April 27, 2007
Tag Clouds for the Democratic Debate
Janet Harris, a friend and Pollster reader, sends an interesting bit of analysis she did on last night's MSNBC Democratic debate. Using the free site TagCrowd, she created a set of "tag clouds" that provides a visual depiction of the words used most often last night by each of the candidates.
For those not familiar with the term, you have probably seen tag clouds appearing on many web sites (and hopefully, very soon here on Pollster). They were apparently first implemented on the photo sharing site Flickr, and typically provide a visual representation of the most popular "tags" assigned to web pages. The type size of each word varies according to its frequency of usage. The larger the type size, the more often each candidate used that word.
Here are the clouds Janet created:
She also created a PDF version suitable for printing.
Now of course, this is a quick blog post, which probably raises as many questions for me as it answers. Each of the clouds consists of the 50 words used most often, omitting common words like "and," "of," "the," etc. I am not sure if the scale of the words is comparable across clouds -I suspect that Professor Franklin will feel strongly that they should be. Finally, for what it's worth, Janet also sends along this total word count for each candidate:
- 1,872 - Senator Obama
- 1,766 - Senator Clinton
- 1,518 - Senator Edwards
- 1,281 - Governor Richardson
- 1,180 - Representative Kucinich
- 961 - Senator Biden
- 912 - Senator Dodd
- 753 - Senator Gravel
A few quick observations, with an assist from Janet (who is the president of the media analysis firm, Upstream Analysis):
- Notice the more frequent use of wonkier language by Chris Dodd, particularly the use of "administration," "multinational," "stateless," etc.
- Now contrast that to John Edwards, whose answers tend to use everyday language and deliver a message loud and clear message: "America," "believe," "united."
- The one-issue emphasis of Kucinich and Gravel - "war" -- is obvious.
Obviously, this feature is a bit off-topic for a site devoted to polling methodology, but it does deal with the graphic analysis of political data. I can certainly see potential applications of this sort of graphic for those that conduct and transcribe focus groups and other "qualitative" analysis.
But enough wonkiness. Readers, what do you see in these clouds? Our comment section is wide open...
-- Mark Blumenthal
April 27, 2007 in The 2008 Race
TrackBack
TrackBack URL for this entry:
http://www.pollster.com/cgi-bin/mt/mt-tb.fcgi/3643.
Listed below are links to weblogs that reference Tag Clouds for the Democratic Debate:
Comments
Elliott Larkfield:
A tag cloud for Joe Biden that doesn't include the words "facetious" or "solicitous"? I'm amazed.
Posted on April 27, 2007 12:15 PM
oddjob:
I see that Obama's manner of thinking about things is utterly unlike any of his competitors, and that, as you point out, Biden, Clinton, & Dodd share a "wonky" way of approaching national issues.
Posted on April 27, 2007 12:15 PM
Metadata:
Obama's tag cloud is almost devoid of words with powerful emotional content. Contrast that with Hillary. Edwards isn't bad, either. This is a weakness for Obama. If his words were not powerful, how will people remember him, unless his personal presentation made up for the weak content.
One of the key aspects of right-wing framing is the use of emotionally loaded language, in particular words that code for powerful images.
It would be interesting to observe the tag-cloud for a random George Bush speech, or a random Tom Tancredo speech.
Posted on April 27, 2007 12:29 PM
Keith Murphy:
Obama's tag cloud has many more process words than the others - adverbs vs. nouns. I think it is telling that among his biggest nouns are "families" and "women."
Also telling is the total word count = the amount of time to speak them. The more words, the more time to talk. It seems apparent that MSNBC gave time roughly equivalent to the candidates place in the polls with Obama and Clinton swapped.
Posted on April 27, 2007 1:26 PM
j:
"Constitution" only appears in one of the clouds.
Posted on April 27, 2007 1:32 PM
Mark and Janet:
Thanks for the example and the link to TagCrowd -- I'm definitely going to use this in my statistics classes!
Posted on April 27, 2007 1:33 PM
Tim Blank:
"energy", "oil", "warming" do not appear at all. Very discouraging.
