CHAPTER

7
Audit sampling
SUGGESTED SOLUTION TO EXERCISE 7.1

1. Audit sampling: The application of audit procedures to less than 100% of items within a population of audit
relevance such that all sampling units have an equal chance of selection in order to provide the auditor with a
reasonable basis on which to draw conclusions about the entire population.
2. Anomaly: A misstatement or deviation that is demonstrably not representative of misstatements or deviations in a
population.
3. Population: The entire set of data from which a sample is selected and about which the auditor wishes to draw
conclusions. (E.g. all items included in an account balance or a class of transactions are populations. 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.)
4. Sampling risk: Arises from the possibility that the auditor’s conclusion, based on a sample, may be different from
the conclusion that would be reached if the entire population were subjected to the same audit procedure. There
are two types of erroneous conclusion, namely the risk the auditor will conclude, in the case of a test of controls
• that controls are more effective than they actually are, or in the case of a test of detail, that no material
misstatement exists when in fact it does. This type of risk affects audit effectiveness and is more likely to lead
to an inappropriate opinion; and
• that controls are less effective than they actually are, or in the case of a test of detail, that material
misstatement does exist when in fact it does not. This type of risk affects audit efficiency and would usually
lead to additional work to establish that the initial conclusion was incorrect.
5. Sampling unit: The individual items constituted within a population, e.g. deposit slips, credit entries on bank
statements, sales invoices or debtors’ balances.
6. Statistical sampling: Any approach to sampling that has the characteristics of
• random selection of a sample; and
• use of probability theory to evaluate sample results, including measurement of sampling risk.
A sampling approach that does not have these characteristics is considered non-statistical sampling.
7. Stratification: The process of dividing a population into sub-populations each of which is a group of sampling units
with similar characteristics.
8. Tolerable misstatement: A monetary amount set by the auditor in respect of which the auditor seeks to obtain an
appropriate level of assurance that the monetary amount set by the auditor is not exceeded by the actual
misstatement in the population.
9. Tolerable rate of deviation: A rate of deviation from prescribed internal control procedures set by the auditor in
respect of which the auditor seeks to obtain an appropriate level of assurance that the rate of deviation set by the
auditor is not exceeded by the actual rate of deviation in the population.
10. Non-sampling risk: The risk that the auditor reaches an erroneous conclusion (on the population) for any reason not
related to sampling risk, e.g. the auditor interprets the results of a valid sampling test incorrectly, due to a lack of
knowledge.

SUGGESTED SOLUTION TO EXERCISE 7.2

1. False Audit sampling is the application of audit procedures to less than 100% of items within a population such
that all items have a chance of selection. The selection of items can be by statistical method or non-
statistical method.

1
 Graded Questions on Auditing

2. False Professional judgement is an important part of statistical sampling, e.g. the auditor applies professional
judgement in setting the parameters for the sampling exercise and will still have to apply judgement in
interpreting the results of the exercise.
3. True Risk assessment procedures are performed to obtain an understanding of the entity and its environment,
whereas sampling is conventionally used to test units of a population so as to gather evidence about the
population as a whole.
4. False Sampling and analytical procedures are fundamentally different, conducting analytical procedures means
evaluating financial information through analysis of plausible relationships among both financial and non-
financial data (and subsequent follow up) of identified fluctuations or relationships which are inconsistent
with other relevant information or expectation. Sampling is as described in point 1 above.
5. False The sampling approach and methodology should be consistently applied throughout the audit.
6. True To be a valid statistical sampling approach, the approach must have the two characteristics described in
the statement, any other approach will be non-statistical.
7. True Monetary unit sampling selects every n th rand in the population (based on predetermined parameters).
The nth item is more likely to fall within a large amount (or balance) than a small amount (or balance).
8. False The lower the tolerable rate of deviation (i.e. the rate of deviation from prescribed internal control
procedures which the auditor is prepared to ‘live with’ the larger the sample needs to be. Thus, a higher
tolerable rate will require less testing, i.e. a decrease in sample size.
9. False Computer-assisted audit techniques can be used without testing the IT general controls and can
technically be classified as substantive audit procedures with the added ability to test the whole
population.
10. True When testing automated application controls, the auditor should complete a walkthrough and a sample
test of 1. (IT system consistently performs the calculation/error and only 1 is necessary.)

SUGGESTED SOLUTION TO EXERCISE 7.3

1. Deviations and misstatements are both ‘errors’ uncovered when conducting an audit sample. However, they differ in
that
1.1 deviations may be identified when testing controls, the auditor is testing for evidence that a control
procedure has not been applied (control); and
1.2 misstatement may be identified when the auditor is conducting a test of detail (amount).
2. The auditor’s professional judgement and experience. Materiality figure can also play a role.
3. 3.1 Random selection of the units in a sample.
3.2 Use of probability theory to evaluate sample results, including sampling risk.
4. No. Some procedures will be carried out on 100% of the applicable population. E.g. all loans to directors will normally
be audited, and all minutes of directors’ meetings will normally be reviewed.
5. 5.1 Audit effectiveness is about conducting procedures to gather evidence and drawing the correct conclusion
(and ultimately) giving the appropriate audit opinion.
5.2 Audit efficiency is about doing the correct amount of work to be in a position to give an appropriate opinion
(not too much, not too little).
5.3 An effective audit (arriving at the appropriate opinion) will not necessarily be an efficient audit; e.g. larger
samples than were actually required might have been drawn, resulting in the correct conclusion being
reached but at an unnecessary cost, e.g. a debtor’s circularisation may be a costly procedure to obtain
evidence of the existence of debtors but testing of subsequent receipts may have been cheaper for the same
(or better) results.
6. The auditor must ensure that the population is
6.1 appropriate for the objective of the audit procedure; and
6.2 complete.
7. False. The intention of sampling whether it be statistical or non-statistical, is to draw conclusions about the
population. If the sample contains R5 000 of misstatement, it is an indicator that the population will contain far more
than R5 000 in misstatement. It is obviously wrong to conclude that the R5 000 misstatement in our sample is the total
misstatement in the population.

2
 Graded Questions on Auditing

8. False. If the auditor expects a decrease in non-compliance, it will be because he or she believes the internal control is
‘working’ based on other information (e.g. more experienced employees now involved in the system). The assessed
risk of misstatement is reduced, and the extent of testing (sample size) can be reduced.
9. True. Data-orientated CAATs are concerned mainly with substantive testing, i.e. obtaining evidence to support the
assertions relating to balances in the statement of financial position and totals of transactions which underlie the
statement of comprehensive income. Also usually deemed as ‘testing with the computer’.
10. False. When using CAATs to perform substantive testing, you need to reconcile the population extracted to the clients’
population to confirm that the population is complete and accurate.

