Robert (Bob) Kubinec
@rmkubinec.bsky.social
Political scientist University of South Carolina. Interests: business politics/corruption, Middle East, measurement, Bayesian statistics
The Bayesian Hitman: https://a.co/d/e4QmtKo
Website: www.robertkubinec.com
The Bayesian Hitman: https://a.co/d/e4QmtKo
Website: www.robertkubinec.com
🚨 New draft 🚨
We built an LLM-enabled system to measure greenwashing scores in 1 million worldwide Facebook ads.
We found vast networks of Facebook pages sharing pro-fossil fuel messages & show that ads are targeted at left-leaning areas with fossil fuel investments.
Link: doi.org/10.48550/arX...
We built an LLM-enabled system to measure greenwashing scores in 1 million worldwide Facebook ads.
We found vast networks of Facebook pages sharing pro-fossil fuel messages & show that ads are targeted at left-leaning areas with fossil fuel investments.
Link: doi.org/10.48550/arX...
November 20, 2025 at 2:04 PM
🚨 New draft 🚨
We built an LLM-enabled system to measure greenwashing scores in 1 million worldwide Facebook ads.
We found vast networks of Facebook pages sharing pro-fossil fuel messages & show that ads are targeted at left-leaning areas with fossil fuel investments.
Link: doi.org/10.48550/arX...
We built an LLM-enabled system to measure greenwashing scores in 1 million worldwide Facebook ads.
We found vast networks of Facebook pages sharing pro-fossil fuel messages & show that ads are targeted at left-leaning areas with fossil fuel investments.
Link: doi.org/10.48550/arX...
My #rstats cheat code for today is the binom.confint function in the binom package that will spit out *12* different ways of calculating a CI for a proportion.
Also, this is why you use R for statistics...
(and of course the correct CI method is bayes 😎)
Also, this is why you use R for statistics...
(and of course the correct CI method is bayes 😎)
November 16, 2025 at 2:52 AM
My #rstats cheat code for today is the binom.confint function in the binom package that will spit out *12* different ways of calculating a CI for a proportion.
Also, this is why you use R for statistics...
(and of course the correct CI method is bayes 😎)
Also, this is why you use R for statistics...
(and of course the correct CI method is bayes 😎)
People are dunking on the AI agents in this story, but isn't this exactly what a real startup is like? 😅
www.wired.com/story/all-my...
www.wired.com/story/all-my...
November 13, 2025 at 7:11 PM
People are dunking on the AI agents in this story, but isn't this exactly what a real startup is like? 😅
www.wired.com/story/all-my...
www.wired.com/story/all-my...
Look, social scientists, it's a 2x2 table! In the wild!
November 13, 2025 at 2:25 AM
Look, social scientists, it's a 2x2 table! In the wild!
The recent movement in Trump's approval may not look like a big deal, but you have to remember that Trump's base stands to gain a lot from his regime. For 1% of his base to say, "no, we don't like what he's doing," that means those benefits don't matter anymore.
And frankly, that's a big deal.
And frankly, that's a big deal.
November 7, 2025 at 2:56 PM
The recent movement in Trump's approval may not look like a big deal, but you have to remember that Trump's base stands to gain a lot from his regime. For 1% of his base to say, "no, we don't like what he's doing," that means those benefits don't matter anymore.
And frankly, that's a big deal.
And frankly, that's a big deal.
We've all been there
November 7, 2025 at 12:57 AM
We've all been there
Virginia governor Democratic vote share is the same as New Jersey.
Truly insane.
Truly insane.
November 5, 2025 at 3:00 AM
Virginia governor Democratic vote share is the same as New Jersey.
Truly insane.
Truly insane.
Political scientists will find it interesting (and perhaps no one else 😅) that a Columbia SC mayoral candidate endorsed rank choice voting, saying it would reduce polarization.
She may have been influenced by the NYC mayoral primary earlier this year...
www.thestate.com/opinion/arti...
She may have been influenced by the NYC mayoral primary earlier this year...
www.thestate.com/opinion/arti...
November 5, 2025 at 1:34 AM
Political scientists will find it interesting (and perhaps no one else 😅) that a Columbia SC mayoral candidate endorsed rank choice voting, saying it would reduce polarization.
