Max Falkenberg
@maxfalken.bsky.social
Physicist. Computational Social Scientist. Interested in the structures of polarization.
Central European University & Institute for Sustainable Resources, UCL.
maxfalkenberg.com
Central European University & Institute for Sustainable Resources, UCL.
maxfalkenberg.com
@galessandro.bsky.social presenting our Complexity72H work at CS2Italy 🥳
Solid 6/10 presentation - would collaborate again 😅
Solid 6/10 presentation - would collaborate again 😅
January 17, 2025 at 3:19 PM
@galessandro.bsky.social presenting our Complexity72H work at CS2Italy 🥳
Solid 6/10 presentation - would collaborate again 😅
Solid 6/10 presentation - would collaborate again 😅
Contrary to some previous research, we find that abusive content does not drive higher engagement:
In 8 of 9 countries, toxic posts receive substantially less engagement. However, we have no way of knowing if this is an authentic result, or algorithm driven...
In 8 of 9 countries, toxic posts receive substantially less engagement. However, we have no way of knowing if this is an authentic result, or algorithm driven...
November 14, 2024 at 1:42 PM
Contrary to some previous research, we find that abusive content does not drive higher engagement:
In 8 of 9 countries, toxic posts receive substantially less engagement. However, we have no way of knowing if this is an authentic result, or algorithm driven...
In 8 of 9 countries, toxic posts receive substantially less engagement. However, we have no way of knowing if this is an authentic result, or algorithm driven...
But that does not mean interactions with political opponents are rare!
In each of the countries studied, political out-group mentions are common and significantly more toxic than in-group mentions. This is true on both the political right and the left!
In each of the countries studied, political out-group mentions are common and significantly more toxic than in-group mentions. This is true on both the political right and the left!
November 14, 2024 at 1:42 PM
But that does not mean interactions with political opponents are rare!
In each of the countries studied, political out-group mentions are common and significantly more toxic than in-group mentions. This is true on both the political right and the left!
In each of the countries studied, political out-group mentions are common and significantly more toxic than in-group mentions. This is true on both the political right and the left!
Across 9 countries (Canada, France, Germany, Italy, Poland, Spain, Turkey, UK, USA) we show just how ubiquitous toxic political communication is on X.
Our data shows that the vast majority of interactions are with political allies (structural polarization / homophily).
Our data shows that the vast majority of interactions are with political allies (structural polarization / homophily).
November 14, 2024 at 1:42 PM
Across 9 countries (Canada, France, Germany, Italy, Poland, Spain, Turkey, UK, USA) we show just how ubiquitous toxic political communication is on X.
Our data shows that the vast majority of interactions are with political allies (structural polarization / homophily).
Our data shows that the vast majority of interactions are with political allies (structural polarization / homophily).
The exodus from X/Twitter has finally started, but maybe Twitter should have died a long time ago?
Our new research published today shows how political abuse on X is a global, widespread, and cross-partisan phenomenon.
Out now in Nature Communications: www.nature.com/articles/s41...
Key results🧵
Our new research published today shows how political abuse on X is a global, widespread, and cross-partisan phenomenon.
Out now in Nature Communications: www.nature.com/articles/s41...
Key results🧵
November 14, 2024 at 1:42 PM
The exodus from X/Twitter has finally started, but maybe Twitter should have died a long time ago?
Our new research published today shows how political abuse on X is a global, widespread, and cross-partisan phenomenon.
Out now in Nature Communications: www.nature.com/articles/s41...
Key results🧵
Our new research published today shows how political abuse on X is a global, widespread, and cross-partisan phenomenon.
Out now in Nature Communications: www.nature.com/articles/s41...
Key results🧵
One option is to regulate changes in banks' exposure to fossil fuels rather than absolute exposure.
The stricter we limit changes in exposure, the fewer banks need to withdraw from FFs to enable a transition from inefficient to efficient phase out of lending.
The stricter we limit changes in exposure, the fewer banks need to withdraw from FFs to enable a transition from inefficient to efficient phase out of lending.
September 16, 2024 at 10:45 AM
One option is to regulate changes in banks' exposure to fossil fuels rather than absolute exposure.
The stricter we limit changes in exposure, the fewer banks need to withdraw from FFs to enable a transition from inefficient to efficient phase out of lending.
The stricter we limit changes in exposure, the fewer banks need to withdraw from FFs to enable a transition from inefficient to efficient phase out of lending.
What can we do?
Existing regulations are not enough - they limit the fossil fuel assets a bank can hold, but do not restrict finance substitution or opportunistic behaviour.
US and Canadian banks are heavily exposed but Japanese banks have loads of headroom to continue lending
Existing regulations are not enough - they limit the fossil fuel assets a bank can hold, but do not restrict finance substitution or opportunistic behaviour.
