research paper on political system

Political polarization and its echo chambers: Surprising new, cross-disciplinary perspectives from Princeton

Artwork by Egan Jimenez

Much like an overexploited ecosystem, the increasingly polarized political landscape in the United States — and much of the world — is experiencing a catastrophic loss of diversity that threatens the resilience not only of democracy, but also of society, according to a series of new studies that examine political polarization as a collection of complex ever-evolving systems.

Fifteen interdisciplinary teams of political scientists and complex systems theorists in the natural sciences and engineering explored how polarization is produced and influenced over time by the actions and interactions of individual voters, people in power, and various social networks. Ultimately, as social interactions and individual decisions isolate people into only a few intractable camps, the political system becomes incapable of addressing the range of issues — or formulating the variety of solutions — necessary for government to function and provide the services critical for society.

The studies were published Dec. 6 in a special issue of the Proceedings of the National Academy of Sciences that stemmed from a collaboration between Princeton University and Arizona State University (ASU) and includes several papers led by Princeton researchers.

“The complex systems perspective demonstrates that the loss of diversity associated with polarization undermines cooperation and the ability of societies to provide the public goods that make for a healthy society,” according to an introduction by issue editors Simon Levin , Princeton’s James S. McDonnell Distinguished University Professor in Ecology and Evolutionary Biology , Helen Milner , the B.C. Forbes Professor of Public Affairs and professor of politics and international affairs at Princeton, and Charles Perrings , professor of environmental economics at ASU.

“Polarization is a dynamic process and that is what complexity theory can best help us understand,” they wrote. “As environmental and complexity scientists have shown in other contexts, diversity maintenance is critical for many systems to thrive, and often to survive at all.”

“The polarization currently being experienced both within and among nations  undercuts efforts to deal with critical issues facing societies , none more so than those related to the environment — from climate change and extreme weather, to the emergence of pandemic viruses such as COVID-19,” said Levin, who is director of the  Center for BioComplexity  based in Princeton’s  High Meadows Environmental Institute  (HMEI) and associated faculty in HMEI.

Complex adaptive systems — which are widespread in fields from physics and financial systems to natural systems driven by evolution and socioeconomic-political systems — allow scientists to understand the multiscale interactions that result in specific structures and outcomes, Levin said. At that point, efforts to mitigate negative results can be implemented more effectively.

“These systems are composed of individual agents, in which there is an interplay, and perhaps a coevolution, between the attitudes and actions of individual agents and the emergent properties of the systems to which they belong,” he said. “Similar challenges exist across these applications, involving the need for a statistical mechanics to scale from individuals to collectives, to the emergence of patterns and processes such as social norms.”

Despite the rise of partisanship, populism and polarization, these phenomena have not been thoroughly studied as dynamic systems consisting of multiple interacting components and large-scale features, Milner said.

“James Madison had hoped that the system devised in the Constitution would avoid the sorts of polarization that political parties can produce and that can undermine the workings of government,” Milner said.

“Sadly, we are seeing polarization today and a subsequent loss of diversity in the range of positions in society within the United States and globally,” she said. “The papers in this issue demonstrate from a systems perspective the forces that lead to polarization — and some of the consequences of it — with the hope that understanding them will lead to better governance.”

The studies from Princeton researchers are summarized below. The papers explored issues from how people unwittingly isolate themselves into partisan networks through social media and how to ensure successful electoral reforms using models, to how public opinion fuels extremism among political elites, as well as the potential benefits of polarization under the right circumstances.

People unwittingly polarize themselves by ditching followers considered untrustworthy 

A computational model tested with Twitter data showed that social media users may inadvertently sort themselves into polarized networks by “unfollowing” users they consider untrustworthy news sources. Princeton researchers  Andy Guess , assistant professor of politics and public affairs, Corina Tarnita , professor of ecology and evolutionary biology, and first author  Christopher Tokita , who received his Ph.D. from Princeton in 2021, found that when people are less reactive to news, their online environment remains politically mixed.

When users constantly react to and share articles from their preferred news sources, however, they are more likely to develop politically isolated networks, or what the researchers call “epistemic bubbles.” Once users are in these bubbles, they miss out on more news articles, including those from their preferred media outlets. 

“It’s not hard to find evidence of polarized discourse on social media, but we know less about the mechanisms of how social media can drive people apart,” Guess said. “Our contribution is to show that polarization of online social networks emerges naturally as people curate their feeds. Counterintuitively, this can occur even without knowing other users' partisan identities.”

Conservative swings in public opinion ramp up lawmaker extremism

While it is well-documented that Americans are not as polarized as the people they elect, a study led by  Naomi Ehrich Leonard , Princeton’s Edwin S. Wilsey Professor of Mechanical and Aerospace Engineering , and Keena Lipsitz , associate professor of political science at Queens College, CUNY, with Princeton doctoral student Anastasia Bizyaeva shows that Americans are still partly to blame for the extremism of their elected officials.  