She may have been influenced by the NYC mayoral primary earlier this year...
www.thestate.com/opinion/arti...
She may have been influenced by the NYC mayoral primary earlier this year...
www.thestate.com/opinion/arti...
The Virginia governor's race isn't even remotely close. It looks like a governor race in a "deep blue" state.
We'll see the outcome, of course, but right now the GOP is in some boiling hot water.
We'll see the outcome, of course, but right now the GOP is in some boiling hot water.
November 3, 2025 at 5:32 PM
The Virginia governor's race isn't even remotely close. It looks like a governor race in a "deep blue" state.
We'll see the outcome, of course, but right now the GOP is in some boiling hot water.
We'll see the outcome, of course, but right now the GOP is in some boiling hot water.
ordbetareg has now fallen out of the top 10% of CRAN downloads.
I mostly blame @chelseaparlett.bsky.social for not teaching enough classes to enough students in order to properly inflate my download statistics.
I mostly blame @chelseaparlett.bsky.social for not teaching enough classes to enough students in order to properly inflate my download statistics.
November 1, 2025 at 8:42 PM
ordbetareg has now fallen out of the top 10% of CRAN downloads.
I mostly blame @chelseaparlett.bsky.social for not teaching enough classes to enough students in order to properly inflate my download statistics.
I mostly blame @chelseaparlett.bsky.social for not teaching enough classes to enough students in order to properly inflate my download statistics.
Per @gelliottmorris.com robust polling aggregator, President Trump's approval rating has a hit a new low at 55% disapproval to 41% approval.
After three months of stasis, this public opinion shift--small yet noticeable--could be a sign of anger from the shutdown.
fiftyplusone.news/polls/approv...
After three months of stasis, this public opinion shift--small yet noticeable--could be a sign of anger from the shutdown.
fiftyplusone.news/polls/approv...
October 30, 2025 at 4:31 PM
Per @gelliottmorris.com robust polling aggregator, President Trump's approval rating has a hit a new low at 55% disapproval to 41% approval.
After three months of stasis, this public opinion shift--small yet noticeable--could be a sign of anger from the shutdown.
fiftyplusone.news/polls/approv...
After three months of stasis, this public opinion shift--small yet noticeable--could be a sign of anger from the shutdown.
fiftyplusone.news/polls/approv...
From my co-author group chat today
October 29, 2025 at 1:30 PM
From my co-author group chat today
From @gelliottmorris.com substack this morning 👇
There should be a fire alarm going off right now in the GOP HQ. The GOP made up losses with educated voters by picking up less-educated ones. Cracks emerging in that coalition could be fatal as the GOP is locked into Trumpism.
There should be a fire alarm going off right now in the GOP HQ. The GOP made up losses with educated voters by picking up less-educated ones. Cracks emerging in that coalition could be fatal as the GOP is locked into Trumpism.
October 24, 2025 at 1:13 PM
From @gelliottmorris.com substack this morning 👇
There should be a fire alarm going off right now in the GOP HQ. The GOP made up losses with educated voters by picking up less-educated ones. Cracks emerging in that coalition could be fatal as the GOP is locked into Trumpism.
There should be a fire alarm going off right now in the GOP HQ. The GOP made up losses with educated voters by picking up less-educated ones. Cracks emerging in that coalition could be fatal as the GOP is locked into Trumpism.
October 22, 2025 at 4:20 PM
In this study, people wore a VR headset while a realistic simulation of a spider crawled towards their hand. The subjects all had a pre-existing fear of spiders.
I'm not exactly sure of the purpose, but wow, talk about big treatment effects--and they used ordered beta regression!
I'm not exactly sure of the purpose, but wow, talk about big treatment effects--and they used ordered beta regression!
October 22, 2025 at 1:21 AM
In this study, people wore a VR headset while a realistic simulation of a spider crawled towards their hand. The subjects all had a pre-existing fear of spiders.
I'm not exactly sure of the purpose, but wow, talk about big treatment effects--and they used ordered beta regression!
I'm not exactly sure of the purpose, but wow, talk about big treatment effects--and they used ordered beta regression!
