Classifying Altmetrics by Level of Impact

Introduction In the light of current knowledge we can conclude that altmetrics do not present an alternative for traditional citation-based analysis of research impact (e.g., Haustein et al., 2014). Altmetrics have instead the potential to show some other aspects of research activities and provide a more nuanced view of the impact research has made on various audiences (Liu & Adie, 2013; Piwowar, 2013). Altmetrics come in many forms and from many different sources, all of which can represent different aspects of the online activity or of the different levels of impact that various research products have made on different audiences. What exactly the different altmetrics represent we do not yet know, but the greatest advantage of altmetrics may be exactly in this diversity. Aggregating all altmetrics to a single indicator would remove this advantage. With aggregation of different altmetrics we are just creating another impact factor, another indicator that in the worst case is used for something that it is neither designed for nor capable of indicating. However, because of the wide variety of different sources for altmetrics, some form of aggregation or classification is needed and different types of classifications are already used by some service providers. Here we present another approach, one based on the level of impact. With this we hope to stimulate further discussion about the actual meaning of altmetrics.