The researchers found that over time, conservative swings in public opinion — which are typically slightly larger and more prolonged than liberal swings — exacerbate the self-reinforcement processes for Republican lawmakers, wherein legislators respond to favorable public opinion by further bolstering their own positions. They identified a tipping point beyond which the process of polarization speeds up as the forces driving it are compounded and the forces mitigating polarization are overwhelmed . They report that Republicans may have passed this critical threshold while Democrats are quickly approaching it.    

“By combining our expertise on political processes together with our expertise on feedback and nonlinearity in complex time-varying processes, we were able to make new discoveries about the mechanisms that can explain, and potentially mitigate, political polarization,” Leonard said.

“Until now, the ways in which public opinion changes over time had not been implicated in the political polarization of lawmakers,” she said. “Yet, by accounting for nonlinearity in how lawmakers respond to public opinion, we show that these differences matter significantly and small differences in public opinion swings can in fact lead to large changes in polarization. I am hopeful that the analytical tools we developed for this study will prove useful in finding ways to slow down the trend.”    

Progressive taxation could reduce economic hardships, social tensions fueling polarization

Intergroup conflict triggered by economic hardship can reduce social and economic interactions, which in turn further exacerbates economic decline and political polarization, according to a paper coauthored by Nolan McCarty , Princeton’s Susan Dod Brown Professor of Politics and Public Affairs, and  Joshua Plotkin , a professor of natural sciences at the University of Pennsylvania who received his Ph.D. from Princeton, with first author Alexander J. Stewart, senior lecturer in mathematical biology at the University of St. Andrews in Scotland.  The findings suggest that progressive taxation designed to ensure an adequate social safety net could  help prevent the economic anxieties that fuel ethnic and racial conflict.

“During the past 20 years, the United States and many other countries have experienced profound economic, social and political upheaval — including economic crises, escalating inequality, the exacerbation of racial and ethnic conflicts, and deepening political polarization,” McCarty said. “Our paper is an attempt to understand the complex dynamics that link these developments and explore ways to break the negative cycle.”

Diversity of social networks can intensify or moderate personal attitudes

The social networks to which people belong can “rewire” their personal attitudes over time to reflect the opinions of the people they’re linked to, according to a study led by former Princeton postdoctoral fellow Fernando Santos , an assistant professor at the University of Amsterdam, with Simon Levin , Princeton’s James S. McDonnell Distinguished University Professor in Ecology and Evolutionary Biology , and Yphtach Lelkes , associate professor of communication at the University of Pennsylvania.

The researchers found that when people preferentially connect to people with similar opinions, they create an echo chamber that increasingly polarizes the views of everyone in the network. On the other hand, people who are part of a network consisting of a variety of viewpoints tend to moderate one another. Understanding that social networks influence polarization — rather than merely reflect it — could be crucial in developing interventions to curb polarization online and the spread of political extremism, the researchers report.

“This is a relatively new phenomenon and, like other internet and media mechanisms, has likely sped and reinforced the segmentation of our societies,” Levin said.

Polarization can benefit society when opposing sides consist of diverse populations

Polarization may actually benefit society  when opposing viewpoints each represent a variety of people and communities with shared values, according to research led by  Vitor Vasconcelos , assistant professor at the University of Amsterdam and past postdoctoral research associate at Princeton, with  Elke Weber , the Gerhard R. Andlinger Professor in  Energy and the Environment  and professor of  psychology  and the School of Public and International Affairs, Princeton associate research scholar  Sara Constantino , and  Simon Levin , Princeton’s James S. McDonnell Distinguished University Professor in  Ecology and Evolutionary Biology , as well as Professor  Astrid Dannenberg  and research fellow  Mar­cel Lum­kow­sky  at the University of Kassel in Germany.

Polarization becomes harmful when it segregates social networks and excludes information about the preferences of people other than close neighbors. Cooperation becomes less likely when these local networks distort or undermine the value of working  with opponents, which can result in a number of effects including the weakening of democratic processes.

“Pluralistic societies thrive when members with different values and beliefs manage to discuss these differences and leverage them to generate win-win solutions,” Weber said. “Our paper shows that collective benefits are reduced by the polarization of social networks that restrict communication and negotiation across partisan lines, not the fact that we disagree on values.”

Contrarians at the gate: How strong local attitudes can breed opposition

Local variations in political attitudes can lead to polarization, particularly after political unrest, according to research led by Olivia Chu , a Princeton graduate student in quantitative and computational biology, with coauthors Grigore Pop-Eleches , professor of politics and international affairs, and Jonathan Donges , a visiting research collaborator in HMEI from the Potsdam Institute for Climate Impact Research. They deployed an adaptive voter model — which is used to study opinion dynamics — across Ukraine to determine how people’s perceptions of the European Union differed based on how people in their communities and social circles discussed revolutions, mass protests and other political shocks.

