VERBAL CONCEPT LEARNING AMONG THE DISADVANTAGED AS A FUNCTION OF STIMULUS PREDIFFERENTIATION Thesis for the Degree of Ph. D. MICHIGAN STATE UNIVERSITY {HAYWORTH L BRADLEY 1970 THESIS LIBRARY Michigan State University This is to certify that the thesis entitled Verbal Concept Learning Among The Disadvantaged As A Function Of Stimulus Predifferentiation presented by Hayworth L. Bradley has been accepted towards fulfillment of the requirements for Ph . D. degree inEducational Psychology Date February 11, 1970 0-169 I ABSTRACT VERBAL CONCEPT LEARNING AMONG THE DISADVANTAGED AS A FUNCTION OF STIMULUS PREDIFFERENTIATION By Hayworth L. Bradley The present study examined the effect of relevant stimulus pretraining on the acquisition of easy and difficult verbal concepts by disadvantaged s5. Eighty Negro 7th graders who were matched on IQ, reading level, chronological age, and socioeconomic status were randomly assigned to one of four pretraining conditions, namely: equivalence labeling, distinctive labeling, observation, and no pretraining. Following pretraining, gs were given a concept learning task in which they were required to indicate what common concepts a set of words referred to. The results indicated a significant main effect due to pretraining. Equivalence labeling was superior to all other conditions in facilitating concept learning. Although significant differences did not occur in comparisons involving the other three pretraining conditions, trends indicate that the observation pretraining was superior to Hayworth L. Bradley distinctive labeling and no pretraining in promoting concept learning. Moreover, trends indicated that no pretraining was superior to distinctive labeling in facilitating concept learning. Other results obtained indicated a multiplicative relationship between pretraining and concept complexity. The source of this interaction is related to the fact that si in the distinctive labeling condition learned more of the easy concepts, but a lesser number of difficult concepts than the no pretraining group. Moreover, the overall results indicated a significant interaction between pretraining, concept complexity, and trials. Significant main effects were also observed for concept complexity, and trials. No significant interactions, however, were observed between pretraining and trials, or between concept complexity and trials. The main conclusion which derives from the present findings is that verbal concept learning among the disad- vantaged can be greatly facilitated by stimulus pretraining. Additional research was suggested to clarify the relationship between pretraining and other learning variables. VERBAL CONCEPT LEARNING AMONG THE DISADVANTAGED AS A FUNCTION OF STIM LUS PREDIFFERENTIATION By K Hayworth Lleradley A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Educational Psychology 1970 6'62 7 6 9/ 7—/~ 7 0 To Jo and Paulette ii ACKNOWLEDGEMENTS Grateful acknowledgement is given to Dr. Clessen J. Martin, Chairman, and Drs. Donald M. Johnson, Joe Byers, and Robert L. Green, Guidance Committee Members, for their invaluable assistance in planning and executing this thesis. Moreover, a special expression of gratitude is given to the principal, faculty, and students of Scotlandville Junior High School for their cooperation and magnanimity in the face of repeated intrusion by the author. For this I shall always be grateful. iii TABLE ACKNOWLEDGEMENTS . . LIST OF TABLES . LIST OF FIGURES . . . LIST OF APPENDICES . Chapter I. PROBLEM . . Theory . . Hypotheses . . Definition of Ter II. REVIEW OF LITERAT III. METHOD IV. RESULTS V. DISCUSSION . . VI. SUMMARY AND CONCL REFERENCES . . . . APPENDICES . . . . . OF CONTENTS ms URE . . USIONS . . iv Page iii vii viii ll 13 1A 31 38 58 70 72 75 Table 2. 10. ll. l2. I3. 14. LIST OF TABLES Subjects' Characteristics Word Groups, Concepts and Dominance Level Mean Chronological Age of Se By Sex and Treatment Group . Means SES By Sex and Treatment Group Summary of the Analysis of Variance On The Mean Number of Concepts Learned . . . . . Means and Standard Deviations of Number of Concepts Learned by Pretraining Group . . Multiple Comparison of Group Means by Newman- Keuls Procedure . . . . . . . . . . Analyses of Variance on Simple Effects for Pretraining . . . . . . . . . Multiple Comparison of Mean Number of Easy Concepts Learned at Each Level of Pretraining Multiple Comparisons of Mean Number of Difficult Concepts Learned at Each Level of Pretraining Analysis of Variance on Simple Effects for Concept Complexity . . . . . . . Means and Standard Deviation of Number of Easy and Difficult Concepts Learned by Pretraining Groups . . . . . . . . . . . . Mean and Standard Deviations of Number of Concepts Learned by Pretraining Groups and Trials . Multiple Comparison of Mean Number of Concepts Learned as a Function of Blocks of Trials 31 33 38 39 Al A2 ’43 45 A6 47 50 51 52 52 15. Means and Standard Deviations of Number of Easy Concepts Learned by Pretraining Groups and Blocks of Trials . . . . . . . . . . 16. Means and Standard Deviations of Number of Difficult Concepts Learned by Pretraining Groups and Blocks of Trials . . . . . l7. Weights for Computing Social Class When Factor is Missing . . . . . vi 57 57 95 LIST OF FIGURES Figure l. 2. Pretraining x Concept Complexity Interaction Number of Easy Concepts Learned by Pretraining Groups . . . . . . . . . . . Number of Difficult Concepts Learned by Pretraining Groups . . . . . . . Number of Concepts Learned by Pretraining Groups and Blocks of Trials . . . . Number of Easy Concepts Learned by Pretraining Groups and Blocks of Trials . . . . Number of Difficult Concepts Learned by Pretraining Groups and Blocks of Trials vii AA A8 49 53 55 56 Appendix A. B. STIMULUS MATERIALS USED IN THIS STUDY LIST OF APPENDICES Instructions--Learning Task Difficult Concepts . Easy Concepts . . . . Instructions--Pretraining Study Study Study Study Study Study Study Study Study List List List List 'List List List List List I (Equivalence) . II (Equivalence) III (Equivalence) I (Observation) II (Observation) III (Observation) I (Distinctive) . II (Distinctive) III (Distinctive) Test List I . . . . Test List II . . WARNER SOCIO-ECONOMIC SCALE viii CHAPTER I PROBLEM N_ee_<_i_ The basic assumption which undergirds most thinking in remedial or compensatory planning is that there is a discoverable relationship between an individual child's learning deficit, and the method of instruction by which he learns most efficiently. While there has been disagree- ment (De Hirsch, Jansky, and Langford 1965; Cohn, 196A; Frostig, 1965; Kirk and McCarthy, 1961; Harris, 1965) as to what are the proper methods and practices to pursue in remediating learning deficits, almost all have implicitly agreed that the "instructional method" is a significant variable. Unfortunately, there is little direct evidence which will permit rigorous conclusions to be drawn regarding the effect of a specific instructional technique on a specific learning deficit. This is a serious psychological and educational omission. Moreover, while most of the theoretical and research efforts regarding instructional methods have been directed toward remediating learning disabilities among mental retardates with basic neurological and physical impairments, the findings have been generalized to other nosological or exceptional groups. This practice has been particularly evident in'programs structured for the "educationally" and "socially" disadvantaged. It should be noted that most of the instructional techniques used in Head Start and Job Corp programs are slight variations of those used in remediating learning difficulties among the mentally retarded. While culturally and socially disadvantaged groups doubtlessly manifest many of the learning deficits common to the mentally retarded, such groups may have deficits which are unique to their social condition, and therefore, may require special instructional techniques. Recent psychological research (Bernstein, 1961; Deutsch, 1963; John, 1963) has suggested that one of the most critical factors which delineates the culturally and socially disadvantaged from other groups is concept deficiency. Further, other research (Zigler and DeLaby, 1962) has demonstrated that this exceptional group shows difficulty in attaining or learning concepts. However, in spite of the commonality of these findings, little research is available which concerns the relationship between specific instructional methods and the concept deficiencies which culturally and socially disadvantaged groups manifest. The need to expand educational and psychological technology and theory with respect to the "socially and educationally deprived" is exacerbated by the fact that his group comprises a significant proportion of the national school population. In 1950, one child out of every ten in the fifteen largest cities in the United States was educa- tionally deprived (Riessman, 1962, p. 1). By 1960, this figure had risen to one in three. From all indications this trend is continuing, and by 1970, it is estimated that there may be one deprived child for every two enrolled in the large city schools (Riessman, 1962). Among the most salient implications which these figures suggest.is the need for efficient methods by which concepts can be effectively taught. Stimulus predifferentiation.has been shown to be an effective technique for improving concept learning among subjects of normal intelligence (Norcross, 1958; Norcross and Spiker, 1957; Schaeffer and Gerjuoy, 1955; Spiker, 1956; Spiker and Norcross, 1962). Moreover, Prehm (1966) has recently shown that verbal labeling is an effective procedure for facilitating the acquisition of geometric concepts among disadvantaged subjects. However, the usefulness of stimulus predifferentiation in facilitating the learning of verbal concepts has not been investigated. The failure to examine the relationship between verbal concept learning and stimulus predifferentiation has precluded the development of a technique which might be useful in ameliorating the conceptual deficits of these children. The growing number of disadvantaged youngsters in the nation's school population, and the urgent need to develop new techniques for teaching these youngsters make the investigation of stimulus pretraining important. Purpose The purpose of this study is to examine the acquisition of verbal concepts among the disadvantaged as a function of different stimulus pretraining procedures. More specifically this study will focus on the acquisition of easy and difficult verbal concepts as a function of three types of stimulus pretraining methods: namely, distinctive labeling, equiv- alence labeling, and observation. Theory Stimulus predifferentiation in general refers to stimulus practice designed to facilitate transfer in a subsequent learning task. Transfer, in consonance with this definition, is predicated on the assumption that the stimuli which subsequently become associated with new responses have been predifferentiated in the pretraining task. Two theories have been suggested to account for predifferentiation.as a factor in facilitating transfer, namely the acquired distinctiveness and equivalence of cues formulation (Dollard and Miller, 1950; Goss, 1955), and the differentiation theory (Gibson and Gibson, 1955). Although these two theories provide different explanations for the influence of predifferentiation, they lead to similar predictions. However, Gross' extension of the Dollard and Miller position seems to provide a better explanation of the influence of stimulus predifferentiation on verbal concept learning than the Gibson and Gibson or Dollard and Miller positions. The differentiation theory of Gibson and Gibson suggests that organisms learn to distinguish various components which are inherent in a stimulus. Prediffer— entiating stimulus materials prior to a learning task directs the organism attention to those components of the stimulus situation which are relevant for the learning task. Stimuli, therefore, become more distinct or equivalent not as a result of some "enrichment" process, but as a result of differentiation. Dollard and Miller (1950) have postulated that "attaching a common label to dissimilar objects gives them a certain learned equivalence increasing the extent to which instrumental and emotional responses will generalize from one to the other . . . conversely, attaching distinctive cue-producing responses to similar objects tends to increase distinctiveness." The assertion here is that attaching labels to stimuli is an