chapter 11

Evidence and sampling

Contents

Introduction
Examination context
Topic List
1 Evidence
2 Selecting items to test
3 Drawing conclusions from sampling
Summary and Self-test
Technical reference
Answers to Self-test
Answers to Interactive questions

© The Institute of Chartered Accountants in England and Wales, March 185

2009
 Assurance

Introduction

Learning objectives Tick off

 Understand the procedures for obtaining evidence

 Identify when tests of controls and substantive procedures will be used

 Recognise the strengths and weaknesses of particular forms of evidence

 Understand how much evidence to obtain

 Recognise when sufficient appropriate evidence has been obtained such that a conclusion can
be drawn

Specific syllabus references for this chapter are: 3b, c, d, e, g.

Practical significance
Evidence is what the assurance conclusion is based on. Therefore, in practice, knowing what tests to carry
out and how many items to test is an important practical skill. Getting it wrong may cost the firm money, if
too much work is done, or leave the firm exposed to negligence claims, if insufficient work is done. It is an
area where considerable judgement will have to be exercised.

Stop and think

How do you think an auditor determines how much evidence to obtain?

Working context
If you are training in practice, you are likely to be involved in obtaining evidence, but less likely to be
involved in the more judgemental areas of how to obtain it and how much of it to obtain. However, as you
progress with your training, you will be given opportunity to make judgements in this area. You will be
expected to draw conclusions from the audit work you have undertaken from an early stage. It is important
therefore to understand why you are carrying out a particular test and what you intend to achieve by it.

Syllabus links
In Audit and Assurance you will focus on the drawing conclusions part of evidence, based on the collection
of evidence that we focus on in this Assurance manual.

186 © The Institute of Chartered Accountants in England and Wales, March 2009
 EVIDENCE AND SAMPLING 11

Examination context

Exam requirements
This is a very important part of your syllabus and the issues discussed here and previously in Chapter

4
underpin the following two chapters as well. You can expect a number of practical and theoretical
questions
in the assessment covering audit evidence.
In the assessment, candidates may be required to:

 Distinguish between tests of controls and substantive procedures

 Identify and evaluate procedures used to obtain evidence

 Select the appropriate means of obtaining evidence in a given scenario

 Determine whether an audit test would be a test for overstatement or understatement in a given
scenario

 Identify factors affecting sample sizes

 Identify weaknesses in methods of obtaining evidence

© The Institute of Chartered Accountants in England and Wales, March 187

2009
 Assurance

1 Evidence

Section overview
 Evidence must be sufficient and appropriate.
 Evidence is obtained in the form of substantive procedures and/or tests of controls.
 Evidence can be obtained by inspection, observation, inquiry and confirmation, recalculation,
reperformance and analytical procedures.

 Substantive procedures will test for evidence of understatement or overstatement of account

balances.

1.1 Overview of evidence from Chapter 4

You studied the basic principles of evidence in Chapter 4. These are the key points:

Evidence includes all the information contained within the accounting records underlying the financial
statements, and other information gathered by the assurance providers, such as confirmations from third
parties. Evidence is obtained in relation to the financial statement assertions which were set out in Chapter
4. There are two types of test; tests of controls (which we have looked at in detail in Chapters 5 to 9) and
substantive procedures (which we will look at in more detail in Chapters 12 and 13).
BSA 500 states that evidence must be sufficient and appropriate.
 Sufficiency is the measure of the quantity of audit evidence.
 Appropriateness is the measure of the quality or reliability of the audit evidence.
We will look at the quantity of evidence obtained in section 2 below.

There are some general principles relating to the quality of evidence which were set out in Chapter 4.

Quality of evidence

External Evidence from external sources is more reliable than that obtained from the entity's
records

Auditor Evidence obtained directly by assurance providers is more reliable than that
obtained indirectly or by inference

Entity Evidence obtained from the entity's records is more reliable when related control
systems operate effectively

Written Evidence in the form of documents (paper or electronic) or written

representations are more reliable than oral representations

Originals Original documents are more reliable than photocopies, or facsimiles

188 © The Institute of Chartered Accountants in England and Wales, March 2009
 EVIDENCE AND SAMPLING 11

1.2 Procedures to obtain evidence

Assurance providers obtain evidence by one or more of the following procedures outlined in BSA 500.
client
Procedures Explanation
management or staff or external sources and
Inspection of Inspection (physical examination) of evaluating responses.
tangible assets
tangible
assets that are recorded in the accounting
records confirms existence, but does not
confirm rights and obligations or valuation.
For example, machinery recorded in asset
register can be inspected by assurance
providers.
Confirmation that assets seen are recorded

in
accounting records gives evidence of
completeness. However, this is limited to
assets assurance providers can see – if
Inspection of
assets
documentation
have been taken off site (hidden) they
might
not be picked up.
Inspection of documents involves

examining
records or documents, for example, looking
at a sales contract or a share certificate.
What inspection of documents achieves
depends on the nature of the document.
For
example, looking at a share certificate
gives
evidence of the existence of the
investment.
Looking at source documents (e.g. sales
invoices) and tracing to financial
statements
gives evidence of completeness (e.g. of
revenue).
Observation Inspection also provides evidence of

valuation
(for example, a purchase invoice gives
evidence of the cost of inventory), rights
and
Inquiry obligations (for example, a hire purchase
agreement gives evidence in relation to
ownership of non-current assets) and the
nature of items (presentation and
disclosure).
It can also be used to compare documents
(and hence test consistency of audit
evidence)
and confirm authorisation.
This involves watching a procedure being
performed (for example, post opening).

