Many people think that “ivory tower” intellectuals make little difference in the world. But some of the highest impact people in history have been researchers, and if you have good personal fit with academic research, we think it can be one of the highest-impact paths.
Alan Turing was an academic – a mathematician, the ultimate ‘ivory tower discipline’ – yet he developed code breaking machines that allowed the Allies to be far more effective against Nazi U-boats in WW2. Some historians estimate this enabled D-Day to happen a year earlier than it would have otherwise.1 Since WW2 in Europe resulted in 5 million deaths per year, Turing’s work contributed substantially to saving about 5 million lives.
And, at the same time, he developed the fundamentals of modern computing.
In this article, we’ll cover why we think a career in academia has the potential to be very high impact in the right circumstances, how to figure out whether this option is for you, and how to maximise the impact you can have as an academic.
To work on some important fields of research that are hard to fund commercially, academia is the natural place to go. A single outstanding researcher can move a field forward and make a significant contribution to solving key global problems. Beyond research, academics also have other avenues for impact, such as by influencing government policy, the priorities within their field, and the culture of society at large.
On the other hand, entering academia is a lengthy process, involving 5-15 years of study and post-docs, or even more. Most people who make the attempt – already a talented bunch – are rejected or give up getting a permanent position. While it has many positives, it’s common for people to continue with academic research by default when they’d be better suited elsewhere, so we encourage you to be self-skeptical and consider other options before starting a PhD.
Raw intelligence, hard work and curiosity are the key indicators you may be able to beat the odds. Salaries are lower than elsewhere, though job satisfaction nevertheless seems high.
Choosing the right research questions is essential to maximise your expected social impact and we offer advice on how to do that, but PhD students often face a trade-off between working on the questions they think are most useful, and those which will best advance their career.
• If you're a good fit, there's a large potential for social impact by working on pressing research questions
• Prestigious platform for advocacy
• Autonomy and job satisfaction
• High competition for a limited number of positions
• Lengthy training time
• You often have to teach, and academic incentives remain imperfectly aligned with having social impact
Key facts on fit
High intelligence, conscientiousness and need for intellectual stimulation. Ability to work independently. Deep interest in the area and willingness to focus on particular questions for long periods.
Explore different fields to find one you’re interested in which can be used to work on important problems. Get very good grades at the undergraduate level. Look for PhD submission dates and search for an academic supervisor working on valuable questions at a prestigious university. Work hard to publish in good journals during your PhD or during postdocs.
If you are well suited to this career, it may be the best way for you to have a social impact.
This profile is based on 10 interviews with successful academics (including in statistics, economics, computer science and biomedicine), as well as numerous more informal conversations with people pursuing academic careers (including some of our colleagues and trustees who have an academic background), and looking into the careers of some of history’s most impactful academics. We’ve also looked for other sources of advice on maximising your research, as well as research into what makes a successful academic career, statistics of progression within academia, and the distribution of research impact.
As a result, we think that research seems like a high-impact career choice in general.
But why go into academic research in particular?
Academia isn’t the only place where it’s possible to do research on important problems. Instead, you could do research in the private sector – as a biomedical researcher you could also work for a pharmaceutical company, or as a computer scientist you could do research at a big technology company like Google or Amazon, for example. As a social scientist, you could work for a think tank or as a researcher for government.
Academia has some advantages over research in industry or the public sector, as well as some disadvantages, as we’ll discuss below. It’s certainly not true that academia is always better than these alternatives – it will depend on your personal situation, opportunities, and priorities. However, we do think that in many cases it’s one of the best environments for doing valuable research.
One downside of academia is that it can be slow-moving: academics often report frustrations with the efficiency and speed of university administration, especially relative to the private sector. A second is that academics may have to spend a lot of time applying for grants to fund their research. A third commonly recognised challenge is that academics often have many other responsibilities beyond their research, such as teaching, which can take up a lot of time. A typical teaching load during term-time is 10-12 hours a week – though in reality it’s probably substantially more, as this doesn’t account for time spent preparing, marking, and supervising students. This seriously limits the amount of time you have to spend doing your own research.
However, it’s not clear the situation is necessarily any better in other research jobs. In industry or the public sector, researchers are likely to have more meetings and more short-term deadlines to meet, giving them less time to really explore questions deeply. Academia also gives you more time to focus on learning and research early in your career than any other job will. Graduate school gives you ~4-7 years (depending on your course and country) where you’re able to focus almost entirely on studying and learning to do research. It’s pretty impossible to get this in any other job, even if it’s research-focused – as there’s much more day-to-day pressure to deliver.
More generally, academia provides a huge amount of freedom to research what you choose. Nick Feamster, a computer science professor at Princeton who has worked at both universities and large industry labs, says of why he ultimately chose the academic path:
As a professor, you do not have quarterly deadlines to meet, monthly reports to file, or a boss you are regularly accountable for…. Freedom from the constraints of short-term deadlines and having to answer to others who are setting the agenda really presents the possibility of thinking about the “right” solution to a problem, rather than hacking together something that just works well enough to get the job done. In my opinion, this aspect of being a professor presents the greatest opportunity for adding value.
Academia is also the best place to do research that is unlikely to get commercial funding because the benefits are long-term or abstract. Academics often have a lot more freedom to pursue whatever research they think is most valuable (especially at later stages of an academic career) than you would get doing research in a commercial setting. “In most companies, research topics are largely chosen by the business and marketing departments… pressures to show immediate and positive results can also challenge the best ethical and professional practices”, says Jaime Teevan, a scientist at Microsoft Research.
The flipside of this is that it can be harder to get academic research applied in “the real world.” Explaining why he chose to leave Harvard and join Google, computer scientist Matt Welsh writes “Academics have a lot of freedom, but this comes at the cost of high overhead and a longer path from idea to application.” And Nick Feamster agrees: “I find that I spend a lot of time ‘evangelizing’ my ideas to industry to try to transfer them into practice… industrial researchers do not have this problem.” However, just because the impact of academic research is often less immediate or obvious than the impact of private or public sector research, doesn’t mean it’s necessarily less valuable.
Finally, academia attracts some of the world’s brightest minds, so you can find great collaborators to work with and learn from.2 What’s more, in many fields there’s a perception that the most serious researchers work in academia, and this means being in academia is an important signal of ability, putting you in a better position to find collaborators.
