Ramping up the academic rat race: Moore’s law of scientific publishing?

To conclude this series of posts, I’ll start with a quote that my good friend Chris is fond of.

C: “Ben, do you know who wins the rat race?”

B: “I don’t know, who wins the rat race, Chris?”

C: “A rat.”

Of course, Chris has long realized that to even play in the rat race that is our modern society, one must become a rat in order to succeed. Joining the rat race is certainly essential to a career in academia in its current state.

This idea crystallized for me this morning when I found a new paper by Brischoux and Angelier. This paper has suggested something crazy, that I suspect might be broadly true. During the time period that I have been a graduate student and a postdoc (2005-2015), the number of scientific papers required to obtain a permanent academic job in evolutionary biology in France has doubled. Doubled. What’s more depressing is that the number of years spent in temporary employment has also doubled during this time. If their findings are more broadly applicable across the sciences, this represents a fundamental shift in academia that I’m fairly certain nobody would have predicted a decade ago. From the time that I decided academics might be an appropriate career choice for me, until now, the rat race has changed entirely. The cause of the increase in publication stats and temporary work is complex, but I suspect it doesn’t have to do with increased absolute productivity. People applying for faculty positions today aren’t twice as smart, or twice as productive as people ten years ago.

One of the obvious reasons for the increase in number of papers is the relentless drive toward quantitative metrics for scientific output. Indeed there has been a great deal of discussion about scientific metrics (references within the linked manuscript), and essentially the problem boils down to one idea: it’s really hard to measure the impact of scientific output. However, hiring committees, promotion and tenure committees, and awards committees must all conduct their business, and being quantitative scientists, they are looking for ways to score, or rank scientists’ output, and justify their decisions to administration. There are flaws in every scoring system, whether one ranks candidates by number of papers, patents and reports produced, the number of times these works are cited (h-index), or some other arbitrary assessment of quality. However, despite the flaws of this approach, metrics provide something that’s easy to understand for university administration—the numerical rank of Candidate A is higher than Candidate B.

Consider Job Candidate A who has written an impressive 16 papers that have been cited by 135 other papers. On the surface, Candidate A appears to have greater scientific impact than Candidate B who has only written 8 papers with 92 citations. Candidate A has an h-index of 10, whereas Candidate B has an h-index of 8. Using quantitative metrics alone, one is unable to determine that most of the citations of Candidate A’s work were from other scientists refuting the findings of a single controversial manuscript in a more popular sub-discipline, and the remainder of his/her papers were only cited a handful of times. However, Candidate B had made a series of important discoveries in a smaller sub-discipline which redefined the field, and are indicative of a truly rising star. This is an extreme example, of course, but highlights some of the problems with quantitative metrics of scientific output.

Second, an evaluation system based on number of papers and citations inevitably splits science into thinner “salami” slice manuscripts and “least publishable units.” Additionally, some papers may be important but not widely cited in the academic literature. An important discovery in applied science may have broad reaching implications for an industry, but little “importance” in terms of scientific literature citations.

Third, and interestingly, the most cited papers in science consist almost entirely of methods papers. While these papers are most certainly important, they didn’t directly change our understanding of the world we live in.  Hence number of citation isn’t always a measure of creativity or thought provoking work. So the quantitative system may not be measuring what a hiring committee might want, which is groundbreaking scientific work that makes an impact on society in the near or distant future. My favorite fortune cookie this year read “The purpose of education is not knowledge, it is action.” Impact on broader society (on an admittedly nebulous time scale) is essential for good scientific work.

Finally, there is a demographic component to this problem that I think boils down to an availability of cheap labor. The number of postdocs and grad students has increased substantially relative to the number of faculty over the past couple decades, and it appears that we’re near the breaking point of the academic ponzi scheme that has developed. It  means that early career scientists are spending many more years in postdoctoral purgatory than in the past. Living contract to contract, moving from city to city. Perhaps this is broadly reflective of larger society where full time permanent employment seems like a relic of the 20th century. Quote from a senior faculty member, “As I got older and developed more outside responsibilities . . . it became easier to have more postdocs than graduate students because they didn’t need as much supervision. You could have a bigger lab that way without occupying more of your time.” And maybe more importantly, “In 1970, scientists typically received their first major federal funding when they were 34. In 2011, those lucky enough to get a coveted tenure-track faculty position and run their own labs, at an average age of 37, don’t get the equivalent grant until nearly a decade later, at age 42.” This is unsustainable, and the end result will be a lost generation of scientists who mistakenly thought there was a career path laid out for them because the positions to train them existed. Picture a medical school that asked students to go hundreds of thousands of dollars in debt, only to have jobs for about 15% of its graduates. There certainly wouldn’t be students lining up to pay tuition after a few years. So why do people continue to start PhD and postdoctoral programs? I suspect it’s because lost opportunity cost is generally invisible to a green 22 year old potential graduate student. It’s hard to think that far into the future and see what might lie ahead, especially when you have a professor that you like and respect telling you how good you can be at science if you only went to grad school. What else could you have done with those 20 years (between grad school and first grant) that would have contributed more to society? There is an enormous lost opportunity cost to the individuals and to society that they might have served. In retrospect, there are no easy solutions for the current supply glut, given the time and resources that have been invested into training these talented people. However, maybe we should stop training so many people in the first place?