This involves seeking information from

 to give evidence. For instance,
Strengths and weaknesses
inspection of a purchase invoice
Inspection of assets is a good gives better quality evidence than
procedure, particularly in the case inspection of sales invoice, because This procedure is relatively weak,
of assets that the entity could not a purchase invoice is created by a as it only confirms that the
function without (for example its third party. procedure is being performed
production plant), but the weakness correctly when the assurance
associated with inspection is that provider is watching.
assets not used in daily production The strength or weakness of this
could be hidden from the assurance procedure will depend on of whom
providers and not included in the inquiry is being made – a
financial statements. member of client staff could
misrepresent matters to the
assurance provider if they
misunderstand the nature of the
question, or they are seeking to
The strength of this procedure conceal an error or fraud.
depends on what is being inspected

© The Institute of Chartered Accountants in England and Wales, March 189

2009
 Assurance

Procedures Explanation Strengths and weaknesses

Confirmation This involves seeking confirmation from a This can be a very strong
(a particular third party, e.g. confirmation from bank of procedure but there may be
form of bank balances. instances where the third party is
inquiry) motivated to misrepresent, for
example an understated receivables
balance might be confirmed because
it favoured the customer.

Recalculation Checking mathematical accuracy of client's Recalculation is evidence created by

records, for example, adding up ledger the assurance provider so is strong
accounts. evidence.

Reperformance Independently executing procedures or Again, the fact that the assurance
controls, either manually or through the provider carries out the
use performance of a control himself
of computer assisted audit techniques makes it strong evidence.
(covered below).
Analytical Evidence here is limited by the
procedures Evaluating and comparing financial and/or strength or weakness of the
non-financial data for plausible underlying accounting system.
relationships
and investigating unexpected fluctuations.
However, this can be a strong
procedure if comparison is made to
items that do not rely on the same
accounting system or that the
assurance provider can corroborate
outside the accounting system.

Often these procedures will be used in conjunction with one another to provide a greater quality of
evidence. For example, an assurance provider might observe controls in operation and then reperform the
control himself to confirm that it operates as he has observed. Auditors will gather detailed evidence but
other assurance providers may need less evidence.

1.3 Computer assisted audit techniques

With so many accounting systems now held on computer, the assurance provider may wish to make use of
computer assisted audit techniques (CAATs). These have been mentioned before in your Study Manual,
particularly in Chapter 5. There are two main types of CAAT that can be used:
 Test data
 Audit software

1.3.1 Test data

Under this test of control, the assurance provider supervises the process of running data through the
client’s system. The stages in the use of test data are as follows:
 Note controls in client’s system
 Decide upon test data, the options include:
– Dummy data (the assurance provider must be very careful to reverse all effects)
– Real data (the data may not contain all the errors necessary to test the controls rigorously)
– Dummy data against a verified copy of the client’s system (much safer)
 Run the test data
 Compare results with those expected
 Conclude on whether controls are operating properly

190 © The Institute of Chartered Accountants in England and Wales, March 2009
 EVIDENCE AND SAMPLING 11

Worked example: Test data

Test data makes use of the client’s own system. To carry out such a test the assurance provider

identifies a
control (or series of controls) in the client’s system. The assurance provider then predicts the system’s
reaction to the test data. For example:
 An invoice which does not cast should be rejected when entered in the system
 An invoice with an invalid supplier code should be rejected
 Dates outside the current year should be rejected
 Valid data should be posted to the correct account
The assurance provider then runs the test data through the client’s system (or a copy thereof) and
compares the results with those expected. The results tell the assurance provider whether the controls
within the system are operating correctly; the test is therefore a test of control.

1.3.2 Audit software

Audit software makes use of the assurance providers’ own specialised software. There are a number of off-
the-shelf packages available, or the assurance provider could have a tailor-made system. Audit software
works on the basis of interrogating the client’s system and extracting and analysing information. It can
therefore carry out a whole range of substantive procedures, across all sorts of different data.
Examples of what audit software can do include:
 Extract a sample according to specified criteria:
– Random
– Over a certain amount
– Below a certain amount
– At certain dates
 Calculate ratios and select those outside set criteria (e.g. more than five per cent different from last
year)

 Check calculations and casts performed by the system

 Prepare reports (e.g. comparison of actual against budgeted figures)
 Follow items through a system and flag where they are posted
The procedures listed above are mostly substantive procedures, because they are substantiating the figures
in the accounts. To generate more procedures that can be done using audit software, just think of the
substantive tests that you may wish to carry out, and consider whether the information is held on the
client’s computer (you can normally assume that it is). If the test does not require judgement, then it can
almost certainly be carried out by audit software.

