GB700 Unit 8 Research Integrity Analysis Example

Nine Graduates Removed: A Unit 8 Research Integrity Analysis of a Conference Paper on Tuition and Retention

[Student Name]

Purdue University Global

GB700: Doctoral Studies in Business

Unit 8 Research Integrity Analysis

[Instructor Name]

[Date]

The paper, the firm, the people and every figure are composite and illustrative; the policies, principles and sources cited are real.

The Case

A people analytics manager at a composite logistics company and a business professor at a nearby university presented a paper at a regional management conference. Using the company's records for 410 salaried employees eligible for its tuition program over four years, they reported that program participants were 35 percent less likely to resign than eligible non-participants. The paper listed three authors: the manager, the professor and the company's chief human resources officer.

A data analyst who had built the dataset later raised five concerns with the professor's university. First, in an earlier version of the analysis, participants and non-participants had shown nearly identical resignation rates. Before the final analysis, nine employees who had resigned within a month of completing their degrees were removed, with a note that they were "outliers," and the paper did not mention the removal. Second, the chief human resources officer had approved access to the data but had not contributed to the design, analysis or writing. Third, the analyst, who had cleaned and matched the records and run the first analyses, was not listed as an author or acknowledged. Fourth, a paragraph in the literature section matched a paragraph in an education benefits vendor's white paper almost word for word, with the white paper cited at the end but no quotation marks. Fifth, employees had not been told that their education and personnel records might be used for published research.

Definitions Before Judgments

The federal policy on research misconduct defines misconduct as fabrication, falsification or plagiarism in proposing, performing or reviewing research, or in reporting results (Office of Science and Technology Policy, 2000). Falsification includes "manipulating research materials, equipment, or processes, or changing or omitting data or results such that the research is not accurately represented in the research record." The policy excludes honest error and differences of opinion, and a finding requires that the conduct be committed intentionally, knowingly or recklessly. Although the policy governs federally funded research, universities commonly apply the same definitions to all research by their faculty. These definitions are applied below one problem at a time, because the problems differ in kind and seriousness.

Problem 1: Nine Employees Removed

Removing observations is not falsification in itself. Researchers routinely exclude cases for stated reasons, such as data errors. What matters is why the cases were removed, when, and whether the removal was reported. Here, nine employees were removed after an analysis showed no effect, the removal changed the result, the stated reason, "outliers," does not fit employees whose only distinguishing feature is that they resigned soon after graduating, and the paper did not disclose it. Those employees are precisely the group the Benson et al. (2004) study found most likely to leave, which makes their removal look less like cleaning and more like omitting the inconvenient part of the finding. The removal fits the policy's description of omitting data so that the research is not accurately represented. Whether it meets the intent standard depends on what the authors knew, which an inquiry would need to establish.

Problem 2: An Honorary Author

Listing the chief human resources officer as an author, when she contributed no design, analysis or writing, misrepresents responsibility for the work. It is not misconduct under the federal definition, but it violates widely accepted authorship standards, which require substantial contribution to the work and accountability for it. Steneck (2007) notes that honorary authorship dilutes the meaning of authorship and can make named individuals answerable for work they cannot defend.

Problem 3: A Missing Author

The analyst's work, building the dataset and running the first analyses, is a substantial contribution. Omitting the analyst from the byline, and not acknowledging the work at all, is a failure of credit. Like honorary authorship, it falls outside the misconduct definition but within the scope of authorship standards and institutional policies on credit.

Problem 4: Words Taken, Source Cited

Copying a paragraph nearly word for word with a citation but without quotation marks presents another's words as one's own, even though the source is credited. It meets the policy's definition of plagiarism, the appropriation of another person's words without giving appropriate credit, because credit for words requires quotation, not only citation. Its seriousness is lower than wholesale copying, since the source was named.

Citing a source credits its ideas; only quotation marks credit its words.

Problem 5: Three Principles, One Workforce

The 1979 report that still frames human research ethics in the United States sets out three duties for anyone studying people: respect their right to decide, balance harm against benefit, and distribute burdens fairly (National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research, 1979). Employees whose personnel and education records were used for a conference paper were not told and had no chance to object. Analysis done only to manage the business is generally treated differently from research meant for publication, but once the authors set out to publish, the employees became research participants. Respect for persons would require, at minimum, review by the university's institutional review board and notice to employees, and possibly consent. The employees were also in a dependent position, since their employer held the records, which makes that review more important, not less.

Why the Problems Arose Together

The five problems share a cause. The project began as internal analysis for the human resources department and turned into research for publication without anyone stopping to apply research standards. Internal analysis does not require disclosed exclusions, authorship rules or ethics review, so the habits of internal reporting carried over into a paper that needed all three. Many business research integrity problems arise at this transition, when a company analysis becomes a conference paper.

Remedies, Proportionate to Each Problem

Problem 1. The university should conduct an inquiry. The paper should be corrected to report all 410 employees and the original result, or withdrawn.

Problems 2 and 3. The byline should be corrected: the chief human resources officer removed, the analyst added or acknowledged according to her contribution.

Problem 4. The paragraph should be rewritten or quoted and the correction noted.

Problem 5. Any future use of the data should go through institutional review, with notice to employees.

Treating every problem as misconduct would be unjust to the authors; treating none as serious would be unjust to the employees and to readers. The remedies above match each response to the kind of failure it addresses.

References

Benson, G. S., Finegold, D., & Mohrman, S. A. (2004). You paid for the skills, now keep them: Tuition reimbursement and voluntary turnover. Academy of Management Journal, 47(3), 315-331. https://doi.org/10.5465/20159584

National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research. (1979). The Belmont report: Ethical principles and guidelines for the protection of human subjects of research. U.S. Department of Health, Education, and Welfare.

Office of Science and Technology Policy. (2000). Federal policy on research misconduct. Federal Register, 65(235), 76260-76264.

Steneck, N. H. (2007). ORI introduction to the responsible conduct of research (Rev. ed.). U.S. Government Printing Office.