"Our research shows that rather than sweeping everyone along, the effect of revolutions on how people think about politics depends in part on the attitudes of the people with whom they talk about politics,” Pop-Eleches said. “Those who mostly talk to supporters of the revolution are likely to change their opinions in the opposite direction from those who talk to opponents. This can lead to pockets of increased polarization even in countries where most people support the goals of the revolution."

Partisan interpersonal interactions can weaken Madison’s cure for factions

A study led by Corina Tarnita , professor of ecology and evolutionary biology, and doctoral student Mari Kawakatsu in Princeton’s Program in Applied and Computational Mathematics examined how partisan interpersonal interactions can weaken processes that the framers of the US Constitution viewed as safeguards against factions and polarization. Kawakatsu and Tarnita co-authored the study with Simon Levin , Princeton’s James S. McDonnell Distinguished University Professor in Ecology and Evolutionary Biology , and Yphtach Lelkes , associate professor of communication at the University of Pennsylvania.

The researchers were inspired by James Madison’s essay, “Federalist No. 10,” in which he argued that a republic mitigates the dangers of factions by fostering a diversity of political interests. But Americans today care about many more political issues than they did 75 years ago, yet polarization is worse. The authors developed a theoretical model of cultural evolution to investigate the possible role that interactions among partisan opinionated citizens play in this puzzle.

Their analysis confirmed Madison’s intuition that societal cohesion increases when individuals care about a greater diversity of issues. But there is a twist — under extreme partisanship, individuals’ openness to learning from peers with a different political ideology is diminished. This leads to greater tribalism that drastically diminishes interest diversity, which leads to high within-ideology camaraderie and heightened polarization.

But the researchers also found a silver lining: the harmful effects of extreme partisanship are only substantial when individuals are primarily relying on social peers to shape their opinions and strategies and are limited in their independent exploration. “Our model suggests that actively pursuing learning from beyond one’s social network is crucial to maintaining a cohesive society,” Tarnita said. “Although both opinion formation and cooperation are well-explored topics, we understand relatively little about the coupled dynamics of cooperation and polarization,” Kawakatsu said. “The unexpected interactions we found between partisanship, cooperation and independent exploration highlight the need to study polarization in a coupled, multi-level context.”

Complex systems theory can lead to deeper understanding, better design of lasting reforms to American democracy

The implications of democracy reforms  such as ranked-choice voting and citizen redistricting may be better understood using dynamic systems theory based in engineering and biology, according to an analysis led by  Sam Wang , professor of  neuroscience  and director of the Electoral Innovation Lab at Princeton, with political scientists  Keena Lipsitz  at the City University of New York,  Jonathan Cervas  from Carnegie Mellon University, and  Bernard Grofman  at UC Irvine.

Wang and a multi-institutional team of political scientists report that systems-based theory typically used in the sciences can help understand the myriad of interactions that lead to current weaknesses in American democracy — particularly polarized institutions, unresponsive representatives and the ability of a faction of voters to gain power at the expense of the majority. Concepts such as nonlinearities and amplification, positive and negative feedback, and integration over time can help identify problems in representation and institutional power.

Similarly, the effectiveness of any proposed reform is difficult to predict against a backdrop of complex network interactions. A mathematically rich description of how electoral mechanisms interact can maximize the impacts of reforms in the context of the politics and procedures of individual states.

"Our core objective was to translate the American political system into a mathematical complex-systems framework that fosters participation by scholars of the natural sciences,” Wang said.

“We want to encourage natural scientists to build models that reproduce political phenomena, create simulations to explore alternative scenarios, and design interventions that may improve the function of democracy,” he said. “These goals are analogous to those of engineers — to understand a system of many parts well enough to make repairs or improvements."

The special issue, “The Dynamics of Political Polarization,” was published Dec. 6 by the Proceedings of the National Academy of Sciences.

B. Rose Huber in the Princeton School of Public and International Affairs, Molly Seltzer in the Andlinger Center for Energy and the Environment, and Molly Sharlach in Engineering Communications contributed to this story.

Related Stories

From muddy boots to mathematics: Advancing the science of ecosystems and biodiversity .

Princeton: A half-century at the environmental forefront

Fefferman and Levin receive BBVA Frontiers of Knowledge Awards .

Two professors, mathematician Charles Fefferman (also a 1969 Ph.D. alumnus) and ecologist Simon Levin, have been awarded the BBVA Foundation Frontiers of Knowledge Awards.

McCarty explores economic roots of today’s political strife .