enriching, cue-producing process. Goss (1955) extended the Dollard and Miller formulation in two directions by (l) postulating that the acquired equivalence and distinctiveness of cues model is a special case of an interaction between external cues, response- produced cues and instrumental responses; and (2) by applying this model in conjunction with known principles of learning and retention postulates certain relationships between the acquisition of cue-producing responses and the subsequent learning of instrumental responses. He has suggested that with respect to stimulus similarity the more similar the external cues, the smaller will be the contribution of the common mediating responses. Conversely, the more dissimilar the external cues, the more dependent §§ will be on a common mediating response. Although Goss' extension of the Dollard and Miller position was developed to account for discrimination, motor learning, and to some degree generalization, it appears to be relevant for verbal concept learning. However, in order to extend this theoretical formulation to verbal concept learning, several definitions, assumptions, and parallels must be given. The first requirement necessary to extend Goss' theoretical model to verbal concept learning is to postulate a definition of concept learning and more particularly of verbal concept learning. Concept learning in general may be defined as the acquisition of a common response to a number of different objects or stimuli. Diagrammatically, this definition of concept learning per se can be represented in the following manner: ’ l I. I ,’ ,’52 R / ” x,’ C1 S 2. ————— 3 ““Su In traditional concept learning experiments, the stimulus arrays have been geometric designs from which subjects were required to deduce rules for forming concept classes or attributes. However, it is possible to place verbal concept learning in this diagrammatical representation with little or no modification. Given that verbal concept learning is the acquisition of a common verbal response to a number of structurally different, but conceptually related words, a verbal concept may be represented by: ,"Sl (Baseball) ’ I / R(round)- E;-——--"S2 (Orange) ‘ ~ ~"S3 (Knob) Concept learning when defined in the above way is said to be facilitated by mediational processes which are induced in an experimental situation or which are assumed to be products of the §L§ extra-experimental experiences. Thus, the schematic representation of concept learning then becomes: ",, "Sl (Baseball) ’ I R - — _.: ______ (reind)s \ a ‘ S2 (Orange) ‘ \ “s3 (Knob) This mediating process (r - s -) according to some theories is the sine qua non of concept learning and facilitates acquired equivalence. Given a verbal concept learning task of the kind defined above, the structure of English words, and data from word association research, it becomes possible to relate Goss' theoretical formulation to verbal concept learning. Goss has postulated that: (l) the more similar the external cues (stimuli, words, etc.), the smaller will be the contribution of mediating responses, and (2) the more dissimilar the external cues, greater will be the contribution of the mediating responses. In view of the fact that there are very few sets of English words which are structurally similar and which at the same time are conceptually similar, Goss' first postulate would appear to have limited application in verbal concept learning. Thus, it becomes necessary to deal only with the second postulate. The second postulate suggests that when stimulus objects including words are physically or structurally dissimilar, learning and concept formation is dependent -on mediational processes. Consider the following verbal concept: Baseball Fang Word Group = concept "white" Paste Sugar According to Goss' second postulate, the formation of this concept should be heavily dependent on a common mediating process since the stimulus words are structurally different. Assuming that word association norms indirectly reflect the influence of mediation processes, there is some evidence to support this contention. Underwood and Richardson (1956) have demonstrated that verbal concepts are formed more eaSily from words that have high dominance.than from words of low dominance.l The explanation of this difference is based on the assumption that the high dominant words have a common mediation process as a result of their frequent use and association in this culture, while the low dominant words do not since they are infrequently associated. Based on the postulation of Goss, and the evidence provided by Underwood and Richardson, it appears that one of the most important limiting variables in verbal concept learning is the mediating process. Given that mediating processes can be effectively manipulated, verbal concept learning can be facilitated. lDominance refers to the frequency with which a common response is given independently to a set of words in a free association task. 10 The Relevance of Stimulus Predifferentiation for the Theory of Verbal Concept Learning and for Aiding the Disadvantaged Stimulus predifferentiation theoretically would appear to be a useful technique for influencing verbal concept learning. Preliminary activity with stimuli which are to be used in subsequent learning tasks might possibly provide the sufficient condition for the development of what has been called the limiting variable in verbal concept learning, namely mediation processes. Verbal labeling and other kinds of preliminary activity with the stimulus materials should provide not only response-produced cues, but stimulus familiarity as well. That disadvantaged subjects experience difficulty encoding stimulus materials and forming concepts is well documented (Covington, 1962; Deutsch, 1963; Jensen, 1963; John, 1963; Katz and Deutsch, 1963). These difficulties in part have been attributed to the very narrow range of experiences to which these subjects have been exposed. As a result of these narrow experiences, it has been suggested that disadvantaged subjects lack the mediating processes and stimulus familiarity which are required in mastering a structured conceptual task. Given that the absence of useful mediating processes is the chief cause of concept learning difficulty among the disadvantaged, stimulus predifferentiation would appear to offer a unique possibility for assuaging this problem. II Hypotheses Broadly interpreted the theoretical positions of Dollard and Miller, and Goss imply that when stimulus predifferentiation promotes accurate cue utilization relevant mediating processes develop and postive transfer is facilitated. Research with normal and retarded subjects suggests that various types of stimulus predifferentiation differ widely in the degree to which positive transfer is facilitated. In consonance with the theoretical implication and, research findings, the basic hypotheses which will be examined in this study is that: 1. Verbal concept learning among the disadvantaged will vary as a function of the type of stimulus pretraining received. 2. Verbal concept learning among the disadvantages will vary as a function of concept complexity. Goss has suggested that one dimension of the relation- ship between cue-producing, and instrumental responses is the interaction between the complexity of external cues or stimuli and type of stimulus predifferentiation. In consonance, with this assertion, it is hypothesized that: 3. Stimulus predifferentiation and concept complexity will interact. Several ordered hypotheses can be deduced as corollaries from the theory. The theory suggests that a common mediating 12 response is effective in facilitating transfer. Distinctive labeling of stimulus materials should produce the same number of mediating responses as labels used. Thus, the greater the number of mediating responses the greater the amount of 'inhibition and negative transfer. Equivalence labeling, on the other hand, should faCilitate the development of one common mediating response for all stimuli given a common label. Therefore, it is hypothesized that: A. Equivalence labeling will better facilitate the acquisition of verbal concepts among the disad- vantaged than distinctive labeling. Further, based on the theoretical assumption that distinctive labeling produces inhibitory mediators, the following hypotheses will be examined: 5. Observation will better facilitate the acquisition of verbal concepts among the disadvantaged than distinctive labeling. 6. No predifferentiation will better facilitate the acquisition of verbal concepts among the disad— vantaged than distinctive labeling. Moreover, since ong priori grounds it appears that equivalence labeling provides more useful cues than either observation or no pretraining, these hypotheses will also be examined: 7. Equivalence labeling will better facilitate the acquisition of verbal concepts among the disad- vantaged than observation. l3 8. Equivalence labeling will better facilitate the acquisition of verbal concepts among the disad- vantaged than no pretraining. Finally, since observation provides more useful cues than no pretraining, it is hypothesized that: 9. Observation will better facilitate the acquisition of verbal concepts among the disadvantaged than no pretraining. Definition of Terms The following terms were used in this study: Disadvantaged - the term disadvantaged refers to subjects whose families fall in the low and lower-lower classes as measured by the Warner Social Index Scale and whose reading grade placement score is 1 to 3 years behind actual grade placement as measured by the Metropolitan Achievement Test, Form B.1 Distinctive Labeling - attaching a single meaningful verbal label to a single stimulus pattern. Equivalence Labeling - attaching the same meaningful verbal label to several different stimulus patterns. Observation - attending to stimulus patterns without the aid of verbal cues. lDeutsch and his associates (1967) have reported data which suggest that level of reading is a sensitive measure of deprivation. Moreover, most research on the disadvantaged which involves levels of reading suggest that the reading grade placement of these §§ is l to 3 grades below actual grade placement. CHAPTER II REVIEW OF LITERATURE The effects of various types of predifferentiation on transfer are well established. However, the amount of transfer produced seems to very as a function of the type of predifferentiation technique employed. Arnoult (1957) has identified five categories of predifferentiation which have been used by investigators in verbal, discrimination, and concept learning. These categories are relevant stimulus-response, irrelevant stimulus, attention or observation, no pretraining, and relevant stimulus training. Since the focus of this study is on relevant stimulus and attention training only those studies pertinent to these two types of predifferentiation will be reviewed. Review of Relevant Stimulus TrainingrResearch The most important characteristic of this type of predifferentiation is that the stimuli which are used in the pretraining task are identical to those used in the transfer task. However, the responses used in the pre- training task are qualitatively different from those involved in the learning task. Several investigators have 14 15 examined this type of predifferentiation as a factor in facilitating transfer. Robinson (1955) in attempting to test the distinc- tiveness of cues hypotheses examined the effect of verbal labeling on discrimination learning. Three groups of 14 Se were given three different kinds of predifferentiation training. One group, the distinctiveness group, learned distinctive verbal labels to 10 fingerprints; a second group, the equivalence group, learned one verbal label to the first five fingerprints, and a second verbal label to the other five prints; a third group of Se, the Same- Different group, was required to compare the ten prints with other prints, and to indicate whether the ten standard prints were the same or different. During the criterion task, all g; were required to compare a set of prints with those which were used in the pretraining task. Those S5 with preliminary training were told that the prints used in the pretraining task would also be used in the criterion task; however, a fourth group, a no prior training, was not so instructed. The §§ in this group were told that the purpose of the experiment was to determine discrimination ability and that they were to make their judgments as quickly as possible. Two measures, errors and latency, were used as dependent