University of Southern California Dissertations and Theses

The Role Of Intellectual Abilities In Concept Learning

(USC Thesis Other)

The Role Of Intellectual Abilities In Concept Learning

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content This dissertation has been microfilmed exactly as received 67-2106 DUNHAM, Jack Lewis, 1938- THE ROLE OF INTELLECTUAL ABILITIES IN CONCEPT LEARNING. U niversity of Southern California, Ph.D., 1966 Psychology, general University Microfilms, Inc., Ann Arbor, Michigan THE ROLE OF INTELLECTUAL ABILITIES IN CONCEPT LEARNING by Jack Lewis Dunham A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (Psychology) September 1966 UNIVERSITY O F S O U T H E R N C ALIFO R NIA T H E G R A D U A T E S C H O O L U N IV E R S IT Y PA R K L O S A N G E L E S , C A L IF O R N IA 9 0 0 0 7 This dissertation, written by ......... .J.a.c.L.Lew.is...Dunham......... under the direction of h..ls..Dissertation Com mittee, and approved by all its members, has been presented to and accepted by the Graduate School, in partial fulfillment of requirements for the degree of D O C T O R OF P H I L O S O P H Y ..... Dean Date September 3 . * . . . 19.66 DISSERTATION COMMITTEE ACKNOWLEDGMENTS The author would like to express his appreciation and gratitude to the members of his dissertation committee, professors J. P. Guilford, Norman Cliff, and Richard Wolf, for their inspiration and guidance. This study was made possible by the financial sup port of the Office of Naval Research Contract Nonr-228(20) and by the staff of the Aptitudes Research Project, Univer sity of Southern California, who provided the necessary facilities and assistance. The author is grateful to Mr. George Prince, Assistant Principal, Mr. Herbert Abrams, Counselor, and Mrs. Marie Sander, Coordinator of Guidance, of Mayfair High School, Lakewood, California, for providing a sample. Thanks are also due Dr. Ralph Hoepfner, Los Angeles City College, and the Psychology Department, University of Southern California, for providing subjects for the pilot studies. Computing assistance was obtained from the Health Sciences Computing Facility, University of California, Los Angeles, sponsored by National Institute of Health Grant FR-3, the Western Data Processing Center, University ii of California, Los Angeles, and the Computer Sciences Laboratory, University of Southern California. TABLE OF CONTENTS Page ACKNOWLEDGMENTS ............................ ii LIST OF TABLES............................ v LIST OF FIGURES............................ vi Chapter I. THE PROBLEM........................ 1 II. THE HYPOTHESES.................... 21 III. PROCEDURES........................ 28 IV. ANALYSIS AND RESULTS............. 32 V. DISCUSSION........................ 70 VI. SUMMARY AND CONCLUSIONS........... 81 REFERENCES................................. 84 APPENDIX A ................................. 93 APPENDIX B ................................. 97 APPENDIX C 106 LIST OF TABLES Table Page 1. Means, Standard Deviations, and Reliabilities of Test Scores............. 33 2. Correlation Matrix of the Tests.............. 35 3. Unrotated Factor Matrix ...................... 36 4. Rotated Factor Matrix ......................... 38 5. Correlation Matrix of the Trials of the Symbolic-Concept-Learning Task with the Ability T e s t s ................... 49 6. Correlation Matrix of the Trials of the Figural-Concept-Learning Task with the Ability Tests ................. 50 7. Correlation Matrix of the Trials of the Semantic-Concept-Learning Task with the Ability Test .................... 51 8. Extended Matrix: Factor Loadings of the Trials on the Ability Factors......... 52 9. Means and Standard Deviations of the Concept-Verbalization Scores ............. 67 10. Correlation Matrix of the Concept- Verbalization Scores and Ability Tests....................................... 68 11. Extended Matrix: Loadings of Concept- Realization Responses on the Ability Factors..................................... 69 v 13 47 53 54 55 56 57 58 59 60 61 LIST OF FIGURES Theoretical Model for the Complete "Structure of Intellect" . . . . Mean Number of Correct Responses on 12 Trials for the Symbolic-, Figural-, and Semantic-Concept Learning Tasks ............... . Factor Loadings of the Trials for the Symbolic-Concept Task on the Divergent Production Factors . . Factor Loadings of the Trials for the Symbolic-Concept Task on the Convergent-Production Factors . . Factor Loadings of the Trials for the Symbolic-Concept Task on the Cognition Factors ............... Factor Loadings of the Trials for the Symbolic-Concept Task on the Memory Factors ................. Factor Loadings of the Trials for the Figural-Concept Task on the Divergent-Production Factors . . Factor Loadings of the Trials for the Figural-Concept Task on the Convergent-Production Factors . . Factor Loadings of the Trials for the Figural-Concept Task on the Cognition Factors ............... Factor Loadings of the Trials for the Figural-Concept Task on the Memory Factors ................. Factor Loadings of the Trials for the Semantic-Concept Task on the Divergent-Production Factors . . vi 12. Factor Loadings of the Trials for the Semantic-Concept Task on the Convergent-Production Factors . 62 13. Factor Loadings of the Trials for the Semantic-Concept Task on the Cognition Factors ............................ 63 14. Factor Loadings of the Trials for the Semantic-Concept Task on the Memory Factors .............................. 64 CHAPTER I THE PROBLEM The Relationship of Intellectual Abilities to Concept Learning T. S. Kendler (1961) divided the theories of con cept learning into two categories. One type of theory is concerned primarily with the stimulus and its response, and treats concept learning as an extension of discrimi nation learning. The other focusses on the processes intervening between the stimulus and the response. The present author believes the latter type to be more promising. In the complex learning situations such as a concept-learning task the subject does not sit pas sively, learning only at the whim of the experimenter's manipulation of frequency of reinforcement, interval of reinforcement, number of relevant dimensions, etc. Grant ed that these are important variables affecting the learn ing process, they do not tell the whole story. An S recognizes attributes common to the stimuli, he produces and tests hypotheses concerning which attributes are relevant, he remembers what occurred on previous trials, etc. In the investigation of these processes that S 1 2 performs between the receiving of a stimulus and the pro duction of the response is the promise for the understand ing of concept learning. The theorist who ignores process generally ends up with highly arbitrary constructs, such as the proportion of relevant to irrelevant cues as in the Bourne-Restle Model (Bourne § Restle, 1959). On the other hand, the constructs of the theorist concerned with process seem to be of two types: internalized S-R bonds which mediate between the stimulus and the response (e.g., H. H. Kendler 6 D*Amato, 1955) and strategies employed by S (e.g., Bruner, et al., 1956). What the theorists have ignored is that the processes which appear to be relevant to learn a concept can, in part, be identified with the intellectual abili ties measured in the psychometric laboratory for decades. The value of the use of abilities with regard to learning theory has been recognized by some theorists. Ferguson (1954, 1956) assumes that abilities "... are attributes of behavior, which through learning have at tained a crude stability or invariance in the adult." He asserts that, by means of transfer, abilities exert their effects differentially in learning situations. Abilities that transfer and produce an effect at one stage of learn ing may differ from those at another. Guilford (1967) has argued for the use of abili ties as constructs in learning theories. He has outlined 3 possible relations between certain abilities and various kinds of learning. Hunt (1962, pp. 164-167) realized the importance of measures of individual differences as an avenue to the strategies involved in concept learning. However, he con cluded that there was a paucity of evidence relating the two domains. The lack of relevant data is not only prominent with regard to roles of abilities in concept learning, but with regard to abilities and learning in general. Possibly this is due to the limited success of the early attempts to establish relationships between abilities and learning. Correlations of gains or other learning measures with mea sures of abilities were generally found to be low, leading many investigators to conclude that learning and ability are unrelated and that each learning task requires its own specific learning ability (Duncanson, 1964). Factor-Analytic Investigations of Abilities and Learning Working in the area of motor learning, Fleishman and Hempel (1954) made an important methodological ad vance. They factor analyzed a number of ability tests along with scores made at various stages of practice on a complex psychomotor task. The factor loadings indicated that abilities did contribute to the learning performance, thus establishing a relationship between abilities and 4 learning. In particular they have shown that performances on the Complex Coordinator (Fleishman $ Hempel, 1954), discrimination reaction time (Fleishman 4 Hempel, 1955), and rotary pursuit (Fleishman, 1960) are related to abili ties, e.g., rate of movement and psychomotor