Twitter has created new hierarchies in academia rather than serving as a democratising force, according to a study that raises questions about the use of social media data to judge research impact.

Researchers who analysed the Twitter activity of 469 postgraduates and academics found that follower counts were significantly skewed towards a small elite of users. The top 1 per cent most popular scholars accounted for 21 per cent of all followers in the study, with an average of 15,059 each.

The most popular 5 per cent accounted for 43 per cent of followers, while the top 50 per cent scooped up 91 per cent of followers.

The study, which covered Twitter users who had tweeted about the 2014 American Educational Research Association conference, found that lecturers and professors tended to have significantly more followers than postgraduates: 557, on average, versus 36.

³ÉÈËVRÊÓƵ

But further analysis found that the size of a scholar¡¯s social media following was strongly influenced by a number of factors that may have a more limited relationship with academic merit.

For example, there was a strong relationship between the number of accounts a user followed and their own follower count. A user following 100 accounts had 91 followers, on average, while someone following 1,000 accounts was likely to have 870 followers.

³ÉÈËVRÊÓƵ

Other significant factors were the number of tweets that the user had posted, and the length of time that they had been a member of the social network.

The analysis, by George Veletsianos, Canada research chair in innovative learning and technology at Royal Roads University, and Royce Kimmons, assistant professor in instructional psychology and technology at Brigham Young University, is .

Dr Veletsianos told ³ÉÈËVRÊÓƵ that the research demonstrated how Twitter was ¡°not an equalising force¡± and instead ¡°may recreate or foster alternative hierarchical structures¡±.

Dr Kimmons said that the use of social media data to measure the impact of scholarly activity ¨C so-called altmetrics ¨C was inevitably called into question by their findings.

³ÉÈËVRÊÓƵ

¡°Though follower count might be used as an altmetric for impact, signifying the reach of the academic, its connection to other academic metrics of success, such as rigour and prestige, is dubious,¡± Dr Kimmons said.

¡°The implications of social media as an altmetric platform for scholars should be considered carefully, because it is difficult to isolate behaviours and qualities reflecting scholarly value from those that merely reflect one¡¯s ability to flourish in a social platform, which means that metrics may favour those scholars who can game the social system.¡±

chris.havergal@tesglobal.com