Cognitive Daily

Suppose you’re a granting agency, and you have $1 million to spend to help foster research in your area of interest. Would you be better off giving ten grants for $100,000 each, in hopes that one or more of the funded projects might produce results, or just announce a $1 million prize — to be awarded after a researcher achieves the desired goals. Instead of gambling on whether the discovery can be made, you let the researchers do the gambling — and you only have to pay if they produce the goods.

The Wall Street Journal’s David Wessel discusses the growing trend of offering prizes instead of outright grants:

“‘Prize philanthropy’ is useful for breaking a bottleneck where government bureaucracy and markets are stuck,” says Thomas Vander Ark, who recently left conventional philanthropy at the Bill & Melinda Gates Foundation to run the X-Prize Foundation. While Gates and similar foundations “push” money on people to solve problems or meet social needs, he says, prizes “pull” people to problems.

Such prizes, newly popular and possible in an age of instant, cheap global communication, have a venerable history. In 1714, Britain offered £20,000 (roughly equivalent to £2.5 million, or $5 million, today) for a way for mariners to determine their longitude. Sir Isaac Newton was convinced the solution lay in astronomy. He was wrong: John Harrison, a working-class joiner with little formal education, built a clock that did the job. In 1919, hotel owner Raymond Orteig offered $25,000 for the first nonstop flight between New York and Paris. Eight years later, Charles Lindbergh won.

Prizes prompt a lot of effort, far more than any sponsor could devote itself, but they generally pay only for success. That’s “an important piece of shifting risk from inside the walls of the company and moving it out to the solver community,” says Jill Panetta, InnoCentive’s chief scientific officer. Competitors for the $10 million prize for the space vehicle spent 10 times that amount trying to win it.

It seems to me that one problem with this logic is that the extra funding has to come from somewhere. If all research were rewarded with prizes instead of pre-funded with grants, then there would be no way for work to get started. Large prizes for research results depend on grants, and without the grants to begin with, nothing would get done.

Comments

#1 Harlan
January 29, 2007

Huh? Prizes work fine where there is investment money. The X Prize worked because investors were willing to finance the project, in hope of getting the prize, and also in hope of reaping benefits of commercialization afterwards. In the market, the “prize” is market share and the ability to sell your product or service. If I’m studying wing development in fruit flies, with no obvious practical uses for decades, there’s not going to be much or any investment money, and I won’t be able to even start the experiments in hope of winning a prize.

Also, a prize is useful when the end state of the research is known. (“Build a better mousetrap/rocket/cancer drug”.) But not so useful for pure research when the goal is just to find stuff out. How do you judge whether Lab A or Lab B has learned more about the hippocampus?

This said, I rather like the use of prizes for engineering goals…
#2 John
January 29, 2007

Of course this is the way to go! As long as the government is providing grants, there’s no incentive to cure anything or solve any problem. If someone cured cancer today, the grant money would dry up tomorrow and a lot of people would be looking for work.

However if the government announced prizes for a cure or solution I’m totally convinced cancer and AIDS would be a thing of the past.
#3 Josh
January 29, 2007

Woah, talk about a recipe for disaster, and an explosion of scientific fraud.

We researchers have enough trouble keeping ourselves honest without directly tying funding to results. cf. Cloning in South Korea.
(check out Judson’s “The Great Betrayal: Fraud in Science” for a couple of other case studies, remarkably fun book)

Aside from plain old fraud there’s also the question of what is considered “useful” research. My wife is in the field of Psychology, and a perennial problem is that psychological studies often generate interesting [b]NEGATIVE[/b] results, but they’re understandably difficult to publish. Having a prize-based system would only exacerbate the problem.

I’m an engineer, and as Harlan mentioned, a prize paradigm may be a bit more achievable here. However, it already exists. It’s called Venture Capital. Check out how many top engineering professors sit on the boards of start-ups.