Finally, as we’ll cover in the next section, having academic credentials also opens up lots of other ways to have an impact.
Research isn’t the only way academics can have a large impact
When we think of academic careers, research is what first comes to mind, but academics have many other pathways to impact which are less often considered. Academics can also influence public opinion, advise policy-makers, or manage teams of other researchers to help them be more productive.
If any of these routes might turn out to be a good fit for you, then it makes the path even more attractive. We’ll sketch out some of these other paths:
1. Public outreach
Peter Singer’s career began in an ordinary enough way for a promising young academic, studying philosophy at Oxford University. But he soon started moving in a different direction from his peers, by seriously trying to change the views and behaviour of the general public on important moral issues.
Singer’s first book, Animal Liberation, is one of the most widely-read books published by any philosopher. Many consider the book to have given birth to the animal liberation movement.
Singer has also argued strongly that we have a moral obligation to donate much more to those living in poverty. His writings on global poverty inspired the creation of the organisation Giving What We Can, whose 3,000 members have donated over $10 million to the most cost-effective charities, and pledged over $1 billion of future donations.3
Early-career economist Max Roser made a website clearly presenting data on how the world is changing that now gets over a million visits a month and earned him introductions to politicians and philanthropists, as well as a large Twitter following.
Other academics venture far outside of their original fields of research. Noam Chomsky managed to convert an academic position focussed on linguistics into being one of the most prominent commentators on US foreign policy.4
As we argue in the career guide, advocacy seems like a promising approach in general to have a large social impact. And academia seems to be an especially good position from which to advocate for important ideas. Successful academics have developed the expertise required to understand complex questions, and the credibility to have people listen to what they say.
This is especially true if you want to advocate for issues concerning extinction risks, because some of them are emerging from scientific and technological progress itself. We need experts in the relevant fields to shape public understanding — to prevent the risks and benefits from being overhyped, while helping people to understand genuine concerns with, and the real potential of, these technologies.
However, it’s important to be careful with public outreach – it wouldn’t be that difficult to do harm by spreading slightly the wrong idea, in slightly the wrong way, to slightly the wrong audience. For example, writing popular articles about threats from artificial intelligence might lead to widespread fear and alarm, which could be counterproductive.
This suggests that getting the biggest possible platform for your ideas might not be the best route to impactful outreach – sometimes targeting communication at smaller, more influential groups might be a better approach.
For instance, academic macroeconomists have a great deal to say about appropriate use of monetary or fiscal policy, and are regularly appointed to decision-making roles in central banks. But communicating their ideas to the general public is difficult and likely to result in misunderstanding. As a result macroeconomists more often write books and papers aimed at a limited audience with the goal of influencing politicians or bureaucrats within government or central banks themselves.
2. Apply your research to problems outside of academia
John Beddington was a successful academic biologist, but his biggest impact seems to have come through his service as a respected science advisor to the government.
Beddington was the UK Government’s Chief Scientific Adviser from 2008 to 2013, and played a key role in helping the government to navigate the challenges of the Fukushima nuclear disaster, eruptions of the Icelandic volcanoes, and ash dieback disease in the UK. He also pushed to establish a network of “Chief Scientists” across all government departments.
Often the people in charge of making important decisions — about how to allocate funding, or what the appropriate policy response to a given challenge is — won’t necessarily have expertise in the most relevant fields. This means they need to call on external experts who do have this specialist knowledge, who can help them to make the best possible decisions.
These experts are very often academics, since academia is one of the best ways for people to develop the credentials needed to be taken seriously by policymakers.
And a need for guidance from academics isn’t limited to government – you could do a lot of good if you have the relevant expertise to advise NGOs like the Gates Foundation or World Bank, or tech companies that have a large influence like Google or Amazon.
Policymakers in many countries often seek the advice of experts in some relevant fields such as economics, data science and machine learning, and some areas of social science such as behavioural science. That demand means it’s not necessarily as hard to end up in a position of advising decision-makers as you might think.
For example, Prof Bruce Chapman was an academic economist at the Australian National University who consulted for the Australian Minister for Employment, Education and Training in 1987. While there he helped to devise a system of government-backed ‘income contingent’ student loans which enabled students to pay for more of the costs of higher education without taking on any significant financial risk. He had a large influence over the design of the policy, while the public advocacy was largely left to others. This approach to student loans was later copied by the UK, Thailand and Ethiopia, among others.5
If you want to be Chief Scientist of a government department you’ll need to be a pretty senior and established researcher, but we know several people who have advised government early in their academic careers, even during a PhD, because they have valuable expertise and have been proactive about building connections and looking for opportunities.
Some academics also notice that insights from their field could be applied in the world more concretely; they might be used to build new products or companies that help solve important problems.
This could lead to them leaving academia to do these projects, but you might also be able to do this kind of work while continuing to do research. For example, leading AI researcher Yoshua Bengio cofounded Element AI with an established entrepreneur, in order to apply cutting-edge AI research to tackling important business challenges. Google was founded by researchers at Stanford trying to improve online search.
Other academics have gone on to found valuable non-profits based on their research. For example, GiveWell recommended charity the Schistosomiasis Control Initiative emerged from a group of academics at Imperial College London, such as expert in tropical parasitology Prof Alan Fenwick, who coined the term ‘neglected tropical diseases’ to draw attention to the problem. Similarly the Centre for Pesticide Suicide Prevention emerged from academic researchers at the University of Edinburgh. You can learn about that history in our interview with one of its founders, Dr Leah Utyasheva. The credibility they gain from their work in academia appears very helpful in getting these projects off the ground.
Anyone with expertise can set up these kinds of non-profits – you don’t necessarily need to be an academic – but we think academia is a particularly good place to develop expertise for many people, as you get to work alongside some of the most knowledgeable people in an area, free from commercial pressure.
3. Influencing academia and research management
When Nick Bostrom was at university, he was expelled from the psychology department for trying to study too many subjects: he was taking classes from anthropology and literature all the way to science and psychology.
Bostrom’s highly interdisciplinary approach, and his tendency to take a step back and look at the “bigger picture”, has led him to ask questions almost no-one before him had, such as whether we might face a genuine threat of extinction in the next century, or how technology might lead to drastic changes in the human condition.