The political scientists are having quite a week
October 20, 2025 at 11:35 PM
The political scientists are having quite a week
Great, so now @gelliottmorris.com knows that I cyber-stalk him 😬
October 16, 2025 at 8:20 PM
Great, so now @gelliottmorris.com knows that I cyber-stalk him 😬
It's strange to see these "cost estimates" of charging cars in the NYT that don't take into account the cost of actually doing the charging.
Driving to a gas station, especially in places where they are crowded/rare, is a big time cost vs. charging at home.
Driving to a gas station, especially in places where they are crowded/rare, is a big time cost vs. charging at home.
October 8, 2025 at 1:20 PM
It's strange to see these "cost estimates" of charging cars in the NYT that don't take into account the cost of actually doing the charging.
Driving to a gas station, especially in places where they are crowded/rare, is a big time cost vs. charging at home.
Driving to a gas station, especially in places where they are crowded/rare, is a big time cost vs. charging at home.
The US investor community is an absolute legend 🫣
September 29, 2025 at 2:38 PM
The US investor community is an absolute legend 🫣
now we can calculate your expected life span
sorry/not sorry
sorry/not sorry
September 22, 2025 at 5:33 PM
now we can calculate your expected life span
sorry/not sorry
sorry/not sorry
The weird spike in ordbetareg package downloads continues.
I have no idea what is going on -- similar spike d/n exist for other CRAN packages.
Presumably it's some kind of mechanical process... but what exactly?
#rstats
I have no idea what is going on -- similar spike d/n exist for other CRAN packages.
Presumably it's some kind of mechanical process... but what exactly?
#rstats
September 21, 2025 at 2:22 AM
The weird spike in ordbetareg package downloads continues.
I have no idea what is going on -- similar spike d/n exist for other CRAN packages.
Presumably it's some kind of mechanical process... but what exactly?
#rstats
I have no idea what is going on -- similar spike d/n exist for other CRAN packages.
Presumably it's some kind of mechanical process... but what exactly?
#rstats
The plot below shows pages from Facebook Ad Library that are connected by ad similarity (at least 5 ads with high similarity scores). We see a dense network of pages with high greenwashing/climate misinformation.
Some with known funders (API) and others with fronts (Partnership for Energy).
Some with known funders (API) and others with fronts (Partnership for Energy).
September 19, 2025 at 3:23 PM
The plot below shows pages from Facebook Ad Library that are connected by ad similarity (at least 5 ads with high similarity scores). We see a dense network of pages with high greenwashing/climate misinformation.
Some with known funders (API) and others with fronts (Partnership for Energy).
Some with known funders (API) and others with fronts (Partnership for Energy).
The guys who complain about merit-based hiring put someone on the Fed board who hasn't published an academic paper in 13 years (and as far as I can tell, has only ever published one study).
One.
Lisa Cook, the "DEI" hire? More than 30 studies, each with >10 citations from other researchers.
One.
Lisa Cook, the "DEI" hire? More than 30 studies, each with >10 citations from other researchers.
September 16, 2025 at 6:07 PM
The guys who complain about merit-based hiring put someone on the Fed board who hasn't published an academic paper in 13 years (and as far as I can tell, has only ever published one study).
One.
Lisa Cook, the "DEI" hire? More than 30 studies, each with >10 citations from other researchers.
One.
Lisa Cook, the "DEI" hire? More than 30 studies, each with >10 citations from other researchers.
Package downloads for ordbetareg have gone up 10x in the last two weeks (!!). It is now in the top 10% of all CRAN packages.
I honestly have no idea why. If someone could turn off that infinite for loop with an "install.packages" line inside, that would be great... 😅
I honestly have no idea why. If someone could turn off that infinite for loop with an "install.packages" line inside, that would be great... 😅
September 16, 2025 at 5:46 PM
Package downloads for ordbetareg have gone up 10x in the last two weeks (!!). It is now in the top 10% of all CRAN packages.
I honestly have no idea why. If someone could turn off that infinite for loop with an "install.packages" line inside, that would be great... 😅
I honestly have no idea why. If someone could turn off that infinite for loop with an "install.packages" line inside, that would be great... 😅
me to my co-author this morning... 🙄
September 15, 2025 at 3:20 PM
me to my co-author this morning... 🙄