Lin Jiang University of Missouri Jerry Thursby Georgia Institute of Technology Marie Thursby Georgia Institute of Technology amp NBER 1 Prior Research Sharing Among Competing Researchers Paper I HJTT ID: 188439
Download Presentation The PPT/PDF document "Communication, Collaboration, and Compet..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Communication, Collaboration, and Competition in Scientific Research
Lin Jiang
University of MissouriJerry ThursbyGeorgia Institute of TechnologyMarie ThursbyGeorgia Institute of Technology & NBER
1Slide2
Prior Research
“Sharing
Among Competing Researchers” Paper I: HJTT Survey of UK and German Bio-scientists’ willingness to shareSimple game theoretic framework of 2 situations
Private sharing of materials, methods, algorithms
Public presentation of intermediate resultsDifferent trade-offs & incentives in the two contextsReciprocity important in formerCredit & feedback vs. risk of misappropriation in latterCompetition in both reduces sharing, ceteris paribus
Life cycle and lab size effects differ in the two contexts
2Slide3
Communication, Collaboration, and Competition in Scientific Research
Game Theoretic Models:
Competition/Collaboration Model Disclosure can lead to competition which is undesirable. Collaboration can result from disclosure.Mathematician Model No credit for a discovery until entire problem solved. Competition can lead to completion so competition is desirable.
Research Leader Model
Entry of other researchers (competitors) enhances the reputation of the focal researcher. They are seen as leaders so competition is desirable.Survey: Survey of academic researchers across disciplines that mirrors the above models.
3Slide4
Competition/Collaboration Model
Focal Researcher makes
discovery of partial value in solving a problem. There are 2 types of other researchers Trusted colleagues – can be trusted not to compete General colleagues – who might decide to competeTwo decisions: Whether to share with general colleagues Whether to collaborate with either a trusted or a general colleague (if they are willing)Tradeoffs in generally sharingBenefits: reputational credit from general colleagues for the initial discovery possibly find a collaborator Risks: engaging rivals who may have better talents or resources. 4Slide5
The Competition/Collaboration Model
Players: Focal Researcher, Trusted Colleagues, General Colleagues
Can vary in quality/resources (i, t, g)Each working on different trajectoryV base value of a publication (Vi, Vt, Vg)Focal Researcher has made a random discovery r
Decides to share with t or g (there are n general colleagues)Decides whether to collaborate with t or g
5Slide6
Competition/Collaboration Model
Includes opportunity costs of working on the project (base value is k) and that cost reflects time saving if collaborate and a cost of coordination
If collaborate, then value of solution is the maximum quality times V.Competition reduces chances focal researcher “wins”Competitors face a cost of replication (s)
6Slide7
Possible Outcomes
If GS, then there is a possibility that a general colleague will compete.
There are n general colleagues.Slide8
Simulation with parameters given numerical values. without loss, can consider the ratio k/V Small -> important or novel discovery Large -> incremental discovery
Vary over qualities i, g, tVary over “amount of initial discovery” rVary over replication costs sVary over number of general colleaguesTotal cases considered > 4 million
Consider results for three conditions:ISMARTER: i > t and i > gIDUMBER: i < t and i < gIBET_GIT: g < i <tIBET_TIG: t < i <g Simulation Results: Competition ModelSlide9
Outcomes
ISMARTER
IBET_GITIBET_TIGIDUMBERGS_G2.20%0.10%23.40%13.10%GS_T2.20%28.00%0.00%20.80%
GS_N82.40%
61.80%55.20%49.20%TS_T0.00%0.40%0.00%
3.20%TS_N
13.20%9.90%21.50%13.70%If collaborate with trusted colleague, then typically GSSlide10
Outcomes
ISMARTER
IBET_GITIBET_TIGIDUMBERGS_G2.20%0.10%23.40%13.10%GS_T2.20%28.00%0.00%20.80%
GS_N82.40%
61.80%55.20%49.20%TS_T0.00%0.40%0.00%
3.20%TS_N
13.20%9.90%21.50%13.70%Collaborate less likely than not collaborate (IDUMBER close)Collaborate typically when t or g greater than iSlide11
The Mathematician ModelFocal researcher benefits from introducing competition since credit for initial discovery comes only with a complete solution.Do not distinguish between trusted and general colleagues since competition is welcomed.
Does not consider collaboration.Slide12
The Research Leader ModelFocal researcher gains reputation if there is entry (competition): the focal researcher is a research leader (they have caused others to change their research)No distinction between trusted and general colleagues and collaboration not modeled.
Unlike the prior models the disclosure decision is not “now or never.” The focal researcher decides on how much of the project to complete before disclosure.Entry decision depends on how much of the problem has been solved.All else equal, disclosure by a focal researcher who are much more capable than the general colleagues will lead to more entry.Slide13
Importance of the Projectk/VCompetition/CollaborationIncremental projects more likely to lead to collaboration. Incremental discovery more likely to be generally disclosed
Mathematician ModelNovel projects more likely to be disclosed.Research Leader Model(Complicated) When i is much greater than g, a very novel project is less likely to be disclosed, otherwise a very novel project is more likely to be disclosed.