Friday, April 07, 2006

Fixing the System

I was doing some brainstorming today after talking to a non-science friend. She was asking me the obvious questions:

1. How did the system get so fucked up?
2. What could fix it at this point?

I wrote the following equations (this is how I think these days, rather than complete sentences):

science= technology + people

technology= people + training + resources

people= training + technical proficiency + creativity + communication skills + desire + rewards

training= books + lectures + experience + place to do it + toys to do it with + making mistakes + feedback

What's the goal?

- greater output (faster progress): cure disease, fix pollution, space travel, easier day-to-day, understand the meaning of life and how everything works. Fix the world.

- waste less money

- employ trained scientists instead of having them go to waste (and the money we spent training them)

Where does massive change come from?

from within: revolution of the proletariat

from without: government mandate (fascism?)

Better to do it fast or slow?

realistic: will be slow to implement even if it's easy to throw out the old system all at once.

idealistic: slow risks not going through with it or not making drastic enough changes.

Why would it be slow to implement?

Because it would involve bureaucrats, and they can't do anything quickly.

Why do we need bureaucrats? What do bureaucrats do? Why would it be so bad if we got rid of them all?

- handle money distribution and arrange for infrastructure (construction, cleaning, building maintenance); handle lawsuits from uneducated public/offended religiosers.

- push paper around.

- it's a self-renewing system: they're good at keeping themselves in business. formula: create rules --> hire people to enforce them --> create conflicts --> hire people to resolve them

Why do we need rules?

because people would kill each other without them. i get that.

Why do we need so many rules and bureaucrats?

Because nobody can know it all in detail and they need something to help guide their decision-making (or prevent them from having to make decisions based on all the information at any given juncture).

Who should be in charge?

- not the bureaucrats!
- not the politicians!
- not solely the old established scientists
- input from young scientists is critical.

Labels: ,


At 8:36 AM, Anonymous Anonymous said...

In spite of the unfinished thoughts and curt bullet points I think I can see where you're headed with your solution for fixing the system. And, I think that you're headed totally in the wrong direction. The problems that you describe on your blog are totally the result of there being too many PhD scientists and not enough jobs. I too am in your situation, so I've spent a good deal of time thinking about the situation.

The issue is that science as we know it is structured incorrectly. Billions in NIH and NSF dollars go into funding of the biosciences annually, yet, it's still not enough. The NIH budget was recently DOUBLED over a five-year period, yet, it's still not enough. So right now we see pain for all young scientists whenever the budget is not growing at >5% annually. This is not sustainable. So the question is: why is $30Billion in NIH funding not enough to insure a tolerable life for scientists? The answer is that a large part of this funding currently goes to fund students. Most R01s go to fund student tuition, insurance, and stipends. So when the Congress decides that the US should produce more science they increase the number of scientists without asking where these new scientists will work. Essentially the system currently pushes more supply of scientists without doing anything about the demand for them. Think about the implications for this. 1) the job market at all levels is totally oversaturated. Competition is fierce for every opening. Degree + post-doc times will rise as the number of papers required for consideration in tenure track (permanent) positions rises. If you’re hiring a new professor why not take the best available candidate. 2) because of the saturation above, salaries will fall. This is why post-docs only make $35k annually to start. Many BS degrees in engineering pay double this these days. 3) because of the poor market those with other options avoid the field. This explains why interest is low among American students. They have more attractive options in law, medicine, and business. Don’t give me this nonsense about ‘the students need more exposure to science’. High school students and undergrads are making rational decisions here – the job market sucks. 4) because there is ample funding for new students yet inadequate domestic interest these fully-funded slots are given to foreign students. Because these students will not find jobs in the US they will eventually go home. We’re training folks to eventually compete against us. While I have no issue with high percentage of foreign students – I’m friends with many of them – one must admit that the current situation makes little sense from a policy prospective. The NIH budget was doubled in an attempt to ‘retain America’s lead in the biosciences’. Paradoxically this action may have had the opposite effect – the new student slots will be filled by foreign students because American students are still uninterested.

So what to do? My solution: phase in a restriction on NIH/NSF funding for students. Start with restricting 10% of government funding so that it cannot fund students in any form. (This would INCLUDE the post-doc as a ‘student’. Essentially funding would be restricted to permanent professional researcher positions.) Over a 10 year period raise this percentage until it hits say 90%. What would be the result of this? What now are large educational / research centers would become primarily research only. The large facilities that are filled with PhD students and post-docs would be filled with professional research teams. The obvious question: without funding, how would we get new scientists? The answer – they’d take out loans just like everyone in medicine, law, and business. Over the 20 year transition period the number of new PhDs would drop, yet the total funding for the scientists would remain the same. Salaries would go up as the positions became more difficult to fill. Eventually salaries would be high enough so that it makes sense to take out large loans – look at medicine where they routinely take $200k in loans – when you’re making $250k a year this amount seems trivial. So with funding coming externally degree times would drop. Domestic interest in the sciences would rise with salaries. A professional class of researchers would rise – the professional bench scientist would become a viable option as a long term career. Instead of training students from all over the world and then sending them home we could do the opposite – make the most vibrant job market for scientists on the planet. As salaries for permanent bench scientists rose we would start to poach the world’s best to stay here for good. (This phenomenon is observed in medicine where the best from all over the world flock to immigrate to the US because of stratospheric salaries.) Added bonus: the taxpayer would get more productivity with their money as students were replaced with more expensive professionals – I’ll argue that one seasoned researcher is more productive than four students. (Grad students are essentially useless in their first two years.) Last bonus: funding for the academic sciences would go up. They would retain 100% of current government funding while gaining tuition money from students (or banks on their behalf) as the career became more profitable.

