Hacker News Reader

Comment by alexandros | Hacker News Reader

"A glance at scholarly journals or university-press catalogs might make one wonder how much of this "research" is advancing knowledge and how much is part of a guild's need to credentialize its members."

A glance? Really? Judging the life's work of a few thousand people who got into academia to do research is easily judged in a 'glance'? It's interesting that some would like universities to be little more than vocational training programs, but I've never seen such dismissal of the work produced by academia in such a high profile outlet, with no evidence to back it up whatsoever.

EDIT: Reading further, I see that they did look more closely.. at sociology journals. Why the strange choice? Biology, computing, medicine, physics not interesting enough?

View on HN

See another article currently on the front page:

Realtime Linux: academia v. reality http://lwn.net/SubscriberLink/397422/27eef125e03b8a2d/

"I actually took the time to spend a day at a university where I could gain access to IEEE papers without wasting my private money. I picked out twenty recent realtime related papers and did a quick survey. Twelve of the papers were a rehash of well-known and well-researched topics, and at least half of them were badly written as well. From the remaining eight papers, six were micro improvements based on previous papers where I had a hard time figuring out why the papers had been written at all. One of those was merely describing the effects of converting a constant which influences resource partitioning into a runtime configurable variable. So that left two papers which seemed actually worthwhile to read in detail. Funny enough, I had already read one of those papers as it was publicly accessible in a slightly modified form.

That survey really convinced me to stay away from IEEE forever and to consider the university ranking system even more suspicious."

That's funny. If you took a look at startups, you'd find that 99% of them are me-too crap. Badly designed websites for Yet Another Social Networking Gimmick, bad mobile apps, outlandish bad business ideas, etc.

1% of them ever truly go anywhere. 0.1% of them change the world.

My point is that anytime you randomly sample a field of work of any kind, you're going to get an exponential distribution of quality with a few hits and a long tail. All papers are not the Church-Turing thesis, and all startups are not Google or Apple.

A lot of academia-bashing amounts to criticism of academia for having the same sorts of failings that are present in equal measure in all other societal institutions public or private. Things like: elitism, know-nothing overpaid leadership, poor quality control, loss of focus, abusive labor practices, perverse incentives, poor success ratios for new ventures, etc.

Sturgeon's Law: Applicable Again!

http://en.wikipedia.org/wiki/Sturgeons_Law

Looking at journals isn't the only way to do it. You can take encyclopedias and biographical dictionaries, and check for citations and mentions there.

Funnily enough, one such attempt (http://en.wikipedia.org/wiki/Human_Accomplishment) finds that per capita scientific achievement seems to start declining around the 1920s or 1930s, which is also (with WWII and whatnot) the beginning of the modern university system.

Correlation != causation.

http://www.seanbonner.com/blog/archives/piratesarecool.jpg

I think per capita scientific achievement has declined, but I think it's because we did all the easy stuff already. We are starting to hit a wall where new discoveries increasingly require one or more of the following:

1) A lot of expensive tools. (accelerators, genomics cores, supercomputers, etc.)

2) A lot of labor. (The "armies of graduate students and postdocs" phenomenon is a product of this.)

3) Intelligence beyond the capacity of a single individual. (Supercomputers, large numbers of people over time, etc. are required to "grasp" something. Anything ending in "omics" is a perfect example of this.)

You can think of it as being loosely analogous to peak oil, though the underlying causes are different. With peak oil, it's because the resource itself is limited. With scientific discoveries, you might have an infinite ultimate supply but you have a finite horizon of vision. It's like if there actually was infinite oil deep in the earth, but we were strictly incapable of building drills longer than one mile.

Personally I think #3 is the limit we're really hitting. I sometimes doubt that we have the IQ power as a species to truly grasp the genome or to unify QM and gravity. We either don't have the IQ or we don't yet have the language or philosophical framework to think about those things.

We're not infinitely smart. Can your dog read Hacker News? Maybe aliens are watching us now and thinking "well, they're sort of clever but they're really just still animals... they only see two sides to every conflict, they can't think in more than three dimensions, and they can't hold more than seven or eight simultaneous ideas at once..."

I'm not saying academia is perfect. It needs to change. If we really are nearing our IQ limits, we need an environment that removes distractions and allows people to really focus to get that last bit of creativity and intelligence. Academia is too bureaucratic, noisy, interrupt driven, and stressful. All those are IQ-killers.

"Correlation doesn't imply causation, but it does waggle its eyebrows suggestively and gesture furtively while mouth 'look over there'" http://xkcd.com/552/

I might buy the End of Science argument, but this would be far from the first time that it has been made, and I'm not sure it really works. Look at computer science; how much of that was or could have been worked out in the absence of an actual computer? And as much computer development was aided by government funding, the commercial imperatives were just as compelling. (Look at IBM's pre-digital computer success.)