variables. In terms of errors in discrimination, 16 the Same-Different group was superior to the other groups. Moreover, the distinctive labeling and equivalence groups were not significantly different. The no prior training group was significantly inferior to the distinctive, equivalence, and Same-Difference groups. No significant differences, however, were found among the four groups in terms of latency scores. Another attempt to provide a test of the acquired distinctiveness of cues hypothesis through the use of relevant stimulus training has been conducted by Norcross and Spiker (1957). The stimuli were pairs of faces; pair A was female faces and pair B was male faces. Three groups of gs were differentiated with respect to pretraining experience. Subjects in Group R were given only Pair A and learned to say the name "Jean" each time one member was presented and "Peg" each time the other member was presented. A second group, Group I, was given only pair B and learned to say "Jack" each time one member was presented and "Pete" each time the other member was presented. Subjects in Group D were also given pair A. However, they learned to say "same" when the two pictures were identical and to say "different" when both members were presented together. Immediately following pretraining the transfer task was administered. This was identical for all Ss and involved having §§ select the correct faces from among faces which had not been used in the pretraining task. 17 The response measure was the number of correct choices in 30 trials of the transfer task. The difference between the means of Groups R and I was 3.00 with a t-ratio of 1.96 (p= .05). The Group R - Group D difference was 3.98 with the t-ratio equal to 2.36 (p < .02). There was, however, no significant difference between the D and I groups. These results tended to confirm previous findings that verbal labeling facilitates positive transfer. Moreover, Cantor (1955) has examined the effects of three types of pretraining on the discrimination learning in preschool children. The Ss were 60 children from the State University of Iowa Preschool and the Iowa City Parents' Preschool. Ten of the subjects were 3 years old, 38 were A years old, 12 were 5 years old. An upper and lower age group were constituted and §§ within each age level were randomly assigned to three treatment groups. Subjects in the relevant stimulus group received relevant stimulus training in which they learned to associate the names "Jean" and "Peg" with female faces. Subjects in the irrelevant group were given irrelevant pretraining ' involving male faces and the names "Jack" and "Pete." The attention group received no verbal labels, but §§ were required to attend to various parts of the relevant stimuli. All §§ were treated identically in the transfer task. Subjects were required to choose one of two cars which had one of the faces which had been used in pretraining task 18 mounted on the sides. The response measure used was the number of correct choices made in the transfer task. Analyses of the transfer task data revealed that the relevant stimulus group performed significantly better than did the irrelevant and attention groups. No significant difference, however, was observed between the irrelevant and attention groups. Schaeffer and Gerjuoy (1955) conducted three experiments to test the contention that "naming" influences discrimination learning by providing auditory and proprioceptive stimulation which facilitates differentiation of the stimuli. In the first experiment, three groups of children were compared when one was presented with no names for the stimuli, another with similar names, and third with dissimilar names. No significant differences were observed. The second experiment was a repeat of the first with only two groups, a "no names" group, and a "dissimilar names" group. Again no differences were found. The third experiment was a repeat of the second without the instructions. In this experiment, a significant difference between the two groups was found in favor of the one which received the dissimilar names. The authors interpret these data as indicating that names served to facilitate the acquisition of discrimination set without which solution did not occur. When instructions obviated the need for this set to be acquired during the experiment, naming had no effect. l9 Norcross (1958) in a partial replication of the Schaeffer and Gerjuoy study (1955) had gs learn highly similar names for one pair of stimuli (pictures) and dissimilar names for a second pair. The transfer task, the same for all §§ required S to push a different one of four‘buttons to each of separately presented stimulus pictures. Norcross predicted that learning performance, in terms of correct responses, would be superior for the pair of stimuli with dissimilar names as compared with the pair with similar names. The results obtained confirmed this hypothesis. The mean number of correct responses to the dissimilarly labeled stimuli was 20.20, while the mean number of correct responses of similarly named stimuli was 19.90. An examination of verbal responses during the transfer task revealed that 103 incorrect dissimilar verbal responses were given while 3A1 incorrect similar responses were made. The mean difference of 8.27 is significant at .001 level. Norcross performed a second experiment to examine the possibility that the greater number of motor errors to the similarly named pairs occurred because the similar verbal cues were more inconsistently associated with their respective experimental stimuli thanwas true for dissimilar verbal cues. Experiment II differed from the first experiment only in that S was corrected for misnaming in the transfer task, and was required to verbalize the correct name prior to the 20 button pressing response. The results again revealed a difference in favor of the dissimilar named pair. A mean of 22.73 correct responses occurred to the pair with dissimilar names while a mean of 19.62 correct responses occurred to the pair with similar names. Spiker in collaboration with Norcross (1962) has further examined discrimination learning as a function of stimulus predifferentiation in preschool children. The three groups differed in the kind of predifferentiation received in the pretraining task. Groups D1 and D2 learned appropriately to say "same" or "different" when presented with settings of the pietures. Group D1 had been shown the pictures in simultaneous settings while D2 had responded following successive, single presentations of the stimuli. Group N learned discrete names for the pictures. An analysis of correct responses in 30 transfer trials indicated that in the later stages of learning Group N was significantly superior to either Group D1 or D2. A further attempt to investigate the relative effects of certain kinds of verbal labeling has been carried out by Ellis and Muller (196A). Ten §§_were randomly assigned to each of 2A conditions of the experiment. Two levels of stimulus complexity (6 or 2A point random shapes), three kinds of predifferentiation (distinctiveness, observation, or equivalence practice) and four levels of practice (2, A, 8, or 16 trials) were employed. 21 During predifferentiation training, §§ received one of the three types of predifferentiation. Subjects given distinctiveness practice were required to learn relevant meaningful labels to each of eight random shapes. Similarly, §§ given equivalence practice were given an equal number of pretraining trials; the label "wide" was learned for four shapes and the label "narrow" was learned for the remaining four shapes. Subjects given observation practice were given the same number of predifferentiation trials except that they were given no labels to attach to the stimuli and were inttructed only to inspect the shapes and differentiate among them. For the labeling groups each stimulus shape was exposed for a A see. period consisting of a 2 sec. anticipation period and 2 sec. simultaneous presentation with the response. Labels were pronounced by S; correct anticipation and errors were recorded. The observation group observed the shapes for the same time interval as did the labeling groups. Following predifferentiation training, all §§ were immediately given a recognition test which consisted of presenting S with 16 cards, each containing a set of five shapes mounted in a row. Eight of the cards contained a prototype, a shape learned or observed during prediffer- entiation, and the remaining four shapes on each card were variations of the prototype. On the remaining eight 22 cards all five shapes were variations of the prototype. The §§ were instructed to point to a shape if they thought it was one which they learned or observed during predifferentiation training. Five types of responses were recorded: correct selection of a prototype shape, incorrect selection of a variation when a prototype was present; incorrect rejection of shapes when one of them was a prototype; incorrect selection of variation when all shapes were variations, correct rejection of all the shapes when all were variations. The results revealed that with complex shapes, the distinctiveness group made more correct recognitions than the observation groups after 2, A or 8 trials. In short, when labels were attached to stimuli of high complexity, recognition was superior to that provided by observation practice; when labels were attached to stimuli of low complexity, recognition was poorer than that provided by observation practice. The interaction which the finding suggests is supported by a significant complexity X training interaction (F = 3.95). Moreover, practice led to a reduction of incorrect rejection responses (IR) for all types of differentiation training and to a reduction of selection of variations when all shapes were variations (SIV) for the distinctiveness and observation groups. Similarly, practice led to an increase in correct rejection responses (CR) for 23 distinctiveness and observation groups. No systematic changes occurred in SIV or CR responses with increasing amounts of equivalence training. The analysis of variance resulted in a significant variance for practice effects for all responses. Further, the distinctiveness and observation groups tended to make fewer SOV (selection of variations when the prototype was present) responses to either 6 or 2A point shapes. A second experiment was performed to determine if positive transfer to an instrumental motor task would occur following practice under two conditions, distinctive labeling and observation. Fifteen §§ were assigned to each of two conditions of the experiment. One group received distinctiveness training and the other received observation, identical to that given in the first experiment. Following predifferentiation training, §§ were given a criterion task which required them to learn to press a switch for each of the eight shapes they had labeled or observed. Correct responses and latency were the response measures used. The analysis of the data revealed that §§ given distinctiveness practice in labeling 6 point shapes were superior in the acquisition of a motor switching task when compared with §§ given practice in observing the shapes. An analysis of the latency data further confirmed this finding (F = 6.53, p < .05). 