Week 3

Question 1

Definition Computational Politics

Answer 1

Applying computational methods to large datasets derived from online and offline data sources for conducting outreach, persuasion and mobilization

Question 2

Computational Politics (6 bulletpoints)

Answer 2

1. Big Data
2. Emerging computational methods
3. Modeling
4. Behavioral Science
5. Experimental science in real-time environments
6. Power of algorithmic platforms

Question 3

Big Data

Answer 3

1. increase in amount and variety of data on each individual (microscopes telescope)
2. Amount: increase in depth and reach
3. Nature: user generated data & latent data (cfr. explicit vs implicit data)
4. Added up with data from data brokers
5. Databases with large amounts of datapoints per individual

Question 4

Emerging computational methods

Answer 4

1. Developments in storage and database systems
2. Extraction of semantic information
3. Social network analysis
4. Correlation data analysis

Question 5

Modeling

Answer 5

1. Predicting new information through computational data analyisis
2. Model individual attributes without asking the voter direct questions (with accuracy)
3. Subtle persuasion - the voter has no idea about this modeling

Question 6

Behavioral science

Answer 6

1. Rational voter
2. Models and theory on how to persuade & influence
3. Psychographic data

Question 7

Experimental science

Answer 7

1. Stepping aside from “good feeling”
2. Large, real-time field experiments (became cheap)
3. Randomized A/B testing

Question 8

Power of algorithmic platforms

Answer 8

1. News feed algorithm (organic content)

2. Sponsored or promoted algorithm (paid content)

Question 9

News feed algorithm

Answer 9

• organic content
• proprietary algorithms (undisclosed)
• result: the feed is driven by opaque algorithms (black box)
• users see different political content, but we don’t know why

Question 10

Sponsored or promoted content

Answer 10

• paid content
• brands (political actors) can tailor a message to specific individuals (more control about who sees what
• political micro targeting

Question 11

Results study political microtargeting

Answer 11

• Extraverted people will be more persuaded when the ad-voked affect is positive (enthusiasm)
• Introverted people will be more persuaded when the ad-voked affect is negative (fear)
• Message elaboration was the mediator

Question 12

Promises political microtargeting

Answer 12

Democratic society

• increase political participation
• increase political knowledge
• reach hard-to-reach societal groups

Individual voter

• receive relevant political messages
• easier to get a sense of standpoints of political parties

Question 13

Threats political micro targeting

Answer 13

Democratic society

• Power transfer towards well funded parties
• fragmented public sphere
• difficult to fact-check or counter (fake news)

Individual voter

• manipulation/bias perception of voters
• privacy threats
• tapping into human weaknesses and vulnerabilities

Question 14

Higher ranked results generate…

Answer 14

1. more fixations in eye-tracking research
2. more clicks (91,5% on the first page, 32,5% on the first result)
3. more time spent on webpages

Question 15

Biased candidates had… (study Epstein)

Answer 15

1. higher trust
2. higher liking
3. better overall impression

Question 16

Conclusion study Epstein

Answer 16

1. Biased search rankings can shift the voting preferences of undecided voters by 20% or more
2. Without their awareness (covert algorithm) - subtle persuasion
3. Ranking algorithms can be a tool influencing elections

Question 17

Google’s algorithms & politics

Answer 17

• Search results: the set of sites ranked on the first page of search results
• Visual framing: candidates are visually framed in search results as a consequence of the image selections
• In the news: the presentation of news information about each candidate as framed in the “In The News” section

Question 18

Political bots (what, why, how, who)

Answer 18

What: social media accounts equipped with algorithms that post, tweet, or message of their own accord (mimic human user)
Why: to influence or manipulate the public opinion
How: by spreading propaganda in support of, or against particular issues or people
Who: an increasing amount of political candidates in election campaigns

Question 19

How work political bots (6 bulletpoints)

Answer 19

1. Write or acces a pre-made script for a bot
2. Automatically setting up an account
3. Mimicking an actual person
4. Crawling through content and scanning posts on social network (observe the environment)
5. Posting content to engage with human users
6. Network of bots act together (botnet)

Question 20

Tactics of political bots during elections

Answer 20

1. Zombie electioneering = gives the appearance of broad support for an issue or candidate
2. AstroTurf campaigning = makes an electoral or legislative campaign appear to be a grassroots effort
3. Hashtag Hijacking = appropriates an opponents hashtag to distribute spam or otherwise undermine support
4. Retweet storm = simultaneous reposts or retweets of a post by hundreds or thousand of bots