原文链接：https://genomebiology.biomedcentral.com/articles/10.1186/s13059-018-1584-0?utm_source=other&utm_medium=other&utm_content=null&utm_campaign=BSCN_2_DD_GenBio_TPP_Scinet

微信链接：https://mp.weixin.qq.com/s/La53EnujFWzIEThOBG2v3g

在过去的一年里，Genome Biology测试了透明评审制度对审稿流程的影响，并在近期的一篇Editorial里面公布了测试结果。

Genome Biology进行测试的透明评审制度，变态天剑狂刀，指的是每一轮的审稿报告以及作者在修回期间针对审稿意见进行回复的点对点内容都在稿件最终见刊上线之时作为最终稿的附件之一一同发表，读者们通过下载稿件的最后一份附属文件就可以找到稿件从送审之日起的同行评审历史信息。

本次测试共有45篇透明评审测试稿，68篇传统评审对照稿（审稿报告仅对编辑、审稿人以及作者开放），我们针对最为重要的审稿周期，审稿意见质量等情况进行分析。实行新制度测试之初，我们最担心的是审稿人会由于报告最终会公开而拒绝参与审稿流程，从测试的结果看，我们并没有遇到这个问题，而平均审稿周期在测试组与对照组中并没有明显差异。同样的，我们也并没有发现审稿人由于报告最终会被发表而对稿件放松要求，从而减低审稿质量。

再次，Genome Biology向参与此次测试的审稿人与作者们表示感谢，并决定于2019年1月1日起正式实行透明审稿制度，也就意味着每一篇在Genome Biology 经过同行评审的文章，都会有评审报告及修改回复一起公开。

摘要：

We describe the results of our year-long trial of transparent peer review and announce the adoption of transparent review as permanent policy.

A year ago, Genome Biology published an editorial describing a trial of transparent review at the journal (where transparent review is the process of publishing reviewer reports anonymously alongside the published article) [1]. Other journals already use this model of peer review and have reported reasonably high author satisfaction [2], but we could find no quantitative reports about the effects on manuscript turnaround times or the engagement of peer reviewers. We were concerned that reviewers might be more reluctant to agree to review if they knew that their reviews would be published. This would lead to increased turnaround times for authors, since a large factor in determining the time elapsed between submission and the initial decision is the length of time taken to find sufficient numbers of reviewers to agree to review. Thus, we started a trial to compare submissions assessed under traditional single-blind review (where the authors’ identities are known to reviewers, the reviewers remain anonymous, and their reports are not published) with those reviewed transparently.

We have now assessed the data we collected over the year. In the trial, 45 submissions underwent transparent review for at least one round of review, and 68 went through traditional single-blind peer review. Eligible authors could opt out of the trial before the review procedure commenced either explicitly or passively (if they did not reply to our emails), which accounts for the difference in numbers between the two arms. Our main finding is that there was no significant difference in the mean time to first decision between transparent review and single-blind review. For both sides of the trial, the average number of reviewers we needed to invite for each agreed reviewer was 3.1. Thus, we can see no evidence that reviewers are more reluctant to agree to review transparently. This is also borne out in the responses we received from reviewers—those who commented at all were positive; no reviewer declined to review explicitly because of transparency, although we cannot rule out that some declined for this reason without letting us know.

A low rate of reviewer engagement is only one potential downside of transparent review, though. Another concern might be that reviewers would be reluctant to be critical knowing that their criticisms would be published; however, we see no evidence of that in our data. There is no significant difference between the two arms of the trial in the proportions of manuscripts given a first decision of outright rejection, rejection with the option of resubmitting a revised version, or major revisions. Of those manuscripts that have been through the trial to a final decision (published or rejected), there is no difference in the overall rejection rate.

Based on these results, it seems to us that transparent review does not affect the speed or outcome of the peer-review process. It should be noted, however, that the pool of transparently reviewed manuscripts had some element of self-selection, as authors whose manuscripts qualified to be transparently reviewed were offered the option to take part (or not) in the trial. This might introduce some bias to the results. For example, authors who suspect that their studies might be reviewed more negatively might opt for traditional review, as they might not want criticisms of their paper to be made public. Whether or not this bias affected the results, it is true that 13 out of 98 authors chose not to take part in the trial. Somewhat to our surprise, three of those 13 explicitly stated that they opted out because review would not be fully open (suggesting that they would wish the reviewers’ names to be published alongside the reviews). Moving to transparent review permanently would therefore run the risk of putting off some authors and reducing the number of manuscripts submitted to the journal.