Using bibliometrics to detect questionable authorship and affiliation practices and their impact on global research metrics: A case study of 14 universities

https://www.webofscience.com/api/gateway/wos/peer-review/10.1162/qss_a_00339

In today’s increasingly competitive higher education landscape, universities face immense pressure to improve their global rankings. These rankings influence institutional decisions, from student admissions and faculty recruitment to funding allocation and strategic policy formulation (Hazelkorn, 2015; Hazelkorn & Mihut, 2021; Rovito, Kaushik, & Aggarwal, 2021). Central to these rankings is the focus on research, which prompts many institutions to prioritize publication and citation metrics (Rhein & Nanni, 2023; Sheeja, Mathew, & Cherukodan, 2018; Vernon, Andrew Balas, & Momani, 2018). While many efforts to enhance research metrics are legitimate, some involve questionable and potentially unethical practices, raising concerns about the integrity of academic publishing and the reliability of the ranking systems that depend on these metrics. Practices such as gift authorship, ghost authorship, guest authorship, honorary authorship, sold authorship, paid affiliation, and multiaffiliation have become increasingly prevalent, distorting the evaluation of university research performance (Bhattacharjee, 2011; Biagioli & Lippman, 2020; Halevi, Rogers et al., 2023; Hottenrott, Rose, & Lawson, 2021; Moosa, 2024; Pachter, 2014; Trung, 2020).

These practices and their impact on academia align with Goodhart’s Law, which states, “When a measure becomes a target, it ceases to be a good measure” (Dezhina, 2022; Fire & Guestrin, 2019), further undermining the value of research metrics as indicators of academic success. The proliferation of these questionable behaviors or practices signals an urgent need for reforms to maintain research integrity and preserve the credibility of ranking systems (see Box 1 for definitions of these practices).

Recent analyses have identified significant declines in first authorship rates and sharp increases in research output, hyperauthorship (excessive numbers of authors per paper), multiaffiliation (excessive numbers of affiliations), and hyperprolific authorship (individuals with high numbers of articles per year) at specific institutions, raising concerns about authorship attribution and its implications for research evaluation (Biagioli & Lippman, 2020; Catanzaro, 2024; Fire & Guestrin, 2019; Gureev, Lakizo, & Mazov, 2019; Ioannidis, Klavans, & Boyack, 2018; Jakab, Kittl, & Kiesslich, 2024; Teixeira da Silva & Dobránszki, 2016). Despite the establishment of ethical guidelines, such as those from the International Committee of Medical Journal Editors (ICMJE), questionable authorship and affiliation practices persist, often evading detection due to their complexity and the scale at which they occur (Resnik, Rasmussen, & Kissling, 2015; Teixeira da Silva, 2023). Current detection methods focus on individual misconduct (Barta, 2022; Whetstone, Ridenour, & Moulaison-Sandy, 2022), leaving a gap in identifying institutional-level practices that distort global research metrics.

While bibliometric techniques cannot definitively prove questionable authorship and affiliation practices, they serve as valuable methods for identifying anomalies in research output and performance (Delgado López-Cózar & Martín Martín, 2024). These anomalies may appear as sudden spikes in publication volume, high instances of hyperauthorship, concentrated coauthorship in specific venues, excessive multiaffiliation, or sharp increases in hyperprolific authors, among others (Biagioli & Lippman, 2020). Bibliometric analyses, therefore, play a crucial role in safeguarding research integrity by detecting deviations from normative authorship and affiliation practices and providing data for studying relevant questionable behaviors.

Using bibliometric analysis, we examine 14 universities that demonstrate significant anomalies in research output and authorship and affiliation patterns. We assess the consequences of these deviations on institutional rankings and research performance.

The study group was selected through a structured multistage process to identify universities exhibiting extraordinary growth in research output and significant declines in first authorship rates between 2019 and 2023. The selection of the 2019–2023 period was driven by the need to capture recent considerable shifts in research output and authorship dynamics over a short period, often influenced by institutional policies, government incentives, or the pressures of global ranking systems (Ahlers & Christmann-Budian, 2023; Lee, Liu, & Wu, 2020; Shattock, 2017; Shen, Zha, & Liu, 2023). This period is important because it coincides with the rise of hyperprolific authorship, a growing phenomenon in academic publishing (Moreira, Meira et al., 2023).

Additionally, using these years helped minimize distortions that might or could have been caused by the COVID-19 pandemic, which disproportionately impacted disciplines and institutions (Raynaud, Goutaudier et al., 2021; Zammarchi, Carta et al., 2024). Indeed, global research output experienced irregular annual growth rates during the pandemic, increasing by 10% in 2020 and 8% in 2021 before slowing to 3% in 2022 and declining by 3% in 2023 (based on data from Scopus and corroborated via Web of Science). By selecting 2019 and 2023 as focal points, we aimed to isolate genuine shifts in research productivity and authorship and affiliation practices from pandemic-related anomalies (Ioannidis, Salholz-Hillel et al., 2021).

Our initial selection process began with identifying all academic institutions that published at least 2,000 documents during the 5 years from 2019 to 2023, using data from SciVal. This produced a data set of 2,740 institutions worldwide. For each institution, we retrieved two key bibliometric indicators: total journal articles published annually from 2019 to 2023 and the first 9 months of 2024 and first authorship rates for these years (via InCites). These two metrics—research output and first authorship rates—were chosen because they are critical for understanding institutional productivity and changes in research contribution patterns. See Box 2 for a description of the data sources and software used in the study.

To focus the analysis on institutions with a substantial impact on global research metrics, we narrowed the data set to the top 1,000 universities based on the number of journal articles published in 2023. This step ensured the exclusion of smaller or less research-intensive institutions. We then refined the selection by identifying institutions with publication growth rates exceeding 100% between 2019 and 2023, more than five times the global average growth of 20% (per SciVal). Eighty institutions met this criterion, signifying exceptional growth and potential deviations from typical global research trends.

