Today’s New York Times mentions my work on the history of prizes and grants: Back in the 1700s, prizes were a fairly common way to reward innovation. … Eventually, though, prizes began to be replaced by grants that awarded money upfront. Some of this was for good reason. As science became more advanced, scientists often needed to buy expensive equipment and hire a staff before having any chance of making a discovery.
Time to Praise Al Gore and Richard Branson-Updated
I was watching CBS this morning and got treated to a really constructive effort initiated by Al Gore on climate change.Former presidential candidate Al Gore has joined forces with the British financial heavyweight Sir Richard Branson to offer up a $...
I've suggested a prize for predictions of global climate change, as an alternative to what is essentially paying the IPCC (Intergovernmental Panel on Climate Change) to lie.
A method for achieving honest climate predictions
I've also proposed technology prize for developing (non-tokamak) fusion, since tokamak fusion is pretty clearly headed nowhere:
Carl's suggestion is pretty much the model used by Innocentives and other "open innovation" platforms in the corporate sector. Task definition is indeed "decentralized" but it's hardly unbiased, since rewards are offered only for innovation whose benefits are privately valuable (i.e. free-riders can be excluded, etc.)
Those interested in philanthropic uses of prize mechanisms might want to check out this proposal:
The system proposed there does not pre-specify a technology, but rather pre-specifies a way to measure technologies' value, and a sum of money to be paid out in proportion to measured value. This solves a lot of the problems associated with fixed-target prizes.
What makes this fix possible is having a well-specified measure of something that the donor values: in this case, increases in agricultural production in Africa, to help farmers overcome poverty and malnutrition. These increases are possible to measure, are desired by donors, and could be rewarded in a market-like way.
Task definition can be both decentralized and debiased: funding entities can accept and pay for experimental designs submitted to them from independent scientists, just as they accept grant proposals today, but without the direct connection to implementation. The set of submitted ideas should be improved in several respects over the current set:
1. Reduced bias towards fancy status-enhancing equipment (fMRIs anyone?).2. Reduced bias against against experiments that are boring to run.3. Smart academics' reallocation of time towards experimental design and away from tasks like interviewing the 50th subject in a psychological study.4. Scientists designing experiments to be done by others should be less influenced by self-serving biases, and thus more likely to propose replication and measures to prevent falsification/exaggeration. The symmetry=attractiveness idea would have had more difficulty spreading (http://www.overcomingbias.c... if replication had been simultaneous.
David Leonhardt has a near-excellent column in the New York Times today on why prizes are 1) A great way to foster innovation, but; 2) far less popular than grants or other compensation schemes: in the 1700s, prizes were a...
I agree that prizes are underemployed. I've also thought for some time that it would be useful to supplement the grant system with registries (as in wedding registries) of studies to be performed. Using the language of 'fillers' and 'framers' (http://www.overcomingbias.c..., you pay framers to identify needed studies, and then use competitive bidding to have fillers in professional for-profit firms execute them. Some advantages:
1. This would be a handy vehicle for replication, since important studies can appear multiple times in the registry, with each firm only allowed to bid for one.2. Firms could be subjected to reputational tracking and auditing to ensure strong internal controls aimed against data falsification.3. Firm scientists would have less incentive to exaggerate results, or to selectively report positive results, than scientists invested in particular hypotheses and with their social status at stake.4. Division of labor and economies of scale improving the efficiency with which experiments are conducted by the experimentation firms.
Time to Praise Al Gore and Richard Branson-Updated
I was watching CBS this morning and got treated to a really constructive effort initiated by Al Gore on climate change.Former presidential candidate Al Gore has joined forces with the British financial heavyweight Sir Richard Branson to offer up a $...
Hi,
I've suggested a prize for predictions of global climate change, as an alternative to what is essentially paying the IPCC (Intergovernmental Panel on Climate Change) to lie.
A method for achieving honest climate predictions
I've also proposed technology prize for developing (non-tokamak) fusion, since tokamak fusion is pretty clearly headed nowhere:
Fusion technology prizes, October 8, 2006, 20:07
Mark
Carl's suggestion is pretty much the model used by Innocentives and other "open innovation" platforms in the corporate sector. Task definition is indeed "decentralized" but it's hardly unbiased, since rewards are offered only for innovation whose benefits are privately valuable (i.e. free-riders can be excluded, etc.)
Those interested in philanthropic uses of prize mechanisms might want to check out this proposal:
www.agecon.purdue.edu/prizes
The system proposed there does not pre-specify a technology, but rather pre-specifies a way to measure technologies' value, and a sum of money to be paid out in proportion to measured value. This solves a lot of the problems associated with fixed-target prizes.
What makes this fix possible is having a well-specified measure of something that the donor values: in this case, increases in agricultural production in Africa, to help farmers overcome poverty and malnutrition. These increases are possible to measure, are desired by donors, and could be rewarded in a market-like way.
Task definition can be both decentralized and debiased: funding entities can accept and pay for experimental designs submitted to them from independent scientists, just as they accept grant proposals today, but without the direct connection to implementation. The set of submitted ideas should be improved in several respects over the current set:
1. Reduced bias towards fancy status-enhancing equipment (fMRIs anyone?).2. Reduced bias against against experiments that are boring to run.3. Smart academics' reallocation of time towards experimental design and away from tasks like interviewing the 50th subject in a psychological study.4. Scientists designing experiments to be done by others should be less influenced by self-serving biases, and thus more likely to propose replication and measures to prevent falsification/exaggeration. The symmetry=attractiveness idea would have had more difficulty spreading (http://www.overcomingbias.c... if replication had been simultaneous.
I want more prizes
David Leonhardt has a near-excellent column in the New York Times today on why prizes are 1) A great way to foster innovation, but; 2) far less popular than grants or other compensation schemes: in the 1700s, prizes were a...
Carl, a thoughtful suggestion, though there is the issue of how clearly the filling tasks could be defined.
I agree that prizes are underemployed. I've also thought for some time that it would be useful to supplement the grant system with registries (as in wedding registries) of studies to be performed. Using the language of 'fillers' and 'framers' (http://www.overcomingbias.c..., you pay framers to identify needed studies, and then use competitive bidding to have fillers in professional for-profit firms execute them. Some advantages:
1. This would be a handy vehicle for replication, since important studies can appear multiple times in the registry, with each firm only allowed to bid for one.2. Firms could be subjected to reputational tracking and auditing to ensure strong internal controls aimed against data falsification.3. Firm scientists would have less incentive to exaggerate results, or to selectively report positive results, than scientists invested in particular hypotheses and with their social status at stake.4. Division of labor and economies of scale improving the efficiency with which experiments are conducted by the experimentation firms.