So there’s my idea for ‘fixing’ the sciences. Basically, we need to push money to increase demand for our services. We currently have a system that pushes the supply of scientists up while ignoring the market demand. If anyone has a better idea I welcome it. Note that the solution cannot involve increasing funding indefinitely. And any one time increase does not fix anything as the new funding will just create more of us.

At 9:07 AM, Blogger Bill Hooker said...

Most R01s go to fund student tuition, insurance, and stipends

Huh? How do RO1s fund students?

At 9:22 AM, Blogger TW Andrews said...

I get the feeling that only young scientists who are committed to an academic career feel that there is such a problem that the foundations of current scientific practice should be overthrown.

Maybe young scientists should just accept that there aren't an infinite number of academic jobs, and use their years of learnin' to start businesses, or work in industry, or make themselves useful in the same way that folks with other degrees need to.

At 10:12 AM, Anonymous Anonymous said...

Have to agree with TW... I've been working in Biotech for 10 years. Sure, there are layoffs periodically, but no one I know who has been laid off has been out of work for more than a month or two, and in some cases they've moved on to even better positions. Even people who have been FIRED from my company have moved on to new positions very quickly. I get the feeling that a lot of these new PhDs have blinders on - they are so focused on tt postions that they don't see the huge number of satisfying career options for PhD scientists. A new MBA who will take nothing less than CEO of a fortune 500 company is foolish because it isn't possible for every one of the thousands of MBAs to secure this type of position. How is this any different than a PhD not taking anything less a tt job at a major research institution?

At 10:33 AM, Blogger NeuroChick said...

In my grad program, R01s are used to fund students once they have picked a dissertation lab. Before then, the program covers tuition, insurance etc. If students win their own fellowships then they're no longer supported by the R01.

I think a big difference between grad school and the other professional schools (law, medicine, business etc.) is the length of time required for the degree. A law or business degree takes 2-3 years. Asking students to cover tuition for such a short time is reasonable, in my opinion. Med school takes 4 years, but the salaries afterwards are so high that taking out loans isn't that big of a deal. Grad school, on the other hand, can take at least 5 years, and usually more like 6-8. Asking a student to cover high tuition for that entire time, when salaries for PhDs are not that high once they graduate, would make it even worse.

I think it would be fine to have PhD students pay tuition while they're in the masters portion of their studies (basically, when they're taking classes), but once they've passed to candidacy and are doing full-time research they should be treated like research fellows and compensated accordingly.

The main difference between law/medicine/business and PhD schools is that in a PhD you are doing work that will benefit the school, both in the short and long term. If you publish a great paper or make a great discovery, your school gets noticed as well. Law, business, and medical students do not have to do original research. It's essentially take these classes, pass these exams, and you get your degree. I think it's completely reasonable to ask the students to pay for that. But if part of my degree requires doing something that has benefits for not just me, but my school, my mentor, my program, and hopefully science in general, I don't think I should be required to be hundreds of thousands of dollars in debt.

At 11:02 AM, Anonymous Anonymous said...

Ok, R01s are the main source of funding for many academic labs in the biosciences. The majority of these funds flow to the salaries/tuition/benefits of the graduate students and postdocs. So, in short, the majority of student funding ultimately comes from NIH or NSF funding. As R01 grants are a significant slice of the NIH pie it was used as an example. Granted, other grant categories do exist, but the same issue applies – a lot of the funding is going to fund students. This funding is artificially inflating the labor pool beyond what the market would otherwise support.

I agree that most scientists should accept that they will not be able to work in academia. But even in industry the situation is poor. Industrial post-docs and starting scientist salaries are way below the salaries of peer groups with similar education. As an example, a top-10 PhD graduate will not receive close to $100k a year to start. A top-10 law graduate can expect to recieve $130k on average from law firms. ( )

Why does this disparity exist? It’s not because our society doesn’t value science – it’s because the supply of scientists is being artificially inflated. I’m not advocating cutting the funding for the biosciences, but reallocating it so that it creates demand for scientists. The supply of PhD bioscientists will then be determined by the same market forces that determines the number of JDs and MBAs.

