You really shouldn't start a PhD without doing the PhD simulator first [0]. Sadly, people who haven't yet been through the PhD experience think that this game is exaggerating for comic effect. It really isn't.
That simulator hits home. The discouragement and anguish was real.
I finished my Ph.D. in 6 years rather than the usual 4 1/2 because ideas just didn't work out. My topic was much harder than those of my peers.
That said, I felt the 2 extra years I spent made me a much more solid researcher in my narrow field, because I spent more time learning and relearning the foundations of my craft.
I relate to what Winston Churchill said about being a dunce at school (who later become a incomparable wartime orator distinguished by his use of simple English):
"By being so long in the lowest form I gained an immense advantage over the cleverer boys... I got into my bones the essential structure of the ordinary British sentence–which is a noble thing. Naturally I am biased in favor of boys learning English; I would make them all learn English: and then I would let the clever ones learn Latin as an honor, and Greek as a treat."
During my PhD I realized the longer you can extend it, the better (if you intend to stay in academia). If you do 6 years, you'll have the experience of someone who has already done one post doc, but you'll have easier time applying for grants since you're judged as a fresh PhD (possibly with a CV much better than the other applicants due to your extra time).
Wow, I lasted exactly as long in the simulator as I did in real life, with many of the exact same circumstances (less a global pandemic and family tragedy plunging the hope meter into the negatives).
That was fun. I earned my PhD in 5 years 11 months. Got so engrossed I didn't even notice the hope meter but finished with a 56. That seems like a long time but I also enjoyed reading all those papers.
I got it in 6 years and seven months and got filthy rich on PhD coin. Bear in mind, I think not being a PhD actually made the game MUCH easier for me because I had no empathy for the character and approached it as semi-predictable inputs and outputs.
Many CS PhDs work on a small piece of a larger funded project. You work normal hours, read papers in the evening, take Sunday off. It’s not a big deal. So many naive students think they are going to solve AGI. No, you’ll publish an insignificant twist on a loss function and get a job.
Seems accurate. I had 1 conference paper after 1 year, same as during my actual PhD. The expiration of the initial enthusiasm modifier was brutal indeed. I dropped out at that point, but there is no such option in the simulator (perhaps it shows up only once the hope drops low enough)?
My big issue with the PhD is that it was designed to treat me as an employee in exchange for an annual salary equal to 1-2 months worth of earnings as a software freelancer. But the work was interesting. So I wondered why not be an amateur researcher instead.. Of course once I quit, real life intervened and I did little of substance during the following couple of years. I neglected the value of the focus that doing a formal program facilitates. However, I had some ideas recently and was able to establish a dialogue with a relevant research group, so the whole idea may work out after all.
Neat, my personal experience was that there were way more opportunities to submit conference abstracts and you needed little more than an idea, you can figure out the actual presentation by the time the conference actually happens. You also can get rejected a lot more times before you get accepted, if you submit to high impact journals.
For non-research related stuff, that's always my approach :-) My general flow seems to be: genesis of idea, proposal, acceptance, panic, coalescence, creation, (usually) decent delivery.
The simulator didn't really mirror my experience but I do think folks should think twice about going to grad school. Economically, it is often not the best decision... plus you spend the best years of your life with less personal freedom (or financial ability) for long vacations/adventures, etc.
Going to grad school was the wrong economic decision for me... and perhaps the wrong 'life' decision as well.
i think this depends on what the alternative career/job would be compared to grad school.
if you had the option of getting into a well-payed, cushy tech job grad school would result in less personal/financial freedom.
if (like me) you didn't have that option out of undergrad, grad school was comparatively a period of great freedom.
* i made enough on research/TA stipends that i lived a slightly-fancier-than-my-undergrad lifestyle that wasn't too far behind what my classmates that became teachers were living.
* TONS of freedom with respect to how I wanted to work and having full control of my schedule
I feel like grad school gave me a pretty idealistic way to spend my mid 20s. And (luckily) in that time I was able to develop enough skills that I could jump into one of those high-paying, cushy tech jobs when it came time to realize that academica sucks and I wanted to leave.
>i think this depends on what the alternative career/job would be compared to grad school.
This, the economy was shit when I graduated, I wasn't interested in a phd, but strongly considered getting a masters, and would have likely been financially better off had I stayed in school those extra two years instead of graduating into a horrible job market and losing that fresh graduate advantage when applying for jobs once it finally got moving again. In retrospect, I would have much rather lived a college lifestyle and did research/ta type stuff instead of doing the tech support type jobs I ended up with to make ends meet.
I know someone who was in a PhD program for 10 years in Europe, without graduating. After year 6, it got very difficult for the university to legally employ him, and after year 9, it was simply impossible and he lived on his savings, but was still allowed to use his old office.
European PhD programs generally assume you already have a Masters in the subject so they skip most of the grad level courses and they typically don't do things like rotations either. You have a lab and some project to start on day 1. It's still overall faster to finish in Europe but the difference isn't as extreme as it sounds, assuming you do go for the Masters first.
Australia too. But difference is you at least get paid (albeit a small stipend) in our countries, I think US the student pays tuition or has to also undertake seminars etc.
My doctorate was essentially; run studies for 2-3yrs, write up papers for submission, smash them together with an intro and general discussion, graduate.
Same, basically. In fact, in the UK the requirements were low enough that after my first paper published a year into my PhD (in theoretical quantum physics) my supervisor was like: "I consider this enough for graduation, now you can work on whatever you are interested in"
I liked the aspect of it where you get an idea, and then work on it to get a result, and then a major result, supporting figures etc.; and while this rigidity in finishing the ideas one starts may not be optimal there's at least an element of it which could motivate people to finish things first, before moving on.
