The Audit Process: Use of Statistics (18)

Question 1

Mean

Answer 1

Average of numerical dataset - sum of values divided by number of items, denoted with μ

Question 2

Median

Answer 2

Central point in the dataset when arranged in ascending order

Question 3

Mode

Answer 3

Most frequently occurring number in dataset

Question 4

Outlier

Answer 4

Data point significantly different from other data points in dataset (might indicate misstatement)

Question 5

Standard Deviation

Answer 5

Measure of dispersion of dataset showing average distance between values and the mean, denoted with σ. Datasets with a wider spread will have a higher standard deviation.

Question 6

Benford’s Law

Answer 6

Probability distribution of likelihood of first digit in set of numbers (1 is most common first digit in a dataset, descending to 9 - requires large dataset)

Question 7

Regression analysis

Answer 7

A statistical method for estimating the relationship among variables based on past relationship.

The aim is to understand the relationship between an independent variable and a dependent variable and is often performed through ADA.

Question 8

What happens when correlation is 0?

Answer 8

There is no correlation between the two factors.

Question 9

Dependent variable

Answer 9

The variable we are trying to understand or predict

Question 10

Independent variable

Answer 10

The variable that may impact dependent variable

Question 11

Regression line

Answer 11

Statistically determined relationship between dependent and independent variables in historical data that we can compare new data against to identify anomalies.

Question 12

How can Benford’s Law be applied within the audit process?

Answer 12

It can be used to identify where anomalies (including incidents of fraud) appear in the dataset. Allowing the auditor to identify a misstatement in relation to error or fraud during the audit process.

Question 13

How can regression analysis be applied within the audit process?

Answer 13

Historical data can be collected from a client’s (financial data and non-financial data) system, allowing the auditor to understand a historical relationship between variables.

The variables will be plotted on a scatter chart and a line (regression line) will be drawn through the middle.

The current year data can be compared with this line and any outliers that appear may indicate misstatements in the dependent variable being audited (Sales being overstated or cost of sales being understated).

Question 14

Correlation

Answer 14

Correlation is the statistical association or relationship between two variables. Calculated between +1 and -1.
Correlation of 0 indicates no relationship.
Correlation of +1 indicates a strong relationship
Correlation of -1 indicates an inverse relationship.

Question 15

Correlation vs Causation

Answer 15

Important to remember that correlation does not necessarily imply causation.

Both variables could alter dependent on a shared independent variable i.e. sales of Christmas trees and mulled wine are correlated but there is no causation between them.

Question 16

Statistical Sampling

Answer 16

Statistical sampling is defined in ISA (UK) 530 Audit sampling with the following characteristics:

i. Random selection of sample items
ii. The use of probability theory to evaluate sample results, including the measurement of sampling risk.

A sampling approach that does not have the above characteristics is considered non-statistical sampling

Question 17

Advantages of Statistical Sampling

Answer 17

1. Can allow the auditor to select a more targeted and efficient sample
2. It allows a measure of the sufficiency of the audit evidence obtained.
3. The use of statistics can reduce the risk that differences in audit judgement (non-statistical sampling) result in significant differences in sample sizes selected by different auditors.
4. It allows for errors identified in the sample to be quantified and extrapolated to the full population.

Question 18

Disadvantages of Statistical Sampling

Answer 18

Where data is not available in electronic format, statistical sampling may not be an efficient way of obtaining an appropriate sample.

Statistical sampling requires additional expertise within an audit team, or software provided to an audit firm, to obtain a sample.

Question 19

Random Seed

Answer 19

The starting point when generating a random sample. Random seeds are often produced using a random number generator to ensure the auditor has not influenced the sample by choosing a starting point.