Another Sample Size Q. 2183

  • Hey fellow SOXers…%0AThis is the current situation I’m in:%0ADaily IT Control%0AProjected sample size for year:25%0ACurrent samples tested as of Q3: 6 (entire population)%0AUnfortunately, my performance as a tester is based on % of work complete. So in this situation I’m not even close obtaining required %. This was raised as an issue. After long deliberations my manager came up with an idea of adjusting my projected sample size to 20% of entire population available. So doing the math this would yield annual population of 10, with me required to do only 2 to declare the control effective or ineffective. Great news, less work, right? NO because as an auditor I feel unsure about making my decision about the operativeness of control based only on two samples.%0ASo far I have been trying to make an argument to leave 25 as it is in projections. I’m totally fine with testing as many samples as possibly close to 25, however as it is management only cares about the final number (above mentioned percentage) to be met.%0AIt’s a pickle… any input on this situation this will be appreciated.%0AThanks.

    1. You can’t test any more than 100% and any performance assessment should be sufficiently flexible to show what you have done to date as being on-track.%0A2. Generally I would advocate having minimum sample sizes as well. Therefore if the inferred sample is below, say, 5 you should still test 5 or 100% of the population (as appropriate)%0A3. Based on your population size the control is obviously not a daily one - it falls more into the transactional/as required bucket. Typically these are dealt with as x% of population with minimum and maximum parameters e.g. 10% of population, minimum 5 and maximum 25

  • That makes perfect sense. Thanks Dennis…

Log in to reply