SUGGESTED SOLUTION TO EXERCISE 7.4

1. The population being sampled.

1.1 The intention of the company appears to be to draw a conclusion about the use of its software by home
computer users.
1.2 To draw conclusions about this population it would be necessary to select a sample from the entire ‘home
computer user’ population.
1.3 The population from which Interphace (Pty) Ltd has chosen to select its sample is the company’s own debtors
master file. This population is retail suppliers and only the retail suppliers with which Interphace (Pty) Ltd does
business.
1.4 Therefore conclusions can only be drawn about that population and not about the ‘home computer user’
population.
2. The sample size and sample itself.
2.1 To be of any value, the sample should be representative of the population itself.
2.2 In an attempt to get a ‘good response’ Joe David has picked out those customers whom he thinks will respond.
This introduces bias (Joe David’s opinion) and negates the representativeness of the sample.
2.3 It appears that no parameters or thought was given to the sample size – this was determined by an opinion of
which customers might or might not reply.
3. ‘Statistically pretty conclusive’
3.1 For a result to be ‘statistically pretty conclusive’, a statistical sampling approach must have been carried out.
3.2 At a minimum this would require that sample items (home computer user) be randomly selected, and the use
of the probability theory be used to determine sample sizes and evaluate sample results. As can be seen from
the above, this has not happened.
3.3 To be able to affect a statistical sampling approach, careful consideration must be given to setting parameters,
e.g. confidence levels, extrapolating results etc. None of this was done and at best this could only be described
as a non-statistical sample.
3.4 Thus the conclusion that the result is ‘statistically conclusive’ is totally invalid.
4. The conclusion drawn.
4.1 The conclusion drawn (Interphace software is the most commonly used software . . .) does not in any way
relate to the questions asked and responded to.
4.2 Asking a wholesaler of the company’s products whether the product meets user requirements and if they
outsold other brands, is a long way from concluding that Interphace software is the most commonly used
software by home computer users.
5. Overall.
5.1 This sampling exercise was very poor in all aspects. None of the basic requirements for meaningful sampling
were met, e.g.
• minimal planning: clear objectives, defining the population etc.;
• inappropriate selection of sample units; and
• proper evaluation of the results.

SUGGESTED SOLUTION TO EXERCISE 7.5

1. Decrease: 1.1 If we are able to tolerate (accept) a greater number of deviations (instances where a
control has not been applied as prescribed), we can do less testing; thus, sample size is

3
 Graded Questions on Auditing

reduced. This situation suggests that the risk associated with the internal control
procedure not being applied, has reduced.
2. Increase: 2.1 The higher the expected rate of deviation, the larger the sample size needs to be for
the auditor to be in a position to make a reasonable estimate of the actual rate of
deviation. In effect, we consider that there may be an increased risk of
unauthorised/valid overtime payments.
3. Increase: 3.1 The risk surrounding the material misstatement of inventory by the inclusion of non-
existent inventory has increased substantially. We must therefore increase our testing
to make sure we do not let material misstatement go undetected.
4. Decrease: 4.1 As we expect to be in a position to conduct successful subsequent receipt tests, we will
be going some way to gathering sufficient appropriate evidence about the existence of
debtors. This will reduce the amount of evidence we need to gather from the
circularisation. Hence a smaller sample size.
5. Increase: 5.1 If the auditor wishes to be in a position to rely to a greater extent on the internal
control system to identify, prevent, detect material misstatement, it will be necessary
to increase the tests of controls on the system. This should provide the level of
assurance required by the manager.
6. Various: 6.1 There are two considerations here. Firstly, the fact that there are 3 000 more inventory
items will have little effect on the sample size. (Increase in the population; this is part
of probability theory.)
6.2 However, with the introduction of new products and the associated spare parts, there
could be increased risk of material misstatement in inventory. E.g. physical storage of
the new inventory lines may be chaotic, not all new lines may have been entered
correctly on the database (existence or completeness) etc. Increased risk would lead to
larger samples as more evidence relating to existence would have to be gathered.
7. Decrease: 7.1 In general and with all other things being equal, the aggregate of the sample sizes from
the strata will be less than the sample size that would have been required if the total
population was sampled. As we are stratifying in terms of monetary value, better
coverage of the population is obtained by selecting say, five high value items from the
top stratum, than from selecting 15 or more items from the bottom stratum.
8. Decrease 8.1 It would not be feasible to test approximately 15 million home loans through a sample.
Stratifying the entire population through CAATs to test interest rates granted would
decrease the risk and would be an efficient way to test. This would indicate interest
rates granted lower than the repo rate and exceptionally high interest rates.

SUGGESTED SOLUTION TO EXERCISE 7.6

(a) Sampling risk

1. The intention behind sampling is to be able to draw a conclusion about the population from which the
sample is drawn once the necessary procedures have been conducted on the sample units.
2. In selecting the sample, certain parameters are set with the intention of providing the auditor with a
sample which will provide the auditor with a level of assurance at which the conclusion can be expressed,
e.g. 90% assurance that . . .
3. However, there is always the risk that the sample selected, despite the auditor doing everything correctly,
will not be representative of the population from which it is selected.
4. This is termed sampling risk which may lead to the auditor drawing an incorrect conclusion on the
population sampled. It is a risk which is inherent in sampling.
5. Essentially it is the risk that the auditor’s conclusion based on a sample may be different from the
conclusion if the entire population were subjected to the same audit procedure.
Non-sampling risk
1. With any auditing procedure, there is the possibility that the person carrying out the procedure will make
some kind of error which could lead to an incorrect conclusion being drawn on the population being
sampled.