I think another basic criticism is that prize-type funding would turn Universities from more “pure” research towards more a of production environment, which would actually decrease innovation. The standard procedure currently is, Hypothesize, research & development, maybe prototype (if you’re lucky enough to get that kind of funding), then extend/move-on. Unless prize conditions are carefully worded, focus would shift to the prototyping phase, arguably the least important for the research community, in order to achieve the prize goals. I’d argue that industry tends to be much more efficient at actually building things than the academy.

Also, as you mentioned, the funding for prizes would have to be ramped up quite a bit. If I recall correctly, Spaceship One cost a heck of a lot more than Ansari paid out, and that was a relatively large prize.

Phew, and that’s just off the top of my head. I’m not saying that the current system is ideal. The weird combination of Political Scientist, Spinmeister and Scientist required for effective grant writing makes my head spin, not to mention the time involved in the process. (If anything, I’d say that’s the one thing a Prize-type system would have going for it, the shift away from grant-writing to real-work). But I think it’s a step in several wrong directions at once.

–Josh

P.S. I’m a grad-student currently (so any talk of grant-writing/obtaining funding is from a sort of outsider looking in. In Horror.)
#4 Larry Moran
January 29, 2007

I’m with Josh. It’s a horrible idea because it will destroy curiosity-motivated research and keep people from trying to falsify hypotheses.

Science is not technology. If you want to fund technology by offering prizes then that’s fine by me but don’t confuse the making of things with finding out about the natural world.

Janet Stemwedel and I have been discussing the difference between science and technology at Teaching Ethics in Science: Science v Technology (Part 2).
#5 Dave Munger
January 29, 2007

From a public-policy standpoint, I agree with Josh and Larry. But what I was wondering about is if the perspective is turned around: from the perspective of an individual granting agency, aren’t the payoffs greater when research is funded by prizes, rather than grants?

Sure, on a larger scale, the whole system breaks down, and you can argue that therefore no agency should work on a prize system rather than a grant system. But I just wonder if you don’t get more bang for your research-funding dollar if you offer prizes rather than grants, since you only have to fund the “winner.” Ultimately, this reward system might break down, too, but I’d be interested if any readers can suggest a mechanism (cognitive, economic, or otherwise) by which it would.
#6 Josh
January 29, 2007

Dave,

Even put that way, I’m still not sure that I agree. There are a couple of kinds of value you get from the grant system. (look to DARPA as an example of high-risk type of funding scheme that seems to provide decent payoff)

From the grantor’s point of view, one thing the grant system has going for it, at least ostensibly, is that it gives the grantor substantial control over the grantee. As I understand it, the level of detail as to what’s to be done with the money varies, but may be quite involved.

In a prize system, at the least, you sacrifice that level of control. Whether this is a good or bad thing is open to debate, but you’re losing whatever value that brought you.

In addition, the costs for the prize-keeper are not limited to the prize itself. In a field where there are a couple of major players, evaluation isn’t that expensive. Large scale, dramatic projects (X-prize, Sequencing the Human Genome, the DARPA Challenge) might fall into this category. But what about when there are plenty of labs (and probably hundreds if not thousands of homegrown projects), depending on how the ‘rules’ are worded and the payoff?

Validation/reproduction of results isn’t necessarily cheap, and passing the costs of this onto the participants may limit your field too much, or involve conflicts of interest.

Another thing you get with grants, that you may not with a prize (or at least a winner-take-all prize) is redundancy. If a few labs are funded for a similar problem, you get different approaches. One may go so far, and work better in the initial domain, while the other sits on a shelf, then sometime later the approach that didn’t work is re-discovered, and found to be more applicable to problems of today. Computer Architecture has had this sort of oscillation between serial and parallel approaches with a period of 5-7 years.

And you also get viability, if say, the mother of the PI passes away and she or he loses 6 months work taking care of things. (At work, we try to maintain a “In case I get hit-by-a-bus folder” for projects). Some fields are small enough, where further winnowing of funding recipients might put too many eggs in one basket for comfort.

So while grant-based funding may not be the most efficient in purely monetary terms, there are other, slightly less tangible benefits.

That being said, prizes certainly have their place alongside grant-based. The DARPA challenge brought all sorts of people out of the woodwork and into the spotlight, which is a very good thing for otherwise insular communities.