I agree... like an economic bubble, I don't think it's possible to ever "call the top" except in retrospect. But it does worry me. I don't think human beings like to think about it because it bruises our collective ego to think that there might be a bunch of stuff that we're too dumb to get.

But I do wonder if we've passed "peak innovation." The biggest reason is the 1960s. It really feels like virtually every piece of technology that we use outside of maybe biotech was invented in the 50s and 60s.

Looking back, I see two peaks. I see a peak in per capita innovation in the early 20th century, and then a second peak in absolute terms centered at approximately 1960.

The difference is that startups go out of business when they fail, whereas publishing papers like this is considered success in academia.

You fail in academia too, but the floor is different. With a startup, you cease to exist or drop off into an abandonware project. In academia, failure means you get lackluster appointments, little grant money, etc. You kind of stagnate.

I've noticed that smart self-aware people in academia can right this by recognizing that kind of stagnation for what it is, namely failure, and responding by changing course, increasing effort, etc.

The dangerous thing about academia is that failure just looks like mediocrity... mediocrity is failure. This isn't the case in the real world, where mediocrity is just mediocrity and failure is... well... you see them sleeping on the sidewalk.

Your dopamine system is not calibrated this way. So to succeed in academia I think you have to realize that if you're not a star you're sleeping in a gutter and respond to your situation accordingly. Respond to mediocre results the way you would respond to sleeping in the gutter in the "real world."

A random sample of an entire field or industry should have an exponential distribution, yes. Journals, though, are quality-gated: there are supposed to be review processes in place so that only the interesting/important/hopefully-seminal stuff gets published. That they reflect the same exponential curve is a testament to the fact that the review processes don't work.

I agree that journals do not have the quality control they claim to have. I was responding to the larger idea that academia is somehow unique in having such a quality distribution.

Some years ago, I saw a 60 minutes episode on issues in higher education. The reporter picked a biology journal off a shelf, read its arcane title, and exclaimed "who actually reads this stuff?" to back up her claim about how useless journals are... as if she were somehow qualified to judge the importance of a paper by its title.

I couldn't help but imagining what she would have said if she picked up a paper titled "A Mathematical Theory of Communication" or "On formally undecidable propositions of. Principia Mathematica and related systems".

Fortunately it's not up to her to decide. That said, it seems that some papers go out of their way to hide what the paper is really about by looking at the title.

Usually you just don't know. I got shamed in front of a prof once by saying that a given paper title was a good way to make people not want to come to your talk, and he responded by explaining why he thought it was an interesting talk.

There are plenty of examples of things that might be meaningless to a graph-theory professor that are a core part of the language of a combinatorics specialist.

And of course, undergrads and humanities majors best keep their mouths shut.

One of the things taught to undergrads in the humanities—specifically, in communications design—is that you should make the metadata of something interesting-seeming to a wider audience than the data it represents, so that it spreads further—that way, more of the people interested in the data will see it.

As an example, one of my friends told me about a very interesting paper called, and I quote, "Three Monkeys something something." He didn't know what it was about, but he said it "sounded like something I'd be interested in." He was quite right—but if it hadn't been for the monkeys, he wouldn't had anything vivid about the paper to recall at all.

Wasn't that an Andy Rooney sound bite? I.e. a comedian's opinion on the world.

There's a Mathematical Theory of Communication? Wow!

I noticed that too, and I wondered if they were even capable of making that judgement. I looked up their backgrounds: the review writer is an English professor at Emory; one of the book authors is a Professor Emeritus in the Political Science Department at Queens College in NY.

This continues the tendency I've seen elsewhere, which is to lump science and engineering education and research in with humanities when criticizing higher education.

Yes, but only humanities types do that, quite deliberately in the hope that the hard subjects will reciprocate when defending them.

It's a fair point, but he's probably focusing on the arts and soft sciences because those degrees are the most overpriced for the value they deliver.

To put it another way, an MIT grad with an engineering degree is going to be much better off career-wise than a Harvard grad who majored in gender studies.

The long term studies of technical degrees vs 'soft science' hasn't been looked at in depth on a national scale with relation to earnings and salary. These studies are just beginning.

From AACRO: after 10 years the earning potential for an engineering degree drops below someone with an English degree. Technical degrees are way more valuable out of the gate while soft science take time to develop.

