Rising political activism in prestige scientific journals and magazines

A preliminary look at the change in political content of some prestige scientific journals and magazines over time

TLDR:

• The testimony of the leaders of Harvard, MIT and University of Pennsylvania has caused widespread concern about the politicisation of our research and education institutions. An element of this which is not getting consideration is the role that academic magazines may have in providing information and sculpting opinions.

• We performed a pilot study comparing the political content in prominent scientific journals and magazines Nature, Science, and Scientific American in a 3 month period in 2002 and in 2022, and found that there was a marked politicisation of these highly prominent publications.

• Science journals and magazines 20 years ago used to deal almost exclusively with science - specifically scientific results, discoveries and inventions. They have now become far more focussed on broader politics. Relative to the same time period in 2002, the frequency of political articles in Science increased ten fold, Nature by six fold, and in the window chosen infinitely so in Scientific American due to a denominator of 0 back in 2002, going from averaging zero political articles per issue to 3.3 per issue in 2022.

• Whilst some of this is unbiased reportage of events, or a discussion of how science could be used in society, there is a major increase in pieces advocating particular political viewpoints. When this occurs, they are almost always from the left of center and only very rarely give differing viewpoints.

• In particular, there is a sharp rise in coverage of issues broadly defined as ‘identity politics’. Some of these identity-related articles would have broad agreement across the population, such as supporting Afghan girls being in school, but there is also a major element of what could loosely be described as ‘hard-left’ identity politics, that may be out of step with public opinion. For example, in the examined time period in 2002 Nature and Science had no pieces at all in 3 months focussing on a specific human identity, whilst by 2022 they were averaging almost 2 identity-oriented articles per issue.

• Even after excluding identity-related articles, there are many more articles advocating particular political viewpoints. The frequency of this doubled in Nature, whilst it has gone from being essentially completely absent from Science and Scientific American to occurring in most issues, particularly in the latter.

• We also noted a strongly growing tendency for journals to speak about and state what the views of researchers are on issues, as though they are globally homogenous (ie, ‘scientists want this’ or ‘scientists think this about issue Y’). Yet we could not find any examples where this was actually based on or reported representative polling of scientists and technology-related people, or comparison with views of the broader public. Rather it appears to be simply based on the opinions and preferences of the staff of these scientific journals, whose political opinions may not be representative of the community or society as a whole.

---

Most of us have now seen the deeply chilling testimony to congress of the leaders of Harvard, MIT, and University of Pennsylvania. If you have not I highly recommend watching this 3 and a half minute excerpt.

There has been an enormous backlash, and a lot of people are concerned about what this tells us about our institutions of research and education. Even many news outlets and writers which rarely report on this space are covering what is a substantial crisis for confidence in our research and education institutions, exposing concerning politicisation of them. Already one of the University leaders has resigned, and a bipartisan letter from members of congress has called for the boards remove all of them. Many people are speaking out who were not before, with even Patrick Collison and Sam Altman tweeting deep concern, as has Laurence Tribe.

It’s important in trying to understand how we got to this point to consider what influences are acting on these institutions and the people in them. Specifically I want to explore scientific journals and magazines - for now the main communication channels for scientists. These are usually controlled by companies, nonprofits, and learned societies.

Whilst people outside science may think the role of a scientific journal or magazine is to convey scientific discoveries or inventions in an impartial and accurate way, they now cover a lot more politics and take specific views on it. On this note, its relevant that in contrast to the major attention that the Congressional testimony and fallout has had across the old media and social media, there has been has been absolutely no pick up on this issue at all in the world’s two most prominent scientific magazines. As I tweeted earlier:

As outlined below, there is no shortage of Nature or Science wading into politics. Quite the opposite, in fact. So it’s curious that they don’t think it is a topic worth covering at all for their readers, as it’s one of the biggest reputational crises the sector has ever faced.

The lack of coverage raises questions about whether or not these journals are reporting news and views in an unbiased, impartial way, and whether they are providing a full picture of what is relevant news and information for their readership. The alternative is they are creating an information bubble that effectively operates to indoctrinate.

Last year I was part of a small group that did a pilot study addressing this question.

In a nutshell, we found that the most prominent venues in science have become quite politicised and are not giving a balanced and impartial view of the world to scientists, nor exploring complex issues in a way that presents different viewpoints. We originally published this in March of this year through an anonymous blog called ‘researchers for impartiality’ as some of our authors and analysts did not want to be identified, but given recent events I’m republishing it on my blog hoping it might be useful to people trying to understand what is happening in the research world.

I think even those of us who share the viewpoints consistently adopted by these magazines should be concerned about what this may be doing to science and its reputation. My own view is it’s damaging to science long term. Already confidence in US universities has almost halved since 2015 [gallup poll].

