The TV debate: What did Twitterers think and feel?
So we've seen a new record for a political event: last night's debate was the most-tweeted ever. Over 180,000 tweets in 90 minutes by more than 38,000 Twitterers.
"So what?" I hear you say. "Nearly 10 million people watched it on television; those numbers are pretty insignificant." Perhaps, although there may well have been far more people watching Twitter passively than the numbers who actually tweeted.
An analysis of what was being said on Twitter as the debate unfolded comes up with some interesting conclusions. We asked a firm called Lexalytics to put its "sentiment-analysis" machine to work on the debate last night. In the end, we decided to narrow the results down to what was said on Twitter - it produced far more data than any other source. Here's the chart:
What you're seeing is a chart of the ratio of positive to negative language used about each leader over a series of five-minute periods.
The first thing you notice is that right at the start, at 2040, Gordon Brown dips sharply into negative territory, perhaps during his answer on immigration.
His rating recovers and leaps into positive territory - at about the time that he launched a jokey attack on David Cameron over airbrushed posters.
Mr Cameron falls into negative territory at 2110 somewhere between the constitutional reform and education questions, where all three take a dip. But from then on, Nick Clegg moves into positive territory and stays there for the rest of the debate.
So, just like the other measures of opinion, our Twitter sentiment tracker appears to show that Nick Clegg had the best night as part of its almost-live representation of public reactions. Now, we must be cautious about reading too much into this experiment; the twitterverse is by no means certain to be representative of the electorate as a whole. But this could be a useful addition to the toolbox of any political analyst.
Comment number 1.
At 16th Apr 2010, minuend wrote:Complain about this comment (Comment number 1)
Comment number 2.
At 16th Apr 2010, Lallands Peat Worrier wrote:I wonder if your cunning group of Twitter analysts bothered to differentiate among the great wealth of data where the Twitterers hailed from? After all, as a Scot, much of what they had to say on domestic policy was of at best marginal relevance to my life. Eduction? Healthcare? Most interesting. In the spirit of brotherhood, I have a care and interest in England picking good choices, being well educated and healthy. You can't be treating the whole sample as if they're some homogeneous whole, all taking in the debate in the same way. Investigate the use of the #ScotlandSpeaks hashtag on Twitter for a blindly clear demonstration.
Complain about this comment (Comment number 2)
Comment number 3.
At 16th Apr 2010, Mark - Hove wrote:I liked this debate. It’s a good Idea and we need more like it.
Unfortunately this first debate has left me very very angry. How can a candidate to become PM, or even simply an MP, say something like “I was in Plymouth recently, and a 40-year-old black man made the point to me..". It’s an absolute scandal. This sort of expression is offensive, unacceptable in a decent society and has real consequences for real people.
Complain about this comment (Comment number 3)
Comment number 4.
At 16th Apr 2010, Stuart wrote:Why have you failed to point out the level the leaders started from???? Almost all polls up untill last nights debate have shown Cameron to be the most popular of the three yet before the debate starts cameron is well behind the other two, this clearly shows that the Tweeters are not representative of the population!
Complain about this comment (Comment number 4)
Comment number 5.
At 16th Apr 2010, Neilly wrote:Why do they not all start out at '1' (equal positive to negative) from 20:00?
Complain about this comment (Comment number 5)
Comment number 6.
At 16th Apr 2010, Bryn The Cat wrote:Interesting but let's be honest, the leaders answers were pre-prepared and heavily rehearsed. Not knowing the questions that were coming up, allowing audience interaction and better chairing of the debate would have livened the whole event up. As it stands there was little insight into their policies above what we've heard already. There was very little "New" news.
I think we can over analyse the debate (which is an interesting term to describe the spectacle last night - not sure debate is fully accurate). Sure it gave Clegg good exposure but is that likely to translate into more votes? Who knows. And will those votes be likely to affect the result? Unlikely I suspect as the voting system is stacked against the Lib Dems. This would only really be consequential if we had PR or AV, but with FPTP it will make little difference.