Nasty clashes on hot topics such as the Iraq war, social issues and corruption may dominate the current political discourse, but the answer to why America is so polarized ultimately lies in the widening gulf between its richest and poorest citizens. That is the message of “Polarized America: The Dance of Ideology and Unequal Riches,” written by Princeton’s Nolan McCarty, a Congressional and presidential scholar.

Of ants and men: Ant behavior might mirror political polarization, say Princeton researchers .

A team of Princeton biologists found that division of labor and political polarization — two social phenomena not typically considered together — may be driven by the same processes in ant societies.

Less knowledge, more power: Uninformed can be vital to democracy, study finds .

A Princeton University-based research team reports in Science that uninformed individuals — as in those with no prior knowledge or strong feelings on a situation's outcome — can actually be vital to achieving a democratic consensus. These individuals tend to side with and embolden the numerical majority and dilute the influence of powerful minority factions who would otherwise dominate everyone else. This finding — based on group decision-making experiments on fish, as well as mathematical models and computer simulations — challenges the common notion that

Princeton faculty members receive grants for COVID-19 research from C3.ai Digital Transformation Institute .

The C3.ai Digital Transformation Institute has awarded $5.4 million to 26 projects to accelerate artificial intelligence research to mitigate COVID-19 and future pandemics. Princeton faculty members Matthew Desmond, Simon Levin, Stefana Parascho, H. Vincent Poor, Corina Tarnita and Mengdi Wang are among researchers to receive funding for their projects.

Vaccine stockpiling by nations could lead to increase in COVID-19 cases, novel variant emergence .

The allocation of COVID-19 vaccine between countries has thus far tended toward vaccine nationalism, wherein countries stockpile vaccines to prioritize access for their citizenry over equitable vaccine sharing. The extent of vaccine nationalism, however, may strongly impact global trajectories of COVID-19 case numbers and increase the potential emergence of novel variants, according to a Princeton University and McGill University study published Aug. 17 in the journal Science.

Academia.edu no longer supports Internet Explorer.

To browse Academia.edu and the wider internet faster and more securely, please take a few seconds to  upgrade your browser .

•  We're Hiring!
•  Help Center

Political System

• Most Cited Papers
• Most Downloaded Papers
• Newest Papers
• Last »
• Iraq after 2003 Follow Following
• political instability and Economic Growth Follow Following
• Ilmu Negara Follow Following
• Pengantar Ilmu Hukum Follow Following
• Judicial independence Follow Following
• Universitas Sriwijaya Follow Following
• Politics of New Zealand Follow Following
• Accountability Follow Following
• Experimental Methods Follow Following
• Multidisciplinary Follow Following

Enter the email address you signed up with and we'll email you a reset link.

• Academia.edu Journals
•   We're Hiring!
•   Help Center
• Find new research papers in:
• Health Sciences
• Earth Sciences
• Cognitive Science
• Mathematics
• Computer Science
• Academia ©2024

We've detected unusual activity from your computer network

To continue, please click the box below to let us know you're not a robot.

Why did this happen?

Please make sure your browser supports JavaScript and cookies and that you are not blocking them from loading. For more information you can review our Terms of Service and Cookie Policy .

For inquiries related to this message please contact our support team and provide the reference ID below.

A new ‘AI scientist’ can write science papers without any human input. Here’s why that’s a problem

Dean, School of Computing Technologies, RMIT University, RMIT University

Disclosure statement

Karin Verspoor receives funding from the Australian Research Council, the Medical Research Future Fund, the National Health and Medical Research Council, and Elsevier BV. She is affiliated with BioGrid Australia and is a co-founder of the Australian Alliance for Artificial Intelligence in Healthcare.

RMIT University provides funding as a strategic partner of The Conversation AU.

View all partners

Scientific discovery is one of the most sophisticated human activities. First, scientists must understand the existing knowledge and identify a significant gap. Next, they must formulate a research question and design and conduct an experiment in pursuit of an answer. Then, they must analyse and interpret the results of the experiment, which may raise yet another research question.

Can a process this complex be automated? Last week, Sakana AI Labs announced the creation of an “AI scientist” – an artificial intelligence system they claim can make scientific discoveries in the area of machine learning in a fully automated way.

Using generative large language models (LLMs) like those behind ChatGPT and other AI chatbots, the system can brainstorm, select a promising idea, code new algorithms, plot results, and write a paper summarising the experiment and its findings, complete with references. Sakana claims the AI tool can undertake the complete lifecycle of a scientific experiment at a cost of just US$15 per paper – less than the cost of a scientist’s lunch.

These are some big claims. Do they stack up? And even if they do, would an army of AI scientists churning out research papers with inhuman speed really be good news for science?

How a computer can ‘do science’

A lot of science is done in the open, and almost all scientific knowledge has been written down somewhere (or we wouldn’t have a way to “know” it). Millions of scientific papers are freely available online in repositories such as arXiv and PubMed .

