The paper, which has been cited more than 6,000 times, has fatal flaws, but Wall Street executives, government officials, and even a former US vice president have referenced it, and the academic community claims to take no action.
Andrew King of Boston University and Andrew Gelman of Columbia University criticized the authors of the highly cited paper, alleging that it was flawed and that some of the results were not reproducible. King and his colleagues also blamed the academic community for not responding appropriately.
This paper in Management Science has been cited more than 6,000 times. Wall Street executives, top government officials, and even a former US Vice President have all referenced it. It's fatally flawed, and the scholarly community refuses to do anything about it. | Statistical Modeling, Causal Inference, and Social Science
The paper King and his colleagues chose to discuss was ' The Impact of Corporate Sustainability on Organizational Processes and Performance,' written by Robert Eccles et al. The paper states that 'companies that prioritize sustainability significantly outperform their competitors in both stock market and accounting performance.' The paper has been cited over 6,000 times in the past to highlight the significance of sustainability.
However, the paper was found to have several flaws, including 'important results were falsely declared significant when they were not,' 'some measurements were irreproducible,' 'important statistical tests appeared to be missing,' and 'the sample was highly unusual.'
In 2023, King pointed out the mistake to the authors, but was ignored. He then sought advice from colleagues and respected scholars, but was brushed off with comments like, 'We can't do anything because it would lead to conflict.' He then contacted the journal Management Science, which published the paper, but his appeal was rejected.
So King decided to go beyond the usual process and report the issue on LinkedIn, which prompted the authors to respond and publish a correction in Management Science. King then published a replication study in the Journal of Management Scientific Reports, censoring the authors.
'In the process of revising the manuscript for the replication study, I realized that the methods reported in the original paper were not the methods actually used, and that the true methods could not support the authors' 'findings.'' King decided that a research ethics complaint was necessary and filed it with the Harvard Business School and the London Business School, where the authors are affiliated, arguing that the reported methods could not be performed as described and therefore the results were uninterpretable.
Shortly after this was pointed out, the authors admitted to misreporting the methodology, but did not correct the error. They responded, 'The section was removed during the writing process, but the incorrect statement was inadvertently included and was unintentional.' However, King pointed out, 'The error appears multiple times in early and subsequent drafts, and there appears to have been no attempt to remove it. It is unclear whether the research referenced in the paper actually existed.'
In response to King's complaint, Harvard Business School said it would not disclose how it will address the issue. London Business School responded, saying the issue is minor because the authors did not have access to the raw data and did not fabricate it. King also criticized the issue, saying, 'The false claims are not minor; they are material differences that affect the usefulness of the research.'
King and his colleagues point out that the paper in question has only been partially corrected in Management Science, leaving thousands of readers misled. They also point out that journals should disclose comments, complaints, corrections, and requests for retraction. They also point out that universities should report complaints about research integrity and their outcomes.
In addition, King and his colleagues argue that even peer-reviewed papers will contain errors, and they suggest that we stop citing single studies as definitive and publish corrections when we find errors in publications.
Related Posts:
in Note, Posted by log1p_kr