By Dr Giulia Stefenelli, Scientific Communications Lead, MDPI and Dr Enric Sayas, Business Analyst, MDPI– Silver Sponsor at University Press REDUX Conference 2025
The context of self-citation scrutiny
Self-citations are a key topic in
academic publishing. While self-citations can naturally reflect a journal’s
specialization and research depth, concerns arise when they artificially
inflate metrics like the Impact Factor (IF) and CiteScore, potentially
distorting a journal’s perceived value.
This topic garnered significant attention
following the release of Clarivate’s 2024 Journal Citation Report (JCR), which included
the suppression of 17 journals from a range of publishers for excessive
self-citation patterns (see, Clarivate
Suppression List 2023).
This prompted MDPI to analyze the citation patterns of its 237 JCR-indexed
journals, ensuring alignment with industry norms and reinforcing our commitment
to transparency, improvement, and integrity.
Methodology
As part of this assessment, we
compared MDPI journals to others found in the same JCR category. Using data
from Web of Science (WoS), which aligns with the data impacting JCR scores and is
independently accessible, enabled an accurate and open comparison with other
publishers’ datasets.
The
original data and python scripts used can be found here: 10.5281/zenodo.14844342
Key Findings: Alignment with industry norms
Findings
reveal MDPI’s citation practices are sound, consistent, and comparable to
prevailing industry trends and established norms.
The evaluation includes 237 MDPI
journals indexed in the JCR and analyzed their Impact Factor (IF) both with and
without self-citations for the latest available year, 2024. Self-citations
rates (SCRs) were also examined relative to category averages
and assess how self-citations influence their overall ranking within their
respective categories.
Findings show some variability in self-citation. Overall, however, MDPI journals align to industry norms and the limits established by WoS. MDPI has very few journals with high SCRs in their category compared to other leading publishers.
In 2024,
MDPI’s average SCR was 14.85%, positioning the publisher 6th among the top ten
publishers by SCR and 4th by publication volume (Figure 1).
SCRs across MDPI’s journals were calculated
individually and compared to the average rates within their respective JCR
categories. For journals indexed in multiple categories, the first category was
selected.
Beyond Citations: Emphasizing broader research impact
We also examined the overall
distribution of SCRs across all JCR journals and compared them to MDPI journals.
Figure 2 shows SCRs are not normally distributed but exhibit a right-skewed
distribution. This indicates that while most journals have relatively low SCRs,
some exhibit significantly higher rates.
Based on
this distribution, we analysed SCRs using percentile rankings, defining outlier
journals as those exceeding the 95th percentile (top 5%). The median self-citation
rate of all JCR journals was 0.12 while the median for MDPI-only journals was
0.14. This difference was shown to be not significative using a Wilcoxon
rank-sum test (P-value 0.33374); there is no statistically significant
difference in SCRs between MDPI journals and all journals in the dataset. Any
observed difference in medians is due to random variation rather than a true
difference in behaviour.
A similar
approach can be applied within each JCR category, identifying outliers based
solely on SCRs within their respective categories.
This was performed by grouping journals by their primary JCR category and generating SCR distributions within each. This allows for identifying outliers at the category level based on SCRs within the same field.
Per Figure 3, outlier SCRs vary
widely, from 20% to nearly 100%. For MDPI journals, however, the range is more
contained, between 20% and 40%. Notably, some MDPI journals with relatively
high SCRs are not classified as outliers, suggesting they belong to JCR
categories with inherently high SCRs. This highlights the need for
category-specific assessments rather than a universal threshold.
Overall, we found that among the top
10 publishers (excluding MDPI), 4.7% of journals were classified as outliers
based on their SCRs, whereas for MDPI, this figure was 2.3%.
Minimal impact on Rankings
We also
examined whether excluding self-citations from the IF calculation significantly
affects journal rankings within their categories. To do this, we ranked all the
journals per category based on their IF both with and without self-citations,
then observed the changes in ranking and quartile distribution. The analysis
revealed that any category
ranking changes were modest, with shifts averaging 7% based on 2024 data.
Key findings from the 237 MDPI journals analyzed reveal:
- 105 journals (44%) saw a ranking change between 10% and -10%, when self-citations were excluded.
- 74 journals saw a ranking decrease, and 49 journals saw a ranking increase when not counting self-citations.
- Six journals had a ranking decrease higher than 20%, while eight journals had a ranking increase higher than 20%.
The
analysis also looked at whether any MDPI journals experienced changes in
quartile rankings due to self-citations. As Table 2 shows, 14 journals decreased
quartile when self-citations were removed: five journals from Q1 to Q2 and nine
journals from Q2 to Q3. However, we also observed seven journals that increased
their quartile when self-citations were removed: one journal from Q4 to Q3, three
journals from Q3 to Q2, and three journals from Q2 to Q1. Removing self-citations from MDPI journal IFs
left 91% of journals in their original quartile. While some moved out of Q1,
others moved in, with 97% of MDPI’s indexed portfolio unaffected by these
shifts.
The
results reinforce that self-citations have minimal influence on MDPI journal
rankings, emphasizing the integrity of editorial practices and the absence of
any deliberate efforts to manipulate impact.
Reinforcing transparency and integrity
As part
of our continued commitment to integrity, transparency and protecting the
scientific record, we welcome further scrutiny of the data used. We believe
research impact goes beyond metrics to advancing knowledge, addressing societal
challenges, and creating meaningful academic conversations. Broader indicators,
such as research quality, societal relevance, and overall scholarly influence
are equally important.
We invite our authors, editors, and readers to explore this analysis further and engage with us in conversations about responsible publishing. Together, we can continue to uphold the high standards of scholarly communication.
About MDPI
Headquartered in Switzerland, MDPI is an Open Access
publisher with a portfolio of more than 440 journals across all scientific
disciplines. To date, MDPI has published the works of over 3.7 million
researchers, collaborating with an extensive network of academic institutions
and scientific societies worldwide. Above all, MDPI is committed to ensuring
that high-quality research is freely accessible to readers across the globe.
About the Authors
About University Press Redux 2024
The 6th ALPSP University Press Redux returns as an
in-person event on 15 & 16 May 2024, in partnership with Oxford University
Press. Book your place.
No comments:
Post a Comment