Description

THREE MODELS

. Fixed Sample Size Attribute

. Stop-Or-Go

. Discovery

CHARACTERISTICS (some, general):

. Dichotomous Test - Answers the question yes or no

. Answers the Question How Often

. Test internal controls via estimating rate of deviation

. Quality characteristics (vs. quantifying under variable models)

. Use when $ estimates not required (can estimate $ exposure to some extent)

. Use on compliance tests

. If controls are weak, do substantive testing, not compliance

. Cannot use if controls do not leave an audit trail

FIXED-SAMPLE-SIZE ATTRIBUTE SAMPLING (attribute estimation):

. Test for compliance with controls by estimating deviations

. Examples: inventory; cash receipts; cash disbursements; payroll

. Use when expect errors in N

. Estimate Rate of Occurrences in Population

. Estimate (& calculate) Achieved Upper Precision Limits of Occurrences for attribute tested based on results of sample

STOP-OR-GO (sequential attribute sampling)

. Test for compliance with controls when estimated deviation rate not known

. Examples: inventory, cash receipts; cash disbursements; payroll

. Expect zero/few errors in population

. Efficient

. Determine if projected error is > predefined error

. Stop sampling if zero deviations found; # deviations within limits based upon defined CL and acceptable UPL

DISCOVERY

. Expect no/extremely low errors in population: fraud, ghosts, etc.

. Examples: confirm account balances when have strong IC, expect few errors (banks, etc.)

. Sample sizes are large so at least one deviation will be revealed

. Find one deviation; doesn't tell you much

. Useful in substantive testing (looking for one material error) & special studies