How the Twittersphere Helped Donald Trump Win
Ask about Knowledge at Wharton
Episode Description
Wharton research shows that sensational tweets have staying power, aiding Donald Trump’s ascent. More: http://knlg.net/2kn1y97 📂 Technology 🏷️ Research
Read the full transcriptHide the full transcript
- make America tweet again that's the provocative title of a new research paper out of won that looked at the
- Tweet surrounding the 2016 Republican primary debates here to talk about the paper and its surprising findings are
- Wharton Professor Ron Burman and doctoral candidate Coleman Humphrey so welcome thank you deorah so tell me why
- did you decide to analyze Twitter versus other types of communication so that's actually a very interesting story so our
- co-authors professor Robert mayor and shirim maluma do a PD candidate at Columbia business school they were
- starting to work on a project using Twitter data um and the focus was how emotionality and the device he use
- changes the patterns but then uh we realized that uh in um elections and debates what happens is that people
- today use Twitter as the main forum for pushing their opinions and debating with
- everyone Etc so this has become like the Twitter election and as a result we wanted to focus very much to see what do
- voters think given what they see on Twitter but also what do voters do on Twitter and how they interpret that now
- we this was about in August at that time it was clear that Trump has a big advantage on Twitter and he's uh getting
- a lot of uh voter opinion and voter advantage through Twitter so combined with that it was very very interesting
- to look at Twitter as does it reflect voter opinion but also can you use Twitter to also influence voter opinion
- and how that works during a debate and can you give us an overview a quick overview of how you conducted the
- research well sure so first um to decide what data to collect we decided to focus
- on three very pivotal debates so the first one was the August debate which was the most viewed primary debate of
- all time um and people were clamoring to see Trump versus bush and who would do better we also um had the February
- debate um which was just before super Tuesday so it was kind of like the last chance for uh one of the other
- candidates to try and derail Trump and his uh amazing poll numbers and then finally the March debate which um people
- were buzzing about because it was back U Trump versus Megan Kelly the the second
- time I guess so within there we collected all the relevant uh tweets from all debates so from an hour
- beforehand to uh two hours after the debates just all tweets that that had the main hashtag in them so we knew they
- were tweeting about the debate uh and from there from this data we were able to you know we have um who tweeted and
- who retweeted what time that all happened at and we also then we applied this program called Luke which um it
- finds was there a negative tone in the Tweet was there a positive tone did you use like um oops sorry did you use
- personal pronouns um was a kind of uh based on like was it a power tweet or a reward tweet and then for a subset of
- the data the kind of the most popular tweets we also sent them off to MK to get humans to look a little bit further
- at the tweets we got um humans to decide is there a picture or a video in the Tweet or does it contain humor or
- sarcasm uh as well so what are the key findings of your paper um so so we found a a few very interesting things some of
- them just didn't conform with past research results and some of them were just brand new the first thing we looked
- at was um what do people see during the debate do they mostly see new tweets or do they
- see more and more old tweets because of retweets of other people and we found that over time you basically have a more
- dated view of of the debate which means if you look very towards the end of the debate actually you're going to see
- mostly tweets from very early on in the debate the second thing we looked at was
- what was the sentiment during the debate uh for different candidates and we found
- that if you looked after the debate versus during the debate you would get a very different view of the sentiment for
- so example for Donald Trump the uh sentiment after the debate was uh very positive while during the debate
- sometimes there were controversies maybe Sensations that generated a negative um
- sentiment uh and finally another thing we noticed is that the Tweet stream becomes more and more Sensational and
- less substantive and in the words of of Bob Mayer another professor o on this colleagu is that during the debate you
- get the New York Times after the debate you getting the New York Post and this is something thing we found out which
- was very interesting do you think this explains why Trump defied expectations to get the Republican nomination for
- president sure this we feel like this what Rono said definitely speaks to um to one of the reasons why perhaps he was
- so popular um Twitter kind of like a lot of other media I guess just really focuses on tabloid or Sensational stuff
