Four experiments were conducted to investigate the development of stimulus classes using matching-to-sample procedures. The first three experiments involved teaching subjects to select four different comparison stimuli in response to a single sample stimulus. In the fourth experiment, the subjects were taught to select a single comparison stimulus in response to four different sample stimuli. Only one of the six subjects who received single-sample, multiple-comparison training demonstrated stimulus class development. All three subjects who were given training in selecting a single comparison stimulus in response to multiple samples demonstrated stimulus class development.
In the present study, we explored the feasibility, effectiveness, and acceptability of teaching seriously mentally handicapped individuals to provide interactions and materials to multiply handicapped retarded persons during leisure periods. With simple and economical teaching procedures, five seriously retarded individuals were taught to engage in levels of interaction which exceeded standards established by exemplary employees. Their efforts systematically resulted in higher levels of material availability to recipient clients. The behavior of the client providers was maintained under realistic supervision conditions, which staff reported were practical in the settings in which these clients functioned. Information from questionnaires indicated that client providers enjoyed their work, that staff valued their contributions and that human rights representatives viewed the effort as worthwhile to client providers and recipients alike. The potential benefits and limitations of such an approach are discussed and areas of future research are described.
In each of two experiments, two teachers, Teachers 1 and 2, taught subjects to respond to requests for the opposites of several different spoken words with answers from one of two different response repertoires, Repertoires 1 and 2. In the main experiment, two retarded adolescents and one normal 4-year-old served as subjects. Teachers 1 and 2 were puppets. Two other puppets, prober-puppets, never taught the subject to respond to those questions about opposites. Through visual-visual matching-to-sample, auditory-visual matching-to-sample, and oral-naming teaching, two classes of puppets were established. One of the classes contained Teacher 1 and one prober-puppet, Prober A. The other contained Teacher 2 and the other prober-puppet, Prober B. The formation of those classes was demonstrated on unreinforced probe trials in which the two prober-puppets asked the subjects the questions about opposites previously taught by the Teachers 1 and 2. On those trials, Prober A was responded to with Repertoire-1 responses (taught by Teacher 1) and Prober B was responded to with Repertoire-2 responses (taught by Teacher 2). Subsequent replications (two with the adolescents and one with the 4-year-old) reversed the teacher-prober classes through matching-to-sample and oral-naming teaching and found similar results. In a second experiment, normal adults served as Teachers 1 and 2. A retarded adolescent was the subject. Matching relations were not established in this study. Instead, generalization of Repertoire 1 or 2 to a third person, the Teacher-Prober, was produced by having the Teacher-Prober teach only portions of Repertoires 1 or 2, respectively. Apparently, the Teacher-Prober could form a stimulus class with a particular teacher by teaching only some of the same responses as that teacher. These experiments show the kinds of reinforcement contingencies that can establish stimulus classes of listeners and thereby enable language repertoires to generalize across those listeners.
In recent years, concern has developed regarding the degree to which training with a limited set of objects will result in generalized responding to novel objects of the same category. The purpose of this study was to expand our understanding of the benefits of training with multiple exemplars. Specifically, training with five as opposed to three good examples was investigated. The difference in the averaged correct response to novel objects between conditions was not significant (p = .08); however the arithmetic difference between the two conditions indicated higher levels of generalization following training with five examples for five of the six subjects. The results are discussed in reference to implications for teaching and for further research regarding the benefit of using multiple exemplars when teaching natural object categories.
Analyses of response relationships offer the potential for designing treatment procedures that are more economical and entail less risk for misuse than many we currently employ. The present study investigated whether previous results indicating the reduction of problem behaviors by reinforcing compliance in fact occurred because of the contingent relationship between compliance and the presumed reinforcing events. A clinic analogue procedure consisting of a standard set of adult requests was employed in a multiple baseline design across subjects with three children receiving three conditions — no reinforcement, noncontingent reinforcement, and reinforcement contingent on compliance — and a fourth child receiving the first and third conditions to assess possible order effects. The data indicated that: (1) compliance covaried inversely with some, but not all, of the problem behaviors measured; (2) sustained covariation occurred only when reinforcement was contingent on compliance; (3) substantially large but transient increases in compliance and decreases in problem behavior during noncontingent reinforcement occurred for those children who initially demonstrated high compliance and low levels of aberrant behavior during baseline; and (4) the problem behaviors that inversely covaried with compliance were idiosyncratic across children. The results suggest that treatment procedures based on response relationship strategies should consider that children's individual learning histories may determine which behaviors covary and the likelihood that large, initial changes can occur with noncontingent reinforcement.
This study evaluated the effects of access to opportunities to earn edibles for performing a high probability stereotypic behavior on the public genital stimulation and stereotypy of a severely retarded male in a special education class. The treatment involved the use of a series of increasing DRO lengths in which the absence of genital stimulation was reinforced with edibles and a stereotypic behavior. Later, we attempted to reduce the rate of stereotypy by withdrawing the edibles that had been provided for it during the reinforcement period. The results indicated that (a) genital stimulation was nearly eliminated with the graduated DRO procedure; (b) withholding the edibles had little effect in reducing the stereotypic behavior; and (c) edibles alone were enough to maintain the reductions in genital stimulation. These findings suggest that public genital stimulation is amenable to treatment with a positive treatment procedure, but more research with the reinforcer displacement technique is needed before any firm conclusions can be drawn about its effectiveness in an applied setting.