1.4 Analytical procedures

BSA 520 Analytical Procedures gives more detail on the use of analytical procedures as substantive
procedures (optional) and at the overall review stage (compulsory) of an audit. We covered the use of
analytical procedures in planning (compulsory) in Chapter 3. This BSA applies to audits only, but all
assurance providers may be able to use analytical procedures (indeed, they will be an important tool where
less detailed evidence is required) and will need to consider the same general principles.
The BSA states that auditors must decide whether using available analytical procedures as substantive
procedures will be effective and efficient in reducing detection risk for specific financial statement
assertions. Auditors may efficiently use analytical data produced by the entity itself, provided they are
satisfied that it has been properly prepared.

© The Institute of Chartered Accountants in England and Wales, March 191

2009
 Assurance

There are a number of factors that the auditors should consider when using analytical procedures as
substantive procedures:
 Objective of the analytical procedures (for example analytical procedures may be good at indicating
whether a population is complete)

 Suitability of analytical procedures

 Reliability of analytical procedures

Suitability factors Example

Materiality of the items When inventory balances are material, assurance providers do not
involved and the assessment solely rely on analytical procedures.
of inherent and control risk
If internal controls over sales order processing are weak, and control
risk is high, assurance providers may rely more on tests of individual
transactions or balances than analytical procedures.

Other audit procedures Other procedures assurance providers undertake in reviewing the
directed towards the same collectability of accounts receivable, such as the review of subsequent
financial statement cash receipts, may confirm or dispel questions arising from the application
assertions of analytical procedures to an aged profile of customers' accounts

Example
Reliability factors
Analytical procedures may be more effective when applied to financial
The degree to which information on individual sections of an operation
information can be analysed
Financial: budgets or forecasts
The availability of
Non-financial: e.g. the number of units produced or sold
information
Assurance providers normally expect greater consistency in comparing
The accuracy with which the the relationship of gross profit to sales from one period to another than
expected results of analytical in comparing discretionary expenses, such as research or advertising
procedures can be predicted
A pattern repeated monthly as opposed to annually might be more
The frequency with which a likely to identify a material misstatement
relationship is observed
Whether budgets are established as results to be expected rather than
The relevance of the as goals to be achieved
information available
Independent sources are generally more reliable than internal sources,
The source of the and internally, sources independent of the accounting system are more
information available reliable than those which are not
Broad industry data may need to be supplemented to be comparable
The comparability of the with that of an entity that produces and sells specialised products
information available
Whether budgets are prepared with sufficient care and whether
The reliability of the controls exist over their preparation, review and maintenance
information available
The effectiveness of the accounting and internal control systems
The knowledge gained
The types of problems giving rise to accounting adjustments in prior
during previous audits
periods

Auditors will also consider the plausibility and predictability of the relationships being tested. Some business
relationships are strong, for example between selling expenses and sales in businesses where the sales force
is mainly paid by commission.
When analytical procedures identify significant fluctuations or relationships that are inconsistent with other
relevant information, or that are not the results that were expected, this should be investigated further.
The auditor should make inquiries of management about the inconsistency or unexpected result and then
corroborate those replies with other evidence.

192 © The Institute of Chartered Accountants in England and Wales, March 2009
 EVIDENCE AND SAMPLING 11

If management responses cannot be corroborated or are unavailable, the auditor should consider
extending
audit testing.
Auditors will need to consider testing the controls, if any, over the preparation of information used

in
applying analytical procedures. When such controls are effective, the auditors will have greater
confidence
in the reliability of the information, and therefore in the results of analytical procedures.
The controls over non-financial information can often be tested in conjunction with tests of
accounting-related controls. For example, in establishing controls over the processing of sales
invoices,
a business may include controls over unit sales recording. In these circumstances the auditors could
test the
controls over the recording of unit sales in conjunction with tests of the controls over the processing of
sales invoices.
Reliance on the results of analytical procedures depends on the auditors' assessment of the risk that
the
procedures may identify relationships (between data) as expected, whereas a material misstatement
exists
(i.e. the relationships, in fact, do not exist).
The BSA states that 'the auditor should apply analytical procedures at or near the end of the audit
when
forming an overall conclusion as to whether the financial statements as a whole are consistent with
the
auditor's understanding of the entity'.
The conclusions from these analytical procedures should corroborate the conclusions formed from
other
audit procedures on parts of the financial statements. The auditor should consider whether the
assertions
made are consistent with their understanding of the entity, or, in particular, whether they reveal new
factors or undue influence by management. However, these analytical procedures may highlight areas
that
require further investigation and audit.
As we have discussed, analytical procedures should be used at the risk assessment stage. Possible
sources of
information about the client include:
 Interim financial information
 Budgets
 Management accounts
 Non-financial information
 Bank and cash records
 Sales tax returns
 Board minutes
 Discussions or correspondence with the client at the year end
Auditors may also use specific industry information or general knowledge of current industry conditions to
assess the client's performance.
As well as helping to determine the nature, timing and extent of other audit procedures, such analytical
procedures may also indicate aspects of the business of which the auditors were previously unaware.
Auditors are looking to see if developments in the client's business have had the expected effects. They will
be particularly interested in changes in audit areas where problems have occurred in the past.