At 11:16 AM, Anonymous Anonymous said...


Yes, your department funds your first year and fellowships do exist. But these funds also ultimately come from the government from the most part. Your department gets 'overhead' or 'indirect' funding from each of the lab's grants. This indirect funding is often 40%+ of the grant. Most fellowship money can be tied to the NIH or NSF. Yes, private entities do fund students, but not at the same level the government does.

Your point that PhD students deserve funding and 'shouldn't' have debt is well taken. But what if this funding is the cause of the excess of temporary assignments and low pay for bioscientists in later life? Look at physicians, they take on hundreds of thousands of dollars in debt. Yet with average physician salaries well in excess of $100k a year ($200k?) the debt is not that big of a deal. The point is that debt is relative. With a high enough PhD salary the existence of debt would be a non-issue.

At 12:48 PM, Blogger Bill Hooker said...

a lot of the funding is going to fund students. This funding is artificially inflating the labor pool beyond what the market would otherwise support

OK, I get it, thanks for clarifying.

There seems to be something of a chicken-and-egg problem with your proposal, Anon: why would a student take on debt for grad school in order to go into a science career as it's currently constituted? To get students willing to take out loans to do science PhD's, you need an attractive career on the other side, or at least one sufficiently lucrative to make the loans a supportable burden.

To live on loans and pay tuition at the same time, for the 5+ years of a PhD, would cost X; the repayments on such a loan would be Y; this would be an acceptable burden on someone whose income was Z. I don't know what those numbers are -- and they will vary by state and school -- but I know that Z is much higher than current postdoc salaries. You propose to reserve funding away from students incrementally, but it's not clear that there's an initial level (your suggested 10%, or any other) which would provide savings sufficient to increase postdoc salaries to a level commensurate with Z.

At 1:12 PM, Anonymous Anonymous said...


Yes, for sure there would be a huge dropoff in new PhD students over what I call the 'transition period'. Eventually students would return though as salaries rose until an equilibrium is reached. Initially, salaries in the 'professional bench scientist' positions would not be high enough to justify the costs of graduate education. (note that these jobs are not post-docs as that implies a transitional position. I'm proposing the formation of a new professional class of bench scientists. This job would be your ‘final’ career destination, similar to the senior scientist in industry.) So what would happen as the number of young scientists dropped over a 10-20 year period? The salaries of the existing bench scientists would rise as competition for their services increased. In time, that salary would be high enough to justify full funding of education via loans.

There would also probably be downward pressure on the length of a PhD - a good thing by most accounts as it is currently at 7.8 years on average. ( ) Lastly, the education itself would be enriched for the student as they’d be surrounded by highly experienced researchers who are working as a cohesive team. (Not to say that the current situation is all that bad in this area, but that being surrounded by more experience is always a good thing when one is learning.)

What I am proposing is certainly not a small change. A ten year transition period might not be enough. Hell, there may be another idea that is far superior. But I do know that the current situation is sub-optimal. Just take a look at this list ( of articles – many of them paint pictures of despair and desperation. Our devoted and intelligent scientists deserver better from the system.

At 8:00 PM, Anonymous Anonymous said...

Anonymous 1,
thanks for your well thought-out proposal. I agree with it. Maybe, you should write it up and send a copy to every politician in DC.

d. (still a postdoc)

At 10:08 PM, Blogger Bill Hooker said...

It's an interesting idea. Let's run some numbers.

What's the average grant? According to this page, it was about $355K/yr, including indirect costs, in 2005. According to this page, the average facilities and admin rate is 52%, so the PI gets to spend (would actually have requested) about $234K/yr.

What do lab supplies cost? This varies widely, but this page says to budget $15-25K/yr per person at the bench. I remember $1000/month/person being a rough guideline some of my mentors have put forward, so that sounds about right.

What about salaries? According to this page, institutional benefits costs run 17-30% of salary, so let's say 20% to make it easy. The same page says that salaries typically account for 80% of requested funds, but I think they're including part of supply cost in that figure. What does a postdoc cost? Here is the NRSA scale, on which postdoc pay is loosely based (although many schools and institutions pay less). At $40K/yr for salary, the cost to the PI is $50K/yr. What does a student cost? At $20K for a stipend (NRSA guidelines again), a student costs about $25K/yr plus tuition (which varies widely, but would actually be income under Anon's system).

So, we have $234K to spend. Lab supplies will run us $45-$75K; let's say we already have all the equipment we need (established lab) and just ignore maintenance. Two postdocs is $100K, three students another $75K and that puts us towards the low end of the supplies estimate, so we're done.

Now, current student stipends will likely be grandfathered in, so suppose these are all new students and Anon's scheme is firmly in place. We save the entire $75K and earn ourselves (say) $10K/year tuition from each student; then we have enough money to top up each bench scientist's salary costs to $102.5K, of which they will get about $80K.