coordination, which were previously identified by traditional psycho metric procedures. The same approach has since been applied to cogni tive tasks. Stake (1961) investigated the relationship of rote- and relational-learning tasks with scholastic achievement and various mental abilities. Learning curves were fitted to the learning scores and parameters were determined for each person for each task. Values on these parameters were then factor analyzed together with the ability and achievement scores. Twelve factors were iden tified, eight of which were common to the learning tasks and the reference variables, and four of which were speci fic to the learning tasks. The results indicate that learning is related to the ability and achievement mea sures, but there are also contributing determiners inde pendent of these measures. There was not a general learn ing factor. Allison (1960) administered 13 learning tasks hypothesized to represent three learning processes--rote, conceptual, and motor learning. He derived parameters from fitted learning curves and investigated their relationships with ability and achievement measures. He interpreted seven learning factors and five ability fac tors, four of which were common to both domains. The three learning factors that were independent of the abili ty measures were interpreted as rote learning, spatial rote learning, and early vs. late learning. Again there is evidence that learning is partially related to ability measures, and that there are learning factors which are independent of the ability measures, but which are common to more than one learning task. Games (1962) investigated the learning of a number of verbal tasks, both paired-associates and serial lists. The task varied in method of presentation (anticipation or recall), type of response (oral or written), and order of presentation. He factor analyzed the scores on six refer ence tests designed to measure rote and span memory and then incorporated the trial scores by a projection proce dure. He found that learning performance was related to the measured memory abilities, but that the evidence about method, response, and order was inconclusive. Duncanson (1964) administered a battery of ability tests in conjunction with three types of learning tasks, concept formation, paired associates, and rote learning. Each type included one task with verbal, one with numeri cal, and one with figural material. He employed a proce dure developed by Tucker (1958, 1960), which permits the 6 decomposition of the learning records of a number of peo ple on a single learning task into component curves and the determination of the contribution of each curve to each person's learning performance. Each learning task was subjected to a separate factor analysis in order to determine the number of factors necessary to describe the learning performances of the subjects on that task. Fac tor scores were then calculated for the subjects, and the factor scores of all the subjects on all the tasks were entered in a factor analysis together with the scores on the ability measures. Seven factors were extracted and rotated to an equamax solution. One factor was found to be restricted to the ability measures. It was interpreted as a speed factor. Three factors were common to the abil ity and learning measures. They were interpreted as ver bal ability, rote-memory ability, and reasoning ability. Three factors were specific to the learning measures. They were interpreted as verbal learning, nonverbal learn ing, and concept formation. An explanation for the success of the studies employing factor analysis to relate abilities and learning performance is offered by Duncanson (1964). The investi gations prior to the use of factor analysis possibly sought too high a magnitude of correlation between mea sures of learning and measures of abilities. The relia bilities as well as the common variance of the tests and 7 tasks determine their correlation. Tilton (1949) has crit icized many of the early studies on the basis of low relia bilities for the learning measures used. Factor analysis is a particularly valuable tool in this respect in that it permits the discovery of common variance even when relia bilities are low. Although it is now apparent that mental abilities do contribute to performance in some learning situations and that some learning tasks are related to some others, the nature of the interrelationships is not clear. It is also possible that the factors which have been found speci fic to learning tasks simply represent abilities which were not differentially sampled by the tests included among the reference variables. Factor-Analytic Investigations of Abilities and Concept Learning If we consider the relationship of abilities to concept learning, then three of the above studies are pertinent. One of Stake's (1961) tasks required the sort ing of cards into four categories, white things, household things, common edibles, and living things. It had no ap preciable loadings on any factor and did not enter into the interpretation of any factor. In addition, it had the loweat communality of the 12 learning tasks he used. This task is clearly different from tests of the abilities mea sured and from the other learning tasks. 8 Duncanson (1964) used three concept-formation tasks which were essentially the same except for the type of stimulus material employed, verbal, numerical, and figural. Each task consisted of a series of stimulus dis plays, and each display was or was not an instance of the concept. Each display varied along four dimensions, with three values possible along each dimension. A subject was to decide whether a display was or was not an instance of the concept, and then was informed as to the correctness of his choice. All three concept-formation tasks loaded on one factor and no others, and no variables loaded on that factor. Allison (1960) included four concept-formation tasks in his learning battery. All involved assigning one of four letters to sets of four words or four drawings. Two of the tasks used verbal materials, while the other two employed figural material. The subjects were shown a set, assigned one of the four letters, and then were in formed which letter was correct. In two of the tasks, one of each type of stimulus material, the same 16 sets, four for each concept, were repeated; while in the other two tasks the same set was never repeated. The concept learn ing tasks loaded predominantly on one factor which was interpreted as conceptual learning. Through an inter battery factor analysis it was found that this conceptual learning factor was related to most of the reference 9 factors. Allison's (1960) finding is obviously at variance with the results of Duncanson (1964) and Stake (1961), who found concept-learning tasks unrelated to ability vari ables. However, the tasks were quite different in the three studies, Allison's being the most complex. Duncanson (1964, p. 41) points to this complexity as being responsible for the relationship with the ability mea sures. He concludes that "... the variables affecting performance on concept-formation tasks have not yet been identified." Recently two investigators (Bunderson, 1965; Manley, 1965) have concentrated solely on the relationship between concept learning and abilities. Bunderson (1965) administered 30 reference tests and 26 concept~attainment problems to 145 university undergraduates. His concept task involved eight stimuli that contained eight geometric figures, a diamond, a circle, a square, a spade, a cross, a club, a heart, and a triangle; an example of a concept might be "black-triangle." Combinations of positive and negative instances were used to build into the tasks spec ified kinds of cognitive activity. The S was shown a stimulus and was asked to determine whether it was a posi tive or negative instance of the concept; wrong responses were recorded by S and later scored by grouping into blocks of trials defined as stage 1, 2, 3, etc. The 10 reference test data were factor analyzed by the principal- components method and then rotated to an equamax solution. Ten mental ability factors were defined: three reasoning, two flexibility, three memory, and two visual-speed fac tors. A factor extension procedure was used to determine the relationship between performance at different stages of practice on the concept tasks and the reference factors. Bunderson postulated three high order processes: problem analysis, search process, and organization. The reference abilities transferred differentially at different stages of practice and supported his postulates of higher order processes. Bunderson found relationships between concept tasks and three kinds of reasoning as well as perceptual abilities. Manley (1965) administered a battery of mental- ability tests, representing seven previously identified factors, to 119 male ninth-grade students. The seven fac tors or ability measures represented were flexibility of closure, induction, associative memory, number facility, general reasoning, syllogistic reasoning, and verbal com prehension. After the reference battery was given, sub jects were tested on 12 concepts that represented three types of tasks: nonverbal concepts restricted by the attributes of the stimuli, nonverbal concepts not restric ted by the attributes of the stimuli, and verbal concepts. The ability measures and the concept-attainment 11 error scores were factor analyzed. Five factors were clearly defined by the reference tests and were inter