A leading research career is influenced by many factors: a researcher’s interests, yes, but also funding conditions, institutional expectations and the often-shifting priorities, policies and attitudes of wider society. Add into that mix the complexity and pressures of today’s interconnected global science system, and it’s clear that the research itself is just one aspect determining success — albeit by far the most important one.

Against this background of institutional, structural and cultural influences, what is stopping researchers from reaching greater heights? Where would they like to see changes that enable them to pursue the breakthroughs of tomorrow? And is there any divergence between their perceptions and the reality of current trends in research?

Research Leaders survey white paper

To examine some of these questions in more detail, Nature Research Intelligence, which manages the Nature Index database, surveyed more than 6,000 authors of articles published since 2020 in the high-quality natural- and health-science journals tracked by the index (see ‘Who was surveyed and how?’). This ‘Research Leaders’ survey — which invited responses from a representative sample of authors across regions and disciplines, and was analysed by demographic factors, such as gender and research experience — reveals exactly how some of the world’s most successful scientists view the research landscape at present. Further results are also available in a ‘white paper’ report on the Nature Index site.

One major focus of the survey was research impact, whether in the ‘academic’ sense of scientists reading and citing one another’s work — types of impact that have a long history of being tracked in global science — or in a wider societal and economic context. The results suggest that there is substantial variation in how measures of impact are viewed by leading researchers.

Respondents were asked, for instance, which three factors they thought contribute most strongly to a research paper achieving high academic impact. Across the survey, “novel, innovative and original findings” was cited most frequently (selected by 84% of Nature Index authors as one of their top three factors), followed by research being “rigorous and methodologically robust” (70%). Publishing in a high-impact journal was chosen by 39% of all survey participants; 29% cited the availability of open-access publishing and 27% pointed to research being interdisciplinary or intersectoral. Confirming or replicating previous findings was selected least often, with just 15% of all respondents identifying it as a contributor to high academic impact.

Breaking these results down by seniority, using the number of papers published by the respondents as a proxy for experience, reveals clear differences in how impact is understood (see ‘Fresh perspectives’). The more experience researchers had, the more likely they were to prioritize journal prestige and methodological rigour. Among those who had authored more than 100 papers since 2020, almost half selected publishing in a high-impact journal — something usually determined using an ‘impact factor’ or average citation rate for the journal — as a top-three factor, compared with just under one-third of respondents who had authored one to five papers. Similarly, three-quarters of those with the highest publishing rate highlighted methodological rigour, compared with two-thirds of the group with the least experience.

Junior researchers, by contrast, placed greater emphasis on open access. Among respondents with up to ten papers, around one-third selected the availability of open-access research as a key contributor to impact, compared with 21% of those with 51–100 papers and 20% of those with more than 100 publications.

Attitudes towards the importance of reproducibility and replication also varied with publishing experience. The most experienced respondents were least likely to select replicating previous findings as a contributor to impact; only 12% of them chose this factor, compared with 20% of those with the least experience.

Health psychologist Marcus Munafò, deputy vice-chancellor of the University of Bath, UK, says the pattern suggests a generational shift in how impact is understood. It seems that “using journal impact factor as a proxy for the quality of an article or as an indicator of prestige is becoming less relevant for early-career researchers”. He adds, however, that it would be valuable to repeat the survey to assess whether these differences persist as individual researchers progress in their careers. “If those early-career researchers become tomorrow’s senior researchers” and continue to hold such views, “that could lead to a real shift in culture that I believe would be quite positive”, Munafò adds.

He also notes that such perspectives on academic impact seem to align with the shifting priorities of UK funders. Internationally, he says, the picture is “a little bit more mixed”, but the overall direction of travel still seems to be towards more transparent and nuanced approaches to assessing research quality, particularly among early-career researchers. He points to the growing influence of networks and initiatives, such as the Declaration on Research Assessment and the Coalition for Advancing Research Assessment, which challenge the use of metrics such as the journal impact factor. He adds, however, that “people are at very different stages of the journey” in various parts of the world.

Although survey response numbers varied around the world (see ‘Who was surveyed and how?’), that unevenness is reflected in regional differences. Researchers in Asia, and in Africa and South America, were more likely to value publication in high-impact journals than were those in North America and Europe. Respondents in Europe, by contrast, were more likely to pick open access as a strong factor for impact than were those in Asia. Researchers in Asia were less likely to cite methodological rigour as a driver, compared with respondents in Europe, North America and Africa and South America.