Would you take out 5-7 years' worth of loans to cover living expenses and tuition fees on the promise of an $80K/yr job? Would anyone lend you the money on the basis of that earning potential? With that prospect on the horizon, would you even study science as an undergraduate?

Note: the numbers here could be way off. I'm just thinking out loud, not making any strong claims.

At 8:48 AM, Anonymous Anonymous said...


An excellent start on running the numbers. I'd like to start from your assumptions and do a little more with them, but unfortunately, I have to do lab work and taxes today. I'll be back this evening with my own take based on your assumptions. (I'd also like to find a place to host a spreadsheet....hmmm...)

Anyway, in short, yes - in an off the cuff analysis it appears that given your numbers a student would be rational in taking out loans to gain this theoretical job. I'll be back this evening. And again, thanks for engaging me in this little exercise - I believe it's more instructive for me than anyone...


At 8:54 AM, Blogger NeuroChick said...

Thanks for putting those numbers together, Bill. One thing that bothers me about Anon's proposal is that tuition goes to the university, not to the PI, and many PIs make their salary solely off of the grants they bring in, not from the university. So requiring students to pay their own tuition doesn't necessarily mean it goes back to funding science, or raising PI salaries. It might fund the new music building (not saying that's a bad thing, I'm just pointing out that it's not like we're talking about a set pool of money that will always remain in science, no matter what).

At 8:49 PM, Anonymous Anonymous said...


When you stated that under the new system the lab would ‘earn’ itself $10k per year per student, I presume that you mean that this cost is forgone as the lab would no longer be paying student tuition as is the case now. In other words, I presume that your theoretical situation represents a lab that now has zero graduate students and a single professional scientist.

So if we use your numbers, the cost to a student to attend school would be $35k per year. Let’s assume a five year degree time – granted this is well below the current average (see above post) but under the new system the time to degree would need to come down to something more reasonable. Five years at $35k per year would result in a total debt of $175k upon graduation. I will assume that these loans are ‘subsidized’ by the government, that the interest payments are covered by them until the student graduates. Upon graduation, the student will begin payment at 6% interest - something close to the current rate I believe. If one sets up a repayment table with the $175k principle compounding at 6% annually one would need to repay just under $24k per year to pay the loan off in 10 years. (The standard length of student loan repayment.) I’ll grant that $2,000 per month sounds like a hefty loan payment. But this student can now get a job at your $80k rate. I’ll grant that higher salaries result in higher taxes so I’ll apply a flat 30% tax on salary. The true take home pay would be more like $56k a year. I will assume that there are no raises given to this scientist over the entire decade. (Clearly this would not happen in reality, but I’m trying to make this analysis as favorable to the current system as I can.) After removing the loan payment, the new scientist has $32k in take home pay. This is $320k in take home cash earned over the ten-year loan repayment period. Sounds meager, but let’s remember that this loan will be repaid within a decade while the $80k salary will pay out over the entire career. Anyway, let’s look at the alternative in today’s system.

I’ll assume that today’s graduate also exits in 5 years – more of a rarity apparently these days if one believes the statistics. (Note that if I lengthen the student time to graduation the overall calculation would be weighted toward the new system. So I’m being as favorable to the current system as possible.) I’ll assume that the new graduate exits school with no debt whatsoever. Using the NRSA guidelines posted above the postdoc would earn between $36k and $51k per annum over years one to eight and holding steady through year ten. The total pre-tax income over this period would be $455k. After applying the above 30% flat tax this becomes $319k in take home pay over the ten year period. Interestingly, this is $1,000 less than the ‘new’ scenario in which loans were utilized. So it would appear that the student would benefit by working under a system where loans were utilized as long as it allowed a higher salary upon graduation.

If we push the analysis timeline out over the entire career while allowing the ‘professional scientist’ to gain pay raises the result would be skewed toward the new system by a wide margin. If we ran the calculation using today’s average graduation time of 7.8 years the result would favor the new system even further. Lastly, I did not apply an interest rate while compounding the take home pay earned over the ten year period. As the professional scientist is making way more money than the postdoc in the early years then the effect of applying an interest rate would be to widen the margin in favor of the new system. (Indeed, if one compounds take home pay annually at 6% the result is a $414k gap in favor of the new system. Certainly nontrivial. One can argue that as the majority of money earned will be spent immediately one shouldn’t apply interest earned to take home pay. This is why I did not include it in my above analysis.) Lastly, as noted by Neurochick, the system as written gives incoming tuition money to the university. In reality, surely some of this would flow back to the laboratory. If this is the case then this money could be used to further sweeten the pot for the professional scientist. Or buy supplies. But as 100% of today’s grant money would remain in either system the lab could only gain monetarily.

Anyway, as with Bill, my numbers are very rough and may have errors. But I think that they’re in the ballpark enough to show that the above proposal isn’t totally ludicrous. Which makes it the only viable alternative to the current regime that I’ve ever heard of. Food for thought. As before, I invite other proposals…

At 11:26 PM, Blogger Bill Hooker said...