I got interested in this issue as I’d been particularly concerned by the behaviour of Nature and Science giving highly misleading and selective analysis of the COVID lab leak debate, an issue I had been following very closely when I was inside government. I felt their misrepresentations were meaningfully altering decision making on answering important questions. Recently a bipartisan group including very senior US National Security professionals wrote an open letter raising concerns about the journals’ conduct and I agree with what they wrote. I think this phenomena of an illusion of consensus in the research community on politically sensitive issues being created is happening much more broadly than just this one issue.

Stuart Ritchie, author of ‘Science Fictions’, wrote an excellent argument as to why it is so damaging for science to be politicised. I encourage you to read it if you read what’s below and simply shrug and think it isn’t consequential.

Science is political - and that's a bad thing

Piece as published 22nd March 2023

---

Political activism in prestige scientific journals

A preliminary look at the change in political content of some prestige scientific journals and magazines over time

Nature just published a news piece that has gotten significant attention.

It’s based on a study from Nature Human Behaviour, available here, by Floyd Jiuyun Zhang, showing that Nature’s endorsement of Joe Biden reduced Trump supporters’ trust in Nature and in science more broadly, while having no effect on voting intention. Nature’s editorial on the subject begins - ‘We live in an era in which there are many attempts to politicize science.’ 

This raises an interesting question: to what extent has Nature itself–along with other prestige scientific journals– become politicized?

Anecdotally, many people have noticed what appears like an increasing amount of political content in journals like Nature and Science in recent years. This content tends to have a common ideological standpoint, favoring broadly left-of-center perspectives. It is notable that the political content typically appeared not in the journals’ peer-reviewed scholarly articles, but in their editorial, news and correspondence sections.

We do not believe this will come as a surprise to long-standing readers of Nature and Science. But nevertheless, rather than relying on subjective impressions and anecdote, we thought it would be useful to try to observe and quantify this change.

We conducted a short pilot study in May 2022, which we revisited in late November 2022. In this study, we looked at the contents of the prestigious scientific journals Nature and Science and the popular-science magazine Scientific American over two three-month periods, in 2002 and 2022, and attempted to identify the number of politically oriented articles and their content. The study confirmed our impression that there had been a very significant increase in the amount of political content in the three periodicals.

Readers will no doubt have their own views on whether it is problematic that prestigious journals have increased their political content, and that it seems to have a common ideological slant. 

We hope that this study brings some useful facts to an important debate, and provides a basis for those who want to build on, expand or refine the analysis. 

What follows is the text we had, with light revisions, in November 2022.

—----

Key findings:

• Science journals and magazines 20 years ago used to deal almost exclusively with science - specifically scientific results, discoveries and inventions. They have now become far more focussed on broader politics. 

• Whilst some of this is unbiased reportage of events, or a discussion of how science could be used in society, there is a strong increase in pieces advocating particular political viewpoints. When this occurs, they are almost always from the left of center.

• In particular, there is a sharp rise in coverage of issues broadly defined as ‘identity politics’. Some of these identity-related articles would have broad agreement across the population, such as supporting Afghan girls being in school, but there is also a strong element of what could loosely be described as ‘hard-left’ identity politics, that may be out of step with public opinion.

• We could not find examples where this was based on or reported representative polling of scientists and technology-related people, or comparison with views of the broader public. Rather it appears to be based on the opinions and preferences of the staff of these scientific journals.

For example, in the examined time period, the editorials in the top US science journal Science had far greater advocacy and coverage of political and identity issues than it did coverage of science issues.

There were 13 weeks in the examined period. 4/13 weeks of editorials covered topics which can be deemed primarily science focussed, such as changing science education, discussing the new ARPA-H agency, the growing African space enterprise, and ‘mental health 2.0’.

However, 5/13 weeks of editorials covered identity related topics such as complaining that Biden hasn’t appointed a black person as his chief science advisor and in our view implicitly accusing him of racist motives, another advocated removing standardized cognitive testing in college admissions to increase racial equity, another focussed on ‘faculty must lead inclusion efforts’, another editorial complained about the Supreme Court Abortion decision making process, and another examined how equity is a new goal for global science philanthropy. In no cases were counter-arguments offered. 

There were also 3 weeks of predominantly political advocacy, such as complaining about Biden calling an end to COVID, calls for more environmental targets, and calling for peace and engagement with China to solve climate change.

So 9/13 of the editorials in Science in this period were primarily political. Whilst some of these would be supported by a broad base of people, it nonetheless represents an overt politicization of the editorial base of a scientific journal.