LLMs trained with this data capture the language of science and its patterns. It is therefore perhaps not at all surprising that a generative LLM can produce something that looks like a good scientific paper – it has ingested many examples that it can copy.

What is less clear is whether an AI system can produce an interesting scientific paper. Crucially, good science requires novelty.

But is it interesting?

Scientists don’t want to be told about things that are already known. Rather, they want to learn new things, especially new things that are significantly different from what is already known. This requires judgement about the scope and value of a contribution.

The Sakana system tries to address interestingness in two ways. First, it “scores” new paper ideas for similarity to existing research (indexed in the Semantic Scholar repository). Anything too similar is discarded.

Second, Sakana’s system introduces a “peer review” step – using another LLM to judge the quality and novelty of the generated paper. Here again, there are plenty of examples of peer review online on sites such as openreview.net that can guide how to critique a paper. LLMs have ingested these, too.

AI may be a poor judge of AI output

Feedback is mixed on Sakana AI’s output. Some have described it as producing “ endless scientific slop ”.

Even the system’s own review of its outputs judges the papers weak at best. This is likely to improve as the technology evolves, but the question of whether automated scientific papers are valuable remains.

The ability of LLMs to judge the quality of research is also an open question. My own work (soon to be published in Research Synthesis Methods ) shows LLMs are not great at judging the risk of bias in medical research studies, though this too may improve over time.

Sakana’s system automates discoveries in computational research, which is much easier than in other types of science that require physical experiments. Sakana’s experiments are done with code, which is also structured text that LLMs can be trained to generate.

AI tools to support scientists, not replace them

AI researchers have been developing systems to support science for decades. Given the huge volumes of published research, even finding publications relevant to a specific scientific question can be challenging.

Specialised search tools make use of AI to help scientists find and synthesise existing work. These include the above-mentioned Semantic Scholar, but also newer systems such as Elicit , Research Rabbit , scite and Consensus .

Text mining tools such as PubTator dig deeper into papers to identify key points of focus, such as specific genetic mutations and diseases, and their established relationships. This is especially useful for curating and organising scientific information.

Machine learning has also been used to support the synthesis and analysis of medical evidence, in tools such as Robot Reviewer . Summaries that compare and contrast claims in papers from Scholarcy help to perform literature reviews.

All these tools aim to help scientists do their jobs more effectively, not to replace them.

AI research may exacerbate existing problems

While Sakana AI states it doesn’t see the role of human scientists diminishing, the company’s vision of “a fully AI-driven scientific ecosystem” would have major implications for science.

One concern is that, if AI-generated papers flood the scientific literature, future AI systems may be trained on AI output and undergo model collapse . This means they may become increasingly ineffectual at innovating.

However, the implications for science go well beyond impacts on AI science systems themselves.

There are already bad actors in science, including “paper mills” churning out fake papers . This problem will only get worse when a scientific paper can be produced with US$15 and a vague initial prompt.

The need to check for errors in a mountain of automatically generated research could rapidly overwhelm the capacity of actual scientists. The peer review system is arguably already broken , and dumping more research of questionable quality into the system won’t fix it.

Science is fundamentally based on trust. Scientists emphasise the integrity of the scientific process so we can be confident our understanding of the world (and now, the world’s machines) is valid and improving.

A scientific ecosystem where AI systems are key players raises fundamental questions about the meaning and value of this process, and what level of trust we should have in AI scientists. Is this the kind of scientific ecosystem we want?

• Artificial intelligence (AI)
• Computer science
• Research integrity
• Paper mills

Project Manager SSTP

Head of Evidence to Action

Supply Chain - Assistant/Associate Professor (Tenure-Track)

OzGrav Postdoctoral Research Fellow

Casual Facilitator: GERRIC Student Programs - Arts, Design and Architecture

IBM China said to be laying off more than 1,000 employees as it closes research labs

• IBM is the latest global tech giant to cut jobs in China this year, as an intensified Sino-US rivalry threatens their mainland businesses

US computing giant IBM has reportedly shut down its research and development (R&D) operations in China, joining a slew of global Big Tech firms in trimming their mainland businesses amid geopolitical headwinds.

IBM is closing its China Development Lab and China Systems Lab, while laying off more than 1,000 employees in cities including Beijing, Shanghai and the northern port city Dalian, according to reports by local news outlets.

IBM’s China-based R&D employees over the weekend found themselves blocked from accessing the company’s intranet system, Chinese news website Jiemian reported on Saturday. The Armonk, New York-headquartered company announced the job cuts during an internal meeting on Monday morning, according to posts by multiple employees on Chinese social media platforms.

“IBM adapts its operations as needed to best serve our clients, and these changes will not impact our ability to support clients across the Greater China region,” an IBM representative said in an email to the Post, without providing details of the lay-offs.