US and Canadian banks are heavily exposed but Japanese banks have loads of headroom to continue lending
September 16, 2024 at 10:44 AM
What can we do?
Existing regulations are not enough - they limit the fossil fuel assets a bank can hold, but do not restrict finance substitution or opportunistic behaviour.
US and Canadian banks are heavily exposed but Japanese banks have loads of headroom to continue lending
Existing regulations are not enough - they limit the fossil fuel assets a bank can hold, but do not restrict finance substitution or opportunistic behaviour.
US and Canadian banks are heavily exposed but Japanese banks have loads of headroom to continue lending
This means that if $1 of financing to the fossil fuel sector is withdrawn by banks, it translates to less than $1 of lost financing at the project level.
We say that this generates an "efficiency gap": The value of impacted deals vastly exceeds failed deals.
We say that this generates an "efficiency gap": The value of impacted deals vastly exceeds failed deals.
September 16, 2024 at 10:41 AM
This means that if $1 of financing to the fossil fuel sector is withdrawn by banks, it translates to less than $1 of lost financing at the project level.
We say that this generates an "efficiency gap": The value of impacted deals vastly exceeds failed deals.
We say that this generates an "efficiency gap": The value of impacted deals vastly exceeds failed deals.
But haven't many major banks reduced their fossil fuel lending?... Yes, but in an uncoordinated manner!
Major banks in Europe have reduced their lending...
But, the withdrawal of finance has been replaced by finance from other banks, mainly in Canada and Japan!
The finance has been "substituted"!
Major banks in Europe have reduced their lending...
But, the withdrawal of finance has been replaced by finance from other banks, mainly in Canada and Japan!
The finance has been "substituted"!
September 16, 2024 at 10:41 AM
But haven't many major banks reduced their fossil fuel lending?... Yes, but in an uncoordinated manner!
Major banks in Europe have reduced their lending...
But, the withdrawal of finance has been replaced by finance from other banks, mainly in Canada and Japan!
The finance has been "substituted"!
Major banks in Europe have reduced their lending...
But, the withdrawal of finance has been replaced by finance from other banks, mainly in Canada and Japan!
The finance has been "substituted"!
From 2010-2021, banks underwrote $7 TRILLION of fossil fuel loans and bonds to fossil fuel companies.
Most of the deals involve syndicates - a group of partner banks sharing risk.
Across that period, underwriting has remained stable, averaging $584bn before Paris agreement and $592bn post-Paris.
Most of the deals involve syndicates - a group of partner banks sharing risk.
Across that period, underwriting has remained stable, averaging $584bn before Paris agreement and $592bn post-Paris.
September 16, 2024 at 10:40 AM
From 2010-2021, banks underwrote $7 TRILLION of fossil fuel loans and bonds to fossil fuel companies.
Most of the deals involve syndicates - a group of partner banks sharing risk.
Across that period, underwriting has remained stable, averaging $584bn before Paris agreement and $592bn post-Paris.
Most of the deals involve syndicates - a group of partner banks sharing risk.
Across that period, underwriting has remained stable, averaging $584bn before Paris agreement and $592bn post-Paris.
New dataset for anyone interested in social media beyond the West 👇
We provide 400m+ user interactions and 1.4m user profiles for the microblogging platform Koo which has gained traction in India, Brazil and Nigeria.
Paper: arxiv.org/abs/2401.07599
Data: zenodo.org/records/10476212
We provide 400m+ user interactions and 1.4m user profiles for the microblogging platform Koo which has gained traction in India, Brazil and Nigeria.
Paper: arxiv.org/abs/2401.07599
Data: zenodo.org/records/10476212
January 17, 2024 at 12:34 PM
New dataset for anyone interested in social media beyond the West 👇
We provide 400m+ user interactions and 1.4m user profiles for the microblogging platform Koo which has gained traction in India, Brazil and Nigeria.
Paper: arxiv.org/abs/2401.07599
Data: zenodo.org/records/10476212
We provide 400m+ user interactions and 1.4m user profiles for the microblogging platform Koo which has gained traction in India, Brazil and Nigeria.
Paper: arxiv.org/abs/2401.07599
Data: zenodo.org/records/10476212
For countries where we have reliability ratings, the right are reference more unreliable media than the left.
Finally, we find a common ally-enemy structure in how the left and right interact with each other, relative to how they interact with apolitical accounts...
Finally, we find a common ally-enemy structure in how the left and right interact with each other, relative to how they interact with apolitical accounts...
November 30, 2023 at 1:52 PM
For countries where we have reliability ratings, the right are reference more unreliable media than the left.