Yves Gingras, scientific director of the Observatory of Science and Technology at the University of Quebec in Montreal, Canada, says these differences reflect structural inequalities in the global research system.

For instance, differences in attitudes towards open access can be shaped by resources, he says: the “policy push toward ‘open science’” is something that is “possible only for rich countries and scientists”, who can afford costs such as article-processing charges.

Chaoqun Ni, who studies research evaluation and bibliometrics at the University of Wisconsin–Madison, says the survey seems to capture “not only what researchers think should drive academic impact, but also what they perceive to be rewarded in their own institutional and regional contexts”.

In that sense, the regional differences are likely to “reflect variation in research evaluation systems, publishing infrastructures and career incentives”, she says.

“In settings where journal prestige remains a strong signal for hiring, promotion and recognition, it is not surprising that researchers place greater emphasis on publication in high-impact journals,” she adds.

She also cautions against interpreting differences in opinion on methodological rigour “as evidence that researchers in some regions value rigour less”.

“Rather, it may reflect a distinction between what researchers believe makes work scientifically strong and what they believe actually translates into academic impact under current reward structures,” she says. “That distinction is important.”

Views on impact also varied by gender. Women were more likely than men to prioritize open access (38% compared with 31%) and collaboration (32% compared with 30%).

Lynn Nygaard, a specialist in academic writing at the Peace Research Institute Oslo, says this aligns with patterns she has observed in her qualitative research. “I did notice that women were much more interested in being collegiate and being perceived as being collegial,” she says, adding that this might reflect wider social expectations.

Experience level also seems to be a key factor when researchers are thinking about leaving research entirely or moving institution. Although this might not be surprising, given that employee mobility is often more fluid earlier in a career, the data are stark: Nature Index authors with fewer publications overall were more than twice as likely as were highly published researchers to say that they intended to leave research or move institution in the next two to three years (see ‘Exit strategy’).

Among respondents with one to five publications since 2020, 22% said they were likely or very likely to leave research. This fell to 17% among those with 6–10 papers, 16% among those with 11–20 papers, 12% in those with 21–50 papers and 9% in those with 51 or more publications.

Nisheet Patel, previously an early-career researcher in machine learning at the University of Geneva, Switzerland, was one survey respondent who said they were “very likely” to leave research. He told Nature Index in a follow-up interview that he thought working in industry was more likely to deliver a better work–life balance and salary. Although he “loved” working in academia and said his PhD experience was positive, the outlook for the next five to seven years was much less appealing. The bureaucracy of the next stage — such as applying for grants and submitting papers — would “take most of my time away from what I actually wanted to do”. In 2024, he left his postdoc and later founded Axy, a start-up that builds artificial intelligence (AI) tools for scientific conferences and research collaboration. He also works as an AI scientist in the private sector in Switzerland.

In his ideal system, Patel says, a research career would be less constrained by existing structures, in which projects are often tied to grants. Instead, funding would flow more flexibly to allow researchers to follow their interests and collaborate easily across disciplines.

Mridula Bhalla, a postdoctoral researcher in neuroscience at the Institute for Basic Science in Daejeon, South Korea, agrees that difficulties in achieving a good work–life balance underpinned her survey response of “very likely” to leave academia. She describes the progression from postdoc to principal investigator to professor as “particularly demanding in Asia”.

In neuroscience, Bhalla says, the “publish or perish” culture — in which researchers feel they must publish research papers, and often in certain journals, to keep their career on track — is well established, and intensifying. “People who publish well, get received well; people who don’t publish well, don’t get received well, and it’s just a lot of pressure,” she says. “It kind of takes away the joy from science.”

Respondents who flagged their intention to move institutions followed a similar pattern. than half of the authors with one to five publications (55%) and of those with six to ten publications (52%) said they were likely or very likely to move in the next two to three years. This dropped to 44% among those with 11–20 publications, and to 31% among those with 21–50 publications.

Paul McCarthy, who runs League of Scholars, an academic data and recruitment firm based in Sydney, Australia, says that the results are unsurprising given the lack of job security that often exists for early-career researchers. However, he says that this was not necessarily always the case. In Australia, he points to a structural shift in the mid-1990s, when the number of PhD graduates began to outnumber available academic posts considerably — a gap that has since widened. In 2000, the number of PhD completions in Australia was less than 4,000 per year; by 2023, it had reached 10,000. At the same time, the number of academic positions in the country shrank from 54,086 in 2016 to 46,971 in 2021.