The chapter discusses the controlling factor in conditional discriminations and tests of equivalence. A conditional discrimination arrangement is very complex, but the chapter is confined to procedures in which the number of samples (in successive trials) equals the number of comparisons. All comparisons are presented in every trial, and the relation between each sample and the respective comparison is arbitrary. Assuming that a conditional discrimination establishes a relation between each sample and the comparison stimulus selected in its presence (abbreviated, in the simplest case, as AlrB1 and A2rB2), this relation would have the property of reflexivity if each stimulus bore the relation to itself. Symmetry is demonstrated by the reversibility of sample and comparison functions—that is, conditional discrimination. A necessary step toward solving the controversies over the nature of equivalence and its necessary prerequisites would be to examine the nature of controlling relations in trained conditional discriminations and in untrained conditional discriminations. This shows the need to re-evaluate data showing apparent failures or successes in equivalence tests. Similar performances are brought about by different controlling factors, including, for language-able humans, verbal behavior. Verbal behavior controls subsequent behavior, generating response topographies similar to those that are shaped by contingencies, and this also happens in the conditional discrimination paradigm.
Two normal children and four retarded youths were taught auditory-visual and visual-visual conditional discriminations and were then tested to determine whether the conditional relations were also equivalence relations. All subjects showed that conditionally related sample and comparison stimuli had become equivalent; the conditional discriminations generated two sets of three-member classes, one set containing both auditory and visual stimuli and the other containing only visual stimuli. In the auditory-visual classes, two of the retarded subjects failed to apply the name of the auditory stimulus to each visual stimulus in the same class; in the visual classes, all but one normal subject failed to apply a consistent label to all stimuli in a class. The formation of classes of equivalent stimuli does not, therefore, require that common names be applied to each member of a class. Although common naming responses were not necessary for equivalence, relations denoting symmetry and transitivity in the original conditional discriminations were shown to be prerequisites, as is demanded by the definition of equivalence.
A stimulus-equivalence paradigm was used to establish match-to-sample relations between members of two separate stimulus classes through a procedure involving direct training and observational learning. Four retarded adult males were taught to conditionally match arbitrary visual stimuli. Each subject was taught one sample- comparison relation within each class (e.g., B1A1, B2A2) and given an opportunity to observe another subject perform a second sample-comparison relation within each class (e.g., C1A1, C2A2); each observed relation had one stimulus in common with a sample-comparison relation acquired through direct training (A1, A2). When subjects acquired the trained and observed relations, they were tested on the sample-comparison relations that were never taught directly (AB, AC, CB, BC). Initial testing on the BC and CB relations revealed chance performance for two subjects. Subsequent testing on the AB and AC relations, however, resulted in immediate acquisition of the CB and BC derived relations. Results indicated that stimulus-equivalence teaching procedures used in academic group settings have promising educational implications for producing learning through observation of others.
Five rehabilitation clients were trained to use written prompts to (a) plan nutritious meals for 1 week, (b) write accurate grocery lists based on their menus, (c) stay within a $25.00 weekly food budget, and (d) shop effectively and systematically. Training occurred in a classroom using group instruction for the first three response classes above. The shopping response classes were trained in a community grocery store. Training consisted of experimenter modeling, subject rehearsal, as well as experimenter and peer feedback in a multiple probe design. Subjects' meal plans were nutritionally deficient in three of the four food groups during baseline, but met or exceeded the four food group criterion after instruction. Grocery list writing skills improved from a mean of less than 30% correct during baseline to more than 90% correct after training. Similarly, subjects' grocery lists cost an average of 72% over the $25.00 budget before training, and 19% under the budget after training. Subjects' mean percent of effective and systematic shopping increased to 91% and 100%, respectively, following instruction. Generalization of shopping skills from the training store to two stores in which no training occurred was observed. Follow-up tests for all subjects as well as community probes for two subjects living independently showed that all skills maintained at acceptable levels.
Three mentally retarded subjects were taught with matching-to-sample procedures and geometric shapes as stimuli to make matching and mismatching responses, respectively, in a two-component, multiple schedule. The subjects also made generalized matching and mismatching responses to unreinforced probe samplees in the respective components. During each trial the subjects were given the opportunity to choose the sample stimulus to which to respond. Choice responding was compared under conditions of no intertrial interval and 5-sec and 10-sec interial intervals. Generally, longer intertrial intervals resulted in a low probability of probe-sample selection and a high probability of reinforced-sample selection. Nevertheless, the subjects continued to display generalized matching and mismatching behavior when probe samples were selected, demonstrating that generalized matching and mismatching could be maintained concurrent with a discrimination between the reinforced and unreinforced sample stimuli as measured by the choice procedures. Further, the patterns of responding in the mismatch-reinforced component for all subjects were observed to change as intertrial interval conditions were changed across sessions. Under longer intertrial interval conditions, subjects responded to specific choice stimuli as mismatches for specific samples, but reverted to other strategies under no intertrial interval conditions. The results suggest that intertrial interval length may be an important variable in both matching-to-sample and choice behavior.