Ten years ago, he founded the Future of Humanity Institute at Oxford, an academic research department which he still runs full-time. Rather than just conduct his own research, he’s recruited and managed other researchers working on important topics, who would have struggled to find a place to do this work otherwise, or might not even have considered doing so.
Other academics have had a huge impact on a field by suggesting improvements to the methods used in a field. Psychologist Brian Nosek and statistician Andrew Gelman are two examples of academics who have led the push for more rigorous methods and openness in social science, which could vastly improve the quality of research in future.
The recent “replication crisis” in social science has found that many of the most widely accepted findings do not hold up under scrutiny – in many cases, repeating the original experiments does not yield the same result. This suggests that a huge amount of resources have been wasted on research using poor methods – fixing this could massively boost the reliability of all future work in the entire field.
Another way to have a large impact as an academic, then, is to influence the direction of other researchers’ work, and the trajectory of a field as a whole – maybe even establishing entirely new fields.
Influencing the direction of others’ research often follows from having amazing ideas yourself. But it’s also possible to increase the output of other researchers or a field in more subtle, less glamorous ways, if you’re more entrepreneurial.
We’ve written before about how jobs in research management can be incredibly valuable: if you can help a whole team of researchers to be a little more productive by doing the neglected work of management, then you can have more impact than you would alone. We recently made the case in more detail in our article on operations management.
Projects like keeping track of and prioritising research within a department, fundraising, managing researchers to enable them to be as productive as possible, and recruiting new, talented researchers, are really important but often neglected within academia.
The skill set needed to do this kind of work effectively is also relatively rare: you need to have a thorough understanding of the relevant research, but also administrative and management skills. Many great researchers are uninterested in management and do as little of it as they can, so if you have this rare combination of skills and a willingness to focus more on the management side, you could be very valuable to a department.
Research management jobs can be part of an academic career – working as head of a department, for example – but sometimes they’re a separate job, such as being a project manager in a research group. This latter case could also be a good option if you decide that research isn’t quite the right path for you.
A related path is academic journal editing. Journal editors determine which articles are published, and therefore which are considered credible and newsworthy. Editors therefore shape the public conversation both inside and outside of academia — e.g. by determining which studies science journalists write about, and to some extent, how the findings are presented. Editors are often senior academics, but some journals are run by people for whom editing is their main career. For example, Editor-in-Chief Magdalena Skipper of the journal Nature went full-time into an academic editorial career after her postdoctoral work.
For many of these kinds of roles you need a PhD, and even where you don’t it helps to have experience in the kind of research you’re working with. But you also don’t need to be a leading professor. For example, after finishing his PhD in genomics, Seán Ó hÉigeartaigh decided he could likely have a much greater impact doing project management at the Future of Humanity Institute, which he believed was doing really valuable work.
Seán is now executive director at the Centre for the Study of Existential Risk in Cambridge, an organisation which almost certainly wouldn’t exist without him.
The career paths we have discussed above are competitive, and success is not guaranteed. It’s natural to doubt whether you’ll be able to do or enable high-impact research, be in a position to advise policymakers, and so on. This leads us to:
If you’re in doubt, it’s often worth staying in academia
Keeping options open
If you haven’t yet done a PhD and are unsure whether an academic career is right for you, then it may be worth spending a year or two exploring other options. It’s not unusual for people to start a PhD in their mid-to-late twenties, and this shouldn’t hinder your career prospects much.
But if you already have a PhD and are uncertain about whether to continue in academia, it’s probably better to lean towards staying (especially if you don’t have clear alternative options.)
Spend a couple of years focused on building up a good publication record and applying for academic jobs, and then reassess based on your success and what other options are available to you.
Almost everyone in academia advises against taking a break once you’ve done your PhD, because it’s generally harder to enter or re-enter academia later. Postdoc and permanent positions are highly competitive and will go to those with the best track record and strongest publications. If you take a career break, your publication record may not be strong enough to get a good position, and it gives the impression you’re not serious about pursuing research. It’s also generally easier to transition from academia to other industries than the reverse, as many jobs look positively upon academic experience, but academia tends to place little value on non-academic experience.
Good backup options
Moreover, even if you leave academia later, you’ll probably have good back-up options, especially if you’ve chosen certain fields and leave before your early 30s. The table below shows some good backup options if you’re coming from different academic backgrounds:
Quantitative fields e.g. maths, statistics, computer science, economics, physics
Quantitative skills are in-demand in most career areas, but some particular quantitative areas include: - Effective non-profits - Think tanks and government research - Quantitative asset management - Tech startups - Data science - Software engineering
Social sciences e.g. economics, political science, sociology, public policy
- Government - Think tanks - Foundation grantmaker
Lab sciences e.g. biology, chemistry
- Scientific research in industry e.g. drug development - Tech startups with a specialist focus e.g. biotech - Public health research for NGOs/foundations - Consulting
Academic research can be very satisfying if you’re a good fit and find the right position: there’s probably no other job that gives you the same degree of autonomy to do intellectually satisfying work, and you’ll be surrounded by other smart and curious people.
However, the amount of freedom you have to research what you think is most important will vary depending on the specific field and department, and is likely to be lower earlier in your career, when there’s more pressure to work on whatever is most likely to get you published.
One study on job satisfaction in academia suggests that satisfaction is higher among full professors than more junior faculty members, which is consistent with this.7
One downside of academia to be aware of is that sometimes it can be quite solitary. Especially during a PhD, if you’re not working on projects with other people regularly, you’ll be working on your own quite a lot. However, this varies a lot by field – in the humanities and some social sciences highly collaborative research groups are less common, whereas lab sciences are generally a lot more collaborative.
This means you’ll need to be happy working fairly independently, and be able to manage and motivate yourself to a large degree. However, you’ll at least be based in a university with lots of smart and interesting young people – so even if the work is solitary, there are likely to be good social opportunities.
Overall, having a large degree of autonomy seems to be a good thing – at least for people who enjoy research, and want the freedom to pursue their own ideas – and so is another point in favour of academic careers.
Having a very engaged and supportive supervisor (so not necessarily the most busy and senior academic in the field!) can make a big difference, as can talking to other PhD students and postdocs regularly to share ideas. (For AI research specifically, our job board curates a list of potential PhD supervisors working on AI alignment.)
You’ll also want to make sure you’re working on questions you feel very intrinsically motivated to answer, since you’ll have to drive yourself to keep making progress with much less external pressure than most jobs.
