Variation in Herrnstein's r(o) as a function of alternative reinforcement rate.

In a test of Herrnstein's (1970, 1974) equation for simple schedules, 15 pigeons pecked a key that produced food delivered according to variable-interval schedules. One group of birds was water deprived, and food-reinforced key pecking occurred in the presence of free water. Two other groups were not water deprived; water was present for one and absent for the other. As predicted by Herrnstein, the parameter r(o) was significantly higher in the water-deprived group than in the two nondeprived groups. Contrary to Herrnstein's interpretation of r(o), the rate of drinking varied across schedules. Herrnstein's interpretation can be salvaged by considering r(o) to be an average. However, if r(o) is an average, the equation is not a good explanation of behavior because this average is not valid until all schedules have been sampled. In addition, low percentages of variance accounted for suggest that Herrnstein's equation may be of limited usefulness even as a descriptive model for these situations.