ABSTRACT
Linear contrasts are comparisons between linear combinations of different groups. Suppose we want to know whether students are more bored on Wednesdays than other weekdays, because Wednesday is statistics day, and whether they're more bored on weekdays than weekends. We could measure their boredom on all days of the week, and use DayOfWeek as a factor (with 7 levels) in an ANOV A. If this turned up significant, we would know that all days were not the same-but it wouldn't answer our original questions. We can do that with linear contrasts. See also Howell (1997, pp. 354-369); Myers & Well (1995, ch. 6).