2A Ellis, Bessemer, Devine, and Trafton (1962) have (examined the recognition of random tactual shapes as a function of distinctive, equivalence, and observation predifferentiation. The stimuli were eight random shapes selected from six point shapes scaled by Vanderplas and Garvin (1962). Sixteen sets of three shapes were prepared and used for the recognition task. In 8 sets, all three shapes were variations of the prototype; in the remaining 8 sets, two shapes were variations and the third the prototype. The distinctiveness training group learned to label the random shapes; the equivalence training group was given an identical number of predifferentiation trials except that the label "wide" was learned for four shapes and the label "narrow" was learned for the four other shapes. A third group, the observation group, also was given six predifferentiation trials except that SS in this group did not overtly apply verbal labels to the shapes and were instructed to inspect the shapes and differentiate among them. Following predifferentiation, all §§ were immediately tested for recognition. This task consisted of presenting S with 16 sets of shapes and asking him to select shapes which were identical to those he had experienced in the predifferentiation task. The response measures used were selections of prototype, selections of variation of prototype, and 25 incorrect rejections of prototype. Only the group differences in prototype selections were found to be statistically significant (F = 15.55, df = 2/A2, p < .001). Although no multiple comparisons were made, it appears that the equiv- alence group was significantly poorer than the observation and distinctiveness groups on this response measure. However, there was apparently no difference between the distinctiveness and observation groups. The mean number of prototype selections for the equivalence, distinctiveness, and observation groups respectively was 2.93, A.27, and A.AO. There have been several attempts to influence the acquisition of concepts by retarded S§_through the use of stimulus predifferentiation. However, most of these attempts have failed to reveal any significant influence on the performance of these §§ (Miller and Griffith, 1961; Dickerson, Girardeau, and Spardlin, 196A). Nevertheless, Prehm (1966) has demonstrated the effectiveness of verbal pretraining on the acquisition of concepts by disadvantaged §§- Twenty-seven high and twenty-seven low risk disad- vantaged SS, ranging in chronological age from four to seven years, were used in the experiment. Nine §§ within each risk group were randomly assigned to one of three pretraining conditions: a verbal, an attention, and a control condition. Those S§,in the verbal label group were told the names of the stimulus cards and were 26 directed to sort them into two piles, saying the name of each as they placed it on the pile. Those in the attention group sorted the cards into two piles according to the relevant stimulus dimension, but were neither told, nor required to say, the names for the stimuli. §§ in the control group sorted the stimuli into two piles in an unsystematic fashion. All §§ sorted the cards into two piles three separate times, with the examiner shuffling the cards between each sorting. §§ were randomly assigned to receive pretraining on either set one or set two of the stimuli in Transfer Task 1. Immediately following pretraining, §§ were presented with the experimental task. Each one was given a stake of five pennies and told that when he made a correct choice, the examiner would take one of his pennies from him. Pennies were withdrawn from §§ in order to prevent them from obtaining so many pennies that motivation would be reduced. Subsequent to reaching criterion on Transfer Task 1, each S was immediately given Transfer Task 2. Errors and trials to criterion on each transfer task were the dependent variable. However, because of the high correlation (r = +.98) between trials to criterion and errors, error data were not analyzed. The analysis of trials to criterion data indicated that the main effects of pretraining was statistically significant, indicating that performance varied as a 27 function of the pretraining conditions. Secondary analyses of the data revealed that the verbal label group was significantly superior to the attention group or control group in trials to criterion. The attention group was significantly superior to the control group. Review of Attention or Observation TrainingiResearch Observation training involves requiring SS not to make overt responses such as naming or labeling during the pretraining task. Subjects, however, are required to attend to relevant characteristics of the stimuli. Several studies alluded to in the previous section included observation or attention as one of the types of prediffer- entiation. Cantor (1955) has examined the relative effects of attention training, irrelevant and relevant stimulus training on discrimination learning. Attention training consisted of having §§ attend to relevant parts of the stimulus materials (male and female faces). Relevant stimulus consisted of requiring some §§ to label female faces "Jean" and "Peg"; a second group of §§ was given irrelevant stimulus training in which they were required to label a set of male faces "Jack" and "Pete." The analysis of the transfer data revealed that the relevant stimulus training group made significantly more correct choices than either the irrelevant training or 28 attention groups. There was no significant difference between the attention and irrelevant training groups. Ellis, Bessemer, Devine, and Trafton (1962) in a previously cited study examined the relative merits of attention, distinctiveness, and equivalence labeling on tactual shape discrimination. Analyses of the transfer tasks revealed that those §§ receiving equivalence labeling were significantly inferior to those receiving distinctive labeling and attention. There was, however, no significant difference between the distinctiveness and observation groups. Further, Ellis and Muller (196A) have demonstrated that observation training is superior to distinctive labeling in facilitating discrimination of simple random shapes. Distinctive labeling, however, better facilitated discrimination of complex shapes when compared with obser- vation or equivalence labeling groups. There was no significant difference found between the observation and equivalence group at any level of task complexity. Lastly, Prehm (1966) has demonstrated that distinctive labeling was superior to attention training in facilitating the acquisition of simple concepts among disadvantaged SS. The attention training group was also superior to a control group which received no training. 29 Discussion of Relevant Stimulus and Attention Training Research Several conclusions regarding stimulus predifferen- tiation appear possible from the review of the preceding studies. The first conclusion which can be drawn is that when distinctive labeling is compared with observation in carefully controlled experiments, distinctive labeling is superior to observation in facilitating transfer. Moreover, it appears that distinctive labeling is also superior to equivalence labeling in promoting transfer. There appears, however, to be no difference between observation and equivalence labeling in the degree to which transfer is facilitated. I Although these results appear consistently throughout the literature, it should be noted that such findings are based on studies which involved discrimination learning and "normal" subjects. Because of this, caution must be exercised in generalizing these results to disadvantaged populations. A second conclusion which emerges from the review is that for the most part researchers have not examined the prediction made by Goss (1955) concerning an interaction between stimulus predifferentiation and stimulus or task complexity. Only Ellis and Muller (196A) have examined this relationship. Their results confirmed this predicted interaction (F = 3.95 or 2/216,p < .05). Prehm (1966) did not examine this relationship in his research with 30 disadvantaged §§ nor did Miller and Griffith (1961) or Dickerson, Guardeau, and Spradlin (196A) in their studies of predifferentiation. The failure to examine this relationship obfuscates any attempt to fully explore stimulus predifferentiation as an instructional technique. Moreover, most of the studies reviewed here and those which abound in the literature have been concerned with stimulus predifferentiation as a factor in discrim— ination learning. A modicum of research has been done relevant to stimulus predifferentiation as a factor in concept and verbal learning. This is a serious shortcoming from an educational point of view since concept and verbal learning are the kinds of learning most frequently found in schools. CHAPTER III METHOD Subjects The §§ were 80 Negro 7th grade students who attended school in Baton Rouge, Louisiana, whose family socio- economic levels were in the low or lower-low category as measured by the Warner Social Index,1 and whose reading grade placement as measured by the Metropolitan Achievement Test, Form B, was 1 to 3 grades behind actual grade placement. Forty of the §§ were females, and A0 were males. The mean IQ for §§ as measured by the Otis Group Intelligence Test, Form EM, was 89.6, while the mean CA was 12.A. Table l summarizes SSY characteristics. TABLE l.--Subjects' characteristics. Mean Socio— Mean Reading - v Sex Mgin Economic Grade “ism Status Placement Male 12.3 74.25 5.2 89.5 Female 12.5 71.25 5.5 89.7 1 See Appendix B for a description of this scale. 31 32 Stimuli Predifferentiation The stimulus material used in the predifferentiation phase of this study consisted of 6 study lists of 32 word- pairs, 3 study lists of 32 single words and 2 test lists, constructed from material by Underwood and Richardson (1956). Three of the study lists, the Equivalence Lists (EQ) contained 32 stimulus words and 8 response words. The 32 stimulus words were selected from the Underwood and Richardson material in such a way that there were 8 different A word groups from which eight different concepts could be learned. Four of the word groups with an average dominance level of 5A.2 were labelled "easy," while the other four with a dominance level of 12.87 were labelled "difficult." Table 2 contains the word groups, dominance levels, and 8 concepts. The 8 response words for the 32 stimulus words in the EQ Lists were selected in such a way that the response words would suggest the relevant dimension which could be used in deducing the correct concepts from the 8 four word groups. The same response word was paired with each of the four stimulus words which made up a particular word group. For example, the response word "taste" was paired with Lemon, Pickle, Grapefruit, and Vinegar, and the response word "feel" was paired with the stimulus words Eel, Oyster, Lizard, and Seaweed, etc. (See Appendix A). The 3 EQ Lists were the same except that the 32 word pairs were arranged in different random orders such that TABLE 2.--Word groups, 33 concepts and dominance levels. Dominance Word Groups Concepts Levels Badge Buckle p. 67% Diamond Shiny (easy) Aluminum Eel Oyster . A8% Lizard Sllmy (easy) Seaweed Lemon Pickle 57% Grapefruit Sour (easy) Vinegar Sulphur A % Cigar 5 Onion Smelly (easy) Pine Snail A% Cherry 1 Grape Round (difficult) Skull Earthworm 67 Closet , 1 , Freckle Small (difficult) Tack Rigil k 9 5% a esna e . Telephone Noisy (difficult) Zoo ~ Baseball % Fang l2 Paste White (difficult) Sugar 3A no more than two of the four words belonging to the same word group appeared together. The other 3 lists of word pairs, the Distinctive Lists (DI) contained the same 32 stimulus words which comprised the EQ Lists. However, the response words which were selected to maximize response competition differed from those in the EQ Lists in that: (1) each stimulus word had a different response word, (2) the response words were related to the stimulus words but not to the relevant dimensions of the concepts to be learned, and (3) none of the response words was the same as any of those in the EQ Lists. Thus, there were 32 different response words for each of the 32 stimulus words. The same 8 concepts which were common to the EQ Lists were also common to the DI Lists. Each of the 3 DI Lists differed only in the random order in which the word pairs were arranged. The order of arrangement for the DI Lists conformed to the same restriction which governed the EQ Lists. The 3 single word lists, the Observation Lists (OB), again contained the same stimulus words as the EQ and the DI Lists. There were, however, no response words for any of the stimulus words in these lists. The order of arrangement in the 3 lists was random and followed the same restriction which governed the EQ and the DI Lists. 35 The 2 test lists contained random arrangements of the 32 stimulus words which were common to the study list. One test list, TL contained the 32 stimulus words 1) and 32 blank spaces in which could be written the apprOpriate response words. The second test list, TL contained the 23 32 stimulus words plus 3A additional words which were not included in either of the study lists. Concept Learning The stimulus material used in this phase of the study consisted of a 3 x 5 inch cards on which were printed the 8 four word groups. Procedure Prior to predifferentiation the 80.