over 80 years, professional pollsters have been trying to predict election
outcomes. Now, an unlikely new technology may bring them one step closer to
predicting the next President of the United States: social media.
A new study released today by OhMyGov
Research – the media analytics arm of OhMyGov Inc. – showed a strong association
between a politician’s popularity on social media and political polling
The study analyzed whether Facebook and
Twitter fan and follower numbers and account growth rates correlate with
political polling data by examining the leading Republican presidential
candidates over a six-month period.
Results from the study showed a statistically significant correlation
between the total number of social media Facebook fans and Twitter followers a candidate
had and his/her popularity in the polls.
the past six months leading up to the upcoming 2012 Republican primaries, many
voters have been seeing campaign messages, debates, and political commentary
—not only through traditional media, but also on social networking sites like
Twitter and Facebook. Voters react instantly to the candidates' appearance,
words and policies on such platforms. But do reactions on social media mirror
reality? This study certainly suggests it does, especially on Facebook, where
the correlation between fans (“likes”) and polling numbers was even stronger
than that between Twitter followers and polling data.
“Federal, state and local elections are
all using social media to engage voters and the media, so it’s no surprise to
me that we found a correlation between social data and polling,” said Andrew B.
Einhorn, CEO of OhMyGov Inc. “We saw similar results in our 2010 study of
Congressional elections. This study just confirms what we’ve been seeing all
The new analysis examined the total
number of Facebook fans and Twitter followers of the eight Republican
candidates: Michele Bachmann, Herman Cain, Newt Gingrich, Jon Huntsman, Ron
Paul, Rick Perry, Mitt Romney and Rick Santorum from June 2011 to December 13,
2011, and the social media growth rates from July 2011 to December 13, 2011.
Interestingly, no relationship was
found between growth rates on social media and polling numbers, though
researchers state it is possible that the sample size was simply too small or that
prevailing views of candidates did not change enough over the study period to
create a relationship between growth rates and polling data. Researchers go on
to state in the report that the sample size was intentionally small to
determine if similar size analyses could be helpful for augmenting polling data
for campaign managers during election cycles.
the science of social media prediction matures, issues like social media
manipulation by spammers
and propagandists will need to be addressed, said Einhorn. “But the
implications of harnessing social analytics for the purpose of making
predictions clearly holds a lot of promise. It also signifies a growing trend
within the tech space: he who controls the data, controls the future.”
GOP Social Media Analysis 21 Dec 2011 Final