I think this is often useful. Maybe it's obvious, but it can be very tempting to develop ideas, or develop new shiny results, when you still have other ideas that haven't yet been turned into definite packages of well-supported results.
I agree, but my estimate 90% of my peers' PhDs (including mine) were essentially 3 years of work on an idea that did not work, or trying and failing to find an idea that did work. Then writing up your work as if it was great. The 10% of "successful" PhDs were on ideas that were almost guaranteed to work - more development than research.
This was in a proper hard engineering field though. I think in other fields can be much more likely to be things that can't really fail. For example in computer science, a lot of PhDs are just like "I implemented this thing" where there's very little risk of it simply not working.
An exception in computing is AI research where it is very much like the "try some stuff; it didn't work" experience of engineering and science research. I imagine a PhD in AI is not a fun experience...
In hindsight it should've been a huge red flag to me when the only "what's it like to do a Ph. D" students at the induction were comp. sci.
My advice to most people would just be "don't". My second run of advice would be "find the most boring project imaginable" since it's likely to succeed on the basis of "do a bunch of fairly predictable experiments and publish them".
I don't think a lot of commenters here have realized that the author was doing a PhD in Europe, not the US. The dynamics are very different. In many European countries, you have funding for only 3 years. You either finish or you're out (typically no coursework, though). I've known cases where after 3 years they simply lose access to the buildings.
As a result, the criteria for graduation is often softer than in the US.
> As a result, the criteria for graduation is often softer than in the US.
Also bear in mind in Europe you need a masters before entering a PhD program (or significant research experience), and you spend 100% of your time on your research, there's not courses or teaching. I think that better explains why they're finished in 3-4 years.
"Europe" is not a useful description of the location in this case. There are too many differences between countries, universities and even faculties within universities
Someone should sit every new PhD student down and explain to them that the goal isn't (or shouldn't be) getting publications or making an impact or getting a credential or a job. Of course you have to keep an eye on those things for practical reasons, but you can't lose sight of the goal which is to improve your understanding. Not even "our" understanding, just your own. If you pursue that and have a bit of talent (and luck) you will eventually push your understanding past that of the community on a given topic and you will write it up so that they may improve their understanding by reading your work.
4 first author pubs? Maybe this is field dependent but that sounds like a pretty extreme requirement. The most I've seen formally required in my field is 1 first author peer reviewed pub, but many programs don't even set that as an official requirement anymore.
Makes sense, it'd still be kinda harsh as a strict graduation requirement for EECS though wouldn't it?
In biology I've seen a lot of more recent theses where a couple of the chapters are from middle author works with a bit of extra context on what the student did, and only 1 first author chapter. But besides being slower to do generally, bio still has a lot of people that don't believe in shared first authorship. Sometimes what is technically a second author chapter was pretty close to 50/50 in what the student actually contributed.
That's sad. That makes your piece of paper worth even less, in my opinion. How is that different from a degree mill. By the way I have my piece of paper but what I value (and what my colleagues value) is my understanding
That's certainly the systemic pressure. And it's hard to avoid having that approach as a postdoc. I think PhD students, especially in the first couple years, can benefit in the long term from OC's advice if they have a supportive department. Even if the long term goal is an academic career, but perhaps moreso if not. Faculty like to think of PhD students as employees but really they are trainees, and the primary goal should be strong capability to do good independent research by the end of the program. Immediately pushing for pubs is often not the best way to gain an independent research vision.
Nice story about perseverance. But I do not know how helpful this is to young researchers.
My advice for young researchers is read more articles - like at least a solid month of reading and journaling full time (40 hrs per week) before you even start to think about what you want to start off with by replicating.
The other related mistake I see young researchers make a lot is not leveraging pre-existing work / results and wasting weeks or months reinventing the wheel.
This is very useful for laboratory science but not too much for computer science research for example.
You can typically make nice prototypes in a weekend and try out ideas. If one or two of these work out, you get a paper after a bit more work polishing it.
Of course you have to be competent and able to build prototypes, which is something I have seen about half of the students lack. In that case, you can typically also get a PhD by basically doing minor tweaks on existing tools and putting a lot more effort in benchmarking and story telling. Both are useful to science.
This is why you have more than one project. And also backup projects for those projects. Every mature lab should have a simple turn the crank project for each student that's got the biochemistry worked out, and is just a matter of collecting observations and doing lengthy data analysis. Those are the lifesavers when the ambitious ones fall through.
I must have been involved in ~20 projects in my PhD. Only 3-4 will ever be published.
This exactly. I filled hundreds of pages of behavioral data during my PhD. Although I don’t know the exact numbers, I’d wager that my publications came from only 20-25 pages of data.
If you anticipate low success rates, you can find success either through luck or through diligent experimentation with an expectation of null results. Null results which, by the way, can often improve your odds of downstream success.
Every lab usually has a few “I’ll bet it’ll work but I don’t have time to find out” projects.
But that’s actually a big part of doing independent research - using your time the most efficiently, so students need to get to the point where they have a dozen or so paths and they plan out the next year or two thinking about the best way to tackle them and “fail fast”.
They make a good point: By focusing only on positive results in publication, there is little common knowledge of pathways that don’t succeed. Back when I was doing grad studies in math, I remember thinking that this could be a useful thing in that field as well: compiling plausible pathways towards proving theorems that end up to be dead ends. I almost feel like something like this could in its own way spark new discoveries in mathematics.
I would be interested to hear more about how much they were discussing their failures/challenges with others. It sounds like this might be a case of, "I finally talked to someone about it and they unblocked me in an hour."