I have to do lab work and taxes

The postdoc version of "death and taxes". Heh.

I presume that your theoretical situation represents a lab that now has zero graduate students and a single professional scientist.

No, it has three graduate students and two bench scientists. The lab is actually getting that $10K/yr tuition fee from each student (although, NeuroChick, I think you're right, and the tuition is likely to get the same F&I treatment as the grant money, ie max 2/3 goes to the PI). Thus we had $75K of grant left (because we didn't have to pay the students anything) and $30K in extra income. I split that between the two postdocs to make them bench scientists; they cost us $102.5K, of which they actually get 80% or around $80K, the remainder being paid out as benefits.

income = $234 (grant) + $30 (tuition) = $264K
two bench scientists = $205K
supplies = 264 - 205 = $59K.

I did a little more figuring -- I took a $200K loan at fairly generous terms and more than ten years to pay off, and got something like $1500/month in payments. If taxes come to about 30%, I make the take-home around $38K, in the same ballpark as your $32K and considerably better than the kind of income many postdocs can expect now.

I started out skeptical, but I am coming around.

Here's a very important point, for me: I'm assuming that with the new job, bench scientist, we will see a different set of assumptions -- that a bench scientist, unlike a postdoc, will not be expected to be aiming for tenure, willing to relocate all over the damn place for a shot at it, and so on. In other words, I'm viewing "bench scientist" as a permanent position, a career in itself. If that's the case, I think the new system looks considerably better than the old, from a postdoc point of view.

Some of the questions that remain focus on the effect on students. My wife, who grew up poor, points out that devolving the cost of graduate studentship onto the students themselves will mean no poor people ever become scientists. It's not because they couldn't get the loans, or because it's not economically rational (the above analysis indicates that it certainly is economically better with the loan system). The thing is, poor people do not even think about getting such jobs -- taking such loans. It will put "scientist" in the same category as "lawyer" and "US Senator" -- out of consideration.

Here's another: what level of student debt do US scientists currently incur? My impression is that it's small because it's all from undergraduate study, and postgrad study does not involve further debt. Am I right about that? If so, we could say that science handicaps itself relative to other fields (by paying students instead of making them take loans), or we could say that science gives itself a considerable advantage in attracting students in the first place.

Another observation from my wife: we're making the debt load to become a scientist approach that which lawyers and doctors take on, without making the remuneration anywhere even close. I'm not sure how different this is from the current situation (who goes into science to get rich?), but it's a consideration.

And one more from me: how are faculty going to react to this? Assistant and even some associate professors don't earn $80K/yr! We are going to have to fund an increase in PI salaries as well.

At 6:25 AM, Blogger NeuroChick said...

This comment has been removed by a blog administrator.

At 6:55 AM, Blogger NeuroChick said...

Let's say currently a PI gets an R01 funded and needs X human hours to complete the research proposed. Using Bill's paradigm, this lab has 3 grad students and 2 postdocs.
g=grad student hours
p= postdoc hours
and let's say a postdoc can put in 1.5x the hours of a grad student
3(g) + 2(p) = X
X = 3(g) + 2(1.5(g))

So let's say altogether the three grad students can put in as many human hours as 2 postdocs. Now let's say Anon's proposal is in place, and the PI doesn't pay for students. If postdocs cost $40k/yr, but students cost $0 to the PI (or just $20k if students retain a stipend), what's to stop the PI from just filling their lab up with free students, and have no postdocs? Or have just one postdoc and 4-5 students? The PI would now have more money to spend on their own salary, or on supplies/equipment for the lab. But fewer postdocs (or bench scientists in Anon's proposal) would have jobs.

At 4:30 PM, Blogger Bill Hooker said...

what's to stop the PI from just filling their lab up with free students

The fact that these students are taking out a mortgage's worth of loans to get through grad school: why would they go to a lab in which they are seen as cheap labor? Labs will be competing to attract grad students, and having professional mentors available will be an important factor.

At 4:45 PM, Blogger Ms.PhD said...

Wow, I'm amused at how this conversation went on a major tangent and really never came back.

re: lab work and taxes, I laughed at that too.

re: tuition, whoever thinks tuition goes back to the PI or the lab in any way is WRONG. Tuition goes to pay overhead for the school. There is a tiny trickle-down effect in the form of infrastructure, but that is all the lab gets. So there is no benefit, in that sense, from having a student in the lab.

Several other points I wanted to make...

Yes, grad school and postdoc both incur debt for American students. First off, most of the top schools are located in expensive cities where rent is high and you have to have a car. So most grad students have car loans. Many postdocs are of the age where they want to have children. Who can afford to have children on $25-35k a year? People on welfare, that's who. There was a big uproar at my grad school a few years ago when a Mormon chemist started lobbying hard for Federal assistance. Our school wasn't qualified to even have students apply for these kinds of loans, because they were unwilling to go through the paperwork of becoming accredited that way, I don't know the names but some of you may have heard of these things.

