In each baseline session, pigeons were exposed to a multiple schedule in which each of five distinctive stimuli was correlated with a different frequency of reinforcement. In one component, responses were reinforced with a probability of 0.10 (random-ratio schedule); in the other four components, responses were reinforced with different scheduled temporal frequencies averaging 30 to 240 sec between reinforcements (random-interval schedules). For periods lasting 30 sessions, contingent reinforcement was discontinued and reinforcement was presented independent of responding at irregular intervals averaging 30, 60, or 120 sec, while the sequence of stimuli continued. After each such period, the baseline was reinstated for 30 sessions. The data indicated that: (1) The rate of responding in the presence of all stimuli decreased as exposure to the non-contingent reinforcement procedure was prolonged, at all the frequencies of reinforcement employed; (2) The rate under the random-ratio schedule declined faster than the rates under all the random-interval schedules, presumably because the decrease in reinforcement frequency under this stimulus condition was greatest; (3) The decline in rates of responding under the stimuli correlated with the random-interval schedules tended to be greatest for the stimuli paired with the lowest frequencies of reinforcement.Implicit in the concept of a reinforcement contingency, is the notion of a specific temporal relationship between a response and a reinforcer. Since reinforcement must always occur in close temporal contiguity with some response, it can be assumed that conditioning is taking place whenever a reinforcer is delivered. As Skinner (1948) pointed out, whether or not a response is pre-specified in no way alters the power of reinforcement to exert its effect. When reinforcers are delivered without reference to the ongoing behavior, the resultant effects are said to be produced by "superstitious" or non-contingent reinforcement.More specifically, a non-contingent schedule of reinforcement may be defined as one in which the distribution of interresponse times in no 'These data are based on a dissertation submitted to

Two White Carneaux hen pigeons were exposed to a 60-sec random-interval baseline procedure. Six different exteroceptive stimuli were successively correlated, within a single session, with blocks of 10 reinforcement presentations. Following this training, a noncontingent reiniforcement procedure was instated with inter-reinforcement intervals of 5, 15, 30, 60, 120, and 240 sec. Within a single session, each non-contingent frequency was correlated with one of the previously presented discriminative stimuli. After an initial increase in the rate of responding as the result of a high density of non-contingent reinforcements, the rate declined as exposure to each non-contingent frequency was prolonged.

The effect of amount of student-proctor interaction was investigated within the framework of Keller's (1968) method of personalized instruction. College students enrolled in introductory psychology were randomly assigned to five groups: 0%, 25%, 50%, 75%, and 100%, reflecting the percentage of units on which each student was proctored. The results indicated that (a) the proctored students were superior to the nonproctored students as measured by final examination performance, (b) for the proctored groups, the amount of proctoring did not differentially affect final examination performance, and (c) the major effect of increased proctoring was an acceleration of the rate of progress through the course.The variables emphasized in traditional teaching methods have recently been subjected to critical scrutiny. Both in the laboratory and the classroom, behavior change can be effectively evaluated and controlled only after some objective behavior has been selected and reliably measured. Such objective or active responding is not emphasized in the lecture system in which the student is treated as a passive recipient of information . More importantly, events or procedures capable of producing those changes in behavior that are the concern of any teaching method have seldom been detailed in terms of their schedule of prelPortions of this paper were presented at the meetings of the Eastern Psychological Association, Washington, D. C., April, 1968. 2The final version of this manuscript was completed without the guidance of its originator, Dr. John Farmer. His untimely death deprived his coauthors and colleagues of his innovative and constructively critical approach to the teaching methods discussed and evaluated here. The co-authors wish to acknowledge their immeasurable debt, express their gratitude for this opportunity of collaboration, and assume full responsibility for the statements to which any criticism may be addressed. sentation; nor have investigators described the effects of these procedures on the behavior of the individual student, behavior from which academic achievement is then inferred (Skinner, 1968). Personalized instruction (Keller, 1966), however, focuses upon the specific objective behaviors of the individual student. These behaviors are differentiated and maintained by presentation schedules of classroom events (such as a passing test grade) that function as reinforcers. Recent investigations of personalized instruction have shown this procedure to be more effective than the traditional lecture system in the following ways: (a) students earned higher grades in personalized instruction courses than in lecture courses (Keller, 1966(Keller, , 1968; (b) final exam performance was better after personalized instruction courses than after lecture courses (McMichael and Corey, 1969); (c) in a retest one full semester after completion of the course, superior performance was maintained among students who had received personalized instruction (Corey, McMichael, and Tremont, unpublished).

No abstract