Page 117 - DCOM509_ADVANCED_AUDITING

Advanced Auditing

Notes

Task Prepare a report on the different types of sampling errors and the situations where
they are most likely to occur.

Self Assessment

Fill in the blanks:
9. The sampling error can be found by deducting the value of a ………………. from the value
of a statistic.
10. The ………………. and the ……………………. can be considered as an example of random
sampling error.

11. It is more difficult to quantify ……………….. errors than the ……………………… errors.

7.4 Statistical vs. Non-Statistical Sampling

The difference between statistical and non-statistical sampling is that statistical sampling allows
the user to measure the sampling risk associated with the procedure. Statistical sampling applies
the laws of probability to determine the percent likelihood that the sample does not accurately
reflect the population.

In essence, the laws of probability say that large, relatively homogeneous populations have
similar distributions and other features so that if a random sample is taken, it will consistently
reflect the population within certain limits. In order for the sample to be a “statistical” sample,
the results must be evaluated and two calculations made. These calculations tell the user how
likely it is that the sample results are within a given range of the actual population.

Example: A statistical sample would not only tell you that disallowed deductions are
estimated at $5,000, but that you have a 95% likelihood (confidence) of being within $50 (precision)
of the actual disallowed deductions.

A properly designed and applied non-statistical sample can provide results that are accurate and
effective, but will not measure the sampling risk.
Generally, the decision to apply a statistical or non-statistical sampling application to a particular
audit test is a matter of cost effectiveness. The department utilizes non-statistical random sampling
procedures.
Many audit programs in both the public and private sector have developed standard sampling
approaches which use predetermined allowances for acceptable sampling risk, expected and
tolerable error. These approaches reduce the time and effort required to determine the sample
size and evaluate results. The cost of these approaches is that they usually generate sample sizes
somewhat larger than necessary to allow a margin for variance. The New Mexico Taxation and
Revenue Department has elected to use fixed sample sizes of 250 and 500 items of interest for
variable sampling. The sample sizes apply to homogeneous and non- homogeneous populations
respectively. Stratification of a population can reduce sample size in most cases.
A provision for sampling units based on time periods is also provided. In any sampling approach,
the auditor must evaluate the population that is being tested, must determine if any stratification
should be done, must evaluate the cause of any exceptions and must apply the results from the
sample to the remaining portion of the population.

112 LOVELY PROFESSIONAL UNIVERSITY