To put it into other words: where an engineer can make 65k right after graduation an english major will start at 35k. The engineer's salary increases on par with inflation while the english major increases substantially over time and overtakes it on average at the 10 year point. The paraphrased quote from AACROs lobbyist: 'we're seeing through earnings study that technical degrees go stale over the career of an individual, this is shown when comparing the average english major vs the average electrical engineer' There's alot of background nuance (politics) there behind the study but the idea hasn't really been broached before in higher ed. It's really quite interesting to look at on a US scale, but that data hasn't been collected yet.

Oh: I'm not contesting that some soft science degrees aren't total wastes of money. Just that there will always be a place for a gender studies Harvard grad. We just don't need 10,000 of them.

Do you have a link to back this claim up? I don't believe it.

If you look at PayScale.com, for example, you'll see a ranking of undergrad majors, with starting and mid-career salaries, where mid-career is defined as "full-time employees with at least 10 years of experience in their career or field who hold a bachelor's degree and no higher degrees."

ref: http://www.payscale.com/best-colleges/degrees.asp

English Majors start at $37,800 and get to $67,500 after 10 years. Electrical Engineering Majors start at $60,800 and get to $104,000 after 10 years.

AACRO has an incentive to overstate the benefits of non-technical college degrees, as these are the most common types of degrees, and these are the degrees that are leading more people to question the value of a college education.

Does anyone know of any studies that take into consideration labour-force dropout as well? It's a bit less enlightening if you compare the 95% of science grads who are employed doing science with the 60% of soft studies grads (both numbers purely hypothetical), to compare the most career-driven X-students with all the Y-students is unfair to Y-studies.

Tenored professorships are nicely paying, hard to get but exist in every discipline (let's pick on Gender Studies) so if 2/100 GS grads become 100K/year tenured profs and 98/100 opt out of the labour force, you're really measuring the wrong thing.

I do not have any reports to back this up as the transcript from the meeting is yet to be released. I've been curious to dig into the distinctions and locations used for the selection populations.

The discussion I paraphrased was made at an AACRO tech conference during a talk about the creation of a federal student tracking database - the speaker was Barmak Nassirian who is a lobbyist in DC for higher ed. I'm going off of the data presented in that talk.

On the same note, I've never used payscale.com for more than a glance. Their methodology sounds valid, but I don't know how they make their money.

I bet you have some tremendous selection bias going on.

I bet they only looked at the english majors who still had a job in the field after many years. If you include all the english majors, including the failed ones who work in a different field then the numbers might make more sense.

Definitely, it's from an AACRO conference which is full of collegiate registrars and provosts. The message may ignore contrary data to get a political agenda across. Until I can review the actual reports the argument is based upon, I can't really say for certain. To truly analyze the information across the nation the feds will have to create a US student tracking and lump it in with K-12 & IRS data. There's a quiet movement to do so - but there's resistance to that idea hence the discussion. Resistance stems from the orwellian aspects while proponents trend along "we spend federal dollars on these diplomas, we need to make sure they're doing GOOD for the society somehow." A figure quoted at the same talk was that we spend 3% of our GDP on education.

I didn't say I agreed with what was said, it's just something that higher ed is talking about at the top of the greater organization and was worth a mention.

Personally I feel state academia is bloated and backwards and many changes need to be made. Admissions should be highly curtailed where the whole nation makes an effort towards math and science instead of 'getting a degree so you can get 10k$ extra starting out.' I'm still not convinced that the federal dollars lost to defaulting loans wouldn't be better spent on k-12 education.

Payscale make their money by helping companies figure out what salary ranges to offer in new markets, offering benchmarking and other compensation guidance.

This is very true. Most young male engineering grads like to gloat the fact that they make a lot more than their humanities counterparts (perhaps to make up for the misery and lack of female companionship during their undergrad?)

But the reality is that if you graduated with a humanities/social science degree from a ivy league or sub-ivy school, you are bound to do pretty well. As opposed to their state school counterparts, English majors from Harvard have the academic clout to get into a top law/med school which will put them with far greater earning power over engineers in 10 years.

But if you are a true-minded humanities major, you might even become a true humanitarian for a couple of years; maybe travel abroad to a Peace Corps or get a masters at a International Studies program. Then only later, you get on the fast-track to a management program at some Fortune 500 to manage some engineers.

> But if you are a true-minded humanities major, you might even become a true humanitarian for a couple of years; maybe travel abroad to a Peace Corps or get a masters at a International Studies program. Then only later, you get on the fast-track to a management program at some Fortune 500 to manage some engineers.

Keep telling yourself that.

Only a small fraction of humanities grads make that transition.

Oh, and the Peace Corps wants folks who can actually do things. Food, shelter, etc come from applied technology.

Ahhh this is one of the issues I've heard tossed around - what makes for a successful individual post graduation? How can we measure it?