Complain about this comment (Comment number 6)
Comment number 7.
At 16th Apr 2010, Phillip E Jones wrote:Despite coming from a generation that doesn't have a brilliant grasp on the new technologies I have been inspired to join Facebook, Twitter and Dugg in order to participate in this election, hopefully to bring about real change in our country. For me Nick Clegg stands head and shoulders above his opponents and offers Britain a new vision for the future
Complain about this comment (Comment number 7)
Comment number 8.
At 16th Apr 2010, SirBedevere wrote:Twitterers seems like an awfully skewed sample. Wouldn't it be just as useful to find out what foxhunters or coalminers thought of the debate?
Complain about this comment (Comment number 8)
Comment number 9.
At 16th Apr 2010, TimmyNorfolk wrote:Where is the moderator?
Also, do you have any info as to why all 3 were at 0 at 21:20? did all tweeters put the kettle on?
Complain about this comment (Comment number 9)
Comment number 10.
At 16th Apr 2010, Andy Bando wrote:Vote Labour - get taxed; vote Lib Dems - get taxed more. Don't the moronic Britons get it? Thickos really!!
Complain about this comment (Comment number 10)
Comment number 11.
At 16th Apr 2010, _Ewan_ wrote:What's 'zero' on this scale? It's clearly not the sentiment at the start of the debate, and over the time scale chosen David Cameron seems to be the biggest winner, going from -3 to +2.5ish, with the other two mostly finishing where they started.
Complain about this comment (Comment number 11)
Comment number 12.
At 16th Apr 2010, FishFingers wrote:The ³ÉÈËÂÛ̳ Editorial Guidelines state:
We must avoid any undue prominence or giving the impression that we are promoting or endorsing products, organisations or services. To achieve this we must:
- ensure that references to trade names, brand names and slogans are clearly editorially justified.
- not linger on brand names or logos and use verbal references sparingly unless there are very strong journalistic reasons for repeated references to a brand.
Doesn't this post (and the MANY others across ³ÉÈËÂÛ̳ News) about Twitter go completely against these guidelines?
Complain about this comment (Comment number 12)
Comment number 13.
At 16th Apr 2010, PaulK wrote:When will the ³ÉÈËÂÛ̳, along with other media outlets, get over Twitter? The opinions of a few people with time on their hands and access to the internet does not count as a movement... Please, stop it.
Also, once upon a time a blog was known quite simply as an article...
Complain about this comment (Comment number 13)
Comment number 14.
At 16th Apr 2010, barry white wrote:Twitter?
I think in reality as a poll it will prove to have too many variables to be viable in prediction.
After all look at the posts here.
Maybe it just shows off viewing habits instead of how votes will go.
Complain about this comment (Comment number 14)
Comment number 15.
At 16th Apr 2010, Omnibusologist wrote:I realise it might not seem proper for the ³ÉÈËÂÛ̳ to support the idea, but the simple fact of the matter is that the experience was pretty well ruined by the poor production - in particular Mr. Stewart's ineptitude.
Complain about this comment (Comment number 15)
Comment number 16.
At 16th Apr 2010, Nicolapse wrote:It seemed quite clear to me, from watching the live Twitter feed, that Labour had a lot of people on there. I find it puzzling that the ³ÉÈËÂÛ̳ has wasted our money on this totally skewed segment of society.
Complain about this comment (Comment number 16)
Comment number 17.
At 16th Apr 2010, Arnold White wrote:I know online polls are extremely unreliable but they are usually interesting all the same. I have been using a chat site at to discuss the election for a while and their poll has been consistently conservative, labour, libdem. Apart from yesterday evening and today when LibDems strongly overtook Labour.
I think the LibDems have a really strong internet following which could easily alter the election result.
Complain about this comment (Comment number 17)
Comment number 18.
