Lessons · Accounting · a screening tool, and the data it does not work on
A screening tool, and the data it does not work on
In many naturally occurring sets of numbers, leading digits are not evenly distributed: small digits appear far more often than large ones. A large departure from that pattern is a reason to look closer. It is a screen, never proof, and it only holds for data of the right kind.
Hone is a place to practise a career, one idea a day. This is one of its lessons, written out in full and free to read without an account.
In beta. This lesson was written for Hone and has not yet been checked by a qualified accountant. Practice material, not professional advice. What that means.
What it is for
It is the most misused technique in this field. It does not apply to assigned numbers such as invoice or account numbers, to data confined to a narrow range, or to data with a built-in floor or ceiling. Run it on the wrong data and you get a result, because you always get a result, and a confident conclusion from it is expensive for whoever is on the other end.
How to think about it
First decide whether the data qualifies: naturally occurring, spanning several orders of magnitude, not assigned and not bounded by a rule. Only then run it, and treat any departure as a place to look rather than as a finding.
Worked example
Naturally occurring amounts across a wide rangeThe kind of data it holds for.
Invoice numbers, account numbers, identifiersAssigned. It does not apply.
Expenses under a policy limitA ceiling. It does not apply.
A departure is a place to lookNever a conclusion on its own.
Your turn
Name what a departure from the expected pattern gives you.
A departure is a place to , not a finding
Solve one, graded on the server
The trap
Running it on data with a policy ceiling and reading the spike as fraud. The spike is caused by the rule, and the pile of claims just under the limit is a real signal found a completely different way, which the next lesson is about.