Focus groupsFocus groups
Frank Luntz's dial-meter focus groups measures people's reactions to the debates in real time. allow for real time tracking of people's responses to debates. We use Twitter to set up an "online" version of focus group, based on three sub-groupsSub-groups
Each group has been sampled so that they have equivalent size (about 2500 each). Research suggests that individuals don't spend much time paying attention to information that opposes their point of view, so we expect that individuals' twitter behavior reveals something about their political preferences. This method gives us a unique view of how people are reacting to the candidates' statements during the debate.:
The curves below show the tweet rate generated by each group (updated every 10 seconds).
The tag clouds show the most frequently used words in these tweets. The R, D and B clouds capture the emergent topics (if any) from the watchers. The words containing names, hashtags, and handles of the candidates are removed for visual interestingess.
The Winning index tracks the victory declaration words from the groups. Scores above zero on the index indicate members of the group tweet "winning" words when mentioning Obama/Biden, such as "winning" or "victory," or losing words when mentioning Romney/Ryan. Scores below zero indicate the reverse: that tweets contain winning words for Romney/Ryan and losing words for Obama/Biden.
The Sentiment scoreboard tracks how the groups express sentiment when talking about the candidates. Scores above zero on the index indicate members of the group tweet positive words when mentioning Obama/Biden, and negative words when mentioning Romney/Ryan. Scores below zero indicate the reverse: that tweets contain positive words for Romney/Ryan and negative words for Obama/Biden.
Each index is normalized to control for group activity and accumulated over the past hour. This means the index value is not influenced by one group becoming more vocal than another. For detailed indices, see breakdown by group.
Also check out tweet meter via Tweet elites!
We invite you to watch these real time streams with us during the debate. After the debates we'll present results analyzing the content and timing of these tweets in more detail.
Some interesting things to look for (hover the headline to get a description):
Zingers vs. Gaffes
Zingers vs. Gaffes
Which gets more attention on Twitter -- applause lines from the candidate's base or gaffes from their opposition? Watch the tweet volume and the tag clouds to see which groups react to which messages. An interesting question is whether fatigue plays into these reactions. Perhaps applause is strong early but only gaffes play toward the end of the debate.
"You're no Jack Kennedy" moments... "You're no Jack Kennedy" moments...
Which zingers or gaffes become memes and take off on their own? Which groups pick them up? How long do they last?
Support, attack or defend? Support, attack or defend?
Research suggests people will tune out when they expect to hear information that doesn't support their views, but if they are confronted with a message they can't avoid, they will react strongly. We predict that people will generally tweet to support their own candidate, but will attack in bursts when the other candidate interjects with assertions they cannot avoid.
Polarization vs. Consensus Polarization vs. Consensus
Which points in the debate lead to polarization of the groups, expressed either through candidate evaluation or sentiment? What issues appear to bring them together? Which groups show more loyalty to one side or resistance to change?