I have spent a lot of time recently thinking about careers in science, so I’d like to ask you one question if you’ve made it this far without closing your browser: if you were advising an undergraduate student about careers after graduation, could you honestly recommend a PhD and academia as a viable option, given the current state of science? Could you recommend a prolonged underpaid adolescence (grad school, postdoc), while working toward a rare, and even more intensely competitive (but equally underpaid) junior faculty position that may or may not exist when you’re done? I’m certain I could not recommend this path to a student starting today. If I were a tenured faculty member right now, I could not justify taking on a graduate student. I’m not sure it would be ethical. Ten years ago, when I started on this path, things were a little different. I saw postdocs in my lab, and in my large interdisciplinary project getting good jobs, and saw scientists doing interesting work (one out of three attained a TT position). I saw a reasonable career pathway that would be challenging but rewarding. Today’s students should (hopefully) see the numbers from studies like the one I’ve highlighted today, and run away screaming from an academic path that offers little in the way of job prospects, salary, or future job security. The current labor supply glut is far greater than the system can absorb into permanent positions in the near future, and a silent majority of trained scientists have dropped off the academic track along the way. The tenure track has become a war of attrition, rather than a fair competition among young scientists with the best new ideas, and that is a big problem.

So given my current position as a postdoc, I suppose the question is: what now? Before I started my job, I laid out a set of concrete rules that I’ve followed perfectly so far. These rules help me manage and accept the major structural problems in academic science that I see around me. I took my current position with three ideas in mind: 1) this is an exciting project, in a field I always wanted to explore (I changed disciplines entirely following my PhD), 2) this position would be the only postdoc I’d ever have, and 3) if I didn’t have an academic TT job by the time this postdoc was over, I’d find a different career. So far #1 has been great success. I’ve learned a ton, contributed a solid paper that helps understand “Snowball Earth” (an undergraduate fascination of mine), and will continue to work hard to understand triple oxygen isotopes in the geologic record, at least until funding runs out. I’m still awaiting results of #2, as I’ve applied for a number of jobs recently and interviews are still being conducted at a number of schools. As for #3, I’m funded until the end of July, and we’ll see where things go after that. As much as I’d love an academic career, there are a lot of other paths in life, and I’m not so one-track focused that I’d do “whatever it takes.” At the end of the day, it’s just a job like any other, and there are things that are important in life outside of science.

Perhaps most of all, when I’m old and grey and look back on my career, I hope I can confidently say that I didn’t win the rat race. I think Chris would be proud of me for saying so.

~B

Advertisements

Science and nationalism – how does Canada fund early career scientists?

I’ve been debating whether or not to write a post about my recent NSERC experience, mostly because in doing so I may sound like I’m whining. I promise I’m not. I’m just trying to help others understand how science is funded in Canada, and help other young Canadian scientists navigate the complexities and difficulties of building a scientific career. Maybe a sharp politician or bureaucrat will read this and something might change? Nah I’m dreaming about that… but I do promise I’m not whining, just want to put this story out there because I know there are at least two more people in my position, and maybe we can form a support group or something. I’m fortunate to have continuing funding from my current lab, and we have a super strong NSF proposal in review right now to continue the awesome work we’ve been doing. Science will go on without support from my home country. Here’s the story:

I recently received results from the Banting Fellowship competition, a postdoctoral award that pays a young scientist’s salary for two years to conduct advanced research after completion of a PhD. The competition is unusual because unlike regular NSERC awards, this one is not separated by discipline – biochemists, geneticists, engineers of all kinds, mathematicians, physicists and geologists compete for the same 24 awards each year. Interestingly, in the ~5 year history of the program, a geologist had never won the award (there have been two or three awardees studying modern climate, surface water, and groundwater that broadly might be grouped into Earth sciences, but nobody who regularly looks at rocks, or thinks about Earth’s history before the Holocene has won this award, which seems odd in such a natural-resource-rich country). That is, until this year, when I placed in the top 24. I opened the “Competition Results” PDF, saw the high scores, and was ecstatic. I’d worked hard on this proposal “The Geological History of Oxygen* ” (can you believe the amount of oxygen in the air we breathe has changed through Earth history and we know practically nothing about it? Sort of important to understand how oxygen might be affected by climate change…), and given my recent work in the oilsands that has recently been published and getting reasonable amounts of attention from industry and the press, believed that I had a great chance of winning this fellowship. To score so high was a great feeling, and it was nice to have some success in a business that successes are few and far between. What this meant for me was two more years of funding at Harvard to actually carry out this research (the first year of my postdoc I spent building and testing a new laser instrument to measure oxygen isotope ratios). Then I read the second PDF, “notification of decision”…

Please note that the proportion of Banting Postdoctoral Fellowships that may be awarded annually to individuals who apply in collaboration with a foreign institution is capped at 25% overall. While your high ranking would normally achieve funding, due to this 25% cap we are unable to offer you an award at this time.