4
 Graded Questions on Auditing

2. This is termed non-sampling risk, and in effect means that the erroneous conclusion reached is caused by
reasons other than the non-representivity of the sample (sampling risk). E.g. the auditor may
2.1 make mistakes in calculating sample size;
2.2 select the wrong sampling units;
2.3 apply an inappropriate procedure for the assertion for which evidence is being gathered; and
2.4 misinterpret the results.
3. Essentially it is the risk that the auditor reaches an erroneous conclusion for any reason not related to
sampling risk.
(b) 1. Non-sampling risk
1.1 The sampling plan, if applied properly, would have detected the problem.
1.2 The trainee made an error in failing to select precisely the requisition indicated by the sampling
plan.
2. Non-sampling risk
2.1 The trainee made an error in his choice of procedures: he omitted a procedure essential when
testing to the completeness of sales assertion.
2.2 There was nothing wrong with the sampling plan. The failure to identify the unrecorded sales
was due to the incompetence of the trainee.
3. Sampling risk
3.1 The trainee has not done anything wrong: he chose, and applied, his procedures correctly and
implemented the sampling plan correctly.
3.2 The undetected non-existent inventory items clearly were not selected in the sample of items
from the master file as identified by the sampling plan, i.e. their omission was strictly by chance –
a risk inherent in any procedure involving sampling.
4. Non-sampling risk (most likely)
4.1 The senior made an error of judgement in choosing his tolerable rate of deviation.
4.2 Presumably setting the tolerable rate too high was a result of the senior underestimating the risk
of the controls not operating (control risk). This resulted in the selection of a sample which was
too small.
4.3 The sampling plan would quite possibly have detected the breakdown in controls if the sample
size had been bigger, which it would have been had a lower tolerable rate of deviation been
applied.

SUGGESTED SOLUTION TO EXERCISE 7.7

(a) 1. Statistical sampling means any approach to sampling that has the characteristics of
1.1 a random selection of the sample; and
1.2 the use of the probability theory to evaluate sample results, including measurement of sampling
risk.
2. Non-statistical sampling is any sampling approach that does not have both of the characteristics in point 1
above.
(b) I will consider the following:
1. The results of any tests of controls carried out on the payroll system, e.g. if the controls over master file
amendments are strong, I may consider a smaller sample (and vice versa).
2. The extent of any other substantive tests we may plan to do in respect of wages, e.g. interrogation of the
master file to identify fictitious employees.
3. The amount of error I expect to find based on the nature and frequency of errors detected in prior tests,
e.g. mistakes in PAYE deductions.
4. How confident I want to be in the results of the sample, e.g. if I want to be 95% confident that there are
no fictitious workers, my sample size will be larger than if I want to be 80% confident. (These are difficult
decisions especially when statistical sampling is not being used.)
Note: It will not actually be possible for you to conclude in terms of a confidence percentage.

5
 Graded Questions on Auditing

5. Whether I can stratify the population to get a more meaningful result while reducing the size of the
sample.
6. My assessment of risk that there will be deviations/misstatements (see also 1 and 3) resulting in material
misstatement.
7. The size of the population would have some influence on the sample size, e.g. we have a population of
400; a sample of 10 would clearly be insufficient but 150 would be unnecessary.
Note: If, once we are conducting tests on the sample, it appears that matters are not as we anticipated, the
sample size can easily be extended, e.g. we detect numerous errors.
(c) I will select from the entire population on the master file because
1. we are testing the validity and accuracy of the payroll (including fictitious employees). If a person’s name
appears on the payroll, we want to confirm that that person exists, whether they actually earned or not;
and
2. we also want to confirm that the employee actually did not work (no clock card etc.) and that the
employee’s existence can be corroborated by the ‘sick leave’ system.
(d) I would consider:
1. Stratifying in terms of unskilled and skilled. It is more likely that fictitious workers will be included as
unskilled as it is easier to ‘hide’ them in the much bigger number of unskilled employees.
2. Stratifying by cost centre. This gives me the opportunity to select employees from all cost centres for the
specific reason of testing the authorisation of overtime by all ten foremen.
(e) Fictitious workers
1. For each employee selected, verify his/her physical existence by contact during working hours and inspect
the employee’s staff card and have the employee sign against his/her name on the sample workpaper.
2. Examine the employee’s file in the personnel department and
2.1 confirm the signature;
2.2 scrutinise the employment contract; and
2.3 inspect the copy of the ID document.
3. Look for evidence of a unique bank account number on the master file (EFT) by using GAS to search the
bank account number field for duplication.
4. Inspect the union deduction return schedule for the name of the employee.
5. The above tests would be supplemented by using GAS to search the employee master file for
5.1 any employee whose record does not reflect a SARS tax reference number;
5.2 an identity number; and
5.3 any duplications of the above (numbers).
Hours worked
1. Agree the hours worked per the payroll (sample) to the times recorded on the clock card.
2. Reperform the calculation of hours and confirm the split between normal and overtime hours.
3. Inspect the clock card for signatures of
• the wage clerk who calculated the hours; and
• the foreman for authorising the overtime.
4. If there is a planned overtime schedule (e.g. prepared by the foreman), agree the overtime on the clock
card to the schedule.
(f) Should the IT general controls environment have limited findings and the control environment is effective,
automated application controls may be tested. If the IT general controls environment is not effective, the auditor
may still rely on automated application controls but will need to test the access and change management around
the automated application control embedded in the application.
(g) Pay rate
• Determine who has access to change rates within the application or make changes on the master file.
• Have any changes been made to the configuration during the period under review?
• Have changes been authorised in the application?

6
 Graded Questions on Auditing

• Determine whether these rate changes were approved by the authorised individual.

SUGGESTED SOLUTION TO EXERCISE 7.8

1. Inventory count
1.1 There is no reference to any risk assessment pointing to a particular risk of overstatement.
1.2 There is usually risk associated with the existence of inventory (overstatement) but the level of testing being
suggested does not seem warranted unless there is very material and specific risk to existence.
1.3 In a general sense, emphasis can be placed on higher value items provided there is also some coverage of
low and nil value inventory items. A sample should be representative of the population but can be
stratified.
1.4 No rationale is given for aiming at 50% of the total value. This seems to be a particularly poorly considered
approach with no indication of how many test counts will be needed!
1.5 By testing only from the records to the physical inventory, no evidence will be gathered about the
completeness of inventory. A small sample for testing the other way should be selected. In addition, the
trainee’s method and reasoning do not make it absolutely clear whether he will be selecting from the
records or from the physical inventory.
2. Purchases
2.1 There seems to have been no evaluation of the risk of misstatement in ‘purchases’.
2.2 There is no indication of what testing the junior trainee will be carrying out – is he conducting tests of
controls or tests of detail?
Note: This is a very important point: the sampling plan must match the objective of the test.
2.3 Haphazard selection is not a totally random method of selection so this immediately negates the
possibility of introducing statistical evaluation ‘should we want to’.
2.4 Furthermore, statistical analysis requires that a proper statistical sampling plan be used. This requires
setting out objectives, setting parameters, defining the population etc. None of this has been done.
2.5 There is no indication of the population size or what document is actually being tested. Even without this
information, 200 items selected is a very big sample and
• boredom is very likely to set in, resulting in non-sampling error;
• a proper statistical sample is likely to be smaller and by using random tables, a far more
representative sample will be obtained; and
• audit efficiency and effectiveness are improved if statistical sampling is used on large populations.
2.6 There is no indication of how the audit senior arrived at R5 000 as a cut-off. Again, while stratification is
perfectly acceptable, there must be some justification for setting the different strata. As it stands, every
purchase under R5 000 will be ignored and when combined, could make up a large percentage of
purchases.
3. Completeness of creditors
3.1 Completeness is the assertion relating to creditors which tends to attract the greatest audit risk, so the
senior is correct with regard to its importance.
3.2 However, MUS (monetary unit sampling) is not a sampling plan which matches to the objective of testing
completeness.
3.3 MUS tends to ‘hook out’ the larger creditor’s balance from the list of creditors. In testing for
completeness, we are more interested in creditors who are not on the list at all or who are understated.
Therefore, MUS is not a suitable sampling plan.
3.4 The audit team should be concentrating on all the other completeness of creditor’s tests, e.g. comparison
with prior year balances, nil balances, cut-off testing for purchases etc.
3.5 The results of purchases testing will usually have a bearing on completeness of creditors testing (e.g.
purchases cut-off). The senior does not consider this but that is probably because he has not decided on,
or defined, what tests he is conducting on purchases (see point 2.2).
3.6 Successful statistical sampling (including MUS) is not just a matter of following step-by-step instructions in
the audit manual. It is essential that the audit team in the field has a good understanding of the technique.

