Very early in my career, I was teaching the research methods class required of our majors. The class required students to complete several demonstration research projects, each requiring collection of a small amount of data. One of the projects involved a demonstration of the illusion of control. It involved having participants express their confidence in the predicted outcome for a roll of a die – a chance-determined event over which they could have no actual control. On each of 10 rolls, participants rated their confidence in a predicted outcome, witnessed the roll of the die, and then recorded the result. The manipulations involved a 2 × 2 factorial. Some participants chose the number they thought would show up on each roll; other participants had the number predicted for them (actually the sequence chosen by a previous participant). Independently, some participants rolled the die themselves; other participants had the die rolled for them by the experimenter. The prediction was that participants would be more confident of the predicted result if they chose it themselves and rolled the die themselves – a replication of previous research. Each student in the class collected data for one participant in each of the four conditions.

When the data were assembled, one glaring problem emerged – the data set submitted by one student matched exactly the data set submitted by another. Given the number of trials involved, the probability of this actually occurring by chance are vanishingly small. I called the students in for a meeting, presented my concerns, and after first denying any wrongdoing, they finally confessed to having fudged the data (and then simply copying it for the second set). I referred the matter to our university academic integrity committee, and following a hearing of the case, that committee recommended that the students fail that particular project (I recommended to the committee that they fail the class, given the nature of the offense and class).