A Proposal for Science Journals
Scientific journals have a bias toward positive results and are hesitant to publish replications. To anyone familiar with the scientific method, that is a major problem, as science requires unbiased and repeated tests to verify results. Therefore, I propose the following reform for evidence based scientific papers:
1.) Papers are accepted or rejected before the data are collected. The scientists submit their abstract, data collection methods and the tests they will run. Ideally, the paper will be accepted before the data itself is even collected so that the researchers themselves are unbiased. Failing that, at the minimum, journal editors will not be able to select for studies which agree with their ideological biases, because they won’t know the results until it is too late to reject.
2.) Immediately upon accepting a paper, the journal releases the paper for other scientists to verify. Anyone who wanted to could begin running the paper with their own data and try to pre-replicate the study. At this point, they would know the methods, but not the “original result”. Thus, replicators would not be biased in either direction. The more common bias would be to confirm the original study, because the researchers typically do not want to rock the scientific boat and offend the original researchers. Conversely, ideologically opposed groups might want to counter the conclusions of the original researchers for the notoriety of contradicting their opponents.
3.) Both papers would be published simultaneously regardless of their outcomes. The original paper’s authors would get access to both datasets and research notes so they could reconcile any differences. Perhaps the discrepancy would be a difference in testing or in data sampling and seemingly contradictory results could be reconciled.
4.) All data and methods would be published for independent re-replication. Scientists who did not publish their data and methods would be treated as frauds and shunned (as they should be!).
The way journals are run today is shocking. That so called scientists could refuse to publish their data and not face widespread condemnation/dismissal is mind boggling to me. That’s pretty much explicitly saying you falsified it, or you have something to hide. That journals can reject a paper just because it disagrees with their biases is likewise troublesome. To quote one of my favorite philosophers, “You don’t use science to show that you’re right, you use science to become right.”