IBM’s local strategy is “focused on having the right teams with the right skills” to help Chinese companies – especially privately owned firms – co-create hybrid cloud and artificial intelligence (AI) solutions by drawing on its “considerable technology and consulting expertise”, the representative said.

IBM is the latest multinational tech giant to shed jobs in China, as an intensified Sino-US rivalry forces global businesses to adjust their operations on the mainland.

IBM’s sales in China have steadily declined in recent years.

In 2023, IBM’s revenue in the country dropped 19.6 per cent compared to a 1.6 per cent rise in revenue across Asia-Pacific, according to the company’s annual report. Sales in China in the six months ended June 30 this year fell 5 per cent, while revenue in Asia-Pacific increased 4.4 per cent, IBM’s financial statement showed.

Still, IBM China credited its Development Lab for making “important contributions” to the development of the company’s enterprise-facing generative AI development platform WatsonX, in a blog post published on WeChat last November.

IBM announced WatsonX in May last year and made it available to customers in China in the following August.

The China Development Lab had “more than 24 years of outstanding development experience” and was behind hundreds of main and innovative products, IBM said in the WeChat post.

IBM reported 2 per cent growth in global revenue for the second quarter, with software sales up 7 per cent. Its shares have jumped 21 per cent since the beginning of this year.

• All Research Labs
• 3D Deep Learning
• Applied Research
• Autonomous Vehicles
• Deep Imagination
• New and Featured
• AI Art Gallery
• AI & Machine Learning
• Computer Vision
• Academic Collaborations
• Government Collaborations
• Graduate Fellowship
• Internships
• Research Openings
• Research Scientists
• Meet the Team
• Publications

Kilometer-Scale Convection Allowing Model Emulation using Generative Diffusion Modeling

Storm-scale convection-allowing models (CAMs) are an important tool for predicting the evolution of thunderstorms and mesoscale convective systems that result in damaging extreme weather. By explicitly resolving convective dynamics within the atmosphere they afford meteorologists the nuance needed to provide outlook on hazard. Deep learning models have thus far not proven skilful at km-scale atmospheric simulation, despite being competitive at coarser resolution with state-of-the-art global, medium-range weather forecasting. We present a generative diffusion model called StormCast, which emulates the high-resolution rapid refresh (HRRR) model—NOAA’s state-of-the-art 3km operational CAM. StormCast autoregressively predicts 99 state variables at km scale using a 1-hour time step, with dense vertical resolution in the atmospheric boundary layer, conditioned on 26 synoptic variables. We present evidence of successfully learnt km-scale dynamics including competitive 1-6 hour forecast skill for composite radar reflectivity alongside physically realistic convective cluster evolution, moist updrafts, and cold pool morphology. StormCast predictions maintain realistic power spectra for multiple predicted variables across multi-hour forecasts. Together, these results establish the potential for autoregressive ML to emulate CAMs – opening up new km-scale frontiers for regional ML weather prediction and future climate hazard dynamical downscaling.

Publication Date

Research area, uploaded files.

Numbers, Facts and Trends Shaping Your World

Read our research on:

Full Topic List

Regions & Countries

• Publications
• Our Methods
• Short Reads
• Tools & Resources

Read Our Research On:

1. Democracy and government, the U.S. political system, elected officials and governmental institutions

Americans are generally positive about the way democracy is working in the United States. Yet a majority also says that the “fundamental design and structure” of U.S. government is in need of “significant changes” to make it work today.

Republicans are more likely than Democrats to say U.S. democracy is working at least somewhat well, and less likely to say government is in need of sweeping changes.

And far more Republicans than Democrats say the U.S. political system is “best in the world” or “above average” when compared with political systems of other developed nations.

Overall, about six-in-ten Americans say democracy is working well in the U.S. today (18% very well, 40% somewhat well); four-in-ten say it is not working well (27% not too well and 13% not at all well).

About seven-in-ten (72%) Republicans and Republican-leaning independents say U.S. democracy is working very or somewhat well, compared with 48% of Democrats and Democratic leaners. Relatively small shares in both parties (30% of Republicans and just 7% of Democrats) say democracy in the U.S. is working very well.

While a majority of Americans say democracy in this country is working well, about six-in-ten (61%) say significant changes to the fundamental design and structure of government are needed to make it work for current times; 38% say the design and structure of government serves the country well and does not need significant changes.

By roughly two-to-one (68% to 31%), Democrats say significant changes are needed, while Republicans are divided (50% to 49%) over whether or not extensive changes are needed.

Although the view that significant changes are needed is widely held, those with higher levels of political engagement are less likely to say this than people who are less politically engaged.

Overall, those with high levels of political engagement and participation are split over whether significant changes are needed or not (51% vs. 48%). Views that the American system of government needs far-reaching reforms are more widespread among those with lower levels of engagement: 60% of those with a moderate level of engagement say this, along with 71% of those who are relatively unengaged with politics.