What are the main downsides?
Highly competitive and low chances of progression
As we mentioned earlier, academia is highly competitive:
A study by the U.S. Bureau of Labor Statistics found that in 2010, less than 15% of new PhDs in science, engineering, and health-related fields found tenure-track positions within 3 years after graduation. For PhDs in the life sciences, the figure was a grim 7.6%.
The NSF estimated that, in 2010, only 11% of PhDs in the biological sciences held tenure-track positions 3 to 5 years after graduation, down from 55% in 1973.8
One way to measure how competitive an academic field is to look at its “reproduction rate”: what is the mean number of new PhD students a typical faculty member will graduate during his or her career? This gives us an estimate of how many PhD graduates there are per faculty position. One study found that there is roughly one tenure‐track position in the US for every 6.3 PhD graduates in the biomedical sciences.9
Another looking at engineering finds that “a professor in the US graduates 7.8 new PhDs during his/her whole career on average … This implies that [barring growth in academic positions], only 12.8% of PhD graduates can attain academic positions.”10
The situation seems similar in other countries. For instance, in the UK, for example, only 3.5% of people with a science PhD make it to permanent research positions in academia and just 0.45% of STEM PhD holders in the UK become tenured professors (though note that the title “professor” is awarded much less often in the UK as in the US, as it is reserved for senior academic staff). Only around 20% end up in any sort of research roles.11
However, these figures vary a lot by institution – a 2015 study of 19,000 faculty members in business, computer science and history found that 25% of institutions produced 71-86% of all tenure-track faculty depending on the field.12 This means that if you’re able to do your PhD at one of the most elite universities, your chances of getting tenure will be substantially higher than 15%. (If we assume that all these universities produce the same number of PhD graduates, and that 15% on average get academic roles, then around 47% of graduates from the top 25% of universities would be successful but only 4% from the remaining universities. If the top 25% of institutions graduate twice as many students as the rest, then the figures move to 29% and 5%. This suggests your chances are 2-3x higher than the overall average at top institutions.)
Over the last 20 years, it has also become increasingly common to do one or more “postdocs” – temporary non-tenure-track research positions, normally lasting 1-3 years each – before getting a faculty appointment. According to the National Research Council’s report, “Bridges to Independence”, the share of recent PhDs in postdoc positions rose from 13 to 34 percent between 1972 and 2003.13 Scientists doing postdocs in the US spend an average of 3 years in this holding pattern and only about 17% ultimately land tenure‐track positions.14 A typical postdoctoral research associate salary is $45-55,000.
This seems to be a result of the fact that the number of PhD graduates has dramatically increased – in 1994, 7,800 people received doctorates in the life sciences in the US, whereas by 2014 there were 11,335 – while the number of tenure track and tenured professorship positions has stayed constant.
All of this comes after a large fraction of people who start PhDs fail to complete them. Between 41 and 78% of people who start PhDs have finished them after ten years, depending on their discipline. Computer and Information Science has among the lowest completion rates at (41%), Economics is in the middle (52%) and life sciences fairly high (63%).15 Though note completion rates are better among the most prestigious institutions.
Ten year completion rate (%)
Electrical and Electronics Engineering
Genetics, Molecular Genetics
Microbiology and Immunology
Molecular and Cellular Biology
Mathematics & Physical Sciences
Computer and Information Sciences
Physics and Astronomy
Anthropology and Archaeology
English Language and Literature
Foreign Languages and Literatures
Competitiveness also means that you might not have that much flexibility to decide where you live and work, or even what research you do. It’s common to have to move to a different university at each different stage of your academic career, which might not suit some people, especially if they need to bring their families with them.
Academic incentives aren’t always aligned with doing valuable research
One of the biggest problems in academia is the “publish or perish” mindset. As we showed in the section above, academia is extremely competitive. Progression is mainly decided based on your publications, which means there are powerful incentives to publish many papers, to do so in the most prestigious journals, and to be as highly cited as possible.
These incentives make it hard to do the most valuable research within academia. The research that gets you lots of publications isn’t necessarily the research that’s most valuable. The research that’s most likely to get published often depends on what’s popular in the field at the time, or research that suggests novel, exciting results.
In psychology, for example, the pressure to publish exciting and novel results appears to have led to a slip in the standards of research methods, which has now become clear as many of the most popular findings fail replication.16 If you need to publish regularly, this also pushes towards producing lots of small, incremental findings, rather than deeper work that might result in more valuable breakthroughs.
If you think that the most valuable research questions are far higher impact than average, then having to work on other questions might significantly reduce the impact of this path.
However, we still think academia has a lot of advantages for doing high-impact research – as we discussed earlier, you’ll generally have more freedom than doing research for a company, and academia attracts some of the smartest people.
How easy it is to do valuable research can vary a lot by the field and lab/research group you’re working with. One way to make this easier is to explicitly try and work with academics and groups who you think have a track record of working on important topics in the past, or who are clearly in the middle of trying to solve an important problem. If you can go and work in an AI group that has been producing cutting-edge, well-respected research over the last few years, a biomedical research lab who have a track record of producing effective disease treatments, or a public policy group who regularly engage with and advise government, for example, the incentives you face on a day-to-day basis are much more likely to align with doing valuable work. Some positions might carry fewer teaching responsibilities, and if you can get to a tenure-track position relatively quickly, you’ll have much more time and autonomy.
That said, if you can’t find a PhD opportunity with a research group working on an important problem, you may want to reconsider. We discuss this issue under How to establish your career early on.
Academia has lower salaries than industry in general – for example, being an academic in a quantitative field is much less well paid than being a data scientist or programmer which requires similar skills. A Nature Jobs salary and career survey found “a big disparity in industry versus academic salaries”, with average industry salaries exceeding average academic salaries by 50% in Asia and by 40% in Europe and North America (though note that many academics get far more annual leave and flexibility than those in corporate jobs).17
These averages vary significantly by academic field – with disciplines such as computer science, economics, and law paying higher salaries than average. Professors in those fields can earn as much as you might by your 30s in software engineering or even some parts of finance.18 However, people capable of becoming computer science professors can often earn far more than average in industry, and rise up the ranks more quickly. All considered, most academics have the impression that they could have earned significantly more in the commercial sector.