The purpose of this study was to determine the generality of Sidman's (1971) equivalence model when applied to vocabulary acquisition in deaf preschool children. The model specifies six vocabulary tasks involving auditory and visual sensory modalities. The data confirmed that lip reading (visual modality) could be substituted for auditory input (speech). However, for novice lip readers, words with auditory-visual confusions depressed performance on receptive vocabulary tasks, in contrast to the emergent expressive performances. The results are discussed with respect to the applicability of the equivalence model to education of deaf children using a computerized system. It is suggested that the computer system provides a standardized method for assessment of mechanical and cochlear implant prosthetic devices.
Anecdotal and scientific reports of the speech of autistic children suggest a possible differential responsiveness to limited components of speech stimuli. Previously nonverbal children who acquire speech seem to acquire content, but their speech is dysprosic, characterized by inappropriate intonation, pitch, rhythm, and modulation. In contrast, echolalic autistic children often display modulation and varied intonation but may demonstrate no comprehension of the verbal stimulus being echoed. The present experiment was conducted to systematically determine whether echolalic and nonverbal autistic children respond overselectively to the intonation or to the content of a complex auditory stimulus differing along these two dimensions. Ten autistic children (five echolalic and five nonverbal) and six normal children, matched for MA, were trained to bar press at the sound of an auditory complex stimulus designated as S+ and to withhold response to an auditory complex stimulus designated as S-. The auditory complex stimulus consisted of two components, content and intonation. Test trials consisting of various combinations of the intonation and content components were then presented. Results indicated that whereas the normal children responded to both components or to content, the nonverbal children showed evidence of selectively responding to content, and the echolalic children showed evidence of selectively responding to intonation. These results are discussed in terms of relating autistic stimulus overselectivity to deficits in speech production and to implications for designing language training programs for these children.
Little is known about the acquisition of stimulus control in autistic children. Many studies comment on the incidental, unintended, and often bizarre stimulus control acquired when normal discrimination learning procedures are used. Earlier studies suggested the problem may involve autistic children's inability to respond to multiple, simultaneous cues. While this has been clearly demonstrated in some instances, subsequent data have shown that autistic children sometimes do respond to multiple cues. The question remains, then: What variables influence stimulus control in autistic children? The present study examined one alternative, the possible effects of the absolute distance between cues, in an attempt to better understand (and program) stimulus control in autistic children. Three autistic children were trained to select a card containing a stimulus array comprised of three visual cues. The distance of each cue from the center cue was varied, using small, medium and large distance conditions. Stimulus control probes were conducted for each distance condition to assess which of the features the children learned about. The results revealed that the absolute distance between the cues determined the number of stimulus features to which the autistic child responded; decreasing the distance between cues served to increase the number of cues that controlled responding. The distances used in this study did not, however, affect the responding of MA-matched and CA-matched normal children. The possible role of “tunnel vision” and “the distance between cues” in treating and understanding autistic children are discussed.
Eight handicapped youths were taught arbitrary matching between visual sample stimuli and visual comparison stimuli: Given Sample A1, selecting Comparison B1 was reinforced; given A2, selecting B2 was reinforced. During nonreinforced test trials consisting of novel samples, novel comparisons, and familiar B comparisons, subjects selected the novel comparisons. Compared to intervening preference tests between novel comparisons, these results suggested that selecting the novel stimuli was controlled by the B comparisons, a phenomenon called exclusion. Some subjects continued to select the same novel comparisons during subsequent preference tests, suggesting that control by exclusion may result in arbitrary matching in new contexts. emonstrated symmetry of AB matching (i.e., BA matching) suggested that A stimuli should also control novel comparison selection; however, these subjects did not show exclusion, but most often selected the A comparisons.
To determine how stimulus control of eye orientation was established and how that process affected learning, an analysis of normal preschool children's eye orientations and pointing responses to each of two simultaneously presented stimuli (S+ and S-) was conducted. Using a probe design to test for acquisition, each subject was trained on two tasks. In one task, S+ (selections of this stimulus produced reinforcement) was held constant at the final criterion level while S- (selections of this stimulus never produced reinforcement) was manipulated using stimulus shaping procedures. The second task was programmed with the opposite method (S+ shaped while S- held constant). Results showed that more pointing response errors, more occurences of eye orientations, and longer durations of eye orientations characterized conditions where S+ was held constant and S- was shaped. The individual analysis of the eye orientation data indicated that differences in eye orientation to stimuli under the two programs were largely due to relatively higher frequency of orientation to the S- stimulus under program condtions that shaped S- and held S+ constant.
This experiment evaluated parent and teacher use of reinforcement-based, reductive procedures to treat aggressive pinching in a severely retarded child. A withdrawal and a multiple baseline design were used for the analysis. Substantial and lasting treatment effects were obtained at home and school. The benefits and problems associated with parent use of reinforcement-based, behavioral procedures with severely handicapped children, as well as the comparative effects of DRO and DRI are discussed.
