• Media and Violence
  • Persuasion
• Evidence:
  • Nazi Radio
  • Russian Social Media
  • Trump on Twitter

Persuasion

Framing of events might change how people perceive/what people believe…

• what “problems” are there in society (what moral transgressions have occurred? what moral order ought be imposed?) diagnostic frames
• what are the “solutions” to this problem? prognostic frames
• what are the “facts” of what has happened? (what frames are credible?)

Coordination

Transaction costs of organizing violence

• How to coordinate collective violence
• How to acquire information/weapons/targets

Consequences of violence

• Social pressure (within social network) to not/do violence
• Supportive voices offsets risk of condemnation

Coordination definitely explains some effects of media on violence:

• Rwandan Genocide:
  • reduced costs of organizing violence (large increases in collective violence)
  • reduced perceived social costs of violence (spillovers effects?)

Is there evidence of persuasion?

Difficulties:

People with pre-existing ethnic bias/hatred select into media messages promoting violence

Did exposure to the radio increase violence against Jews? (Adena et al 2015)

Examine two time periods:

• 1928-1932 (Moderate control over radio)
• 1933-1945 (Nazi control over radio)

During 1920s, radio, a new technology, regulated by Germany government

• When Weimar government ruled, forced moderate, pro-democracy messages on radio
• When Nazis take control, anti-semitic propaganda on the radio

What are the effects?

• on voting for Nazi party?
• on anti-semitic activities?

How do Adena et al get around selection into radio exposure?

• Compare effects of same radio exposure when message changes
• Local, arbitrary variation in exposure to radio signals:
  • exposure to radio signal, net of expected radio given population, proximity to major cities, elevation

In analysis of radio exposure and…

Nazi Voting

Increasing radio exposure when messages are Moderate

\(\rightarrow\) fewer Nazi votes

Increasing radio exposure when messages favor Nazis

\(\rightarrow\) more Nazi Votes

In analysis of radio exposure and…

Antisemitic Behavior (1929-1934)

Increasing radio exposure when messages are Moderate

\(\rightarrow\) less anti-Semitic violence/harassment (not statistically significant)

Increasing radio exposure when messages sent by Nazis

\(\rightarrow\) more anti-Semitic violence/harassment

Effects on voting is likely persuasion, since it is individual choice

Effects on anti-Semitic violence could combination of persuasion and coordination:

• persuasion if it induced more people to adopt anti-Semitic views, perceive imminent threat posed by Jews
• coordination if it facilitated rallies, “normalized” anti-Semitic violence

Coordination or Persuasion?

Antisemitic Behavior (Late 1930s)

Increasing radio exposure when messages favor Nazis

\(\rightarrow\)

• increased denunciations/deportations of Jews
• increased letters to anti-Semitic newspapers
• no effect on attacks on Synagogues

Persuasion or Coordination?

1. No effects on attacks on synagogues (collective action)

2. Exposure to radio caused more denunciations of Jews, open letters to Nazi newspapers (individual actions)

Persuasion or Coordination?

1. Effects of Nazi radio different in places with different histories of anti-Semitism

14th century pogroms against Jews predict Nazi-era violence, and

• Radio exposure increased anti-Semitism in places with medieval anti-Jewish pogroms
• Radio exposure decreased anti-Semitism in places without medieval anti-Jewish pogroms

Most likely Persuasion and Dissuasion

persuasion: frames are resonant (consonant) with local norms and lead to violence, more anti-Semitism

• Nazi radio elicits anti-Semitic violence in places with history of antisemitism

dissuasion: frames that are dissonant with local norms provoke a backlash, less anti-semitism

• Nazi radio elicits less anti-Semitic violence in places with history of Jewish inclusion

Serbian Radio in Croatia

Della Vigna et al (2014) examine local variation in exposure of Croation communities to Serbian nationalist Radio due to topography

Croatian villages with greater exposure to Serbian nationalist Radio see:

• Increased vote for extremist Croatian parties
• Increased anti-Serb graffiti

Must be dissuasion, backlash against messages, not coordination.

