Tuesday, December 29, 2015

Peer Review in Science and the Science Classroom

I think that peer review has a place in the teaching of scientific thinking. Science hangs its credibility on peer review. A scientist performs an experiment or a statistical test, draws a conclusion, writes it up as a journal article, and submits it for publication. An editor who is assigned the paper looks over it, identifies a short list of other scientists whose skills overlap with those relevant to the paper, and sends personal invitations to them to review it. Invitees then accept or decline, and those who accept volunteer their time to read the paper and consider its arguments. After usually a few weeks, the reviewers then provide summaries to the editor and the authors, including lists of their concerns or other feedback. Reviewers typically recommend that the editor accept or reject the paper. Reviewers who recommend accepting papers typically do so after first recommending revisions. Reviewers typically do not reveal their identities to authors, although occasionally they sign their reviews. In light of the reviews, and sometimes his or her own evaluation of the manuscript, the editor then makes a decision. In the event that the paper is not rejected outright, the authors typically then have the opportunity to respond to the reviewers’ concerns, with revisions or with further evidence substantiating their claims. The result often includes formal written discussion between the author and the reviewers through the editor. 

Just like authors, reviewers are human and subject to folly. Sometimes their recommendations are uninformed or wrong. Sometimes papers that perhaps should have been accepted get rejected, as well as the other way around. The peer review process is full of flaws from the initial conception of an idea through the time of acceptance or rejection, yet peer review has supported growth of the scientific enterprise and improved its capacity to build our understanding of nature.

My own experience with peer review has been interesting. As a mid career scientist, I am approaching having 50 peer reviewed papers in print. I have also served three years as an editor for a major journal in atmospheric sciences. I have personally reviewed hundreds of articles submitted by other scientists, and I received a formal editors’ award for those efforts. I have seen the process from all sides. In spite of all of its flaws, I love it!  

Peer review prevented publication of one paper I submitted as a graduate student. My conclusion turned out to be wrong, and I did not resubmit. A handful of other papers I wrote were also initially rejected, but after some additional analysis and revision of my conclusions, they were ultimately published after resubmission. I expect that others of my own papers should have been rejected, but were not. Sometimes reviewers are simply wrong. Their wrong conclusions can lead to poorly motivated recommendations of both rejection and acceptance. Yet, I have found that even when reviewers turned out to be completely wrong, my papers ultimately improved for having undergone the process. Critical reviews are essential to good science. A critical review based on good arguments is a good review. I think of a “bad” review as based on an incorrect conclusion about what results actually showed. Bad reviews sometimes lead to improper rejection of ideas. Yet, even after a bad review, I think the process as a whole still works (though not efficiently). I find that I can reflect on a bad review and revise my arguments to reduce potential that future readers come to similarly poor conclusions.

In recent years, as a professor, I have assigned research project reports to my students. I ask them in their projects to relate their own thesis research topics to those of the course. The weight placed on me to review tens of student papers at a time became too great to handle. I had to find another way to provide students with the feedback they needed. After I started serving as a journal editor overseeing the peer review process for scientists, I decided to treat the student papers as journal submissions. I receive the papers as an editor, then I invite other students in the class who wrote on related topics, to provide a formal peer review to the authors, including a recommendation to me, the editor, to reach a decision about “publication” of the paper. Regardless of my decision, the author then responds to the reviews, revises the manuscript, and sends the final product to me for a grade. The peer review process always improves the papers, improves student-learning outcomes, and reduces my workload.

I think that there is place for peer review in k-12 science and math education. In fact, I don’t think complete science education is even possible without it. Several attributes of good science education are well known. Learning to do science requires learning facts. It benefits from learned skills in technical mathematics and computer programming. It benefits from technical writing practice. But it also requires learning how to give, receive, and respond to criticism, whether the criticism is right or wrong. Effective direct criticism is essential to effective science. Science education should provide students with experience giving and receiving peer review, and responding constructively to reviews without taking them personally. Yet, peer review is almost entirely absent from the science classroom. Implementing peer review in the classroom is not without complication, but I think it is worth the effort required. In the real world, personal insults occasionally get exchanged in peer review, though they are out of place in the process. Receiving direct criticism as if it were a personal insult is also out of place in peer review. Yet, the process of making and receiving criticism could dramatically improve many aspects of the learning process. It can also lead to healthy conversation among students.

Peer review in the classroom could be blind or direct. Opportunity for direct criticism is optimum when students demonstrate techniques to each other, similar to what scientists do when they meet together for conferences. The feedback students give to each other, under loose supervision from teachers or mentors, can produce positive learning outcomes. In these settings, teachers should not usually simply jump in to correct errors made, but should help catalyze the process, letting interaction between students lead to correction. Opportunities for blind peer review could include things like inviting students to review each other’s homework. Possibilities are endless. Controlling bullying would be an important part of the process of instituting peer review in the classroom, but preventing or damping criticism should not be. Helping students to learn to not take it personally is part of the process. Peer review is full of flaws, but full of promise, and if it works for science, it should work for science education.