1.5 Directional testing

Broadly speaking, substantive procedures can be said to fall into two categories:
 Tests to discover errors (resulting in over or understatement)
 Tests to discover omissions (resulting in understatement)
Tests to discover errors will start with the accounting records in which the transactions are recorded
and check from the entries to supporting documents or other evidence. Such tests should detect any
overstatement through causes other than omission.

Worked example: Test for errors

If a test is designed to ensure that sales are priced correctly, the test would begin with a sales invoice
selected from the receivables ledger. Prices on that invoice would then be checked to the official price list.

© The Institute of Chartered Accountants in England and Wales, March 193

2009
 Assurance

Tests for omission must start from outside the accounting records and then check back to those records.
Understatements through omission will never be revealed by starting with the account itself as there is
clearly no chance of selecting items that have been omitted from the account.

Worked example: Test for omission

If the test is designed to discover whether all raw material purchases have been properly processed, the
test would start, say, with goods received notes, to be checked to the inventory records or payables ledger.
Omission is usually more of a problem for certain account balances, such as liabilities, than others, such as
sales, as a company might be more inclined to understate liabilities to create a good position in the financial
statements. However, this is where the auditor's understanding of the entity is important, as the auditor
must make judgements about the situation of the entity at the moment – for example, a company with a
large VAT commitment might want to understate sales.

An assurance firm may use directional testing to identify over- and understatements.
The concept of directional testing derives from the principle of double-entry bookkeeping, in that for every
debit there is a corresponding credit (assuming that the double-entry is complete and that the
accounting records balance). Therefore, any misstatement of a debit entry will result in either a
corresponding misstatement of a credit entry or a misstatement in the opposite direction, of
another debit entry.
By designing tests carefully the assurance providers are able to use this principle in drawing conclusions, not
only about the debit or credit entries that they have directly tested, but also about the corresponding credit
or debit entries that are necessary to balance the books.
Directional testing is particularly appropriate when testing the financial statement assertions of existence,
completeness, rights and obligations, and valuation.
Tests are therefore designed in the following way.

Test item Example

Test debit items (expenses or If a non-current asset entry in the nominal ledger of CU1,000 is
assets) for overstatement by selected, it would be overstated if it should have been recorded at
selecting debit entries recorded in anything less than CU1,000 or if the company did not own it, or
the nominal ledger and checking indeed if it did not exist (e.g. it had been sold or the amount of
value and existence CU1,000 in fact represented a revenue expense)
Test credit items (income or Select a despatch note and check that the resultant revenue has
liabilities) for understatement by been recorded in the nominal ledger revenue account. Sales would
selecting items from appropriate be understated if the nominal ledger did not reflect the transaction
sources independent of the at all (completeness) or reflected it at less than full value (say if
nominal ledger and ensuring that goods valued at CU1,000 were recorded in the sales account at
they result in the correct nominal CU900, there would be an understatement of CU100).
ledger entry

A test for the overstatement of an asset simultaneously can give comfort on understatement of other
assets, overstatement of liabilities, overstatement of income and understatement of expenses.

Worked example: Overstated non-current asset

A company might have sold an asset for its book value of CU1,000. However, this sale was not reflected in
the financial statements, which still show the original asset of CU1,000. Therefore cash (the other side of
the entry for a sold non-current asset) is understated by CU1,000, depreciation charged is overstated and
depreciation accumulated is overstated.

194 © The Institute of Chartered Accountants in England and Wales, March 2009
 EVIDENCE AND SAMPLING 11

Interactive question 1: Evidence [Difficulty level: Exam standard]

In respect of an assurance engagement, which one of the following is the least persuasive method of
gathering evidence?
A Inspection of a purchase invoice
B Inspection of a sales invoice
C Inspection of inventory by the auditor
D Reperformance of a supplier statement reconciliation undertaken by the client
See Answer at the end of this chapter.

1.6 Audit of accounting estimates

The auditor often has to audit estimated figures, such as those for product warranties, depreciation,
inventory or receivables provisions, where the values included in the financial statements are not the result
of transactions with third parties (which are fairly reliable) but result from judgements made by
management. Yet these figures can have a very significant effect on reported profits.
The audit approach required is set out in BSA 540 Audit of Accounting Estimates. Essentially the auditor has
three methods to use:

Method Example

Test the process that management Management may use a formula to calculate the provision for
doubtful accounts receivable. The auditor can test this by:
used
to estimate the figure  Looking at past experience

 Checking the calculation

 Considering if anything this year is likely to have changed

the estimate
If a provision is required in respect of legal action against the
company, the auditor can use evidence from the company’s
Use an independent estimate legal advisors.
If a settlement is reached after the year end regarding a claim

Review subsequent events

against the company which requires a provision, the auditor
can use the evidence of the agreement to establish the
correct figure for the financial statements. In this case there is
usually no need to use the other two methods.

Having done the detailed work on the accounting estimate, the auditor checks the reasonableness of the
figure and then reaches a conclusion about whether it is fairly stated.
This sort of work is clearly needed in an audit assignment, where estimates such as provisions required for
damages in a lawsuit might be required, but the work is also very relevant to a number of other types of
assurance engagement. Reports on a business plan often require an accounting estimate to be checked. The
techniques used in these assignments will be the same as for audit assignments.