Next, we focused on institutions with notable shifts in authorship practices, particularly first authorship rates. First authorship is a recognized marker of scholarly leadership, often signifying the primary intellectual contributor to a study (Chinchilla-Rodríguez, Costas et al., 2024; Kharasch, Avram et al., 2021; Marušić, Bošnjak, & Jeroňcić, 2011). The first author typically leads research design, data collection, and analysis. Therefore, a sharp decline in first authorship rates can indicate changes in research contribution dynamics, such as the rise of hyper-authorship and possibly questionable authorship and affiliation practices like gift, honorary, or sold authorship.

Accordingly, from the 80 high-growth universities, we applied a rigorous selection criterion to identify institutions experiencing significant shifts in authorship dynamics. Specifically, we focused on those whose first authorship rates declined by more than 15 percentage points between 2019 and 2023. This threshold, which is over five times the global average decrease of 3%, was carefully chosen to pinpoint institutions undergoing substantial changes in research leadership patterns. Such a dramatic decline often indicates a fundamental shift in how research contributions are distributed within an institution. It may suggest an influx of external collaborators taking lead roles or, potentially, the emergence of questionable authorship and affiliation practices. By setting this stringent criterion, we ensured that our study group comprised only those universities demonstrating the most pronounced and potentially concerning changes in authorship dynamics. Ultimately, this selection process yielded 14 universities that met both criteria: exceptional growth in research output and a marked decline in first authorship rates, both exceeding five times their respective global averages from 2019 to 2023 (see Figure 1 for an outline of the study group selection process).

The two key metrics—research output growth and first authorship decline—are complementary, representing different facets of institutional research behavior. A substantial increase in research output typically signals institutional productivity, driven by strategic responses to rankings, funding, or external incentives. However, rapid growth in publication volume can also be associated with potential engagement in practices designed to inflate research metrics artificially. Conversely, declines in first authorship rates provide insight into institutions’ evolving structure of research contributions. A marked reduction in first authorship may indicate shifts towards authorship and affiliation practices that could obscure genuine academic leadership. This imbalance between research output growth and first authorship decline offers a critical lens through which to examine the implications of current authorship and research practices across institutions.

The 14 universities comprising the final study group represent diverse geographic and institutional backgrounds. Six are public universities in Saudi Arabia and eight are private universities in Egypt, India, Iraq, and Lebanon. One is relatively young, having been established less than 15 years ago; eight were founded within the last 15 to 25 years, while four have a history spanning over 5 decades. According to several ranking agencies (e.g., QS and U.S. News), these universities also vary in size, with six institutions each employing between 500 and 1,000 academic staff members, four somewhere between 1,000 and 2,500 academic staff members, and four hosting over 2,500 such members each.

To ensure a reliable and meaningful comparison, we adopted a two-pronged approach for benchmarking: a national benchmark based on aggregated data from the study group countries—Egypt, India, Iraq, Lebanon, and Saudi Arabia—and a global control group comprising well-established universities known for their adherence to high-standard conventional authorship practices. This approach allowed for a broad, stable comparison, mitigating the distortions that the peculiarities of national institutions in the study group countries might introduce.

For the global control group, we selected four internationally renowned universities that consistently rank among the top 10 in the Academic Ranking of World Universities (Shanghai Ranking): California Institute of Technology (Caltech), Massachusetts Institute of Technology (MIT), Princeton University, and the University of California, Berkeley. These universities were chosen for several reasons:

• ▪ 

  Research stability: These institutions have consistently high research output, providing a solid benchmark against which to assess the rapid and significant growth seen in the study group universities.

• ▪ 

  Global influence: As global leaders in research, their practices set a high bar, making them ideal comparators for assessing the broader implications of questionable authorship and affiliation behaviors.

• ▪ 

  Disciplinary balance: According to Scopus, the study and control groups share the same top seven subject categories as their main research strengths, reinforcing the comparability of their research portfolios (see more below).

In addition to using national data to make comparisons, we compared the study group and control group universities on several key metrics: overall and subject-specific publication growth, first authorship trends, number of authors per article, hyperprolific authorship, multiple institutional affiliations, and levels of international collaboration. We concentrated on two primary types of documents—articles and reviews—excluding other formats such as conference papers and editorials to maintain consistency and focus on core academic output across disciplines. For each country represented in the study group, we excluded publications where the study group university was an author. This exclusion ensured that national benchmarks remained independent of the study group’s output, providing a more transparent and objective comparative framework.

Unless otherwise stated, we sourced our data from Scopus and its analytical tool—SciVal—employing data collection methods similar to those outlined in Halevi et al. (2023). We tracked each author’s publication counts, institutional affiliations, and country affiliations. In addition to 2019–2023, we collected publication and authorship data for 2014–2018 (and January–September 2024 to use where needed). We performed sensitivity analyses to validate the robustness of our findings, cross-referencing data between Scopus and Web of Science and utilizing both SciVal and InCites to ensure comprehensive coverage and accuracy.

A significant aspect of this analysis was the examination of hyperprolific authorship, a concept introduced by Ioannidis et al. (2018) to describe researchers who publish a high number of papers in any one calendar year (in their case, 72 articles, conference papers, substantive comments, or reviews). This phenomenon has raised concerns regarding adherence to the four ICMJE authorship criteria. Ioannidis found that 70% of hyperprolific authors admitted to not fulfilling these criteria more than 25% of the time. Moreira et al. (2023) further characterized these authors as “anomalous” due to their sudden and unsustainable spikes in productivity, often concentrated in specific venues and collaborations. In this study, we adopted a more conservative threshold of 40 articles in a calendar year between 2019 and 2023 to capture a broader spectrum of potential anomalies.

To identify hyperprolific authors, we initially downloaded data for the 500 most published authors during 2014–2024 in each of the 27 major academic disciplines represented in SciVal (e.g., mathematics, psychology). For large disciplines, such as engineering and medicine, we supplemented this with data on the top 500 most published authors within their specific subdisciplines (e.g., biomedical engineering, oncology). Additionally, we extracted the top 500 authors from each major global region (e.g., Central Asia) and countries with large numbers of productive authors (e.g., Canada, India, Saudi Arabia). After removing duplicates, we compiled a comprehensive list of 53,630 highly published authors. We then tracked the annual publication count of each author from 2014 to 2023 and during the first 9 months of 2024. Using these data, and excluding physicists, we identified 11,838 hyperprolific authors globally, 9,011 of whom achieved this status at least once from 2019 to September 2024. These 9,011 hyperprolific authors represent 0.05% of the 18 million authors during this time frame.