§§ were randomly divided into A groups of 20, and randomly assigned to one of four treatment conditions, namely: equivalence labeling, distinctive labeling, observation, and no pretraining. Ten males and ten females comprised each of the four groups. During the predifferentiation phase of the experiment, those §§ assigned to the Equivalence Labeling Group (ELG) were given one of the EQ List and instructed to study the list for 2 minutes (Appendix A). At the end of the 2 minute period the §§ were given test list TLl and instructed to write the response words which belonged with each of the stimulus words on the list. If the §§ did not correctly supply the correct response for each of the stimulus words, 36 they were given another of the EQ Lists and permitted to study for another 2 minute period. Again, the SS_were given test list TLl and instructed to supply the missing response words. This procedure was continued until the §§ supplied the correct response words for all the stimulus words. All SS were individually administered the predifferentiation tasks. Following predifferentiation, each §§ was individually given the concept learning task. The 8 cards containing the 8 four word groups were presented in the following order: (1) one half of the SS received the "easy" word groups first, and then the "difficult" word groups, (2) the other half of the §§ were given the difficult word groups first and then the easy word groups. At the presentation of each card, the §§ were instructed to read each of the words carefully and to indicate in what way the words were alike. The presentations were further restricted in that all SS had to supply the correct concept for each of the four word groups in a given category of difficulty or exhaust the allotted trials before they proceeded to the next category. That is, if §§ were given the easy word groups first, they had to supply the correct concepts for each of the four easy word groups or exhaust all trials before they proceeded to the difficult word pairs and vice versa. However, within the 2 categories of difficulty, the presentation of the word pairs was random. One half of 37 the §§ in each group received the easy concepts first, and then the difficult ones, while the other half received the difficult-easy sequence. Subjects were given 15 trials to correctly supply the concepts to each of the 8 four word pairs. The procedure for SS assigned to the Distinctive Labeling Group (DI) was the same as that for the Equivalence LabelingGroup. .The procedure for the Observation Group was also the same as that for the Equivalence and Distinctive Labeling Groups except that the Observation Group was given the OB Lists to study and was required to circle those words on TL which were on the OB Lists. 2 The No—Pretraining Group received no predifferentiation, but was given the concept learning task in the same way the other 3 groups received it. Statistical Design The statistical design used in this study was a A x 2 x 3 factorial repeated measures design (Winer, 1962). Two levels of concept complexity, "easy and difficult," 3 blocks of 5 trials, and 3 levels of predifferentiation, "equivalence labeling, distinctive labeling, and observation" plus a control group constituted the experimental arrangements. CHAPTER IV RESULTS Subjects' Characteristics The validity of the results of this study is of necessity predicated on a demonstration of the equalization of the characteristics of §§ within the treatment groups. These preliminary data are presented to accomplish this objective. Table 3 summarizes EEL chronological age (CA) by sex and treatment group. It is apparent from Table 3 that §§ did not differ in CA by sex or treatment group. TABLE 3.--Mean chronological age of EB by sex and treatment group. Male Female Treatment Group Mean SD Mean SD Equivalence Group 12.20 1.60 12.10 1.30 Observation Group 12.AO 1.A8 12.30 l.A5 No Pretraining Group 12.30 l.AA 12.20 l.A5 Distinctive Group 12.30 l.A5 12.50 1.50 Table A summarizes the socio-economic status (SES) of SS by sex and treatment group. Again the results suggest that §§ did not differ in SE8 by sex or treatment group. 38 39 TABLE A.--Mean SES by sex and treatment group. Treatment Group Male Female Mean SD Mean SD Equivalence Group 80.10 10.A1 79.20 12.A0 Observation Group 78.20 11.31 78.A0 11.56 No Pretraining Group 79.00 10.50 78.50 10.80 Distinctive Group 78.30 11.61 78.90 10.15 The mean IQ of §§ in the four groups ranged from 89.1 for the No Pretraining Group to 90.1 for the Equivalence Group. Comparable mean IQ's for the Observation and Distinctive Groups were 89.7 and 90.0. Moreover, the mean reading grade placement for the four groups ranged from A.9 for the No Pretraining Group to 5.A for the Equivalence Group. Mean reading grade placements for the Distinctive and Observation Groups were 5.3 and 5.1. Pretraining Initial analysis of the pretraining data indicated a significant relationship between the number of trials to criterion on pretraining tasks and the number of concepts learned (r = .78, df = 58, p < .01). The mean number of trials to criterion for the Equivalence (EQ), Observation (OB), and the Distinctive (DI) Groups was 6.60, 5.05, and A.20 respectively. The standard deviations for the three A0 groups were 2.A3, 2.0A, and 1.80. Further analyses of the mean difference in trials to criterion revealed that the Equivalence Group required significantly more trials to reach criterion than the Distinctive Group (t = 3.52, df = 38, p < .01). Moreover, this analysis revealed that the Equivalence group also required significantly more trials to reach criterion than the Observation Group (t = 2.18, df = 38, p < .01). There was, however, no significant difference between the Distinctive and Observation Groups. Concept Learning One of the basic hypotheses examined in this study was that verbal concept learning among the disadvantage would vary as a function of stimulus pretraining. An analysis of variance on the mean number of concepts learned was performed to examine this and other hypotheses. Table 5 summarizes the results of this analysis. The significant A main effect suggests that the mean number of concepts learned varied significantly as a function of pretraining (F = 31.93, df = 3/76, p < .01). Table 6 contains the means and standard deviations for the pretraining groups. A1 TABLE 5.--Summary of the analysis of variance on the mean number of concepts learned. Source _ df MS F Between Subjects A (Pretraining) 3 136.69 31.93** Subjects within groups 76 A.28 Within Subjects B (Concept Complexity) 1 50.10 39.AA** AB 3 8.90 7.01** Bx Subjects within groups 76 1.27 C (Trials) 2 12.25 A.26** AC 6 .58 .20 Cx Subjects within ’ groups 152 2.87 BC 2 .19 .29 ABC 6 10.58 16.27** BCx Subjects within groups 152 .65 ** p < .01 TABLE 6.—-Means and standard deviations of number of concepts learned by pretraining group. Pretraining Group Mean SD Equivalence 8.00 0.00 Observation A.75 2.AA No Pretraining 3.50 1.56‘ Distinctive 3.A0 2.13 A2 In View of the significant A main effect, multiple comparisons of the mean number of concepts learned by pretraining condition were performed by means of the Newman-Keuls procedure. These comparisons permitted the examination of the following hypotheses: (l) Equivalence labeling would better facilitate the acquisition of concepts than distinctive labeling. (2) Equivalence labeling would better facilitate the acquisition of concepts than observation. (3) Equivalence labeling would better facilitate the acquisition of concepts than no pretraining. (A) Observation would better facilitate the acquisition of concepts than distinctive labeling. (5) Observation would better facilitate the acquisition of concepts than no pretraining. (6) No pretraining would better facilitate the acquisition of concepts than distinctive labeling. Table 7 summarizes the comparison of group means. TABLE 7.-—Mu1tiple comparison of group means by Newman- Keuls procedure. EQ OB NPT DI Pretraining Group ‘ Ordered Mean Differences Equivalence (EQ) 3.25** A.50** A.60** Observation (OB) _ 1.25 1.35 No Pretraining (NPT) 0.10 **p < .05 43 An inspection of this table reveals that significant differences were found in only the comparisons involving the Equivalence Group. In accord with predictions, the Equivalence Group was superior to the Observation, Distinc— tive, and No Pretraining groups. Contrary to predictions, no significant differences were observed between the Observation, No Pretraining, and Distinctive Groups. How- ever, the results obtained for these three groups were in the predicted direction. A second basic hypothesis examined in this study was that verbal concept learning would vary as a function of concept complexity. The significant B main effect suggests that concept learning varied significantly as a function of concept complexity (F = 39.AA, df = 1/76, p <.01). Subsequent analysis of the mean difference between the number of easy and difficult concepts learned revealed that significantly more to the easy concepts were learned than difficult ones (t = 5.11, df = 79, p < .001). The mean number of easy and difficult concepts learned was 2.80 and 2.11 respectively. Moreover, the overall analysis of variance in accord with predictions revealed a significant pretraining x concept. complexity interaction (F = 7.01, df = 3.76, p <.01). Figure 1 shows the geometric representation of this interaction. Inspection of the interaction profile suggests that the LEARNED NUMBER OF CONCEPTS MEAN AA .. EQ EASY DIFFICULT CONCEPT COMPLEXITY FIGURE I. PRETRAINNG X CONCEPT COMPLEXITY INTERACTION 4" tom-m m1 .1. J “E: A5 slope of the acquisition line for the Equivalence Group is different from the slopes of the other three groups. An additional source of the A x B interaction would appear to be related to the performance of the Distinctive and No Pretraining Groups. The profile suggests that the Distinctive Group learned more of the easy concepts but a lesser number of difficult concepts than the No Pretraining Group. Additional analyses were conducted to further determine the source of the A x B interaction. Tests of simple effects for pretraining were conducted in accordance with procedures suggested by Winer (1962). Table 8 summarizes this analysis. TABLE 8.--Analysis of variance on simple effects for pretraining. Source df MS F A at B1 3 99.76 A7.96* A at B2 3 318.33 153.0A* Within Cell A56 2.08 s p < .01 The results of this analysis suggest that the number of easy concepts learned varied significantly as a function of pretraining (F = A7.96, df = 3/A56, p < .01). The mean number of easy concepts ranged from A.00 for the Equivalence Group to 2.10 for the No Pretraining Group. The mean number of easy concepts learned by the Observation and Distinctive Groups was 2.70 and 2.A0 respectively. This analysis further showed that the number of difficult concepts learned also varied as a function of pretraining (F = 153.0A, df = 3/A56, p < .01). The mean number of difficult concepts learned ranged from A.00 for the Equivalence Group to 1.00 for the Distinctive Group. Comparable means for the Observation and No Pretraining Groups were 2.05 and l.A0 respectively. In View of the significant simple effects for pretraining at both levels of concept complexity, multiple comparisons of the mean number of easy and difficult concepts learned at each level of pretraining were performed by means of the Newman-Keuls procedure. Tables 9 and 10 summarizes the results of this analysis. TABLE 9.--Mu1tip1e comparison of mean number of easy concepts learned at each level of pretraining. EQ OB DI NPT Groups Ordered Mean Differences Equivalence Group (EQ) 1.30* 1.60* 1.85* Observation Group (OB) 0.30 0.65 Distinctive Group (DI) 0.25 * p < .05 A7 Inspection of Table 9 reveals that the Equivalence Group learned significantly more of the easy concepts than the Observation, Distinctive, and No Pretraining (NPT) Groups. No significant differences existed among the other three groups. The results obtained, however, were in the predicted direction. !F} Table 10 contains a similar analysis of the mean ‘ number of difficult concepts learned at each level of ' pretraining. , TABLE 10.--Mu1tip1e comparisons of mean number of difficult b; concepts learned at each level of pretraining. EQ 0B NPT DI Group Ordered Mean Differences Equivalence Group (EQ) 1.95* 2.60* 3.00* Observation Group (OB) 0.65 1.05 No Pretraining Group (NPT) 0.A0 * p < .05 Again, an inspection of this table reveals that the EQ Group learned significantly more of the difficult concepts than the other three groups. There was, however, no significant differences between the OB, NPT, and DI Groups. Figures 2 and 3 show the mean number of easy and difficult concepts learned as a function of pretraining. ~MEAN NUMBER OF CONCEPTS LEARNED A8 / ‘ I . . . J EQ ‘ OB NPT DI PRETRAINING GROUPS "FIGURE 2. NUMBER OFEASY CONCEPTS LEARINED BY PRETRAINING GROUPS R or CONCEPTS LEARN-:20 '- I.