When I was in grad school I was very hesitant to ask others for help or feedback. Big mistake! I see similar things with interns: they'll wait until a daily or weekly check-in to raise problems.
My approach now is to set a time budget; if I don't figure it out myself within X hours, then I have to ask someone for help.
I think some of this is fine, but doesn’t match the pressure people put on themselves in Europe, and the expectations people put are put under in the US.
For me in the U.K. there were 7 chapters to my thesis that I worked on for roughly 3.5 years. Two of those were “failures” in that they weren’t publishable as positive results but the others worked OK. One was an algorithm paper and I published the source code to GitHub because while it wasn’t useful to me (slower than the competing algorithm we already used) it was applicable in more general situations and it has proved to be so as people have emailed me about it at various times and it’s been cited.
That's not really true. They don't have to finish and graduate by the time funding runs out - nearly half of students simply fail to graduate (in the us).
It's also extra untrue because this author is describing running into a null result and turning that into a PHD when usually null results don't go anywhere, either to a thesis or a publication.
But it becomes true again in the context that the null result is nice framing for this article but isn't the framing in the intro/contents of the thesis. And also because:
"A turning point came when another graduate student suggested a dramatic change to my protocol. I was skeptical, but I thought it was worth a shot. It turned out they were correct: After trying yet another experiment, the results started to look better—and after a few more changes, I eventually got the protocol to work. "
Oddly enough, my PhD was also saved by a fellow grad student suggesting a different experiment that used the same equipment. I was ready to drop out and apply to the engineering school.
well, also luck. Plenty of people have a PI change which can be devastating especially on a visa, or wouldn't have gotten the recommendation for the other protocol to get it to work, and also for the current crop - giant global pandemic shutting down all the labs - did you pick a topic that's essential research/not lab based? And strategy. If you chose a field based on passion instead of practical considerations like grant availability - life might be more complex in infectious disease research than in oncology lol. And academia politics. Though for some obviously politics is a plus
Can you explain how you would determine or know about grant availability?
Especially that "life might be more complex in infectious disease research than in oncology" - as an outsider this really caught my interest, I'd have imagined the two have similar levels of grants? Maybe more than a field like Psychology or Developmental Biology, about the same?
Does this mean more people are going into in disease research because A) it's more interesting to them? (vs Oncology)... or B) The grants from COVID mean it's a better field?
You can see NCI (National Cancer Institute) leading the pack there, with NIAID (National Institute of Allergy and Infectious Diseases) coming in second.
It surprises me how few people, even in my field(s) (which are overwhelmingly funded primarily by NIH), know of or use these tools. Particularly RePORTER and the advanced project search functions.
That's part of academic politics. Navigating grant funding is arguably one of the most important skills that a PHD program is supposed to teach given how important that is to the process of research.
Oncology has way way way more funding because there's a much bigger market for it - infectious diseases primarily affect poor people who aren't Americans so not much money in it. Even at the peak of covid I remember advisors warning phd applicants passionate about infectious diseases that the funding surge wasn't going to last and to be careful (which was right). Consider how moderna was originally founded as a vaccine company before almost completely shifting focus to oncology because that's where the money was. Obviously they swung back to vaccines when the pandemic showed up but now they're back focused on oncology.
Funding priorities are also set by the government - there's the cancer moonshot etc which means tons and tons of cancer funding. For universities new cancer treatments are quite valuable IP. And so on. Within a field it becomes very obvious where the funding and momentum is - what are the buzzwords everyone is putting in their grants and papers (nanotech? AI?), what types of publications are getting published int he big journals. Gene therapy came out of its winter after luxturna for example. Car-t was booming after the first approval there, entire departments were set up for it. Mrna stuff is super hot right now, but one of the researchers behind the core technology (that made her university like a billion dollars in royalties) was denied tenure and pushed out of academia when she originally developed the technology because it was considered not useful. First lab I worked in - the PI was fresh out of his postdoc and got hundreds of thousands of dollars in funding and his own lab and bought his own flow cytometer and everything (absolutely not the normal path lol) because he was a first author on a particularly good CRISPR paper. On one hand - trend chasing and bubbles have their obvious downsides. On the other hand - it does mean a lot of resources get funneled into new interesting areas when something cool pops up. But on the original hand - trend chasing and bubbles, welcome to humans on earth i guess. But yeah in terms of how you know about it - you'll know it's quite obvious who's getting money and tenure.
I don't think the author was trying to position herself as "special". It's a pretty universal experience to try something unkonwn and feeling like a failure. I think the author is trying to convey that it's ok to feel that way.
> This is just a regular PhD experience, but the author has made it sound like some special sort of skill only they and a few others know how to master.
Making a regular experience sound like a special sort of skill only you have is itself a special sort of skill that it's valuable to have if you want to apply for grants.
> How do you effectively apply to PhD grants, is it different to winning say Startup / Innovation grants?
I'm afraid that I don't have the faintest idea, both because my grant activity hasn't been noticably effective (and I held no grants while pursuing my Ph.D.), and because I've never applied for a Startup / Innovation grant.
I don’t think she’s saying “only _I_ can do this!!” - it just read to me like a “this is a what happened and what I learned from it”. Actually, I think I’ll send this article to my high school physics teacher - he was a good mentor, and he might enjoy giving it to his students.
Agree this seems like basic things you would figure out doing anything uncertain. The first blocked out quote of the article is:
“If an experiment doesn’t go as expected, it doesn’t necessarily mean you did something wrong.”
That’s why it's called an experiment… I don’t get why this is insightful.
I guess there is a lesson about carefully noting down failures, in the corporate world handing over a document detailing all the things you tried that failed is much better than just failing silently.