This pattern is evident within both partisan coalitions: 40% of Republicans and Republican leaners who are highly engaged with politics say the fundamental design and structure of American government needs significant reform, compared with 60% of low-engagement Republicans. Similarly, while a 57% majority of highly engaged Democrats and Democratic leaners say significant changes are needed, that share rises to 78% of the least politically engaged Democrats.

Across demographic groups, there are only modest differences in the shares saying that democracy is working at least somewhat well, but there are more pronounced differences on whether changes are needed to the fundamental design and structure of government.

Whites (54%) are less likely than blacks (70%) and Hispanics (76%) to say the government needs significant change, but the three groups have similar assessments of American democracy’s performance.

There also are significant age gaps over whether extensive change is needed to the structure and design of government, with 66% of adults younger than 50 saying this, compared with 58% of those ages 50 to 64 and 50% of those 65 and older. But age groups differ little in their evaluations of how well democracy is functioning.

Educational groups also differ little in their overall opinions of how well democracy is working. But those without a bachelor’s degree (65%) are more likely to say the government needs significant change than those with a college degree (54%) or a postgraduate degree (45%).

Americans give their political system mixed grades

When asked to compare the U.S. political system with others in developed countries, only about four-in-ten Americans (41%) say it is “best in the world” or “above average.” Most (57%) say it is “average” or “below average.”

Several other national institutions and aspects of life in the U.S. are more highly rated than the political system. Nearly eight-in-ten (79%) say the U.S. military is either above average or the best in the world compared with militaries in other developed nations – with 38% calling it best in the world.

Larger shares also say the U.S. standard of living, colleges and universities, scientific achievements and economy are at least above average internationally than say that about the political system. Only the nation’s health care system (30% best in the world or above average) and public schools (18%) are rated lower.

Republicans and Republican-leaning independents generally give the U.S. better marks for its performance on these issues than Democrats and Democratic leaners. About six-in-ten Republicans say the country’s political system is above average or the best in the world (58%), compared with about a quarter of Democrats (27%). Republicans also give the country much higher marks than Democrats on its standard of living, health care and economy.

The shares of Republicans and Democrats giving the U.S. high marks on several of these national institutions and aspects of American life have diverged sharply since 2014.

Today, Republicans are about twice as likely as Democrats to say the U.S. political system is above average or the best in the world (58% vs. 27%).

In 2014, about four-in-ten members of both parties gave the political system a positive rating (37% of Republicans, 36% of Democrats); in 2009, identical shares of Republicans and Democrats (52% each) said the U.S. political system was at least above average.

Partisan divides are growing in other areas as well. For example, 61% of Republicans and just 38% of Democrats describe the U.S. economy as best in the world or above average. Partisan differences in these assessments were much more modest in 2014 and 2009.

Little public confidence in elected officials

Americans express little confidence in elected officials to act in the best interests of the public. Just a quarter say they have a great deal (3%) or fair amount (22%) of confidence in elected officials.

That is by far the lowest level of confidence in the six groups included in the survey. Large majorities say they have a great deal or fair amount of confidence in the military (80%) and scientists (79%). In addition, higher shares express confidence in religious leaders (49%), business leaders (44%) and the news media (40%).

Overall public confidence in these groups is little changed since 2016, but in some cases – including elected officials – the views among Republicans and Democrats have shifted.

Though majorities of both Republicans and Democrats continue to express little or no confidence in public officials, Republicans (36%) are more likely than Democrats (17%) to express at least a fair amount of confidence in elected officials to act in the public interest. Two years ago, more Democrats (32%) than Republicans (22%) had confidence in elected officials.

The partisan gap in confidence in the news media also has widened considerably. Today, 58% of Democrats and just 16% of Republicans are confident in the news media to act in the public interest. Since 2016, the share expressing at least a fair amount of confidence in the news media has increased 12 percentage points among Democrats, while falling 13 points among Republicans.

And more Republicans have confidence in business leaders than did so two years (62% now, 51% then). Far fewer Democrats express confidence in business leaders (32%), and their views are little changed from two years ago.

Republicans also express more confidence in the military (92%) than do Democrats (73%), and the gap has not changed much since 2016.

State, local governments viewed more favorably than federal government

Americans have more favorable opinions of their state and local governments than the federal government in Washington. Two-thirds say they view their local government favorably, and 58% have favorable views of their state government. Only 35% of adults report a favorable opinion of the federal government.

Views of federal, state and local government have changed little over the past decade. Favorable opinions of the federal government have fallen significantly since peaking in the wake of the 9/11 terrorist attacks.

While overall views of the federal government in Washington are largely unchanged from late 2015, Republicans and Democrats have moved in opposite directions since then.