I don’t think there’s a magic bullet here– TAANSTAFL.

You’ve got competing interests, and as far as I can tell it’s somewhat zero-sum. What value you get ultimately depends on the specific vision you have. SETI provides an interesting data set (maybe not cost effective, but useful) without results in its main mission. If I’ve got a rare, terminal disease and a bucketload of money, then I want a cure ASAP, and I could care less about control over how the researchers do it.

–Josh
#7 Corby
January 29, 2007

In addition to the explosion of fraud and bias toward more “immediate gratification” scientific projects, I can’t even fathom how investigators, postdocs and graduate students at Research I universities will manage to fund their salaries from a winner-take-all prize system when the next awarding of the “prize” to that lab is completely unknown.

The professional instability based on the current grant climate is bad enough here, going totally prize-based would really ruin it. Talk about promoting brain drain…
#8 Clark Goble
January 29, 2007

Wouldn’t any prize be contingent on the demonstration? If so, then wouldn’t that lead to less fraud and not more? After all the alternative method seems based upon a lot of honesty on the scientists – something prizes tends avoid.
#9 Larry Moran
January 29, 2007

Dave Munger asks.

From a public-policy standpoint, I agree with Josh and Larry. But what I was wondering about is if the perspective is turned around: from the perspective of an individual granting agency, aren’t the payoffs greater when research is funded by prizes, rather than grants?

I can’t see how you would get any payoffs at all with a prize system. It takes several hundred thousand dollars a year to run a research lab in my field. Where is that money going to come from during the few years that it takes to do the research?

There’s another problem. Who’s going to decide on the prize? What if I’m interested in doing research on evolution and junk DNA? Do I have to convince someone to offer a prize or do I have to work on someone else’s idea of what’s important in science? Directed research doesn’t have a great track record.
#10 Sunil Bajpai
January 29, 2007

I should think that prizes and grants go together.

While John Harrison did go on to win the meridian prize with his clocks, his efforts were partly supported by grants to build them.

An important choice is between prizes and patents, as rewards for successful research. A point argued by Stiglitz.
#11 Dave Munger
January 30, 2007

There’s another problem. Who’s going to decide on the prize? What if I’m interested in doing research on evolution and junk DNA? Do I have to convince someone to offer a prize or do I have to work on someone else’s idea of what’s important in science? Directed research doesn’t have a great track record.

That’s an excellent point. Consider the prize for proving supernatural powers. It’s been counterproductive, because parapsychology organizations link to it as “evidence” their “science” is real. And the administration costs are high because they spend a lot of time weeding out wackos instead of making their larger point.

Even prizes that grantors really do expect to give away would suffer from this problem — though arguably this may be no more work than reviewing grant applications in the first place.
#12 Michael
January 30, 2007

Mr Goble’s comment, wouldn’t any prize be contingent on the demonstration?“, bring up an interesting point. However, one mustn’t confound intependant verification of hypotheses with demonstrating one’s own results-if you can fake your results you can fake your demonstration as well.

Just take a look at a recent Nature article (Nature, 445, 18 January 2007, pages 240-1) entitled “Misconduct? It’s all academic…” which is not co-incidentally hanging on my office door. The introduction briefly recounts an alleged incident of misconduct that was all but completely overlooked by the university until a large, highly visible part of the community outside the university rallied. If there had been a multi-million dollar prize on the line? I imagine nothing short of a court order would have much effect.

Incidentally I also agree that the grant system, while in some respects problematic, is still better than awarding a large prize. In (my field,) linguistics, I can’t honestly see anyone offering a 1M $ prize for discovering that a hitherto unknown sub-Saharan African dialect does *in fact* present an unvoiced velar plosive (a sound believed not to exist in any language and impossible to procude).
#13 Hans
February 1, 2007

There’s also an issue of scientific openness and intermediate or partial results. In a grant system there is lots of motivation to publish results, if only to improve the chances to get the next grant. In a prize system, you’d want to keep your results as close to the vest as possible, so others can’t take your work to solve part of the problem and combine it with their own on another part to take the whole prize, leaving you with nothing.