We excluded physicists to avoid skewed results, as many of these scientists participate in large international collaborations with over 1,000 members, where authorship is granted based on team membership rather than individual contributions. This decision aligns with the rationale provided by Ioannidis et al. (2018), ensuring that the data more accurately reflect individual research productivity and potential authorship anomalies. Notably, the number of hyperprolific authors worldwide (i.e., those who published 40 or more journal articles in a calendar year) increased by 66%, from 2,517 in 2019 to 4,189 in 2023, against an increase of 15% for all authors during the same period (InCites).

In addition to hyperprolific authorship, we also explored the growing trend of multiple institutional affiliations—authors listing affiliations with numerous institutions, often across different countries (Hottenrott et al., 2021). This phenomenon has seen a marked increase over the past decade, with some authors listing affiliations with more than 20 institutions (Halevi et al., 2023), raising significant concerns regarding the legitimacy of their contributions to each institution (Gök & Karaulova, 2024). These individuals often serve as “external authors” to those institutions frequently listed as secondary affiliates. The motivations behind these affiliations, such as whether institutions provide financial compensation in exchange for listing, fall outside the scope of this study.

Research output is a core indicator in global university rankings, with institutions often ranked based on the volume of their publications. A substantial increase in research output can propel universities up the rankings, influencing decisions around faculty recruitment, funding, and international collaborations. Figure 2 highlights the contrasting trajectories of research output between the study group and the control group from 2019 to 2023. The 14 universities in the study group demonstrated substantial increases in publication output, ranging from 100% to 1457%, resulting in an overall increase of 234%, nearly 12 times the 20% world average. In contrast, the control group universities exhibited only modest changes in publication output, with increases ranging from 1% to 6%, leading to an overall increase of 3% over the same period, significantly below the world average but slightly above the 2% average for the United States, where all the control group members are based.

To provide a deeper perspective, Table 1 compares each institution’s growth in the study group with their respective national averages. For example, Future University in Egypt exhibited a 977% increase in research output, significantly outpacing the national average of 71%. Similarly, Indian universities in the study group, such as Chandigarh University and GLA University, grew by 530% and 485%, respectively, both far surpassing India’s national average of 36%. In Iraq, Al-Mustaqbal University saw its research output soar by 1,457%, dwarfing the national average increase of 63%. The Lebanese American University, with a 723% growth, exceeded Lebanon’s national average of 18%, while Saudi institutions in the study group also performed well above the national average of 109%, with Prince Sattam Bin Abdulaziz University growing by 485% and Princess Nourah Bint Abdulrahman University achieving an 819% increase. These figures highlight the extraordinary research output growth at study group universities compared to their national contexts and global trends.

By comparison, the universities in the control group, based in the United States, showed much smaller increases in research output. For instance, the University of California, Berkeley, recorded only a 1% increase, while MIT and Princeton saw a modest 6% rise. These figures align with the US national average growth rate of 2%, reflecting stable research output levels at top-tier institutions.

The substantial increases in research output among the study group universities have led to a significant rise in their global rankings, underscoring the strategic focus on expanding publication volume. In 2019, only one study group institution ranked within the top 300 most-published universities globally, per Shanghai Ranking. By 2023, this number had risen to five, reflecting a dramatic surge in research productivity. This trend is further corroborated by the Shanghai “research output” (PUB) rankings, where the number of study group universities listed among the world’s top 500 most-published institutions increased from one in 2020 (which is based on articles published in 2019) to seven in 2024 (based on articles published in 2023). Conversely, control group universities showed a modest decline in research output rankings. The median rank of the control group universities fell from 165th in 2019 to 210th in 2023, with two institutions dropping out of the top 300 altogether. Additionally, their Shanghai rankings declined from an average of 126 in 2020 to 158 in 2024.

The dramatic increase in research output among the 14 study group universities is further underscored when analyzed by discipline. Discipline-specific research productivity impacts subject area rankings and can notably affect a university’s standing. In 2019, only one of the study group universities ranked among the world’s top 100 most-published institutions in any of the 21 subject categories tracked by Clarivate’s Essential Science Indicators. By 2023, however, seven universities had broken into the top 100 a total of 35 times across 13 subject categories, with significant representation in science, technology, engineering, and mathematics (STEM). In contrast, the control group declined in several areas, having its representation among the world’s top 100 most-published universities reduced from 36 appearances in 17 subject categories in 2019 to 28 appearances in 13 categories in 2023 (Table 2).

The study group’s presence becomes even more pronounced by expanding the analysis to include the 300 most published universities globally in each subject category. Moving from 17 representations in 2019 by two universities to 13 universities appearing 77 times across 18 subject categories in 2023, the study group showed substantial gains in fields such as chemistry, computer science, engineering, environment and ecology, materials science, mathematics, pharmacology and toxicology, and physics, with six or more universities in each category. Shifts in research output rankings were less pronounced in the medical, health, and social sciences. Conversely, the control group saw a decrease in representation among the world’s 300 most published universities by field, dropping from 54 appearances in 2019 to 48 in 2023.

First authorship in publications is a key indicator of intellectual leadership within research projects. It offers insight into the balance between internal and external contributions from researchers at a given institution. Globally, first authorship rates have slightly declined, with the average among academic institutions dropping from 53% in 2019 to 50% in 2023 (InCites). This trend varies across disciplines, with rates ranging from 42% in clinical medicine to around 47–49% in environmental science, ecology, geosciences, and physics, 52% in computer science and social sciences, and 56–57% in engineering and mathematics. The rate in physics would decline to 38% when counting articles with over 1,000 coauthors.

The 14 universities in the study group experienced a dramatic decline in first authorship rates, from an average of 50% in 2019 to just 27% in 2023—an eightfold drop compared to the global average. This downward trend reflects a significant shift in how research contributions are attributed, with several universities falling far below the global and disciplinary averages. Notably, eight of the 14 study group universities saw the most significant declines in first authorship rates among the 1,000 most published universities worldwide. By 2023, 11 of these institutions were ranked among the world’s 15 lowest for first authorship rates, a considerable drop from their median rank of 662 in 2019 (Table 3). For instance, Saveetha Institute of Medical and Technical Sciences, which had the highest first authorship rate in the world in 2019 (89%), saw its rate plummet to 25% in 2023, ranking it among the lowest globally. Similarly, GLA University dropped from 78th to 969th place during the same period.