Today, 44% of Republicans and Republican leaners have a favorable opinion of the federal government, compared with 28% of Democrats and Democratic leaners. In 2015, views of the federal government were reversed: 45% of Democrats had a favorable view versus 18% of Republicans. Republicans’ and Democrats’ views of the federal government also flipped between 2008 and 2009, when Barack Obama won the presidency.

There are much smaller partisan differences in favorability toward states and local government. Majorities in both parties (61% of Republicans, 55% of Democrats) have favorable impressions of their state government; similar shares in both parties (69% of Republicans, 68% of Democrats) view their local governments favorably.

Views of Congress and the Supreme Court

Views of Congress remain extremely negative: Two-thirds of Americans say they have an unfavorable view of Congress, compared with 30% saying their view is favorable. The share expressing unfavorable views has increased slightly from a year ago (62%).

With their party in control of both houses of Congress, Republicans’ views are slightly more favorable than Democrats: 37% of Republicans and Republican leaners say this versus 24% of Democrats and Democratic leaners. Republican’s attitudes are more negative than a year ago, when 44% had a favorable opinion. Views among Democrats are mostly unchanged.

Attitudes toward the Supreme Court continue to improve after reaching 30-year lows in 2015 . Republicans’ views, in particular, are now more positive than three years ago.

Two-thirds of the public says they view the court favorably, and about three-in-ten (28%) hold unfavorable views. The share of the public saying it has a favorable view of the Supreme Court has increased 18 percentage points since 2015 (48%).

Most Republicans viewed the Supreme Court unfavorably after its decisions on the Affordable Care Act and same-sex marriage in summer 2015: Just a third of Republicans viewed the court favorably, compared with about six-in-ten Democrats (61%). Today, more Republicans (71%) hold a favorable view of the Supreme Court than Democrats (62%). Favorable views among Democrats have fallen since 2016.

Sign up for our weekly newsletter

Fresh data delivery Saturday mornings

Sign up for The Briefing

Weekly updates on the world of news & information

Most Popular

901 E St. NW, Suite 300 Washington, DC 20004 USA (+1) 202-419-4300 | Main (+1) 202-857-8562 | Fax (+1) 202-419-4372 |  Media Inquiries

Research Topics

• Email Newsletters

ABOUT PEW RESEARCH CENTER  Pew Research Center is a nonpartisan fact tank that informs the public about the issues, attitudes and trends shaping the world. It conducts public opinion polling, demographic research, media content analysis and other empirical social science research. Pew Research Center does not take policy positions. It is a subsidiary of  The Pew Charitable Trusts .

© 2024 Pew Research Center

Advertisement

Supported by

The Radical Strategy Behind Trump’s Promise to ‘Go After’ Biden

Conservatives with close ties to Donald J. Trump are laying out a “paradigm-shifting” legal rationale to erase the Justice Department’s independence from the president.

• Share full article

By Jonathan Swan Charlie Savage and Maggie Haberman

When Donald J. Trump responded to his latest indictment by promising to appoint a special prosecutor if he’s re-elected to “go after” President Biden and his family, he signaled that a second Trump term would fully jettison the post-Watergate norm of Justice Department independence.

“I will appoint a real special prosecutor to go after the most corrupt president in the history of the United States of America, Joe Biden, and the entire Biden crime family,” Mr. Trump said at his golf club in Bedminster, N.J., on Tuesday night after his arraignment earlier that day in Miami. “I will totally obliterate the Deep State.”

Mr. Trump’s message was that the Justice Department charged him only because he is Mr. Biden’s political opponent, so he would invert that supposed politicization. In reality, under Attorney General Merrick Garland, two Trump-appointed prosecutors are already investigating Mr. Biden’s handling of classified documents and the financial dealings of his son, Hunter.

But by suggesting the current prosecutors investigating the Bidens were not “real,” Mr. Trump appeared to be promising his supporters that he would appoint an ally who would bring charges against his political enemies regardless of the facts.

The naked politics infusing Mr. Trump’s headline-generating threat underscored something significant. In his first term, Mr. Trump gradually ramped up pressure on the Justice Department, eroding its traditional independence from White House political control. He is now unabashedly saying he will throw that effort into overdrive if he returns to power.

Mr. Trump’s promise fits into a larger movement on the right to gut the F.B.I., overhaul a Justice Department conservatives claim has been “weaponized” against them and abandon the norm — which many Republicans view as a facade — that the department should operate independently from the president.

We are having trouble retrieving the article content.

Please enable JavaScript in your browser settings.

Thank you for your patience while we verify access. If you are in Reader mode please exit and  log into  your Times account, or  subscribe  for all of The Times.

Thank you for your patience while we verify access.

Already a subscriber?  Log in .

Want all of The Times?  Subscribe .