7
 Graded Questions on Auditing

Steven Denham clearly lacks the understanding to make the sound professional judgement calls which are
also required when applying MUS.
4. Directors’ minutes
4.1 The reason that all directors’ minutes are reviewed is that important decisions, which the auditor may not
discover any other way, may be contained in the minutes.
4.2 Although 24 sets of minutes may seem excessive (it is not really) it is not possible to extract any
meaningful samples, statistical or otherwise because
• directors’ meetings do not necessarily deal with the same matters from meeting to meeting, i.e. the
sampling units are too dissimilar, and their characteristics are too varied;
• the population is far too small to perform a suitable sample; and
• furthermore, before the auditor reviews minutes, he does not know exactly what he is looking for.
Of course, he has an idea of matters which may arise, but there could be any number of matters
which the auditor did not expect.
5. Cost of plant and equipment
5.1 Again no mention of any risk evaluation in respect of plant and equipment.
5.2 It is not necessary to do such extensive verification of additions unless there is serious risk of
overstatement.
5.3 It appears as if only substantive work will be done on fixed assets. Is there any rationale for this decision?
Why will there be no tests of controls over the acquisition of fixed assets including plant and equipment.
E.g. if the authorising controls over additions (and disposals) of fixed assets are good and strictly applied,
substantive work can be significantly reduced.
5.4 Whatever the amount of testing is decided on, there should be some physical inspection of a sample of all
assets, not just additions, for existence (and physical condition/impairment purposes). Assets included in
the opening balance may no longer exist or be owned by the company. The cost of any such assets should
be removed from the plant and equipment accounts.
5.5 Furthermore, a sample of disposals will need to be verified to ensure that costs are properly removed
from the records where necessary.

SUGGESTED SOLUTION TO EXERCISE 7.9

1.1 When we come to perform our year-end procedures on trade accounts receivables, we will almost certainly stratify
the population.
1.2 Firstly, we will probably stratify the wholesale trade debtors from the retail trade debtors and evaluate them
separately as they exhibit different characteristics, e.g. wholesale debtors tend to be better payers than retail
debtors because the client is selling to established companies, as opposed to the man-in-the-street.
1.3 We must thus place more emphasis on debtors’ recoverability in the latter case as there is greater risk of material
misstatement.
1.4 Secondly, we will probably stratify the trade debtors according to value (to a large extent this will have been done
already in point 1.2 above, as wholesale debtors will tend to be larger than retail debtors), e.g. trade debtors
• below R500;
• between R500 to R50 000; and
• R50 001 and above, etc.
1.5 This means that we can audit more extensively those strata (or sub-populations) which represent the greater value
of total debtors (though we would not, of course, ignore completely the strata of lower values because we still need
our sample to be representative of the population).
1.6 Your present activities are part of identifying and assessing the risks of material misstatement of trade accounts
receivable, so anything that emerges suggesting that further stratification would be beneficial could be considered:
e.g. are there foreign debtors or a particular category of debtors who appear to be longer outstanding than the
others?
2.1 In this particular instance one would not sample from the population but would examine the whole population.
2.2 This is because
• the population is so small; and

8
 Graded Questions on Auditing

• of the importance of the matter being audited (i.e. legality, related parties etc.).
2.3 Strictly speaking, only statistical sampling is based on probability theory and some statistical sampling plans can only
be used on large populations. (The benefits of statistical sampling are usually associated with large populations.)
2.4 For non-statistical sampling, the decision whether to examine all or only part of a population depends entirely on
the auditor’s experience and his knowledge of the client and the population involved, as well as the assessed risk of
material misstatement.
2.5 As alluded to in point 2.2 above, directors’ interests in contracts are usually regarded as inherently risky and, to
reduce this risk to an acceptable level, it would be appropriate to ‘audit’ the whole population.
2.6 Don’t forget that you need to gather evidence about the completeness of directors’ interests in contracts, i.e. that
all such contracts have been recorded in the register. So, although you are examining the total population of entries
in the register, you will need to perform further procedures to satisfy yourself that all contracts which should be
entered have been entered.
3.1 As you are concerned with the expected amount of misstatement in the population, you must be conducting
substantive tests.
3.2 Setting the expected amount of misstatement is a matter of professional judgement (so you are quite right in saying
you will need assistance), but the following information would be helpful to us when determining the amount:
• the results of the risk assessment relating to the account heading you are testing (greater risk means we are
likely to expect more misstatement);
• the results of any tests of controls which may have been carried out, which are relevant to the account
heading you are testing (obviously if there were numerous deviations, we would expect more misstatement);
• the results of any related audit tests carried out in the prior periods (e.g. we may have found material
misstatements in the account heading on last year’s audit); and
• the results of any other substantive tests we might have carried out on the account heading (e.g. preliminary
analytical procedures may indicate a potential problem with the account).
3.3 The estimate we come up with will have a direct effect on the sample size:
• the more misstatement we expect, the greater the sample size; and
• if the level of expected misstatement has declined, say, from the prior year’s audit, the size of our sample will
decrease.
3.4 The more misstatement we expect, the greater the audit risk and we will need to do enough work to reduce the risk
to an acceptable level. As the level of risk increases, so must the size of the sample.
4.1 It is certainly possible to test more than one control using the same sample of transaction documents.
4.2 However, you should remember that if the factors that will influence your sample size for testing a particular control
are different to the factors that will influence the sample size for testing another control, you must extend (or
reduce) your sample. E.g. you may be looking for greater assurance (say 95%) for one control than for another. This
will require different sample sizes.
4.3 It is also possible to test for deviation and misstatement using the same sample. It is very important, however, to be
quite clear about the objectives you are trying to reach and how you evaluate the sample.
5.1 No. For a statistical sample to be ‘valid’ and defendable, you need to carry out the procedures exactly as required.
Statistical sampling uses statistics and the theory of probability to calculate sample sizes.
5.2 In effect statistical sampling states, ‘if you want to draw such and such a conclusion about the population (which will
be reflected and affected by the parameters you set) you must test a sample of a particular size’. If you don’t select
the correct number of items, you cannot draw a statistically based conclusion.
5.3 In effect you have made a non-sampling error which invalidates the sample conclusion. Strictly speaking, you should
reperform the exercise on a sample of 63 items.
5.5 The fact that you found no deviations in your sample is positive but does not ‘validate’ your smaller sample.
6.1 You do not need to test the full population when performing computer assisted audit techniques (CAATs), however,
using CAATs provides you with the capability to scale and the majority of the population should be considered.
6.2 Testing for ‘ghost employees’ means testing for employees that do not exist, where ‘employees’ have fictitiously
(and fraudulently) been created on the payroll (with valid bank details) and are being paid erroneously every month
as they do not really exist.
6.3 There are a few ways to test for ‘ghost employees’. You can