Crushed doesn’t begin to describe how I felt that day. I’m someone who’s been rejected for about 100 proposals along the way, and usually handle rejection letters pretty well, this one was different. Now fair is fair, the rules about foreign held awards are set beforehand, and I’m under no delusion that the 23rd (or 4th, or 15th) ranked scientist is substantially superior to the 25th or 26th ranked people in this competition. Comparing interdisciplinary scientists must be a challenge of epic proportions, it would be a very difficult committee to chair. There are so many great young scientists in Canada, it’s an honour to be mentioned in this group, and this was a sort of silver lining to an otherwise depressing Friday. However, I do question why the restriction exists in the first place, and after doing a little digging, I think that the distribution of these awards is misaligned with the purpose of the award:

The Banting Postdoctoral Fellowships program provides funding to the very best postdoctoral applicants, both nationally and internationally, who will positively contribute to the country’s economic, social and research‑based growth.

In a nutshell, the Banting Fellowship exists to train the next generation of people who will become the next leaders in scientific fields in Canada. Nobody should undertake a postdoctoral fellowship unless they want advanced training in research – unlike the PhD which can have a broader outlook for jobs outside of research, postdoc work is purely research focused. Canada wants to fund good people to develop their skills and hopefully contribute to the country’s scientific expertise over the next couple decades. However, awarding only 25% of these to Canadians who pursue postdoctoral work in a foreign country is directly counter to the reality of who is hired in Canadian institutions for faculty positions, and this is where I have my beef, and where it may sound like I’m whining. I’m not.

A very wise friend advised me a few years ago it would be academic suicide for me to stay in Canada for a postdoc. As I demonstrated in my last blog post, she was right; exactly one Assistant Professor of geology in Canada did a PhD and Postdoc in Canada (and she’s a very special case, with a very special resource—Alberta’s dinosaur fossils—that are essential to her work). The pathway to an academic career in geology in Canada is either a) do your PhD in the US or Europe, OR, b) do your postdoc in the US or Europe, OR c) a & b. In Geology, 66% of new hires for tenure track positions did a PhD outside of Canada, and 73% did a postdoc outside of Canada. There’s a lot of merit to this – Canada is small, and we can’t possibly be on the cutting edge of every field in every sub-discipline. Hiring people who have been trained at top global institutions is great for students, cities, and yes, Canada as a whole. The breadth of thought these people bring to their institutions is incredibly valuable, and should be encouraged. It’s probably worth mentioning that Canada doesn’t have many top-ranked institutions, despite lots of chest thumping from smaller schools, who certainly have pockets of outstanding people, we have four “world class” institutions: U of T, McGill, U of A, and UBC. Counter this with the US that generally holds 18 out of the top 20 in most international rankings (with only Cambridge and Oxford in the UK holding top 5 positions). It’s entirely understandable that Canadian universities want to hire the best people, period. Andrew Leach made a very good point this morning that universities shouldn’t be exempt from Canadian hiring laws, but they do have more selective criteria for hiring than most regular jobs. This can be paraphrased as, “do you want universities to hire the best available person, the top 1% of the field, or do you want the best person above a minimum standard.” This has a certain Office Space “pieces of flair” ring to it that made me chuckle.

Anyway, back to the point of my post – if nearly everyone who ends up contributing long-term to Canadian research as faculty members of a university spends some time outside of the country, why not reallocate funding at more junior stages to reflect this? Perhaps more appropriately, 50% of the Banting awards could be allocated to Canadians with Canadian PhD’s to pursue scholarship abroad, and 50% of the awards to expat Canadians and foreigners with foreign PhD’s to “come home” and bring their expertise back to Canada. This would better reflect the reality of who is hired at our universities, and support those who will ultimately contribute to building a stronger Canadian scientific enterprise. An alternative might be to allocate the awards to the best candidates, period, regardless of location. Food for thought.

Ultimately, I’m insanely disappointed by the results of the competition, but science is all about striking out and getting back up to bat the next time and trying to get on base. I’ve had a couple weeks to digest this one, and I’m feeling better about it. After giving a talk at Princeton last week that received very positive feedback, I’m excited to get back in the lab and work on Earth history questions that will help answer where our planet is heading into a century of rapid climate change. Telling Earth’s history has never been as sexy to fund as “cure for cancer” research. However, if the geological past is our best predictor of our planet’s future course, studying Earth’s history has never been more relevant as it is today.