- especially afterwards like that's that's what sticks um one of the things we analyzed was was sentiment people
- positive or negative towards certain candidates and you can see for Trump that even as something like the Trump
- University is being discussed um like quite heavily and he's really getting hammered for it in the debate his
- sentiment just like doesn't really seem to drop like uh and and especially after
- debates you can see his sentiment actually goes right back up and and is generally positive after the debate so
- people just really don't care about this this the negative stuff he says just doesn't seem to matter um and again
- building on on what Ron's saying it seems like Trump is the perfect cidate for the the Twitter election um since he
- is so used to generating like Sensational um news and and generating controversy and kind of leading towards
- all the Tabloid stuff and and also avoiding policy where he's probably weakest out of all the candidates on the
- stage and that's what sticks so absolutely um Twitter seems to seems to be the medium the medium for him so what
- are the implications of your findings on voter opinions so there are a few implications
- we think are are playing like are play there there there implications on trying to to gauge or to understand voter
- opinions and and one of the findings is that if you analyze Twitter data you might get the wrong idea so if you look
- during the debate maybe you'll get one sentiment after the debate you might get a different sentiment so it really
- matters when you look at the debate and the other thing is um that voters who there are some voters who are very
- activing the debate and watch it and tweet about it Etc but most of the people probably open the newspaper the
- next morning or open their Twitter and say what happened during the debate yesterday and then they look and they
- see this backwards view which means the influence on them is probably much stronger U than what you would get if
- you watch during the debate one for example one thing we notice is that tweets that came from news sources
- during the debates were very influential during the debate and were retweeted a lot Etc but after the debate they were
- kind of going down if you think about it it used to be that there was a debate and news sources maybe would uh um kind
- of summarize or comment on the results after that and you would get what happened Deb during the debate in in a
- newspaper after that if you look at Twitter you would get a very different idea which means that perhaps Twitter is
- not either a good measurement or maybe Twitter is a reflecting uh Forum of the results and the news media is not a good
- measurement and reflecting uh Forum of the results which might cause surprises later
- on did any conclusions surprise you so one interesting uh finding we had which was surprising was how media was almost
- um inactive during the debate in actually commenting on it what they were doing is picking up mostly quotes from
- TV and just tweeting them they weren't saying who they think is winning they weren't commenting on facts they were
- just quoting and as a result during the debate people were retweeting them a lot
- but after the debate they had very very very little influence this was one surprise finding the second one is that
- for many tweets so this comes to academic research typically we assume there's something called a diffusion
- pattern you would have an idea that goes up gets retweeted and goes down and dies
- for many things we saw they barely had any influence during the debate they basically died during the debate and
- then had the Resurgence after the debate so for example um this um basically fight almost between Megan Kelly and
- Trump in August at the beginning generated some attention during the debate but then died off and but then
- after the debate it became the primary thing everyone was talking about and just building on that what was also
- perhaps surprising is that strong positive or negative emotions didn't didn't help much at all in fact what
- actually in in our case hurt the popularity of tweets both before and after the debate um and just kind of
- building on what we've been saying like being being everything being so Sensational and not being the most
- popular stuff um uh like it's it was especially clear afterwards so for example beforehand one thing that
- actually did stir some interest was the economy when people some people did tweet about the economy but afterwards
- it was just nothing there was just no no substantial policy not even policy that
- was popular during the debate um this was so it was a bit almost surprisingly sad that
- Twitter you might expect that Twitter is finally the way for everyone to debate and argue and and maybe get an
- understanding of of the issues but actually everyone towards the end was focused more on Sensations which means
- maybe Twitter is not the place you want to look at and kind of understand what is important in this election at at
- least in terms of the issues so how is your research different from prior work in this area um so so research is
