Et markant nybrud er, at den demokratiske samtale flytter fra forsamlingshuset til sociale medier.

Vi mangler viden & data. Derfor præges debatten ofte af myter fx om ekkokamre.

Men vi skal ud af ekkokamret om, ja, ekkokamre.

🧵 om de store spørgmål & forskningens svar. (1/15)
Når man skal forstå et fænomen på sociale medier, fx ekkokamre, misinformation eller had, bør man altid stille 3 spørgsmål:

1) Hvor udbredt er X?
2) Hvad er årsagerne til X?
3) Hvad er effekterne af X?

Hvad siger forskningen om de 3 spørgsmål ift. ekkokamre? (2/15)
Et første skridt er at erkende, at forskningen om sociale medier er i sin vorden. Mange antagelser har vist sig forkerte. Og et enkelt studie giver sjældent sandheden.

Man må basere sig på opdaterede reviews fra internationalt førende forskere. (3/15)
Tråden baseres på 4 reviews fra de bedste forskere i verden:

#1 @andyguess et al (2018): researchgate.net/profile/Benjam…

#2 @p_barbera (2020): cambridge.org/core/books/soc…

#3 @STWorg et al (2021): osf.io/preprints/soca…

#4 @rasmus_kleis et al (2022): reutersinstitute.politics.ox.ac.uk/echo-chambers-…
(4/15)
1) Hvor udbredte er ekkokamre?

Ekkokamre findes i ekstreme miljøer, men de er ikke udbredte.

Alle reviews slår fast, at ekkokamre i form af en ensidig eksponering for nyheder, der passer til ens overbevisning, fylder mindre end antaget.

Her fra review #4.👇

(5/15)
Review #3 understreger, at der er "homophilly" i folks sociale medie netværk. 👇Venstreorienterede er mere tilbøjelige til at følge venstreorientede.

Men som påpeget i selvsamme review samt review #2, så er spørgsmålet om, der er mere "homophilly" online end offline? (6/15)
Det leder til spørgsmål 2: Hvad er årsagerne til ekkokamre?

Begreberne om "filterbobler" og "ekkokamre" adskiller sig ved årsagen. "Filterbobler" er algoritme-drevne ekkokamre, mens begrebet om "ekkokamre" holder årsagen åben. (7/15)
Review #4 (jf. 👆) afviser "filterbobbel"-tesen kategorisk ift. nyheder. Det betyder ikke, at algoritmer ikke kan påvirke strukturen i online netværk. MEN: "homophily" afspejler sig i alle interaktioner vi har 👇(annualreviews.org/doi/abs/10.114…). Selv-selektion er nulhypotesen. (8/15)
Review #2 argumenterer faktisk for, at "homophily" må forventes at være *mindre* online end offline. 👇 (9/15)
Men selvom ekkokamre er mindre udbredte & mindre algoritmer-drevet kan de godt være problematisk. Derfor spørgsmål 3: Hvad er effekten?

Forskning viser, at debat i lukkede grupper af ligesindede kan skabe polarisering. Ekstremisternes ekkokamre *er* uhensigtsmæssige. (10/15)
Men forskningen tyder også på, at nogle af udfordringerne med had og polarisering på sociale medier hænger sammen med *fraværet* af ekkokamre.

Studierne i review #2, der viser, at sociale medier polariserer, fokuserer netop på eksponeringen af modsatte synspunkter. 👇(11/15)
Det er konsistent med forskning fra mit eget @ROPHproject, som viser, at folk i højere grad oplever had fra fremmede online end offline (cambridge.org/core/journals/…) & at hadefulde individer indtager mere centrale positioner i online netværk (psyarxiv.com/tp93r/) (12/15).
Hvis ekkokamrerne i mindre grad er udbredte på sociale medier end mange tror, hvorfor er der så stadig et "ekkokammer om ekkokamre"?

Review #1 giver 3 grunde. 👇 (13/15)
Der mangler stadig masser af viden. Den viden hindres af lukkede platforme, data & algoritmer. Og der er ingen tvivl om, at vi særligt mangler viden om danske forhold. Forskningen er stadig i sin vorden. Netop derfor skal man passe på med forsimplede forståelser. (14/15)
De sociale medier er formentlig de mest komplicerede sociale økosystemer som findes. De kan ikke blive reguleret hensigtsmæssigt uden opdateret viden & uden åbne data. Det vil være som at sætte en blind kioskejer til at regulere det finansielle system. (15/15)

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with Michael Bang Petersen

Michael Bang Petersen Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @M_B_Petersen

Feb 26
Individuals who fight in armed conflict tend to be anti-democratic.

