SU 08 Statistical Sampling Flashcards
When sampling is used
- when evidence to support the test is not in electronic form
- when the audit population is small and it is efficient to test with traditional procedures
- when relevant data is not reliable and internal controls over reliable data are weak
- when relevant data is in multiple formats/ not easy to use
When is data analytics used
- when the evidence to support the audit test is available in electronic form
- when the audit population is large, and the auditor’s tests are supported by reliable/relevant data in electronic form
- relevant data is reliable and internal controls over reliability are strong
- ## when relevant data is already clean or can be cleaned up easily
Where is data analytics used in the confirmation process of evidence gathering?
In selecting items to confirm (previously used sampling)
Sample selection methods
- non statistical sampling
- statistical sapling
Non-statistical sampling
when auditor uses their subjective judgement to determine sample size and sample selection
Statistical sampling
aka probability or random sampling
randomly selects sample items and uses a statistical method to evaluate results
Attribute sampling
used on tests of controls - tests a binary statement: are the controls working or not
Variables sampling
used for substantive testing of specific details (such as dollar values or quantities). Considering the accuracy of a variable
Sample risk
the risk that the auditor’s conclusion based on the sample is not correct if the procedure was applied to the whole population (risk sample is not representative)
Nonsampling risk
risk of an erroneous conclusion caused by a factor besides sampling risk (such as inappropriate procedures used, evidence misinterpreted, failure to recognize misstatements and control deviation
Types of sampling risk
Type I (Alpha) Risk aka audit efficiency error
Type II (Beta) Risk aka audit effectiveness error
Type I risk
Sample is overly negative so auditor assumes controls are not/ less effective or that a material misstatement exists when that is actually not the case for the population
risk of incorrect rejection
Causes unnecessary audit effort doing follow-up
Type II Risk
Sample is overly positive, so auditor assumes that the controls ARE effective, or that there are no material misstatements when that is not the case for the population
Risk of incorrect acceptance
Could potentially cause audit failure
Confidence level
Reliability level of sampling
Per statista: indicates the probability with which the estimation of the location of a statistical parameter (e.g., an arithmetic mean) in a sample survey is also true for the population.
COMPLEMENT of sampling risk
Relationship between confidence level and sampling risk
complementary: sum to one
Factors affecting sampling size
Population size: direct relationship not linear
Acceptable risk: inverse relationship
Variability in population: direct relationship
Tolerable deviation rate
acceptable (sampling) risk for attribute sampling
Tolerable Misstatement
Acceptable misstatement that may exist without causing the financial statements to be materially misstated - generally a value
expression of performance materiality
Expected deviation rate
in attribute sampling, an estimate (potentially based on a prior year or pilot sample) of the deviation rate in the population
deviation = control failure rate
should be less than tolerable rate - if more control should be omitted
Standard deviation
for variable sampling, estimate of population SD based on pilot sample or prior year
Determining sample size for attribute sampling
Use a table with axis of:
- expected population deviation
- tolerable deviation rate
If expected population deviation is greater than tolerable deviation
Cannot rely on that control, control risk for that control is 100%
if tolerable deviation rate is 0
auditor cannot afford any deviations, entire population must be assessed
Allowable risk of over reliance
allowable type II risk (risk of audit failure)
complement to confidence level
Inputs for determining sample size for variable sampling
- Confidence coefficient (C)
- Population standard deviation (S)
- items in population (N)
- tolerable misstatement (TM)
- allowed % for sampling risk aka Precision (taken from table) (P)
- allowance for sampling risk = precision x tolerable misstatement
A = TM x P
equation for variable sampling sample size
n= ((C x S x N) / A) ^2
Sampling approaches
-Random sampling
- systematic sampling
- block sampling (Cluster sampling)
- Stratification & money unit sampling
Systematic sampling
Random start + every nth item in population (determined by dividing the full population by the necessary sample size
Block sampling
Selecting groups rather than individual items
Stratification
Divide population into sub groups - results in less variability within groups than within whole population
can use a smaller sample size to get the same desired confidence level
Monetary unit sampling
stratification based on dollar value - risk-based
Is sampling with replacement or without replacement more conservative
with replaces
which sampling approach is assumed by the tables
with replacement (requires larger sample size)
Which sampling approach do most auditors use
without replacement - more efficient
Statistical sampling procedure steps
- Define objectives (account and assertion being tested)
- Define population
- Define acceptable level of sampling risk
- calculate the sample size needed for significant results
- select sampling approach
- collect samples
- evaluate sample results
- document procedures and results
Population in test of controls
- what control is being tested for what period
Population in substantive procedures
What items are being tested