Monday, September 5, 2011

How to publish junk

There is a discussion in climate change online forums about the resignation of a journal editor who accepted a paper that he now decided should not have been published.

BBC: “The editor of a science journal has resigned after admitting that a recent paper casting doubt on man-made climate change should not have been published. The paper, by US scientists Roy Spencer and William Braswell, claimed that computer models of climate inflated projections of temperature increase. It was seized on by “sceptic” bloggers, but attacked by mainstream scientists. Wolfgang Wagner, editor of Remote Sensing journal, says he agrees with their criticisms and is stepping down.”

This gave me some perverse ideas. When you do bad work, how do you get it published? Here are the two key points:
(1) Choose a journal that is not specialized in the topic, and for which the subject of your paper is only marginally relevant. This avoids confrontation with experts.
(2) Carefully put together an incomplete bibliography, so as to make sure that the editor doesn't accidentally send the paper for review to people who would shoot it down.

So simple!

One may imagine other helpful steps:
(3) Get a reputed scientist to sign on as a coauthor.
(4) Pressure behind the scenes. Name-dropping; allude to large grants.
(6) Spell-checking, careful avoidance of those irritating details that so annoy reviewers.
(7) Use your own terminology to make sure that the editors doesn't stumble upon relevant papers when using a search engine to look to researchers who have worked on the subject.

So much of the field depends on people's integrity! How can we guard against such techniques?


  1. I think scientist do that to get famous and take the first step to any discovery if their study works.

  2. In my field, the Acknowledgment section usually have the line ... "thanks to our shepherd (fill in big shot Prof. here) so-and-so ...". I.e., big shot Prof. says the paper is OK, so it must be OK.

  3. I would think journals need to operate in reject-is-the-default mode. That is "the paper needs a supporter to get in", not "unless somebody shoots it down, it gets in".

    If it's submitted to a journal where nobody knows the subject matter and no expert can vouch for the paper, it should be safely out.