Sebastian Ramirez-Ruiz
@seramirezruiz.bsky.social
Max Weber Postdoctoral Fellow at @eui-eu.bsky.social
Interested in causal inference, evidence in policy- and decision-making, #rstats, and most importantly, bicycles | Ph.D. @hertiedatascience.bsky.social | 🇨🇴
🌐 https://seramirezruiz.github.io/
Interested in causal inference, evidence in policy- and decision-making, #rstats, and most importantly, bicycles | Ph.D. @hertiedatascience.bsky.social | 🇨🇴
🌐 https://seramirezruiz.github.io/
Across policy domains:
📚 Some lean more on scholarly research, others on policy sources (think of lit on 'cultures of evidence').
Yet across all domains, 🌍 governments mainly cite knowledge from high-income countries.
Bottom line: origins of references stay stable despite domain differences.
📚 Some lean more on scholarly research, others on policy sources (think of lit on 'cultures of evidence').
Yet across all domains, 🌍 governments mainly cite knowledge from high-income countries.
Bottom line: origins of references stay stable despite domain differences.
June 12, 2025 at 9:27 AM
Across policy domains:
📚 Some lean more on scholarly research, others on policy sources (think of lit on 'cultures of evidence').
Yet across all domains, 🌍 governments mainly cite knowledge from high-income countries.
Bottom line: origins of references stay stable despite domain differences.
📚 Some lean more on scholarly research, others on policy sources (think of lit on 'cultures of evidence').
Yet across all domains, 🌍 governments mainly cite knowledge from high-income countries.
Bottom line: origins of references stay stable despite domain differences.
Who's cited most across borders?
🇺🇸 The U.S. leads gov-to-gov refs by a wide margin. 🇬🇧, 🇩🇪, 🇨🇦, 🇦🇺 follow. Also, 30 countries—mostly Least Developed—were never cited.
📚 Same for academic refs: 17 of top 20 gov-cited countries also top in academia.
🤯 43% of scholarly works include a U.S.-based author.
🇺🇸 The U.S. leads gov-to-gov refs by a wide margin. 🇬🇧, 🇩🇪, 🇨🇦, 🇦🇺 follow. Also, 30 countries—mostly Least Developed—were never cited.
📚 Same for academic refs: 17 of top 20 gov-cited countries also top in academia.
🤯 43% of scholarly works include a U.S.-based author.
June 12, 2025 at 9:27 AM
Who's cited most across borders?
🇺🇸 The U.S. leads gov-to-gov refs by a wide margin. 🇬🇧, 🇩🇪, 🇨🇦, 🇦🇺 follow. Also, 30 countries—mostly Least Developed—were never cited.
📚 Same for academic refs: 17 of top 20 gov-cited countries also top in academia.
🤯 43% of scholarly works include a U.S.-based author.
🇺🇸 The U.S. leads gov-to-gov refs by a wide margin. 🇬🇧, 🇩🇪, 🇨🇦, 🇦🇺 follow. Also, 30 countries—mostly Least Developed—were never cited.
📚 Same for academic refs: 17 of top 20 gov-cited countries also top in academia.
🤯 43% of scholarly works include a U.S.-based author.
Foreign v. domestic:
🌍 Global South govs rely mainly on foreign policy sources; Global North mostly on domestic.
🇺🇸 60% of scholarly refs in the U.S. docs cite papers with only U.S.-based academics as authors 🤯
🌐 Elsewhere, foreign or mixed international author make-ups dominate.
🌍 Global South govs rely mainly on foreign policy sources; Global North mostly on domestic.
🇺🇸 60% of scholarly refs in the U.S. docs cite papers with only U.S.-based academics as authors 🤯
🌐 Elsewhere, foreign or mixed international author make-ups dominate.
June 12, 2025 at 9:27 AM
Foreign v. domestic:
🌍 Global South govs rely mainly on foreign policy sources; Global North mostly on domestic.
🇺🇸 60% of scholarly refs in the U.S. docs cite papers with only U.S.-based academics as authors 🤯
🌐 Elsewhere, foreign or mixed international author make-ups dominate.
