Identifying Anomalies Using the Relative Size Factor Test
The previous chapter introduced tests to identify abnormally large subsets and subsets that had experienced explosive growth. The tests concluded with a test of the dollar totals for all the days in a year. The focus in the previous chapter was on size. In this chapter we compare large amounts to a benchmark to see how large they are relative to some norm, hence the name the relative size factor test. The relative size factor test is a powerful test for detecting errors. The test identifies subsets where the largest amount is out of line with the other amounts for that subset. This difference could be because the largest record either (a) actually belongs to another subset, or (b) belongs to the subset in question, but the numeric amount is incorrectly recorded.
The relative size factor (RSF) test is an important error-detecting test. An airline auditor reported at an IATA conference that his airline had found errors that amounted to around $1 million as a result of running this test on their accounts payable data. This test was developed in the mid-1990s after I learned of a case where a company in Cleveland wired $600,000 in error to the bank account of a charity. The $600,000 was supposed to have gone to a vendor. Once the “wrong bank account” error was discovered the company contacted the charity, which claimed that the money had already been spent and was largely unrecoverable. The $600,000 was significantly ...