Pre-registration:Reproducibility Crisis And The Journal Editing System Revolution

Reproducibility crisis has broken out in psychology,neuroscience,biomedicine and other scientific fields since 2010.In order to quell the crisis,the European and American journal editor's community and the scientific community actively explored the way out of the crisis.Pre-registration is a new approach to reproducibility crisis.Firstly,Different reproducibility connotation determines different situations of crisis.Daniel T.Gilbert et al.denied the existence of the crisis and argued that there were many differences between reproduced and original experiment.Liu Chuanjun et al.attributed the reproducible crisis to embodied cognition and argued that the crisis could not be solved.Christopher D.Chambers advocated pre-registration of new practices to restrain the crisis.Secondly,the transformation of submission and publication standards.Pre-registration is closer to the scientific essence of concept innovation.It only takes the degree of concept innovation before experiments and the reasonable degree of experiments design as the criterion for judging the publication.It improves the quality control of the whole process of scientific research,including the idea,experiments design,experiments process and writing process.Thirdly,create scientific standards that are traceable throughout.Pre-registration explicitly requires labeling of confirmatory experiments that conform to the original experimental design and exploratory experiments that are discovered after experiments,complete recording and rational explanation of all changes in ideas and experiments,making it easier to trace the original experiments,and improving the level of scientific reproducibility.The new pre-registered editing system can control the reproducibility crisis,and the key is to contain the bias,help to reveal many kinds of experimental bias,and construct the full data chain from scientific research to paper publication,providing the data base of reproducibility of experiment and even bias.Firstly,register and upload the pilot data to build a data base for preserving and revealing confirmation bias.The formally published papers also include the pilot data,thus improving the data reproducibility before experimental design.Secondly,incremental registration of all hindsight findings,which both controls hindsight bias and provides a reproducible data basis for hindsight findings;Thirdly,the second peer review assessed the degree of data analysis and execution after the experiment according to the statistical method of pre-registration and uploading.The selective reporting bias was controlled in a systematic way to avoid problems such as floor effect,ceiling effect,P-value effect,file drawer and so on.Fourthly,the pre-registration system allows the publication of invalid or nullresults,which can control the publication bias and improve the reproducibility of results.As China increasingly becomes an innovative country,it is facing more and more irreproducible problems.Therefore,China should start the pilot pre-registration of sci-tech journals,and build a pre-registration system from data registration platform to a new peer review system,to curb the crisis thoroughly.