Share this page!

Andreas Backhaus Profile picture
Twitter logo
1 Dec, 4 tweets, 5 min read
Many papers are now investigating whether YouTube is creating "echo chambers", mostly regarding the (far) right. In principle, an interesting topic. However, if one of the labels for "far right" is "IDW", it's starting from false premises. arxiv.org/pdf/2011.12843… @EricRWeinstein ImageImage
Consequently, the paper lumps people like @SamHarrisOrg @BretWeinstein @joerogan @GadSaad @reason @ScottAdamsSays @benshapiro together with white nationalists and neonazis under the label "far right". Note that "Real Coffee with Scott Adams" is only labeled as "right" though. ☕️ ImageImage
How did this happen? The paper provides sources for the labels. So it's just repeating what others have already done, without checking whether it makes sense. Playing it safe, one may say. ImageImageImage
The main source for the "far right" label is a 2018 study that created these interesting networks that were used to call people "far right" who had talked to someone who had talked to someone on the "far right". Even the author seems disappointed by the YouTube papers by now. ImageImageImage

• • •

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

Keep Current with Andreas Backhaus

Andreas Backhaus 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 @AndreasShrugged

Andreas Backhaus Profile picture
8 Nov
Feeling quite compassionate for the people who apparently have been going through hell, psychologically, during the past 4 years. They've been dragged from one mass delusion to another, with their emotional states being played like a fiddle by the media. Scary stuff.
By the beginning of Trump's term, they believed Trump won in 2016 due to Russia collusion, by the end, they believed the U.S. was on the brink of becoming a fascist dictatorship. Both beliefs, and many more in-between, have been refuted by the facts.
In retrospect, why don't they realize that? I suppose because they're getting a big reward now from the current mass delusion that they've prevailed over the forces of evil. If the happy ending to a story is that you've slain Darth Vader, you'll like that ending.
Read 4 tweets
Andreas Backhaus Profile picture
6 Nov
Sehr gute Darstellung der Tatsache, dass es auch dann noch viele weitere Infektionen geben wird, wenn R<1, die Zahl der aktiven Infektionen aber bereits hoch ist. Wenn R=0,9 ist gilt: Aktive Infektionen x 9 = zusätzliche Infektionen bis Stillstand. Rechenbeispiel für Deutschland:
Wir haben aktuell ~200.000 aktive Fälle. Seien wir mal zu optimistisch und unterstellen, es gäbe keine Dunkelziffer der Infektionen. Bei R=0,9 würde dies 1,8 Millionen weitere Infektionen in Deutschland bedeuten. Bei einer Infektionssterblichkeitsrate von 0,65% -> +13.000 Tote.
Bei einem Dunkelzifferfaktor von 2,5 (immer noch optimistisch) wären wir aktuell bei 500.000 aktiven Infektionen und 4,5 Millionen, die bei R=0,9 noch dazukommen würden. Zusammen 5 Millionen Infektionen -> 0,65% Sterblichkeit -> +32.500 Tote.
Read 4 tweets
Andreas Backhaus Profile picture
19 Oct
Some quick fact-checking on @SKMorefield's text:
- Charts about mask mandates and covid cases don't tell us much because, as noted, they don't have a control group showing what would have happened without mandates
- Causality between mandates and cases could run both ways
- Once you add reasonable control groups, evidence points in favor of masks, e.g. nber.org/papers/w27891
- Denmark, Norway and Finland all recommend masks by now, not everywhere, not all the time, but reasonably
- Linked report about pneumonia due to masks is totally anecdotal
- The global infection fatality rate reflects that much of the world is younger than the West, IFRs in our countries are higher and we have large older populations at high risk
- Sunetra Gupta has so far been wrong about everything, en.wikipedia.org/wiki/Sunetra_G…
Read 4 tweets
Andreas Backhaus Profile picture
10 Oct
Notion: Gap between excess deaths and confirmed Covid-19 deaths could be "lockdown deaths". Let's test this: (a) States with lockdown but low Covid should have excess deaths from lockdown (b) States without lockdown but Covid shouldn't have excess deaths unexplained by Covid.
For (a), New York had a long lockdown with stay-at-home orders extending into June. But as New York suppressed the virus, both Covid-19 deaths and other excess deaths fell down a cliff simultaneously. So that part of the lockdown deaths theory doesn't hold up.
For (b), Arkansas adopted only mild measures, no SAH orders. But as Covid-19 deaths went up (blue), also other excess deaths (green) went up above average (orange line). CDC estimates non-Covid excess deaths are 158-1203, compared to 1500 Covid-19 deaths. So (b) doesn't hold up.
Read 5 tweets
Andreas Backhaus Profile picture
9 Oct
What's wrong with this notion? Many things: First, it seems to be based on the very recent increase in cases that has been happening mostly in French-speaking Brussels and Wallonia. Before that, Flanders had actually been more affected (en.wikipedia.org/wiki/COVID-19_…).
So, Yascha is looking only at the current pattern, but not at the dynamics. Second, he singles out one variable that might explain the pattern ("cultural attitudes") and claims that's it. That's generally speaking wrong. What about commuting? French-speaking Belgians commute...
...mostly in the French-speaking parts. So some infections introduced into Wallonia and Brussels are very likely to spread only within these areas because almost no one would carry them over to Flanders. That's not culture but potentially very important. Third, Flanders had a bad
Read 4 tweets
Andreas Backhaus Profile picture
19 Aug
Two months later and we can apply @nataliexdean's thread to Europe, too: Germany is back to reporting more than 1,000 new daily cases on average, while the median age of cases has dropped to a new low of 32 years (German source: @rki_de situation reports).
The COVID-19 incidence among the two oldest age groups has not risen (yet), as this @rki_de graph shows (h/t: @christoph_rothe). Also number of patients in intensive care is low and stable so far. Image
Part of the increase in cases due to testing? Yes - testing in Germany has just hit a fresh high of 875,000 tests per week. But cases have mostly been rising stronger than testing, pushing the positivity rate from 0.59% to 0.99% temporarily, last week at 0.96%.
Read 7 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

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

Donate via Paypal Become our Patreon

Thank you for your support!

Follow Us on Twitter!

Like this author's thread?

Get emails when @AndreasShrugged has new unrolls!

Check out other threads from @AndreasShrugged!

Read all threads by @AndreasShrugged