In contrast, the universities in the control group experienced only modest declines in first authorship rates, dropping from 46% in 2019 to 42% in 2023. The median rank for this group remained relatively stable, moving slightly from 736th in 2019 to 759th in 2023. Control group members rank 759th because a significant proportion of their research output is published by sizeable research groups (23% are articles with over 10 coauthors each, compared to 8% among the study group members). The relatively stable performance of the control group suggests that, while there has been a general decline in first authorship globally, the sharp drop in the study group universities may indicate institutional strategies prioritizing output quantity over research leadership.

Hyperprolific authorship, defined in this study as publishing 40 or more journal articles in a calendar year, is becoming a significant feature of global academia. However, the universities in the study group have exhibited a far more pronounced increase in hyperprolific authorship compared to global trends. Between 2019 and 2023, the number of hyperprolific authors at these institutions rose by 670%, increasing from 23 authors in 2019 to 177 in 2023. This growth rate is 10 times the global average increase of 66% over the same period, underscoring the distinct authorship dynamics within the study group.

The increase in hyperprolific authorship varied across the institutions in the study group. For example, King Saud University in Saudi Arabia experienced an extraordinary rise, from four hyperprolific authors in 2019 to 63 in 2023, a remarkable growth of 1,475%. Similarly, the Lebanese American University, which had no hyperprolific authors in 2019, had 18 by 2023. Chandigarh University in India, which also had no hyperprolific authors in 2019, saw the emergence of eight by 2023 (Table 4).

In contrast, the universities in the control group, including Caltech, MIT, Princeton, and UC Berkeley, showed far smaller increases in hyperprolific authorship, adhering more closely to traditional authorship norms. For instance, Princeton University had just two hyperprolific authors in 2019 and 2023, while Caltech maintained a steady number of three in both years. Overall, the control group saw a modest increase of 100%, with hyperprolific authors growing from seven in 2019 to 14 in 2023—substantially lower than the rates observed in the study group.

Further data analysis revealed a notable presence of noncore hyperprolific authors (HPAs)—researchers who published 40 or more articles in a calendar year, with their publications attributed to multiple institutions in different periods. This phenomenon often arises when authors affiliate with or depart from institutions during a calendar year, leading to their contributions being split across different affiliations. These noncore HPAs differ from core HPAs, who contribute at least 40 articles to a single institution during a calendar year. Notably, an author can simultaneously be a core HPA at one institution and a noncore HPA at another, and their status may shift over time—for instance, transitioning from noncore to core in subsequent years. Moreover, while some noncore HPAs contribute only a handful of articles annually to an institution, others may contribute as many as 39. For this study, any noncore HPA contributing more than 10 articles per year to an institution was deemed to have a significant impact on that institution’s research output. This nuanced categorization underscores the complexity of hyperprolific authorship and its implications for institutional research metrics, highlighting the need for further scrutiny of these patterns to ensure academic integrity and transparency.

Our data show that noncore HPAs significantly impact the research output of several study group universities. For instance, at the Lebanese American University, the 48 noncore HPAs accounted for 29% of the university’s total research output in 2023–2024, while the 30 core HPAs contributed 47%. At Chandigarh University, 20 noncore HPAs were responsible for 14% of the research output, compared to 19% for the 13 core HPAs. These figures underscore noncore HPAs’ critical role in enhancing overall productivity, even at institutions where core HPAs are already present in substantial numbers.

In contrast, institutions in the control group, such as MIT and UC Berkeley, also had noncore HPAs—10 and 11, respectively—though their contribution was relatively modest, reflecting more traditional authorship patterns at these universities. This distinction highlights the amplifying effect of noncore HPAs on research output, particularly at institutions with high concentrations of core HPAs, as seen in the study group.

The analysis of hyperprolific authorship from January 2019 to September 2024 reveals notable differences between the study and control groups regarding raw numbers and the nature of their productivity trends. During this period, the study group had 396 hyperprolific authors compared to 33 in the control group, highlighting a striking disparity in prevalence (see Table 4). However, the contrast extends beyond these figures and reflects fundamental differences in the authors’ publication histories and productivity growth.

A key observation in the study group is the relatively recent emergence of hyperprolific authorship. Only 7% of hyperprolific authors in the study group consistently published more than 10 articles annually during the 5 years before 2019, and just 9% were classified as hyperprolific during that earlier period. This indicates that most hyperprolific authors in the study group achieved this status only recently, reflecting a sharp, sudden surge in their output.

In contrast, hyperprolific authors in the control group displayed a far more consistent history of high productivity. Among control group institutions, 45% of hyperprolific authors published more than 10 articles annually between 2014 and 2018, and 21% had already attained hyperprolific status in that period. This indicates a steady and sustained pattern of scholarly output, demonstrating that hyperprolific authors in the control group were long-term high producers before the study period.

The differences in productivity growth further highlight the contrast between the two groups. Hyperprolific authors in the study group saw their average output soar from seven articles per year between 2014 and 2018 to 33 articles per year between 2019 and 2023, marking a staggering 371% increase. By comparison, the control group’s hyperprolific authors experienced a more moderate rise in productivity, increasing from 18 articles per year in 2014–2018 to 40 articles per year between 2019 and 2023—a 122% increase.

When viewed globally, the growth of hyperprolific authorship in the study group is even more pronounced. Globally, excluding the study and control groups, hyperprolific authors increased their output by 81%, from an average of 21 articles per year during 2014–2018 to 38 articles per year during 2019–2023.

Furthermore, among the study group universities, 57% of the hyperprolific authors show sudden surges from a maximum of 20 articles in any year before 2021 to an average of 50 articles annually during 2022–2024. For example, one engineering faculty member’s output spiked from one article annually from 2001 to 2021 to 231 articles in 2022 and 541 articles in 2023, contributing 40% of their institution’s research output in the latter year. Five percent of the study group’s 396 hyperprolific authors exhibit similar trends. By contrast, none of the control group’s hyperprolific authors exhibited similar sudden spikes.