- MEAN NUMB- T-Ax— FIGURE 3. L E0 49 ca . NPT PRETRAINING GROUPS NUMBER OF DIFFICULT CONCEPTS LEARNED BY PRETRAINING GROUPS DI 50 A comparison of these figures suggests that the acquisition curves are different. Further analysis involved an examination of simple effects for concept complexity at all levels of prediffer- entiation. Since only two levels of concept complexity were involved, this analysis facilitated the determination of whether differences existed between the mean number of easy and difficult concepts at each level of pretraining. Table 11 summarizes the results of this analysis. TABLE 11.--Analysis of variance on simple effects for concept complexity. Source df MS F B at Al 1 0.00 B at A2 I l A.20 3.07*** B at A3 1 A.90 3.85** B at AA 1 19.20 15.A3* B x §§ with Groups 76 1.27 *p < .01 **p < .05 ***P < .10 This analysis revealed a number of significant results. Significantly, more of the easy concepts were learned than difficult concepts under the Observation (A2), No Pretraining (A3), and Distinctive (AA) conditions. There was, however, no significant difference in the number of easy and difficult concepts learned under the Equivalence (Al) condition. Table 51 12 shows the mean number of easy and difficult concepts learned at each level of pretraining. TABLE 12.--Means and standard deviation of number of easy and difficult concepts learned by pretraining groups. Concepts Pretraining Group Easy Difficult Mean SD Mean SD Equivalence A.00 0.00 A.00 0.00 Observation 2.70 0.92 2.05 0.A0 No Pretraining 2.10 0.82 1.A0 0.28 Distinctive 2.A0 0.50 1.00 0.60 Moreover, the overall analysis revealed a significant main effect due to trials (F = A.26, df = 2/152, p < .01). These results suggest that the mean number of concepts learned varied significantly as a function of trials. Table 13 contains the mean number of concepts learned by blocks of trials and pretraining group. Further analysis of the mean number of concepts learned as a function of trials was conducted by means of the Newman-Keuls procedure. Table 1A contains the results of this analysis. 52 TABLE l3.—-Mean and standard deviations of number of concepts learned by pretraining groups and trials. Blocks of Five Trials Pretraining Group 1 2 3 Mean SD’ Mean SD Mean SD Equivalence 7.34 0.47 8.00 0.00 8.00 0.00 Observation 3.46 2.l7 4.75 2.44 4.75 2.44 No Pretraining 3.10 1.48 3.50 1.56 3.50 1.56 Distinctive 2.22 1.49 3.40 2.13 3.40 2.13 TABLE l4.-—Multiple comparison of mean number of concepts learned as a function of blocks of trials. Blocks of l 2 3 Trials Ordered ' Mean Difference l .99* l.Ol* 2 .02 * p < .05 The results of this analysis suggest that significantly more concepts were learned during the second and third blocks of trials than during the first block. No significant difference, however, occurred between the number of concepts learned on the second and third blocks of trials. Figure 4 shows the mean number of concepts learned by blocks of trials and pretraining groups. "'1‘“ . Vt: -1 I) ‘1 I‘m-pm . NUMBER 0? CONCEPTS LEARNED MEAN 53 E0 Ux Ii I l I I 2 3 BLOCKS OF TRIALS FIGURE 4. NUMBER OF CONCEPTS LEARNED BY PRETRAINING GROUPS AND BLOCKS OF TRIALS 54 The overall analysis of variance permitted an examination of two additional two factor interactions (Pretraining x trials; Concept complexity x trials) and one 3 factor inter- action (pretraining x concept complexity x trials). The results of this analysis revealed that neither the pretraining x concept complexity nor the concept complexity x trials interaction was significant. These results suggest that the relationship between trials and pretraining and trials and concept complexity is additive rather than multiplicative. The analysis, however, revealed a significant pretraining x concept complexity x trials interaction (F = 16.27, df = 6/152, p < .01). Figures 5 and 6 show graphically this interaction. Inspection of Figure 5 suggests that one source of this triple interaction is related to the fact that the Distinctive Group learned a smaller number of easy concepts on the first block of trials than the No Pretraining Group, but a greater number on the second and third blocks of trials. Another source of the interaction is one which is not so apparent from an inspectiOn of Figures 5 and 6 is the fact that §§_in the Equivalence Group learned more difficult than easy concepts on the first block of trials. Tables 15 and 16 contain the means and standard deviations of number of concepts learned by level of concept complexity, pretraining groups, and blocks of trials. CONCEPTS LEARNED MEAN NUMBER OF 55 O 0 -——L———_—l————r 3 I 2 BLOCKS OF TRIALS FIGURE 5. NUWER OF EASY CONCEPTS LEARNED BY PRE- TRAINING GROUPS AND BLOCKS OF TRIALS MEAN NUMBER OF CONCEPTS LEARNED 4- 4 3.. fl 2.. 2f» 4 : 0% I 2 3 BLOCKS OF TRIALS FIGURE 6 NUMBER or DIFFICULT CONCEPTS LEARNED 3v PRETRAINING GROUPS AND BLOCKS or TRIALS 57 TABLE 15.—-Means and standard deviations of number of easy concepts learned by pretraining groups and blocks of trials. Blocks of Trials Pretraining Groups 1 2 Mean SD Mean DS Mean SD Equivalence 3.64 0.51 4.00 0.00 4.00 0.00 Observation 2.10 0.50 2.54 0.81 2.70 0.92 No Pretraining 1.81 0.70 2.10 0.82 2.10 0.82 Distinctive 1.80 0.40 2.40 0.50 2.40 0.50 TABLE 16.—-Means and standard deviations of number of difficult concepts learned by pretraining groups and blocks of trials. Blocks of Trials Pretraining Groups 1 2 Mean SD Mean SD Mean SD Equivalence 3.70 0.52 4.00 0.00 4.00 0.00 Observation 1.34 0.30 2.05 0.35 2.05 0.35 No Pretraining 1.29 0.25 1.40 0.28 1.40 0.28 Distinctive 0.42 0.16 1.00 0.60 1.00 0.60 CHAPTER V DISCUSSION The Relationship between the Findings and Theory Although the results of this study appear to be obscured to a degree by a possible ceiling effect, they suggest that the pretraining conditions used can be ordered in terms of their effectiveness in producing positive transfer. This finding is in agreement with theoretical assertions by Dollard and Miller (1950), Gibson and Gibson (1955), and Goss (1955). In terms of the present study, equivalence pretraining facilitated the greatest amount of transfer while distinctive training produced the least. While significant differences did not occur in comparisons involving the observation, distinctive, and no pretraining conditions, the trends observed were strong and congruent with theoretical predictions. These trends indicate that observation pretraining was superior to distinctive and no pretraining while no pretraining was superior to distinctive training. The difference in the effectiveness with which the pretraining conditions facilitated transfer would appear to be related to the degree to which the pretraining 58 59 conditions provided relevant cues or promoted accurate cue utilization. This assertion would probably be concurred in by Gibson and Gibson (1955), Goss (1955), and Dollard and Miller (1950). The process, however, by which pretraining produces accurate cues or cue utilization has been debated by these theorists. Gibson and Gibson (1955) suggests that pretraining does not provide additional cues, but directs the organism to the relevant aspects of the stimulus complex which are essential for acquiring a given learning task. Thus, the chief function of pretraining according to this position is to facilitate accurate cue or stimulus utilization. Dollard and Miller, and Goss, on the other hand, suggest that the pretraining functions chiefly in producing relevant cues for use in subsequent learning tasks. Stimulus pretraining, then is an enriching cue producing activity. Unfortunately, the results of this study do not permit a substantiation of either of these positions. The resolution of this theoretical issue must await studies or developments which will permit a more adequate quantitative description of the stimuli to be predifferentiated. Such studies would provide detailed knowledge of the functional relationship between the relevant stimulus attributes, and the processes which account for transfer. However, the lack of this detail knowledge should not preclude the use of stimulus predifferentiation as a method for facilitating positive transfer. 60 The significant pretraining x concept complexity interaction observed in this study is consistent with the theoretical prediction of Goss (1955). The interaction observed would appear to be related to the fact that the distinctive pretraining group learned more of the easy concepts than the no pretraining group, but a lesser number of difficult concepts. Although the differences in the number of concepts learned by the Distinctive and No Pretraining Groups at each level of concept complexity were not significant, the trend is suggestive. It appears that while distinctive training does not either facilitate or inhibit the acquisition of easy concepts when compared with No Pretraining, it appears to inhibit the acquisition of difficult concepts. Although an explanation for this finding is obscure, the implication is that distinctive pretraining may under certain conditions have no influence on the acquisition of verbal concept learning tasks while under other conditions it may have an inhibitory effect. Relationship between the Results and Previous Findings The findings relating to the effect of pretraining on the concept learning generally appear to be inconsistent with previous findings. Previous research (Robinson, 1955; Ellis and Muller, 1964; Ellis, Bessemer, Devine, and Trafton, 1962) has generally shown that distinctive labeling or pretraining is superior to observation, 61 equivalence, and no pretraining in facilitating positive transfer.. Moreover, this same research has shown that while observation, and equivalence pretraining provides more transfer than no pretraining, there is generally no difference in the amount of transfer produced when obser- vation is compared with equivalence training. In terms of the present study, equivalence pretraining facilitated the greatest amount of transfer followed by observation, no pretraining and distinctive labeling respectively. Moreover, the trends observed suggest that observation pretraining was superior to both no pretraining and distinctive labeling, while no pretraining was superior to distinctive labeling in facilitating transfer. One possible explanation for these findings is that the effectiveness of Various types of pretraining is determined to a degree by the nature of the subsequent learning taSks. It should be recalled that the studies conducted by Robinson (1955), Ellis and Muller (1964), and Ellis, Bessemer, Devine, and Trafton (1962) examined the effect of pretraining on the acquisition of discrimi- nation tasks while in the present study a verbal concept learning task was employed. The implication here is that differences in effectiveness of pretraining conditions in this and the other studies is related to the fact that discrimination and verbal concept learning are different tasks. 