We also don’t think our findings are a product of the time window analyzed. In just the few days in which we originally wrote this article [November 2022], many similar articles turned up in Nature alone:

• ‘Seeding an anti-racist culture at Scotland’s botanical gardens: Botanical gardens are re-examining their collections’ colonial roots — botanists of colour say keep going.” 22nd November.  Note the way it writes as though ‘botanists of color’ are a homogenous group who all agree with this. This was part of a ‘decolonizing science toolkit’ series that the journal has run in recent weeks, including ‘Imperialism’s long shadow: the UK universities grappling with a colonial past - Many research institutions in the United Kingdom gained prominence as the British Empire reached its zenith — bringing them a painful reckoning as they attempt to decolonize.’

• Decolonizing the biosciences: Turning lip service into action- Prevent ‘equity washing’ by making research collaborations with level partnerships. 18th Nov 2022.

• The rise of scientific racism in palaeoanthropology - A forensic anthropologist unmasks insidious interpretations of fossil finds. 14th Nov 2022.

• How science museums can use their power: A trawl through exhibition halls and storage rooms reveals a drive to do better. 21st November. This examines museums as coming from the ‘payday of empire’ and how they can use their power to fight ‘colonialism’.

• Institutions must acknowledge the racist roots in science: Funders and universities should move beyond talk to actions. 24th November 2022.

• Tackle systemic racism to diversify health care and clinical research: Solving structural-racism problems in health will require everyone, from community members to heads of university departments, to be engaged. 25th November.

All of this in little over a week, just in Nature. And this followed, just a month before, a whole issue of Nature dedicated to ‘anti-racism’, whilst the month before its lead editorial covered “How Nature contributed to science’s discriminatory legacy”. 

We found no articles advocating counterviews. For example,  Nature’s coverage of the Wellcome Trust’s ‘systemic racism declaration’ featured no dissenting views, and stated that ‘Ultimately, diversity experts want to see action, rather than lip service, on racism and equality’. We do not know what constitutes a ‘diversity expert’ - some of the ‘diversity experts’ cited in the article appear prima facie from internet searches to be activists.

We selected these venues for analyses due to their prominence and prestige. But this is by no means unique to these journals. For example, the HHMI and Wellcome-funded eLife journal recently ran a feature on ‘Equity, Diversity, and Inclusion: how to write an anti-racist tenure letter’, including “Here we propose a set of practical steps for writing inclusive, anti-racist tenure letters…..We are…working toward Black liberation in academia……Whether we are aware of it or not, academic culture is steeped in the beliefs and values Okun associates with ‘white supremacy culture’....”. It went on to list things like a sense of ‘urgency’, ‘perfectionism’, and ‘defensiveness’ as signs of white supremacy culture in academia, and made recommendations “Based on Okun, 2021 antidotes to ‘white supremacy culture’” for letter writers. Again, there was no attempt at balance or presenting alternative interpretations.

Pilot study on rising politicization of science journals

Background

Society depends on the authority of science as an independent search for truth free from politics or agenda. Scientific journals remain the primary way in which such information is shared. However, we anecdotally noticed a sharp rise in political material in the most prestigious scientific journals.

This led us to ask: are scientific journals becoming politicized? This pilot study examines that question.

Methods 

We take three of the biggest scientific journals and magazines - Nature, Science, and Scientific American - and look at the contents of their articles over a three-month period - March, April and May - in both 2002 and 2022. We note that Scientific American is not a scientific journal, but rather a magazine about science.

By highlighting articles which are on political issues we can count the number of such articles in both periods and then see if the number has increased between 2002 and 2022. This was done via inspection of the title, and then examination of the text if it was deemed likely political or if the title was ambiguous. 

Findings

The data we’ve collected shows a marked increase in politicization across major science publication venues. Whilst the judgment call of including an article as political or not could in some instances go one way or the other, we think the results are dramatic enough to mitigate such concerns. The raw data are provided below, showing every categorisation made.

In the preliminary phase of our study, we examined the presence of articles that could be deemed ‘political’, either in terms of covering a political issue in a partisan way, or advocating for a particular policy. As this was an exploratory analysis, this was not tightly defined, and based on the subjective opinion of a single analyst (the second phase of analysis had multiple analysts, see below).

In this analysis, relative to the same time period in 2002, the frequency of political articles in Science increased ten fold, Nature by six fold, and in the window chosen infinitely so in Scientific American due to a denominator of 0.

Figure 1 - Preliminary screen showing rising political coverage in science journals and magazines

In doing this early analysis, we found that the notion of what was ‘political’ coverage could be quite subjective, but also that it was clear there had been a sharp increase in politicisation. We also found that some articles, though political, were balanced and simply reportage.

This motivated us to dig deeper and do a more thorough study. We therefore repeated our analyses, this time dividing the articles between those which were politically contentious and those which were driven by identity politics, and excluded articles which appeared to be balanced political reportage. This means the figures below is based on different categorisations to the preliminary one above.

