Win or go home
I like my job. I like it so much that I can't really imagine doing anything else. But there's also not a lot else that a PhD in the evolution of animal sexual behaviors will get you. There's not a lot of call for animal behavior in industry. And I have zero interest in finding an industry niche. So, I have a lot to gain by succeeding and a lot to lose by failing. Let me explain.
The weird thing is that the structure of academia precludes partial success. For every PhD who wants to stay in academic science, there is only one goal: to be a professor. If you don't manage to get hired as an assistant professor, that feels like failure. And after between 5 and 7 years spent earning a PhD and between 3 and 5 years of postdocing, during which you may move around once per year. All the moving comes at great financial expense and tremendous cost to one's personal life. For instance, you can give up hope of living near your family. If you have a spouse, they have to change jobs as regularly as you do. And if you have kids, they're changing schools, daycares, and social networks on, sometimes, only a few weeks' notice. That's a lot of drawbacks. I think the benefits outweigh the costs, which is why I'm still interested in academia, but the benefits need to be discounted by the probability of success.
At research universities that train PhDs, though, each professor is likely to be training at least 3-5 PhD students at any one time in their career. If professors work between, say, 35 (not an unreasonable age to be hired as an assistant professor these days, though it seems like it's getting older all the time) and 70 years of age (at which time they may be relatively likely to stop training students, if not likely to stop doing research), and if a PhD in my field (in the US) takes on average 6 years to complete, a single professor could easily generate around 35 "replacements".
This is not feasible. Some of the pressure is relieved by the need for professors at institutions that do not train PhD students -- institutions that grant only bachelors and masters degrees nevertheless have professors. And it is also true that not everyone who earns a PhD wants to be an academic (and that's great!). But the rest of us continue to take jobs that are probably underpaid relative to our skillsets and accept that we'll likely live in constant stress about our job prospects for, on average 22-35 -- or however long it takes you to get from the beginning of your PhD to tenured professor.
In the meantime, many people leave at each stage. This is probably death, in my evolution metaphor. Some PhDs age out of the system as postdocs. There is a maximum amount of time in which you must finish your PhD, and if you fail to finish, or fail to find a postdoc, you're out. In many countries there's actually a maximum number of years you can be a postdoc before you are no longer allowed to accept such work. Sometimes, this is meant to be a pro-labor measure, in that if salaries for postdocs are lower than for those PhD-holders with permanent positions, it protects workers from being exploited by universities who are purportedly paying them with experience. But taking a non-postdoc job to work as a lab tech is the end of the academic road.
As an evolutionary biologist, I can tell you that one of the necessary things for evolution to happen is that not all individuals are equally successful. Check. The second thing is variation among individuals that influences their success. There's definitely a bunch of that good variation for selection to act upon. So here we are, in a system that selects for the most productive scientists. And the field is evolving in response to that selection. People are publishing more collaborative papers than they used to. For instance, when my PhD adviser finished, it was common for all PhD students to publish single-author papers for their entire thesis -- now, you are expected to include your PhD adviser on all your publications. People are being hired at much later stages -- it used to be possible to land an assistant professorship without having published a single paper, which really blows my mind. You might be asking yourself if variation among scientists in productivity is heritable, and I might argue that it is, in the sense that having a productive adviser at a prestigious institution opens doors for you and allows you to be way more productive.
Okay, so you get rewarded for being productive and punished for being less so. That's okay, right? It's a meritocracy. How do we measure productivity, though? As scientists, we're extremely concerned with how we can quantify multidimensional traits, so we've given this some thought. We'd like something quantitative. But our metrics are fairly imprecise for how to measure a great scientist. How many publications do you put out each year? Were they in high-impact journals, of the kind that are widely read and disproportionately likely to be cited? How many times per year are those published papers cited by other scientists? Who are the other scientists that are citing you? How many grant dollars did you receive? How many students who worked with you graduated? These are our best metrics by which to judge the quality of scientific output.
(There are a lot of problems with this idea related to access to resources, which intersect with various aspects of privilege, but that's a topic for a future post.)
