A procedure to effectively and efficiently train moderately and severely retarded individuals to make fine visual discriminations is described. Results suggest that expectancies for such individuals are in need of examination. Implications for sheltered workshops, work activity centers and classrooms are discussed.
Sorting tasks are very common in school and workshop programs for the retarded. In schools, trainable retarded children learn to sort on many dimensions including size, shape, color and category. The use of systematic training rather than merely exposure is, however, seldom found. The level of difficulty, in addition, is usually kept very low.
In sheltered workshops and work activity centers, sorting is usually done as a subcontract. When the required discriminations are perceived by staff as difficult, one of two alternatives is usually followed: either the subcontract is rejected on the grounds that it is too difficult for the clients, or a jig is made to insure the correct decision. (A jig is a device used as a guide or template.) For example, to sort piles of bolts which are identical in every way except one-eighth inch difference in length, a board with a hole the depth of the shorter bolt might be used. The sorting decision would be made on the basis of whether or not the bolt went all the way down in the hole. The position taken here is that the use of jigs in sheltered workshops and vocational training programs often results in reducing the habilitative value of the contract. Instead of teaching clients the discriminations and movements required to perform the task, they are allowed to complete the contract using skills already in their repertoire. In addition, jigs are often slower because of the added movement (Maynard, Stegemerten & Schwab, 1948). One other reason for the frequent use of jigs in habilitation programs is that professionals would not expect the retarded to be able to accurately and consistently sort a pile where the difference between bolts was one-eighth of an inch, especially if the professionals had attempted the task and experienced difficulty.
There were several purposes to the present study: (1) to demonstrate the ability of moderately and severely retarded individuals to learn a difficult visual discrimination; (2) to apply a piece of information from basic learning research to a task found in applied settings; and (3) to demonstrate the effectiveness of the particular training procedure. The procedure used was based on information from the Attention Theory of Zeaman and House (1963). According to the Theory a series of easy-to-hard discriminations is learned more quickly than a hard-only discrimination (House & Zeaman, 1960; Shepp & Zeaman, 1966; Zeaman & House, 1963). For a practitioner-oriented description of the theory the reader is referred to Gold and Scott (1971).
The independent variable was procedure (easy-to-hard sequence or hard task only). Criterion for learning was defined as one errorless sorting of the pile on the one task, for the Hard Group, and on each of three tasks for the Easy Group. The dependent variable was trials to criterion, where the decision on each bolt constituted a trial.
Sixteen moderately and severely retarded adolescent and adult individuals were randomly selected from the population of a work activity center located in an industrial midwestern city of about 40,000 residents. The mean IQ score for the sample was 47.5; the age range was 17 to 59. Following selection, subjects were randomly assigned to one of the two groups. A replication of this study was conducted subsequently, using mildly and moderately retarded clients from a workshop in Peoria, Illinois. For this group, the mean IQ score was 58; the age range was 20 to 57.
The apparatus was a 2 by 3 foot plywood base with three trays—one for the mixed and two for the sorted Cap-Screws-Hex Head bolts. Twenty-five of each of four length bolts were used—one and one-half inches, one inch, seven-eighths of an inch, and three-quarters of an inch. All were one-quarter of an inch in diameter.
Each subject was seated at a table on which the apparatus was placed. The experimenter sat across the table facing the subject. The experimenter picked up one bolt, looked at it, and placed it in one of the two empty trays; then he picked up a bolt of the other length and placed it in the remaining empty tray. The experimenter then said to the subject, “Now, you do it.” When the subject made an error, the experimenter removed the bolt from the wrong tray, handed it back to the subject and said, “Try another way,” while pointing to the correct tray. No other verbal directions or cues were provided.
Subjects in the Hard Group were brought to the criterion of one errorless sorting of a pile of 50 bolts containing 25 each of one inch and seven-eighths inch bolts. Subjects in the Easy Group were brought to criterion on each of three piles, starting with one and one-half inches and three-quarters of an inch, then one inch and three-quarters of an inch, and ending with one inch and seven-eighths of an inch. Subjects failing to reach criterion by 15 sortings of the pile (750 trials) were stopped and given a score of 750. For the Easy Group this meant having to learn all three tasks within 750 trials. An independent groups design was used.
For the moderately and severely retarded work activity center clients, no subjects in the Hard Group learned the task. All subjects in the Easy Group learned all three tasks; their mean trials to criterion for the three tasks combined was 394; the range was 200–750. The difference was analyzed using a single classification Analysis of Variance, and found to be highly significant (F1,14 = 32.72, P < .001). All of the subjects in the Hard Group were subsequently re-run, using the easy-to-hard sequence. Two subjects still failed to reach criterion. The other six reached criterion, with a mean of 358 trials (Range = 250–450, not including the two failures).
For mildly and moderately retarded subjects in the replication study, the mean trials to criterion for the Hard Group were 281 (Range = 50–750). Mean trials to criterion for the Easy Group were 413 (Range = 200–750). The difference was not significant (F1,14 = 1.47, N.S.). Two subjects in the Easy Group and one subject in the Hard Group failed to reach criterion. None of the subjects in the replication study were rerun.