© The Institute of Chartered Accountants in England and Wales, March 195

2009
 Assurance

2 Selecting items to test

Section overview
 Assurance providers usually seek evidence from less than 100% of items of the balance or transaction
being tested.

 Every item in the population of items being sampled must have an equal chance of being selected in
the sample.

 The greater the risk of the area being sampled, the higher the sample size will be.

 When drawing conclusions from sampling, the auditor must identify which discovered errors affect
the overall balance.

2.1 The concept of sampling

Assurance providers do not normally examine all the information available to them; it would be impractical
to do so and using sampling will produce valid conclusions provided it is carried out properly.
BSA 530 Audit Sampling and Other Means of Testing states that 'when designing audit procedures, the auditor
should determine appropriate means for selecting items for testing so as to gather sufficient appropriate
audit evidence to meet the objectives of the audit procedures'. Remember that the BSA relates specifically
to audits, but all assurance providers may use sampling.

Definitions
Audit sampling involves the application of audit procedures to less than 100% of the items within an
account balance or class of transactions such that all sampling units have a chance of selection. This will
enable the auditor to obtain and evaluate audit evidence about some characteristic of the items selected in
order to form or assist in forming a conclusion concerning the population from which the sample is drawn.
Population is the entire set of data from which a sample is selected and about which an auditor wishes to
draw conclusions. A population may be divided into strata, or sub-populations, with each stratum being
examined separately. The term population is used to include the term stratum.

Some testing procedures do not involve sampling, such as:

 Testing 100% of items in a population (this should be obvious)
 Testing all items with a certain characteristic, as selection is not representative
Assurance providers are unlikely to test 100% of items when carrying out tests of control, but 100% testing
may be appropriate for certain substantive procedures. For example, if the population is made up of a small
number of high value items and there is a high risk of material misstatement then 100% examination may be
appropriate.
The BSA requires assurance providers to 'determine appropriate means of selecting the items for testing'. It
distinguishes between statistically-based sampling and non-statistical methods.

Definitions
Statistical sampling is any approach to sampling that involves random selection of a sample, and use of
probability theory to evaluate sample results, including measurement of sampling risk.
Non-statistical sampling is a subjective approach to inference, in that mathematical techniques are not
used consistently in determining sample size, selecting the sample, or evaluating sample results.

196 © The Institute of Chartered Accountants in England and Wales, March 2009
 EVIDENCE AND SAMPLING 11

The auditor may alternatively select certain items from a population because of specific
characteristics they
possess. The results of items selected in this way cannot be projected onto the whole population but
may
be used in conjunction with other audit evidence concerning the rest of the population.
 High value or key items. The auditor may select high value items or items that are suspicious,
unusual or prone to error.

 All items over a certain amount. Selecting items this way may mean a large proportion of the
population can be verified by testing a few items.

 Items to obtain information about the client's business, the nature of transactions, or the client's
accounting and control systems.

 Items to test procedures, to see whether particular procedures are being performed.

2.2 Design of the sample

When designing the sample, the BSA requires the auditor to 'consider the objectives of the audit
procedure and the attributes of the population from which the sample will be drawn', and to consider
'the sampling and selection methods'.
Assurance providers must consider the specific audit objectives to be achieved and the audit procedures
that are most likely to achieve them. The assurance providers also need to consider the nature and
characteristics of the audit evidence sought, possible error conditions and the rate of expected error.
This will help them to define what constitutes an error and what population to use for sampling.

Definitions
Error means either control deviations, when performing tests of control, or misstatements, when
performing substantive procedures.
Expected error is the error that the auditor expects to be present in the population.

The population from which the sample is drawn must be appropriate and complete for the specific
audit objectives. The BSA distinguishes between situations where overstatement or understatement is being
tested.
Assurance providers must define the sampling unit in order to obtain an efficient and effective sample to
achieve the particular audit objectives.

Definition
Sampling units are the individual items constituting a population.

Example: Sampling units

 Cheques listed on deposit slips
 Credit entries on bank statements
 Sales invoices
 Receivables' balances
 A monetary unit (an example of monetary unit sampling is given in section 2.3)

© The Institute of Chartered Accountants in England and Wales, March 197

2009
 Assurance

The BSA requires that the auditor 'should select items for the sample with the expectation that all sampling
units in the population have a chance of selection'. This requires that all items in the population have an
opportunity to be selected.
As we saw above, in obtaining evidence, the auditor should use professional judgement to assess audit risk
and design audit procedures to ensure this risk is reduced to an acceptably low level. In determining the
sample size, the auditor should consider whether sampling risk is reduced to an acceptably low level.

Definitions
Sampling risk arises from the possibility that the auditor's conclusion, based on a sample of a certain size,
may be different from the conclusion that would be reached if the entire population were subjected to the
same audit procedure.
Non-sampling risk arises from factors that cause the auditor to reach an erroneous conclusion for any
reason not related to the size of the sample. For example, most audit evidence is persuasive rather than
conclusive, the auditor might use inappropriate procedures, or the auditor might misinterpret evidence and
fail to recognise an error.