9
 Graded Questions on Auditing

– perform CAATs on the complete employee list and compare the ID’s and/or any other proof of identity to
ascertain if there are any errors or missing data;
– test for duplicate bank account details; and
– test for clerical errors in birth dates, hire dates or termination dates.

SUGGESTED SOLUTION TO EXERCISE 7.10

(a) I would not be satisfied with the work of Elvis Adams for the following reasons:
1. Despite the selection of the 12 clients to be circularised being made on a random basis, the size of the sample
is far too small to be representative of the population.
2. In effect Elvis Adams has disregarded (or never considered) sampling risk.
3. The population from which he has selected his sample is incorrect; the full population of sets owned by
Screens (Pty) Ltd (14 000) should have been used; not the population of rented sets (alternatively, further
existence tests should be carried out on sets on hand).
4. He clearly gave no thought to the appropriate sampling plan to be used to achieve his objective, nor did he
give any considerations to the parameters which should be set, e.g. confidence level.
5. This further suggests that he gave very little consideration to the risks that may affect the existence assertion
for television sets.
6. His response to the sample results is totally inadequate
6.1 Although it would be virtually impossible to meaningfully extrapolate the results of the sample over
the population, Elvis Adams is not even aware of the need to do this.
6.2 He has failed to follow up adequately on the no replies – it could be that these rentees are fictitious
and that the televisions do not in fact exist – precisely what the test is attempting to prove.
6.3 The follow up procedure for the returned television was not appropriate, he should have
• established details of the return (e.g. date, documentation) number and inspected the television
set if it was ‘in stock’ or sought positive confirmation from the new rentee.
6.4 83% return is excellent for a circularisation but
• the circularisation is not, on its own, sufficient evidence;
• the results must be analysed in relation to the population as a whole not in terms of the sample;
and
• it is a very small sample.
7. Further audit procedures must be performed
7.1 There is a chance that the sample is representative of the population of rented out sets – in which case
there could be as many as 1/6 of 12 321 sets which do not exist, i.e. are with fictitious rentees. This
would mean a possible overstatement of inventory of approximately R5 million (total value
R35 million) which is material. Further tests must be performed to reduce this risk to an acceptable
level.
(b) 1. Further procedures
1.1 Physically verify the existence of the television sets on hand.
1.2 In preparation for point 1.3 (below) set strict sampling parameters (high confidence, low tolerable
misstatement) to address the risk relating to existence and obtain a sample size.
1.3 Draw a random sample from the total population of rental agreements.
1.4 For each rental agreement selected, trace by name and number to the most recent receipt of a rental
payment. This would provide corroborative evidence that the set exists as clients will not pay for
something they don’t have.
1.5 By inspection of deposit slips/bank statements, confirm the amount was banked. This would provide
evidence that receipts are not simply fictitious book entries.
1.6 Any rental agreement that cannot be tied to a payment would be regarded as a potential fictitious
rentee.
• Each of these would be followed up by phoning or faxing the rentee for positive confirmation of
the existence of that particular television set (e.g. serial number) and rental agreement number.

10
 Graded Questions on Auditing

• The unsubstantiated rental agreements which remain should be extrapolated over the
population and compared to the maximum tolerable misstatement. This will give us an
estimation of the number of rental agreements for which no payment has been recently
received.
1.7 The results of the test should be referred to the management of Screens (Pty) Ltd.
• If the extrapolation provides an unacceptable result, the client should be requested to
investigate further and provide us with detailed explanations.
• We should substantiate these explanations.
• If the extrapolation falls within our tolerable misstatement, the matter is less of a problem with
regard to our report on fair presentation, but it will suggest an internal control breakdown which
must be reported to management.

SUGGESTED SOLUTION TO EXERCISE 7.11

(a) 1. Yes, I would agree with my senior.

2. Justification
2.1 The authorisation process is inadequate.
• As only one of three authorised signatories is required, it would be a reasonably simple
matter for an employee to have the same journal entry passed twice by presenting the
genuine JER to one signatory and then later presenting the false JER to another signatory.
• This weakness is compounded by the fact that the signatory does not indicate in any way
that he or she has reviewed the supporting documentation; it should be signed or stamped
accordingly otherwise it can be presented a second time, to support a false JER.
2.2 Returning the JER to the requester provides the opportunity for the requester to alter the JER
before passing it on for capture.
2.3 It appears that the clerical assistant does not check the JER for authorisation but simply captures
what s/he is given.
2.4 There is only one copy of the JER, which weakens the audit trail and makes it much easier for
evidence of an unauthorised journal entry to be destroyed. (Compounded by the fact that it is filed
in the accounting department’s general space.)
2.5 There appears to be no reconciliation of journal entries that have been entered into the system
and journal entries that were authorised by a senior accounting staff member (transaction file/log
to JERs):
• the clerical assistant could be passing journal entries, e.g. writing off a debtor’s account in
collusion with other members of staff; or
• members of staff could be bypassing the authorising step and simply sending the JER through
for processing (see 2.3. above).
2.6 In other words, there is no control carried out to ensure that only authorised journal entries have
been entered.
(b) Step 1 : Define the objective of the test.
: We wish to gather evidence on whether all journal entries are authorised, i.e. that they are signed
and supported by appropriate original documentation.
Step 2 : Decide on the procedure to be performed.
: We must specify the ‘error (deviation) condition’, in this case unauthorised JERs. The procedure
will be to select a sample of journal entries from the computerised log (or transaction file if it is
available) of journal entries and trace them to the JERs which will be inspected for the signature of
one of the approved signatories and traced to original supporting documentation.
Step 3 : Define the population (appropriate and complete).
: The population will be all journal entries passed for the year under audit, listed (numerically) on
the computerised log of journal entries.
Note: The start and end points of the population can be established as the JERs are sequenced.
Step 4 : Define the sampling unit.