Another key distinction between the groups lies in the affiliation history patterns of hyperprolific authors. On average, hyperprolific authors in the study group had been affiliated with their institutions for 9 years, compared to an average of 17 years for those in the control group.¹ This pattern was especially evident at the Lebanese American University, where 28 out of 30 hyperprolific authors emerged as affiliates in 2022 and 2023, and Chandigarh University, where nine of 13 hyperprolific authors emerged as affiliates during the same period. No comparable surges in recent affiliations were observed within the control group, where hyperprolific authorship remained more stable and was primarily linked to long-standing faculty members.

Both the study and control groups feature hyperprolific authorship concentrated in fields with high publication outputs. However, notable disciplinary distinctions emerge between the two groups, reflecting differences in research practices and institutional strategies.

In the control group, hyperprolific authorship is most prominent in earth and planetary sciences (61%), where large-scale international collaborations, particularly in space sciences and related fields, are common. These collaborations, often involving large numbers of researchers, naturally lead to high publication counts per author. Other fields with significant representation in the control group include medicine, biochemistry, genetics, and molecular biology (24%), where multiauthor papers and high-output research are typical. Engineering represents 12% of the hyperprolific authors in the control group, with a minor presence in materials science (3%).

In contrast, the study group exhibits a broader disciplinary spread, with distinct areas of concentration. Materials science (32%) dominates hyperprolific authorship in the study group. Engineering (24%) and chemistry (20%) also represent substantial portions of hyperprolific authorship. Other significant disciplines include biochemistry, genetics, and molecular biology (15%), pharmacology (14%), and computer science (13%), highlighting the influence of fields with high potential for interdisciplinary research and fast publication cycles. Mathematics (10%) is also prominently featured in the study group. Additionally, more minor but notable contributions in the study group come from agricultural and biological sciences, chemical engineering, energy, and environmental science, each accounting for 5–10% of hyperprolific authors. Eleven other subject categories have between one and 11 hyperprolific authors, underscoring the study group’s diverse and multidisciplinary nature of high-output research. The sum of the percentages is over 100% due to overlap.

Multiaffiliation, especially across institutions in different countries, can boost an institution’s perceived research collaboration and international reach, which are critical metrics in global rankings. While these affiliations can signify genuine collaborations, such as joint appointments or research partnerships, a sharp and rapid rise in multiaffiliations within a short period, as seen among some study group universities, may point to strategic efforts to amplify research output.

As with other indicators examined earlier, the analysis of multiaffiliation practices between the study and control groups revealed significant differences. The study group exhibited marked changes in the proportion of multiaffiliated publications between 2019 and 2023, signaling a divergence in author affiliation behaviors. In 2019, 18% of articles authored by study group institutions involved multiaffiliated authors. By 2023, the percentage remained stable; however, notable increases occurred at specific universities.

Three institutions demonstrated particularly sharp rises in multiaffiliated publications: the Lebanese American University saw a 60 percentage point jump (from 15% in 2019 to 75% in 2023), Chandigarh University increased by 30 percentage points (from 9% to 39%), and the University of Petroleum and Energy Studies saw a 29 percentage point rise (from 5% to 34%).

Conversely, two other universities—Prince Sattam Bin Abdulaziz University and Al-Imam Mohammad Ibn Saud Islamic University—experienced moderate declines in multiaffiliated publications, by five and seven percentage points, respectively. However, despite these reductions, both institutions reported relatively high levels of multiaffiliation in 2023 (24% and 21%), indicating that the practice remains prevalent. All other institutions in the study group recorded around 10% levels in 2023.

An illustrative example of multiaffiliation practices can be seen in the case of a faculty member affiliated with multiple study group institutions. This researcher, averaging five articles per year from 2016 to 2020, saw their output surge to over 50 articles in 2021 and over 100 annually since 2022. Notably, they listed multiple affiliations in 80% of their publications, with an average of 3.5 affiliations per article, often including two to six international institutions.

In contrast, the control group maintained a stable proportion of multiaffiliated publications throughout the same period, consistently at 6% in 2019 and 2023. This stability reflects the adherence of control group institutions to more conventional and consistent authorship and affiliation practices.

This section limits the analysis to articles with up to 100 coauthors. The exclusion of articles with more than 100 coauthors is justified by the unique nature of large collaborative projects, particularly in fields like physics, where the number of contributors can reach thousands. Such large-scale collaborations often follow different authorship norms and would distort the trends observed in smaller-scale research efforts.

From 2019 to 2023, the study group demonstrated a considerable increase in the average number of authors per article, rising from 4.8 in 2019 to 6.4 in 2023, an increase of 33%. The control group also exhibited growth in the number of authors per article, though at a slower pace. The average number of authors per article in the control group increased from 7.5 in 2019 to 9.0 in 2023, representing a 20% increase over the same period. These differences suggest that while both groups are trending toward greater multiauthorship, the study group institutions are exhibiting a more rapid shift. It is worth noting that, according to InCites data, the global average number of authors per article increased by only 5%, from 3.7 authors per article in 2019 to 3.9 in 2023. This comparison highlights the distinctive rise in multiauthorship within both the study and control groups, though the study group appears to be diverging more sharply from global increase trends.

The period between 2019 and 2023 showed substantial increases in international research collaboration among the universities in the study group, contributing to significant improvements in their global rankings in this category. International collaboration, measured by the percentage of publications coauthored with researchers who list foreign affiliations, became a crucial metric in global university rankings, reflecting the increasing globalization of academic research. The study group universities exhibited notable gains, with international research collaboration rates rising by as much as 64 percentage points. In contrast, the control group universities experienced only marginal increases or remained static (Table 5).

The universities in the study group demonstrated marked increases in international collaboration rates, leading to significant shifts in their global rankings in this area. For example, the Lebanese American University (LAU) saw an increase of 39 percentage points in its international collaboration rate, rising from 55% in 2019 to 94% in 2023. This surge propelled LAU to first place globally in international collaboration rankings, substantially improving from its rank of 260th in 2019. Similarly, the University of Petroleum and Energy Studies (UPES) in India improved its international collaboration rate by 48 percentage points, from 12% in 2019 to 60% in 2023. As a result, the university climbed from 992nd in 2019 to 121st globally in 2023.