- different on both methodology and a bit on findings so methodology it's interesting because we looked both at
- who tweets and who retweets and who gets gets followed by whom but also we looked
- at the sentiment and the topic of the tweets so this is one methodology difference and another one is the
- setting most people that analyzed uh tweets were looking at you know are people tweeting more about uh pop
- culture versus music or something like that if it's an event which is very active and you get more than 200 tweets
- a m a second right a second then there's this fight for attention and everyone is
- trying to say something this creates a very different Dynamic um and we think this is very unique because no one has
- has previously analyzed Tweets in a very short period of time in a very um kind of energetic active uh event so how will
- you follow up this research so building on what Ron just said we we or others might hope to um analyze other similar
- situations such as um sports or um any any other situation where it's very highly focused not over a long period of
- time um another area of focus we're interested in is um this the phenomenon of a lot of tweets being picked up by
- this by different users and what we mean by that is not just that something is retweeted but is is tweeted like not a
- retweet an original tweet um as opposed to just being like a very popular tweet like someone says something funny and
- everyone thinks it's hilarious and retweets so one example is is um quotes So sometimes a quote is so
- instantaneously interesting that everyone picks it up and just to give you one example from the debate U Mike
- hucke in one of the debates said um uh the purpose of the military is to kill people and break things and and this was
- retweeted this was tweeted excuse me not retweeted tweeted out by just over a 100
- people just immediately people were thought this was fascinating it actually wasn't that popular of of an idea like
- it didn't um last for ages and ages afterwards but it was initially very popular as opposed to this some quotes
- are only interesting because they get brought up by a person in context so actually in the same example um Pat and
- oswal um tweeted out um kill people and break things Mike Cooke just described the GOP and that itself is a very
- popular tweet and there are definitely some quotes that get picked up in that way just someone someone gives them more
- context that becomes super popular or sometimes just the quote itself just goes on and that's not the only thing we
- have um you know token tweets like who won is of course very popular but but it's it's interesting to see that in
- certain circumstances some tweets are either completely ignored of course some are picked up because one one user
- tweets something funny and that gets retweet a lot and sometimes just one idea or one moment who just gets picked
- up so much like I guess another moment that was a big deal was um uh the debate between Rubio and Trump's uh who had the
- bigger uh hands so to speak so these are and the other thing we didn't analyze so this was um a
- debate where people were trying to VI for the Republican nominations so these are not opposing ideas or issues this is
- just who gets the most attention we also have data from um the the primary debates right between Hillary Clinton
- and Donald Trump and I think it would be very interesting to understand if you had the same phenomenons and the same
- patterns there and would it be would have have been able to allow you to predict maybe who's going to win or or
- do issues matter or Sensations matter more and basically applying the same findings from this research to another
- data set to see if you get the same things does it generalize or is it very very specific well we can't wait to find
- out what you discover in your new research but thank you very much for joining us today thank you very much d
- [Music]
Brief Summary
Wharton researchers analyze Twitter's influence on voter sentiment during the 2016 Republican primary debates.
Key points
- Make America Tweet Again. A new research paper explores the impact of Twitter on the 2016 Republican primary debates.
- Surprising Findings on Sentiment. The study reveals how sentiment changes before and after debates, particularly for Trump.
- The Role of Media During Debates. Media's influence wanes post-debate, focusing more on sensational quotes than analysis.
Episode highlights
- Make America Tweet AgainA new research paper explores the impact of Twitter on the 2016 Republican primary debates.“Twitter has become like the Twitter election.”1:00Feb 3, 2017
- Surprising Findings on SentimentThe study reveals how sentiment changes before and after debates, particularly for Trump.“Sentiment after the debate was very positive while during the debate it was negative.”4:01Feb 3, 2017
- The Role of Media During DebatesMedia's influence wanes post-debate, focusing more on sensational quotes than analysis.“Media was almost inactive during the debate, just quoting TV.”7:35Feb 3, 2017
Tension
45/ 100Some tension