Long training time
Training for an academic career takes a long time – at least 3 years for a PhD, and closer to 5-7 years in the US (though you generally get much more training in the US, which prepares you better for the academic job market.) Because of the competitive nature of academia we discussed earlier, getting from a PhD to an academic position can also take a long time – in biomedical research, for example, the age of first independent faculty appointments has risen from 34 in 1979 to 38 in 2003.19
Since these are some of your most potentially productive years, the opportunity cost of doing a PhD and then postdocs is pretty high if you don’t end up following an academic career.
Furthermore, many of the problems we recommend focusing on are urgent, and work on them today is more useful than work on them in the future. Delaying your directly valuable work by 7 years, is itself a cost.
For this reason we generally recommend people think thoroughly about whether to do a PhD before committing to one, and ideally spend some time exploring other options first, rather than just doing one as a default. We cover how to do that later.
Establishing yourself as a specialist in a specific academic discipline may make it harder for you to be flexible and change the cause area you work on later in your career, though this varies from field to field. Mathematicians and social scientists, for example, often have lots of flexibility to apply their expertise to different problems, whereas biologists are often more focused on a narrow area.
What’s more, since relevant expertise is so important in many of the top problem areas, we think it’s often worth giving up this flexibility in order to gain specialist expertise if you pick a promising field. This is especially true if you’re coordinating with a community – if everyone specialises in something, the community learns more and some of those bets are sure to pay off.
How to assess your personal fit
As we’ve discussed, research seems to be an area where the most successful people have far more impact than the rest. This means that personal fit is an even more important factor than normal in deciding whether to enter this path.
What are the entry requirements?
To get started down this path, you’ll need:
A strong undergraduate degree (at least a 2.1 in the UK, though most people recommend a first, or a GPA of 3.5 or higher) in a relevant field.
Many PhD programs require a Master’s degree first (especially in Europe), depending on the field and your background.
The best predictor of success in academia, as in many fields, will be your existing track record. If you have managed to publish respectable papers before or during your PhD, especially as first author, that is a great indicator that you could succeed in academia. Similarly, your ranking in your classes or the competitiveness of the graduate school you got into are leading indicators.
Within biomedical research, researchers built a statistical model that predicts chances of becoming a principal investigator based on factors including publication record, gender, and university rank.
Below we will focus on other factors whose predictive value is not so obvious, and which don’t require you to already have pursued a PhD.
How important is intelligence?
In general, high IQ seems to provide a significant advantage in doing scientific research. A study of 64 eminent scientists (physicists, biologists, and social psychologists) by Harvard psychologist Anne Roe found that their median scores on tests of verbal, spatial and mathematical reasoning corresponded to IQ scores well above the median IQ of PhD scientists (though some have contested this, as we discuss below).20 If IQ were irrelevant beyond a threshold, we’d expect this group of successful scientists to have average scores similar to the average population of scientists which are already very high. Another line of support for this comes from the fact that intelligence is correlated with job performance more generally21, and the correlation is stronger for more complex jobs22. Since research is among the most complex careers, this suggests intelligence will be strongly predictive of success in research.23
Beyond this, more specific abilities – in verbal, quantitative, and spatial reasoning – seem to be important predictors of which fields a person is most likely to succeed in. A more recent longitudinal study of “mathematically precocious youth” over 35 years found that ability level (as measured by SAT scores aged 13) contributed significantly to academic accomplishments (securing a doctorate, tenure-track position, patents or noteworthy publications), but that ability tilt (the difference between math and verbal SAT scores) was highly predictive of the kind of domain these achievements occurred in. Subjects who reached high levels of achievement in the humanities were more likely to score high on the verbal SAT relative to the math SAT, and the reverse for those whose achievements were in the sciences.24
Results from the same 35-year study of talented youth found that spatial reasoning ability (the ability to match objects seen from different perspectives, judge what cross-section will result when an object is cut in different ways, etc.) is predictive of academic success in addition to verbal & quantitative reasoning. A 2013 analysis found that verbal & quantitative reasoning jointly accounted for about 11% of the variance in the number of patents & peer-reviewed publications a subject had, and that spatial ability accounted for an additional 7.6%.25 David Lubinski, one of the study’s co-directors, suggests that spatial reasoning “may be the largest known untapped source of human potential… no admissions directors I know of are looking at this, and it’s generally overlooked in school-based assessments.”26
However, this doesn’t mean that you need to have an IQ or test scores in the top 0.01% in order to have a chance of contributing valuable research. In A Question of Intelligence, Daniel Seligman reports that the correlation between IQ and elementary school grades is 0.65.27 This is a high correlation, but far from perfect – meaning how hard you work and other personality factors are also likely to be important, as we’ll discuss below. Seligman points out that if becoming a tenured professor is a one in a thousand level accomplishment, then we’d expect the average tenured professor to have an IQ of around 150 if the correlation between IQ and academic success were perfect. But if the correlation is 0.65, then we should expect the average tenured professor to have an IQ around 133, with quite a bit of variability around that.
There’s also reasonable variability by field – high intelligence, and high verbal/quantitative/spatial reasoning ability will be more important in some areas than others. Here are the average GRE scores of applicants to the following PhD courses:
Computer Science (1862)
Chemical Engin. (1845)
Material Science (1840)
Electrical Engin. (1821)
Mechanical Engin. (1814)
Earth Sciences (1761)
Industrial Engin. (1745)
Civil Engin. (1744)
English Lang. / Lit. (1702)
Religion / Theology (1701)
Political Science (1697)
Art History (1681)
Anthro. / Archaeol. (1675)
Public Administrat. (1460)
We worry about the reliability of this data, which is purportedly from 2002, and would like to find a better source, but so far it is the only one we have found.
We can also get a sense of how important IQ is in a field by looking at the age of peak performance in that field. Since IQ declines sharply with age, fields where researchers make their biggest breakthroughs early in their careers are likely to rely more on intelligence — in physics and pure mathematics, the age of peak output is around 30, for example, suggesting intelligence is highly important for contributions in these fields. In medicine and history, by contrast, the age of peak output is closer to 50 — suggesting that accumulated knowledge and effort play a much larger role in making contributions to these fields. Psychology falls somewhere in between.28
So while intelligence is important, and will certainly increase your chances of being able to contribute valuable research, it varies by field, and you don’t have to be a genius to contribute. While IQ does predict academic success on average, the correlations are weak enough that there are lots of exceptions, and measures of IQ aren’t perfect.29 This means you shouldn’t necessarily give up on research if you don’t get an off-the-charts IQ score, especially if you have other strengths or can find important but neglected areas of research.
