When people work together, there can often be disputes about who originally generated an idea. This has been formally tested in experimental studies that show that people often recall other people’s ideas as their own, despite being instructed not to. This is called unconscious plagiarism, and has been suggested to stem from a self-serving bias. It has shown to be pervasive across a range of tasks, to be more likely with a longer delay between the original idea and the recall, to be more likely if the originator of the idea is the same sex as the person recalling, and more likely if the person who originated the idea is absent when the person recalls it as their own.
In a series of experiments were conducted to reappraise our understanding of unconscious plagiarism. First, we challenged the notion that it is a self-serving bias, by showing that people are more likely to give away their own ideas (when they try to recall their partner’s ideas) than they are to steal their partners ideas (when trying to recall their own ideas). This pattern was replicated in a series of experiments that explored other aspects of unconscious plagiarism. Both forms of error were shown to increase with a delay, to be greater if the partner is absent at retrieval, and to be greater if the partner was of the same sex.
One hypothesis is that the tendency to steal (or give away) ideas is driven by the goals of the recall task: that is when trying to recall one’s own ideas people accidentally steal because the goal is to recall own ideas. Conversely, when recalling a partner’s ideas, ideas are given away because the goal is to recall the partner’s ideas. However, we ruled out this hypothesis by demonstrating that both the rate of errors, and the bias in errors (with more given away than stolen) replicated in studies which asked participants to recall ideas from both sources simultaneously.
We also explored the extent to which the errors could be traced to faulty retrieval processes, or to a failure to properly monitor the source of each idea that comes to mind. The data strongly suggest that faulty generation causes the difference between ideas being stolen and given away, and not monitoring. When people were asked to list all ideas that came to mind, they were more likely to generate their own ideas (when trying to recall their partner’s ideas) than their partner’s ideas (when trying to recall their own). For the ideas that were generated, the ability to distinguish where they came from was equal across sources. Thus, people give away more ideas simply because their own ideas come to mind more readily, even when unwanted.
Finally, we explored the effects of delay and partner presence on the cause of errors. The presence or absence of a partner had no impact upon the generation of ideas initially, but did influence how these were monitored. When a partner was present, people were more cautious about reporting items, and thus avoided errors. In contrast, delay both decreased the number of ideas that came to mind, and decreased the ability to monitor their accuracy. Thus, despite fewer ideas coming to mind, more errors resulted.