62 That discrimination and concept learning are different tasks has been argued persuasively by Gagné (1965). According to Gagné discrimination learning is essentially a matter of establishing a number of different (response) chains. That is, the organism develops the capacity to make different responses to different members of a particular stimulus collection. Concept learning, on the other hand, involves the development of the capacity to respond in a single way to single collection of stimulus objects. The implication of these two definitions is that for discrimination learning the primary process involved is discrimination, while for concept learning the chief process is generalization. Given that this implication is valid, it becomes possible to account for the present and previous findings. According to Gagné (1965) one of the optimal conditions for facilitating discrimination learning is to lessen the similarity between the stimuli to be discriminated. This may be accomplished (l) by choosing stimuli which have disparate attributes, and (2) by providing some kind of cue, or activity which will maximize the discrimination of the stimuli (e.g., pretraining). In the previously cited .studies some §§ were required to learn a distinctive label for each stimuli presented during pretraining, while others learned the same label for two or more of the stimuli 53 (equivalence labeling). According to Gagné's assertion, those Ss receiving distinctive labeling would be superior on a subsequent discrimination task to those receiving equivalence labeling. The results from the previously cited studies support this prediction. The superiority of Ss receiving distinctive labeling would appear to be related to the fact that distinctive labeling maximized stimuli discrimination, while equivalence labeling inhibits it. The effect of equivalence labeling, it can be said, is to make disparate stimuli functionally or perceptual equivalent. If the effect of equivalence labeling is to make stimuli equivalent, (i.e., to facilitate generalization among stimuli), then it should facilitate concept learning. On the other hand, if distinctive labeling increases the disparity between stimuli (i.e., inhibits generalization) then it should inhibit concept learning. The results of the present study support this assertion. Equivalence labeling was superior to all types of pretraining in facilitating concept learning. Previous research has shown that observation or directed attention is second only to distinctive labeling in facilitating transfer in discrimination learning tasks. The results of the present study indicate that observation was also the second most effective pretraining method. Although the general effectiveness of observation 64 pretraining in promoting transfer in both discrimination and concept learning tasks is the same, two different explanation appear possible. The effectiveness of observation or direct attention pretraining in facilitating transfer in discrimination learning appears related to the fact that this type of training directs Ss' attention to stimulus attributes which can be used in making correct discrimination. However, the effectiveness of observation pretraining as used in this study might be related to the fact it afforded s; an opportunity to engage in a kind of facilitative clustering or organizing similar to the kind described by Bousfield and Cohen (1953). Bousfield and Cohen have found that when gs are required to recall words which are conceptually or semantically related, they tend to place those words which are related in the same cluster. It should be recalled that in the present study, §§ in the observation condition were: (1) required to study randomly arranged lists of the stimulus words, and (2) to select the stimulus word which they had learned from a longer list of words. Since the 32 stimulus words which comprised the study list were related, it is possible that §§ in this condition learned through the process of clustering. Given that §§ in this condition engaged in clustering during pretraining, all that the §§ were required to do during the concept learning phase was to 65 verbalize the basis on which the words had been clustered. That is, since concept formation occurred during pretraining, all that was necessary in the second phase was to verbalize the concept. The fact that the requirements of the con- ceptual task were considerably reduced increased concept learning performance. First - Order Interactions The significant pretraining x concept complexity is consistent with both the theoretical predictions of Goss (1955) and previOus findings reported by Ellis and Muller (1964). The source of the interaction in the Ellis and Muller study obtained in the finding that distinctive labeling produced better recognition of complex shapes, but poorer recognition of simple shapes than observation. The interaction observed in the present study is related to the fact that S3 in the distinctive pretraining condition learned more of the easy concepts, but a lesser number of difficult concepts than S3 in the no pretraining group. Both the previous findings and the present results suggest that the relationship between pretraining and complexity is multiplicative irrespective of whether the stimuli are verbal or non-verbal. The failure to obtain significant interactions between pretraining and trials, or between concept complexity and trials would suggest that these interactions need not be considered in future studies involving these variables. That 66 is, these results suggest that sources of variances due to these interactions may be pooled to yield more powerful tests on main effects and other interactions. The Pretraining x Concept Complexity x Trials Interaction 'Previous researchers have not structured experiments which permit an examination of the relationship between pretraining, complexity of stimuli, and trials. The lack of previous findings hampers any attempt to interpret the meaning of the significant triple interaction observed in this study. Whether this is a meaningful finding or whether it is a statistical artifact must be determined by further research. Significant Main Effects due to Concept Complexity and Trials The only productive point which emerges from the significant main effect due to concept complexity is that it verifies the presently used operational definition of complexity. This finding suggests that the dominance level is a useful and valid method for determining the difficult level of verbal concepts. The significant main effect due to trials confirms previous findings, and simply suggests that concept learning improved as a function of practice. 67 Significance of The Findings as They Relate to The Disadvantaged The overall results of this study indicate that those engaged in teaching the disadvantaged would be well advised to develop syllabi, study guides, or other kinds of preliminary activities which facilitate transfer of training. The design and the use of these preliminary devices should provide these exceptional Ss with the necessary cognitive pegs or mediating devices for acquiring and effectively using concepts. Moreover, such activities could indirectly influence the conative behavior of these §§~ Given that such activities or devices provide success experiences for these gs, motivation and the willingness to engage in learning tasks are likely to be increased. Implications for Further Research That pretraining is an effective procedure for facilitating concept learning among the disadvantaged has been convincingly demonstrated by the results of this study. However, several interesting and valuable research questions remain unanswered. One researchable question which needs to be answered is: To what degree does the kind of pretraining used in this study facilitate the acquisition of other concepts? Detailed research studies designed to attack this question could in addition to answering the basic question provide 68 information relating to the nature of the transfer process and the variables which influence it. Such information would be invaluable in developing educational technology and theory for disadvantaged as well as advantaged pupils. A second researchable question and one which is related to the first is: To what degree does pretraining influence retention among disadvantaged S3? Research designed to answer this question would of necessity deal with the influence of pretraining on both short and long term retention. Given that adequate information regarding the influence of pretraining on retention is acquired, and given that previously mentioned research question is answered, it would be possible to generate a model which would relate in detail the relationship between concept learning, retention, and pretraining. The development of such a model would have both theoretical and practical value. Another research question which should be pursued is: Can the kinds of pretraining used in this study be effectively used with younger disadvantaged S3? If further research demonstrates that pretraining can be used effectively with younger S3, a valuable psychological and eduCational goal can be reached. The effect of being able to use pretraining with younger S3 would be to reduce considerably not only the number of disadvantaged S3 in school populations, but also the number of learning difficulties experienced by this exceptional group. 69 A fourth research question which should be examined is: What influence would the kinds of pretraining employed in this study have on the acquisition of concepts of medium difficulty? If this research question is answered, it would become possible to specify and predict the acqui- sition of all verbal concepts across the complexity dimension. This information would conceivably then permit the most effective kind of pretraining to be selected for teaching Verbal concepts of varying levels of difficult. Limitations of the Present Study One of the possible limitations of the present study would appear to be related to the fact that an insufficient number of concepts was used. AThe effect of the limited number of concepts used appears to have manifested itself in reducing the variance among the groups, thus obscuring or preventing the discovery of more significant differences. However, in spite of this apparent shortcoming, the findings have meaningful implications for those who instruct the disadvantaged. CHAPTER VI SUMMARY AND CONCLUSION The present study examined the effect of relevant stimulus pretraining on the acquisition of easy and difficult verbal concepts by disadvantaged S3. Eighty Negro 7th graders matched on IQ, CA, socio-economic status, and reading grade placement were randomly assigned to one of four pretraining conditions namely: equivalence labeling, distinctive labeling, observation, and no pretraining. Following pretraining, S3 were individually given a concept learning task in which they were required to indicate what common concepts a group of words referred to. The results obtained indicated: (1), that concept learning varied significantly as a function of pretraining. A. Equivalence labeling was superior to all other types of pretraining in facilitating transfer. (2) that concept learning varied signigicantly as a function of concept complexity. A. More of the easy concepts were learned than difficult concepts. 70 71 (3) that a multiplicative relationship exists between pretraining and concept complexity. (4) that a multiplicative relationship exists between pretraining, concept complexity and trials. (5) that an additive relationship exists between pretraining and trials, and between concept complexity and trials. The main conclusion which derives from these findings is that verbal concept among the disadvantaged can be greatly facilitated by stimulus pretraining. Additional research was suggested to clarify the relationship between pretraining and other learning variables. REFERENCES Arnoult, M. O. Stimulus predifferentiation: some general- izations and hypotheses. P3ychological Bulletin, 1957, 54, 339-350. Berstein, B. Social class and linquistic development: a theory of social learning. In A. W. Halsey, J. Floud, and C. A. Anderson (eds.) Education, Economy and Society. Glencoe: The Free Press, pp. 288-314. Bousfield, W. A. and Cohen, B. H. The occurrence of clustering in the recall of randomly arranged words of different frequency of usage. Journal of General Psychology, 52, 83-95. Canter, G. N. Effects of three types of pretraining on discrimination learning in preschool children. Journal of Experimental Psychology, 1955, 49, 339-341. Cohn, R. The neurological study of children with learning disabilities. Exceptional Children, 1964, 31, 179-85. Covington, M. V. Some Effects of Stimulus Familiarization on Discrimination. Unpublished doctoral dissertation, University of California, 1962. DeHirsh, Katrina, Jansky, J., and Langford, W. The Prediction of Reading, Spelling, and Writing Disabilities in Children: A Preliminary Study. Final report to Health Research Council of the City of New York, Columbia University Contract U-l270. New York: Columbia University, 1965. Deutsch, M. The Role of Social Class in Language Development and Cognition. Institute for Developmental Studies, New York, 1964. Deutsch, M. et al. The Disadvantaged Child. New York: Basic Books, 1967. Dickerson, D. J., Guardeau, F. L., and Spradlin, J. E. Verbal pretraining and discrimination learning by retardates. American Journal of Mental Deficiency, 1961, 66, 270-276. 