2.2.1 Factors influencing sample sizes

The BSA gives examples of factors which influence sample sizes for tests of controls and tests of details:

Tests of controls

Factor Stratification of the

An increase in the extent to which the risk of material misstatement is population where
reduced by the operating effectiveness of controls appropriate
An increase in the rate of deviation from the prescribed control activity The number of
that sampling units in
the auditor is willing to accept (tolerable error)
the population
An increase in the rate of deviation from the prescribed control activity

that
the auditor expects to find in the population (expected error)
An increase in the auditor's required confidence level (or a decrease in

the
risk that the auditor will conclude that the risk of material
misstatement is
lower than the actual risk of material misstatement in the population)
An increase in the number of sampling units within the population

Tests of details

Factor

An increase in the auditor's assessment of the risk of material

misstatement

An increase in the use of other substantive procedures directed at the

same
assertion
An increase in the auditor's required confidence level

An increase in total error that the auditor is willing to accept (tolerable

error)
An increase in the amount of error the auditor expected to find in the
population (expected error)
Effect on sample size
Effect on sample size
INCREASE
INCREASE

DECREASE DECREASE

INCREASE INCREASE
DECREASE
INCREASE
INCREASE

NEGLIGIBLE EFFECT DECREASE

NEGLIGIBLE EFFECT

198 © The Institute of Chartered Accountants in England and Wales, March 2009
 EVIDENCE AND SAMPLING 11

The greater the degree of confidence that the auditor requires that the results of the sample are in
fact
indicative of the actual error in the population (confidence level), the larger sample sizes have to be.
In
other words, if the auditor is placing a great deal of relevance on this (it is not corroborating other
evidence, for example) the higher the sample size will have to be.

Definition
Tolerable error is the maximum error in the population that the auditor would be willing to accept.

Tolerable error is considered during the planning stage and, for substantive procedures, is related to
the
auditor's judgement about materiality. The smaller the tolerable error, the greater the sample size will
need
to be.
(a) In tests of control, the tolerable error is the maximum rate of deviation from a prescribed
control procedure that assurance providers are willing to accept in the population and still conclude
that the preliminary assessment of control risk is valid.
(b) In substantive procedures, the tolerable error is the maximum monetary error in an account
balance or class of transactions that assurance providers are willing to accept so that, when the results
of all audit procedures are considered, they are able to conclude, with reasonable assurance, that the
financial statements are not materially misstated.
Larger samples will be required when errors are expected than would be required if none was expected, in
order to conclude that the actual error is less than the tolerable error. The size and frequency of errors is
important when assessing the sample size; for the same overall error, larger and fewer errors will mean a
bigger sample size than for smaller and more frequent errors. If the expected error rate is high then sampling
may not be appropriate. When considering expected error, the assurance providers should consider:
 Errors identified in previous audits
 Changes in the entity's procedures
 Evidence available from other procedures

Worked example: Designing the sample

Sarah is planning the audit of receivables at Manufacturing Company Limited (MCL). MCL makes all its sales
on credit, and the receivables ledger is extensive. However, Sarah has adjudged the area to be low risk as
most customers pay their debts promptly and controls over the receivables ledger and credit control are
good. In previous years, testing has revealed that few errors are discovered. She therefore applies a low
sample number.
During the course of testing, Sarah discovers a much higher number of errors than she was expecting. She
therefore increases her sample and extends her test.

In practice, most auditing firms use computer programs to set sample sizes, based on risk assessments and
materiality.

2.3 Selecting the sample

There are a number of selection methods available.
(a) Random selection ensures that all items in the population have an equal chance of selection, e.g. by
use of random number tables or computerised generator.
(b) Systematic selection involves selecting items using a constant interval between selections, the first
interval having a random start. When using systematic selection assurance providers must ensure that
the population is not structured in such a manner that the sampling interval corresponds with a
particular pattern in the population.
2

© The Institute of Chartered Accountants in England and Wales, March

199
 Assurance

(c) Haphazard selection may be an alternative to random selection provided assurance providers are
satisfied that the sample is representative of the entire population. This method requires care to guard
against making a selection that is biased, for example towards items that are easily located, as they may
not be representative. It should not be used if assurance providers are carrying out statistical sampling.
(d) Sequence or block selection. Sequence sampling may be used to check whether certain items have
particular characteristics. For example, an auditor may use a sample of 50 consecutive cheques to
check whether cheques are signed by authorised signatories rather than picking 50 single cheques
throughout the year. Sequence sampling may, however, produce samples that are not representative
of the population as a whole, particularly if errors only occurred during a certain part of the period,
and hence the errors found cannot be projected onto the rest of the population.
(e) Monetary Unit Sampling (MUS). This is a selection method that ensures that every CU1 in a
population has an equal chance of being selected for testing. The advantages of this selection method
are that it is easy when computers are used, and that every material item will automatically be
sampled. Disadvantages include the fact that if computers are not used, it can be time consuming to
pick the sample, and that MUS does not cope well with errors of understatement (as the computer
cannot select a CU which is not there) or negative balances.