11
 Graded Questions on Auditing

: The sampling unit will be a (each) journal entry.

Step 5 : Determine the sample size.
: If we are to use a statistical package (as requested), we will need to define
• the confidence level;
• the precision;
• the tolerable rate of deviation rate;
• the expected rate of deviation rate; and
• the population size.
These parameters can then be applied to the appropriate formulae tables to give sample size.
Note: If a non-statistical package is used, the sample size will be made judgementally.
Step 6 : Decide on the sample selection method. (Select the sample.)
: As we require a statistical sample, the selection must be done on a random basis to afford every
journal entry an equal chance of selection.
: As we have software available, we would use a ‘random number generator’ and select by JER
number.
Step 7 : Perform the audit procedures.
: We will extract the journal entries whose numbers appear on the random selection.
: We will then obtain the JER and inspect it to determine whether it was authorised and supported.
Step 8 : Analyse the nature and cause of errors (deviation).
: We will analyse the sample results and consider the cause of the deviations including whether the
deviations are anomalous or not.
Step 9 : Project the sample results over the population.
: We will then calculate the number of deviations (see step 2) in the sample and using the formula/
tables and parameters set in step 5, calculate the estimated rate of deviation for the population.
Step 10 : Conclude.
: Compare the projected deviation rate to the tolerable deviation rate and conclude whether our
results fall within acceptable (pre-set) parameters.
: Decide on the need for further audit procedures.
(c) 1. No, we cannot assume that fraud has occurred.
2. We did not test for fraud: we tested whether journal entries were authorised. Even though we discovered
an unacceptable level of unauthorised journal entries, we cannot conclude that they were fraudulent/that
there was any intent to act fraudulently. It may simply be that an employee forgot to have the JER signed
(perhaps in a hurry) and the weak internal controls did not pick it up.
3. Of course, the results have to be carefully followed up because we have been alerted to potential fraud.
The risk of material misstatement has increased, and we would have to respond appropriately.
(d) Depending on the outcome of the general IT control evaluation, the auditor will be in a position to proceed with
automated control testing.
The entity’s information system may include the use of manual and automated elements, which also affect the
manner in which transactions are initiated, recorded, processed and reported. In particular, procedures to initiate,
record, process and report transactions may be enforced through the IT applications used by the entity and how the
entity has configured those applications. In addition, records in the form of digital information may replace or
supplement records in the form of paper documents.
ISA 315 advises that automated controls may be more effective than manual controls in the following circumstances:
• High volume of recurring transactions, or in situations where errors that can be anticipated or predicted can be
prevented, or detected and corrected, through automation.
• Controls where the specific ways to perform the control can be adequately designed and automated.
(e) 1. To audit around the computer
• This approach treats the computer system and programs as a black box and relies on review and comparison
of the input and output documents. The rationale behind this approach is that if the source documents are
valid, accurate and complete, and the output produced by the computer system as a result of processing these

12
 Graded Questions on Auditing

source documents is correct, the processing functions of the computer system are being performed correctly.
The manner in which these processing functions are performed is deemed to be of little consequence. This
approach assumes that the computer-generated output can be traced back and compared to the input.
• The audit is performed by selecting a sample of transactions that have already been processed and then
tracing these transactions from their point of origin as source documents to the output documents or records
produced by the computer system.
• This approach is only feasible if the computer system under consideration is a simple, batch-oriented system
with no significant controls or automated/integrated functions built into the system.
• Additional requirements for the adoption of this approach are that control is maintained by segregation of
duties, independent checks and management supervision together with the maintenance of a clear audit trail.
2. To audit through the computer
• This approach is concerned with testing the computer system and controls that are built into the system.
• Simplistically, this is achieved by the auditor sending transactions (test data), some of which will contain errors
which the system’s program controls should detect, through the system. In this way the auditor tests whether
controls are working as expected.
3. To audit with the computer
• There are two aspects to ‘auditing with the computer’:
– using the computer to assist in the performance of audit procedures (mainly substantive testing); and
– using the computer to produce electronic/automated workpapers, audit programs and financial
statements.
• Using this approach for substantive testing involves gaining access to a client’s files and using audit software
(programs which help the auditor to do what he has to do) to read, sort, compare and analyse data on the file,
very quickly and extensively.
• The idea behind using the computer to automate the audit is to make it a more effective and efficient audit by
harnessing the power of the computer.

SUGGESTED SOLUTION TO EXERCISE 7.12

(a) Existence of inventory.

(b) 1. A sample of 30 boxes from a population of 15 000 units may appear to be very small.
2. As it is a non-statistical sample, the number was arrived at judgementally and would have been based
primarily on our assessment of the risk of material misstatement in the inventory balance arising from the
inclusion of non-existent inventory.
3. Based on the evidence available at the time of making the assessment, this particular risk was probably
assessed as low because of
3.1 a strong control environment;
3.2 sound controls over the receiving and issuing of inventory;
3.3 sound physical controls over inventory; and
3.4 in prior years no instances of empty boxes were found and there is no indication that there has been a
weakening of controls during the year.
The fact that gaming consoles are attractive items to steal would also have been considered.
4. Had a statistically based sample been selected, it would probably not have been much bigger.
(c) Peter Dean 1. He is making the basic mistake of drawing a conclusion on the population without
extrapolating the results of the sample over the population. Even a simple proportional
extrapolation would come out at 500 (potentially) empty boxes.
2. It is not one empty box in 15 000; it is one empty box in 30, as only 30 boxes were selected.
3. His conclusion is completely invalid.
Roy Sidhu 1. The principle of extending the sample is sound on the grounds that this may just be a ‘one-
off’ error which happened to be identified in our initial sample but is not indicative of the
population as a whole.
2. However, it is not as simple as just selecting another 30 boxes.

