Full Text Available
Note: Clicking the button above will open the full text document at the original institutional repository in a new window.
An investigation of the effectiveness of a two-stage risk-based ranked-set sampling approach in external audits : a substantive audit approach perspective
Thesis (PhD)--Stellenbosch University, 2025.
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Language: | en_ZA |
| Published: |
Stellenbosch : Stellenbosch University
2025
|
| Subjects: | |
| Tags: |
No Tags, Be the first to tag this record!
|
| _version_ | 1867613800788656128 |
|---|---|
| access_status_str | Open Access |
| author | Swanepoel, Elmarie |
| author2 | Von Wielligh, Simon Petrus Johannes |
| author_browse | Swanepoel, Elmarie Von Wielligh, Simon Petrus Johannes |
| author_facet | Von Wielligh, Simon Petrus Johannes Swanepoel, Elmarie |
| author_sort | Swanepoel, Elmarie |
| collection | Thesis |
| dc_rights_str_mv | Stellenbosch University |
| description | Thesis (PhD)--Stellenbosch University, 2025. |
| format | Thesis |
| id | oai:scholar.sun.ac.za:10019.1/133242 |
| institution | Stellenbosch University (South Africa) |
| language | en_ZA |
| last_indexed | 2026-06-10T12:41:54.752Z |
| license_str | Other — see source repository |
| provenance_str_mv | Harvested via OAI-PMH from SUNScholar — Stellenbosch University Repository |
| publishDate | 2025 |
| publishDateRange | 2025 |
| publishDateSort | 2025 |
| publisher | Stellenbosch : Stellenbosch University |
| publisherStr | Stellenbosch : Stellenbosch University |
| record_format | dspace |
| source_str | SUNScholar — Stellenbosch University Repository |
| spelling | oai:scholar.sun.ac.za:10019.1/133242 An investigation of the effectiveness of a two-stage risk-based ranked-set sampling approach in external audits : a substantive audit approach perspective Swanepoel, Elmarie Von Wielligh, Simon Petrus Johannes Steel, Sarel Johannes Stellenbosch University. Faculty of Economic and Management Sciences. School of Accountancy. Logistic regression analysis Decision making -- Statistical methods Auditing -- Statistical methods Risk assessment -- Mathematical models UCTD Thesis (PhD)--Stellenbosch University, 2025. Swanepoel, E. 2025. An investigation of the effectiveness of a two-stage risk-based ranked-set sampling approach in external audits: a substantive audit approach perspective. Unpublished doctoral dissertation. Stellenbosch: Stellenbosch University [online]. Available: https://scholar.sun.ac.za/items/2d4fd1b2-0097-4717-9190-22c491b39dd3 ENGLISH SUMMARY: The global auditing profession has faced significant criticism due to high-profile audit failures in recent years. Investigations by governments, regulatory bodies, and standard setters have identified the use of sampling in the audit evidence-gathering process as one of the root causes of these failures. While emerging technologies hold the promise of eliminating the need for sampling, they have not yet reached the necessary maturity level. This study aimed to use technology at its current maturity level to integrate risk-assessment procedures and audit sampling as a single construct on a scalable basis. This study started by investigating the behaviour and performance of existing audit sampling approaches through simulation by focusing on two widely used methods: simple random sampling (SRS) and the monetary unit sampling (MUS) approach. The simulation studies focused on tests of details performed on a class of transactions (revenue). Interestingly, existing literature lacks a comprehensive performance evaluation approach for audit sampling methods. The study introduced a standardised performance evaluation model that combines individual performance measures related to tightness, variability, and decision accuracy into a single performance score to address the identified gap in the literature. The simulation study showed that MUS is more sensitive to changes in population characteristics than SRS. A second simulation study provided insight into the performance of SRS and MUS when selecting items from a subset of high-risk items identified by automated risk-classifying procedures. Notably, an optimal risk-threshold limit exists and setting it beyond this level will negatively affect the performance of both SRS and MUS. The study also introduces a novel audit sampling approach called risk-based extreme ranked-set sampling (RSSext). Adapting traditional ranked-set sampling (RSS), which inherently stratifies the population based on a specific variable (the concomitant variable), RSSext uses the variable that contains the results of automated risk-assessment procedures as the concomitant variable. A further adaptation ensures that line items with higher estimated risks have a higher probability of inclusion. A second step was added to RSSext to enhance its robustness, which resulted in a novel two-stage risk-based ranked-set sampling approach (RSStwo). In the first stage, a training dataset is created from an initial sample to fit a logistic regression model. This model refines the estimated risk of misstatement for each population line item. In the second stage, a sample is selected using the RSSext approach, and the results are projected to the population. In the third simulation study, RSStwo outperformed single-stage approaches, including RSSext, which makes it a feasible alternative to existing sampling methods. This research contributes significantly to the understanding of the existing audit sampling methods, their limitations, and the potential of integrated risk-based sampling approaches. The findings have important implications for auditors, regulators, and standard setters in their efforts to improve audit quality. AFRIKAANSE OPSOMMING: Die ouditprofessie is die afgelope paar jaar aansienlik gekritiseer weens hoe-profiel oudit mislukkings. Ondersoeke wat deur regerings, regulerende liggame en standaardstellers onderneem is, het bevind dat die gebruik van ouditsteekproefneming een van die hoofoorsake van hierdie mislukkings is. Opkomende tegnologiee hou die belofte in om die gebruik van steekproefneming uit te skakel, maar is nog nie ver genoeg gevorder nie. Die doel van hierdie studie is om tegnologiee wat tans beskikbaar is te gebruik om ouditrisiko beoordelingsprosedures en ouditsteekproefneming op ‘n skaalbare basis te integreer. Die studie is begin deur die gedrag en prestasie van die bestaande ouditsteekproefneming benaderings deur simulasie te ondersoek, met die fokus op twee metodes wat algemeen gebruik word, naamlik: eenvoudige ewekansige steekproefneming (SRS) en monetere eenheid steekproefneming (MUS). Die simulasie studies fokus op stawingsprosedures wat op 'n spesifieke klas van transaksies (inkomste) uitgevoer word. 'n Omvattende prestasieevaluerings benadering vir ouditsteekproefneming kon nie in die bestaande literatuur gevind word nie. Hierdie gaping in die literatuur is aangespreek deur 'n gestandaardiseerde prestasie-evalueringsmodel te ontwikkel. Die model kombineer die resultate van die volgende individuele prestasiemaatstawwe in ‘n enkele prestasie punt: die akkuraatheid van besluitneming, akkuraatheid en graad van variasie van geprojekteerde foute. Die simulasie studie het getoon dat MUS meer sensitief is vir veranderinge in die eienskappe van ‘n populasie as SRS. 'n Tweede simulasie studie het insig verskaf oor die prestasie van SRS en MUS indien steekproefitems gekies word uit 'n groep items wat deur geoutomatiseerde risiko klassifisering prosedures geidentifiseer is as items met ‘n hoe risiko van wanvoorstelling. Die studie het gevind dat daar ‘n optimale risiko-drempel bestaan. Indien die risiko-drempel wat gebruik word hoer is as hierdie optimale vlak, word die prestasie van beide SRS en MUS negatief beinvloed. Die studie stel ook 'n nuwe ouditsteekproefneming benadering bekend, genaamd “risk-based extreme ranked-set sampling” (RSSext). Tradisionele “ranked-set sampling” (RSS) stratifiseer die populasie inherent op grond 'n gespesifiseerde veranderlike (meegaande veranderlike). RSSext pas RSS aan om die geoutomatiseerde risiko klassifisering prosedures se risiko van wanvoorstelling as die meegaande veranderlike te gebruik. 'n Verdere aanpassing verseker dat lynitems met hoer risiko's 'n groter waarskynlikheid van insluiting het. 'n Tweede stap is by RSSext gevoeg om die robuustheid daarvan te verbeter, wat gelei het tot 'n nuwe benadering, genaamd “two-stage risk-based ranked-set sampling” (RSStwo). In die eerste fase word 'n opleidings-datastel geskep uit 'n aanvanklike steekproef om 'n logistiese regressiemodel te pas. Hierdie model verfyn die beraamde risiko van wanvoorstelling vir elke populasie lynitem. In die tweede fase word 'n steekproef gekies deur die RSSext-benadering te gebruik, en die resultate na die populasie te projekteer. In die derde simulasie studie het RSStwo beter gevaar as die benaderings wat slegs uit een fase bestaan. RSStwo is dus 'n moontlike alternatief vir bestaande steekproef metodes. Hierdie studie dra aansienlik by tot begrip van die bestaande ouditsteekproefneming benaderings, hul beperkings, en die potensiaal van geintegreerde risiko-gebaseerde steekproefneming benaderings. Die bevindinge het belangrike implikasies vir ouditeure, reguleerders en standaardstellers in hul pogings om ouditkwaliteit te verbeter. Doctoral 2025-07-29T09:30:19Z 2025-07-29T09:30:19Z 2025-03 Thesis https://scholar.sun.ac.za/handle/10019.1/133242 en_ZA Stellenbosch University xvi, 213 pages : illustrations, maps, includes annexures application/pdf Stellenbosch : Stellenbosch University |
| spellingShingle | Logistic regression analysis Decision making -- Statistical methods Auditing -- Statistical methods Risk assessment -- Mathematical models UCTD Swanepoel, Elmarie An investigation of the effectiveness of a two-stage risk-based ranked-set sampling approach in external audits : a substantive audit approach perspective |
| title | An investigation of the effectiveness of a two-stage risk-based ranked-set sampling approach in external audits : a substantive audit approach perspective |
| title_full | An investigation of the effectiveness of a two-stage risk-based ranked-set sampling approach in external audits : a substantive audit approach perspective |
| title_fullStr | An investigation of the effectiveness of a two-stage risk-based ranked-set sampling approach in external audits : a substantive audit approach perspective |
| title_full_unstemmed | An investigation of the effectiveness of a two-stage risk-based ranked-set sampling approach in external audits : a substantive audit approach perspective |
| title_short | An investigation of the effectiveness of a two-stage risk-based ranked-set sampling approach in external audits : a substantive audit approach perspective |
| title_sort | investigation of the effectiveness of a two stage risk based ranked set sampling approach in external audits a substantive audit approach perspective |
| topic | Logistic regression analysis Decision making -- Statistical methods Auditing -- Statistical methods Risk assessment -- Mathematical models UCTD |
| url | https://scholar.sun.ac.za/handle/10019.1/133242 |
| work_keys_str_mv | AT swanepoelelmarie aninvestigationoftheeffectivenessofatwostageriskbasedrankedsetsamplingapproachinexternalauditsasubstantiveauditapproachperspective AT swanepoelelmarie investigationoftheeffectivenessofatwostageriskbasedrankedsetsamplingapproachinexternalauditsasubstantiveauditapproachperspective |