How much difference does hard work make?
K. Anders Ericsson, a leading researcher studying expert performance, argues against the view that some people are “naturals” in an area, able to attain mastery with ease. His research suggests that the highest-performing people have all done a huge amount of focused practice, usually with top mentors: “even children considered to have innate gifts need to attain their superior performance gradually, by engaging in extended amounts of designed deliberate practice over many years.”30
This suggests that success in academic research may depend to a large degree on your ability to work in focused ways over long periods of time, with good feedback — which may in turn depend on your interest in and motivation to sustain work in an area, and your ability to find others to learn from.
However, the importance of more innate factors like intelligence/talent versus deliberate practice is still debated. One meta-analysis of studies of deliberate practice challenges whether Ericsson’s findings – which mostly look at performance in clearly-defined, predictable areas such as games and music – generalise to less-predictable domains such as science and education (where presumably it’s harder to get the high-quality feedback needed for fast improvement.)31
More generally, Ericsson’s research strongly suggests that world class performance in a domain requires 10 to 30 years of focused practice. But it may be that even this is not sufficient for just anyone to achieve excellence – you may still need to start out with certain genetic predispositions. Simonton, the psychologist who studies scientific productivity we mentioned earlier, suggests that “scientific achievement is not a matter of either talent or training but rather a matter of talent operating in the context of training.” To say that some people are genetically predisposed to be more likely to successful scientists is not to say they are born with some diffuse “gift” for science. Rather, there may be a number of important composite factors – both intellectual and personality characteristics – which are at least partially genetically determined, and which contribute significantly to scientific achievement.32
What about other personality factors?
There’s reason to think that personality factors contribute substantially to academic success in addition to intelligence and deliberate practice.
This makes sense if we consider Shockley’s point that we mentioned earlier: being a successful scientist requires combining multiple distinct skills, including the ability to think of a good problem, the ability to work hard on it, the ability to write well, and the ability to respond well to feedback and persist in making changes. Different personality variables are likely to contribute to these different abilities – creativity and openness might help you to think of a good problem and look for unusual solutions, whereas conscientiousness will help you to persist in working on a problem when it’s no longer exciting.
There’s some evidence on how personality factors influence academic and research performance to back this up.33 A meta-analysis of studies on predictors of academic performance found that “conscientiousness added as much to the prediction of tertiary academic performance as did intelligence”.34 The same paper also reports weaker effects of agreeableness and openness to experience on academic performance.
Other research suggests that intellectual curiosity may be an important determinant of academic achievement (a position we’ve found echoed by almost everyone we’ve spoken to in the field). Another meta-analysis of studies of academic achievement looked at the predictive power of the personality construct Typical Intellectual Engagement (TIE) – a measure of enjoyment of intellectually demanding activities – alongside intelligence and effort. They found that intelligence accounted for the greatest variance in academic achievement, but that the combined effects of TIE and effort equalled those effects of intelligence.35
There’s also some moderate evidence that creativity is important for doing successful research, and in some cases high levels of creativity can compensate for lower levels of intelligence36.
Finally, we expect that advancement in fields which require managing teams of young researchers and applying for grant funding will benefit from strong social skills.
How to assess your personal fit at each stage of your career?
You have a few key opportunities to assess whether an academic career is for you:
Something you can do at any stage is practice research and research-based writing. A relatively easy way to get started is by trying your hand at writing ‘fact posts‘.
During your undergraduate studies, aim to complete at least one summer research project. This will help with graduate school applications, while also giving you a taste of what research is like — it’s pretty different from studying a subject at undergraduate.
After you graduate, if you’re highly confident that academia is your top option (say 80%+ confidence), then aim to continue directly into graduate studies. If you’re unsure (say 40%+), then the 1-2 years between undergraduate and graduate study are a good time to experiment with other options you’re interested in. We recommend experimenting now rather than after your PhD because, as we explained earlier, it’s hard to take any break from academia after your PhD. This could also be a good time to consider a research assistant position or pre-doctoral fellowship, which can allow you to work in a research lab and test your fit for academia while also sometimes allowing you to take classes at your host institution, without committing to a PhD.
At that point, if you still think academia is for you, then apply to graduate studies. Again, if you can get into a top 20 school in your subject, that’s some indication of potential. For graduate schools in the US, you have to take the GRE (Graduate Record Examinations), and your scores in this also provide some indication of potential – a meta-analysis found that GRE test scores predict grade point averages in graduate school, faculty ratings, citation scores, and later career research productivity.37
During your studies, you might be able to experiment with some internships on the side, to keep learning about alternative options. For example, as a graduate student you might be able to get internships in government, think tanks, or industry, depending on your field.
Near the end of your PhD, you face a key decision-point: will you continue? This is a good opportunity to re-assess your fit. If you think there’s a reasonable chance academia is your top option (say over 50%), then it’s worth continuing to keep your options open. You can also apply to postdoc positions to see what you get. If you’re able to get a postdoc in a good department/group without a large teaching load, it’s usually worth taking. In our individual profiles on specific fields, we discuss specific signals of potential at this point — but the conventional advice is that you (i) have some reasonably good publications (ii) have an offer to do a postdoc at a top research centre.
The next reassessment point is when you start applying for permanent positions. See what you get, and if in doubt, continue with academia.
However, it’s also worth mentioning that many people find it difficult or scary to leave academia – even if they don’t enjoy it anymore, or no longer think what they’re doing is valuable. People on the academic track are not taught about, nor encouraged to value, options that compete with academia. One thing that might help with this is to think in advance what your “exit” conditions are – under what conditions you’ll decide to leave academia and try something else (e.g. if you struggle to get a postdoc position in a top group/university a couple of years after your PhD), and commit to reassessing your options at each new career stage (after a PhD, after your first postdoc, when you’re going up for tenure-track positions.) If you’re going to leave academia and transition to another area, it’ll probably be easier to do so before your mid thirties (as a very rough guideline).
As well as your fit for academia in general, you’ll also want to think about your personal fit for specific areas of research, which you might be able to test by:
Doing research projects or working as a research assistant (especially if you’re an undergraduate);
Finding out what the prerequisites/normal backgrounds of people who go into this field are, and comparing your skills and experience to them;
Reading key research papers, trying to contribute to discussions with other researchers, and getting feedback on your ideas;
Talking to professors or other successful researchers in a field and asking what they look for.
