Public Opinion Formation Online
The formation of public opinion is often determined by a multitude of factors, with both internal and external influences such as political parties or social media bias. However, the formation of public opinion was further convoluted with the introduction of the Internet. Now, individuals as well as larger groups of people have far greater access to content that could influence the formation of public opinions. An opinion leader within a network is usually be defined by their power in relation to other members, so opinion leadership would typically be determined by metrics such as degree and closeness centrality. Taking into account the Internet makes this measurement and the factors involved much more complex. In a paper investigating public opinion relating to Barack Obama’s Nobel Peace Price nomination, Juliane Urban and Kristin Bulkow present a new approach to measuring online public opinion to determine the outcome of the nationwide implementation of the Web on the public sphere.
One interesting phenomenon noted was a deviation between two common social network analysis methods: polling and social media analysis. Polls suggested public opinion on Obama’s Nobel Prize nomination to generally be negative, however, analysis of different social media platforms like Twitter and other blogs led to the opposite conclusion. Urban and Bulkow go on to define three levels of public sphere: the encounter public sphere, the public sphere of assemblies, and the media public sphere.
After defining levels of public sphere, they discuss the formation of online public opinion through the replacement of journalists in the real world by online search engines and speakers themselves. Speakers also tend to use hyperlinks, which result in the transportation of opinions between different spheres simply by clicking one button. Hyperlinking provides a measure of public opinion because search engines will rank popular links highly based on how many different public spheres are linked within the source, essentially allowing others to see how well represented the certain online public opinion truly is.
Through further analysis of the hyperlink network for search engine results and in-link pages for the Nobel Prize nomination, Urban and Bulkow were able to depict a visual representation of the public spheres and speakers with the most influence on public opinion.
The results indicated that weblogs are the dominant platform of influence on online public opinion at 65%, while other platforms like journalistic media, websites of institutions, and forums follow by a much smaller percentage. They also found that most of the links posted by individuals from the public sphere of assemblies refer to journalistic media content. The edges of this hyperlink network resemble the structure of a power network in which certain nodes have more influence on others while being in the form of a bipartite graph, where the two disjoint sets are represented by the in-link pages and the Google-hits. The nodes of greater influence point to the way online public opinion is formed and to the amount of influence of this particular node in relation to the othes in its disjoint set. The sentiment of the posts by public spheres was also found for each type of public sphere. In relation to the hyperlink network, edges that showed a greater strength between nodes did not correlate with the overall positive/negative sentiment of the post. It just showed which sources were more likely to cause opinions to spread regardless of whether the opinion on the Nobel Prize nomination was negative or positive.
https://www.sciencedirect.com/science/article/pii/S1877042813042328