In biochem results aren't "success" or "fail", they're more like "probably success" and "either fail or experimental error". And in research the vast majority of results are in the latter category so it can be hard to discriminate between failures of your hypothesis and failures in your experimental method.
Why a failed experiment needs to be turned into success?
The outcome of the research (experiment) could be both success and failure, right? That's why we study it, experiment with it, we do not know it yet, we want to see if it is as believed or not. The important is to grow the body of knowledge here - and knowing if something does not work the way we thought will is knowledge -, not to pretend being successful, right? For pretending there are countless other (much much better paying) occupations anyway.
It's too easy to get unsuccesful results. (Does listening to Macarena for 8 hours per day cure brain cancer in mice?)
So unsuccesful results are very difficult to publish and to be the base of a thesis. So you must find a twist or secondary result and make it the central part. (Does listening to Macarena for 8 hours per day cause brain damage in mice?)
Also, I think I seen that Macarena article. : ) It suggested to extend the research to primates and Spice Girls and finding songs that cure the damage of Macarena.
This column, which appears as the last page in the print magazine, is generally pretty good and often fun to read (https://www.science.org/topic/careers-overline/working-life). IIRC the stories never end on a down note (perhaps not true to life).
But, taken as a whole, they offer some ways out of the single-track-grindset that some people in the academic system have -- and that the system promotes.
It turns out that there are a lot of stories out there of people who had to give up on something, change fields, recognize their strengths or weaknesses, etc. People don't talk about this stuff as much as they should.
I think the real takeaway is that collaborating/bouncing ideas off other people is the most important part of research. Few people can work isolated on their own without idea exchange.
The mistake that a lot of people make is that they think science is all about proving things.
In actuality, science done right is all about disproving things.
You can spend a lifetime tracking down all the white swans to “prove” that “all swans are white”, but you need to find only one black swan - as the British did when they reached Australia - to disprove that hypothesis.
Science is about disproving things, as it is by far the easier path - and frequently, the only possible path - to take. And by disproving things, we push back the darkness more and more, until what little remains must contain the truth.
And when science “changes it’s mind”, it’s simply science obtaining more data that points better towards that truth, more of the darkness has been pushed back and the old position was eliminated by the new evidence.
Fortunately I was just doing a Masters so the fact that my thesis (in Material Science) ended up mostly just being some basic characterization of an alloy (grain growth, hardness, Xray crystallography) didn't really matter. It was sort of a dead-end and my thesis advisor shortly thereafter dropped research in the area to focus on his other interest where he was very successful (water ice). Things would probably have been a lot uglier had it been a PhD.
Genuine question: What's the motivation for getting a PhD in an engineering field? I’m guessing you just really really love researching and have enough money/low enough lifestyle to take the oppertunity cost financial hit of not working in industry for five years?
This is also a recipe for successful companies. Start/work at a company, notice some adjacent problem you have to do in the course of the work that is annoying/difficult/error prone, smooth out the problem and turn that into a new product.
It was refreshing to read. A research world producing (publishing) mostly success and breakthrough is not creadible anyway. I belive this improves the reputation of the academic world too.
You really shouldn't start a PhD without doing the PhD simulator first [0]. Sadly, people who haven't yet been through the PhD experience think that this game is exaggerating for comic effect. It really isn't.
[0] https://research.wmz.ninja/projects/phd/index.html
That simulator hits home. The discouragement and anguish was real.
I finished my Ph.D. in 6 years rather than the usual 4 1/2 because ideas just didn't work out. My topic was much harder than those of my peers.
That said, I felt the 2 extra years I spent made me a much more solid researcher in my narrow field, because I spent more time learning and relearning the foundations of my craft.
I relate to what Winston Churchill said about being a dunce at school (who later become a incomparable wartime orator distinguished by his use of simple English):
"By being so long in the lowest form I gained an immense advantage over the cleverer boys... I got into my bones the essential structure of the ordinary British sentence–which is a noble thing. Naturally I am biased in favor of boys learning English; I would make them all learn English: and then I would let the clever ones learn Latin as an honor, and Greek as a treat."
During my PhD I realized the longer you can extend it, the better (if you intend to stay in academia). If you do 6 years, you'll have the experience of someone who has already done one post doc, but you'll have easier time applying for grants since you're judged as a fresh PhD (possibly with a CV much better than the other applicants due to your extra time).
My adviser insisted that I extended it as much as I could, but I didn't listen to him :(
If you want a non-simulated memoir of someone's PhD, I highly recommend reading "The PhD Grind" by Philip Guo.. [0]
I read it in my last year of college. It was a page turner, and helped me think about whether I wanted to do a PhD.
[0] https://www.dropbox.com/scl/fi/g8mc1lniyf26opxtrqgna/pguo-Ph...
Wow, I lasted exactly as long in the simulator as I did in real life, with many of the exact same circumstances (less a global pandemic and family tragedy plunging the hope meter into the negatives).
Thought about going back?
That was fun. I earned my PhD in 5 years 11 months. Got so engrossed I didn't even notice the hope meter but finished with a 56. That seems like a long time but I also enjoyed reading all those papers.
I got it in 6 years and seven months and got filthy rich on PhD coin. Bear in mind, I think not being a PhD actually made the game MUCH easier for me because I had no empathy for the character and approached it as semi-predictable inputs and outputs.
> I had no empathy for the character and approached it as semi-predictable inputs and outputs.
I dunno about the "filthy rich" outcome but this strategy is actually fairly common.
Many CS PhDs work on a small piece of a larger funded project. You work normal hours, read papers in the evening, take Sunday off. It’s not a big deal. So many naive students think they are going to solve AGI. No, you’ll publish an insignificant twist on a loss function and get a job.
