DIGITAL
ANALYSIS TESTS AND STATISTICSâ
DATASâ
Using
Digit and Number Patterns to Detect Fraud,
Errors,
Biases, Irregularities, and Processing Inefficiencies, AND to
Test New Computer Systems
|
DATASâ 2000 FOR SAS:
DOCUMENTATION
|
Overview
|
1
|
|
Installing the
Programs
|
2
|
|
Preparing a File for
Analysis
|
3
|
|
Data Profile
|
4
|
|
Basic Digit Tests
|
4
|
|
Number Duplication
|
4
|
|
Last-Two Digits
|
8
|
|
Round Numbers
|
9
|
|
Advanced Tests
|
10
|
|
Profile Details
|
10
|
|
Number Frequency Factor
|
10
|
|
Number Duplication Two
|
12
|
|
Relative Size Factor
|
13
|
|
Round Number Subsets
|
13
|
|
Two Subsets (Same-Same-Different)
|
14
|
|
Three Subsets (Same-Same-Same-Different)
|
16
|
|
Three Alike (Same-Same-Same)
|
16
|
|
Four Alike
(Same-Same-Same-Same)
|
16
|
|
New System and Other
Tests
|
17
|
|
My Law
|
17
|
|
Summation Test
|
17
|
|
Writing a File to Excel-readable Format
|
18
|
OVERVIEW
Digital
Analysis is a new audit technology used by auditors to test the validity
of large corporate data sets. With
the ever increasing size of corporate data sets and the competitive need
to be more effective and efficient, auditors have welcomed the new
technology. The tests begin
with high level overview tests and drill deeper and deeper searching for
abnormal duplications of digits, digit combinations, and specific numbers. Statistical theory developed by a physicist, Frank Benford,
60 years ago determines what is normal duplication, and what is abnormal
duplication. The final result
is an audit sample focused on abnormal recurrences of digits, digit
combinations, specific numbers, and round numbers.
The technology could steer you towards,
Finding
fraud (where the fraudster has excessively duplicated their actions)
Finding
errors (where the error is systematic)
Finding
errors (based on Advanced duplication seeking formulas)
Finding
outliers (again based on Advanced anomalous transaction seeking formulas)
Finding
biases (where managers have many purchases just below their authorization
level)
Finding
processing inefficiencies (where there are numerous small invoices from
certain vendors)
A
recent use of Digital Analysis is for continuous audit and systems
improvement, and by value-added auditors seeking to find inconsistencies
in revenue data sets.
The DATASâ
Digital Analysis tests are grouped as follows:
1. Basic Tests,
2. Advanced Tests, and
3. New System & Other Tests.
The Basic
Tests analyze a single field of numbers which is usually dollar
amounts. The Advanced
Tests analyze a field of values plus one or more other fields called
subset fields. The New System & Other Tests are tests that cannot be classified as
Basic or Advanced. DATASâ
includes a test data set, invoices.csv, with 38,176 observations and five
fields for practice.
The DATASâ
suite of programs consists of:
DATASâ
2000 for ACL
DATASâ
2000 for Excel
DATASâ
2000 for IDEA
DATASâ
2000 for SAS
The
DATASâ
suite of programs all have approximately the same functionality.
The differences in functionality and output are minor.
DATASâ
2000 for SAS has a few advantages over the other DATASâ
programs. The programs are
usable on both the PC and the mainframe. Large data sets can be analyzed.
Also, the code is visible making the programs customizable.
Trademarks:
Digital
Analysis Tests and Statistics, DATAS, and the stylized 123456789 logo are
all trademarks of Mark J. Nigrini, 606 Rockcrossing Lane, Allen, Texas,
75002, USA.
ACL is a
registered trademark of ACL Services Ltd., Vancouver, Canada.
IDEA is a
registered trademark of The Canadian Institute of Chartered Accountants,
Toronto.
SAS is a
registered trademark of SAS Institute, Cary, NC, USA.
Copyright
ã
2000 by Mark J. Nigrini. All
rights reserved.
Mark J.
Nigrini
(972)
359-0020, mark_nigrini@msn.com
|
