Just how prevalent is plagiarism? At a meeting devoted to the topic at New York University last month, Alan Price of the Office of Research Integrity (ORI), which primarily handles complaints in biomedicine, reported that in the past 16 years, only 5–12% of its misconduct cases each year involved plagiarism. This is defined by the ORI as “the appropriation of another person's ideas, processes, results, or words without giving appropriate credit”.

On the other hand, James Kroll, head of administrative investigations at the US National Science Foundation, revealed that more than 60% of its misconduct findings concern plagiarism. And earlier this year, the National Natural Science Foundation of China reported that plagiarism accounted for about one-third of its misconduct cases in the past six years.

Human nature hasn't changed recently, but reusing with the intent to deceive seems to be on the rise, both in the literature and in grant proposals. The replacement of pen and paper with software makes it far easier to slip in large sections of text. Internet connectivity, online repositories and sophisticated search tools provide almost irresistible accessibility to the polished thoughts of others.

Students trained today have grown up in an environment where access is taken for granted and attribution only loosely enforced. So they need more rigorous instruction than their predecessors regarding the ethical standards expected of them. Mentors must counter the ever-rising promotion and funding pressures that reward prolific publication rather than support creative quests.

Although the development of web-based tools that can recognize text-based plagiarism will eventually help detection, more can be done before that point. Some common-sense guidelines need stressing at the bench, long before the data or grant application are written up. Copying text, even when supplying new data, is not acceptable without clear reference to the process. One duplicate figure in a paper is one too many, if attribution to the original paper or grant is not noted. Oblique reference to a method in a previous publication in an attempt to hide the paper's intellectual precedents is still deceitful and a form of plagiarism.

Editors are obliged to act if concerns are raised about improper attribution.

Editors have an obligation to act if concerns are raised about improper attribution. If authors do not supply satisfactory explanations, their employers and funding agencies must be notified. It is the responsibility of institutions, who have a legal mandate, to initiate a formal investigation.

Timeliness can be difficult if institutes are reluctant to taint their reputations with negative findings, or if international boundaries are crossed. Editors should nudge investigations that drag, and draw attention to incidents where no satisfactory progress is made.

Where plagiarism is found, the author's previous publications must be examined. The evidence shows that an act of misconduct is usually part of a pattern of behaviour rather than an isolated incident, says Richard Smith, former editor of the British Medical Journal.

Journals should proceed promptly to correct the literature where discovery of misconduct necessitates it. Plagiarized text or figures should be clearly indicated as such within the original content. Nature will play its part where necessary, as will other Nature titles. One might hope that such public humiliation will act as a deterrent to those inclined to pass off another's work as their own.