I gotta say, while I think the Market Forces Model is an important one, in practice expecting grad students to take out loans to pay for the privilege of being slave labor in a lab is bullshit. More to the point, there's little training involved and it really is more like a job, especially after the first year or two when all the classes and the major exams are mostly over.

I've said before and I'll say it again, I agree that having fewer students is the way to go. The problem is, universities have no incentive to limit their graduate programs. It's free teaching for the undergraduates and free labor for the labs. If you can figure out a way to convince everyone that it's in their best interest- and it can be misleading, it just has to sound good- that is a big step in the right direction.

At 5:07 PM, Blogger Ms.PhD said...


You clearly have no concept of what it's like right now.

Grad students go where the fame and money are. That means it doesn't matter how unavailable the 'mentor' is or how many other slave laborers there are. In fact, in some cases it's better to go to a lab with a lot of students, because then at least you have each other! But you go for the name of the PI, not for the quality of the advising, because ultimately the name will get you farther than good training from someone no one's ever heard of (I'm a great example of that myself).

So no, you can't expect grad students to self-regulate based on criteria any sane, intelligent person would use. Grad school priorities are a different animal. That's what the Market Forces Model requires pressure to come from somewhere else- ideally from outside the system. If scientists and universities were going to self-regulate, somebody would be doing something about it by now. And they're not.

At 5:22 PM, Anonymous Anonymous said...

I'm not sure refactoring the system will help much with the money/career aspect. Part of the issue might be an artificial inflation in the supply of scientists, but I think there is a deeper problem.

I think the major problem with scientific research is that it's on the bottom of the money chain (like the food chain, except with money). Other professions that require expensive professional degrees (MD, lawyers, ect..) operate on higher tiers. Take the legal system as an example. I don't know much about the legal system, but I do know that people or companies stand to lose substantial (millions-billions) of dollars if they end up on the wrong end of a legal decision. They have a vested interest in not losing money, and are therefore willing to pay ridiculous amounts of money into the system to ensure that they are protected. Law firms benefit directly from this. They take the money from businesses, keep some, and use the rest to hire lawyers. I bet that lawyers end up seeing more from Big Pharmas money than scientists do.

Law and medicine are service industries. They don't really create anything. What's even more interesting is that they grow via positive feedback - lawyers create a demand for more lawyers. Doctors also create a demand for more doctors. The effect may be indirect, but ultimately the growth of both professions is non-linear.

Research is different. It's not a service industry like law or medicine. Scientists are mostly focused on trying to figure out how things work. That practice alone is not profitable. Someone has to take that information and turn it into a product/service so that they can make money. I think I remember reading somewhere that if you consider the amount of wealth generated from discoveries resulting from NIH spending, the governments return on their investment is quite good. Companies spend money on R&D for the same reason. Unfortunately, like the government, the amount they spend on research is still a relatively small percentage of the total.

How do we create a demand for raw knowledge? If you can create high demand, you can tap into the money flow further upstream, rather than fighting for the last few drops.

At 6:03 PM, Anonymous Anonymous said...

So this entire idea came to me when I began to think about PhD population control. There are two ways to do this: 1) Central control - eg: control the number of schools and the number of slots in each school. By all appearances, this is what MDs do with AMCAS, and it works well for them. I would argue though that they've taken things too far in the other direction, to the point to where they're inflating their salaries and workload beyond what is reasonable. 2) Let the market control the population. This is how JDs and MBAs work. When starting salaries drop in either field you see a drop in applications a year or so afterwards. Self regulation does work in these fields. They too have better career prospects for the most part than those with a PhD.

So which to choose? My example above shows how the second model can be implemented while being revenue-neutral for academia. (If they stand to lose in any given proposal you can bet that the issue is dead.) My thinking was that funding for students is acting like a crutch – it’s funneling people into a field that they would not otherwise enter if they actually sat down and analyzed their future prospects. But if you remove this crutch you cut out the main source of revenue for the sciences. I think we can all agree that more science is a good thing. So the idea was simply to shift it over to permanent positions. This had the added benefit of providing more jobs for the existing pool of scientists. If this plan were to be implemented, enrollment would plummet. No one would be willing to put thousands of dollars of their own money into an education that resulted in a $35k postdoc.

Lastly, I agree that universities have no incentive whatsoever currently to limit enrollment or improve education and treatment of students / postdocs. We’re cheap labour. This is why my proposal does include a single top-down change – the restriction of funds from providing for future students. This would result in population control for PhDs while avoiding an AMCAS style cap – if the need for PhDs grew in future years the population would grow organically without some sort of convoluted bureaucracy. When the job market improved it would provide an incentive for students to jump back into the field. Using the savings from forgone student stipends can start that process, although as seen above it wouldn’t provide an overwhelming source of riches. Either way, we’d all be better off than under today’s system – even if PhDs still were not competitive with JDs.