I beat it in one click.
The only winning move.
That's the most realistic simulator I ever play; and that's not a good thing.
I really enjoyed my PhD (computational physics)
The team was great, the topic was great, I did something really innovative.
I also partied a lot, met great girlfriends, met my wife, made friends.
One of the best times in my life.
> You really shouldn't start a PhD without doing the PhD simulator first.
Or without seeing what Matt Groening thinks of your plan. (Just in case the Ph.D. simulator was not enough to scare you off.)
https://www.pinterest.com/pin/grad-school-so-glad-i-am-done-...
Thanks for sharing this game! It showed me exactly what I'm doing wrong with my current book project (~2.5 years so far).
The winning move is to mine for an idea, then just do that idea and nothing else until you get a paper or reject the idea. Rest when tired.
Got my [simulated] PhD in 5y 4mo with 99/100 hope left by following that algorithm. Perfect teachable moment. Thanks
Seems accurate. I had 1 conference paper after 1 year, same as during my actual PhD. The expiration of the initial enthusiasm modifier was brutal indeed. I dropped out at that point, but there is no such option in the simulator (perhaps it shows up only once the hope drops low enough)?
My big issue with the PhD is that it was designed to treat me as an employee in exchange for an annual salary equal to 1-2 months worth of earnings as a software freelancer. But the work was interesting. So I wondered why not be an amateur researcher instead.. Of course once I quit, real life intervened and I did little of substance during the following couple of years. I neglected the value of the focus that doing a formal program facilitates. However, I had some ideas recently and was able to establish a dialogue with a relevant research group, so the whole idea may work out after all.
Neat, my personal experience was that there were way more opportunities to submit conference abstracts and you needed little more than an idea, you can figure out the actual presentation by the time the conference actually happens. You also can get rejected a lot more times before you get accepted, if you submit to high impact journals.
For non-research related stuff, that's always my approach :-) My general flow seems to be: genesis of idea, proposal, acceptance, panic, coalescence, creation, (usually) decent delivery.
Well, I probably should not have played this, given that I'm (seemingly) in my last year of my PhD.
Back to paper writing...
Don't worry, I was in the last year of my PhD for 3 years.
The simulator didn't really mirror my experience but I do think folks should think twice about going to grad school. Economically, it is often not the best decision... plus you spend the best years of your life with less personal freedom (or financial ability) for long vacations/adventures, etc.
Going to grad school was the wrong economic decision for me... and perhaps the wrong 'life' decision as well.
i think this depends on what the alternative career/job would be compared to grad school.
if you had the option of getting into a well-payed, cushy tech job grad school would result in less personal/financial freedom.
if (like me) you didn't have that option out of undergrad, grad school was comparatively a period of great freedom. * i made enough on research/TA stipends that i lived a slightly-fancier-than-my-undergrad lifestyle that wasn't too far behind what my classmates that became teachers were living. * TONS of freedom with respect to how I wanted to work and having full control of my schedule
I feel like grad school gave me a pretty idealistic way to spend my mid 20s. And (luckily) in that time I was able to develop enough skills that I could jump into one of those high-paying, cushy tech jobs when it came time to realize that academica sucks and I wanted to leave.
>i think this depends on what the alternative career/job would be compared to grad school.
This, the economy was shit when I graduated, I wasn't interested in a phd, but strongly considered getting a masters, and would have likely been financially better off had I stayed in school those extra two years instead of graduating into a horrible job market and losing that fresh graduate advantage when applying for jobs once it finally got moving again. In retrospect, I would have much rather lived a college lifestyle and did research/ta type stuff instead of doing the tech support type jobs I ended up with to make ends meet.
Discussed on HN here: https://news.ycombinator.com/item?id=36597534
Took me 7 years to earn my PhD, but I did publish 4 papers (ex conference papers) instead of 2 :)
5 years and 2 months. I assume this is about average?
In real life? In North America it depends. For some programs it is. For Europe? I think they more reasonably kick everyone out the door in 3 years.
I know someone who was in a PhD program for 10 years in Europe, without graduating. After year 6, it got very difficult for the university to legally employ him, and after year 9, it was simply impossible and he lived on his savings, but was still allowed to use his old office.
European PhD programs generally assume you already have a Masters in the subject so they skip most of the grad level courses and they typically don't do things like rotations either. You have a lab and some project to start on day 1. It's still overall faster to finish in Europe but the difference isn't as extreme as it sounds, assuming you do go for the Masters first.
In the UK it’s 4 years but you stop getting paid after 3 and a half.
Australia too. But difference is you at least get paid (albeit a small stipend) in our countries, I think US the student pays tuition or has to also undertake seminars etc.
My doctorate was essentially; run studies for 2-3yrs, write up papers for submission, smash them together with an intro and general discussion, graduate.
Same, basically. In fact, in the UK the requirements were low enough that after my first paper published a year into my PhD (in theoretical quantum physics) my supervisor was like: "I consider this enough for graduation, now you can work on whatever you are interested in"
Why did I get downvoted for this?? This website is absurd
1 year 1 month for me
The first try I failed after 7 years while the second time I finished it in 5y4m with 99 hope. Hmm interesting
its more amusing than accurate, IMO
I liked the aspect of it where you get an idea, and then work on it to get a result, and then a major result, supporting figures etc.; and while this rigidity in finishing the ideas one starts may not be optimal there's at least an element of it which could motivate people to finish things first, before moving on.
I think this is often useful. Maybe it's obvious, but it can be very tempting to develop ideas, or develop new shiny results, when you still have other ideas that haven't yet been turned into definite packages of well-supported results.