preted as numerical ability, reasoning, verbal ability, memory, and selective attention. Two factors were not clearly defined by either the reference tests or the concept-attainment tasks. The remaining factors were de fined by the concept-attainment tasks, although only one type of concept-attainment task characterized each factor. The above studies offer contradictory evidence. But many of them do confirm a relationship between abili ties and various kinds of learning. Still, we seem a long way from indicating with any precision what abilities are important for a particular type of learning. The most precise results can be drawn from the design employed by Bunderson (1965). From his type of study we not only learn what abilities are important in the type of task employed, we can indicate at what stage they are impor tant. A design similar to Bunderson*s was employed in this study. Intellectual Abilities A basic difficulty with the above studies is that the abilities chosen to be investigated were not selected from any coherent theory of intelligence. Duncanson (1964) points out that his failure to establish relation ships with any concept-learning task could be due simply 12 to the fact that he selected the wrong abilities. Guilford (1956, 1957, 1959; Guilford $ Merrifield, 1960; Guilford and Hoepfner, 1963) has developed a model of intelligence which should serve as a base for the investigation of learning and intellectual abilities. The Structure-of-Intellect Model Guilford calls his model the Structure of Intellect (SI). The SI model is a three-way classification of intel lectual factors designed to encompass known aptitude fac tors and to predict the existence of undiscovered ones. The three dimensions of the SI model specify the content, the operation, and the product of a given intellectual act (see Fig. 1). The theory hypothesizes four kinds of intellectual content: semantic, symbolic, figural, and behavioral. Semantic content is most often encountered in standard aptitude tests that demand recognition of word meanings. Symbolic content refers to information such as numbers and letters, which have no significance in themselves. Infor mation that is concrete or perceivable as space or contour is defined as figural content. A synonym for figural is spatial. The kinds of information subsumed under the fourth content area, behavioral, include feelings, motives, thoughts, intentions, attitudes, or other psychological 13 Units C la s s e s R elations S y ste m s T rans form ations Im plications c6‘ A \0 F ig . 1. - - T h e o r e tic a l m od el for the com p lete "Structure of Intellect. " PRODUCTS 14 dispositions which might affect an individual's social behavior. The operation dimension of the SI model includes intellectual processes or activities, that is, what the organism does to any particular kind of content. Five different intellectual operations are hypothesized: cog nition, memory, divergent production (generation of variety of output), convergent production (generation of the one correct solution), and evaluation (judging in terms of criteria). The production dimension includes the results of intellectual processing. The six product categories are as follows: units (elements having "thing" character), classes (aggregates of units, the members of which have common properties), relations (connections between units), systems (organized or structured information), transfor mations (changes or redefinitions in known information), and implications (extrapolations in the form of predic tions or identification of antecedents). Each intellectual factor may be uniquely described in terms of one of four content areas, one of five opera tions, and one of six products. The structure-of-intellect model predicts the existence of 120 separate intellectual abilities. To this date, 79 abilities have been measured, many of them suggested by the model before they were empirically demonstrated (Guilford 4 Hoepfner, 1966). IS Guilford (1962) utilized these abilities as hypo thetical constructs in what he refers to as an informa tional theory of learning. He regards man as an agent that acquires information, retains it, uses it in generat ing new information, and evaluates information in connec tion with any of these steps. He defines learning as the achieving of information, and regards the various intel lectual abilities as abilities to learn. Guilford (1967) has identified the abilities in volving the product of classes as those important in con cept learning. Since the product learned is a class, the abilities which are a conjunction of a particular opera tion and content with the product class should be relevant to learn a concept. Although all 20 of the abilities dealing with classes have potential relevance to the investigation of concept learning, it was beyond the scope of this study to include the entire classes domain. To maintain generality over content, but to circumscribe the domain, only seman tic, symbolic, and figural materials were utilized in this study. In addition, only four operations were brought into focus, namely, cognition, divergent production, con vergent production, and memory. Some of these abilities have been previously iden tified, some have not. Guilford and Hoepfner (1966) have provided some information regarding a history of each 16 known structure-of-intellect factor. The cognition of classes abilities is well established. The cognition of figural classes was identified in^four different investi gations (Gershon, Guilford, 6 Merrifield, 1963; Guilford, Green, Christensen, Hertzka, 6 Kettner, 1952; Guilford, Kettner, 6 Christensen, 1956; Hoepfner 6 Guilford, 1965) . The cognition of symbolic classes was found in two studies (Guilford, Merrifield, Christensen, $ Frick, 1960; Hoepfner, Guilford, $ Merrifield, 1964). The cognition of semantic classes has been identified five times (Guilford, Hertzka, 6 Christensen, 1953; Guilford, et al., 1956; Merrifield, Guilford, Christensen, § Frick, 1960; Nihira, Guilford, Hoepfner, 6 Merrifield, 1964; O'Sullivan, Guilford, 6 deMille, 1965). The divergent production of classes abilities are less well defined with the exception of the divergent pro duction of semantic classes. It has been found in five different investigations (Guilford, Frick, Christensen, § Merrifield, 1957; Guilford, Merrifield, 6 Cox, 1961; Guilford, Wilson, $ Christensen, 1952; Hoepfner 6 Guilford, 1965; Marks, Guilford, $ Merrifield, 1959). The divergent production of figural classes was identified in two studies (Gershon, et al., 1963; Hoepfner 6 Guilford, 1965). The divergent production of symbolic classes was reported in three studies (Gershon, et al., 1963; Hoepfner 6 Guilford, 1965; Hoepfner, Guilford, 6 Merrifield, 1964). 17 With regard to the convergent-production-of-classes factors, only the convergent production of semantic classes has been identified, and it was found in only one study (Merrifield, et al., 1960). The convergent production of figural classes and the convergent production of symbolic classes have never been investigated. Recent investigations have identified two memory classes factors. Memory of semantic classes was identi fied by Brown, Guilford, and Hoepfner (1966). Memory of symbolic classes was found by Tenopyr, Guilford, and Hoepfner (1966). Design In order to establish more clearly the domain of classes abilities, a major focus of this study was to identify the heretofore unmeasured abilities, and to strengthen the measures of weakly established abilities. It was felt that the more adequately the abilities were measured, the better possibility there would be of estab lishing meaningful relationships with concept learning. Thus this study was designed to meet two goals. First, an attempt was made to confirm a number of hypothe sized classes abilities generated by the structure-of- intellect model. Second, an investigation of the rela tionship among these abilities and concept learning was undertaken. 18 The first step in relating the classes abilities to concept learning was to delineate the meaning of the term. A wide variety of phenomena have been subsumed under this rubric and related verbalizations, e.g., con cept formation, concept attainment, concept identification, etc. T. S. Kendler (1961, p. 447) criticized the use of these various terms because they ” ... are used to de scribe experimental procedures that either overlap or are not discriminably different." She refers to concept learning as " . . . the acquisition or utilization, or both, of a common response to dissimilar stimuli." Con cept learning is considered by Hunt (1962) to be the learning of a decision rule whereby an element can or can not be assigned to a class. Both of these definitions are broad and would include various kinds of concepts, e.g., the conjunctive, disjunctive, and relational concepts de scribed by Bruner, et al. (1956). In this investigation it was decided to focus only on the type of concept in which the stimuli share a common property or attribute. This is consistent with Guilford's (Guilford 8 Merrifield, 1960, p. 5) definition of a class --"Aggregates of items of information grouped because of their common properties." Thus the "dissimilar stimuli" in Kendler*s definition would have to share at least some common aspect. The decision rule that Hunt refers to would simply be whether or not a given stimulus had a 19 particular attribute. The confusing terminology employed with regard to the term "concept" is adequately reflected by the plethora of