摘要

Academic laboratories and private research organizations have something in common. Much of the data processing is done by students or associates who have little to no formal ethics training. Yet their work is complicated, demanding, and critical to the success and integrity of the overall project. Any unwarranted "data trimming" could alter results for the worse, discard real findings, or sully the reputation of the lab or company. One way to encourage good behavior in data analysis is to connect the individual's role to the wider effort of the lab, the field, and the scientific enterprise. By doing so, the underlying responsibilities of the data analyst are highlighted.
We have designed an introductory lecture for new scientists and staff to emphasize these responsibilities in an effective and engaging manner. The lecture is organized around four themes: (1) discovering data manipulation; (2) setting objective rules for eliminating outliers, (3) understanding the far-reaching effects of improper data analysis, (4) confronting injurious scientific fraud in biomedical science. These issues are illustrated with well-publicized episodes and personal anecdotes. Finally, we examine the reactions of the students and employees to the lecture and make recommendations for future refinements.