In contrast, the universities in the control group, such as the California Institute of Technology (Caltech) and Princeton University, exhibited much more minor changes in their international collaboration rates. For instance, Caltech’s international collaboration rate remained steady at 62%, dropping its global rank from 118th in 2019 to 196th in 2023. Similarly, Princeton saw only a one percentage point increase in its collaboration rate, leading to a decline in its global international collaboration ranking from 304th in 2019 to 333rd in 2023.

The increased international collaboration among study group universities has substantially improved their global rankings in this area. In 2019, the median global rank of the study group universities in international research collaboration was 538th; by 2023, this had improved dramatically to 20th, with nine institutions ranked among the top 25 globally. In contrast, the control group universities showed minimal changes in international collaboration rankings. The median rank of the control group fell slightly from 235th in 2019 to 282nd in 2023, reflecting the relative stability of their international collaborations over this period.

VOSviewer-generated maps further illuminate the collaboration dynamics within the study group and beyond. Figure 3 shows the institutional coauthorship network at the beginning of our study period in 2019, highlighting a network involving only 21 external institutions collaborating with any study group member on more than 90 articles (the minimum number of articles published by a group member in that year). In contrast, Figure 4 from 2023 reveals a dramatic expansion, with the number of collaborating institutions skyrocketing to 244, representing a nearly 12-fold increase. This map illustrates a vastly more connected and seemingly collaborative environment, although the validity of these collaborations is questionable, given the unusual collaborative group publication activity patterns.

For comparative context, Figures 5 and 6 depict the institutional coauthorship network maps for the control group in 2019 and 2023, respectively. These universities maintained a more stable collaboration landscape, with a moderate increase in collaborating institutions (from 166 in 2019 to 203 in 2023—a 22% increase).

Publications in high-impact journals are among the important factors influencing university rankings. Institutions that publish more of their research in top-tier journals are often viewed as producing higher-quality research, leading to improved rankings. The study group institutions showed a marked shift in publishing their research in top-tier journals between 2019 and 2023. On average, the proportion of articles published in top 10% journals by impact increased from 15% in 2019 to 26% in 2023, reflecting an overall rise of 11 percentage points. In contrast, the control group experienced a slight decline, with their proportion decreasing from 54% in 2019 to 51% in 2023, a reduction of three percentage points (Table 6).

Individually, the most significant improvements were observed at institutions such as Saveetha Institute of Medical and Technical Sciences in India, where the proportion of articles in top journals surged by 29 percentage points, from 1% in 2019 to 30% in 2023. Similarly, the University of Petroleum and Energy Studies in India saw a significant increase of 24 percentage points, while Future University in Egypt improved by 23 percentage points over the same period.

At the national level, institutions from Saudi Arabia, such as King Khalid University and Prince Sattam Bin Abdulaziz University, outperformed the national average increase of four percentage points, achieving growth rates of 19 and 16 percentage points, respectively. Indian universities, such as Chandigarh University and GLA University, also saw notable gains, though the national average increase was more modest at two percentage points.

In contrast, all control group members recorded declines in the proportion of articles published in top-tier journals. For instance, MIT’s proportion dropped by three points, from 59.6% in 2019 to 56.4% in 2023.

Citation impact is another critical metric in university rankings, with highly cited publications indicative of influential research. A large number and proportion of highly cited papers can significantly affect an institution’s global standing. The study group institutions experienced a notable increase in the proportion of articles ranked among the world’s top 10% most cited, growing from an average of 13% in 2019 to 25% in 2023, a 12 percentage point rise. In contrast, the control group saw a decline in their citation impact, dropping from 28% to 23%, a 5 percentage point decrease (Table 7).

Among the study group, Saveetha Institute of Medical and Technical Sciences again showed the most significant growth, with a 28-point increase, followed by the University of Petroleum and Energy Studies and Future University in Egypt with 27 and 24-point increases, respectively. These increases highlight their growing influence on global research, particularly in regions like India and the Middle East.

Conversely, institutions in the control group faced declines, with MIT seeing the most significant drop of seven points. This trend may indicate a shift in the global research landscape, where emerging institutions are increasing their global citation presence at the expense of top universities.

Study group institutions additionally often outpaced their national averages. For instance, Al-Mustaqbal University College in Iraq experienced a 22-point rise compared to the national average increase of seven points, while Chandigarh University in India saw a 20-point rise against the country’s five-point increase. Such trends underscore the rapid ascent of these institutions in global citation rankings, further distinguishing them from national and global peers.

This study used bibliometric methods to identify or detect indicators of questionable authorship and affiliation practices and examine their impact on global research metrics. The findings reveal significant and concerning trends in authorship and affiliation practices at a subset of universities from 2019 to 2023. The dramatic increase in research output among these institutions, far surpassing global averages, coupled with notable changes in authorship dynamics, suggests a potential reliance on gift authorship, ghost authorship, guest authorship, honorary authorship, sold authorship, and paid affiliation practices. These practices have significant implications for academic integrity and the reliability of global research metrics. Goodhart’s Law becomes particularly relevant in this context, as the focus on metrics such as publication and citation counts as targets for ranking purposes has led to the manipulation and distortion of these measures, ultimately compromising their validity as indicators of genuine scholarly contribution.

Specifically, we investigated the publication practices of 14 universities that exhibited extraordinary growth in research output alongside changes in authorship dynamics during this period. Compared to a control group of world-renowned universities and national metrics, the study group displayed considerable surges in publication output, often accompanied by several notable trends.

• Reduced first authorship: The study group exhibited a pronounced decline in first authorship publications, which may indicate an increased reliance on external contributions. This trend is particularly noteworthy, as first authorship is often viewed as a marker of a significant contribution to research.

• Rise in hyperprolific authors: The study group saw a dramatic increase in the number of individuals credited with an unusually high volume of publications within a short period. This hyperprolific authorship raises questions about the sustainability and authenticity of these contributions, as maintaining such high output levels is generally challenging without resorting to questionable practices (Moreira et al., 2023).

