What flavors are different types of scholarly articles: an investigation of PLOS publications

Using Article-Level-Metrics (ALMs) data from PLOS, we analyze the citations, views, saves, and shares of different types of articles. Using multiple statistical and visualization techniques, we explore scientists' behaviors associated with scholarly works, which could help us better understand their information needs, and provide clues about how different types of scholarly information are valued. We find that scholarly impact could be better interpreted when article types are taken into consideration. Moreover, the correlations between citations, views, saves, and shares also demonstrate the need to consider article types when assessing the validity of altmetrics. Since altmetrics usage data are associated with various information behaviors, we advocate that altmetrics, especially only the scores, should be used with caution. In summary, our approach provides a novel point of view to explore altmetrics and citation data. Future work will adopt quality approaches to look into more details of the current analyses, to get a better understanding of flavors of scholarly impact.