The feds are putting in place systems to track income as the best metric. But... I hesitate to call a 100k$ rockstar EE major ten times more successful than a rockstar chem e doing water purification work in Guatemala for 10k$ as a humanitarian.

Pretty simple answer to that. You don't. Most college grads don't realize that how much you make at a job is a really poor indicator of your self-worth. But you don't realize this until after you've been on the job for awhile - kind of like high schoolers thinking where you go to college is the ultimate determination of your worth.

so how can we defend the money lent and given by the people (from the federal government) is going to a good cause?

  the most overpriced for the value they deliver.

For some values of 'value'. Because you can't put a dollar sign on everything, right? I feel much better about the value I got from my humanities classes than I ever did about my computer science classes, whether or not that translates directly into money-earned.

It's the WSJ editorial page. Punching hippies is one of their 3 reasons for existence.

Ok, I'll bite. What are the other 2?

Well, the number 3 was arbitrary, but let's say..

2) Spending 8 years advocating "deficits don't matter", then changing their opinion 180 degrees within 24 hours of a new President being inaugurated.

3) Writing screeds castigating liberals for being coastal elitists and out of touch on an iPad from the back seat of the author's car service, while drinking a latte on the way to yoga class.

Why are people voting up this idiot? Is it time for yet-another "this isn't reddit" admonishment?

Yes.

This isn't reddit. Please provide some rationale for your namecalling, but preferably please just stay at home.

/me watches the comment thread change over time as those who are more motivated by tribal allegiances start to outweigh those who appreciate a fair joke

I voted you all up because I thought this exchange was irrelevant but pretty humorous.

I appreciate the well placed joke now and again. "Punching hippies" was funny because its truthy enough to be relevant, and relevant enough to be in the thread.

HN goes through a phase every now and again where I think it takes itself a little too seriously. Its for important discussions unlike that red-place with the silly alien (pretentious snorteling and huffing). All attempts at humor are immediately voted down.

I'm not one of the greats, but I've been here for a few laps and I've got some karma so I'm calling it:

Lighten up y'all.

> HN goes through a phase every now and again where I think it takes itself a little too seriously.

hahaha. yes, every now and then. hahahahahahahaha

When I feel like lightening up, I go to that red-place. When I'm serious, I come back here. It's nice to have the separation; if we weren't all so serious all the time on HN—if it was just another Reddit, in other words—would there really be a point to coming here? :)

You don't see the irony in your comment, do you?

Since the Rupertster took over WSJ, hippie whacking seems to have even eclipsed good business and financial reporting over there.

A lot can be discerned in a glance.

just a glance is enough at this table of contents of academic journal - classical and quantum gravity- http://iopscience.iop.org/0264-9381/27/17 to wonder how much of this -research- is advancing knowledge and how much is --about advancing the career of the authors--

It's a very good question, and hardly a trivial one. Quite a lot of science is nearly worthless, and only published to pad the author's cv. A couple of blog posts on the topic:

http://pipeline.corante.com/archives/2010/06/24/all_those_wo...

http://pipeline.corante.com/archives/2010/06/25/what_to_do_w...

Lasers were thought to be a laboratory curiosity when they were invented. On the other hand, you have a ton of academics who are convinced that the "semantic web" is the next big thing.

Part of the problem with science is that it's an activity where it's very hard to have any useful metric. Even experts in the field have a tough time deciding whether a given research direction is going to be fruitful. They're often wrong.

So far the best thing we've figured out for science resembles the Underpants Gnomes' business plan:

1) Fund lots of science.

2) ????

3) Profit!

The ???? part is "pray that some of it is useful and that some of that gets into the hands of people who will use it."

1) Fund lots of science.

Why do we need this? By every metric I know of, governments spend orders of magnitude more on science than they did in the 1800s. Are we getting orders more "Profit!"?

Could you clarify your metrics and sources?

http://www.oecd.org/dataoecd/17/53/41558958.pdf tells me that non-commercial (government or government-subsidized) R&D represents ~1% of GDP in 2006 for the US. I believe equivalent figures for the 1800s would be closer to 0% R&D inasmuch as there were no Departments of Energy, Health & Human Services, etc., but I'm willing to be proven wrong. Maybe the Gilded Age governments in the US actually spent 2% of GDP on R&D.

You can't split atoms with a chisel, nor email the international space station with telegraph lines. The discoveries possible with 1800-vintage equipment are mostly already made, so now we have to spend orders of magnitude more to learn anything new. And I'd much rather we roll the dice again and keep looking, not just wait and see how civilization survives the decline of oil and fertilizer.

Reply on news.ycombinator.com