🌍 Global South govs rely mainly on foreign policy sources; Global North mostly on domestic.
🇺🇸 60% of scholarly refs in the U.S. docs cite papers with only U.S.-based academics as authors 🤯
🌐 Elsewhere, foreign or mixed international author make-ups dominate.
We classify cited sources into two types:
📑 Policy-based (govs, IGOs, think tanks)
📚 Scholarly (journal articles, working papers, preprints)
*And collected additional metadata relevant to them
📑 Policy-based (govs, IGOs, think tanks)
📚 Scholarly (journal articles, working papers, preprints)
*And collected additional metadata relevant to them
June 12, 2025 at 9:27 AM
We classify cited sources into two types:
📑 Policy-based (govs, IGOs, think tanks)
📚 Scholarly (journal articles, working papers, preprints)
*And collected additional metadata relevant to them
📑 Policy-based (govs, IGOs, think tanks)
📚 Scholarly (journal articles, working papers, preprints)
*And collected additional metadata relevant to them
📄 Whose expert knowledge informs policymaking around the world?
@rsenninger.bsky.social and I analyze data from 1.2 million government policy documents from 185 countries—and find a prominent pattern:
🌍 Policy evidence is overwhelmingly sourced from the Global North.
Preprint: osf.io/w8q3y
🧪🧵👇
@rsenninger.bsky.social and I analyze data from 1.2 million government policy documents from 185 countries—and find a prominent pattern:
🌍 Policy evidence is overwhelmingly sourced from the Global North.
Preprint: osf.io/w8q3y
🧪🧵👇
June 12, 2025 at 9:27 AM
📄 Whose expert knowledge informs policymaking around the world?
@rsenninger.bsky.social and I analyze data from 1.2 million government policy documents from 185 countries—and find a prominent pattern:
🌍 Policy evidence is overwhelmingly sourced from the Global North.
Preprint: osf.io/w8q3y
🧪🧵👇
@rsenninger.bsky.social and I analyze data from 1.2 million government policy documents from 185 countries—and find a prominent pattern:
🌍 Policy evidence is overwhelmingly sourced from the Global North.
Preprint: osf.io/w8q3y
🧪🧵👇
- Engagement can change—when demand for expertise shifts (like early in the COVID-19 pandemic), lawmakers' interactions with researchers increase. Notably in that case, engagement was targeted at medical scientists.
June 5, 2025 at 6:27 AM
- Engagement can change—when demand for expertise shifts (like early in the COVID-19 pandemic), lawmakers' interactions with researchers increase. Notably in that case, engagement was targeted at medical scientists.
- Political ideology & education matter:
🚨 A concerning takeaway in the current political environment... Conservative and radical right lawmakers are significantly less likely to follow or engage with academic researchers compared to others.
🚨 A concerning takeaway in the current political environment... Conservative and radical right lawmakers are significantly less likely to follow or engage with academic researchers compared to others.
June 5, 2025 at 6:27 AM
- Political ideology & education matter:
🚨 A concerning takeaway in the current political environment... Conservative and radical right lawmakers are significantly less likely to follow or engage with academic researchers compared to others.
🚨 A concerning takeaway in the current political environment... Conservative and radical right lawmakers are significantly less likely to follow or engage with academic researchers compared to others.
Here are some of the 🔑 findings:
- Legislators across the 12 countries DO follow researchers but seldom visibly engage.
*Also, most of their engagement is with social scientists.
- Legislators across the 12 countries DO follow researchers but seldom visibly engage.
*Also, most of their engagement is with social scientists.
June 5, 2025 at 6:27 AM
Here are some of the 🔑 findings:
- Legislators across the 12 countries DO follow researchers but seldom visibly engage.
*Also, most of their engagement is with social scientists.
- Legislators across the 12 countries DO follow researchers but seldom visibly engage.
*Also, most of their engagement is with social scientists.
I use data from lawmakers’ Twitter from countries across the Americas and Europe with a novel database of 410,000 academic researchers on the platform (paired with information about their academic profiles).
