Publish or Politic: Referee Bias in Manuscript Review1


  • Stephen I. Abramowitz,

    Corresponding author
    1. Vanderbilt University
    Search for more papers by this author
  • Beverly Gomes,

    1. Vanderbilt University
    Search for more papers by this author
  • Christine V. Abramowitz

    1. George Peabody College for Teachers
    Search for more papers by this author
    • 3

      Gratitude is extended to Barbara Wallston, Kenneth Wallston, Donald Bowen, Kenneth Anchor, John Harvey, Carolyn Jackson, Seymour Lipset, Howard Roback, Stanley Sadava, John Siegfried, Brewster Smith, William Smith, Donald Thistlethwaite, Lawrence Wrightsman, Lynn Cocke, and Betty Keeling.

  • 1

    An earlier version of this paper was presented at the 140th Meeting of the American Association for the Advancement of Science, San Francisco, 1974.

2 Requests for reprints should be sent to Dr. Stephen I. Abramowitz, Sacramento Medical Center, University of California at Davis, Sacramento, California 95817.


To clarify the role of cognitive bias in manuscript review, designated more and less politically liberal area specialists and nonspecialists were sent either of two versions of a brief empirical report contrasting student political activists' and nonactivists' psychological well-being. The forms were identical, except that all references to activists and to nonactivists in the results and discussion sections were interchanged. The referees, led to think that they were participating in a study of the usefulness of a closed-ended manuscript evaluation procedure, were asked to rate the article's publishability and the degree to which specific criteria for scientific quality were met. The main results confirmed the expectation that publication verdicts handed down about a paper containing information of social import are sometimes biased by a reviewer's political orientation. However, the results were largely null regarding the susceptibility to value-intrusion of scholarly inferences anchored to discrete criteria. Ego involvement is implicated as a mediator of biasing effects in the editorial process.