Although the data are from Australia, the pattern reflects trends seen across much of the developed world, where PhD students, postdocs and early-career researchers now greatly outnumber tenured academics, says McCarthy, and are often on insecure, short-term contracts.

“Academia is one of the world’s longest apprenticeships and yet there’s no guarantee of a job at the end of it,” he says.

Gail Kinman, a researcher in occupational-health psychology at Birkbeck, University of London, agrees that a combination of job insecurity and limited prospects for career progression is likely to underpin the survey’s findings.

“Highly published, more-senior researchers are also more likely to be part of more-established national and international research networks — this provides access to collaborative opportunities, shared resources and knowledge exchange,” she says. “This is a key issue for job security, given the drastic reductions in research funding in many countries. As funding becomes increasingly competitive, more-established researchers are in a better position to secure grants and institution support.”

A higher proportion of women in early-career roles might partly explain the differences between male and female responses to the survey question about the likelihood of leaving research; 14% of male and 17% of female participants said that they were likely to do so in the next two to three years.

Regional differences were also evident. In the survey, 18% of respondents in both Europe and North America said they were likely or very likely to leave research entirely in the next two to three years. By contrast, only 7% of respondents in Africa and South America and 10% in Asia and Oceania said the same.

McCarthy notes that opportunities outside academia vary widely by country. Moving into industry is more common in the United States, South Korea and Switzerland, he says, which might reflect both frustration with academic systems and the availability of attractive alternatives.

For Evelyn Gitau, chief scientific officer at the Science for Africa Foundation — a non-profit organization based in Nairobi, Kenya, that promotes and supports science on the African continent — the fact that researchers in low- and middle-income countries were less likely to say they intended to leave research was not surprising, because the alternatives are limited. “My PhD instructor would say as long as you have a PhD in Africa, you’re guaranteed a job,” she says. Unfortunately, an academic position won’t necessarily mean you’re going to be paid well, “but you’re definitely guaranteed a job”.

When asked about the freedom they have to decide which research to pursue, survey respondents reported mixed levels of autonomy. Overall, 46% said they had strong or complete autonomy from their institution, whereas 18% said their research direction was set strongly or completely by their institution. A further 22% described their autonomy as moderate.

Kinman, whose research examines job insecurity, workload and mental health in academia, cautions that because this is the first year of the survey, it is not yet clear whether the findings represent a major shift or are broadly consistent with previous patterns. However, she notes that evidence suggests research autonomy might have declined over time, something driven by “a shift towards financial justification, accountability and the growing prioritization of measurable, high-impact research outputs”.

“This may have reduced the freedom to pursue longer-term or more speculative research,” she says.

Perspectives on autonomy also varied by career stage. Those with more publishing experience were less likely to say their research direction was strongly influenced by institutions or funders. Among respondents with one to five publications, 27% said their research was determined completely or strongly by their institution, compared with 15% of those with more than 100 publications.

Kinman says that this pattern is likely to reflect the greater power and autonomy associated with seniority and professional esteem. She adds that ‘survivor effects’ could also be at play, with researchers who are less successful or less able to secure autonomy being more likely to have left research altogether. Those who remain are, disproportionately, individuals who have been able to obtain stable positions, accumulate resources and negotiate high levels of control over their work.

For early-career researchers, autonomy often seems to depend heavily on supervisory relationships. Early-career researchers who responded to the survey told Nature Index they felt “very lucky” to have supervisors who allowed them to pursue their own research priorities, although they noted that peers in other laboratories often experienced much more restrictive conditions. “Autonomy is important because that’s how we will grow as young postdocs and young scientists,” says one early-career researcher.

Influence from funders was, on average, stronger than that from institutions. Overall, 30% of respondents said their research direction was shaped strongly or completely by funder priorities, compared with an equal share who said it was not at all or only slightly influenced. A further 25% reported moderate influence.

Again, early-career researchers reported less autonomy. Among those with one to five publications, 37% said their research was determined completely or strongly by funders, compared with 30% of respondents with more than 100 publications.

Small differences by gender were also evident. Men were slightly more likely than women to report having complete or a large degree of autonomy from both their institutions (55% compared with 50%) and their funders (37% compared with 34%).