Tsinghua University

EXECUTIVE SUMMARY AND STRATEGIC VISION

Tsinghua University demonstrates a robust and globally competitive research profile, reflected in an overall integrity risk score of 0.090. This score indicates a strong foundation of scientific practice, yet also highlights specific areas requiring strategic attention. The institution's primary strengths lie in its profound scientific autonomy, with a near-zero risk of impact dependency (Gap between Impact of total output and the impact of output with leadership), and exceptional diligence in selecting publication venues, as shown by the very low rate of output in discontinued journals. These strengths are foundational to its global leadership, evidenced by its number one world ranking in critical areas such as Computer Science, Engineering, Energy, and Mathematics according to SCImago Institutions Rankings data. However, this pursuit of excellence is accompanied by medium-level risks related to authorship and citation patterns, particularly a high exposure to hyperprolific authorship and institutional self-citation. These vulnerabilities, if unaddressed, could subtly challenge the principles of transparency and external validation that are cornerstones of a world-class academic mission. By proactively addressing these indicators, Tsinghua University can ensure its quantitative achievements are fully matched by qualitative rigor, solidifying its position not only as a leader in research output but also as a global benchmark for scientific integrity.

ANALYSIS BY INDICATOR

Rate of Multiple Affiliations

The university's Z-score for this indicator is 0.224, showing a moderate deviation from the national average of -0.062. This suggests that the institution is more sensitive to the factors driving multiple affiliations than its national peers. While multiple affiliations are often a legitimate result of researcher mobility or partnerships, the university's higher rate signals a need for review. This divergence from the national norm could indicate strategic attempts to inflate institutional credit or "affiliation shopping," a practice that, while not inherently negative, requires monitoring to ensure that affiliations reflect substantive collaboration and contribution.

Rate of Retracted Output

With a Z-score of -0.155, which is even lower than the country's already low average of -0.050, the institution exhibits a prudent profile in managing post-publication corrections. This indicates that its quality control mechanisms are managed with more rigor than the national standard. Retractions can be complex, but a rate significantly below the norm suggests that the university's pre-publication review processes are highly effective at preventing the types of unintentional errors or methodological flaws that often lead to retractions, reflecting a strong culture of integrity and responsible supervision.

Rate of Institutional Self-Citation

The institution shows a Z-score of 0.124, which is higher than the national average of 0.045. Although both the university and the country fall within a medium-risk band, this result indicates a high exposure, suggesting the institution is more prone to this behavior than its environment. A certain level of self-citation is natural, reflecting the continuity of research lines. However, this elevated rate warns of a potential for 'echo chambers' where the institution validates its own work without sufficient external scrutiny. This dynamic risks endogamous impact inflation, where academic influence might be oversized by internal dynamics rather than recognition from the global scientific community.

Rate of Output in Discontinued Journals

Tsinghua University's Z-score of -0.417 is exceptionally low, contrasting with the national average of -0.024. This demonstrates a case of preventive isolation, where the institution does not replicate the minor risk dynamics observed in its environment. The near-total absence of publications in journals that fail to meet international ethical or quality standards is a critical strength. It signifies that the university exercises outstanding due diligence in selecting dissemination channels, effectively protecting its reputation and research investment from predatory or low-quality practices.

Rate of Hyper-Authored Output

The university's Z-score of -0.129, while in the low-risk category, is notably higher than the national average of -0.721. This points to an incipient vulnerability, as the institution shows signals of this activity that warrant review before they escalate. In disciplines like high-energy physics, extensive author lists are legitimate. However, this slight uptick compared to the national baseline serves as a signal to ensure that collaborative patterns are driven by genuine scientific necessity rather than practices like 'honorary' or political authorship, which can dilute individual accountability and transparency.

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

With a Z-score of -1.055, significantly lower than the national average of -0.809, the institution demonstrates a complete absence of risk signals in this area, performing even better than the national standard. This is a powerful indicator of scientific sovereignty and structural strength. It confirms that the university's high-impact research is not dependent on external partners for prestige. This result shows that its excellence metrics are generated from genuine internal capacity and intellectual leadership, mitigating any risk of its scientific prestige being dependent or exogenous.

Rate of Hyperprolific Authors

The university's Z-score of 1.683 is substantially higher than the national average of 0.425, signaling high exposure to this risk. While the national system already shows a medium level of this activity, the university amplifies this trend. Extreme individual publication volumes challenge the limits of human capacity for meaningful intellectual contribution. This high indicator is a significant alert for potential imbalances between quantity and quality, pointing to risks such as coercive authorship, data fragmentation, or assignment of authorship without real participation—dynamics that prioritize metrics over the integrity of the scientific record.

Rate of Output in Institutional Journals

With a Z-score of 0.937, the university shows a moderate deviation from the national average of -0.010. This indicates a greater sensitivity to the risks associated with publishing in its own journals compared to its peers. While in-house journals can be valuable for local dissemination, this higher rate raises potential conflicts of interest, as the institution acts as both judge and party. It warns of a risk of academic endogamy, where scientific production might bypass independent external peer review, potentially limiting global visibility and creating 'fast tracks' to inflate publication counts without standard competitive validation.

Rate of Redundant Output

The university's Z-score of -0.237 represents a slight divergence from the national Z-score of -0.515. While the risk level is low for the institution, it is present, whereas it is almost non-existent for the country as a whole. This suggests the center shows minor signals of risk activity that do not appear in the rest of the country. This finding alerts to a potential, albeit minor, tendency toward the practice of dividing a coherent study into minimal publishable units to artificially inflate productivity. Such 'salami slicing' can distort the scientific evidence base and should be monitored to ensure research is communicated with maximum substance and clarity.