out that they informed and will be ostracized by other employees. There may be potential reprisals not only from the alleged perpetrator but also from the supervisor, who may not find the tip credible or may be involved in the fraud. Also many organizations do not have a whistleblowing or integrity hotline procedures in place to make it easy and anonymous for people to provide the tips. The 2012 ACFE Report to the Nations9 shows tips as the most common way fraud was initially detected in occupational fraud at 43.3 percent. This is an increase over the 2010 report of 40.2 percent. With whistleblowing legislation in place and organizations implementing tools to facilitate this process for people, the volume of tips will hopefully increase.
Employees are in the best place to witness and detect fraud. They are the best source for information. However, it should be recognized that some tips are provided with malicious intent. The allegations may be false and the tips provided to make trouble for the alleged perpetrator. A tip should be recognized as merely a red flag or fraud symptom and require investigation like any other symptom. An open mind and professional skepticism are needed.
Behavioral and lifestyle changes are another area where employees are best positioned to observe these anomalies. Auditors would likely have no base to compare changes to as they do not know the employee, whereas coworkers see and interact with other employees on a day-to-day basis. Lifestyle changes would be obvious to coworkers. While observed assets can be easily explained away by way of lottery winnings, inheritance, or disposition of investments, the explanation can be just as easily verified.
Similarly, behavioral changes – whether detrimental or good – are best noticed by other staff members. Perpetrating fraud is a stressful action that involves a fear of being caught. The stress triggers unusual behaviour that should be looked into by the organization. This may be out of concern for the employee’s physical and mental health, as well as to determine whether it impairs the organization in its day-to-day operation. The employee may be dealing with personal issues that are causing changes in behavior. This proactive approach is beneficial to the employee and may reduce one of the pressures contributing to committing fraud.
BusinessDictionary.com defines data mining as:
sifting through very large amounts of data for useful information. Data mining uses artificial intelligence techniques, neural networks, and advanced statistical tools (such as cluster analysis) to reveal trends, patterns, and relationships, which might otherwise have remained undetected.10
Data mining is the searching of large amounts of computerized data to find trends, patterns, or relationships without testing a hypothesis. No specific results or outcomes are anticipated.
BusinessDictionary.com defines data analysis as:
the process of evaluating data using analytical and logical reasoning to examine each component of the data provided. This form of analysis is just one of the many steps that must be completed when conducting a research experiment. Data from various sources is gathered, reviewed, and then analyzed to form some sort of finding or conclusion. There are a variety of specific data analysis method, some of which include data mining, text analytics, business intelligence, and data visualizations.11
Data mining is a subset of data analysis or analytics. Data analytics starts with a hypothesis that is to be confirmed or proven to be false. A conclusion is made based on inference from the findings.
The definition and types of data analytics can be further refined as to exploratory data analysis (EDA), confirmatory data analysis (CDA), and qualitative data analysis (QDA).
EDA is the initial stage where the data is explored when little is known about the data’s relationships. It is here those hypotheses are formed and new patterns of features of the data are discovered. Most EDA techniques are visual and graphical. They consist of plotting the data in various types of statistical graphs to obtain an insight to the data.
CDA is where testing takes place and the hypotheses are proven correct or false. Results from samples are applied to the entire database. Causal or cause-and-effect relationships are verified. A cause-and-effect relationship is where one variable is independent and the other dependent. That is, the cause is the independent variable that impacts the dependent effect. An example would be citing the amount of rainfall as causing the growth of grass; care must be taken, as many events appear to be associated but may not actually have a cause-and-effect relationship.
Online analytical processing (OLAP) tools are frequently used with the CDA process. They allow the user to extract data selectively and view the data from different perspectives or dimensions.
QDA is used to draw conclusions from nonquantitative or non-numerical data such as images or text. While typically employed in the social sciences, it can be used in organizational audits of controls, procedures, and processes.
Data analytics provide insight into the dataset, discover underlying data relationships and structures, test assumptions and hypothesis, identify variables of causal relationships, and detect anomalies.
There are a number of software programs that analyze data. Software such as Microsoft Access12 or Microsoft Excel13 is familiar to many people and used by many businesses and individuals. Indeed, Excel is favored and frequently use by accountants and auditors. Access and Excel is suitable when the dataset is not large and the analysis not complex. While it is possible to do more complex procedures, many steps are necessary. The user may need to perform operations and formulas that are not commonly used.
These products are not recommended as professional analytic tools; their complex functions are time consuming to learn and lack data integrity. It is easy to inadvertently change the content of cells by accidentally touching the wrong key. Processing speed is also slow and can be cumbersome when applied to large amounts of data.
Professional or dedicated data analysis software, such as ACL,14 Arbutus,15 and IDEA16 are specifically designed for use with large and very large data sets. Features of this type of software include:
• The data source is protected
• Can provide quick analysis
• Retains audit trails
• Built-in data analytical functions
• User friendly
• Can import from various data sources and file formats
• Able to analyze 100 percent of transactions
• Field statistics
• Various types of sampling techniques
• Benford’s Law analysis
• Correlation and trend analysis
• Drill-down features
• Aging
• Stratification
• Fuzzy matching
• Sophisticated duplicate testing
• Auto-run or automated procedures
ActiveData17 is an Excel add-in that has data analytical capabilities. It is a cross between Excel/Access and the more powerful data analytical software. It is feature-rich with an attractive low price.
Data anomalies are a fact of life. There will always be inconsistencies, abnormal, or incorrect