Bursztyn et al (2019) examine the effect of exposure to social media networks on:

• incidence of ethnic hate crimes
• prevalence of ethnic hatred

In Russia, main social media platform is VKontakte (VK)

• similar to Facebook; launched in 2006.
• 90% of social media users in Russia by 2011.

We could compare cities with more VK users vs those with less…

• but, how do we know it is VK that drives differences? Could be other reasons cities have more VK users, have different rates of hate crimes.

“Random” exposure to VK social network:

• founder attended Saint Petersburg State University in 2006
• first users were students attending this University at the same time, their friends and families back home
• cities which sent more students to SPSU during 2006 have more VK users between 2007 and 2016. No relationship between VK users and cities that sent more students before/after 2006.

Effects on Hate Crimes

Hate crimes with ethnic motivation increased in places with more VK users:

• increased more where pre-existing support for xenophobic political parties
• increased hate crimes with multiple perpetrators

This is consistent with coordination.

• people who share xenophobic views, able to collaborate

Effects on Ethnic Hatred

Ethnic hatred increased in cities with more VK users.

• measure ethnic hatred using list experiment (reveals hidden beliefs)
• effects are strongest for men, younger people, people with lower education
• no effects on women
• does not appear to affect public expression of ethnic hatred

This suggests there also is persuasion:

• personal beliefs about ethnic hatred increased
• but we don’t know if those changes in beliefs \(\to\) violence

Mueller and Schwarz (2020) examine the effect of anti-Muslim tweets by Donald Trump on hate-crimes against Muslims.

We can’t observe the US in the absence of Trump tweeting against Muslims, so authors use correlation…

Trump’s Twitter gained attention as he ran for President.

Trump made nearly 300 negative tweets about Muslims.

When Trump gained prominence, anti-Muslim hate crimes increased

Trump’s anti-Muslim tweets, anti-Muslim hate-crimes track each other over time

Nevertheless…

• could be that something other than Trump’s Twitter changed between 2014 and 2015 that increased hate crimes
• Days with more Trump Tweets could be different in other ways; maybe news cycle focused on ISIS terrorism, driving both Trump tweets and hate crimes.
• Could compare changes in hate crime in communities with more vs less twitter users, but… why do they have more twitter users?

“Random” exposure to Twitter due to early joiners in 2007

• Twitter offered incentives to attendees at 2007 SXSW music festival
• Places with SXSW attendees in 2007 have more Twitter users for years afterward
• SXSW attendance likely unrelated to hate crimes.

Counties with more SXSW Twitter Joiners (treated) see larger increase in hate crimes following rise of Trump’s Twitter

Days when Trump golfed followed by more anti-Muslim tweets. These are Hate tweets unrelated to the news cycle.

Mueller and Schwarz compare hate crimes counties with more/fewer Twitter users when Trump tweets about Muslims.

• Compares changes in counties, over time.
• Any patterns cannot be explained by…
  • unchanging differences between counties (same counties over time)
  • different news events (same dates, tweets induced by golfing)
  • different trends in counties with high/low Twitter usage (SXSW attendees)

Trump anti-Muslim tweets (induced by golfing) increase anti-Muslim hate crimes in places with more SXSW Twitter users

Coordination or Persuasion?

No changes in bias against Muslims in places with more twitter users, after Trump gains fame on Twitter

Coordination or Persuasion?

• Effects of tweets only for hate crimes with solo assailants

Coordination or Persuasion?

• Effects of twitter users only in places with hate groups, history of hate crimes.

Coordination or Persuasion?

Twitter effects…

• Likely not coordination in planning violence, b/c effects are on solo attacks, in places where capacity for violence already present
• likely not persuasion, b/c no effects on anti-Muslim attitudes
• possibly coordination effects on perceived social costs of violence

Media may cause ethnic violence by…

• reducing costs to perpetrators to planning and organization violence
• increasing the perceived social acceptability of violence (social cost)
• persuading people to use violence against ethnic out-group

Strong empirical evidence that all three mechanisms operate in the real world.