Worked example: MUS

You are auditing trade accounts receivable. Total trade account receivables is CU500,000 and materiality is
CU50,000. You will select the balances containing each 50,000th CU1 from the ledger below.

CUSTOMER BALANCE CUMULATIVE TOTAL SELECTED

A 30,000 30,000

B 35,000 65,000 Yes

C 45,000 110,000 Yes

D 52,000 162,000 Yes

E 13,000 175,000

F 50,000 225,000 Yes

G 23,000 248,000

H 500 248,500

I 41,500 290,000 Yes

J 47,000 337,000 Yes

K 54,000 391,000 Yes

L 17,000 408,000 Yes

M 80,000 488,000 Yes

N 12,000 500,000 Yes

500,000

Material items are shown in bold and have all automatically been selected. The cumulative column shows
you when the next 50,000th CU1 has been reached.

200 © The Institute of Chartered Accountants in England and Wales, March 2009
 EVIDENCE AND SAMPLING 11

Interactive question 2: Factors affecting sample size

[Difficulty level: Exam standard]
When determining a sample size for tests of detail there are a number of factors that an auditor
should take
into account.
For each of the following factors, select whether it would cause the sample size to increase or

decrease.

(i) Decrease in the assessed level of tolerable error. Increase Decrease

(ii) Increase in the assessed risk level. Increase Decrease
(iii) Discovery of more errors than were anticipated during testing. Increase Decrease

See Answer at the end of this chapter.

3 Drawing conclusions from

sampling

Section overview
 The purpose of sampling a set of items was to enable the assurance providers to project the
conclusion to the whole population.

 Assurance providers must consider the nature of the error and whether it is fair to project that
error.

 If the projected error exceeds tolerable error then sampling risk must be reassessed and further
audit procedures must be considered.

When the assurance providers have tested a sample of items, they must then draw conclusions from that
sample. The purpose of sampling the items was to enable them to project the conclusion they draw from
the sample to the whole population.
To begin with, the assurance providers must consider whether the items in question are true errors, as
they defined them before the test. For example, when testing receivables, a sampled misposting between
customer accounts will not affect whether the assurance providers conclude the valuation of total
receivables is true and fair.
When the expected audit evidence regarding a specific sample item cannot be found, the assurance
providers may be able to obtain sufficient appropriate audit evidence by performing alternative
procedures. In such cases, the item is not treated as an error.
The qualitative aspects of errors should also be considered, including the nature and cause of the
error. Assurance providers should also consider any possible effects the error might have on other parts
of the audit including the general effect on the financial statements and on the assurance providers'
assessment of the accounting and internal control systems.
Where common features are discovered in errors, the assurance providers may decide to identify all items
in the population that possess the common feature (e.g. location), thereby producing a sub-population.
Audit procedures could then be extended in this area.
On some occasions the auditor may decide that the errors are anomalous errors.

Definition
Anomalous error means an error that arises from an isolated event that has not recurred other than on
specifically identifiable occasions and is therefore not representative of errors in the population.
©
A
M
201
 Assurance

To be considered anomalous, the assurance providers have to be certain that the errors are not
representative of the population. Extra work will be required to prove that an error is anomalous.
The assurance providers should project the error results from the sample onto the relevant population.
The assurance providers will estimate the probable error in the population by extrapolating the errors
found in the sample.
For substantive tests, assurance providers will then estimate any further error that might not have been
detected because of the imprecision of the technique (in addition to consideration of the qualitative aspects
of the errors).
Assurance providers should also consider the effect of the projected error on other areas of the audit. The
assurance providers should compare the projected population error (net of adjustments made by the entity
in the case of substantive procedures) to the tolerable error, taking account of other relevant audit
procedures.
If the projected population error exceeds or is close to tolerable error, then the assurance providers
should re-assess sampling risk. If it is unacceptable, they should consider extending auditing procedures or
performing alternative procedures. However, if after alternative procedures the assurance providers still
believe the actual error rate is higher than the tolerable error rate, they should re-assess control risk if the
test is a test of controls; if the test is a substantive test, they should consider whether the financial
statements need to be adjusted.

Worked example: Drawing conclusions from sampling

Adrian carried out a supplier statement reconciliation on Peabody Ltd. This means that he compared the
statements sent by suppliers to Peabody Ltd to the details on Peabody's own payables ledger. Tolerable
error has been set at CU10,000. The sample was 10 payables ledger balances totalling CU35,024 out of a
total of CU375,297. Adrian found that of these, 8 reconciliations proved that the balance on the ledger was
correct, one showed that an invoice had been misposted to a different supplier's account and one showed
that an invoice had not been posted at all.
When considering the results of his sample, Adrian decided that he can disregard the misposting, as,
although it means that two accounts were individually misstated, the overall balance was not affected by this
mistake. In the case of the invoice that had simply been omitted in error however, Adrian had to conclude
that this error of CU250, which does impact the overall total balance, could be repeated in the overall
population with the potential for causing material misstatement. Adrian projected the total population error
based on the sample and compared the outcome to tolerable error. In this case he found that the projected
error of CU2,679 was considerably below the tolerable error of CU10,000 and concluded that no further
action was required. He concluded from his testing that the trade payables balance in the financial
statements was fairly stated.

