedures to the Apply the audit procedures to the
procedures sample items. sample items.
Evaluate the Decide whether the results supported Decide whether to accept account
sample results the planned degree of reliance on balance as fairly stated or to require
internal control further actions