In all seriousness, I think you're lucky then.
(I can only speak for my experience and those of my peers in my field, at the end of the day)
I agree, but my estimate 90% of my peers' PhDs (including mine) were essentially 3 years of work on an idea that did not work, or trying and failing to find an idea that did work. Then writing up your work as if it was great. The 10% of "successful" PhDs were on ideas that were almost guaranteed to work - more development than research.
This was in a proper hard engineering field though. I think in other fields can be much more likely to be things that can't really fail. For example in computer science, a lot of PhDs are just like "I implemented this thing" where there's very little risk of it simply not working.
An exception in computing is AI research where it is very much like the "try some stuff; it didn't work" experience of engineering and science research. I imagine a PhD in AI is not a fun experience...
In hindsight it should've been a huge red flag to me when the only "what's it like to do a Ph. D" students at the induction were comp. sci.
My advice to most people would just be "don't". My second run of advice would be "find the most boring project imaginable" since it's likely to succeed on the basis of "do a bunch of fairly predictable experiments and publish them".
Had to call it quits after year 3 to get back to work
Desert Bus is slightly more realistic
I don't think a lot of commenters here have realized that the author was doing a PhD in Europe, not the US. The dynamics are very different. In many European countries, you have funding for only 3 years. You either finish or you're out (typically no coursework, though). I've known cases where after 3 years they simply lose access to the buildings.
As a result, the criteria for graduation is often softer than in the US.
> As a result, the criteria for graduation is often softer than in the US.
Also bear in mind in Europe you need a masters before entering a PhD program (or significant research experience), and you spend 100% of your time on your research, there's not courses or teaching. I think that better explains why they're finished in 3-4 years.
> not courses or teaching
That's not true for maths at U of Tromso nor true for (as far as I know) all of Hungary.
"Europe" is not a useful description of the location in this case. There are too many differences between countries, universities and even faculties within universities
Its in Germany, here an engineering PhD on average takes 5 years not including masters.
Yeah, but science PhD is 3 years.
It's very variable. 3 years is going to be the shortest duration, but it can be much longer.
Yep, the dynamics of PhD programs in Europe and the US are shaped by different funding structures
Someone should sit every new PhD student down and explain to them that the goal isn't (or shouldn't be) getting publications or making an impact or getting a credential or a job. Of course you have to keep an eye on those things for practical reasons, but you can't lose sight of the goal which is to improve your understanding. Not even "our" understanding, just your own. If you pursue that and have a bit of talent (and luck) you will eventually push your understanding past that of the community on a given topic and you will write it up so that they may improve their understanding by reading your work.
When I started my PhD, they were clear: four publications is a degree. That's the goal. Understanding is a bonus but doesn't get the title!
4 first author pubs? Maybe this is field dependent but that sounds like a pretty extreme requirement. The most I've seen formally required in my field is 1 first author peer reviewed pub, but many programs don't even set that as an official requirement anymore.
EE & CS aim for that many. I've heard maths grads aim for 1.
Makes sense, it'd still be kinda harsh as a strict graduation requirement for EECS though wouldn't it?
In biology I've seen a lot of more recent theses where a couple of the chapters are from middle author works with a bit of extra context on what the student did, and only 1 first author chapter. But besides being slower to do generally, bio still has a lot of people that don't believe in shared first authorship. Sometimes what is technically a second author chapter was pretty close to 50/50 in what the student actually contributed.
Four first author publications, yes. University guidelines say three, but in practice they'd like to see four or five to get it approved.
Yeah, it wildly depends on the maturity of the field it seems
That's sad. That makes your piece of paper worth even less, in my opinion. How is that different from a degree mill. By the way I have my piece of paper but what I value (and what my colleagues value) is my understanding
My goal is explicitly to publish, and I have the impression this is the case on a national level.
That's certainly the systemic pressure. And it's hard to avoid having that approach as a postdoc. I think PhD students, especially in the first couple years, can benefit in the long term from OC's advice if they have a supportive department. Even if the long term goal is an academic career, but perhaps moreso if not. Faculty like to think of PhD students as employees but really they are trainees, and the primary goal should be strong capability to do good independent research by the end of the program. Immediately pushing for pubs is often not the best way to gain an independent research vision.
> Basically, I’m not interested in doing research and I never have been, I’m interested in understanding, which is quite a different thing.
―David Blackwell
If not for the immense pressure to publish or perish...
Tell that to the advisors
I wish this was clear to me before jumping into that ship. Fast forward few years later, I wish I knew.
If you're curious, here's her PhD thesis. (It's dual-language German/English): https://publishup.uni-potsdam.de/opus4-ubp/frontdoor/deliver...
Only the abstract/Zusammenfassung is dual-language, as it is usual with German dissertations written in English.
Nice story about perseverance. But I do not know how helpful this is to young researchers.
My advice for young researchers is read more articles - like at least a solid month of reading and journaling full time (40 hrs per week) before you even start to think about what you want to start off with by replicating.
The other related mistake I see young researchers make a lot is not leveraging pre-existing work / results and wasting weeks or months reinventing the wheel.
This is very useful for laboratory science but not too much for computer science research for example.
You can typically make nice prototypes in a weekend and try out ideas. If one or two of these work out, you get a paper after a bit more work polishing it.
Of course you have to be competent and able to build prototypes, which is something I have seen about half of the students lack. In that case, you can typically also get a PhD by basically doing minor tweaks on existing tools and putting a lot more effort in benchmarking and story telling. Both are useful to science.
I love this advice. But, I find it extremely difficult to just read and jot things down without having any applied work in parallel.