Posted on April 27, 2007 3:25 PM
Paul:
why did you leave out the word "hell" from gravel's tag cloud? he said it maybe 8 times. answers?
Posted on April 27, 2007 3:42 PM
PatriciaV:
Mark, is there way of quantifying the tags, assigning a numerical value to those tags considered more attractive?
Posted on April 27, 2007 3:43 PM
James:
> "energy", "oil", "warming" do not appear at all. Very discouraging.
Obama, Richardson have energy. Kucinich has oil. The tags are in alphabetical order, so even the smaller print words are somewhat decipherable.
Anyway, wouldn't it be dependent upon the questions they were asked?
Posted on April 27, 2007 3:45 PM
J2:
How many times did Richardson use "unsees"?
Posted on April 27, 2007 4:02 PM
Robert P.:
I have trouble believing that word count. Seems off to me.
Also, gee, wonder what is on Hillary's mind? President is as big as a breadbasket. While it doesn't even show up for Dodd.
Posted on April 27, 2007 4:22 PM
Mark Blumenthal:
Paul: We checked, and the problem is that the word "hell" does not appear in the closed captioning of the debate text that Janet used to create the clouds.
Posted on April 27, 2007 4:39 PM
Dave:
This isn't really pertinent to the topic of discussion but Amazon.com has a word cloud option for most of the books on their site. Click the link that says "concordance" and voila!
Posted on April 27, 2007 6:56 PM
blindasabat:
How does this sizing thing work if people have set minimum font sizes for their browsers? Is this cloud sizing images or controlled by CSS? (Ask a Standards Oriented Webdesigner if you don't understand this question ;) )
For example, I have my browser set to never use less than 16 point.
Posted on April 28, 2007 2:24 AM
We will definitely be using this at our www.DebateScoop.com site and, I imagine, in the profession of communication studies more generally.
Thanks for the heads up!
Posted on April 28, 2007 2:34 AM
AJ:
RobertP: I suspect Hillary uses "President" so often because some one decided she needs to seem "Presidential."
I agree with other commenters that Obama's cloud seems very process oriented, though the prominence of "around" may be a verbal tick.
Posted on April 28, 2007 8:28 AM
Jason Platt:
I think people who are commenting are taking this WAY too seriously. Words by themselves are meaningless, it's the ideas they are part of that really count. I'm not saying there is NO value in analyzing the words by themselves, but that value pales in comparison to the ideas and actions they express.
Posted on April 28, 2007 9:32 AM
Jill Martindale:
I wonder, would the readers who have posted already have had the same comments if the questions asked, or better yet, the tag clouds for the questions asked had been posted as well? We must remember that the questions posed tend to dictate the candidates' answers. Could tag clouds influence the speech of our candidates and public oficials in the same way that sound bites do today?
Posted on April 28, 2007 11:02 AM
Jason:
This reminds me of the tag clouds people make on the Many Eyes site. There was a great one with
Gonzales' testimony.
The cool thing about Many Eyes is that you can see two-word phrases, which gives you a whole other view of the text.
Posted on April 28, 2007 5:54 PM
Shaking fist:
Note that Kucinich's is the only one that mentioned the constitution.
Posted on April 30, 2007 1:48 AM
Shane:
Keith,
Each candidate received the same amount of time. NBC used special software to keep track of this so Williams could more fairly balance things.
The fact that Obama said 1900 words and Gravel said 700 isn't NBC's fault.
Posted on April 30, 2007 8:41 AM
Shane:
Shaking,
Kucinich is the only one that mentioned "constitution" enough for the word to make it into his top-50. It's entirely possible that Obama, for example, said "constitution" more times than Kucinich but if he said 50 other words more often than he said "constitution," it won't be on this tag cloud.
Posted on April 30, 2007 8:43 AM
Shane:
Blind,
The clouds are images, not text. This means that your browser will display them the same as my browser. This is evident when you r-click on it or attempt to highlight text in the "cloud"
Posted on April 30, 2007 8:47 AM
Renee:
I find it VERY interesting that the word figuring most prominently in Hilary's speech was PRESIDENT. How's that for subliminal messaging!