tasks investigated under this same rubric. In order to establish continuity with previous factor-analytic inves tigations the concept-learning tasks employed in this study were modeled after Allison's (1960). As mentioned above, he had found general relationships between abili ties and his tasks, and it was hoped that this study would add considerable definition to his findings. It was expected that the abilities would relate differentially at different stages of the learning process. Consequently, a procedure to ascertain the relationships between the abilities and scores at various stages of the learning process was adapted. Although learning was to be inferred by measures of performance on the trials, an additional measure was obtained. At the end of each learning task the Ss record ed verbalizations of the concepts. It was felt that per formance on this measure would demand an additional abili ty, viz, the conceptual naming ability. In summary, the basic plan of this study was to measure selected SI abilities emphasizing those dealing with classes, and then to determine the relationships of the measures taken from the concept-learning tasks to these abilities. Mental tests and concept-learning tasks 20 would be administered to the same group of subjects. The tests would be factor analyzed to identify the SI factors. Factor loadings on these factors for scores representing the trials and the verbalization responses would then be determined to ascertain the relationship between the SI abilities and performance in the learning tasks. CHAPTER II THE HYPOTHESES The hypotheses presented in this chapter are orga nized with respect to the two goals of this study. The first section is concerned with the hypotheses regarding certain intellectual abilities predicted from the structure-of-intellect model and with the experimental tests designed to measure them. In the second section the hy potheses regarding the relationship between the abilities and the concept-learning tasks will be presented. The SI Abilities The major hypothesis is that 11 separate classes factors predicted from the SI model will be identified as distinct from one another and from other factors repre sented in certain other domains of the model. That isf a satisfactory orthogonal factor solution can be found in a factor analysis of tests representing the 11 classes abilities and tests of other abilities already established through studies based on the SI model. The specific tests hypothesized to measure the 11 classes abilities are pre sented below. The name of each ability is given along with its SI code designation. The tests are listed 21 22 underneath the ability they are hypothesized to measure. Descriptions of the tests may be found in Appendix A. CFC Cognition of figural classes Figural Class Inclusion Figure Classification Figure Exclusion CSC Cognition of symbolic classes Letter Classification Letter-Group Exclusion Number Classification Number-Group Naming CMC Cognition of semantic classes Sentence Classification Verbal Classification Word Classification NFC Convergent production of figural classes Figural Hierarchical Grouping Figure-Concept Grouping Figure Grouping Restricted Figural Classifications NSC Convergent production of symbolic classes Letter-Concept Grouping Letter Grouping Restricted Symbolic Classifications NMC Convergent production of semantic classes Concept Grouping 23 Group Classification Largest Class Word Grouping DFC Divergent production of figural classes Alternate Letter Groups Multiple Figural Similarities Multiple Grouping of Figures DSC Divergent production of symbolic classes Multiple Grouping of Nonsense Words Multiple Letter Similarities Name Grouping DMC Divergent production of semantic classes Alternate Uses Multiple Grouping Utility Test MMC Memory of semantic classes Classified Information Picture Class Memory MSC Memory of symbolic classes Memory of Nonsense Word Classes Memory of Word Classes Tests of four reference factors outside of the classes domain were employed. The cognition of semantic units (CMU), the cognition of semantic systems (CMS), the divergent production of symbolic units (DSU), and the convergent production of semantic units (NMU) are 24 abilities that were thought to have possible relations to some of the classes tests. A completely adequate set of reference factors was impossible to include because of the limitation of available testing time. The four selected reference factors have been well established in previous investigations. CMU has been iden tified in several studies (Guilford, et al.f 1953; Guilford, Berger, § Christensen, 1955; Guilford, Kettner, $ Christensen, 1955; Guilford, et al., 1961; Hoepfner, et al., 1964; Merrifield, et al., 1960; Nihira, et al., 1964; O'Sullivan, et al., 1965; Petersen, Guilford, Hoepfner, § Merrifield, 1963) . CMS has also been consistently found (Guilford, Green, Christensen, Hertzka, § Kettner, 1952; Guilford, Wilson, § Christensen, 1952; Guilford, et al., 1953; Guilford, Kettner, 6 Christensen, 1955; Nihira, et al., 1964; O'Sullivan, et al., 1965). DSU has been reported in a number of studies (Guilford £ Christensen, 1956; Guilford, et al., 1961; Hoepfner § Guilford, 1964). NMU has been identified in three investigations (Guilford, et al., 1956; Guilford, et al., 1960; O'Sullivan, et al., 1965). The following is a list of the reference factors with the tests expected to define them: CMU Cognition of semantic units Verbal Comprehension Word Completion 25 CMS Cognition of semantic systems Problem Solving Ship Destination Test DSU Divergent production of symbolic units Suffixes Word Fluency NMU Convergent production of semantic units Naming Meaningful Trends Picture-Group Naming Word-Group Naming The Relationship of the SI Abilities to Concept Learning The major hypothesis is that there will be sub stantial relationships between the SI abilities and perfor mance on the concept-learning tasks. That is, the scores on the concept-learning tasks will load significantly on the ability factors. Although the exact form of the rela tionships could not be specified a priori, a content anal ysis of the tasks led to certain expectations. In order to make consistently correct responses to classes of stimuli, it is first necessary to recognize or to produce the attributes in common among stimuli. Thus cognition and divergent production of classes should play important roles in the early stages of learning. Detailed descriptions of the tasks are presented in Chapter III. See also Appendix C for the instructions used for each 26 task and for examples of the stimuli. The tasks are fundamentally convergent. After a subject has identified or produced possible common attri butes among the stimuli his task is to form a unique classification of the stimuli on the basis of these attri butes. In particular, a subject was to form four indepen dent classes for each task. Therefore, convergent production of classes was expected to be important in the later stages of learning. To perform well on the tasks it is imperative to remember the relevant attributes encountered on previous trials. Thus memory of classes should play a dominant role until the classes are learned. It is hypothesized that only abilities dealing with content similar to that of the tasks will be involved. That is, only the semantic abilities will show substantial relationships to the semantic concept-learning task, figu ral abilities to the figural task, and symbolic abilities to the symbolic task. The verbalization responses required for all three tasks are hypothesized to load on NMU. This factor, tra ditionally known as concept naming, is defined by tests in which the S is to verbalize the common attribute of groups of words and groups of pictures. Also defining the fac tor, and giving it more generality, is a test in which S is to verbalize meaningful trends in series of words. It 27 is assumed that S can learn the relevant common attributes in the stimuli of the task without achieving a specific verbalization. When S is confronted with the task of verbalizing the attributes, however, he is performing a function similar to that involved in the measures of NMU. To summarize, it was expected that the divergent- production- and cognition-of-classes abilities would show substantial relationships in the early stages of learning; the convergent-production-of-classes abilities in the later stages; the memory abilities in all stages until complete learning occurs. Only the abilities concerned with content similar to a particular task were hypothe sized to show substantial relationships. The scores ob tained from the verbalizations of the concepts for all tasks were hypothesized to load on the concept-naming factor, NMU. CHAPTER III PROCEDURES Ability Measures Sixteen new tests were constructed by employing the SI model in two ways. First, specific examples of tasks were deduced from the operation-content-product com bination. Second, tasks were devised that were similar to those for established SI factors having one or two attri butes in common with the factor that was to be measured. Twenty-seven tests were selected from the list of recommended tests for identified SI factors provided by Guilford and Hoepfner (1963). New items were written for three of these tests, Figure Classification, Figure Exclu sion , and Word Classification, with attempts to increase reliability and univocality. The names of three tests were changed to make them more consistent with the names of the abilities they measure. Sentence Evaluation was changed to Sentence Classification, Letter Grouping to Letter-Group Exclusion, Seeing Trends I to Naming Meaningful Trends. The new and revised tests were pretested to deter mine clarity of instructions, appropriate difficulty lev els, test reliabilities, and optimum time requirements. 