Abstract: This study aimed to develop a metric for measuring the diffusion speed of scientific knowledge on social media platforms, using Twitter as an example, and to investigate the differences between short-term and long-term diffusion speeds of scientific papers, as well as their influencing factors. Articles published in Volumes 66-68 of CA: A Cancer Journal for Clinicians were selected as the research sample. The entropy weight method was used to construct a metric for diffusion speed. SPSS was used to analyze the differences between short-term and long-term diffusion speeds, while Eviews was employed to conduct Granger causality analysis to identify the factors influencing these differences. The analysis revealed that the number of keywords and the number of authors significantly influenced the diffusion speed of scientific knowledge on social media. In the short term, the diffusion speed was Granger-caused by the citation count of the first author, whereas in the long term, it was influenced not only by citation counts but also by the first author’s academic impact, as measured by their h-index. These findings suggested that the diffusion speed of scientific knowledge on social media is affected by multiple factors. Specifically, the number of keywords and the number of authors played a significant role in both short-term and long-term diffusion. Moreover, in the short term, the citation count of the first author was a key driver, while in the long term, both citation counts and the first author’s academic impact contribute to sustained dissemination. Therefore, scholars with higher academic influence were more likely to facilitate the long-term diffusion of their scientific papers on social media, primarily through the endorsement of their academic reputation.

Key words: Scientific knowledge diffusion, Social media, Diffusion speed, Speed index, Citation relationship