13
 Graded Questions on Auditing

2.1 The underlying risk assessment should be re-examined.

2.2 The ‘cause’ of the empty box should be followed up, e.g. was it perhaps used as a
demonstration model?
3. Furthermore, the empty box can never be ignored; it must be followed up, recorded and
reported to the warehouse manager.
Bob Mmetha and Creedence Ndlovu
1. Bob Mmetha has no basis for deciding that one out of 30 (3,3%) is a ‘tolerable error rate’
(tolerable misstatement).
2. If he agrees with Peter Dean, he has no understanding of sampling and the conclusions
which can be drawn about the population.
3. Creedence Ndlovu has a better understanding of the situation. The most obvious and likely
explanation is that the contents of the box have been stolen and we would, as auditors, be
in trouble if we did not recognise this as a potential theft (misappropriation of assets).
4. Her suggestion that every box will have to be checked is impractical and unnecessary. Her
argument that you cannot have a ‘tolerable error rate’ for theft is not quite correct.
4.1 Theft is ‘unacceptable’, but, even if there has been some undetected theft,
inventory can still be fairly presented (i.e. contain only immaterial misstatement).
4.2 In essence, there are two issues here: fair presentation and the misappropriation of
assets. The former may be achieved even if there has been an occurrence of the
latter.
Peter Dean 1. Again, a real lack of understanding on his part. If additional testing is required, it must be
carried out. We cannot voluntarily limit our scope to suit the warehouse manager.
2. The warehouse manager has a legitimate concern about opening boxes and it is reasonable
for us to be careful, but we need to satisfy the requirements of an independent audit by
gathering sufficient appropriate evidence.
(d) 1. The discovery of the empty box should be reported to the warehouse manager and an explanation sought.
Although the most likely explanation is theft (the box has been resealed) there may be a perfectly valid
explanation, e.g. the gaming console was taken out and is being used as a demo model.
2. If the warehouse manager supplies an explanation, its validity should be confirmed.
(e) 1. I would treat each of the four types of gaming console as a separate population.
Reason: Each type of gaming console is already stored in its location and the gaming consoles vary in value.
This is, in effect, a form of stratification.
2. I would randomly select a judgementally based sample for each population, selecting a larger sample for the
Nintendo and Sony consoles.
Reason: No real need for a statistically based sample, but some emphasis should be placed on the high value
of gaming consoles which are likely to be more susceptible to theft.
3. I would, with the warehouse manager, determine whether there were any factors (other than value) which
may make a particular type of gaming console susceptible to theft, e.g. less effective custody controls in that
location of the warehouse.
Reason: To determine whether the sample size for that type should be increased.
4. I would carry out the procedure of opening the boxes with the help of the warehouse manager and his team.
Reason: It would be far more cost efficient to incorporate the client’s staff and in addition, the responsibility
for following up on the suspected theft lies with management as it is their responsibility.
5. Another possibility may be to actually use a statistical sampling plan (setting parameters which take into
account the ‘updated’ risk assessment which we should have carried out) to establish a ‘statistical’ sample size
and random selection of sampling units for testing.
Reason: We suspect fraud (misappropriation of assets) and thus we may be well advised to obtain ‘best
quality’ evidence. Should we need to defend our actions at a future date it will be easier to defend a ‘statistical
sample’ than a ‘judgemental sample’.
(f) The risk of an empty box being sold can cause reputational damage with the distributor which sold it and the
customer who ultimately purchased it.
There is also a risk that there may be more of these empty boxes that have been distributed.

14
 Graded Questions on Auditing

Angry or annoyed customers may resort to social media which could cause more reputational damage if the word
spread on social media.
Potential impact on overall sales as customers might avoid purchasing from Nceba (Pty) Ltd.

SUGGESTED SOLUTION TO EXERCISE 7.13

(a)

Random Sampling Every unit has an equal chance of selection. The selection can be made manually by
using random number tables or by computer, using random number generation
software.
Systematic Sampling This involves selecting a random starting point and then selecting every, say,
30th item. As there may be patterns within the population this is a risky, though cost-
effective, method.
Haphazard Sampling Here the auditor attempts to simulate randomness by avoiding conscious bias or
predictability and not following a structured technique.
 In a non-statistical sample, it is an acceptable technique. It is not a valid method of
selection if using statistical sampling as guaranteed randomness is a prerequisite
of the statistical sampling approach.
Block Sampling This involves selection of a block of contiguous (e.g. numerically consecutive) items
from within the population. This is not often an appropriate selection technique
where the auditor wishes to draw valid conclusions about the entire population.
Monetary Unit Sampling MUS is a value-weighted selection method in which the sampling unit is every rand in
(MUS) the population. Every nth rand is then selected. This will result in larger amounts
being selected because larger amounts have more rand units. E.g. if we select a
sample of debtors from the debtors list, we do not consider the individual debtors to
be the sampling unit, we regard each rand in each balance to be the sampling unit.
Therefore, if we select every 1 110th rand, the chances are that the 1 110th rand will
be contained in large balances rather than in small balances. The debtors’ balances
into which the nth rand falls will be selected for the sample.

(b)
1. Audit sampling
We want to draw a conclusion about the inventory population as a whole. Therefore, the means of selection must be
one where every item in the population has a chance of being selected for testing. A sample will be randomly selected
from the nearly 3 000 different products.
2. Select all items (100% examination)
There will be only 12 sets of minutes to inspect. We do not want to take the risk of missing important decisions taken
by the directors which may affect the audit and fair presentation. Such decisions may only be recorded in the minutes.
3. Audit sampling
Again, we want to be able to draw a conclusion on whether the control is being complied with for all purchase orders.
Thus, we need a (random) sample selected in such a manner that each purchase order has a chance of being selected.
4. Select all items (100% examination)
Twenty-nine leases are not a great number to test. The evidence produced by the test is important because it will
confirm whether the property, plant and equipment etc. have been fairly presented.
5. Select all items (100% examination)
The population is all debtors over 90 days which amounts to approximately 70 debtors. In setting the allowance for
bad debts, the credit manager/financial manager is required to consider each debtor, not just make an overall
decision about the allowance. Therefore, information about each debtor should be available. Furthermore, the
population (90-day debtors) is only homogeneous in the sense that they are long outstanding. The reasons for this will
be numerous and the recoverability will vary from debtor to debtor so each debtor should be considered separately. It
is also important to note that the risk of material misstatement of the allowance for bad debts has increased. The

15
 Graded Questions on Auditing

amounts owed by this category of debtor (90 days and over) have increased significantly (by nearly 60%) while sales
have remained reasonably constant.
6. A combination of 100% examination (of a stratum) and audit sampling
We are primarily testing for the existence of debtors and have a situation where 22 of approximately 700 debtors
represent nearly 50% of the debtors’ balance. Therefore, we should stratify the debtors’ population. The top stratum
(22 major debtors) should all be positively circularised and the rest of the debtors should be tested by the random
selection of a sample to be circularised. This will give us a good coverage of the entire population.
7. Selecting all items (100% examination)
The use of our audit software makes it very easy and quick to compare two fields in the inventory master file for every
single product line. Accurate reports will be generated in a cost-effective manner.
8. Selecting specific items
We are testing these transactions from the purchase journal because they display a specific characteristic – they are
transactions with related parties. Such transactions are inherently high risk.