28 29 The pretests were administered to freshman and sophomore students at the University of Southern California, Los Angeles City College, and Pasadena City College. Concept-Learning Tasks Three concept-learning tasks were devised. In order to establish continuity with prior factor-analytic investigations, the tasks were modeled after those used by Allison (1960) and later used by Manley (196S). Each task involved four concepts. Each concept was represented by 24 different stimuli sharing a single common attribute, so that each task had 96 stimuli. The 96 stimuli for each task were group administered in printed booklets. The method of anticipation was used in the presentation of the stimuli utilizing a form of "teach ing book." A stimulus or exemplar appeared on the page along with the letters A, B, C, and D, each letter repre senting one of the concepts. S was to choose a concept by encircling one of these four letters. After encircling a letter, he turned the page and the stimulus appeared again along with the correct letter. He then looked at the next stimulus, encircled a letter, turned the page, looked at the correct answer, etc. The instructions for each task appear in Appendix B. At the end of the task the Ss were asked to describe the concept each letter represented, in spaces provided at the end of the booklet. 30 A "trial" to be scored was defined as a combination of eight successive presentations of stimuli. In each tri al each concept was represented by two exemplars. The order of representations of the concepts was random within trials. The three tasks differed with respect to content. Examples of the stimuli used in the three tasks are shown in Appendix B. The symbolic task consisted of four-letter nonsense words. The four concepts to be learned were: two letters the same, first three letters in alphabetic order, words containing the letter S, and words beginning with a vowel. The figural task involved geometric types of fig ures. The four concepts were: right angle, parallel lines, intersecting lines, and a dotted line. Each stimulus in the semantic task was composed of a set of four English words. One of the words in the group fitted one of the four concepts. The four concepts were: leaders, edible things, animal sounds, and part. Subjects The subjects were 271 male and female, junior and senior students at Mayfair High School, Lakewood, Califor nia. Unfortunately, at the time of administration an in fluenza epidemic was rampant and attrition was unusually high. Since scores for every individual on all measures were needed for the analyses, the final sample was reduced 31 to 177. Administration of the Ability Tests and the Learning Tasks A large auditorium was used for the administration of the test battery and the learning tasks. Three sessions of three hours each were necessary for the administration of all materials. Each session was on a different day. The juniors were administered all the materials during one week, the seniors during the following week. The tests were presented in nine printed booklets, with the restriction that two tests for the same factor should not occur in the same booklet. The order of the administration of the tests and tasks is indicated in Ap pendix C, The symbolic-learning task was given the first day, the figural task the second day, and the semantic task the third day. Scoring All tests were scored by hand and independently check-scored by a different scorer. A scoring formula was applied to multiple-choice tests to correct for guessing. A description of the nature of the score and the scoring formula used for each test appear in Appendix A. Each subject's responses from the learning tasks were punched onto IBM cards. The number of correct re sponses on each trial in each problem was scored by means of a computer. CHAPTER IV ANALYSIS AND RESULTS Ability Measures A frequency distribution, mean, and standard devia tion were obtained for each test. The distributions of four tests, Concept Grouping, Memory for Nonsense Word Classes, Letter Grouping, and Restricted Figural Classifi cations were too severely skewed or truncated to meet the requirements for a Pearson r, so they were dichotomized near the medians. For all tests with two or more sepa rately timed parts, Spearman-Brown reliability estimates were computed. Kuder-Richardson estimates of reliability were computed for all one-part tests that showed no evi dence of speeding. The communality of each one-part, speeded test was used as a lower-bound estimate of relia bility. The descriptive statistics of the tests are pre sented in Table 1. A matrix of the intercorrelations of the 43 tests and Sex was obtained. Sex was included as a variable because of its possible relationship to the ability mea sures. In order to correct for the effects of dichotomi- zation of continuous variables, the point-biserial 32 33 > s f t H H f t A I t f t H 1 1 O t R DO f t f t l i f t U« ( 5 ' H o I J 1 ) (/)Q ^ £ i t i t i t i t r t f l ........ H H C O O Oll/>HNOON<H/)NOIONl/)N^'OI/)OlOOIOHIsoi/)'Ol/)NM ON*l/)Nt,<OiHOO r N O o r « - r > ' ' O N \ O f ^ ' t f i O i o i n r , ' 0 0 ' O o o o o t - '^, L / > ' O M r t ' O t o t s -r'Ouit>^t®Noo'Otk - r > t ^ u , > N r v \ o u i i o 'O M , 'l/ll'M O0>OHCIiOlOfs*rlO)l/)aiNOOOOOO0lOHOIOHOCM]IOlHN»tHlOlO*OOO M O N IO HO O M O O N S f'SyO rl'O ^'tO H lfllO O NO O ^iJlM O O . l/WOinCll/IIOIOIOOIl/lNI/) * ♦ * » !OI/>*IO lfll/HIO *IO NH**L/)NIO O Ni-IIO NIO NN<tN*M tO O C(JlO «*l/)l/)<ftO IO aiNl/) H H H H H NNO'OCllOl/)lOONI'»MONO)NI'«NVIOOOHNN\OOO^COONI/IOOtOOO'tOl('<^N*MO0l/) • lilt I I I I I NNNl/)H0ClNl/INOlOHOOHOOI/)NOMOl/)OOOIOCir'"t^<M/)Ht> >O lN *tlO O N O NHH H H ^ H H H rtl/IH H Nlfl NH H H H H H H rt rU t rt H H « H <t B N 0 U w s I < 1 0 <0 < N B < 0 H « U 0 0(/)U UZW W U U i M < i< 00 <f 0 l / I O U D U aftw Q 00 5 < <« *0 o u uw BZ z I < N 0 u a z < * I o U MHU < ft s o f t 1 0 UftUZ 0 z aft u I <I3U1< S k , rt<Z k0MW « W 3z<««i mu 0 R ft i Mo H iM HftWitUftOft D i U I S ) not Si P.MgUOO cty° '<: RfiOOl MSBZl t BR i iOBHO 3O0S'H S'HftU «RftOU*VlftQ HNORl (AHU'Hft o U' HzwHoauf tRooft ' HW mg i«m.rtoRO x m s u fti/i ftSft 3 i H 3 Z0Uft3i/iZl><tn ' OOgUhftMftfl.u Uitll OQ rt I H Uf> 0 I 'HS + J 0 3 «HW 'HW ft S.Oft HftUH g Uftit.ft i f t ZB M i RU MMMWCMdrigg i KHft i / i« 1 1 * I no mu m I t (01) ftSft U ftUftlH i 'Hn 3 b + j <D f t 0 MR f t H f t o ) BUft o f t 1 ) I / I I S I I / I 5 ) SO < 0 p u *io «<: ZOBiOiO Ui ooo i w o u p U U | 2 V ) V I S so 0'H< • r t ftlO f t BOQ f l UUN UftSO ftH U W f t f t 3 5 f t n i U b NU ONB UOrl S p o USU USB i UH P I o SB < S ' P C Z H HOB ZN 0 Oft 0 SUO I u UftftO 3 s <Sdi/|ft I W MZ < N Q ft I ft M3 JS UO B o d OU U 0 HBRB 3U g i M ftOftRriOURft Qfi 1 ) OfiHUft 0 MU I f t X M W ' O (t f t ' H f t f t iftftBJjZ M S P f t O O f t S l t C f t i ftmftft ft 4 4 ft) -H 3 -rt ftPWfififtftaftftl) MftZ S S:g viftp I ftx U 0 I g ft 1 ) B am dft ftviftft i/ift a ft 0 0 3 3 3 3ft MBft m aftft SmftW ftdftft it M « «H 3 n ft n v 3 aft in vi ti a a z * mo o oft r h m a n 3 > a w « n o ft m ft ft ft x z s BP 0 it l) n uh 3 vi it n u 3 3 ft ft ft ft lift fi m 3 d o h h b vii/iaftiiu ft ’Oft H f t d R U O mH d f i O O f t f t a U U U d a d d O H f t O ' O f l U f t ' l l i t g i / l H Ba a OtlllUUSHOXft dUHOft ftOO sohSubw## f t HHoSaSss ft ft'H UUWUH UUUU(ft(ft II 0 0 II 0 OSUU I ft ft Q | to 1/1 UUdESOO BltftftHH I U ft 11 HHHHH ft I I) 1 ) g U U U 1 ) II >s HOHftft i M 5 !i: £ 5 J!!! frWJWrfSS 1j 2 ft S°.3S33UU“' , ?U 33SSSS2££S6335S3SSSIIilSS!IS2£:a5S3SS5£IIJI5 HNtO*l/)Or, >OOfllOHNlO*l/>\OlsC0010HNIO*l/)0['.BOIOHNftn, l/)OMOOlOHNIO* H H H H H H H H H H N N NN N N N N N N IO IO IO IO IO tO IO tO IO lO d'^'lW ^ > s f t • H H f t A I t f t H I I f t f t 0 I I f t f t I t V ) g I t • H f t V I 1 ) B 0 V I D f t A U U f t U a i i B ■ p i b f t A 0 A 34 correlation coefficients of the four dichotomized tests with the other tests were converted to biserial correlation coefficients (Guilford, 1965, p. 324) and estimates of a Pearson r were determined for the phi correlation coeffi cients of the dichotomized tests with each other (Guilford, 1965, p. 354). The correlation matrix appears in Table 2. Communality estimates for the 44 variables were made by means of an iterative procedure. Multiple R 1s were put in the diagonal cells and 16 factors (the number of hypothesized factors) were extracted by means of BMD03M (Dixon, 1960). The communalities based on these factors were then placed in the diagonal and factors were reex tracted. This procedure was iterated until none of the communality estimates changed more than .05 from one iter ation cycle to the next. The iterated, stabilized communality estimates were put into the diagonal cells of the correlation matrix and 16 principal-axes factors were extracted. The principal-axes matrix is given in Table 3. The principal axes were rotated to psychological interpretability by means of a procedure developed by Cliff (1966). This procedure provides a least-squares fit of a matrix to a specified target matrix. The square root of the communality of each test was entered in a column of the target matrix representing the factor it was hypothe sized to measure. The remaining target loadings were set 1 . 