Maximising your impact within academia
Which field should you go into?
When choosing the best field to focus on, we recommend considering your personal fit, the impact of the field, and your back-up options.
At the beginning of your career, most of the academics we spoke to recommended focusing on developing expertise and building up a good track record of publications. A good publication record will be essential for you to get the best academic jobs and funding later on, which in turn will give you the freedom to pursue whatever research questions you think are most important.
Learning and building up a good track record doesn’t have to be completely at odds with doing valuable research, though. The best people to learn from are likely to be those who you think are doing good work on important questions.
And while building up a good publication record is important, it would also be a mistake to publish in areas completely divorced from the kind of research you want to do long-term, as then you risk getting pigeonholed in the wrong area, and won’t necessarily be building the right expertise. If you want to have a big impact through your research in the long-run, it seems important to spend at least some time early in your career on projects you think could be extremely valuable.
Maximising your impact later on
1. Maximising the impact of your research
Richard Hamming, the mathematician we mentioned earlier, famously went around asking his colleagues, “what do you think are the most important problems in your field?” When they responded with a few specific ideas, he followed up with, “And why aren’t you working on them?”
This may not have made him many friends, but it demonstrated how many academics simply don’t ask themselves this question, and don’t actually spend a great deal of time thinking about how they can do important research.
In a famous speech on ‘why so few scientists make significant contributions’ Hamming reported his observation that: “the average scientist, so far as I can make out, spends almost all his time working on problems which they believe will not be important.”
This raises an important point – in order to do important research, you need to spend a reasonable amount of time thinking about what questions are important.
As we mentioned previously, even within a field that seems highly relevant to important problems, some research questions will be much more valuable than others.
Many people seem to “fall into” research areas due to circumstantial reasons: they inherited a PhD topic from their supervisor, and then looked for research positions in related areas. This leads to crowding and ‘path dependence’ in what topics are addressed by academics.
There can also be a tendency in academia to focus on very narrow areas. That makes it easier to explain what your specific area of expertise is, but it also leads to a lack of big picture thinking.
How can you ensure you focus on important questions throughout your academic career? When thinking about how to choose research questions within a field, you can use the same framework we use to think about cause areas: looking for questions that are important, tractable, and uncrowded. For example, if you’re a biomedical researcher you might try to identify diseases that affect a lot of people, which not many researchers focus on, but where it seems like more research could yield effective treatments.
Here are some rules of thumb, based on the advice we heard in our interviews with top scientists, which may help you to identify high impact research questions:
Look for research questions that seem important but are short on talent. Ask why people aren’t working on these questions – is it because they’re intractable (good reason!) or because there aren’t good incentives to work on it, it’s not fashionable or doesn’t fit neatly into an academic field? (bad reasons!) Sometimes small findings and unglamorous innovations can make a big difference – not all valuable research is paradigm-shifting. See examples of research questions we think could be promising from a longtermist perspective.
Work on really giant problems that are in the process of being solved – even if there’s a lot of attention on a problem, if it’s big enough additional effort could make a big difference, and if there’s a track record of valuable progress that’s promising. For example, global health research gets quite a bit of attention, but an extra researcher could still do really valuable work on developing treatments and vaccines, especially if focused on more neglected areas. Focusing on big problems that seem tractable, and then looking for neglected areas within those areas could be a good approach.
Bring new skills, perspectives or technology to an important area – for example, Daniel Kahneman brought findings from psychology to economics and ended up winning a Nobel Prize. This might also allow you to work on questions that don’t fit neatly in an academic field, that others might miss. Learning maths, statistics, and how to work with data seems likely to be useful in all fields, and might give you an edge or do work others can’t if you’re in a field that’s not typically quantitative.
Ask other experts in the field what they think the most important questions are (like Hamming) – academics who have a lot of experience in a field might be better able to spot important, neglected areas. However, the opposite could also sometimes be the case – if a professor has built their career working in a niche area, then they may well have a bias towards thinking that area is disproportionately important. If you can find more experienced researchers who seem to have thought hard about the most important questions in their field, learning from them could be really helpful.
Some of the researchers we spoke to also emphasised the value of developing the skill of identifying which research questions are important, and regularly trying to get feedback on whether your research is going in the direction you’d hoped. Anders Sandberg, a research fellow at the Future of Humanity Institute, told us that “being able to evaluate what you’re working on, having some kind of importance check, and setting your priorities straight is really important.” He even said he’d rather take a pill that gave him this skill than an intelligence enhancing pill.
2. Maximising your impact in academia beyond research
As we’ve covered, you might be able to have as much or even more impact in ways that go beyond your own research – by doing advocacy, applying your research outside of academia, advising policymakers, and managing or influencing other researchers.
Your opportunity to make a difference via these different routes will depend a lot on your personal strengths, preferences, and the specific opportunities in your field.
Perhaps the best way to maximise your impact in academia beyond research is to stay open to all these possible paths early on, and spend time exploring different opportunities early in your career to learn where you can have the most impact.
For example, during a PhD or postdoc, keep an eye out for opportunities to communicate your research for a popular audience, and see whether you enjoy and are good at communicating complex ideas simply.
You might do this by looking for popular science publications that look for contributions from academics, or finding opportunities to give talks (which isn’t too difficult in academia – even talking about your research to a wider academic audience at conferences could be good practice.)
If you notice that other academics in your field or department seem to be applying their research beyond academia – advising companies or government, or involved in building products – then talk to them about how they got into this area.
Some companies and parts of government also offer internships specifically for PhD students, which could be a great way to explore the potential to apply research from your field outside of academia. For example, the main research councils in the UK run a “policy internship scheme” for students whose PhD research they fund.
Even if you think you might be able to have a large impact through outreach or applying your research, you’ll still probably want to focus on excelling as a researcher, and developing a good publication record, early in your career. This is because being successful in advocacy, as a governmental advisor or other similar paths, requires you to both have genuine expertise and strong credentials.
If you decide that you definitely want to focus on outreach over research – a “public intellectual”-type path, then you may want to focus on getting an academic position that gives you plenty of time and freedom as soon as possible.