If schools were forced out of the system where students were being used as cheap labour they would be forced to improve conditions. School ranking would matter. Quality of education would matter. Time to graduation would matter. Future placement would matter. Insert you measure of choice here.

I agree that in the above scenario there are numerous problems. Professor salary, tuition remission, etc etc etc. But the point is that in general it conforms to one of the two frameworks for population control. I like a market based solution over a central one for numerous reasons, not the least of which is that it may be easier to implement. I see smaller issues like professor salaries and tuition remission to labs to be a minor point – the big issue is PhD population control.

I am surrounded by unhappy students and postdocs. This message board is filled with them. Everyone acknowledges that the system is broken, but I have yet to see any potential solutions. I encourage you all to think of something and toss it around with others as I’ve done above. Eventually one proposal may stick and real change can occur. With that sermon, I’m off to dinner then to finish a northern…


At 6:18 PM, Blogger Ms.PhD said...

As usual, I think Anonymous and I actually agree, since this is essentially what I've proposed in the past: pay the people out of grad school more, and limit the number of people going to grad school. And then the market will take care of itself.

Again where I think we're all stuck is how to make this happen. Lobby Congress? NIH?

Reports like Bridges to Independence make some useful suggestions but don't comment on where the money is supposed to come from (the insinuation is always that it should come from the money currently funding the war in Iraq).

I don't see anyone trying to put together committees to seriously come up with plans, funds and timetables for implementation of changes that have already been suggested in other reports from committees that were actually organized by the National Academies.

Hello, National Academies! You need more people like the ones on this blog!

But what we really need here are some policy wonks who could tell us who to contact on Capitol Hill and at NIH and how to put pressure on them. Do we need a petition? Would that accomplish anything? An open letter to Nature/Cell/Science?

At 8:30 PM, Anonymous Anonymous said...

I think that any viable solution cannot rely on more government funding as part of the plan. Practically speaking, an increased budget is just not going to happen anytime soon. The NIH allocation was recently doubled and that, along with our current Iraq based budget woes, make additional requests a long shot. So for a plan to have any hope it’d have to be revenue neutral for the federal government and preferably revenue-positive for academia (as a way to sweeten the pot.) Lastly, I think that any plan that calls for fewer scientists explicitly is destined for failure. The academy & industry have been cheering for increased interest in the sciences for so long that most people believe that there is a lack of qualified individuals in the biosciences. Diverting student dollars to professional scientists would implicitly cut the number of scientists in the long term while explicitly increasing scientific productivity in the short term. Instead of funding new first year graduate students the system would be funding the 7th year postdoc that would have otherwise left the field. More science gets done for the same dollars.

Anonymous Coward PhD, you raise a really interesting point about proximity to revenue generation. I’ve been thinking along similar lines for a while now, but I haven’t worked out anything concrete. Ultimately, this is a tougher nut to crack. My only thought is that we’re not likely to find a solution in this area soon, so we’d be better off optimizing the situation where possible elsewhere.

Ms.PhD, I think that if the number of PhDs were reduced the pay upon graduation would necessarily rise without need for direct intervention. This assumes that the national science budget (NIH+NSF+etc) would remain the same while the number of scientists dropped. As the glut of talented individuals was reduced, competition in recruiting would heat up - pushing salaries to a level that is commensurate with effort and therefore more equivalent to other professions. I’ll grant that one could maintain the current postdoc to assistant professor pipeline as is by simply curtailing supply so greatly that everyone fit inside the pipeline comfortably. But this would cut the overall national output of science - which is, after all, the point of the federal budget in the first place. So why not rechannel some of the existing postdocs into a new professional scientist position? After all, most scientists I know would rather be at the bench as opposed to performing assistant professor duties (grants, grants, grants…and teaching uninterested 18 year olds.)

Unfortunately, I don’t see any changes happening in the neat future. Maybe I’m a pessimist, but both academia and the government move at glacial speeds. I guess this is why I’m content to discuss the issues (anonymously?), and hope that over time a consensus builds.


At 8:35 PM, Anonymous Anonymous said...

Doooh. Editing. Last paragraph should read "Unfortunately, I don’t see any changes happening in the near future."

At 2:29 AM, Blogger Bill Hooker said...

You clearly have no concept of what it's like right now.

Gee, thanks. I wonder if any of your problems stem from a tendency to dismiss others without bothering to check facts?

At 6:26 AM, Blogger NeuroChick said...

The fact that these students are taking out a mortgage's worth of loans to get through grad school: why would they go to a lab in which they are seen as cheap labor? Labs will be competing to attract grad students, and having professional mentors available will be an important factor.

