Assembling the evidence…
Empirical research on scientific misconduct and questionable research practices (QRPs) has increasingly adopted a systemic perspective, emphasising research climate, supervision, and organisational conditions. This PRISMA-guided scoping review synthesises 229 survey- and interview-based studies published between 2016 and 2025 to examine how misconduct and QRPs are empirically operationalised in perception- and climate-oriented research. Rather than aiming to estimate prevalence, the review focuses on the concepts, instruments, and methodological choices through which integrity problems are rendered measurable. The findings show that aggregate harm is rarely defined explicitly and is instead inferred from perceived prevalence and severity of practices. Severe violations dominate measurement frameworks, while routine, relational, and structurally induced practices remain weakly captured. Methodological patterns further shape what becomes empirically visible, highlighting the need for approaches that better integrate individual perceptions with institutional conditions when assessing integrity risks.
aggregate harm is rarely defined explicitly and is instead inferred from perceived prevalence and severity of practices
the corpus-wide pattern lacks reported counts and an auditable study-level coding trail
Severe violations dominate measurement frameworks, while routine, relational, and structurally induced practices remain weakly captured.
instrument dominance is not quantified and the reported category table is based on title and abstract mentions
Methodological patterns further shape what becomes empirically visible.
descriptive method–respondent alignment cannot establish that methodological choices determine empirical visibility
Survey-based designs dominate the corpus, accounting for roughly two-thirds of all included studies
Table 1 directly reports surveys in 148 of 229 studies
research is concentrated in relatively stable and well-institutionalised systems, predominantly in Europe and North America
nearly half of contexts are unspecified and Asia exceeds North America among named regions
Derived from the full evaluation — not a separate score.
Strengths
Table 1 characterises 229 studies by year, region, respondents, method, misconduct category, and analytical focus, providing a useful overview of the recent perception- and climate-oriented literature.
↳ Results, “Characteristics of included studies”; Table 1
The paper separates the operationalisation of harm from the visibility of institutional conditions and returns to both questions explicitly in the Discussion.
↳ Background, research questions; Discussion, “Answering the research questions”
The Discussion links English-only and two-database coverage, self-reported evidence, and design heterogeneity to underrepresentation, bias, and the decision not to conduct quantitative synthesis.
↳ Discussion, limitations paragraphs
Limitations
The claims that harm is rarely defined explicitly and severe violations dominate are not accompanied by corpus-wide counts or study-level coding. Table 1 states that misconduct categories reflect explicit mentions in titles and abstracts.
↳ Abstract; Results, “Aggregate harm and blind spots in empirical measurement”; Table 1 footnote
No complete search strings, field codes, or run dates are reported, and one author screened the records. These omissions create material uncertainty about corpus completeness and selection consistency.
↳ Methodology, literature search and screening paragraphs
The limitations section does not address the consequences of single-author screening or title/abstract-based categorisation for the headline findings. It therefore understates uncertainty in the qualitative synthesis.
↳ Methodology; Table 1 footnote; Discussion, limitations paragraphs
The review’s strongest reported result is the descriptive profile in Table 1, including the dominance of surveys and the concentration on general researcher populations. The central synthesis goes further by interpreting how harm and institutional conditions are operationalised, but the Methods do not report a sufficiently detailed extraction and coding process for auditing those conclusions across 229 studies. The absence of reproducible search strings and the use of a single screener further constrain confidence in corpus completeness. The paper remains useful as an agenda-setting map, but its main interpretive claims require a clearer study-level evidential trail.
Nabu’s assessment, alongside the field’s view.
Are you an author of this paper?
Limited2.9
Confidence mediumThe review contributes a useful map of 229 perception- and climate-oriented studies and identifies relational, structural, and emerging practices that instruments may miss. The advance is moderated by incomplete positioning against prior syntheses and limited traceability for its principal corpus-wide interpretations.
“aggregate harm is rarely defined explicitly and is instead inferred from perceived prevalence and severity of practices”
The review reports eligibility criteria, a PRISMA-guided flow, descriptive categories, and an open metadata DOI. Reproducible search strings and dates are absent, screening was conducted by one author, and the qualitative synthesis lacks a documented study-level coding trail.
“Screening was conducted by the author based on predefined inclusion and exclusion criteria.”
The sequence from research questions through mapping, synthesis, limitations, and answers is readily followable. However, headline terms such as “show,” “dominate,” and “demonstrates” communicate stronger corpus-wide support than the reported counts and coding trail establish.
“The analysis further demonstrates that methodological choices strongly shape what becomes visible”
The paper engages literature on climate, supervision, incentives, and instrument adaptation, and discusses database, language, self-report, and heterogeneity limitations. It does not trace the implications of single-author screening or title/abstract-based categorisation through its limitations and conclusions.
“the findings should be interpreted as a structured mapping of empirical approaches and recurring blind spots”
Concerns4 of 4 checks
Several headline synthesis claims cannot be traced to reported corpus-wide coding or counts, and one geographic interpretation does not align cleanly with Table 1. These issues could affect whether the interpretive results represent the included corpus.
The review claims availability of structured metadata through a persistent Zenodo DOI, and no conduct issue is established from the supplied text. Repository contents could not be independently inspected, which is an assessability gap rather than a conduct concern.
Flags: 1 declared / 5 total
39 references in manuscript 39 of 39 checkable references found in an index
No retraction notice found in Retraction Watch.
Sources: Retraction Watch ✓
Where this paper’s evidence sits on the path from initial observation to real-world use.
The synthesis can inform the redesign of surveys and institutional risk assessment, and the metadata deposit supports reuse. It does not provide a validated instrument, risk model, or implementation guidance.
“Risk-oriented analytical frameworks offer a promising direction”
AI-generated, human-governed. Something look off? Contact us to request a review.