Your idea may be the ideal case :]
This is why you have more than one project. And also backup projects for those projects. Every mature lab should have a simple turn the crank project for each student that's got the biochemistry worked out, and is just a matter of collecting observations and doing lengthy data analysis. Those are the lifesavers when the ambitious ones fall through.
I must have been involved in ~20 projects in my PhD. Only 3-4 will ever be published.
This exactly. I filled hundreds of pages of behavioral data during my PhD. Although I don’t know the exact numbers, I’d wager that my publications came from only 20-25 pages of data.
If you anticipate low success rates, you can find success either through luck or through diligent experimentation with an expectation of null results. Null results which, by the way, can often improve your odds of downstream success.
Exactly.
Every lab usually has a few “I’ll bet it’ll work but I don’t have time to find out” projects.
But that’s actually a big part of doing independent research - using your time the most efficiently, so students need to get to the point where they have a dozen or so paths and they plan out the next year or two thinking about the best way to tackle them and “fail fast”.
There was a story recently on Science Friday about journals beginning to publish results of experiments that did not prove the hypothesis: https://www.sciencefriday.com/segments/journal-of-trial-erro...
They make a good point: By focusing only on positive results in publication, there is little common knowledge of pathways that don’t succeed. Back when I was doing grad studies in math, I remember thinking that this could be a useful thing in that field as well: compiling plausible pathways towards proving theorems that end up to be dead ends. I almost feel like something like this could in its own way spark new discoveries in mathematics.
I would be interested to hear more about how much they were discussing their failures/challenges with others. It sounds like this might be a case of, "I finally talked to someone about it and they unblocked me in an hour."
When I was in grad school I was very hesitant to ask others for help or feedback. Big mistake! I see similar things with interns: they'll wait until a daily or weekly check-in to raise problems.
My approach now is to set a time budget; if I don't figure it out myself within X hours, then I have to ask someone for help.
How do you do this if you are mostly working remotely though? I am considering a more lab based PhD precisely because of this.
It is definitely harder to grow this muscle when you are remote. Be biased towards hopping on a call/zoom rather than messaging on slack, etc.
I think some of this is fine, but doesn’t match the pressure people put on themselves in Europe, and the expectations people put are put under in the US.
For me in the U.K. there were 7 chapters to my thesis that I worked on for roughly 3.5 years. Two of those were “failures” in that they weren’t publishable as positive results but the others worked OK. One was an algorithm paper and I published the source code to GitHub because while it wasn’t useful to me (slower than the competing algorithm we already used) it was applicable in more general situations and it has proved to be so as people have emailed me about it at various times and it’s been cited.
This is just a regular PhD experience, but the author has made it sound like some special sort of skill only they and a few others know how to master.
Everyone has to turn failed experiments into a successful PhD because they have to finish and graduate by the time funding runs out.
That's not really true. They don't have to finish and graduate by the time funding runs out - nearly half of students simply fail to graduate (in the us).
It's also extra untrue because this author is describing running into a null result and turning that into a PHD when usually null results don't go anywhere, either to a thesis or a publication.
But it becomes true again in the context that the null result is nice framing for this article but isn't the framing in the intro/contents of the thesis. And also because:
"A turning point came when another graduate student suggested a dramatic change to my protocol. I was skeptical, but I thought it was worth a shot. It turned out they were correct: After trying yet another experiment, the results started to look better—and after a few more changes, I eventually got the protocol to work. "
doesn't sound like a failed experiment to me!
Oddly enough, my PhD was also saved by a fellow grad student suggesting a different experiment that used the same equipment. I was ready to drop out and apply to the engineering school.
PhDs are more about perseverance than anything else, as highlighted in this story.
well, also luck. Plenty of people have a PI change which can be devastating especially on a visa, or wouldn't have gotten the recommendation for the other protocol to get it to work, and also for the current crop - giant global pandemic shutting down all the labs - did you pick a topic that's essential research/not lab based? And strategy. If you chose a field based on passion instead of practical considerations like grant availability - life might be more complex in infectious disease research than in oncology lol. And academia politics. Though for some obviously politics is a plus
Can you explain how you would determine or know about grant availability?
Especially that "life might be more complex in infectious disease research than in oncology" - as an outsider this really caught my interest, I'd have imagined the two have similar levels of grants? Maybe more than a field like Psychology or Developmental Biology, about the same?
Does this mean more people are going into in disease research because A) it's more interesting to them? (vs Oncology)... or B) The grants from COVID mean it's a better field?
https://report.nih.gov/
You can see NCI (National Cancer Institute) leading the pack there, with NIAID (National Institute of Allergy and Infectious Diseases) coming in second.
It surprises me how few people, even in my field(s) (which are overwhelmingly funded primarily by NIH), know of or use these tools. Particularly RePORTER and the advanced project search functions.
That's part of academic politics. Navigating grant funding is arguably one of the most important skills that a PHD program is supposed to teach given how important that is to the process of research.
Oncology has way way way more funding because there's a much bigger market for it - infectious diseases primarily affect poor people who aren't Americans so not much money in it. Even at the peak of covid I remember advisors warning phd applicants passionate about infectious diseases that the funding surge wasn't going to last and to be careful (which was right). Consider how moderna was originally founded as a vaccine company before almost completely shifting focus to oncology because that's where the money was. Obviously they swung back to vaccines when the pandemic showed up but now they're back focused on oncology.
