Exam 1 Flashcards
What is the approach of accounting data analytics?
Data analytics can assist in uncovering unseen patterns and knowledge in vast quantities of data and software. Accounting data analytics is a subset of data analytics focusing on: transactions, examines data for fraud and errors, facilitates the audit function
Fraud
The material, purposeful, misrepresentation of information characteristics, in order to gain assets, leverage or anything of value from another
Occupational Fraud
Fraud in the workplace
Abuse
Inappropriate workplace behavior or events to which the definition of fraud does not fit in
Anomalies
Unusual occurrences or data values
Broad Categories of fraud types
- Misappropriation of assets
- Financial statement fraud
3 Corruption Fraud
- What are the approaches by which organizations measure and address risk?
- Avoiding
- Bearing
- Mitigating
- Share
Traditional Fraud Detection approach
=> Do not directly observe occupational fraud
=> Observe indicators, symptoms or red flags of occupational fraud
=> These observations are referred to as anomalies
=> Investigate the anomalies to see if fraud actually occurred
Standard method is to identify a group of anomalies, examine a few transaction in the group to clear the entire group
Significant number of false positives occur among these anomalies
Due to the number of false positives, often do not get the attention these deserve
Accounting Data Analytics approach
=> Use computers and specialized software to examine each transaction
=> Screen all the transactions to identify anomalies
Using specialized algorithms
=> Investigate the anomalies for fraud and errors
Can examine all or most of the anomalies
Describe the Data Analysis Cycle
Three stages of the data analysis cycle
- Evaluation and analysis
- Software and technology
- Audit and investigation
Types of Data
Categorical data
Characteristics or qualities of the data
Can use pivot tables to examine two-way categorization of the data
Quantitative Data
Numeric data
Numbers, values, counts
Types of numeric data
Cardinal data: numeric data with a natural zero
Score data: numeric data without a natural zero
Ordinal data: ranks, categories
May not perform arithmetic operations
Concepts underlying inferential statistics
=>The population is too large or costly to enumerate
=>Draw a sample that represents the population
Representation is the key
=>Calculate statistics of interest in the sample
Infer these to values to the corresponding population statistic
=>With data analytics the entire population may be examined
Search for anomalies
Inspect the anomalies (like a sample from the population)
Measures of central tendency
Mean
Median
Mode
Measures of Dispersion
=>Range =>Deviations from the mean =>Variance =>Standard Deviation =>Standard Deviation of a sample and a population
Sampling
=>Sampling is the act of selecting items from a population
=>Use the sample to represent the population
=>Draw inferences about the population based on the sample
Sampling is identifying the anomalies in the data set of transactions
Study the anomalies for errors and fraud =>In an audit context
Sample transactions across a variety of transaction types
Stratified sampling
Statistical Sampling
=>Use of mathematical calculations for selecting and analyzing a data sample
Selection of items in the sample
Calculation of sample parameters
Determination of precision levels
Non-Statistical Sampling & Methods
=>Often referred to as judgement sampling
=>Researcher knows the population and the data set
=>Select the transactions in the sample based on the researcher’s knowledge
Methods
=> systematic
=>stratified
=> cluster or block
=>Use if these methods provide a more representative sample
Requires researcher or analyst special knowledge
=>Can make into a random sample by
Selecting a random starting point in a systematic sample
Selecting a random sample within each strata for a stratified sample
Randomly select a cluster or block and include all items in the cluster or block
Statistical Sampling Methods of Auditors
=>Probability Proportional-to-Size Sampling
Sampling unit is dollars
Monetary Unit Sampling (MUS)
=>Transactions with higher monetary value more likely to be selected in sample
Rather than frequency of occurrence
=>Use to determine accuracy of financial transactions when size of transaction is most important
=>Errors expected to be few
What is the general process of accounting data analytics tests?
=>Start with general or initial tests to identify anomalies
=>Once anomalies (or outliers or suspicious patterns) are identified
Partition the data to these anomalies (AKA data reduction)
Perform additional tests on the data partition or anomalies for further investigation
Describe Benford’s Law in concept and its use in accounting data analytics
=> Benford’s law analyzes the digits in numerical data, helps identify anomalies, and detects systematic manipulation of data based on the digital distribution in a natural population.
=>Based on observation of the frequency of lead digits in numbers
Empirical rule
Examines the first or first two digits in numbers
=>The lower the digit, the more frequently it occurs
=>Creates a monotonic downward sloping curve when graphed with frequency on the vertical axis and the lead number or numbers on the horizontal axis
What are the “standard” Benford’s Law and what are the objectives of each?
=>First digit
Suitable for use on data sets of less than 300 transactions
=>First two digits
Most practical and used of Benford’s Law tests
=>First three digits
Very detailed and tends to produce many anomalies
May identify too many anomalies to be useful
Last Two Digits
Apply Benford’s Law to the last two digits of numbers in a data set
Number Duplication
=>See if some digits occur too often in a data set
=>Too frequently occurring numbers may indicate fraud or error