It is not a great surprise that many of history’s most influential figures – like Adam Smith, Stephen Hawking, Rosalind Franklin and Jonas Salk – have worked in academia.
On the other hand, most people who set out to become academics will not make it, as the impediments to reaching a position in which you have discretion over what you study are substantial.
But for those who have the intelligence, conscientiousness and curiosity to succeed, working in academia provides an unusual opportunity to work on the problems they think are most pressing in the world – at least so long as they can find grants to fund them and journals to publish in.
Those who are interested can learn more about specific PhD options and find interviews with academics discussing how the path has worked out for them in the Learn more section.
Think you have a good shot at becoming an academic researcher focused on a pressing problem?
We’ve helped dozens of people decide if this is the right path for them, and if so how to go about it. We can offer introductions and funding opportunities, or answer specific questions you might have. If you think you have what it takes, apply for our free coaching service.
See figures 7-12: Hauser, Robert Mason. Meritocracy, cognitive ability, and the sources of occupational success. Madison, Wis, USA: Center for Demography and Ecology, University of Wisconsin, 2002.↩
Full disclosure: Giving What We Can is part of the Centre for Effective Altruism, which also serves as the parent charity of 80,000 Hours. Without Peter Singer, there’s a good chance 80,000 Hours would not exist, either!↩
Wikipedia contributors. “Bruce Chapman (Australian economist).” Wikipedia, The Free Encyclopedia. Wikipedia, The Free Encyclopedia, 12 Oct. 2017. Web. 25 Jul. 2018.↩
A recent survey of leaders in the EA community asked what skills they wanted to see more in the community as a whole – and found that specialist expertise in subjects such as machine learning and biology were perceived to be of medium importance (though not as important as more generalist skills such as management and policy expertise.)↩
Hesli, Vicki L., and Jae Mook Lee. “Job satisfaction in academia: Why are some faculty members happier than others?.” PS: Political Science & Politics 46.2 (2013): 339-354.↩
Larson, Richard C., Navid Ghaffarzadegan, and Yi Xue. “Too many PhD graduates or too few academic job openings: the basic reproductive number R0 in academia.” Systems research and behavioral science 31.6 (2014): 745-750.↩
Taylor, Martin, Ben Martin, and James Wilsdon. The scientific century: securing our future prosperity. The Royal Society, 2010.↩
Roe obtained difficult problems in verbal, spatial and mathematical reasoning from the Educational Testing Service, and created three tests which were administered to the 64 scientists. The same tests were also administered to a cohort of PhD students, who also took standard IQ tests, which were used to normalise the V, S, & M tests. The median “normalised” scores for verbal, spatial and mathematical reasoning amongst the 64 scientists were 166, 137 and 154 respectively – the median IQ of the PhD graduates tested was 141. The median scientists’ score on the spatial test was lower than the median score of PhD graduates, but Roe notes that spatial reasoning scores correlate with age – younger men are likely to get higher scores – so the comparison with PhD graduates is not so direct (i.e. the scientists may have scored higher than the average PhD student when they were the same age.)
Hunter, John E. “Cognitive ability, cognitive aptitudes, job knowledge, and job performance.” Journal of vocational behavior 29.3 (1986): 340-362.↩
Hunter, John E, Frank L Schmidt, and Michael K Judiesch. “Individual differences in output variability as a function of job complexity.” Journal of Applied Psychology 75.1 (1990): 28.↩
“Although GMA predicts performance in all jobs the more complex the job is13, the stronger the relationship between GMA and performance.14 And the more complex the job, the more variation there is between top performers and bottom performers.15 So if you have one of the highest levels of GMA in a highly complex job, you’ll have a high output compared to the average performer.” Intelligence matters more than you think for career success.↩
This is the best place to understand the true relationship between IQ and academic performance rather than later in the education system, because at later stages lower-IQ people have already dropped out, so other factors beyond IQ will account for more of the variance.↩
“At one extreme, some fields are characterized by relatively early peaks, usually around the early 30s or even late 20s in chronological units, with somewhat steep descents thereafter, so that the output rate becomes less than one quarter the maximum. This agewise pattern apparently holds for such endeavors as lyric poetry, pure mathematics, and theoretical physics, for example. At the contrary extreme, the typical trends in other endeavors may display a leisurely rise to a comparatively late peak, in the late 40s or even 50s chronologically, with a minimal if not largely absent drop-off afterward. This more elongated curve holds for such domains as novel writing, history, philosophy, medicine, and general scholarship, for instance.” From Simonton, Dean K. “Age and outstanding achievement: What do we know after a century of research?.” Psychological Bulletin 104.2 (1988): 251.↩
“On a population level, we see that the average doctor is 30 IQ points higher than the average janitor, that college professors are overwhelmingly high-IQ, and we think yeah, this is about what we would hope for from a statistic measuring intelligence. But on an individual level, we see that below-average IQ people sometimes become scientists, professors, engineers, and almost anything else you could hope for.” Against Individual IQ Worries, SlateStarCodex.↩
“Rather than define talent as a mysterious phenomenon that operates independent of domain-specific expertise, talent is best conceived as a process that openly involves that expertise. In different terms, scientific achievement is not a matter of either talent or training but rather a matter of talent operating in the context of that training… A person is certainly not born with a diffuse ‘gift’ for science. Instead, the natural endowment most likely consists of a weighted composite of numerous and highly specific intellectual and personality characteristics.” Simonton, D.K. (2008). Scientific Talent, Training, and Performance: Intellect, Personality, and Genetic Endowment. Review of General Psychology 12(1), 28-46↩
Chamorro‐Premuzic, Tomas, and Adrian Furnham. “Personality traits and academic examination performance.” European journal of Personality 17.3 (2003): 237-250.↩
Poropat, Arthur E. “A meta-analysis of the five-factor model of personality and academic performance.” Psychological bulletin 135.2 (2009): 322.↩
“There are no significant differences in academic achievement between the High IQ – Low Creative and Low IQ – High Creative groups. This supports the findings reported by Getzels and Jackson (1962), Torrance (1959) and Yamamoto (1964a) of equivalent academic achievement among the highly intelligent and highly creative groups.” Palaniappan, A. K. (2007, July). Academic achievement of groups formed based on creativity and intelligence. In The 13th International Conference on Thinking Norrköping; Sweden June 17-21; 2007 (No. 021, pp. 145-151). Linköping University Electronic Press.↩