For clarity, the replication data will be discussed first. The replication data contained considerable variance and no indication of treatment effect. This is especially interesting in light of the magnitude of the difference found in the first study. The interpretation presented here is that the sheltered workshop population from which the replication sample was drawn may have contained considerable heterogeneity on variables, such as attention span, motivation or coordination, which were not assessed, but, which could affect performance on the task.
The same problem has been noted with institutionalized populations on other, more complex tasks. That is, between-subject variance was great, and statistically significant differences were not found. New studies are in progress on the problem of determining the nature of qualitative rather than quantitative differences between populations. This phenomenon has been found in several studies, however, which suggests that the data from the first study can still be used with confidence.
For the work activity center clients in the first study the effectiveness and efficiency of the easy-to-hard sequence was clearly demonstrated. Had the conventional approach been used where the learners were exposed to the criterion task only, the conclusion could have been that they were not capable of learning the task. Such a conclusion is often made in the absence of alternative methods of instruction. The performance of the Hard Group on the easy-to-hard sequence, following failure, further supports the procedure and suggests something else. Two of the subjects never reached criterion. The other six took approximately the same amount of time as the Easy Group, even after 750 trials on the third of the three tasks. This is interpreted as an example of failure set. Failure set refers to a situation where the retarded individual, having experienced prolonged failure on a problem which he cannot solve, finds it difficult to solve very simple problems, which he could solve, prior to the failure experience (Zeaman & House, 1963). The data presented here might be interpreted to suggest the importance of selecting a training procedure in advance of training, rather than after the learner has experienced a procedure and not learned the task. This means identifying and having some idea of the relative strength of applicable procedures. This study and the works by Brown and Foshee (1971), Gold and Scott (1971), Gold (1968; 1972) and Touchette (1968; 1969) provide information on a variety of training procedures.
An important implication of this study relates to the difficulty of the discriminations acquired by the subjects. Differentiating differences of one-eighth of an inch, without having a sample against which to compare has been demonstrated to be an initially difficult task for normal individuals. The position taken by the writer is that intelligence is related to the time it takes to identify what the problem is, but not to the discrimination itself. Therefore, moderately and severely retarded individuals should be expected to make the same visual discriminations as normal individuals, but require more assistance in guiding their attention to the relevant aspects of the problem. The procedure described in this report might have application to other situations where visual or other sensory discriminations are required, such as letter recognition, cooking, and bench assembly work.
This study also suggests several possible changes in current sheltered workshop and work activity center practices. Perhaps tasks which are presently refused or “jigged” because of presumed difficult discriminations would be more habilitative and remunerative if the discriminations were taught. A study is in progress to provide further clarification on this issue.
In the present experiment, the procedures were administered using a one-to-one relationship (teacher-student). The easy-to-hard procedure could also be used effectively in small groups. However, many individuals need the supervision and feedback available only in a one-to-one situation in order to learn this and other tasks. Workshop and work activity center personnel seldom set up systematic one-to-one training programs because of staff shortage and client load, and also because they have been successful doing what training they have done in group settings. As expectancies rise and techniques develop, we will have to move to one-to-one instruction for much of the training to be done. What has been a luxury for reaching simple tasks will be a necessity for training on more complex tasks.
The level of complexity of tasks found in work settings for the moderately and severely retarded must be scrutinized for their non-habilitative nature (Gold, 1972), with the intent of moving to more fulfilling tasks. Studies such as this one should cause us to question all of the assumptions which underlie apparently inappropriate expectancies. Revised expectancies and an emerging technology of instruction should result in a significantly improved outlook for the moderately and severely retarded.
In “The Revolving Door in an MR Hospital: A Study of Returns from Leave” (Mental Retardation, 1973 11 (1), 55) by Keys, Boroskin, and Ross, the total population was 3,334 with 448 in Family Care and 918 on Home Leave; at the end of the study, there were 3,180 total population with 449 in Family Care and 726 on Home Leave.
EDITOR'S NOTE: This 1973 Marc Gold article appeared in volume 11 of Mental Retardation, one of two articles he had published on the parameters of workshop assembly tasks in our journal that year. As I mentioned in my preceding commentary, the inclusion of this work as an illustration of Secretary Swenson's challenge may seem mysterious to most. I ask that you look beyond the dated terminology and emphasis on workshop tasks. Terminology of course must be considered in its historical context and for the sake of historical accuracy the paper is reprinted (as are the other two) as originally published. However, the point of the study—developing fine visual discrimination—and the implications framed in terms of sheltered workshop productivity deserves some additional context. In one of my very first undergraduate classes related to persons with intellectual and developmental disabilities the Marc Gold film “Try another way” was shown (yes, it was actually a film). Illustrated was a package of techniques Gold developed as a special education teacher on the west coast and later empirically studied and refined as a professor at the University of Illinois. There is nothing particularly remarkable in the method as seen through a contemporary lens: gentle guidance, verbal prompts, use of reinforcement. Basic principles of behavior were the point of my class. What I overlooked then and is largely obscured in the article's narrative on experimental protocol, is Gold's fundamental belief that limitations are not inherent in the person but rather a function of our lack of imagination. The study is as much about aspiration as methodology. As Gold noted, question assumptions and revise expectations.
This study was supported in part by NICHHD Program Project Grant No. HD 05951 to Children's Research Center, University of Illinois at Urbana-Champaign. Appreciation is expressed to Progress School, Decatur, Illinois, and Community Workshop, Peoria, Illinois.