What is sampling? - See text, page 364.

When to use sampling

• Tests of controls - Measure the number of times a control deviation occurs. We call this attribute sampling.

• Substantive testing - Measure the dollar amount of error in the population. We call this variables sampling or probability proportionate to size sampling.

What can go wrong when the auditor uses sampling?

• Sample is not representative of the population. (Sampling risk)

• Auditor can make mistates, eg, fail to recognize a misstatement or apply the wrong procedure. (Nonsampling risk)

• Sampling Risk + Nonsampling Risk = Audit Risk

How to prevent errors?, i.e., Control Audit Risk

• Control sampling risk. Either use statistical sampling or a well design nonstatistical sampling plan. Statistical sampling plan provides superior control since it quantifies sampling risk, ie, assigns a value to sampling risk. Since it is measured, it can be controlled. Nonstatistical sampling does not quantitfy sampling risk.

• Control nonsampling risk. Have proper planning and supervision.

Types of sampling risk. See text, p. 367.

• Tests of controls

  • Risk of assessing control risk too high.

     

  • Risk of assessing control risk to low.

   

• Substantive tests

  • Risk of incorrect rejection

     

  • Risk of incorrect acceptance

Steps for attribute sampling

1. Define objective, characteristics that indicate compliance, and deviations.

    

2. Define population.

    

3. Define sampling unit.

    

4. Define 3 variables that determine sample size.

   • Risk of assessing control risk too low.

   • Tolerable deviation rate.

   • Expected population deviation rate.

   Look up sample size using tables on page 374.

    

5. Select and test the sample.

   • OK methods for selection of sample:
     • Random
     • Systematic selection with random start provided population not organized according to sampling interval.

   • Not ok for selection of statistical samples:
     • Haphazard selection
     • Block selection

   • When testing the items in the sample, treat missing times as a deviation.

    

6. Project the sample results to the population - In the case of attribute sampling, use table on page 382.

    

7. Compare upper deviation limit (per table) to tolerable deviation rate (per audit plan).

    

8. Evaluate qualitative characteristics and other evidence.

    

9. Reach overall conclusion.