ABSTRACT
Here’s a story told to me by someone who was at the University of [bleep] when it happened. A grad student defending a thesis put up an overhead of plotted factor scores which showed the desired pattern, except for one recalcitrant subject up in the right quadrant of the plot. When grilled relentlessly by an unsympathetic committee member about the effects of outliers on such solutions, the student responded with a detailed description of how the case had been carefully examined, and every possible explanation considered, resulting in a decision to leave the subject in, especially since it didn’t seem to modify the pattern very much after all. Unfortunately for the graduate student, when he turned off the overhead projector, the fly stayed on the wall.
