Beneish M score model Scopus dataset
Description
Research questions / hypotheses This is a bibliometric , the research was guided by four research questions: What are the publication and citation characteristics of research on the Beneish M-Score model in financial statement fraud detection? Which authors, journals, and publications have made the greatest intellectual contribution to the field? What are the dominant research themes, and how have they evolved over time? What research gaps and future research opportunities can be identified from the existing literature? What the data shows / notable findings The dataset shows a small but fast-growing research field: publications on the Beneish M-Score are increasing at roughly 15.76% annually, with an average of 2.93 authors per document and relatively limited international co-authorship (18.31%). Beneish's original 1999 paper, together with foundational agency-theory and earnings-management literature, How the data was gathered The dataset was compiled through a structured search of the Scopus database (the sole data source), using a Boolean search string combining terms for the Beneish M-Score model and financial statement fraud detection. The search covered 2015–2026 and was restricted to peer-reviewed journal articles and conference proceedings. Records were screened in five stages (Boolean search → subject-area filtering → document-type filtering to journal articles/reviews → no geographic restriction → English-language restriction), following a PRISMA-style selection process, resulting in a final dataset of 71 documents drawn from 55 sources, authored by 195 individual contributors, and supported by 3,622 cited references. The full search string and screening criteria are reported in Section 3.2 of the accompanying article. What's in the dataset The file contains the bibliographic metadata exported from Scopus for the 71 included documents: author names and affiliations, article title, source/journal title, publication year, abstract, author keywords, citation counts, and cited references. This is the raw input used for the bibliometric analysis conducted in Bibliometrix (Biblioshiny) and VOSviewer How to interpret and reuse the data Citation counts reflect totals at the time of Scopus extraction and will change over time as the literature continues to be cited — they should not be treated as fixed values. Co-authorship and country-collaboration figures are based on author-affiliation metadata as recorded in Scopus and may undercount collaboration where affiliations are incomplete or inconsistently indexed. The dataset is restricted to Scopus-indexed, English-language, peer-reviewed journal articles and reviews; it excludes books, book chapters, conference proceedings not indexed as such, and non-English publications, so findings should be interpreted as representative of the indexed.
Files
Steps to reproduce
This study utilised the Scopus database as the sole source of bibliographic data. Scopus was selected because it is one of the largest abstract and citation databases and, provides extensive coverage of peer-reviewed journals in accounting, finance, business, economics, and related disciplines. The database offers comprehensive bibliographic metadata, including citation information, author affiliations, institutional data, keywords, funding information, and references, making it particularly suitable for bibliometric analysis. Its widespread use in bibliometric research also facilitates comparison with previous studies .Bibliographic records relating to the Beneish M-Score model and financial statement fraud detection were retrieved from Scopus using a structured search strategy. The search string combined keywords associated with the Beneish M-Score model and financial statement fraud detection using Boolean operators: (("Beneish M-Score" OR "Beneish M Score" OR "Beneish model" OR "Beneish's M-Score”) AND ("financial statement fraud" OR "financial fraud “OR "fraud detection “OR "fraudulent financial reporting “OR "financial reporting fraud" OR "financial statement manipulation" OR "earnings manipulation")). The search covered the period 2015 to 2026 and was restricted to peer-reviewed journal articles and conference proceedings. Following the removal of duplicated records and records with incomplete biographic metadata, the final dataset consisted of 71 documents published across 55 sources, authored by 195 individual contributors, and supported by 3622 cited references.
Institutions
- Durban University of TechnologyKwaZulu-Natal, Durban