Posted on May 1, 2007 6:59 PM
Jasonic:
Chirag Mehta developed a lovely "Tagline Generator - Timeline-based Tag Clouds".
The tool [php script] is available for others to install and use. Be sure to play around with his splendid interactive demos:
Simple Demo - Mini US Presidential Tag Cloud
Popular version - US Presidential Speeches Tag Cloud
http://chir.ag/phernalia/preztags/
Posted on May 3, 2007 12:28 PM
waitingtoderail:
Hmmm----couldn't you have moved the words "support" and "terrorism" away from each other in Dodd's cloud? Just sayin'.
Posted on May 5, 2007 5:16 AM
Robert:
I see weasels, pigs, snakes, wolves......a veritable zoo.
Gee I can't wait to see whats in the Republican clouds. More of the same no doubt! - lol
Posted on May 5, 2007 5:17 AM
Oh come on. Gravel is not a one-issue guy. That's all he had time to talk about!
Posted on May 5, 2007 5:17 AM
Nick:
I think the word clouds are revealing, regardless of context. The questions asked seem somewhat less relevant, as most of the canidates seemed to pick words consistent with their talking points thus far. Hillary stressing "President", Edwards with "Health Care" and "United", and Kucinich with war.
Posted on May 12, 2007 5:15 PM
dynamic pic cloud of the candidates:
http://www.chainofthoughts.com/cgi-bin/pic.cgi?n=race%20for%202008
associated tag cloud of the candidates:
http://www.chainofthoughts.com/cgi-bin/ex.cgi?n=race%20for%202008
Posted on May 16, 2007 7:29 AM
Metadata:
The assertion that only Republicans use emotional words in their speeches is just exasperating. They're just different words, dummy. :o)
I'm just SHOCKED that the word "fight" didn't show up in Edwards' cloud. I would have bet everything I own on my hearing it at least 12 times. Perhaps I was just dreaming.
Posted on July 25, 2007 3:39 PM
Bookbird:
I find it interesting that only one candidate had words relating to education... and the word education does not show up anywhere. Is this really not a top concern for these candidates or did the questions not provide an opportunity to discuss this topic?
Posted on August 20, 2007 1:29 PM
Voxpublica:
When I think back to what Obama says, the word that sticks with me, and with many Americans, I believe, is HOPE. After the last 16 years, Americans are ready to believe is their country rather than sit around and bicker like the hard left and the hard right. We need to work toward what we have in common and tolerate what we don't. I don't boycott a grocery store because they sell souse!
Posted on January 4, 2008 10:26 AM
Voxpublica:
When I think back to what Obama says, the word that sticks with me, and with many Americans, I believe, is HOPE. After the last 16 years, Americans are ready to believe is their country rather than sit around and bicker like the hard left and the hard right. We need to work toward what we have in common and tolerate what we don't. I don't boycott a grocery store because they sell souse!
Posted on January 4, 2008 10:29 AM
susan78681:
Clinton: Im READY to be PRESIDENT and HEALTH CARE in my ADMINISTRATION will take YEARS because of the WAR I did/didnt support.
OBAMA: I SURE am GOING AROUND showing that WOMEN are INTELLIGENT. FAMILES are MAKING DECISIONS on HEALTH,the REGION and SECURITY. TERRORISTS will go AWAY.
Edwards: AMERICA BELIVES me for PRESIDENT of the UNITED STATES. HEATH care is an IMPORTANT QUESTION.