Finally, we find a common ally-enemy structure in how the left and right interact with each other, relative to how they interact with apolitical accounts...
Finally, we find a common ally-enemy structure in how the left and right interact with each other, relative to how they interact with apolitical accounts...
Using the groups defined by the network of Twitter interactions, we find that out-group interactions are more toxic (panel a), but receive lower engagement (panel b), than in-group interactions.
Most of these interactions are unidirectional. Out-group conversation is rare...
Most of these interactions are unidirectional. Out-group conversation is rare...
November 30, 2023 at 1:52 PM
Using the groups defined by the network of Twitter interactions, we find that out-group interactions are more toxic (panel a), but receive lower engagement (panel b), than in-group interactions.
Most of these interactions are unidirectional. Out-group conversation is rare...
Most of these interactions are unidirectional. Out-group conversation is rare...
We show that political interaction networks on Twitter are segregated between countries (see figure above) and structurally polarized within them. For 8 of 9 countries, this structural division is between the political left and right (in Germany its establishment-populist)...
November 30, 2023 at 1:51 PM
We show that political interaction networks on Twitter are segregated between countries (see figure above) and structurally polarized within them. For 8 of 9 countries, this structural division is between the political left and right (in Germany its establishment-populist)...
New preprint We show that polarized social network structures align with out-group animosity, aka affective polarization, across nine different countries.
osf.io/preprints/ps...
With @baronca.bsky.social @jurgenpfeffer.bsky.social and others.
Some highlights 👇
osf.io/preprints/ps...
With @baronca.bsky.social @jurgenpfeffer.bsky.social and others.
Some highlights 👇
November 30, 2023 at 1:50 PM
New preprint We show that polarized social network structures align with out-group animosity, aka affective polarization, across nine different countries.
osf.io/preprints/ps...
With @baronca.bsky.social @jurgenpfeffer.bsky.social and others.
Some highlights 👇
osf.io/preprints/ps...
With @baronca.bsky.social @jurgenpfeffer.bsky.social and others.
Some highlights 👇
Of relevance 👇
Toxicity of posts authored by the far right as a function of whether the targeted user is mainly quote tweeted or retweeted. And of course Dems are the main group being heavily quote tweeted...
From doi.org/10.1093/pnas...
Toxicity of posts authored by the far right as a function of whether the targeted user is mainly quote tweeted or retweeted. And of course Dems are the main group being heavily quote tweeted...
From doi.org/10.1093/pnas...
October 27, 2023 at 2:19 PM
Of relevance 👇
Toxicity of posts authored by the far right as a function of whether the targeted user is mainly quote tweeted or retweeted. And of course Dems are the main group being heavily quote tweeted...
From doi.org/10.1093/pnas...
Toxicity of posts authored by the far right as a function of whether the targeted user is mainly quote tweeted or retweeted. And of course Dems are the main group being heavily quote tweeted...
From doi.org/10.1093/pnas...
But Gettr not just relevant for US. We found huge increase in Brazilian engagement in the lead up to the Brasilia riots in early 2023 (the third dashed line in the figure)...
October 26, 2023 at 8:31 PM
But Gettr not just relevant for US. We found huge increase in Brazilian engagement in the lead up to the Brasilia riots in early 2023 (the third dashed line in the figure)...
Turns out that tweets mentioning users who are heavily quote tweeted but not retweeted (i.e. Democrats) are targeted with significantly more toxic tweets...
October 26, 2023 at 8:29 PM
Turns out that tweets mentioning users who are heavily quote tweeted but not retweeted (i.e. Democrats) are targeted with significantly more toxic tweets...
Gettr users active on Twitter primarily retweet far-right media sources but quote tweet centre and left leaning sources. They also heavily Retweet Republican politicians, but quote tweet Democrats...
October 26, 2023 at 8:28 PM
Gettr users active on Twitter primarily retweet far-right media sources but quote tweet centre and left leaning sources. They also heavily Retweet Republican politicians, but quote tweet Democrats...
You can see that here:
The blue line is for the verified cohort of users active on Gettr but banned from Twitter. The green line is for those still active on Twitter. The orange for unverified users.
Rogan criticised the platform at the dashed line...
The blue line is for the verified cohort of users active on Gettr but banned from Twitter. The green line is for those still active on Twitter. The orange for unverified users.
Rogan criticised the platform at the dashed line...
October 26, 2023 at 8:27 PM
You can see that here:
The blue line is for the verified cohort of users active on Gettr but banned from Twitter. The green line is for those still active on Twitter. The orange for unverified users.
Rogan criticised the platform at the dashed line...
The blue line is for the verified cohort of users active on Gettr but banned from Twitter. The green line is for those still active on Twitter. The orange for unverified users.
Rogan criticised the platform at the dashed line...