University of Zagreb

EXECUTIVE SUMMARY AND STRATEGIC VISION

The University of Zagreb demonstrates a robust overall scientific integrity profile, reflected in a very low global risk score of 0.001. This strong performance is anchored in excellent control over multiple affiliations and hyperprolific authorship, where the institution outperforms national averages. However, this solid foundation is contrasted by areas requiring strategic attention, particularly a higher-than-average national rate of retracted output and a notable tendency towards institutional self-citation and publication in its own journals. These vulnerabilities require monitoring to ensure they do not undermine the University's outstanding research leadership, which is evident in its top-tier SCImago Institutions Rankings in key areas such as Computer Science, Dentistry, Business, Management and Accounting, and Engineering. The institution's mission to foster "scientific and artistic research" and "professional work" is directly challenged by integrity risks that suggest potential endogamy or gaps in quality control. To fully align its operational practices with its mission of excellence, it is recommended that the University strengthens its pre-publication review mechanisms and encourages broader international validation, thereby safeguarding its reputation and ensuring its significant contributions are recognized globally.

ANALYSIS BY INDICATOR

Rate of Multiple Affiliations

With an institutional Z-score of -1.094, the University shows a complete absence of risk signals in this area, performing even better than the low-risk national average of -0.821. This low-profile consistency demonstrates that the institution's affiliation practices are well-aligned with national standards of integrity. While multiple affiliations can be a legitimate result of researcher mobility or partnerships, the University's exceptionally low rate confirms that there are no signs of strategic attempts to inflate institutional credit or engage in “affiliation shopping,” reflecting clear and effective governance in this domain.

Rate of Retracted Output

The University's Z-score of 0.023 indicates a moderate deviation from the national standard, which sits at a low-risk Z-score of -0.095. This suggests the institution is more sensitive to the factors leading to retractions than its national peers. Retractions are complex events, and while some signify responsible supervision in correcting honest errors, a rate significantly higher than the national average alerts to a potential vulnerability in the institution's integrity culture. This Z-score suggests that quality control mechanisms prior to publication may be failing more frequently than elsewhere in the country, indicating possible recurring malpractice or a lack of methodological rigor that requires immediate qualitative verification by management.

Rate of Institutional Self-Citation

The institution presents a Z-score of 0.554, which, while within the medium-risk band like the national average (0.288), is notably higher. This indicates a high exposure to practices that may not be prevalent elsewhere in the country. A certain level of self-citation is natural, reflecting the continuity of research lines. However, this disproportionately high rate signals a concerning risk of scientific isolation or 'echo chambers' where the institution validates its own work without sufficient external scrutiny. This value warns of potential endogamous impact inflation, suggesting that the institution's academic influence may be oversized by internal dynamics rather than global community recognition.

Rate of Output in Discontinued Journals

The University's Z-score of -0.309 is statistically normal and fully aligned with the national average of -0.284, both falling within the low-risk category. This indicates that the institution's risk level for publishing in discontinued journals is as expected for its context and size. The data shows that researchers are exercising appropriate due diligence in selecting dissemination channels, effectively avoiding media that do not meet international ethical or quality standards and thus protecting the institution from the reputational risks associated with 'predatory' practices.

Rate of Hyper-Authored Output

With a Z-score of 0.190, the University demonstrates differentiated management of this risk compared to the national average of 0.472. Although both fall into the medium-risk category, the institution's significantly lower score suggests it moderates risks of authorship inflation that appear more common in the country. In fields outside of 'Big Science,' extensive author lists can dilute individual accountability. The University's relative control in this area indicates a healthier approach to authorship, serving as a positive signal that it is better at distinguishing between necessary massive collaboration and questionable 'honorary' authorship practices.

Gap between Impact of total output and the impact of output with leadership

The University's Z-score of 0.715 reflects a medium-risk gap, but it also shows differentiated management, as this value is lower than the national average of 0.807. This suggests the institution moderates the risk of impact dependency more effectively than its national peers. A wide positive gap signals a sustainability risk where scientific prestige is dependent on external partners rather than internal capacity. The University's smaller gap indicates a comparatively stronger foundation of intellectual leadership, suggesting its excellence metrics are more closely tied to its own structural capabilities than the national trend.

Rate of Hyperprolific Authors

The institution maintains a prudent profile with a Z-score of -0.850, a value that is not only in the low-risk category but is also significantly lower than the national average of -0.608. This indicates that the University manages its research processes with more rigor than the national standard. While high productivity can be legitimate, extreme individual publication volumes can challenge the limits of meaningful intellectual contribution. The University's low score is a strong indicator of a healthy balance between quantity and quality, with minimal risk of practices like coercive authorship or authorship assignment without real participation.

Rate of Output in Institutional Journals

With a Z-score of 1.773, the University shows high exposure to this risk, exceeding the already medium-risk national average of 1.531. This indicates that the institution is more prone than its peers to relying on its own publication channels. While in-house journals can be valuable for local dissemination, excessive dependence on them raises conflicts of interest, as the institution acts as both judge and party. This high Z-score warns of a significant risk of academic endogamy, where production might bypass independent external peer review, potentially limiting global visibility and indicating the use of internal channels as 'fast tracks' to inflate CVs.

Rate of Redundant Output

The University's Z-score of -0.001, while in the low-risk category, points to an incipient vulnerability when compared to the much lower national average of -0.247. This suggests that while the issue is not widespread, the institution shows early signals of data fragmentation that warrant review before they escalate. Massive bibliographic overlap between publications can indicate 'salami slicing'—the practice of dividing a study into minimal units to artificially inflate productivity. This slight elevation above the national baseline serves as a reminder to monitor publication strategies to ensure that the focus remains on presenting significant new knowledge rather than prioritizing volume.