Correlation of Universities’ Perfor-mances in THE and THE Impact World Rankings with Social Attitudes toward them: Opinion Mining of Tweets

Purpose: This study aims to evaluate university ranking systems by analyzing the attitudes of their social stakeholders. It examines opinions expressed in tweets about universities and analyzes their correlations with the universities' performance scores. The rationale lies in the importance of the evaluating the academic performance of higher education institutions, which is essential but challenging due to the absence of universally accepted standards. For decades, global university rankings have sought to develop methodologies to measure academic performance. However, they face criticism for overestimating certain factors, underestimating or ignoring others, and exhibiting bias due to their reliance on surveys or citation databases. Consequently, their results are often viewed as unrealistic and misleading. Attaining accurate ranking outcomes is essential, as they can initiate a "ranking for ranking's sake phenomenon, prompting individuals and institutions to prioritize ranking criteria over their core missions and the quality of their services. This shift in focus may cause universities to stray from their primary objectives due to unreasonable external evaluations. Therefore, it is necessary to evaluate the quality of ranking system results. Research on ranking systems has primarily concentrated on the correlations of their outcomes. Additionally, these studies have addressed the methodological similarities among various global ranking systems. The findings confirmed partial similarities in both the results and methodologies. However, due to the commonalities in methodologies, it is necessary to evaluate the results based on external benchmarks. As universities aim to address the needs and interests of their stakeholders in order to enhance their societal standing and ensure their survival, the attitudes of these stakeholders can serve as a valuable benchmark. Social media provides a platform for the general public to discuss and comment on various subjects, including university capabilities, services, and activities. Consequently, these platforms offer opportunities to analyze the opinions of university stakeholders worldwide.Times Higher EducationMethodology: The methodology used in this study involved a quantitative content analysis, incorporating opinion mining and altmetric approaches. The research focused on a sample of universities that are ranked in the Times Higher Education (THE) and THE Impact Ranking systems. A distinct and specific name was an inclusion criterion to ensure precise searches on X (formerly Twitter). A sample of 355 universities ranked in the Times Higher Education (THE) rankings from 2019 to 2021 were identified, and their coverage in the THE Impact ranking was verified, resulting in 174 universities being ranked in both systems. Tweets about these universities were extracted using Mozdeh's Big Data Text Analysis, and opinion scores were calculated using the SentiStrength opinion mining tool. The tweet data were collected through an extensive search on Twitter from January 3, 2022, to August 8, 2022. Due to the non-normality of the data distributions, Spearman correlation analysis was employed to analyze the data.Findings: The research findings indicated weak to moderate significant correlations between the frequency of tweets about universities and their overall scores in the Times Higher Education (THE) and THE Impact Ranking systems, as well as their dimension scores in THE. Tweet counts demonstrated the strongest correlation with the overall score in THE Impact and the weakest correlation with industry income in THE. A significant relationship was observed between total tweet counts and tweet counts reflecting opinion polarities, indicating that an increase in a university's performance score is associated with a rise in expressed opinions on Twitter, and vice versa. This finding was consistent across both total scores in the Times Higher Education (THE) and THE Impact Ranking systems, as well as the dimension scores of the latter. Furthermore, the overall THE score showed a direct correlation with the strength of positive opinions, suggesting that improved performance in the United Nations Sustainable Development Goals (SDGs) resulted in higher positive opinion scores in tweets. However, no significant correlation was found between performance scores in THE Impact and negative opinions expressed in tweets. Meanwhile, overall performance scores in THE were correlated with both positive and negative opinion strengths.Conclusion: The results indicate that, while societal perceptions of universities are closely aligned with their performance in sustainable development goals, they only partially correspond with their performance in the Times Higher Education (THE) rankings at both overall and in specific dimensions. Thus, THE ranking system and social attention may have a complementary relationship, offering a more comprehensive understanding of university performance. It is, therefore, possible to enhance ranking systems by incorporating social surveys. However, due to the challenges associated with altmetrics, extensive research is necessary to facilitate the practical application of this metric in evaluations. This research also contributes to the altmetrics literature by reaffirming the distinction between quantitative and content-based approaches in altmetric studies.