(An) analysis of image-related research misconducts in a retraction database

Abstract

This thesis examines image-related research misconduct cases that have occurred more frequently in recent decades with the dissemination of digital image editing technology and attempts to contribute to the policies to address the problem with an analysis of the attributes of the misconduct cases. Scientific images are a tool for communicating research results clearly, but the criteria for their creation and evaluation are often transferred tacitly between researchers. With the increasing frequency of research misconducts by individual researchers, laboratories, or institutes, several countries and institutions have defined and categorized research misconducts in the 2000s. But the actual occurrences of research misconduct are not accurately recorded or measured. By examining the guidelines and policies for ensuring research integrity, it is possible to establish five categories for empirically analyzing misconduct cases. The criteria have been applied to the retraction cases in four countries - Korea, Taiwan, China and Japan - retrieved from an online database. It has been found that image-related research misconducts are not the most common type of misconduct, but still occupy a sizeable portion of all research misconducts. More than half of the image-related cases fall in the categories of plagiarism, fabrication and falsification of images. Through this analysis, this thesis proposes a comprehensive research management policy that can decrease the number of image-related research misconducts in laboratories, research institutes and peer review processes with an emphasis on existing techniques and systems.