*This is the largest study of its kind on politician-academic interactions online.
*This is the largest study of its kind on politician-academic interactions online.
June 5, 2025 at 6:27 AM
I use data from lawmakers’ Twitter from countries across the Americas and Europe with a novel database of 410,000 academic researchers on the platform (paired with information about their academic profiles).
*This is the largest study of its kind on politician-academic interactions online.
*This is the largest study of its kind on politician-academic interactions online.
🤔 How much do politicians engage with academic researchers online?
In my latest paper, I find that politicians from 12 countries rarely engage with researchers on social media, but this can change when expertise gains salience
Preprint: osf.io/preprints/osf/wqbe4_v1
🧵👇
In my latest paper, I find that politicians from 12 countries rarely engage with researchers on social media, but this can change when expertise gains salience
Preprint: osf.io/preprints/osf/wqbe4_v1
🧵👇
June 5, 2025 at 6:27 AM
🤔 How much do politicians engage with academic researchers online?
In my latest paper, I find that politicians from 12 countries rarely engage with researchers on social media, but this can change when expertise gains salience
Preprint: osf.io/preprints/osf/wqbe4_v1
🧵👇
In my latest paper, I find that politicians from 12 countries rarely engage with researchers on social media, but this can change when expertise gains salience
Preprint: osf.io/preprints/osf/wqbe4_v1
🧵👇
For instance, I explore potential violations of exclusion restriction by ruling out that the treatment state affected conceptually unrelated variables (e.g., termination preferences towards residue right-wing paramilitary groups and non-FARC provisions.) THEY DON'T CHANGE! 👌
November 24, 2024 at 9:26 AM
For instance, I explore potential violations of exclusion restriction by ruling out that the treatment state affected conceptually unrelated variables (e.g., termination preferences towards residue right-wing paramilitary groups and non-FARC provisions.) THEY DON'T CHANGE! 👌
ON ROBUSTNESS: I closely follow, and thank for, the recommendations on UESD by @jordimunozm, @afalcogimeno, and @EHernandezPe in their @polanalysis piece. 📜 https://doi.org/10.1017/pan.2019.27 (You can take a look my Supplementary Materials for the checks.)
November 24, 2024 at 9:25 AM
ON ROBUSTNESS: I closely follow, and thank for, the recommendations on UESD by @jordimunozm, @afalcogimeno, and @EHernandezPe in their @polanalysis piece. 📜 https://doi.org/10.1017/pan.2019.27 (You can take a look my Supplementary Materials for the checks.)
On Sep 21, 2016, @JuanManSantos stood at UNGA and told the world: “After more than half a century of internal armed conflict, I come back to the UN today, on the Int. Day of Peace, to announce, with all the strength of my voice and of my heart: The war in Colombia has ended.”
November 24, 2024 at 9:25 AM
On Sep 21, 2016, @JuanManSantos stood at UNGA and told the world: “After more than half a century of internal armed conflict, I come back to the UN today, on the Int. Day of Peace, to announce, with all the strength of my voice and of my heart: The war in Colombia has ended.”
🤯 Today marks the 8th anniversary of the narrow rejection of the peace agreement in 🇨🇴. I would like to share some thoughts from my first solo-authored paper about the peace plebiscite and conflict-termination preferences out at @Res_Pol! 🕊️ 🔗 https://doi.org/10.1177/20531680241233793 🧵
November 24, 2024 at 9:25 AM
🤯 Today marks the 8th anniversary of the narrow rejection of the peace agreement in 🇨🇴. I would like to share some thoughts from my first solo-authored paper about the peace plebiscite and conflict-termination preferences out at @Res_Pol! 🕊️ 🔗 https://doi.org/10.1177/20531680241233793 🧵
2. We use text of the answers to open Q: “Please think again of the questions you just answered. What characteristics of a person would be most important to you in deciding who should have access to a vaccine first?” ≈40% of respondents signaled old age as the most important
November 24, 2024 at 9:25 AM
2. We use text of the answers to open Q: “Please think again of the questions you just answered. What characteristics of a person would be most important to you in deciding who should have access to a vaccine first?” ≈40% of respondents signaled old age as the most important
We examine two additional measures to gather further insights about attribute importance in the respondents' decisions. 1. We report on the results of the attribute importance when using random forests to model the decision-making process in the conjoint experiment.
