The distribution of citation counts across publications is highly skewed, with a small subset of super-cited papers, defined in this study as those ranking in the top 0.1% by cumulative citations, receiving disproportionately high citation counts. Using universities in the engineering field and its subfield, environmental engineering, from the NTU Rankings (2014 to 2023) as the analytical sample, this study applies four counting methods to examine how super-cited papers influence university rankings. The results indicate that high-ranking universities generally experience counting inflation but maintain relatively stable positions, whereas lower-ranking universities are more sensitive to the inclusion of a few super-cited papers. Although inflation occurs less frequently among these institutions, its magnitude is often substantial when it does arise. This phenomenon is particularly evident in the environmental engineering subfield, which exhibits a higher proportion and broader distribution of super-cited papers. Despite these differences, rankings obtained through the four counting methods remain highly correlated, with harmonic counting and straight-first-author counting showing the highest consistency because both emphasize first-author contributions. At the national level, Chinese universities show the most pronounced ranking changes, with their positions improving after the exclusion of super-cited papers, whereas universities in the United States and the United Kingdom display relatively modest variations. Within the environmental engineering subfield, rankings are more strongly affected and display higher sensitivity to these papers than in the broader field. These findings underscore the importance of considering ranking intervals, disciplinary level, counting method, and national context when interpreting university rankings in research evaluation.

Huang Mu-hsuan, Xiao Yuntong, Liu Xiaojuan. Exploring the impact of super-cited papers on counting methods, counting inflation and university rankings: Evidence from engineering [J]. Journal of Informetrics, 2026,20(3).