To be continued…

~B

ps – If you happen to be one of the two other people who received the same letter from the Banting committee that I did, I’d love to hear from you, and get some sort of dialogue going on. Might be a shot in the dark, but the internet makes the world a little smaller these days…

pps – If anyone has better stats on Assistant professors in other disciplines, or better stats on those in Geology than I dug up, I’d love to see them. Having a reasonable idea of career path is important for new PhDs, who should be aware of the long uphill slog ahead of them if they want to continue on this career path.

*while I indeed was the primary author of the proposal, my adviser Dave Johnston provided outstanding comments and feedback, in addition to the rest of the J-lab who supported the development of these ideas. I can’t say enough how fortunate I feel to be a part of this research team. Thanks everyone for your support (and beer).

On the origin of Assistant Professors

Featured

After spending the last fifteen months as a postdoc (and to be honest, mostly enjoying it, it’s been fun despite the brutal frustration of building a new homebrew instrument and getting it running), I’ve come to appreciate first hand how difficult getting a faculty job offer is—not that I previously held delusions that it was easy. I’ve recently gone through my first round of job applications, and I’ve had a few bites, but no offers so far. It’s still early in the game, and interviews are still being conducted, but I’m not holding my breath. To evaluate where I stand compared to other possible candidates, I went on a small quest last night to figure out what the career paths look like for recently hired Assistant Professors of Geology (Earth Science, Geoscience, whatever you want to call it) in Canada. I chose Canada because I’m Canadian, but also because it’s a smaller more manageable data collection effort. Admittedly, this was partly self-serving because this year there were no hires in Canada for which I would be a suitable candidate, and I wanted to scope out where might be hiring in the near future. The findings were intriguing, and while the sample size is small, it may be helpful for students who are trying to plan their careers. The key piece of information that is lacking in this study is nationality—there are restrictions on hiring at all universities in Canada that rank equivalently qualified Canadians and permanent residents higher than foreign nationals in almost every job competition including those of faculty. Therefore, the results should be skewed toward Canadian citizens, regardless of PhD or Postdoc location, but I was unable to test this idea.

Here’s what I found (please note that the data are incomplete,. but I can only spend so much time hacking around trying to dig up details that should be easily displayed on a university website). At the 12 universities I examined, there are 36 Assistant Professors of Geology. Fourteen (39%) were female, and twenty two (61%) were male. An obvious gender imbalance, but small sample size. The median year of attaining a PhD was 2007, and the median number of years spent as a postdoc was three. There was a wide range of postdoctoral experience, ranging from none to ten years. Most Assistant Professors (67%) did their PhD outside of Canada. Most also did their postdoc outside of Canada (73%). There was only one case that did both PhD and postdoc in Canada and currently holds a junior tenure-track position.

Qualitatively, all of the people holding Assistant Professorships had exceptional qualifications and publication records, and I don’t see universities being short on outstanding people anytime soon. However, this led me to another (yet unanswered) question: where do all the students and postdocs go? In any university department there are far more PhD students than faculty. Doing the math, there were 12 people hired in the last ~8 years, and probably 400 (this is a back of the envelope calculation) graduating with a PhD during the same time frame from Canadian schools, the rate of 1 faculty job for ~30 PhD’s seems similar to that of the American system. In grad school, nobody is under the delusion that they are all going to become professors, but that we don’t know the ultimate destinations of these highly qualified people speaks to our general ignorance of science in our society. The data to conduct this study certainly exist—NSERC collects information about trainees from all its supported researchers, and it would be interesting to follow up on this in a more detailed and data rich study.

What I took away from this is probably more of the same things I knew already. Things are hard if you choose science as a career path. Generally low pay, poor job prospects, etc. etc. that we should all be aware of before embarking on that journey. However, I also see opportunity. The 29 out of 30 people (the 97%?) who didn’t get a faculty job (or maybe didn’t want one in the first place) are still very qualified scientists with good ideas and clever ways to solve problems. Perhaps by reversing the trend of training so many students, better science could be done with a more permanent staff. If I were a PI, I’d expect better science to be done by a competent research associate who’s spent ten years thinking about isotope geochemistry than a newly minted undergrad who doesn’t know the difference between alpha, delta, and epsilon. Perhaps by taking advantage of the current glut of qualified researchers Canada can change the way science is done, and get back to leading, instead of lagging behind, the rest of the scientific world.

~B

ps – I’d hoped to get this published on Darwin Day (hence the title), but I’m a day late.
pps- The raw data can be found here, and admittedly it’s a bit rough. Please feel free to use it however you’d like to share, download, make it better, etc.

Location_PhD Years as Postdoc Male_Female Location_Postdoc