With one exception: Violent resistance in defence of democracy.

🧵 on our research on the psychology of violence & why the ultimate defense of democracy requires understanding anti-democratic impulses. (1/14)
Democracy is the principled recognition of equality in power. Autocracy is the opposite.

Psychologically, an orientation to autocracy draws its strength from so-called dominance motivations (doi.org/10.1177/095679…). (2/14)
All humans seek status. But paths differ (doi.org/10.1037/a00303…). *Prestige* is status in exchange for problem-solving. *Dominance* is the desire to acquire status from fear via intimidation & aggression.

In the game of status, dominance is the strategy of the predator. (3/14)
Read 14 tweets
Feb 19
🚨 Preprint of 21 country study

Those not vaxxed against COVID face prejudice, exceeding other targets such as immigrants: psyarxiv.com/t2g45/

A 🧵 on our findings, why they may be controversial, why they in any case are important & how they help remedy the problem.

(1/16)
Debates about COVID-19 vaccines are intense, involving vax mandates (bbc.com/news/world-eur…), strong protests (bbc.com/news/world-us-…) & sharp rhetoric (edition.cnn.com/2022/01/05/eur…). (2/16)
Political cleavages have previously been found to elicit prejudice (doi.org/10.1111/ajps.1…), e.g., exclusion from family relationships. Here we ask if there is prejudice between groups defined by COVID-19 vaccination status? (2/16)
Read 17 tweets
Feb 15
Pandemic fatigue & conflict are rampant

How to motivate people to keep up with continued threats in this context?

In @SciReports, we find that hope built via actionable advice is more effective than fear: nature.com/articles/s4159…

Fear sans hope can defy its own purpose

🧵(1/8)
Media, governments & influencers are often accused of "fear-mongering". Such strategy would clash with insights on risk-communication: doi.org/10.1111/bjhp.1…. Effective communication is clear about the threat *&* how to deal with the threat. Actionable advice builds hope. (2/8)
During the pandemic, we examined the relative efficacy of threat- & hope-oriented communication. The context was the race between the Alpha variant & implementation of vaccines. The need was to "buy time" via stronger interventions until vaccines took effect. (3/8)
Read 8 tweets
Feb 12
Omicron has raised a key question globally:

What is the end goal of pandemic management?

If you as a citizen don't know the answer to this, communication has failed.

Public support requires clear communication about goals & involved uncertainties, costs & benefits.

🧵(1/17)
Why is now more important than ever?

Because omicron has decoupled cases and epidemic severity.

Before all strategies led to suppressing cases.

With omicron, it is possible to suppress severe disease with vaccines & treatment without suppressing cases. (2/17)
Should countries suppress cases or suppress disease?

Hong Kong has imposed the toughest restrictions yet in the face of omicron, choosing to suppress cases: scmp.com/news/hong-kong…?

Scandinavia has lifted all restrictions, choosing to suppress severe disease. (3/17)
Read 17 tweets
Feb 10
PSA to anyone interested in informing public discourse about the severity of the epidemic situation in Denmark (and maybe elsewhere):

Show extreme care when sharing screenshots of cases & admissions from @OurWorldInData or similar

Let me explain in 3 plots. (1/6)
Omicron generates very high case counts but lower severity, disrupting the indicators & patterns we have all been tracking for 2 years.

The plots I'll now show are all from Danish Center for Disease Control's (@SSI_dk) weekly monitoring report: ssi.dk/-/media/cdn/fi…. (2/6)
#1

With a lot of infections many will be admitted with but not because of covid. The plot below shows the development. Red is "because of". Right now only 55 % of admissions with a positive test are because of covid. (3/6)
Read 6 tweets
Feb 3
Twitter is so terrible because with social media even the illiterate masses get to voice their views, right?

Wrong.

In a new preprint, we find that the most hateful in political discussions are more resourceful: Engaged, efficacious & educated: psyarxiv.com/tp93r/

🧵(1/10)
We obtained US survey participants' Twitter IDs in order to connect their psychological & political profile to Twitter activity (N=2012). In addition to toxicity & sentiment, we assessed their tweets' level of political hate with this classifier: psyarxiv.com/8m5dc/. (2/10)
Our preferred measure ("political hate") is significantly predicted by political engagement, political interest, internal efficacy and education. For toxicity and sentiment, results are in the same direction. We see no evidence that hostility is higher among the unengaged. (3/10)
Read 10 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us on Twitter!

:(