Baker made several mistakes in the course of his audit. One of those mistakes was discounting the materiality of some of the errors. For example, he decided that the inversion of the numbers in $10,443 and $10,434 was immaterial and did not count this as a deviation. All deviations should be counted, regardless of size. A deviation of this type, for example, could result in a major error depending on which of the numbers are inverted. Defects can be classed according to their seriousness, but all defects must be recorded (Schilling, 1982).
Discovery sampling is not an appropriate method for this type of audit. What discovery sampling is best used for is for situations where one deviation would require an intensive investigation, such as when fraud is suspected (Kaplan, 2003). However, Baker actually used acceptance sampling (albeit incorrectly with his decision to reject the deviation noted above).
Baker automatically increased the sample size to 100 documents when the population size was increased. This was incorrect. He should have recalculated the sample size again rather than assume that the sample/population ratio was scalable.
Baker also erred when he made the assumption that the sample deviation rate and allowance for sampling risk could be added and compared with the achieved upper deviation rate. This is incorrect -- the calculation should have been made differently. Indeed, his calculation of the allowance for sampling risk was done incorrectly. That is the difference between the expected error rate and the tolerable deviation rate, not the expected error rate and the actual error rate (Accounting Institute, 2009).
Works Cited:
Schilling, E. (1982). Acceptance sampling in quality control. Boca Raton: CRC Press.
Kaplan, J. (2003). How to use statistical sampling. AuditNet. Retrieved February 4, 2010 from http://www.auditnet.org/Guides/AuditNet%20Monograph%20Series%20Audit%20Sampling.pdf
No author. (2009). Auditing dictionary of terms. Accounting Institute. Retrieved February 4, 2010 from http://www.ais-cpa.com/glosa.html
