What Happens Inside a Public Opinion? A Micro-Level Approach

There is rising interest to apply social media data to understand public opinion and even replace traditional surveys (e.g. Gayo-Avello, 2013; Anstead & O'Loughlin, 2014). However, it has been recently observed that constructing public opinion based on social media data is anything but trouble-free (e.g. Cowls 2014), not least because it is so often done based on the content created, overlooking the users creating it. One should not only concentrate on the size, content and the tone of the pool of social media messages, but descend to examine the micro-level and track the actions of individual users. There are remarkable differences in behavior among the social media users producing this macro-level public opinion. Others advocate loudly and actively on various subjects, whilst others express their opinion only very rarely (Vaccari & Valeriani, 2015; Graham & Wright, 2014). Also, while a significant segment of public usually has fixed beliefs, another segment is attentive and changes their views in response to events and the discussions taking place online (Brulle et al. 2012). Due to the differences, I argue, to understand the construction on macro-level public opinion it is vital to analyse how big of an impact do different types of users have on the macro-level public opinion’s formation. In this paper a careful micro-level analysis on individual users’ online actions is conducted to help clarify the factors influencing macro-level opinion movements in parallel. Based on a data set of 1.5M social media posts collected in Finland during the parliamentary elections 2015, the paper analyses the actions of different types of users concerning certain popular topics, and the progress in their micro-level public opinions are compared with the overall trend of the macro-level public opinion. By comparing the changes this paper opens up new perspectives on the representativeness issues related to public opinion.