Things like journal impact factors are not meant to be arbiters of scientific quality. They were developed by librarians to aid individual libraries in deciding which journals to subscribe to. There are other well-documented problems, including that the magnitude of the impact factor can be wildly sensitive to characteristics of individual subfields, so they make pretty unreliable metrics.
I've just spent a bunch of time preparing a manuscript for a high-impact journal and the more I've thought about it, the less I think publishing in high impact papers is a good metric of success, or even a good scientific practice. High-impact journals often have really strict, tight word limits. Some restrict the number of citations of other work that you can include in your submission. Others just count the number of words in your reference section against your total word count. The result is that the highest-impact journals reward scientists for crafting best, cleanest, and most exciting "story", whether or not the paper includes the most carefully considered explanation of the data*. There are no words to spare for qualifications and caveats. And the limited number of citations and extreme preference of these journals' editors for novelty (inadvertently) encourages scientists not to cite the others in their field who did the necessary foundational work upon which their paper is based. We're all standing on the shoulders of, if not giants**, those who came before us. Not acknowledging that makes the whole process of science fall apart.
And prestige is a positive feedback loop. If you're lucky enough to publish in one of these elite journals, it can have an extremely positive impact on your likelihood of getting an assistant professorship (the ultimate goal for many) -- and if I can extend my evolution metaphor, substantially improves your academic reproductive success because you gain access to a vehicle for training students, which also increases your productivity. But publishing in these high-impact journals also increases the likelihood that people read and cite your other papers, the likelihood that your next grant gets funded, the likelihood that you get invited to give a public lecture in a conspicuous slot at a scientific meeting, etc. Success breeds success. And if you don't succeed, you fail.
__
*I should note that I'm very proud of the work I mentioned that I submitted to a high impact journal, and that I assume that other scientists are acting in good faith and not necessarily acting to exploit this system. But I'm also, again, an evolutionary biologist. Cheaters will invade any system, and manipulation is a behavior that evolves repeatedly.
**Not "giants" because that word reinforces the idea that science is done by once-in-a-generation geniuses and not by regular people who are just deeply committed to incremental progress by hard work.
The weird thing is that the structure of academia precludes partial success. For every PhD who wants to stay in academic science, there is only one goal: to be a professor. If you don't manage to get hired as an assistant professor, that feels like failure. And after between 5 and 7 years spent earning a PhD and between 3 and 5 years of postdocing, during which you may move around once per year. All the moving comes at great financial expense and tremendous cost to one's personal life. For instance, you can give up hope of living near your family. If you have a spouse, they have to change jobs as regularly as you do. And if you have kids, they're changing schools, daycares, and social networks on, sometimes, only a few weeks' notice. That's a lot of drawbacks. I think the benefits outweigh the costs, which is why I'm still interested in academia, but the benefits need to be discounted by the probability of success.
At research universities that train PhDs, though, each professor is likely to be training at least 3-5 PhD students at any one time in their career. If professors work between, say, 35 (not an unreasonable age to be hired as an assistant professor these days, though it seems like it's getting older all the time) and 70 years of age (at which time they may be relatively likely to stop training students, if not likely to stop doing research), and if a PhD in my field (in the US) takes on average 6 years to complete, a single professor could easily generate around 35 "replacements".
This is not feasible. Some of the pressure is relieved by the need for professors at institutions that do not train PhD students -- institutions that grant only bachelors and masters degrees nevertheless have professors. And it is also true that not everyone who earns a PhD wants to be an academic (and that's great!). But the rest of us continue to take jobs that are probably underpaid relative to our skillsets and accept that we'll likely live in constant stress about our job prospects for, on average 22-35 -- or however long it takes you to get from the beginning of your PhD to tenured professor.