35 38 18 33 30 35 42 43 27 25 28 34 30 33 39 25 28 25 32 51 2 . 35 30 35 32 25 36 28 29 23 22 38 32 21 26 30 32 26 20 38 40 3 . 38 3 0 3 0 37 25 24 29 3 0 22 31 41 37 38 25 34 39 32 U 18 43 4 , 18 35 30 36 37 34 41 42 36 33 45 47 47 45 51 22 28 18 27 40 5 . 33 3 2 37 36 5 0 51 50 5 0 39 28 53 49 53 37 50 47 45 31 15 47 6 . 30 25 25 37 5 0 45 53 57 43 30 39 45 36 38 4 5 32 28 26 26 42 7 . 35 36 24 34 51 45 37 45 34 20 40 36 33 29 36 33 3 0 25 16 48 8 . 42 28 29 41 50 53 37 49 41 30 43 53 47 4 5 65 45 42 24 28 48 9 . 43 29 30 42 50 57 47 49 37 31 46 53 43 4 5 34 41 40 31 30 48 10 , 27 23 22 36 39 43 34 41 37 17 34 33 28 38 43 43 24 H 16 35 11. 25 22 31 33 28 3 0 20 3 0' 31 17 37 42 32 27 43 21 27 12 25 21 12 . 28 38 41 45 53 39 40 43 46 34 37 41 41 44 39 53 38 25 24 32 13 . 34 3 2 37 47 49 4 5 36 53 53 33 4 2 41 47 5 0 61 44 39 24 20 49 14 . 30 21 38 47 53 36 33 47 43 28 3 2 41 47 45 53 39 37 33 15 41 15 . 38 26 25 45 37 38 29 45 45 38 27 44 5C 45 54 45 23 24 15 3? 16 . 39 30 34 51 50 45 36 65 34 43 43 39 61 53 54 55 36 23 M 50 1 7. 25 32 39 22 47 32 33 45 41 43 21 53 44 39 45 55 70 II 14 45 18. 28 26 32 28 45 28 30 42 40 24 27 38 39 37 29 36 70 25 11 36 19. 25 20 11 18 31 26 25 24 31 U 12 25 24 33 22 28 11 25 21 30 20 . 32 33 18 27 15 26 16 28 30 16 25 24 20 15 15 U 14 11 21 23 21 . 51 40 43 40 47 42 43 48 48 3 5 21 32 49 41 38 5 0 45 36 30 23 22 . 29 23 22 31 37 37 23 44 29 20 36 33 46 41 36 51 24 30 26 27 38 23. 38 25 32 23 30 30 35 32 33 22 25 34 43 41 35 49 37 31 27 30 41 24. 27 31 3 0 27 32 23 24 24 29 17 18 36 31 3 1 27 26 39 22 14 21 39 25 . 24 26 20 16 12 08 18 14 19 07 18 20 18 11 11 13 03 -01 03 24 18 26. 37 35 35 40 52 43 51 45 45 36 34 38 56 54 46 4? 46 41 29 23 51 27. 44 34 39 41 47 31 51 50 35 31 28 35 46 47 40 46 34 37 14 13 47 2 8 . 26 21 21 25 36 31 22 41 24 29 14 3 0 39 31 19 29 25 30 23 13 27 29 . 20 3 5 26 43 40 29 3 2 37 27 26 19 41 36 31 3 0 33 36 2/6 12 24 32 3 0 . 36 24 28 39 51 49 43 57 48 27 25 41 39 40 3 5 3 5 23 3 2 2? 21 36 31. 41 28 35 37 41 40 30 50 48 39 26 38 49 51 45 48 36 32 26 18 49 32. 30 27 39 52 41 51 27 42 45 36 35 36 40 49 46 42 31 30 29 28 36 3 3. 2D 26 40 37 45 33 26 35 35 22 30 4D 40 40 41 37 33 39 10 03 32 34 . 44 34 37 38 54 51 45 59 49 3 2 23 47 43 43 42 52 34 33 18 14 42 35. 30 25 19 37 19 U 21 31 21 18 35 33 31 34 34 40 19 32 14 22 28 36 . 37 5 5 25 20 25 19 24 25 17 18 17 23 21 12 18 37 IS 11 20 44 35 37 . 37 41 48 56 57 41 52 47 49 36 39 60 45 41 35 48 42 48 22 30 46 3 8 . 19 35 36 57 36 33 35 4 0 38 24 31 44 39 31 31 43 36 28 15 2 0 36 39 . 23 30 26 4 5 4 0 3 5 36 41 31 25 40 37 36 32 26 27 22 2? 14 18 31 4 0 . 26 34 40 63 41 36 40 49 36 26 42 49 38 40 36 43 33 29 16 29 36 4 1. 30 40 24 46 28 20 25 30 29 15 32 36 41 35 44 36 27 24 22 27 40 4 2 . 35 44 30 51 44 31 38 44 38 33 32 47 33 42 34 38 26 25 18 27 42 43 . 25 34 48 52 47 33 3 1 38 28 29 37 45 40 41 39 47 35 32 23 14 28 44 . 09 00 -1 4 -09 U 16 22 15 U 10 -1 2 - 05 -05 -0 2 -1 3 -08 -32 -07 08 -0 9 07 29 38 27 24 37 44 26 2 0 36 41 30 20 44 30 37 37 19 23 26 30 35 25 09 23 25 31 26 3 5 34 21 35 24 28 27 26 34 25 5 5 41 35 3 0 34 40 44 34 00 22 32 ,8 20 35 39 21 26 28 35 39 40 37 19 25 48 36 26 4 0 24 30 48 - U 31 28 27 16 40 41 25 43 39 37 52 37 38 37 20 56 57 45 63 46 51 52 -09 37 30 ; 2 12 52 27 36 40 51 41 41 45 54 19 25 27 36 40 41 28 44 47 U 37 30 23 08 43 31 31 29 49 4® 51 33 51 14 19 41 3; 35 36 20 31 33 16 23 35 24 18 51 51 22 32 43 3 0 27 26 4 5 21 24 5 2 35 36 4 0 25 38 31 22 44 32 24 14 45 50 41 37 57 50 42 35 59 31 25 47 40 41 49 30 44 38 15 29 33 29 19 4 5 3 5 24 27 48 48 4 5 3 5 49 21 17 49 3 8 31 36 29 3 8 28 11 20 22 17 07 36 31 29 26 27 39 36 22 32 18 18 36 24 25 26 15 33 29 10 36 25 18 18 34 28 U 19 25 26 35 30 23 35 17 39 31 40 42 32 32 37 -1 2 33 34 36 20 38 35 30 41 41 38 36 40 47 33 23 60 44 37 49 37 47 45 -05 46 43 31 18 56 46 39 36 39 49 40 40 43 31 21 45 39 36 38 41 33 40 -05 41 15 31 11 54 47 31 31 40 51 49 40 43 34 12 41 31 32 40 35 42 41 -02 38 35 27 11 46 40 19 30 35 45 46 41 42 34 18 35 31 26 36 44 34 39 -13 51 49 2c 13 47 46 29 38 35 46 42 37 52 40 37 48 43 27 43 36 38 47 -08 24 37 39 03 46 34 2 5 36 23 36 31 33 34 19 18 4 2 36 22 3 3 27 26 35 -32 30 31 22 -01 41 37 30 24 32 32 30 39 33 32 11 48 28 22 42 24 25 32 -07 26 27 14 08 29 14 28 12 27 26 29 10 18 14 20 22 15 14 16 22 18 23 08 27 30 21 24 23 13 13 24 21 18 28 03 14 22 44 3 0 20 18 29 27 27 14 -09 38 41 39 18 51 47 27 32 36 49 36 32 42 28 35 46 36 31 36 40 42 28 07 41 35 12 45 35 25 30 42 42 32 28 36 35 32 37 29 29 32 41 24 26 -04 41 30 15 43 39 17 27 34 42 26 29 33 28 22 41 36 22 32 30 26 31 -11 35 30 20 39 32 16 28 28 32 22 26 26 36 4 0 32 39 23 38 49 34 32 1° 12 15 2.0 23 20 10 29 10 13 16 12 19 13 20 27 19 16 20 16 29 25 07 45 43 39 23 61 26 42 47 52 40 44 47 30 25 53 42 43 45 38 42 40 03 35 39 32 20 61 34 44 48 46 39 38 48 30 15 49 41 38 43 33 37 44 12 25 17 16 10 26 34 23 33 29 19 22 33 12 11 26 25 20 20 21 32 27 11 30 27 28 29 42 44 23 29 3 0 33 24 30 24 36 49 3 0 28 41 32 39 41 02 42 34 28 10 47 48 33 29 44 41 35 6 0 29 17 5 0 42 43 46 26 38 38 37 42 4 2 32 18 5 2 46 20 3 0 44 44 42 46 35 16 39 41 35 34 39 38 34 12 32 26 22 16 4 0 39 19 33 41 44 34 47 23 08 38 2 5 3 2 31 27 23 47 07 23 29 26 12 44 38 22 24 3 5 42 34 34 25 14 48 43 43 44 39 37 38 -04 36 33 26 19 47 48 33 30 60 46 47 34 17 2.0 5 0 49 38 47 16 45 3 5 28 35 28 36 13 30 30 12 24 29 3^ 23 25 17 24 28 27 29 36 59 34 37 -07 32 22 40 20 25 15 11 36 17 ] 6 08 14 20 24 27 25 23 33 40 31 25 -01 37 41 32 27 53 49 26 49 5 0 39 38 48 5 0 28 27 5 0 45 59 32 54 5 0 04 29 36 39 19 42 41 25 30 42 41 25 43 49 27 25 50 39 71 33 47 50 -05 29 22 23 16 43 38 20 28 43 35 32 43 38 29 23 45 39 52 34 43 41 03 32 32 38 20 45 43 20 41 46 34 31 44 47 36 33 59 71 52 36 56 56 -07 41 30 49 16 38 33 21 32 26 39 27 39 16 59 40 32 33 34 36 39 f -15 24 26 34 29 4 2 37 32 39 38 38 23 37 45 34 31 54 47 43 56 39 37 13 26 31 32 2,5 40 44 27 41 38 34 47 38 35 37 25 50 50 41 56 36 37 -13 -04 -11 -19 07 03 12 11 02 37 12 07 -04 28 -0 7 -01 04 - 05 03 -07 -15 13 -13 Note.— Decimal points omitted. UNROTATED FACTOR MATRIX 36 N A Z I *OtOIIONH*H4‘ «OIOOOIOIAlAHiOOOlAOOl«OOOHO'OCOiOMOlArtOOIIOOHNSCO*IO*r'> HeO«NO'OOeO*ION*COOOtOHN\OOONiOOliOTf'Ot'.<riOHlOOOtOOOVOlON«'lOtOH* OOOHHOOOOOONOHOHOHOOHOOOHOHHOOOOOHOOHOHOOHOO l i i i I I I I I i i i I ii III ll iOHOtOHHNCllONtO^QNNNHOIHCOtOOIlOlOh'iOH'DOfs'OHIsOtsHlOHNNN*tlOiO OOOOOONOOOOONOOHOOHOOOOOrtOONHOOHOOHHOOOOOHHO ll i i i I I I i I I i i i I l ii ii *i/)ionK)r'>HCiinNcoiO!flOHnt')r'meoHiO0 NVHKWNOi/)«eoiooaNiONMi/)OK>N OOOOHOHOOOrtOHOOHHOOHHOOrlOOHONNOHHOHHOOHOOHOO i I I M I I l I I l ll i 1 I I I i i OlNtOtsOIOinHtOHrtNHOtOlOiO<lt"OOOIOlOOtOiOtOOOtOlNrllOlONeON\ClOlOlO«rllO OOOHrlrlOOOOHOOHOOOONOOHrlrlOOONOOOHHOOrlOHHOOHOO I i i i I I I I I I I I I I I | i i I I I l/|NNrtO*OOIA**IONNO>NiO*OOllOIOHr'HOOON*OIOOiOHOOHIOKKHOOrlN*lflH OOHOHOHOOOINOOOOOOrlHHOOONOHHHOOOHOHHHHHOOOOHO 1,11 ill ll I i I I I I I I I ll i0iO^fflfflOt'HOIM->COf>f'HNN0NNOl«OlHlfllflOOlOOlOTt'<fONN«COOlOlOl£)NO NOHOOONHOOOOOOOOOHHOOHHOONOOHOOOOHNOOOOOOHOH I I i I ii i i i I I I I I I I I I I <fNOlOlNONrlNH*lOHeOOOtO»OOHlOHIOIOIONiOlOlfllOMflNOl'C'OONNOl<HON0 OOOOrtHOONHHOOOHOONNOHHOHOOONrlOHOOOOOHNOOOONO I i i i I I i l l i I I I ill I I I I NH*Ct''NHOOOQIOOOlOVtOClHtOlflOO)NMOOO<tf'OOH'ON<t'^A,^'HIOeOOlHf'CO h h h n o o h h o n n o o o h o o o o h o n h o o o o o n h o h o h o o o n o h h h o o I I I I I I I I I I I t I I I I I I I iOrllOHOOOlONIOIOrl\OlOKH'OOHOMNI'-NOOlt'-<ftO\CtOIOIfl01IONlO\OtOOOlrlrlMfltO HONOOHOONOOrlOOOMONrlHOOHOHONOHHOOOHHOOOOONOHO II ii i I I I i I I I I I I I I I I 0 OHrlf'NVtOI''OltOOIOIOCONr'HNtOHIONOO*|sOMOONOIflOt‘ 'A'NO)NNHOt"0'0 HOMOOOONHHHOOOOMHOOHOHOOHHOHHOHNHOONOHOHOOOO 1)111 I I I I I III I I I I I I I t>.*OailOOOIO\OMOIflCONIO\OOOHi-IO'Orl*lflOONOOO)NIOOHIOlOHONN>ttOtOOM '"0 o o o h o n h o h h h o o o o o o h h i o h o o h o h n o o h h k i h o n o o o o o n o o h I I I I I I I I I I I I I I I I I 1 I 1 1 IOIOOOCOiO*HMOOOOlOlOOlO)lO*rlOtOOOCO«,M''HO<t''tNCl'-ltOiOIOOHOlMOI''OOHiO ONHrlHONOOONrlHHrlH^NHOONOOOOOOHrlHNOOMNNOOONOOO III I I i I I I II I I i I I I I I i I OOiOt'.NKWN0OlON(OCOHt'>NNlO<tMOHOONHIfl<tlO»taiO<tlOtMOOOailOIOOV'f<»N IOHOIOOOOOHOHHOOOrlrlONNNNi-IHOOOOOOrtONOOtOHtONtfHi-INO ill ill i i I I I I I I I I I I IfliOOlNiONiONHCOIOl/lOOOOHtONMVONKlOMOO^riHrlO'OffltOHNOMOlflMiOOOVN iHNOOOOHOQOOOHHNNlAlOOIOOOHONOOOHNOOHrtO*OHHHONO* ll I I I I I I I I i i i ii ii Mi 1 \ONNiONrlNIOIONM'OIOf'NNOllOO^,lOiOI'-NN\ONf'NIOHiOIOIOtOOOHlOtOHtfl''CQN ONHHNtONNNHHOOOOOOOHNOOOtOHOHrlrllOHHOtOlOlOOHON«OHifi I i i i i I I II i l l I I I I I I I I lONnt'-OrlOlH'OOOl'OOIONIOOIOiOMOiOlOOOONMOIOlOlOQNOOeONlOlO'OOi'OlOION < ( * I I I I • • I « t I I t I » I • I I I • ♦ * * » * * * ♦ * * 1 ♦ * • * 1 1 ’ HNt0^l0^r'C0010HINt0^iO\OtvCOCiOHNIO*tlO'OtNOOOlOH<NtO'tlOiOrsCOOlOHNIO^ HHHrlHHHHrlHNNNNNNNNNNKlMIOKItOnKlKllOtO't^^'t^ (0 to 0 1 to t (0 * 1 in * 0 1 * ' i « P 0 in 0 . u fi 0 1 4 1 in 0 0 i 'H 4 1 1 0 •0 U 1 0 1 0 p 0 0 1 0 0 P u • f t c 4 ) 4 ) 1 0 U O C r- X •rl ' 0 w A 1 0 0 4 1 i P P 1 0 H E • 0 H 0 0 1 p to a • H 0 a 0 1 1 0 H t A H E < H 1 0 u 0 1 4 ) 1 Q rl i N ( N 4 1 • P 1 0 0 H z 37 equal to zero. The resulting rotation indicated that some variables would not meet the targeted loadings. These variables were changed in the next target matrix and thus subsequent rotations were performed. These adjustments of the target matrix were continued until the solution pre sented in Table 4 was found. Interpretation of the Factors The interpretation of each of the 16 rotated fac tors is based upon the apparent factor content of the tests loading .30 or more upon the factor. The test load ings for each factor will be listed along with all addi tional significant loadings of the tests, if they proved to be complex. CFC Cognition of figural classes 7. Figure Classification (CFC) .47 5. Figure Class Inclusion (CFC) .46 9. Figure Exclusion (CFC) .44 (.38 NFC) 6. Figural Hierarchical Grouping (NFC) .42 (.41 NFC) 19. Multiple Figural Similarities (DFC) .35 (.35 DFC) The three leading tests on this factor, Figure Classification, Figural Class Inclusion, and Figure Exclu sion, had been hypothesized to define CFC. However, the third test, Figure Exclusion, is complex, loading substan tially on NFC. In the development of the convergent- production-of-classes tests, an effort was made to control ROTATED FACTOR MATRIX 38 u N ygiOlAOOrKHt'OI*OONi/!i/)OiAOOlflG>OlCO«OHOIiOOOlOMOl/IHOOlOOHIOM'***N £ iOy\ON'0'OOt,"OlOl/ll/)'Ol/ll/IOlOl'OyiOi/)iOyi/)N'OiOyi/IMnM/)MOCOM'i <tM''iO'OiO u N HOH#HOOlHNf»HHK|iOONHiO*OOOItOiO*'OH#HGlt'HI/IOIflNNNHO«OIOOl J o O N H H O N r t N O O H O O r l N N O O N O H O H O O H N O N O N O r l O H O O O O r l O O ^ ^ i I M i l l ................................................ I i l l I M I M ,f l i/ | f > M / l N l O l O H l / ) I O y C lK Q N l / l l /lf ' > f ' O I 'O r l l ' » O O O y N O N O O l M O H O O ' O I O N C O O r l N ' 0 5 o O H t O N H O H H O ^ I O H N H N O N N H O H H O H O O N N H O N H O N O M H N N H N ^ t H Z i 11 1 1 3 inOlH\OyHCQ^OlOyXOIOWlONrlCOIOl/INh'0lOClfs HH*fl''>M/IHHCOOOr*OQO'Ol/)ClH J riN H N r l O H H O N O N r lH H r t H H O H H O O r H f r lN H I /) O H O O N H H N H O N H y H H fjj i i I I I rj n\HOOOOOOWIs OMOOiyOHHfflHOh-OOyNIONi/)^>yyNOlN5iO'N(JiyNHl/)fllN5l^ m rtO N N H N O N H N r l O N K i y \O H O H O H N N O O H H O H r tN T f O N H O O O O O O O N O ^ 1 I I I I l l rj ,f t/|IO«)rtrlHO»NiDl/IOO'OlOtOCiyHlON'OHNyyHNN«Cil'"OOOIOO'COHl'C>IOHIOt^ u, H O H r t H < tO ir t M « rt H N O N N H H O N H H O O O O O N O N H H O N O O O O H O O N r tH 2 i I I I i i n j j i O M O « O O I I ' r l C l « C O O I l r t O ! N O M O N f ' » O 0 l N N C l H l/ > H t \ M I O l O i O * O l O O « O O H O O i O ' O T N y H H O H H H O O H O O O O H O O H U 1 N H H N H H O O N O O H O O H M S O H N N r ( H O 0 " (J KlCOI/lsOC60NI/)<, l/Jl/>H^^r^®OiO>NOO^H(*'OlW' flKIOOOKlNNlOOiOlNOIONO><'*H<7l If t H O O O H O O H rl O H N N N rl N H O N H H l/ ll O M O N O H H H N H H H H N H N O H N H O H Q I I I 11 3 Ooyi/)ClOyNHN*HO0lMOOOOlOf,'yHaiOOOyONycQOHIOCO'ONONt'aitO'OCOOOtO I/) o H O N O O O H O O H H O r lN O O H O O H H O N O r l r lO O H N r lr lO ' O H O O H H 'O H H O Q i l l I l l 1 U tvONl/lOtfl/l'Or'NM/lt-'OMHOM/ll'-OMMf'WI'IOrlOCOMC'HOOOOKWWHQKiytOl/l U, ( f l N N O H O N H N H O O H N H N O H r t r t y O N O H H N N H O N O O r l H H r l O O O r l N O H 0 * HI/IOO'OHNCOyMONCieO'O<tl/HtO'OOMO0ll/IOOQOO*l/IOt,>rlNNl/lt'.Hr'Orlt> >N « H H ' t O N r lH H O O O r H H r t O O H H H O rl O O H O r lp U f H H H N N N H H O H H O H H N O P u N^OHlrtNOlNOHHr^yNiOfiONI/lNy0^tOXrMHrs'NyQ|lO'OHK10^'OlflOHOI/l^iOts g H H N O N O H N rl N O N r lr l H N X i O O O H O H H H r l N H H O O O r l H H O N H O H O O H N 0 i l l I I I in (fl n H 0 0 l/l ifl O i 10 H ifl 01K 1 ifl K ) 00 H t-. \0 Ifl X lA £ H O i i0 £ 10 C O 0 * N C l M N 0 0 N \0 l/l 01010 C O t> S N O rl O H rl H N N O O H O r lO O O H O r l O H r tH O H N O O O N N O it H O N N H N O H H lO C I I rl 0 u HMOr^OOlOHH0HiOOOi/li/lNNXMs f^'ONyNt^XIOy0^yi/lUlN(/l0X0Is 0 X N £L y N H K I O N r l N H N O y N N H H H H r t H O H O O H H N H O O H H O ^ N r l H I O H I O N H N H O U I I I *IOIOOIiOHOlCONOOl*<ONeOOIOOIOO*M''OlOM''*00'DOIH01NOIN<'OOl/INN#Oli/l E O H H y O H H H O H O H H H r t N H O O O H O H H O H H O O H O O H N O H N i O N O H N N H >H 0 U | / ) iO r O i O t ' 'O O O H H I O N H I O O lO C ) I O H l /l O 0 l O O I O O t O t O U I i C O C I I ' N C l O N N r l l / ! t O C i y y » f O W H O H N H N N N N H N O y N N H H N O H N K l N H H l / l l / l H N N N N N O H H N r t N r l H O H O U 1 1 u Qt'.HOONMOyiO0lM/)IOCllOO'OIArlt'>fl,M/>NCOCOHyt'.eOiOOOIOait'-yNiOCONCOOIO 11 U, O r l O N y y y O ^ N O N H N H O N H I O O N O O H O N O H H N H N H N O O N H H O H H r t N ¥ U I I I 1 1 ® P I/) ♦ » * i i i * i . . * i » » « • i i ♦ » i » • • • * * * * • « ♦ 1) H N I O ' t l / l ' O M O C l O H N I O y i / l i O r ' ‘ OOOlOHNlOyi/l'OI '1O O O I O H N r O * l / ) ' O M O C l O H N K H (h H H H H H H H H H H N N N N N N N N N N I O I O I O I O I O I O l O l O t O K H H K H H t 39 cognition by utilizing common properties that are readily recognized. Apparently that effort was not entirely suc cessful. A test designed to be a univocal measure of NFC, Figural Hierarchical Grouping, and a test designed to be a univocal measure of DFC, Multiple Figural Similarities, are complex, both sharing cognition variance. CSC Cognition of symbolic classes 26. Number Classification (CSC) .53 27. Number-Group Naming (CSC) .53 13. Letter Classification (CSC) .43 22. Multiple Grouping of Nonsense Words (DSC) .30 (.51 DSC) Three of the four tests designed to measure CSC, Number Classification, Number-Group Naming, and Letter Classification, load univocally on this factor. The fourth test, Letter-Group Exclusion, loads univocally on NSC. In this test the subject is to form a class from a group of figures, excluding the one figure which does not share a common property with the class the S forms. This type of task traditionally has been identified with cog nition. The figural parallel of this test, Figure Exclu sion, is complex in this analysis, loading substantially on both CFC and NFC. Neither test has been included in analyses where the two convergent-production factors were identified. The fact that both tests have large loadings 40 on the convergent-production factors seems to be strong evidence that the traditional concepts of factor composi tion of these tests are not justified. Multiple Grouping of Nonsense Words, a test de signed to measure DSC, has a small loading of .30 on this factor. Again the failure to control cognition is appar ent. CMC Cognition of semantic classes 33. Sentence Classification (CMC) .44 11. Group Classification (NMC) .41 (.43 NMC) 39. Word Classification (CMC) .38 37. Verbal Classification (CMC) .35 (.36 NMC) 3. Classified Information (MMC) .33 (.40 MMC) The three tests hypothesized to measure CMC, Sen tence Classification, Word Classification, and Verbal Classification, load on this factor. Again we see a test designed to measure convergent production, Group Classifi cation, with a large loading on a cognition factor of simi lar content. In the test Classified Information, S is presented with several groups of words, each group sharing a common property, on a study page. He is then given a test page with different groups of words, some of which have the same common property that appear on the study page. His task is to indicate whether or not the common attribute present in each group of the test page is a 41 common attribute of the groups on the study page. Since S has to recognize what attribute is common to the words in the group on both the study page and the test page, the loading of this test on CMC is not surprising. MSC Memory of symbolic classes 17. Memory for Nonsense Word Classes (MSC) .82 18. Memory for Word Classes (MSC) .63 This factor is clearly defined by the tests hypothe sized to measure MSC, Memory for Nonsense Word Classes and Memory for Word Classes. MMC Memory of semantic classes 28. Picture Class Memory (MMC) .40 3. Classified Information (MMC) .40 (.33 CMC) The two tests expected to define MMC, Picture Class Memory and Claisified Information, load substantially on this factor. Classified Information proves to be complex, loading significantly on CMC. DFC Divergent production of figural classes 1. Alternate Letter Groups (DFC) .57 21. Multiple Grouping of Figures (DFC) .40 19. Multiple Figural Similarities (DFC) .35 (.35 CFC) The three tests hypothesized to measure DFC, Alter nate Letter Groups, Multiple Grouping of Figures, and 42 Multiple Figural Similarities, load on this factor. One test, Multiple Figural Similarities, loads partially on CFC, indicating that the cognition aspect was not con trolled. DSC Divergent production of symbolic classes 22. Multiple Grouping of Nonsense Words (DSC) .51 (.30 CSC) 24. Name Grouping (DSC) .39 23. Multiple Letter Similarities (DSC) .37 The three tests expected to define DSC, Multiple Grouping of Nonsense Words, Name Grouping, and Multiple Letter Similarities, have substantial loadings on this fac tor. However, the test having the highest loading, Multi ple Grouping of Nonsense Words, has a small side loading on CSC. DMC Divergent production of semantic classes 36. Utility Test (DMC) .73 20. Multiple Grouping (DMC) .57 2. Alternate Uses (DMC) .46 This factor is clearly defined by the three tests hypothesized to measure DMC, Utility Test, Multiple Group ing, and Alternate Uses. NFC Convergent production of figural classes 8. Figure-Concept Grouping (NFC) .50 6. Figural Hierarchical Grouping (NFC) .41 (.42 CFC) 9. Figure Exclusion (CFC) .38 (.44 CFC) 10. Figure Grouping (NFC) .32 Three of the tests intended to define NFC, Figure- Concept Grouping, Figural Hierarchical Grouping, and Fig ure Grouping, have significant loadings on this factor. The substantial loading of Figure Exclusion on this factor as well as the side loading of Figural Hierarchical Group ing on CFC were discussed under the cognition factors. Restricted Figural Classifications was hypothesized to mea sure