Numbers, Facts and Trends Shaping Your World
About
Follow
My Account
Log in
View Account
Log out
Donate
Internet & Technology
Menu
Research Areas
Home
U.S. Politics
Media & News
Social Trends
Religion
Internet & Tech
Science
Hispanics
Global
Methods
Publications
Topics
Presentations
Datasets
Interactives
Fact Sheets
Experts
April 5, 2018
Bots in the Twittersphere
Automated accounts post the majority of tweeted links to popular websites across a range of domains
Next
→
←
Previous
1
2
3
4
5
6
7
8
9
10
11
12
Download
Automated accounts post the majority of tweeted links to popular websites across a range of domains
Embed
Report Materials
Complete Report PDF
Full Report
Bots in the Twittersphere
Report Infographics
In a sample of 206 accounts coded by a third party, a threshold near 0.43 produces the highest level of accuracy in classifying bot accounts
Verified Twitter accounts have little impact on numbers of automated accounts shared
Including or excluding verified accounts exceeding a Botometer score of 0.43 does not affect results much
Random effects analysis of bot shares by content
Random effects analysis of bot shares by audience score
The most accurate Botometer threshold for classifying bot accounts is below 0.5, based on a sample of 100 hand-coded accounts
Inter-coder testing scores
PI_2018.04.09_Twitter-Bots_M-01
Suspected bot accounts share more links to popular political sites with an ideologically centrist or mixed audience
Automated accounts post the majority of tweeted links to popular websites across a range of domains
The most-active Twitter bots produce a large share of the links to popular news and current events websites
Automated Twitter accounts post the vast majority of tweeted links to popular news and current events sites that do not offer original reporting
Copyright 2019 Pew Research Center
About
Terms & Conditions
Privacy Policy
Reprints, Permissions & Use Policy
Feedback
Careers