Year: 1961
Type: article
Abstract: Either directly or indirectly, measurements of operant behavior are usually in the form of probabilities of occurrence of the behavior.Response rate, resistance to extinction, per- centage of correct choices, and sundry error scores are included in the class.Latency of re- sponse, commonly used where running behavior is under study and also in reaction time and avoidance paradigms, is a related subgroup in which the probabilities of occurrence are supplemented by reference to an initial zero point in time.Although the concept of probability of occurrence has proved useful as an index to the effects of various experimental operations (Skinner, 1957), many properties of learned be- havior are not well elucidated by these measures, and, furthermore, have received little theoretical or experimental treatment.For example, the probabilities of the individual- component response members cannot easily describe serial dependencies among different responses, as in the chaining of behavior.Rather, as Frick and Miller (1951) have shown, the transitional (conditional) probabilities among different response classes must be con- sidered in order to quantify such sequential effects.Furthermore, when the variability in successive members of the saIne response class is of interest, measures such as response force (Notterman, 1959) and speed (Dinsmoor, 1958), spatial location of the response (Antonitis, 1951; Herrnstein, 1959), and bodily orientation at the time of the response (Antonitis, 1951;Guthrie & Horton, 1946) have been adopted.These measures are not entirely without difficulties.Calculation of empirical estimates of transitional probabilities between two or more mutually exclusive response classes becomes exceedingly tedious when carried beyond the second or third order of approximation.All the various response-variability measures present problems of instrumentation.The present report describes the behavior of several more measures, which were observed under the relatively simple procedures of continuous reinforcement and a subsequent extinction ses- sion.The merit of these measures is that they are easily summarized quantitatively and simply instrumented, and furthermore appear to be sensitive to reinforcement operations. METHOD ApparatusThe basic apparatus was an experimental chamber, 9 inches wide by 9 inches long by 8 inches high, constructed of lucite plastic, with a grid floor and a detachable aluminum front wall from which a stainless steel counterbalanced lever 2 inches wide by 0.75 inch long by 0.375 inch thick projected.When depressed with a downward force of about 1.5 grams through a distance of about 1 millimeter, this lever closed a microswitch.Actuation of a solenoid dipper delivered 0.025 milliliter of Nestle's condensed milk diluted with equal parts of tap water through a 0.25-inch hole in the floor of a tray recessed in the front wall below and just to the left of the lever.If nosed open 0.5 inch with about a 5-gram force, a small, hinged transparent lucite door secured access to the milk tray and closed a microswitch.A
Authors: J. R. Millenson, H. M. B. Hurwitz
Institutions: University of London, Birkbeck, University of London
Language: English
Subfield: Developmental and Educational Psychology
Field: Psychology
Domain: Social Sciences
Open Access status: bronze
OpenAlex ID: W1971797508