At 16th Apr 2010, John_from_Hendon wrote:Real people don't tweet!
140 characters is insufficient to express any form of nuanced political view and so tweet polls are pointless.
Complain about this comment (Comment number 18)
Comment number 19.
At 16th Apr 2010, Green Soap wrote:Perhaps, although there may well have been far more people watching Twitter passively than the numbers who actually tweeted.
Ahh, the sound of straws being clutched at to justify the product placing, and worshipping of Twitter.
Look, get it clear Rory. Twitter is not important in the grand scale of things here. Many millions more will have talked about it with their workmates, and friends today.
That is the place where it will make any difference.
Complain about this comment (Comment number 19)
Comment number 20.
At 17th Apr 2010, marek chodnik wrote:Wow, this blog has really gone downhill since I last wrote a response.
It's quite comical of Rory to believe that you can automatically accumulate all the tweets on twitter, pile them into some hastily-written algorithm, and turn the handle, to receive an accurate or even vaguely representative graph model depicting public sentiment towards each of the three main party leaders. Quirky and innovative though it might seem, it is totally dubious and unworthy of any technological merit whatsoever.
This blog is a joke. It knows it's a joke. Amidst the blatant product placement (thank you, Tengsted), it tries to combine cutesy apps and web-based social silliness to basically make people laugh, and help the 'Twitterati' perpetuate their uninspired giddiness.
Complain about this comment (Comment number 20)
Comment number 21.
At 17th Apr 2010, JunkkMale wrote:12. At 4:47pm on 16 Apr 2010, FishFingers wrote:
Doesn't this post (and the MANY others across ³ÉÈËÂÛ̳ News) about Twitter go completely against these guidelines?
But it's news!!!! At least, if it can be made that way by some...
/blogs/thereporters/rorycellanjones/2010/04/the_debate_how_it_unfolded_onl.html
Complain about this comment (Comment number 21)
Comment number 22.
At 17th Apr 2010, PJohnston wrote:Camerons first peak on the Torys popular net migration cap policy is inexplicably shadowed by Clegg. Given that Clegg wants an amnesty for illegal immigrants, (and this policy polls as incredibly unpopular), it shows how the format of the "debate" allowed style and substance to blur and attach equal weight.
Complain about this comment (Comment number 22)
Comment number 23.
At 17th Apr 2010, Chris wrote:Twitter is for people who love the sound of their own voice a little too much (or perhaps, love the sight of their own words). 180,000 Tweets in 90 minutes? At an average of more than 30 Tweets every second, who on earth could possibly be reading it all? Answer: Nobody, except rooms full of hacks who are grateful that they no longer have to get their coats on, go out and meet real people when their editor asks for a vox pop.
All this analysis of people's tweeting habits, and the views they express, goes on even though said hacks candidly admit how unsound it is when you can't possibly know who is included in the sample. Lazy, lazy, lazy.
Complain about this comment (Comment number 23)
Comment number 24.
At 17th Apr 2010, ElephantTalk wrote:Enough already of the election via Twitter. Give us some real technology news, before May the seventh preferably.
Complain about this comment (Comment number 24)
Comment number 25.
At 19th Apr 2010, peejkerton wrote:Twitter's irrelevance is directly proportional to the irrelevance of people who comment on here.
So if you find Twitter irrelevant, maybe you should investigate your own comments on various blogs across the web.
Complain about this comment (Comment number 25)
Comment number 26.
At 4th Feb 2011, phkk wrote:We must avoid any undue prominence or giving the impression that we are promoting or endorsing products, organisations or services. To achieve this we must:- ensure that references to trade names, brand names and slogans are clearly editorially justified.- not linger on brand names or logos and use verbal references sparingly unless there are very strong journalistic reasons for repeated references to a brand Doesn't this post (and the MANY others across ³ÉÈËÂÛ̳ News) about Twitter go completely against these guidelines?
Complain about this comment (Comment number 26)