202 © The Institute of Chartered Accountants in England and Wales, March 2009
 EVIDENCE AND SAMPLING 11

Interactive question 3: Drawing conclusions from sampling

[Difficulty level: Exam standard]
Danielle has carried out a receivables circularisation on Donothing Ltd to gain evidence about the
receivables balance stated in the draft balance sheet. Identify whether the following conclusions
drawn by
her are correct or not.
(a) An amount disagreed by Lazy Limited because an invoice had True False
been
paid two days before the year end and cleared shortly after the
year end, did not constitute an error for the purposes of drawing
a conclusion for the whole population.
(b) An amount disagreed by Sloth Limited because a credit note had True False
been issued by Donothing Ltd a month before the year end did
not constitute an error for the purposes of drawing a conclusion
for the whole population.
(c) An amount disagreed by Busy Limited because they had paid the True False
balance some time earlier, which further enquiry revealed had
been posted to a different customer account, did constitute an
error for the purposes of drawing a conclusion for the whole
population.

See Answer at the end of this chapter.

© The Institute of Chartered Accountants in England and Wales, March 203

2009
 Assurance

Summary and Self-test

Summary

Evidence can be obtained by:

– Inspection
– Observation
– Inquiry
– Confirmation
– Recalculation
– Reperformance
– Analytical procedures

The strengths and weaknesses of

these methods depend on associated
issues relating to the quality of
evidence – for example, of whom the
inquiry was made, client staff or third
parties

Evidence will often be obtained from

a sample of a population, rather than
testing every item within it

Factors affecting sample size To draw a conclusion from a

include: sample, auditors must distinguish
– The degree of assessed risk between true errors and other
– The level of tolerable error errors

204 © The Institute of Chartered Accountants in England and Wales, March 2009
 EVIDENCE AND SAMPLING 11

Self-test
Answer the following questions.

1 Which one of the following procedures would give the most persuasive evidence that a control
operated as the assurance providers had been advised?
A Inspection of the controls handbook
B Inquiry of the staff operating the control
C Observation of the staff operating the control
D Reperformance of the control by audit staff
2 Indicate the purpose of the primary test for each type of account in directional testing.

(a) Assets Overstatement Understatement

(b) Liabilities Overstatement Understatement

(c) Income Overstatement Understatement

(d) Expense Overstatement Understatement

3 Identify the significant relationships in the list of items below.

(a) Payables (b) Interest (c) Purchases (d) Revenue

(e) Amortisation (f) Loans (g) Receivables (h) Intangibles

4 Identify whether the following statements are true or false.

True False

(i) The risk that the assurance providers' conclusion, based on a

sample, may be different from the conclusions that would be
reached if the entire population were subject to the same audit
procedure is sampling risk.

(ii) The risk that the assurance providers might use inappropriate
procedures or might misinterpret evidence and thus fail to
recognise an error is non-sampling risk.

5 Identify whether the following examples of sample selection are random, haphazard or systematic.
Random Haphazard Systematic

(a) Barry is selecting a sample from the list of

receivables balances. He selects the second, and
thereafter every 7th balance.

(b) Carol is selecting a number of purchase invoices

to
carry out a directional test. She selects them by
flicking through the files and selecting an invoice
occasionally.

Now, go back to the Learning Objectives in the Introduction. If you are satisfied you have achieved these
objectives, please tick them off.

© The Institute of Chartered Accountants in England and Wales, March 205

2009
 Assurance

Technical reference

1 Evidence
 Procedures to obtain evidence BSA 500.19 – 38

 Analytical procedures BSA 520.10 – 18

2 Selecting items to test

 The concept of sampling BSA 530.3 – 30

 Design of the sample BSA 530.31 – 41

 Selecting the sample BSA 530.42 – 43 + Appx 3

BSA 530.47 – 56
3 Drawing conclusions from sampling

206 © The Institute of Chartered Accountants in England and Wales, March 2009
 EVIDENCE AND SAMPLING 11

Answers to Self-test

1 D Reperformance by the auditor would give the strongest evidence of this being the case.
2 (a) Overstatement
(b) Understatement
(c) Understatement
(d) Overstatement
3 (a) and (c)
(b) and (f)
(d) and (g)
(e) and (h)
4 (i) True
(ii) True

5 (a) Systematic
(b) Haphazard

© The Institute of Chartered Accountants in England and Wales, March 207

2009
 Assurance

Answers to Interactive questions

Answer to Interactive question 1

B A sales invoice is an internally generated document and therefore provides a poor source of evidence.
It would be better to obtain information about sales from the customers.

Answer to Interactive question 2

They would all cause the sample size to increase.

Answer to Interactive question 3

(a) True – this is just a timing difference.

(b) False – this indicates that the credit note may not have been processed to the sales ledger, which
would be an error that could also be true of other potential credits due on the ledger.
(c) False – this error does not affect the overall balance on the ledger.

208 © The Institute of Chartered Accountants in England and Wales, March 2009