• Increased multiaffiliations: The percentage of publications in which authors listed affiliations with multiple institutions increased significantly. This trend, which often involves authors affiliated with institutions in different countries (Halevi et al., 2023; Hottenrott et al., 2021), raises concerns about these authors’ genuine contributions to each institution’s research output. The sharp rise in multiaffiliated publications, particularly at institutions like Chandigarh University and the Lebanese American University, raises concerns about the strategic use of multiple affiliations to enhance perceived research output and international collaboration. This trend calls into question the integrity of the collaborative efforts and may reflect practices aimed at inflating rankings.

• Inflated authorship: The study group also showed a considerable increase in the average number of authors per publication compared to the control group. This increase, especially when notably higher than national and international levels, may not always reflect genuine intellectual contributions from all listed authors (Hosseini, Lewis et al., 2022).

• Amplified growth in international research collaboration (IRC): The study group experienced a dramatic increase in the proportion of internationally coauthored publications, far exceeding the global average increase. While the IRC is generally beneficial, such a rapid increase suggests possible overreliance on international entities to enhance research output and global network profile. Previous studies have raised concerns about using international coauthorship as a reliable proxy for IRC measurement due to the variability in its dynamics and underlying motivations (Chen, Zhang, & Fu, 2019; Katz & Martin, 1997; Kuan et al., 2024; Luukkonen, Tijssen et al., 1993; Marginson, 2022). Our findings align with the arguments made by Gök and Karaulova (2024), who suggest that questionable forms of IRC, such as IRC via multiple affiliations, are becoming more prevalent and must be scrutinized to ensure accurate assessments of research collaboration. They found that approximately 20% of co-publications are labeled international solely because of an author’s second affiliation, highlighting the need for a nuanced understanding of IRC (Hottenrott et al., 2021). The extraordinary growth in international collaboration, as seen in institutions such as the University of Petroleum and Energy Studies, suggests that while international collaboration is crucial, the speed and magnitude of this growth may indicate that these partnerships are strategically curated to boost ranking metrics.

Our findings highlight the potential of bibliometric evaluations in identifying deviations from expected patterns in scholarly output and activity. By leveraging these evaluation methods, we can promote fair authorship and affiliation practices and uphold the integrity of academic publishing.

The concerning trends in authorship dynamics within the study group suggest the need for further examination and dialog among universities, publishers, policymakers, and ranking agencies. The universities in the study group appear to inflate their publication metrics by relying on external authors. This reliance on external collaborators to boost publication output suggests a shift from internally driven scholarly output toward possibly outsourced work that minimally involves the institution’s faculty members and researchers in primary roles. The examples of a faculty member with an abrupt increase in publication output and another faculty member listing an unreasonable number of international affiliations illustrate potential nominal authorship and affiliation practices where actual research contributions may be minimal. Furthermore, one university from the study group experienced a 10% decrease in full-time professorial rank faculty members from 2019 to 2023. Nevertheless, the university increased its research output by over 700% during these 5 years, providing further evidence of attempts at outsourcing research output.

To foster a robust research culture that is sustainable and impactful, universities need to nurture their talented researchers and cultivate a vibrant, collaborative environment without overreliance on external contributors. Current trends in global ranking manipulations underscore the need for international bodies and academic institutions to critically reassess and refine ranking methodologies to ensure that they reflect true academic excellence and integrity, as emphasized by Goodhart’s Law.

An example of the potential consequences of inflated publication rates is observed in the case of Taif University. According to Scopus, the institution peaked in publications in 2022 with 4,678 articles, up from 516 in 2019, only to see a notable decrease to 2,381 in 2023. This drastic fluctuation suggests that the rapid increases seen across some universities will probably not be sustainable in the long term, potentially leading to a loss of trust in their published research. This instability exemplifies the risks associated with aggressive publication strategies to boost rankings rather than foster genuine academic progress. Moreover, Taif University also experienced a substantial reduction in the number of hyperprolific authors, dropping from 26 in 2022 to seven in 2023, further illustrating the transient nature of such publication strategies. Similar trends are observed in the cases of Al-Mustaqbal University College and Future University in Egypt, where during January–September 2024, they published only 59% and 38%, respectively, as many articles as in 2023, compared to a world average of 85%.

Parallels have been observed in other institutions in the past (Alhuthali & Sayed, 2022; Ansede, 2023a; Bhattacharjee, 2011; Catanzaro, 2023, 2024; Guskov, Kosyakov, & Selivanova, 2018; SIRIS Academic, 2023). For example, Vietnam’s Duy Tan University published 481 articles in 2018, peaking at 2,666 in 2020, but following public scrutiny (Trung, 2020), its output decreased to 848 in 2022 and 852 in 2023. Similarly, the number of publications at Ton Duc Thang University increased from 302 in 2016 to 3,337 in 2020, and after similar public exposure, the number of publications decreased to 756 in 2022 and 551 in 2023. Both universities dropped out of the Shanghai Ranking in 2023 after reaching a peak of 601–700 in 2021.

These cases underscore a challenge where universities may use practices that artificially inflate their research output to enhance their standing in global rankings. This trend not only misrepresents their actual academic contributions but also questions the sustainability and authenticity of their research advancements. Such phenomena, echoing the pitfalls seen at Taif University and possibly other institutions, suggest that initial surges in publication numbers may often be driven by unsustainable practices that do not reflect genuine scholarly activity.

The reliance on networks of external authors and questionable authorship and affiliation practices to increase research output might threaten academic research integrity and undermine public trust in academic institutions. This issue also directly biases the outcomes of ranking systems, compromising their reliability and usefulness. To address this systemic challenge effectively, a concerted effort is required from all stakeholders in the academic community. The proposed measures are as follows:

• Universities: Establish more stringent guidelines and policies for granting secondary affiliations and allowing primary faculty to hold such affiliations. The use of internal review mechanisms to detect and address suspected authorship patterns should be enhanced. Endorse and communicate institutional values that reflect integrity and ethical conduct. Moreover, universities should make more significant efforts to educate their faculty about appropriate and inappropriate authorship in scientific publications (Ali, 2021; Alshogran & Al-Delaimy, 2018).