November 24, 2024 at 9:24 AM
We examine two additional measures to gather further insights about attribute importance in the respondents' decisions. 1. We report on the results of the attribute importance when using random forests to model the decision-making process in the conjoint experiment.
This picture is reinforced in the conjoint. Healthcare worker candidates had about a 30pp higher probability of being selected over a similar unemployed profile. A candidate with preconditions had an about 15pp higher probability of being selected over one without.
November 24, 2024 at 9:24 AM
This picture is reinforced in the conjoint. Healthcare worker candidates had about a 30pp higher probability of being selected over a similar unemployed profile. A candidate with preconditions had an about 15pp higher probability of being selected over one without.
WHAT DID WE FIND? In the drag-and-drop ranking task, healthcare workers, people with medical preconditions, and the elderly were consistently ranked higher for priority access across countries.
November 24, 2024 at 9:24 AM
WHAT DID WE FIND? In the drag-and-drop ranking task, healthcare workers, people with medical preconditions, and the elderly were consistently ranked higher for priority access across countries.
OUR DESIGN We rely on two tasks: a) a priority ranking question of pre-defined societal groups and b) a conjoint experiment with pairs of hypothetical subjects, one of whom must be prioritized for vaccination.
November 24, 2024 at 9:24 AM
OUR DESIGN We rely on two tasks: a) a priority ranking question of pre-defined societal groups and b) a conjoint experiment with pairs of hypothetical subjects, one of whom must be prioritized for vaccination.
2. We pursued a variety of methodological approaches, which allow us to explore in a more nuanced fashion the role of different group and individual characteristics that would not be implementable in a ranking or choice task alone
November 24, 2024 at 9:24 AM
2. We pursued a variety of methodological approaches, which allow us to explore in a more nuanced fashion the role of different group and individual characteristics that would not be implementable in a ranking or choice task alone
Evidence in this camp is rapidly accumulating. These studies look at preference tensions a) globally ("rich vs. poor countries") and b) within countries and across population subgroups. Our study is part of the second set. Here is an overview of studies tackling these questions.
November 24, 2024 at 9:24 AM
Evidence in this camp is rapidly accumulating. These studies look at preference tensions a) globally ("rich vs. poor countries") and b) within countries and across population subgroups. Our study is part of the second set. Here is an overview of studies tackling these questions.
- We explored how political knowledge differed between cheaters and non-cheaters in predicting various related outcomes (Political interest, vote certainty, internal efficacy, and likelihood to vote)
November 24, 2024 at 9:23 AM
- We explored how political knowledge differed between cheaters and non-cheaters in predicting various related outcomes (Political interest, vote certainty, internal efficacy, and likelihood to vote)
4. Cheating does not seem to distort models of political knowledge - We compared estimates for predictors of political knowledge using unadjusted and cheating-adjusted measures
November 24, 2024 at 9:23 AM
4. Cheating does not seem to distort models of political knowledge - We compared estimates for predictors of political knowledge using unadjusted and cheating-adjusted measures
3. We find correlates of cheating. Correcting for it post hoc? Well, that’s another story… - We find robust person- and item-level correlates of cheating. For instance, we find an interesting difference between perceived and actual competence.
November 24, 2024 at 9:23 AM
3. We find correlates of cheating. Correcting for it post hoc? Well, that’s another story… - We find robust person- and item-level correlates of cheating. For instance, we find an interesting difference between perceived and actual competence.
1. Cheating is prevalent in the wild - In the German survey panel, 23% of the respondents engage at least once with outside sites for answers 😲 - Serial cheaters are rare - The prevalence of searches varies across question types
November 24, 2024 at 9:23 AM
1. Cheating is prevalent in the wild - In the German survey panel, 23% of the respondents engage at least once with outside sites for answers 😲 - Serial cheaters are rare - The prevalence of searches varies across question types