In the meantime, many people leave at each stage. This is probably death, in my evolution metaphor. Some PhDs age out of the system as postdocs. There is a maximum amount of time in which you must finish your PhD, and if you fail to finish, or fail to find a postdoc, you're out. In many countries there's actually a maximum number of years you can be a postdoc before you are no longer allowed to accept such work. Sometimes, this is meant to be a pro-labor measure, in that if salaries for postdocs are lower than for those PhD-holders with permanent positions, it protects workers from being exploited by universities who are purportedly paying them with experience. But taking a non-postdoc job to work as a lab tech is the end of the academic road.
As an evolutionary biologist, I can tell you that one of the necessary things for evolution to happen is that not all individuals are equally successful. Check. The second thing is variation among individuals that influences their success. There's definitely a bunch of that good variation for selection to act upon. So here we are, in a system that selects for the most productive scientists. And the field is evolving in response to that selection. People are publishing more collaborative papers than they used to. For instance, when my PhD adviser finished, it was common for all PhD students to publish single-author papers for their entire thesis -- now, you are expected to include your PhD adviser on all your publications. People are being hired at much later stages -- it used to be possible to land an assistant professorship without having published a single paper, which really blows my mind. You might be asking yourself if variation among scientists in productivity is heritable, and I might argue that it is, in the sense that having a productive adviser at a prestigious institution opens doors for you and allows you to be way more productive.
Okay, so you get rewarded for being productive and punished for being less so. That's okay, right? It's a meritocracy. How do we measure productivity, though? As scientists, we're extremely concerned with how we can quantify multidimensional traits, so we've given this some thought. We'd like something quantitative. But our metrics are fairly imprecise for how to measure a great scientist. How many publications do you put out each year? Were they in high-impact journals, of the kind that are widely read and disproportionately likely to be cited? How many times per year are those published papers cited by other scientists? Who are the other scientists that are citing you? How many grant dollars did you receive? How many students who worked with you graduated? These are our best metrics by which to judge the quality of scientific output.
(There are a lot of problems with this idea related to access to resources, which intersect with various aspects of privilege, but that's a topic for a future post.)
Things like journal impact factors are not meant to be arbiters of scientific quality. They were developed by librarians to aid individual libraries in deciding which journals to subscribe to. There are other well-documented problems, including that the magnitude of the impact factor can be wildly sensitive to characteristics of individual subfields, so they make pretty unreliable metrics.
I've just spent a bunch of time preparing a manuscript for a high-impact journal and the more I've thought about it, the less I think publishing in high impact papers is a good metric of success, or even a good scientific practice. High-impact journals often have really strict, tight word limits. Some restrict the number of citations of other work that you can include in your submission. Others just count the number of words in your reference section against your total word count. The result is that the highest-impact journals reward scientists for crafting best, cleanest, and most exciting "story", whether or not the paper includes the most carefully considered explanation of the data*. There are no words to spare for qualifications and caveats. And the limited number of citations and extreme preference of these journals' editors for novelty (inadvertently) encourages scientists not to cite the others in their field who did the necessary foundational work upon which their paper is based. We're all standing on the shoulders of, if not giants**, those who came before us. Not acknowledging that makes the whole process of science fall apart.
And prestige is a positive feedback loop. If you're lucky enough to publish in one of these elite journals, it can have an extremely positive impact on your likelihood of getting an assistant professorship (the ultimate goal for many) -- and if I can extend my evolution metaphor, substantially improves your academic reproductive success because you gain access to a vehicle for training students, which also increases your productivity. But publishing in these high-impact journals also increases the likelihood that people read and cite your other papers, the likelihood that your next grant gets funded, the likelihood that you get invited to give a public lecture in a conspicuous slot at a scientific meeting, etc. Success breeds success. And if you don't succeed, you fail.
__
*I should note that I'm very proud of the work I mentioned that I submitted to a high impact journal, and that I assume that other scientists are acting in good faith and not necessarily acting to exploit this system. But I'm also, again, an evolutionary biologist. Cheaters will invade any system, and manipulation is a behavior that evolves repeatedly.
**Not "giants" because that word reinforces the idea that science is done by once-in-a-generation geniuses and not by regular people who are just deeply committed to incremental progress by hard work.
Comments
Post a Comment