• Policymakers: Develop and enforce regulations that address questionable publication practices and provide clear guidelines for ethical authorship and collaboration (see, for example, Ansede, 2023b). Transparency in reporting research activities and outputs should be encouraged through legislative and administrative measures.

• Funding agencies: Initiate audits to explore the potential misuse of external authors in funded projects and by applicants seeking funding.

• Ranking agencies: Reconsider how to account for the number of publications and citations concerning primary versus secondary affiliations—count the first affiliation for each author and disregard others. Red flag indicators, such as those discussed in this paper, should be developed and incorporated into evaluation frameworks. Qualitative assessments of research and genuine international collaboration should also be incorporated. Consider focusing on metrics that incentivize homegrown research.

• Accreditation agencies: Establish review guidelines and initiate audits to detect questionable authorship and affiliation practices to boost publication records. Enforce strict compliance with the relevant accreditation standards. Consider revoking accreditation for repeated implementation of questionable authorship practices.

• Scholarly publishers: Strengthen peer review processes to more effectively identify questionable authorship and affiliation claims, especially for networks of hyperprolific coauthors with multiple affiliations. Require detailed justifications for multiple secondary affiliations.

• Database producers: Avoid polluting bibliometric databases by ensuring only high-quality publication venues are covered and deleting records of journals discontinued due to quality issues. For example, one of the study group universities had 85% of its articles in 2019 published in journals discontinued by one of these standard databases. As long as these articles in discontinued journals remain in the databases, universities will continue to receive credit for them in ranking systems because ranking systems do not differentiate between active and discontinued journals in the databases they use to generate publication metrics and rankings. This perpetuates distorted research metrics and undermines the integrity of rankings.

• Researchers: To maintain rigorous ethical standards in authorship and promote transparency within their academic environments, researchers should actively report any suspicions of authorship fraud. Resources such as the Center for Scientific Integrity (Retraction Watch, n.d.) can be utilized to provide guidelines on responsible research, including authorship ethics. The Committee on Publication Ethics (COPE) also provides guidelines for responsible authorship (Committee on Publication Ethics, 2019).

Implementing these measures, coupled with a renewed commitment to conventional publication practices, is essential for preserving the integrity of scholarly work. Without intervention, ongoing trends will continue, eroding trust in the academic record, obstructing true scholarly collaboration, and diverting resources at institutions from truly advancing academic excellence.

We analyzed questionable authorship and affiliation practices among 14 universities exhibiting extraordinary growth in research output, declines in first authorship rates, and increases in hyperprolific authors, multiaffiliations, and authorship counts. However, several limitations should be acknowledged:

• Sample size and scope: The study focused on a cohort of 14 universities that met stringent criteria for research output growth and a significant decline in first authorship rates. While these universities provided valuable insights, the study may not capture the full spectrum of questionable practices globally. Future studies should consider expanding the sample size to include a more diverse set of institutions with varying levels of research activity. Lowering the thresholds for selection could offer a broader understanding of authorship and affiliation behaviors across different institutional types.

• Control group comparability: The control group of globally renowned institutions was selected to provide a stable reference point. However, differences between the study and control groups regarding institutional age, size, funding, and governance must be noted. These differences may influence the interpretation of the findings. Nevertheless, the control group offered a high-standard baseline for assessing deviations in research and authorship trends. Future studies could refine control group selection by including more geographically diverse institutions and those from a similar bracket to the study group.

• Temporal scope: The data analyzed spanned from 2019 to 2023, a period capturing recent shifts in research output and authorship and affiliation practices. However, it may not fully reflect long-term trends in these areas. Extending the temporal scope in future research could provide more robust conclusions regarding the persistence or evolution of questionable practices.

• Interpretation of collaborative patterns: The study identified unusual collaborative patterns, but the mechanisms behind these patterns remain unclear. Further research, incorporating qualitative approaches such as interviews or case studies, would provide deeper insights into the motivations driving these collaborations.

• Disciplinary variations: The threshold for defining hyperprolific authorship—40 or more articles per year—may be appropriate for high-output disciplines but too restrictive for fields with lower publication rates, such as the humanities or social sciences. Future studies should consider using discipline-specific thresholds to more accurately identify hyperprolific authorship behavior across various fields.

• Bibliometric limitations: While bibliometric analysis is a powerful tool for identifying patterns in research output, it has limitations in assessing the qualitative aspects of authorship practices. Multiple affiliations, for example, can reflect legitimate collaborations or strategic attempts to boost institutional metrics. Future research should address these nuances to distinguish between genuine scholarly contributions and questionable practices.

While this study highlights bibliometric anomalies suggestive of questionable authorship and affiliation practices, direct evidence is limited. Future research should address this gap through more detailed, individual-level analyses. Interviews with hyperprolific authors and journal editors could reveal more profound insights into the pressures and motivations behind these trends.

• Online marketplaces for authorship sales: Investigating online platforms where authorship is sold could shed light on the prevalence of paid authorship practices. Researchers could track the publication histories of individuals who have engaged in these activities to validate the presence of sold authorship and paid affiliations.

• Cluster analysis: Employing cluster analysis could help identify groups of universities exhibiting similar questionable practices, allowing for more targeted interventions. This method could also facilitate the categorization of different levels of severity in authorship practices.

• Content analysis of highly cited papers: Analyzing the content of highly cited papers from the study group could help differentiate genuine research from metrics inflation. This approach would involve examining the contribution of each listed author and assessing whether the paper reflects true collaborative research or nominal authorship.

• Distortion of citation metrics: Future studies should explore how questionable authorship and affiliation practices impact citation-based metrics, such as the h-index or highly cited researchers. Investigating the extent to which inflated authorship practices distort these metrics would provide valuable insights for reforming global ranking systems and ensuring they reflect genuine academic contributions.

The authors wish to thank the reviewers for their constructive feedback, which is reflected in this version of the manuscript.

Lokman I. Meho: Conceptualization, Investigation, Methodology, Supervision, Visualization, Writing—original draft, Writing—review & editing. Elie A. Akl: Conceptualization, Investigation, Methodology, Supervision, Writing—review & editing.

The authors declare that they are affiliated with a university that is a peer institution to one of the universities included in the study group.

No funding was received for this research.

All data are in the text, tables, and charts.