This post is an extended synopsis of Derks et al. (2019). JASP for Audit: Bayesian Tools for the Auditing Practice. Preprint available on PsyArXiv: https://psyarxiv.com/9f6ub/
Abstract
Statistical theory lies at the core of many auditing guidelines and procedures. Consequently, an auditor needs easy-to-use software that implements the required statistical analyses as well as sufficient knowledge to interpret the results of these analyses. We introduce JASP for Audit (JfA), an open-source and free-of-charge module for JASP built to support the statistical aspects of an audit. Next to the frequentist methods that currently dominate the audit practice, JfA incorporates Bayesian counterparts of these methods that can improve the quality and efficiency of an audit. These Bayesian methods allow auditors to utilize the advantages of knowledge updating by accurately incorporating prior information. JfA is designed with the auditor in mind. This means that the interface is user-friendly, and directly relates to audit processes and International Standards on Auditing. JfA helps the auditor in interpreting, explaining, and reporting the analyses and leaves a transparent audit trail. In sum, JfA performs all the required statistical heavy lifting and enables auditors to plan, evaluate and interpret their statistical analysis in terms of auditing standards and using state-of-the-art Bayesian and classical techniques.
Introduction
In many countries, listed organizations are required to be audited by law. An audit is often complex, quality requirements are high, and the associated costs are substantial. The reason why financial audits are required by law in many countries is that they evaluate the extent to which an organization’s financial statements are presented fairly. It is the auditor’s task to gather appropriate and sufficient evidence to determine whether the organization’s statements contain errors that surpass a specified percentage or monetary amount, the so-called materiality. When the total error exceeds the materiality it is said that the financial statements contain material misstatement, which means that the errors in the financial statements may impact decisions of someone relying on those statements. In a nutshell, an audit is a final check before an organization releases its yearly or quarterly numbers to inform stakeholders of the organization about its financial situation.
The Audit Workflow in JASP
Fig 1. JfA’s audit workflow displayed as a progress chart, showing each stage in the workflow as a block with inputs and outputs. The circles represent the flow of data, emphasizing that the output of previous stages is used as input for the future ones. Most input options have sensible default values in accordance with the International Standards on Auditing, requiring the auditor to make adjustments only when needed.
The case of Buildit
In the case of BuildIt, the auditor wants to state that the fictional BuildIt population of approximately $1.4 million does not contain a material error, where the materiality is set at 5%, or $70161.
Fig 2. Snapshot of the output screen of JfA’s Bayesian evaluation stage containing the
evaluation summary table and an annotation explaining the evaluation process.
Below the evaluation summary table, the prior and posterior plot is displayed with a figure
caption that explains the decision process. At the bottom, the correct conclusion is
displayed.
References
Derks, K., de Swart, J., Wagenmakers, E., Wille, J., & Wetzels, R. (2019, July 18). JASP for Audit: Bayesian Tools for the Auditing Practice. Manuscript submitted for publication. PsyArXiv: psyarxiv.com/9f6ub