I don't think grad students' choices of labs would change if they were responsible for paying for grad school. Most grad students (at least that I know personally) do a pretty good job of deciding which labs are good. There have only been a couple of instances that I'm personally acquainted with where the student really made (in my opinion) a bad decision about a lab.

If grad students were paying tuition (and especially if they weren't receiving a stipend), the PI would likely see very little of that money, or a negligible amount. What's to stop a PI from making things seem all hunky-dory for the first 5 years, only to turn around in the student's dissertation year and make it impossible for them to graduate because they'd be losing their cheap labor? That already happens often now, when grad students are not saddled with the cost of their education. I think requiring students to pay for the privilege of putting their future in the hands of 4-5 people (but really just the PI) is asking them to take a gamble that many aren't willing to take. Currently, when there's a problem such as this the graduate program is there to back the student up, partly because they have been financially supporting the student and want to see their investment work out. If instead the student was paying, I could see the likelihood of the program siding with the PI being much higher.

Anon, in your proposal how is "professional bench scientist" different from Research Associate or Research Assistant Professor? These are generally more long-term positions than an ordinary postdoc, but they are still part of an established lab, and not running their own labs. They don't have to write grants or teach, and spend the vast majority of their time at the bench. Maybe part of the solution is just making it more acceptable to aim for this kind of post, rather than the tenure-track faculty position.

There only seems to be an oversupply of PhDs for tenure track positions -- industry, government, and non-tenure track academic positions seem to be providing a decent demand. Someone in a tenure track position should be able to independently bring in grant money, teach, do top notch research, and contribute administratively to the university (committees etc.). Not everyone who gets a PhD is capable of that, or even wants to do that. I think part of the problem is that grad students and postdocs think they are all good enough to get a tenure-track position (and do well in it once they have it), when in reality probably only 5-10% really are that good. I think if industry and other avenues were not seen as second class, but just other ways of using your PhD, this "crisis" wouldn't seem so big.

At 11:58 AM, Blogger Ms.PhD said...

Fuck. I just wrote a whole long comment, and blogger LOST IT. fuckers!

At 9:36 PM, Anonymous Anonymous said...

What if the system works exactly as intended. Lots of work gets done by very earnest and dedicated people who believe in their own capacity to the point of excluding data to the contrary. If you look at the system from the top, it looks like any other hierarchy. There are lots and lots of people trying to make their way up and fewer and fewer positions as you climb. People who think nothing can keep them from becoming tenured at top-tier institutions are delusional. Period. It takes luck. No one tells you that going in. Sure, you can have the educational pedigree, the top tier publications, well-connected grad and post-doc advisors. Still your CV looks just like the other top 100 that apply for the position. Maybe no one tells you that luck is a major component because they don't want to believe its true. Anyone who has 'made it' can probably string together a rational causality for getting their position without invoking "luck". Don't think that getting a job at a liberal arts school is any easier. Maybe you should write children's books.

At 11:27 PM, Anonymous Anonymous said...

I am now more glad than ever that I double majored in economics and chemistry because this is the stupidist discussion I have ever heard. If I didn't do the degree in economics I would be thinking like all of you, and from the posts above it is easy to see why the science jobs market is bad.
here is the skinny folks: you people have a serious problem with reality. you don't know what it is. The reality is that lab rats are viewed as labor by high level govenemnt officals such as the president (this one and the one before him); they look at you the same as they look at mexicans who pick strawberries (ever use the expresssion fruitful research ?) And they want to keep they suppy up so costs are low. Like Wal-mart. And the culture of "independent thinkers" really helps them, no real unions to deal with. This will not change in our lifetimes. Why? How many post-docs do you know that are u=talking about starting a union? None, they think they are above that.

At 12:01 PM, Blogger Bayman said...

Great discussion, and I agree with most points. But I think the whole scientist's conundrum issue is summed up by a few timeless facts:

1) Demand for knowledge is infinite. Our current economic model will therefore continue to spew out more and more scientists.

2) Real knowledge production is not simply proportional to the number of scientists. One can ramp-up the production of useless information by generating more science workers and more labs, but true discoveries take their own sweet time. If anything, real discovery requires intensive effort directed against a very small and finite set of problems. Increasing the number of researchers beyond a certain small number usually does little to speed the time to solution.

Taken together, 1 and 2 lead to persistent devaluation of the scientist, as the system becomes locked in a cycle of producing more and more scientists but is not getting any payoff in return (a corresponding increase in the rate of beneficial discoveries.)

The bottleneck in today's massively complex and interactive global society is not one of quantitative data production, but the qualitative one of identifying what the problems actually are. In North America particularly, we are very good at throwing money into data generation but very few people understand society or history (even within their fields) enough to actually know where the funding and effort would be useful.

So true science doesn't fit into modern North American economics of supply and demand. You can like doing science and be interested, but will never be guaranteed a job. If you want to be successful, focus on taking the time to understand the context of the world you live in and identifying areas in which innovation and discovery will be useful to societies.


Post a Comment

<< Home