Posted on January 7, 2008 1:06 PM
Very useful rapidshare search. It includes over 4 000 000 files. Fileshunt.com is a best search engine designed to search files in various file sharing and uploading sites. My favorite, rapidshare search engine is Fileshunt.com it`s the most powerful an easy to use. :)
Posted on March 9, 2008 11:48 AM
Post a comment
08 STATEWIDE PRIMARIES
US
Dem,
Rep
Upcoming
Puerto Rico (6/1)
Montana (6/3)
South Dakota (6/3)
2008 POLL DATA
Pres General Election:
McCain vs
US: Clinton, Obama
AL: Clinton, Obama
AK: Clinton, Obama
AZ: Clinton, Obama
AR: Clinton, Obama
CA: Clinton, Obama
CO: Clinton, Obama
CT: Clinton, Obama
DE: Clinton, Obama
FL: Clinton, Obama
GA: Clinton, Obama
HI: Clinton, Obama
ID: Clinton, Obama
IL: Clinton, Obama
IN: Clinton, Obama
IA: Clinton, Obama
KS: Clinton, Obama
KY: Clinton, Obama
LA: Clinton, Obama
ME: Clinton, Obama
MD: Clinton, Obama
MA: Clinton, Obama
MI: Clinton, Obama
MN: Clinton, Obama
MS: Clinton, Obama
MO: Clinton, Obama
MT: Clinton, Obama
NE: Clinton, Obama
NV: Clinton, Obama
NH: Clinton, Obama
NJ: Clinton, Obama
NM: Clinton, Obama
NY: Clinton, Obama
NC: Clinton, Obama
ND: Clinton, Obama
OH: Clinton, Obama
OK: Clinton, Obama
OR: Clinton, Obama
PA: Clinton, Obama
RI: Clinton, Obama
SC: Clinton, Obama
SD: Clinton, Obama
TN: Clinton, Obama
TX: Clinton, Obama
UT: Clinton, Obama
VT: Clinton, Obama
VA: Clinton, Obama
WA: Clinton, Obama
WV: Clinton, Obama
WI: Clinton, Obama
WY: Clinton, Obama
Gov General Election:
Indiana
North Carolina
Washington
Sen General Election:
Alaska
Colorado
Kentucky
Louisiana
Maine
Minnesota
Mississippi (Special)
Nebraska
New Hampshire
New Mexico (Pearce)
New Mexico (Wilson)
North Carolina
Oregon
Texas
Virginia
PUBLIC POLLSTERS
ABC News
AP-IPSOS
CBS News
Democracy Corps (D)
Diageo/Hotline Poll
Economist/YouGov
EPIC/MRA
The Field Poll
FOX News
GWU/Battleground
Gallup
Harris Interactive
IBD/TIPP
ICR - International Communications Research
LA Times/Bloomberg
Mason Dixon Polling and Research
Marist Poll
Market Shares Corporation
Mitchell Interactive
NBC/Wall Street Journal
New York Times
Opinion Research Corporation
Pew Research Center
Polimetrix
Princeton Survey Research Associates International
Public Agenda
Public Policy Polling
Quinnipiac University Poll
Rasmussen Reports
Selzer & Company
Suffolk University Political Research Center
Survey USA
Time/SRBI
Washington Post
World Public Opinion
Zogby International
POLL BLOGS AND SITES
Political Arithmetik
Crosstabs.org
The Polling Report
Electoral-Vote.com
R. Chung's Graphics
Prof. Wang's State Poll Meta-Analysis
Prof. Pollkatz Pool of Polls
Slate: Election Scorecard
Public Opinion Pros
Frank Newport: Gallup Guru
Carl Bialik: The Numbers Guy
Poll Positions: Kathy Frankovic
The Numbers: Gary Langer
Washington Post: Behind the Numbers
SURVEY RESEARCH ORGANIZATIONS
American Association for Public Opinion Research (AAPOR)
The National Council on Public Polls (NCPP)
Council of American Survey Research Organizations (CASRO)
The World Association for Public Opinion Research (WAPOR)
The Council for Marketing and Opinion Research (CMOR)
Marketing Research Association
ARCHIVES
May 18, 2008 - May 24, 2008
May 11, 2008 - May 17, 2008
May 4, 2008 - May 10, 2008
April 27, 2008 - May 3, 2008
April 20, 2008 - April 26, 2008
April 13, 2008 - April 19, 2008
April 6, 2008 - April 12, 2008
March 30, 2008 - April 5, 2008
March 23, 2008 - March 29, 2008
March 16, 2008 - March 22, 2008
All pollster.com archives
MysteryPollster.com archives














mary madderly:
Shouldn't they have rounded edges rather than straight?
Posted on April 27, 2007 11:50 AM