Twitter Commentary Often Diverges Sharply from Public Opinion Data

Twitter commentary may be more partisan and politically biased than polling data according to a Pew research study.

This hardly seems surprising - but it may have both financial and political implications. That the political bias evident in Twitter commentary diverges from the results of polling is probably due to the fact that Twitter is a means of expression directly related to the twitterer's ideology or point of view on an issue. Polling is frequently anonymous and more reactive than expressive. People being polled understand that their opinions are being aggregated  in order to produce knowledge about a given subject. Twitter, by contrast, is a free form and short hand shout out. It is an individual presentation which may or may not be intended to inform but is certainly intended to influence. Those tweeting understand that - perhaps more subconsciously than not - and act accordingly.

The financial or informational implication of this study may give investors pause. Attempts are being made to determine whether aggregated Twitter data may be correlated or causally related to trends that could influence investments. If, the information gleaned from Twitter turns out to be more a reflection of emotional or purely personal opinion rather than considered judgement, it may still have value but its power as a predictor of a more sustainable type may not be as consistent or useful due to the reactive rather than proscriptive nature.

This is further evidence of the fact that more data is not necessarily more illuminating unless its context and interpretation are thoroughly understood. JL

Tom Kludt reports in TPM:

The reaction on Twitter to political events often bears no resemblance to empirical data, according to a year-long study from Pew Research Center released on Monday.
Pew's study focused on eight major events in the last year – including the first presidential debate and President Barack Obama's re-election victory — comparing the tone of tweets to the results of actual polling.
The Twitter response to Obama's re-election, for example, was overwhelmingly favorable: 77 percent of the tweets in response to his victory on Election Night were positive, while only 23 percent were negative. But Pew's own survey research showed a more polarizing reaction, with 52 percent saying they were happy that Obama won a second term compared with 45 percent who said they were unhappy.
After the early-October debate in Denver, which was widely viewed as an unequivocal victory for Mitt Romney, a majority of 59 percent of the tweets were nevertheless supportive of Obama. Forty percent of the post-debate tweets were supportive of Romney, much lower than the 66 percent who declared the former Republican presidential nominee the winner in a subsequent survey.
But some Twitter reactions were more conservative than public opinion. Only 13 percent of responses to Obama's second inaugural on the social media platform were positive, while 21 percent were negative and 65 percent were neutral. But a survey found a plurality of 48 percent expressing a positive reaction to the inaugural speech compared with 22 percent who gave a negative review and 29 percent who said it was neither positive nor negative.
Pew found only two instances in which public opinion comported with the response on Twitter: the Supreme Court's ruling on the Affordable Care Act and Romney's selection of Rep. Paul Ryan (R-WI) as the Republican vice presidential nominee.