SUGGESTED SOLUTION TO EXERCISE 7.14

1. Non-sampling risk: The junior did not receive the latest sampling plan and therefore may arrive at an erroneous
conclusion. This is not related to the sampling methods etc. but due to error from the auditor’s side.
2. Sampling risk: The risk that the auditor’s conclusion based on a sample may be different from the conclusion that
would be reached if the entire population were subjected to the same audit procedure, especially as there will be
millions of transactions in the population considering the size of Lucky’s Convenience Store.
3. Five reports did not have any evidence of review. The daily reports would equate to a sample of 20 that was tested in
total. Therefore, the control failed as only 75% of the test population had evidence of review. The expected
misstatement/rate of deviation is 5% based on past experience and knowledge of the business. The 25% deviation is
therefore an automatic fail, and the sample will not be extended. The control therefore fails, and further substantive
testing will be required to provide assurance over the master product and pricing list.
4. IT general control tests have been performed and access and change management controls have been tested
successfully. For this reason, only one automated application control is required to be tested; the sample of 25 was
excessive and not required.
5. The analytical procedure to compare the trial balance year-on-year will not require sampling. The full trial balance will
be compared (per account), to assess the differences in the balances.
6. The computer-assisted audit technique will be able to test the full population to confirm there are no missing or
duplicated sales orders, and all numbers are sequential.
7. The sample of 25 is correct as there are many employees and termination or change/revoking access (assumption is
more than 250). The access that was not revoked immediately due to the employee suddenly going into labour can be
described as an anomaly – a misstatement or deviation that is demonstrably not representative of misstatements or
deviations in the population. No further testing will be required, as the control is deemed effective.

SUGGESTED SOLUTION TO EXERCISE 7.15

1. The selection method the auditor decides on by which to select items for testing (e.g. sampling per ISA 530, selecting
specific items per ISA 500). With the verification of inventory, we would like to verify a high values of inventory items
based on materiality. Degree/level of assurance is needed from the results.
2. The value of inventory kept at a retail store / the materiality of the inventory at branches.
3. The risks of material misstatement associated with the inventory on hand examples:
a. The nature of the inventory (consignment stock kept by branches / inventory difficult to count). / The existence
of fraud risk factors at a specific branch (e.g. previous frauds, high-risk location).
b. Which retail stores experienced problems with the previous year’s year-end inventory count?
c. Analytical Procedures / gross profit percentage (analytical procedures performed as a risk assessment
procedure) indicates fluctuations which is not in line with expectations.
d. Any results of inventory counts or interim counts during the year which indicate large fluctuations.
e. Significant changes within inventory or the branch or new locations.
4. Any inventory counts already performed by the external audit at the interim stage?

16
 Graded Questions on Auditing

5. Any inventory counts already performed by the internal auditors (if available) and the findings thereof.
6. The adequacy of the controls over the existence of the inventory (do all retail stores have the same controls or do the
controls differ among stores?). Any possible changes in staff/controls during the year over inventory.
7. Specific requests from management.
8. On a random-sample basis (rotation every year) select a few branches seen as not material/risky OR incorporate
unpredictability into your audit response.

SUGGESTED SOLUTION TO EXERCISE 7.16

1. Selection of specific items (high values or key items): Because there are only two assets with a value of 90% of the
value and we are auditing substantively (value), this seems to be the most appropriate method. Selecting some of the
smaller values statistically seems appropriate, too.
2. Sampling (statistical sampling or non-statistical sampling): All purchase transactions will have similar characteristics
(the same controls applied to them). It will be impossible to audit all the transactions and because we are performing
tests of controls, a sampling method seems more appropriate.
3. Selection of all the items (100% test): We will need to audit (inspect) all the minutes. If we leave one minute out, we
could miss an important decision throughout the year. Inspecting minutes also assists in completeness testing and we
therefore need to inspect all the minutes, not to miss any important decision.
4. Selection of all the items (100% test): We are auditing for rand values using substantive procedures. The values are
almost the same and therefore it seems appropriate to audit all three.
5. Selection of specific items (high value or key items): If we select only the R233 886 debtor, we have already tested
60% of the items. If may be advisable to also include the R65 118 debtor. For the smaller debtors, one can use a
sample selection method, but because there aren’t that many debtors and if there are no other risk indicators, these
debtors do not need to be selected. Logically, it seems that it is quite possible to select all the items, but this does not
address the internal audit policy.

SUGGESTED SOLUTION TO EXERCISE 7.17

(a) Population size = 27 668 – 22 119 = 5 549 + 1 = 5 550

(b) Interval = Population / Sample size
= 5 550 / 150 = 37
Starting point = 22 119 + 12 (– 1) = 22 130 (1st sale invoice number to select for testing)
22 130 + 37 = 22 167 (2nd sales invoice number for selection)
22 167 + 37 = 22 204 (3rd sales invoice number for selection)
(c) 1st sales invoice number = 24 409 (Cell E10)
2nd sales invoice number = 23 513 (Cell E27)
3rd sales invoice number = 25 679 (Cell E30)

SUGGESTED SOLUTION TO EXERCISE 7.18

(a) Selection of 100% of the items (all items)

Selection of specific items (high value, high risk or key items)
Sampling selections
(b) 1st tax return number = 72 374 (Cell B2)
2nd tax return number = 59 173 (Cell B7)
3rd tax return number = 68 996 (Cell B10)
(c) Interval = Population / Sample size
= 25 000 / 125 = 200 (nth)
Starting point = 53 001 + 23 (– 1) = 53 023 (1st tax return number to select for testing)
53 023 + 200 = 53 223 (2nd tax return number for selection)
53 223 + 200 = 53 423 (3rd tax return number for selection)

17