Blog

Social Media Analysis -- A battlefield for multi-disciplines researchers
Recently I am digging into the field of social media texts, and want to explore that whether there is any correlations between texts/posts/opinions in the social media, and generally social reality and specifically people's attitude towards political issues or social issues. Gradually I got two gains:

1. there are, without too much surprise, significant correlations between two things (since basically the former one is the consequence and reactions of the latter one in time-sequence and in logically deduction), though not exactly in the form of what people may expect.

2. social media analysis truly is, or about to be, a battlefield for a multiple disciplines researchers. Personally speaking, I think it should be a great field for communication researchers, but dispassionately after wild exploration for several days, I found that it is already a stage for researchers from computer science in the first place (Could anybody tell me that on earth is that because I read too many papers of CS or they just pump up everywhere in the "digital" era ?) and following with the researchers from the linguistic studies and information system... (a quick bold but not necessarily true thought: that's why communication is considered as a field that is going to extinct! )

I'll, briefly summarize the scenario in the opinion mining field. Natural language processing researchers have long stepped into this field, starting with the grow and maturation of the blogosphere. They are interested in the how to identify opinion expressed in texts in the Internet. Their main aim is to settle the research problem how to exactly get the idea from the sentence structure or semantic clues. Their work starts from the linguistic analysis and ends at the accurately summarizing human's language and translating such emotional expression or opinions expression. They are not interested in how such opinions occurs in the certain social contexts and how it would mean to social issues. Naturally in the industry opinions are highly valuable. And consequently that's why the commercial corporations attach importance to consumer's opinions so much to the consumer survey (in a traditional pattern before). Undoubtedly in this sense linguistic analysts/researchers are making millions for corporations, both for the corporations' own sake or for the interest-related corps.

Another group of people eyeing on this field is the leopard-army: computer science researchers. It seems that they occupy each field, fast and thoroughly, in a well-designed scientifically and somehow statistically/mathematically way. Their research interests are ... well ... they just want to know what is going on social media. So they derive data from Internet, compute, and run data and get the result, fulfilling everything in a very easy and smooth way, never attempt to jump to any dangerous conclusions. Their work based on data and their ability to handle the data, in many case, “big data” -- characterized as large-volume, diverse and tagged with time-attribute. For their research interests, in a very superficial way, to precisely describe human activities is their missions, and hobby. They seldom care much about the "social meaning" behind what they've discovered, but they are the men who know exactly what is going on, at least on surface.

To bridge the gap between micro-macro has been a "dream" for social scientists long before the opinion mining (as people always do, intending to widen their territory once they establish their name in the field) To echo the words by Eulau from political science mentioning about the direction of the communication: "the new ‘discipline’ of Communication represents the fulfillment of the dream for... Interdisciplinary behavioral science that can address the gap." (Eulau’s,1996)

So, as you may notice that I wrote the above sentence with a kind of aspiring and respecting to the researchers in the former two fields. But talking about social scientists, what we got?

(Well, well, it seems that we got talents to claim that we know everything and all we need is enough funding... ) Also, as you can see from the above that they are not interested in the social meaning. Opinions does not occur for no reason, it's a reflection of social reality (that's why we can find a job). And there are several roads in front of us:

1. improving our ability to collect data and run data 2. keep playing with all the political and economic analysis...

 please noted that these two roads are by no means parallel All comments and supplementary are welcomed and

to be continued....

Reference: Eulau, Heinz. 1996. //Micro-Macro Dilemmas in Political Science: Personal Pathways Through Complexity//. Norman, OK: University of Oklahoma Press. Is Public Opinion Stable?: Resolving the Micro-Macro Disconnect in Studies of Public Opinion, with Thomas J. Leeper, Daedalus. Forthcoming.