The Selfish Scientist

Can the selection pressures facing scientists be modified so that individual behavior better aligns with the interests of patients, the ultimate beneficiaries – and benefactors – of our research?

Posted 29 December 2008 by Noam Y. Harel (PLEASE COMMENT/CO-AUTHOR THIS PROPOSAL).

To edit collaborative proposals, please register by clicking 'Join this Space' on the left,

then click the 'Edit This Page' button on the right!!!

To make comments without becoming a member, use the 'discussion' tab.

Selection pressures facing scientists today continue to escalate. Funding, promotion, and professional survival itself depend on the credit that goes to those who foster a project from idea through publication. Those who publish first, and publish alone, get more credit. To succeed in this harsh environment, it becomes not only survival of the fittest, but far too often, survival of the most secretive – those who ‘protect’ a novel idea or finding until their lab has sufficient data for publishing first.
Never mind that it may take years longer for one lab to develop, fund, test, and confirm a project alone than it would take for an openly-collaborating community of labs. Never mind that the race to publish first sometimes concludes with data that are at best shoddy and at worst fraudulent, leading to wasted time and resources in attempting to replicate that data.
The real tragedy of the Selfish Scientist approach is that it delays potential treatments for patients facing their own selection pressures in the forms of deadly diseases. Obviously, when scientists share ideas and preliminary findings, the pace of scientific progress quickens. When the pace of progress quickens, the entire world benefits.
Yet, the incentives driving individual scientists conflict with behaviors that would benefit society as a whole. The option of sacrificing individual credit for the good of the community will usually lose out to the opportunity for gaining personal glory, even if at the expense of delaying communal benefits.
As with the Selfish Gene, the drive of the Selfish Scientist is a formidable force – and an adaptable one. What if selection pressures were altered such that behaviors satisfying the Selfish Scientist better aligned with those that benefit society at large? Funding agencies, universities and journals have the power to change the incentives driving both individual and group scientific progress. Below are some of many potential options meant to spark discussion amongst the three pillars supporting science:

Funding Agencies - Can results be produced more efficiently and reliably for less money?

  • Could an open grant submission format improve the efficiency of review and revision?
    • An online repository of grant proposals would mutually increase the pool for both funders seeking scientists and scientists seeking funders, while simultaneously cutting redundant paperwork and deadline pressure.
    • Publicly-posted research proposals would allow constructive criticism to come from the broader scientific community, rather than just a small study section with a higher percentage of members who are not familiar with the technologies used, or who have conflicting interests.
  • Could groups requesting funding for complementary or overlapping projects be convinced to collaborate and divide labor?
    • Using the power of the purse, funding agencies could mandate collaboration, including dividing labor amongst labs with the best expertise in specific phases of a project.
      • This would increase the speed, comprehensiveness, and reliability of project completion.
      • Multiple labs contributing to these types of successful collaborations would earn credit, rather than one or two labs collecting both the glory and the accompanying contempt.
    • Conversely, funding agencies with complementary missions could collaboratively fund/support individual projects that appeal to them.
      • This could spread an individual agency's money more broadly, increasing the number of projects it supports as well as credit it receives.
  • Could funding agencies benefit if individual scientists/labs were eligible for funding even if they could not themselves perform every experiment or phase of a proposed project?
    • This would lessen the incentive for Selfish Scientists to keep their ideas and preliminary findings secret for years until they have the resources to experimentally tackle the entire project on their own.
      • This would speed submission of high-quality ideas and preliminary findings to funding agencies, which could then help organize collaborative efforts/division of labor to achieve the overall projects.
      • Aside from the increased speed of converting ideas and preliminary findings into full-fledged projects, funding agencies would have the luxury of selecting from a higher quality pool of submissions.

Universities - Could mission statements regarding public service be better fulfilled without drastically altering the current system of promotion and tenure?

  • Publications, citations, and awarded grants are important indicators of a scientist's productivity, and should remain so. Furthermore, if funding agencies (above) and journals (below) effect some changes, then grants and publications would inherently incorporate the value of public contributions made by individual scientists. But Universities would still need to alter their evaluation structures to fully quantify and reward an individual's contribution to the communal advancement of science. Some metrics that could potentially be useful:
    • Reproducibility of findings from prior publications.
    • Contribution of novel ideas or helpful suggestions to other research projects in the open format suggested for funding agencies above.
    • Time spent reviewing grants and papers.
    • Time spent teaching.
    • Quality of teaching.
  • Group incentives could drive individual scientists, labs, divisions, departments, and universities to strive for higher cumulative public contributions in a benevolent Ponzi scheme.
    • If resource allocation (funders to universities; universities to individual departments, etc.) were based partly on metrics of cumulative public contribution, this would motivate chairpersons (deans) to push individual scientists (departments) to act in more publicly beneficial ways.
  • Universities could benefit from these changes by better fulfilling mission statements regarding public service, and by decreasing some of the acrimony, subjectivity, and politics surrounding promotion and tenure.

Journals - can the reliability of published results be improved without sacrificing impact factors and profit?

  • An easier path to publication of negative results would provide multiple gains mutually beneficial to scientists, funders, and publishers:
    • For scientists, reduction of the pressure/temptation to ‘massage’ data into a positive story for the sake of publication.
    • For funding agencies, fewer dollars wasted on supporting duplicative research into negative lines of investigation.
    • For journals, a lower risk of receiving submissions with questionable data, without a major drop in impact factor (see below).
    • For everyone, broader awareness of what doesn’t work would prevent enormous amounts of wasted time and resources spent by labs unknowingly pursuing lines of research that lead to nowhere.
  • An easier path to publication of confirmatory (or refutatory) results would provide multiple gains mutually beneficial to scientists, funders, and publishers:
    • Labs attempting to replicate previously published data currently face a Catch-22: unsuccessful replication usually remains unsubmitted. Successful replication gets the dreaded ‘not original’ label that often precludes or delays publication. The rest of us are left to wade through troves of literature, attempting to decide which papers to trust. No wonder then that much of the published literature cannot be faithfully reproduced.
    • What gains could be achieved if for each new breakthrough paper (with the exception of multiple concurring breakthrough papers published simultaneously), journals reserved space in a future issue for publishing the results of one or two confirmatory/refutatory studies?
      • For readers, a better method to learn which published findings are most reliable.
      • For investigators, better incentive for performing the essential function of confirming others' results.
      • For funding agencies, a boost in their efforts to support the growth of independent confirmatory research.
      • For journals, better planning of future issues, and increased anticipation of their readerships for the results of confirmatory studies.
  • Would journal impact factors suffer from publishing more negative and confirmatory results?
    • Though formal citation of negative results may be less likely, citation of 'positive' papers would probably include not just the original paper but the confirmatory paper published later in the same journal.
    • Other measures of 'impact' such as page views and downloads might show that negative results attract readers in similar numbers to positive results.
    • Perhaps a trial study could be performed, comparing citations/page views/downloads of negative versus positive results in a specific field of research? This might show that any potential detriment to a journal's impact factor could be outweighed by the benefit of a more reliable set of published results.
These are just a few proposed options in an attempt to encourage the evolution of today’s scientific ecosystem. With a commitment of effort and time (not necessarily more money), funding agencies, universities, and journals can modify the selection pressures facing individual scientists. Selfish Scientists can adapt to these changes, better aligning our individual behavior with the interests of patients, the ultimate beneficiaries – and benefactors – of our research.
Please add your ideas and modifications directly to this proposal in Wiki-manner. The goal is to eventually submit this collaboratively-authored document to a high-profile print publication, in an attempt to draw interest from individuals not currently a part of the Open Science movement.