72 73 Dollard, J. and Miller, N. Personality and Psychotherapy. New York: McGraw-Hill, 1950. Ellis, H. C., Bessemer, D. W., Devine, J. V., and Trafton, C. L. Recognition of random tactual shapes following predifferentiation training. Perceptual and Motor Skills, 1962, 10, 99-102. Ellis, H. C. and Muller, D. G. Transfer in perceptual learning following stimulus predifferentiation. Journal of Experimental Psychology, 1964, 68, 388—395. Frosting, Marianne. Teaching Reading to Children With Preceptual Disturbances. Philadelphia: F. A. Davis Company, 1955. Gagne, Robert. Conditions of Learning. New York: Holt, 1955. Gibson, James J. and Gibson, E. J. Perceptual learning: differentiation or enrichment. Psychological Review, 1955, 62, 32—41. Goss, A. A stimulus-response analysis of the interaction of cue—producing and instrumental responses. Psychological Review, 1955, 62, 20-31. Harris, A. J. Individualizing First Grade Reading According to Specific Learning Aptitudes. Office of Research and Evaluation, Division of Teacher Education. New York: City University, 1965. Jensen, A. R. Learning ability in retarded, average, and gifted children. Merrill-Palmer Quarterly, 1963, 9, 123-140. John, V. The intellectual development of slum children: some preliminary findings. American Journal of Orthopsychiatry, 1963, 33, 813-822. Katz, P. A. and Deutsch, M. Relation of auditory-visual shifting to reading achievement. Perceptual Motor Skill, 1963, 17, 327-332. Kirk, S. and McCarthy, J. J. The Illinois test of psycholinquistic abilities--an approach to differential diagnosis. American Journal of Mental Deficiency, 1961, 663 399‘412. Miller, M. B. and Griffith, B. C. The effects of training verbal associates on the performance of a conceptual task. American Journal of Mental Deficiency, 1961, 66, 270-276. 74 Norcross, K. J. The effect on discrimination performance of the similarity of previously acquired stimulus names. Journal of Experimental Psychology, 1958, 563 305-3090 Norcross, K. J. and Spiker, C. C. The effects of type of stimulus pretraining on discrimination performance in preschool children. Child Development, 1957, 28, 79-84. Prehm, H. J. Concept learning among the disadvantaged as a function of Verbal pretraining. Exceptional Children, 1966, 599-604. Riessman, F. The Culturally Deprived Child. New York: Harper and Row, 1962. Robinson, J. S. The effects of learning labels for stimulus on their later discrimination. Journal of Exper- imental Psychology, 49, 112-115. Shaeffer, M. S., and Gerjuoy, I. R. The effect of stimulus naming on the discrimination learning of kindergarten children. Child Development, 1955, 26, 233—240. Spiker, C. C., and Norcross, K. J. Effects of previously acquired stimulus names and discrimination performance. Child Development, 1962, 33, 859-864. Underwood, B. J., and Richardson, J. Some materials for the study of concept formation. Psychological Bulletin, 1955, 53, 8“-95. Vanderplas, J. M., and Garvin, E. A. Complexity, association value, and practice as factors in shape recognition following paired-associate training. Journal of Experimental Psychology, 1959, 57, 155—163. Warner, W. Lloyd, Meeker, M., and Ellis, K. Social Class in America. New York: Harper and Row, 1960. Winer, B. J. Statistical Principles in Experimental Design. New York: McGraw-Hill, 1962. APPENDIX A 75 LEARNING TASK INSTRUCTIONS (For All Groups) I am going to show you some cards with four words on them. Read each word carefully and tell me in what way the words are alike. For example: If I ask in what way are Basketball, Tennis, Track, and Football alike, what would you say? If S does not get this example correct: Tell him that all of them are sports. Now in what way are Thomas Jefferson, Abe Lincoln, George Washington, and Richard Nixon alike? If S says they are all men, say yes, that is right, but can you think of any other way in which they are alike. If S cannot answer: Tell him they are all Presidents of the United States. Ask: Are there any questions? If there are no questions: Show first card and ask: In what way are: , , and alike? 77 DIPFICULT CONCEPTS DISTINCTIVE LABELS EQUIVALENCE LABELS SNAIL SLOW SNAIL CHERRY RED CHERRY . GRAPE PURPLE GRAPE SHAPE SKULL B0NY SKULL EARTHWORM CRAWLY EARTHWORM CLOSET DARK CLOSET SIZE FRECKLE FACED FRECKLE TACK SHARP TACK CITY STREET CITY RATTLESNAKE POISON RATTLESNAKE__ SOUVD TELEPHONE CALL TELEPHONE * zoo KEEPER zoo BASEBALL HARD BASEBALL FANG LONG FANG PASTE STICKY PASTE COLOR SUGAR SWEET SUGAR 78 EASY CONCEPTS DISTINCTIVE LABELS BADGE PATROL BUCKLE BELT DIAMOND RING ALUMINUM LIGHT EEL ELECTRIC OYSTER STEW LIZARD SCALY SEAWEED STRINGY SULPHUR YELLOW CIGAR ASHES ONION PEELS PINE TAR LEMON FRUIT PICKLE GREEN GRAPEFRUIT JUICE VINEGAR CARMEL EQUIVALENCE LABELS BADGE BUCKLE DIAMOND ALUMINUM APPEARANCE EEL OYSTER LIZARD SEAWEED SULPHUR CIGAR ONION PINE FUMES LEMON PICKLE GRAPEPRUIT__ VINEGAR TASTE 79 INSTRUCTIONS EQUIVALENCE AND DISTINCTIVE GROUPS PRETRAINING You will be asked to learn certain word pairs from a list I will give you. After you have studied the list you will be given a list which contains one word from each of the word pairs you have learned. You will then be asked to write the words which go with each of the words on this list. Ask: Are there any questions? If S does not understand, demonstrate. EQUIVALENCE GROUP oyster - feel seaweed - feel freckle - size cigar - fumes paste — color eel - feel pine - fumes aluminum - appearance sulphur - fumes cherry - shape rattlesnake — sound city - sound lizard - feel vinegar - taste buckle - appearance badge — appearance 8O STUDY LIST I baseball - color snail - shape skull - shape closet - size lemon - taste diamond - appearance grape - shape earthworm — size fang - color sugar — color grapefruit - taste tack - size pickle - taste onion - fumes zoo - sound telephone - sound 81 STUDY LIST II EQUIVALENCE GROUP fang - color pine — fumes oyster - feel lemon — taste grape - shape freckle - size eel - feel pickle - taste tack - size buckle — appearance aluminum - appearance lizard - feel snail - Shape seaweed — feel cherry - shape sulphur - fumes earthworm - size rattlesnake — sound cigar - fumes telephone - sound diamond - appearance city - sound closet - size' sugar - color badge — appearance baseball - color zoo - sound onion — fumes grapefruit - taste vinegar - taste paste - color vinegar - taste onion - fumes zoo — sound grapefruit - taste grape — shape taste - color freckle - size closet - Size tack - size badge - appearance cigar - fumes seaweed - feel buckle - appearance eel - feel baseball - color pickle - taste 82 TUDY LIST III EQUIVALENCE GROUP aluminum - appearance lizard - feel pine - fumes oyster - feel lemon — taste fang — color sulphur - fumes cherry - shape rattlesnake - sound skull - shape sugar - color telephone - sound snail - shape earthworm - size city - sound diamond - appearance grape lizard baseball rattlesnake_ pickle aluminum sugar skull telephone earthworm vinegar eel zoo fang closet lemon 83 STUDY LIST I OBSERVATION GROUP tack onion oyster buckle sulphur snail pine city freckle paste grapefruit cigar diamond cherry badge seaweed sugar skull aluminum badge rattlesnake pickle sulphur zoo seaweed eel lizard baseball closet city earthworm oyster 84 STUDY LIST II OBSERVATION GROUP telephone freckle grapefruit cigar snail diamond pine buckle fang grape onion tack cherry vinegar lemon paste grape telephone onion cherry freckle diamond buckle lemon snail earthworm baseball cigar lizard fang badge eel 85 STUDY LIST III OBSERVATION GROUP city pickle aluminum zoo oyster sulphur seaweed Skull vinegar sugar grapefruit paste closet tack rattlesnake pine 86 STUDY LIST I DISTINCTIVE LABELING GROUP eel - electric rattlesnake - poison oyster - stew city - street sulphur — yellow grapefruit - yellow paste - sticky zoo - keeper lizard - scaly tack - sharp snail - slow fang - snake closet - dark 3 aluminum — light lemon - fruit telephone - call badge - patrol skull — bony cherry - red grape - purple baseball - hard sugar — sweet seaweed - stringy onion - peels cigar - ashes ' buckle - belt pickle — green freckle - faced diamond - ring earthworm - crawl pine - tar vinegar - carmel 87 STUDY LIST II DISTINCTIVE LABELING GROUP grapefruit - yellow diamond - ring grape — purple pickle - green aluminum - light snail - slow rattlesnake — poison eel - electric earthworm - crawl pine - tar Skull - bony sulphur - yellow onion - peels oyster - stew fang - snake baseball - hard vinegar — carmel lizard - scaly freckle - faced city - street telephone - call lemon - fruit cigar - ashes paste - sticky sugar - sweet 3 - cherry - red zoo - keeper badge - patrol tack - sharp seaweed — stringy closet - dark buckle - belt 88 STUDY LIST III DISTINCTIVE LABELING GROUP badge - patrol freckle - faced paste - sticky tack - sharp oyster - stew sugar - sweet pine - tar vinegar - carmel city - street skull - bony seaweed - strings telephone - call sulphur — yellow earthworm — crawl snail - slow grape - purple cherry - red . fang - snake closet - dark cigar — ashes buckle - belt onion - peels zoo - keeper diamond - ring lemon - fruit aluminum - light lizard - scaly baseball - hard pickle - green rattlesnake - poison eel - electric grapefruit - yellow NAME: 89 SEX: DIRECTIONS: DISTINCTIVE LABELING I3 [11 ST LIS a I PREDIFFERENTIATION In the blank spaces write the word which belongs with each of the given words. EQUIVALENCE LABELING GRAPE LIZARD BASEBALL RATTLESNAKE PICKLE ALUMINUM SUGAR SKULL TELEPHONE EARTHWORM VINEGAR EEL ZOO FANG CLOSET LEMON TACK ONION OYSTER BUCKLE SULPHUR SNAIL PINE CITY FRECKLE PASTE GRAPEFRUIT CIGAR DIAMOND CHERRY BADGE SEAWEED NAME: 90 DIRECTIONS: FANG SKULL TOWN TIN GRAPEFRUIT FIR BADGE FROG GARLIC RADIO NAIL SULPHUR ORANGE MEDAL PEACH CIGARETTE EARTHWORM RUBY SEAWEED TELEPHONE PASTE CLIP PICKLE FLATWORM CITY SNAIL EEL GRAPE CLOSET LEMON CIGAR ZOO TEST LIST 2 P W EDIFFERENTIATION Circle the words which were among those which appeared in the lists of words you have seen. FISH CRAB FOOTBALL CIDER MUSSEL CHERRY RATTLESNAKE DIAMOND CRANIUM FRECKLE PIMPLE PINE VINEGAR ALUMINUM LIZARD FIN ALUM LIME SUGAR PRUNE CABINET OY TEI COTTONMOUTH GLUE PLANKTON CRUCIBLE BASEBALL TACK ONION BUCKLE SALT CUCUMBER APPENDIX B WARNER SOCIO-ECONOMIC SCALE 91 92 NATURE OF TIE SCALE Warner, Meeker, and Eells (1949) developed a social scale which predicts social class status accurately in more than 90 in 100 cases. The formula includes four major factors: occupation, source of income, house type, and dwelling area. Detailed descriptions provide a basis for evaluating and rating the four factors on seven point scales. The following are the four factors and the various points which are assigned to each of the categories under a given factor. FACTOR I - OCCUPATION SCORE 1 Professional and Proprietors of large businesses 2 Semi-professional and smaller officials of large business 3 Clerks and kindred workers 4 Skilled workers 5 Proprietors of very small businesses 6 Semi—skilled workers 7 Unskilled workers FACTOR II SOURCE OF INCOME SCORE l Inherited wealth 2 Earned wealth 3 Profits and fees 93 4 Salary 5 Wages 6 Private relief 7 Public relief FACTOR III - HOUSE TYPE SCORE 33 1 Excellent 53% 2 Very good houses I Good houses Lid Average houses a 5 Fair houses 6 Poor houses 7 Very poor houses FACTOR IV - NEIGHBORHOOD SCORE 1 Most exclusive section of town 2 Area well above average 3 Area "nice and respectable" 4 "Average" neighborhood populated mainly by working men 5 Area close to industry or railroad 6 Edge of slum 7 Strictly slum The Warner et al. formula may be written as: 4 x Occupation = 94 3 x Source of Income 3 x House Type Score 2 x Neighborhood Score = Total score = Social Class Score The total class status score is translated in social class equivalence by the following table: TOTAL SCORE , SOCIAL CLASS EQUIVALENT l2 - 22 Upper class 25 - 34 Upper - middle class 37 - 50 Lower - middle class 54 - 63 Upper lower class 67 - 84 ‘Lower - lower class The reported multiple correlation of the four factors - occupation, source of income, house type, and dwelling area — with Evaluated Participation1 is .972. OTHER PROVISIONS OF THE SCALE The authors have provided data which permit the scale to be used when one of the four factors is missing. The following table shows the weights which must be multiplied by the ratings on each of the three factors when one factor is missing. lEvaluated participation is another technique for evaluating social class perfected by these authors. This technique involves ratings on matched agreement, ratings on symbolic placement, ratings on status reputation, ratings by comparison, ratings by simple assignment to a class and ratings by institutional membership. 95 TABLE l7.--Weights for computing social class when factor is missing. Weight to be Used When One Factor is Missing FACTOR Occupation Source of House Dwelling Missing Income Type Area ‘0 Missing Missing Missing Occupation 5 5 5 Source of Income 5 4 4 House Type 4 4 3 Dwelling Area 3 3 3 FORM OF THE SCALE USED IN THIS STUDY An abbreviated form of the Warner Scale was used in this study. Only three factors—-occupation, house type, and neighborhood scores were used. The reported correlation of this trial with Evaluated Participation is .964. IIIIIIIIIIII 93 II[IIIIIIIIIIIIIIII 3196 3