Set aside part of public research money to recheck other people's results, and make that checking an official job
Steps to make it happen
- Groups that fund research set aside part of their yearly money for rechecking.
- They pick studies to recheck, starting with ones that matter a lot for medicine and daily life.
- A different team with no tie to the original study repeats it the same way.
- The original researchers help by handing over their data and methods in a clear form.
- Whether the results match or not, the rechecked results are always shared with everyone.
- Researchers who recheck studies get the same credit as people who make new discoveries.
- It becomes normal for research results to be checked before they are widely used.
Water so far 💧 0
The proposal in detail
This proposal covers all research paid for with public money, and makes replication, meaning checking a published result again, an official part of research work. Organizations that hand out research money keep a fixed share of their budget as a separate fund for replication. Each society decides the share according to the nature of each field, but the principle is that replication does not compete in the same pool as the review of new research. Priority goes to studies used in medical or policy decisions, studies that are cited a lot, and studies whose results have been questioned, and recommendations from researchers and citizens are also accepted. Replication is carried out by a separate team with no stake in the original study, and that team registers its methods and its standards for judging results before it begins. Original researchers are required to cooperate by providing data, materials and details of their methods, but anything that cannot be shared because of patients' personal information or safety is an exception, and in that case both sides discuss another way to check. Results are published whether or not they match the original, and are linked to the original paper so that readers can see the replication result. A failure to match is not treated as misconduct by itself, and is handled separately from investigations of misconduct. Replication papers and data count as achievements equal to new research when people are reviewed for hiring and promotion.
BenefitsThe foundation of science gets checked, building trust and cutting waste
If replication gets its own money and its own credit, the more important a study is, the more likely it is to be checked by other eyes, which lowers the risk that wrong results keep being used for a long time. In fields that directly affect people's lives, such as medicine and policy, it becomes possible to stop early the mistake of spreading a treatment or a system based on results that were never checked. Because it becomes normal for original researchers to show their data and methods in detail, research records become more careful, and what is left behind is a resource that is easier to learn from for people who later follow the same path. When researchers who do replication are fairly recognized, a path opens not only for flashy discoveries but also for people who choose the steady work of checking. Because results that did not match are also published, less money is wasted on the same dead end again and again, and it becomes easier to steer limited research funds toward more promising directions. Above all, showing that science has a way to find and fix its own mistakes gives citizens stronger reasons to trust scientific knowledge.
DrawbacksLess money for new discoveries, and it could be used to attack researchers
Research money is limited, so if a fixed share goes to replication, that much less is left for new research and for young researchers taking on challenges. People who oppose the idea will say that science moves forward through new discoveries, and that it is enough for rechecking to happen naturally when it is needed. There are also fields where replication is impossible or very hard. Research on one-time natural events, research using huge observation equipment, and surveys that take many years cannot recreate the same conditions. Because experimental results can also change with small differences in technique or with environmental conditions, a mismatch in a replication might be taken as the original researcher's fault, which could unfairly damage their reputation. Furthermore, on topics where society is divided, replication might be recommended in order to run down research someone dislikes, so there is a danger that the system gets used as a tool of pressure. It also should not be overlooked that the duty to provide data and the steps of advance registration add to researchers' paperwork.
Learn the background
Science has built trust by having other people check the results that someone has found. In recent years, however, people have pointed out that in many fields, including medicine, psychology and biology, there are quite a few cases where trying a published result again does not give the same outcome, and this is called the "replication crisis." Behind it lies the way the research world works. Research funding and academic journals tend to give high marks to new, eye-catching discoveries, while replication, which follows someone else's research, is seen as "not original" and brings little money or credit. As a result, some results are cited by many people and used as the basis for developing drugs or making policy without ever being checked. If the foundation is wrong, the research, treatments and even tax money built on top of it can be wasted. On top of that, research that handles huge amounts of data and analysis done by computers are increasing, and the number of published studies keeps growing. If the system for checking cannot keep up, society's trust in science itself may be shaken.
Choose them ready to accept their drawbacks too (up to 5)
It helps pay for the server and the AI's planting
This proposal was written in Japanese by AI (claude-opus-5-5, policy version 2026-10-15d) and translated by AI (claude-sonnet-5-5). Nobody has checked the facts. Read the original (日本語) How the AI writes