Political articles are defined as those which cover political subjects from a partisan perspective or make explicit policy suggestions that could be politically contentious. Identity politics are those featuring a political agenda based on inherent identity, such as race, sex, or sexuality; this includes discussion of equity and the attempt to explain all differences as being rooted in discrimination.

For example, an article reporting on the outcome of COP26 would not be included unless it made politically contested claims; an article saying that climate change can only be solved by socialism or free market incentives would constitute a politically-contested subject; and an article saying that climate change can only be solved through an anti-racist praxis would constitute identity politics.

Articles in this second phase of analysis were labeled by one of two analysts, and then each checked the others definitions. In cases where we could not reach agreement on which category an article should fall into or both analysts were unsure, the third analyst was called in to cast a deciding vote. Highly marginal calls where no one had a confident assessment were left uncategorised. 

As this project inherently relies on subjective judgment calls and there are necessarily many borderline cases, we will give some further examples of how different articles are categorized. We have also provided some comments in []’s to clarify why some articles were or were not categorized.

• Science editor Holden Thorp’s Science Editorial titled “It ain’t over ’til it’s over”espouses a debatable political opinion by complaining about Biden calling an end to COVID, but is not focussing on a particular identity group. This would therefore be counted as a ‘political article’, as would articles advocating a specific policy to address climate change, or supporting a particular political candidate.

• Contrastingly, the lead Editorial by Science editor Holden Thorp titled “Biden doesn’t get it” complains that Biden chose a science adviser with white skin rather than a person with black skin, and so this is deemed to cover an identity related issue. Broader ‘diversity, equity, and inclusion’ related articles would also be deemed as identity-focused. 

• Unlike the ‘political’ designation, an article does not need to take a particular stance to be categorized as covering an identity issue. Nor do all articles deemed identity mean it is taking a hard left position - for example, ‘Correspondence: Vaccinating women against COVID in world’s largest refugee camp’ or ‘Afghanistan’s girls’ schools can — and must — stay open. There is no alternative’ would not be deemed politically controversial, but cover identity and so we would count them as covering an identity issue. However, most articles designated ‘identity politics’ took what would be described as a leftist position usually in line with ‘structural racism’-style arguments, none took a more conservative position.

Articles dealing with a particular nation, such as sanctioning Israeli scientists, are dealt with as ‘political’ rather than ‘identity’.  

We note that Scientific American publishes less frequently (monthly) than Nature and Science (weekly), so it is included illustratively rather than to present definitive conclusions.

We found sharp increases both in politically contentious articles and articles covering identity related issues.

Identity-related articles increase

Figure 2 - Increase in the number of identity-related articles across several science publications

We note that the Science 2022 value has one issue with 10 identity related issues, driving the average up to 1.91 identity articles per issue. Even when that is removed however, it still averages 1.2 identity politics, a 526% increase compared to 2002.  We repeat that not all of these ‘identity politics’ articles would be deemed politically contentious, for example those advocating keeping girls schools open in Afghanistan would not be particularly contentious in the west.

Politically contentious articles increase

Figure 3 - Increase in the number of politically contentious articles across several science publications

Not only are political articles much more common, the focus is more explicitly political. Whereas articles in 2002 at least usually related politics to science, those in 2022 often focus on using science to drive political objectives.  For example, in Nature in 2002 several of the articles were dealing with whether or not sanction Israeli scientists were deemed political. They also tended to be more clustered in correspondence than pushed through editorials, though this was not absolute.

We also noticed that an increasing amount of articles that bordered on covering politics and covering policy solutions, such as ‘Broken bread — avert global wheat crisis caused by invasion of Ukraine’, but we tended to deem these non-political despite them advancing policy solutions as they appear to be a relatively bipartisan concern. Likewise, an article opposing the war in Ukraine, though it technically deals with politics, would not be deemed a debated position politically except by a small subset of Russians, and is therefore not counted as a ‘political’ article by this definition. 

Whilst many topics we have labeled as political would have very broad societal support, such as calls to end the war in Ukraine, there is a strong tendency toward an identity politics agenda. There is invariably no counter view offered in these articles. 

By contrast, in no cases could we find an article which overall advocated a right-wing point of view, and few of non-identity politics leftist issues. This, combined with the increasing prominence of politics in science journals, represents an overt politicization of the editorial stance of independent scientific journals and magazines.

Further work:

This is a preliminary finding, based on the manual analysis of the most prominent routes of sharing scientific information. Such an approach is labor intensive and reliant on individual decision making to categorize work. Whilst this is subjective, the results are pronounced enough to make us confident in our conclusions that there has been an increase in political coverage in scientific journals, and that this has a political bias in it.

Further work could address this question at scale with national language models, which would allow coverage across a wider range of time windows and publication venues.

Data

423KB ∙ PDF file

Download

Download