- Ran logistic and ordinary least squares regressions
- Binary dependent variable of whether the snopee’s reply to the snoper’s tweets are accepted 

University of Michigan - Ann Arbor

- Identifying snopes 
- Types of tweets
- Characterizing snoper-snopee relationships
- Categorizing snopes and responses
- Analysis 
- Results

Political fact-checking on Twitter: When do corrections have an effect? - Study 1

Margolin, D. B., Hannak, A., & Weber, I. (2018). Political fact-checking on Twitter: When do corrections have an effect?. Political Communication, 35(2), 196-219. https://www.tandfonline.com/doi/full/10.1080/10584609.2017.1334018?casa_token=xftR1jtjkdMAAAAA%3ApRSWcFHJ58eoBesKnxkoMR8ri-lVeQWoLPakXv9i_aZri4yWIDtjWBfYeEMOBso1kDpqcja-yXD4LQ 

Political fact-checking on Twitter: When do corrections have an effect?

- Collected tweets from January 2012 to April 2014 that either respond to other users or reference a fact-checking website (Snopes.com, FactCheck.org, or PolitiFact.com)
- Only wanted correction tweets about US politics 
- Tweets obtained from Snopes.com and FactCheck.org had “Obama” or “Romney” in their URLs

Political fact-checking on Twitter: When do corrections have an effect? - Identifying Snopes

- Tweet A0: “snoped” tweet, original snopee tweet containing misinformation
- Tweet B1: “snoping” tweet, correction tweet snoper sent to snopee
- Tweet A2: snopee’s response to the “snoping” tweet 
- Collected 2,044 snopes (A0 tweets), 542 are replies from snopees (A2 tweets)
- Only studied the 451 English tweets where the snopee actually responded to the snoper 

Political fact-checking on Twitter: When do corrections have an effect? - Types of Tweets

- Mutual friends – mutually follow each other 
- Strangers – either follows each other 
- Snoper follows snopee
- Snopee follows snoper 
- Gathered data on the number of followers snopees and snopers have, as well as the number of mutual followers that they have 

Political fact-checking on Twitter: When do corrections have an effect? - Characterizing snoper-snopee relationships

- Non-correcting snope – tweets that simply stated facts or provided irrelevant information such as agreeing to the misinformation rather than corrections
- Accepts – snopees agreed with the corrected information snopers provided  
- Ambiguous or other – snopees does not agree with the corrections or are indifferent 

Political fact-checking on Twitter: When do corrections have an effect? - Categorizing snopes and responses

Political fact-checking on Twitter: When do corrections have an effect? - Analysis

- 322 of the tweets showed agreement between snopee and snoper, but 93 were categorized as non-correction tweets and only 229 were correction tweets
- Out of the 229 correction tweets, 135 (59%) were accepted by the snopee 
- People are more likely to respond to mutual friends (44%) than complete strangers (24%)
- Acceptance rate of 39% when both snopees and snopers are strangers, 73% when mutual friends, 64% when snopees follow snopers, and 79% when snopers follow snopees 
- Acceptance rate is increased most (by 3.4 times) when only the snoper follows the snopee 
- Higher acceptance rate when the number of mutual followers is greater

Learn Before

Related