Funding priorities are also set by the government - there's the cancer moonshot etc which means tons and tons of cancer funding. For universities new cancer treatments are quite valuable IP. And so on. Within a field it becomes very obvious where the funding and momentum is - what are the buzzwords everyone is putting in their grants and papers (nanotech? AI?), what types of publications are getting published int he big journals. Gene therapy came out of its winter after luxturna for example. Car-t was booming after the first approval there, entire departments were set up for it. Mrna stuff is super hot right now, but one of the researchers behind the core technology (that made her university like a billion dollars in royalties) was denied tenure and pushed out of academia when she originally developed the technology because it was considered not useful. First lab I worked in - the PI was fresh out of his postdoc and got hundreds of thousands of dollars in funding and his own lab and bought his own flow cytometer and everything (absolutely not the normal path lol) because he was a first author on a particularly good CRISPR paper. On one hand - trend chasing and bubbles have their obvious downsides. On the other hand - it does mean a lot of resources get funneled into new interesting areas when something cool pops up. But on the original hand - trend chasing and bubbles, welcome to humans on earth i guess. But yeah in terms of how you know about it - you'll know it's quite obvious who's getting money and tenure.
Infectious disease in particular is going to have a really hard time. No market for it, no industry demand. Investments there are already winding down in general: https://www.fiercebiotech.com/biotech/johnson-and-johnson-sh....
I don't think the author was trying to position herself as "special". It's a pretty universal experience to try something unkonwn and feeling like a failure. I think the author is trying to convey that it's ok to feel that way.
> This is just a regular PhD experience, but the author has made it sound like some special sort of skill only they and a few others know how to master.
Making a regular experience sound like a special sort of skill only you have is itself a special sort of skill that it's valuable to have if you want to apply for grants.
How do you effectively apply to PhD grants, is it different to winning say Startup / Innovation grants?
> How do you effectively apply to PhD grants, is it different to winning say Startup / Innovation grants?
I'm afraid that I don't have the faintest idea, both because my grant activity hasn't been noticably effective (and I held no grants while pursuing my Ph.D.), and because I've never applied for a Startup / Innovation grant.
I don’t think she’s saying “only _I_ can do this!!” - it just read to me like a “this is a what happened and what I learned from it”. Actually, I think I’ll send this article to my high school physics teacher - he was a good mentor, and he might enjoy giving it to his students.
Agree this seems like basic things you would figure out doing anything uncertain. The first blocked out quote of the article is:
“If an experiment doesn’t go as expected, it doesn’t necessarily mean you did something wrong.”
That’s why it's called an experiment… I don’t get why this is insightful.
I guess there is a lesson about carefully noting down failures, in the corporate world handing over a document detailing all the things you tried that failed is much better than just failing silently.
In biochem results aren't "success" or "fail", they're more like "probably success" and "either fail or experimental error". And in research the vast majority of results are in the latter category so it can be hard to discriminate between failures of your hypothesis and failures in your experimental method.
Why a failed experiment needs to be turned into success?
The outcome of the research (experiment) could be both success and failure, right? That's why we study it, experiment with it, we do not know it yet, we want to see if it is as believed or not. The important is to grow the body of knowledge here - and knowing if something does not work the way we thought will is knowledge -, not to pretend being successful, right? For pretending there are countless other (much much better paying) occupations anyway.
It's too easy to get unsuccesful results. (Does listening to Macarena for 8 hours per day cure brain cancer in mice?)
So unsuccesful results are very difficult to publish and to be the base of a thesis. So you must find a twist or secondary result and make it the central part. (Does listening to Macarena for 8 hours per day cause brain damage in mice?)
That's true.
Also, I think I seen that Macarena article. : ) It suggested to extend the research to primates and Spice Girls and finding songs that cure the damage of Macarena.
This column, which appears as the last page in the print magazine, is generally pretty good and often fun to read (https://www.science.org/topic/careers-overline/working-life). IIRC the stories never end on a down note (perhaps not true to life).
But, taken as a whole, they offer some ways out of the single-track-grindset that some people in the academic system have -- and that the system promotes.
It turns out that there are a lot of stories out there of people who had to give up on something, change fields, recognize their strengths or weaknesses, etc. People don't talk about this stuff as much as they should.
I think the real takeaway is that collaborating/bouncing ideas off other people is the most important part of research. Few people can work isolated on their own without idea exchange.
Totally agree, collaboration and the exchange of ideas are fundamental to successful research.
The mistake that a lot of people make is that they think science is all about proving things.
In actuality, science done right is all about disproving things.
You can spend a lifetime tracking down all the white swans to “prove” that “all swans are white”, but you need to find only one black swan - as the British did when they reached Australia - to disprove that hypothesis.
Science is about disproving things, as it is by far the easier path - and frequently, the only possible path - to take. And by disproving things, we push back the darkness more and more, until what little remains must contain the truth.
And when science “changes it’s mind”, it’s simply science obtaining more data that points better towards that truth, more of the darkness has been pushed back and the old position was eliminated by the new evidence.
Fortunately I was just doing a Masters so the fact that my thesis (in Material Science) ended up mostly just being some basic characterization of an alloy (grain growth, hardness, Xray crystallography) didn't really matter. It was sort of a dead-end and my thesis advisor shortly thereafter dropped research in the area to focus on his other interest where he was very successful (water ice). Things would probably have been a lot uglier had it been a PhD.
Genuine question: What's the motivation for getting a PhD in an engineering field? I’m guessing you just really really love researching and have enough money/low enough lifestyle to take the oppertunity cost financial hit of not working in industry for five years?
This is also a recipe for successful companies. Start/work at a company, notice some adjacent problem you have to do in the course of the work that is annoying/difficult/error prone, smooth out the problem and turn that into a new product.
It was refreshing to read. A research world producing (publishing) mostly success and breakthrough is not creadible anyway. I belive this improves the reputation of the academic world too.
7 years and a few months for me. I guess I’m not cut out for a PhD :)
[dead]