Abstract

In 2012 there were many proposals for new indicators associated with social web tools. These indicators have been termed as altmetrics (or alternative metrics) and are presented as an alternative to the evaluation of scientific activity. In this paper some possible limitations of the new indicators are presented. More specifically these shortcomings are: 1) a large number of information sources and indicators that make it difficult to establish their classification and relevance; 2) some indicators that have no statistical validity because they generate poor results; 3) the difficulty of determining the meaning of the new metrics –what are we assessing, scientific or social impact?; and 4) the evanescent and ephemeral nature of these new sources and indicators.

Abstract

The extensive use of the web by many sectors of society has created the potential for new wider impact indicators. This article reviews research about Google Scholar and Google Patents, both of which can be used as sources of impact indicators for academic articles. It also briefly reviews methods to extract types of links and citations from the web as a whole, although the indicators that these generate are now probably too broad and too dominated by automatically generated websites, such as library and publisher catalogues, to be useful in practice. More valuable web-based indicators can be derived from specific types of web pages that cite academic research, such as online presentations, course syllabi, and science blogs. These provide evidence that is easier to understand and use and less likely to be affected by unwanted types of automatically generated content, although they are susceptible to gaming.

Abstract

This literature review assesses indicators derived from social media sources, including both general and academic sites. Such indicators have been termed altmetrics, influmetrics, social media metrics, or a type of webometric, and have recently been commercialised by a number of companies and employed by some publishers and university administrators. The social media metrics analysed here derive mainly from Twitter, Facebook, Google+, F1000, Mendeley, ResearchGate, and Academia.edu. They have the apparent potential to deliver fast, free indicators of the wider societal impact of research, or of different types of academic impacts, complementing academic impact indicators from traditional citation indexes. Although it is unwise to employ them in formal evaluations with stakeholders, due to their susceptibility to gaming and lack of real evidence that they reflect wider research impacts, they are useful for formative evaluations and to investigate science itself. Mendeley reader counts are particularly promising.