NFC. This test does not load significantly on any of the factors. No explanation seems apparent for this result. NSC Convergent production of symbolic classes 16. Letter Grouping (NSC) .61 32. Restricted Symbolic Classifications (NSC).47 15. Letter-Group Exclusion (CSC) .41 14. Letter-Concept Grouping (NSC) .30 The tests designed to measure NSC, Letter Grouping, Restricted Symbolic Classifications, and Letter-Concept Grouping, have significant loadings on this factor. The fact that Letter-Group Exclusion loads univocally on this factor was examined under CSC. 44 NMC Convergent production of semantic classes 11. Group Classification (NMC) 43. Word Grouping (NMC) 4. Concept Grouping (NMC) 37. Verbal Classification (CMC) 12. Largest Class (NMC) .37 (.49 CMU) .36 (.35 CMC) .34 .42 .43 (.41 CMC) Four tests were designed to measure NMC. Two of them, Word Grouping and Largest Class, load univocally on this factor. The other two, Group Classification and Con cept Grouping, load significantly on this factor, but each has a side loading on another factor, CMC and CMU, respec tively. Similar to the effort made to control for cogni tion of classes by keeping the attributes at a high level of familiarity at which little individual differences would occur, an attempt was made to control for cognition of semantic units by utilizing words well within the range of high-school students. This effort, although for the most part successful, seems to have been inadequate with respect to Concept Grouping. is somewhat surprising. In this test S is to assign words to one of two classes or neither, the classes being defined by groups of words sharing a common property. This task could be considered as forming a unique classification of the words in three classes, two being defined by the The loading of Verbal Classification on this factor 45 specified groups, the third simply having neither of the two common properties. Regarding the test in this manner, its loading on NMC is reasonable. CMU Cognition of semantic units 38. Verbal Comprehension (CMU) .68 40. Word Completion (CMU) .62 4. Concept Grouping (NMC) .49 (.37 NMC) The two tests intended to define CMU, Verbal Com prehension and Word Completion, have large loadings on this factor. Concept Grouping has a substantial loading on this factor, as mentioned earlier, but no other verbal tests were so loaded, indicating generally good control of CMU. CMS Cognition of semantic systems 30. Problem Solving (CMS) .60 34. Ship Destination Test (CMS) .43 44. Sex .38 (.49 Sex) Problem Solving and Ship Destination Test, the two tests designed to measure CMS, have substantial loadings on this factor. Sex loads significantly on this factor, the boys doing better than the girls. The common name for this factor is general reasoning, which has long been established as significantly correlated with sex member ship. 46 DSU Divergent production of symbolic units 41. Word Fluency (DSU) .66 35. Suffixes (DSU) .62 The two tests hypothesized to measure DSU, Word Fluency and Suffixes, clearly define this factor. No other symbolic tests were related significantly with it. NMU Convergent production of semantic units 29. Picture-Group Naming (NMU) .51 42. Word-Group Naming (NMU) .45 25. Naming Meaningful Trends (NMU) .40 The three tests hypothesized to measure NMU loaded univocally on this factor. Sex membership 44. Sex .49 (.38 CMS) In order to aid in controlling any variance that might be due to sex differences, sex was rotated as a fac tor in this analysis. Concept-Learning Tasks In the learning task each correct response was given one point, so the maximum score for a single trial was 8. For each task, the mean number of correct re sponses for each trial was plotted. The plots for these learning curves are presented in Figure 2. MEAN NUMBER OF CORRECT RESPONSES 47 70 68 66 64 62 60 58 56 54 52 50 .48 46 Symbolic Task Figural Task 44 42 Semantic Task 40 7 5 6 8 9 10 11 12 1 2 3 4 TRIALS Fig. 2.--Mean number of correct responses on 12 trials for the symbolic-, figural-, and semantic-concept- learning tasks. 48 The scores for individuals for each trial for each task were correlated with scores for all the ability tests. The correlation matrices of the trials with ability tests for each of the tasks are given in Tables 5, 6, and 7. An extension procedure developed by Dwyer (1937) and Mosier (1938) was utilized to obtain the loadings of the trials on the ability factors. The results of this procedure are presented in Table 8. The symbolic task has substantial loadings on fac tors DSC and MMC. The figural task has a relatively large loading on MSC. The semantic task is represented by siz able loadings on NMC, MMC, and MSC. Of more importance than the isolated high loadings is the pattern of loadings over trials for each task. It is with these data that there is hope to ascertain the dif ferential contribution of the abilities to each of the learning tasks. Plots of the factor loadings of the trials on the ability factors are given in Figures 3-14. There are four figures for each task; the factors representing a single operation are plotted in a single figure. The two factors most relevant to the symbolic task are DSC and MMC. The curve for DSC climbs systematically to trial 9 and then declines sharply (see Fig. 3). DSC implies the production of several common attributes, which would seem essential to arrive at the attributes in the 49 N H H H 0 H f t I H 01 f t OX f t W< fth 0 0 < ft h (JH ftZW W H h W H ZH Nft • H s o in HWH h JH ft ft 1 f t o ft OH m f t ft f t < xw« H HU ftZW i n H O I <UH S ^ U I Z h H « OftH h OS Hft <1 J H m ft f t f t ft 0 u N H f J ( / ) 4 ) h MO*<MNHr'*ClN\ONnai«,HNClOlH(ONrlHi/!l/)HCOCOOOI')COOOOOOt'>N'OOrt'teO[N NNHHNrtHNHHHrtNl'INNHNHONNNNHNflNHNNHHNHHHNNNNHKIO (O'O^NOOOOO'O'Ol/IN'O'^OflOtOfOOOHr'NO'OO'lOOOOOt^HHOOlrtMH^'dMfl’ C O O i O O t O H ' t NHNHNNNNNHrtNMNNNNNNONNNNHNt0NN«NHNMriOrtNr(HHH(0O r(lflHlrt'O^^MHlflN>ONIfl«OONOOVfflOOHinONOM'tOffl^WKl^*NIONrt^rltO NHNHNfONKlNHHNIOnNMNNHONlOMNHtONNNtOIOHNIOrtOlONNNHHrtO Q N O N lf lr lO O G lC llO G llf lt '’N U W O K } 'O C lN N lO N I f lO O C } f t f t O t O M '0 O 'O H h 'N f lO , J H G ) t '* N NrtNHNtOHNHNOMNMNNNNHONIOIONONHHHIOtOHNnHHNNHNHHNO t '.N O Q if l^ N W C O C u f M M K W ^ O M N N O O I O C O H lf lr lf 'S M O h 'N iO H N K lO M M I 'liO O O lH HHHHNNNNHHONNMNNONNONNNNONNNNKlNrlHNHHIONNNHrinO \ O N f l O O O H ( O N i/ ) f t O O i/> i /I O lO W H r t f lO O N C K ) l O K ) f f lN O O O ^ 't t 'O W I O iO < f O O 'O i n ^ 'O l f lr 'N HrINONNHNHHONNKINNNHNONNNNONHNHNNHHNHHHrtHHHONO N I ^ r l'O t O O O O r 'H Q lN lO N ( O O O O H W O t O N H N O 'O N C O H W K l^ l/lr ( 0 3 'O O ^ t H O N K ) f ^ N NHIOHNr<1HNNHO(ONlONrtNNHHNHNNONNHHNNHHKlHONNNNHHNO M/)M'‘00'OM/)0'C£N\DN*SG)i/>'OOrl5iNI')l')M'W*OOr''NHl/)'OmtONi/t't\OOIOOO HHrtHHNONHHONHlONNHHHOHHNNOHHHHHHHrlHHOHHHHHHNO Tt00N0001Nrl00NfllH»0O'0t0^i0O^NK)l/l0l'ONOI/ l ON VK )l 0i nr ll/lCfl0l/)00'0NK)O00 HOHOONHHOOOHriHHHHHHOHHHHOHOHOOHOOHOOOOOOOOHO HOHONHNCl(NOOO'0\OCOCOlrtOll/ll/lh'^Qlts -HOOOffiN\OOCiCi^ONO'ONHl/)'Oh'iO NHHHONHHHHHNHHHHHOHOOHHHOOHHOrlnQOHHHHHHrtHOrlO I I \ONh'ONOOl^MOOHOOO^rtHOOrlrtNKlS>OOlt'NM< )MflNON^Ml/)KlHNBO>M/)H OOOOOOOOOOOOOOOOOOHOOOOHOOOOOOOOOOHOOOOOOOHO I I I I I I ............................... II I I I t/)KO*l/)«OO»OrtNOOOOll'lMnOlAnONM/)rtff|iOiOt,'OOM'WIOOOitH0t''S'OIOM/) OOOOOOHOOOHOOOOOOOOOHOOHOOOOOOOOOOOOOOOOOOHO I I I I I I I I I I I I I I I HNM'tWiOSCOClOHNKWUl l'Or'CO(J>OrmKHI,l/l\OI''C0010rlNI< Win'OI'''«OOlOHNKW HHHHHrtHHHrtNNNNNNNNNNMM«rtMrtlOlOlO(OV^^«'t ■ d u P p ■ H E 0 « p f i *rt 0 f t so N H p1 H 0 rl w I H 01 u , on Ifl w< J H 00 < m h UH DiZCC hHPJ r l ZH ti rv wk •H E O 0 HWH H JH W W 1 w *0 J OHH a tin < x w < H hU w z w in H O I <UH J I Z < H * OKH H I S HO <H JW 10 W « A 0 U N H 1 0 0 H 'C O r H C O O 'O t,'*,O H O O H C m iA l/1 0 U l\O r l(]1 0 0 0 t'> » O K IO N i/)M C l* N iO H H O lH * O H < J NNNNt'WNTfMNNNMNMflNNNOMWHNHMNNNMMNNrtNrtWIINWNrtflO UI5lK)NMKlM/)OI000N'O'C'^l/lNHinNin00S®00'Of'J)(O00flllNH01OVI0\OCr'fs^OI/l HHNNKllONtONNHNNNNIONKlHONNHHOIONNNNNHrtNKtHtONNNNN^O \OM3OtK)MNtNNNONI,'NNIOlflN0nfiOHOOl(?lffl'ONNOOMOO>O\OI')OI')Ji*l> 'O * ’t NNKlNWrt'ftOrtKHtKinKl^Ki^HHKllONNHlOMt'lNlOrtNIOnKIN^KlKllONIOnO i- H 'K W N C n O H r lN * < 'O C O ( J > r 'O N K 1 f f iO O O t ''H I ‘ > H f f lH 'O O M N l/) l/ia iH O lO 0 f f ir '- N O t N NHKlNMMN^tONNKlKlNNKlKl^NOniOrtNHrtNMNrtNNNNHHtOKlNNHNlOO NKl^SNNOOWHl/l\Ol/)NWlO'OHOfflHHlO\OlOffl^ONNh-rs ON^^K)OOOtOls QlP'NHO rlNNHKWN^MNNNKlNNKlMTfrlNKlKIHNOrtNNNNNNHNNHtONNNrtNlOO HNHNtOKlNflMtOHtONNIOttNtONONNHNHMNNNHNNNNNHKlNNNHNMO 0(7iaiNOHNffitONNMNOOOOVl/)'OVOlHOONO't’t«0)'tOOOI'OOONOffl'0'OHH<tM NHHHlOtONtONNHMNtONKltftOHONNHNHNNHNHNOrlNNHKlHHNHNMO HNNHNtONNNNONNHNHIOtONONHHNHNrtNNHNONHOHNHHHHNNO OHHHHNHKIHHHNNNNNNNNHHKIONHNHNHHNHHHNHHHHNHHNQ 0l/ieO eO *t^ H iO O M N H l/lM lA N T fO 'O N JlV O 'tfl*H O 'O t'*iO *ft'''O O 0)9N »fN £N * OHOOHNHNNHHHHHHNNNHHHHHHHHOHOOHOOOOHOOOOOOHO t^'O0iMH^IONCO8r*'O'O^OHr''HHHOl/)K)'OiA*O^O&h',<tO'OlOOOHHOt'*'flO'fOlOl/IN ohooohohoohohonnhhhohhohohoooohooonhoooohohh Ha)*<Orl0>ON\D0)ttl/>OO«'OOOeOHVriOl/Wn\O*'ONNlAIOHNNOONUlNNNt'>O«tW HOOOHHOHOHHOOOHrtHHNHHHOHOrtOOHOOOOOOHOOOOOOHO I I I I I I I i I HNKWlfl'Ot'-»CiOHNKHfl/!'ONOOOlOHNI'Wl/)'0l, '#OlOHNKKifl'Ot'COmOHNt'W HHHHHHHrtHHNNNNNNNNN(SlOl1l'>rtlOlOtOMIOIO*<m* 0 •H 6 0 1 0 ( J c •H 0 C l 51 N H H H 0 H w I H O t u . 0 X M [/)< JH B < ( 0 hOH KZW HhW H 2H 0 WK J X O N HW H H JH W W i J 0 J OHh P Q M < XW< H hU «zw Ifl HOI iVH A ' U I ZwH * OHh H2S H< JW • 0 ww a a 0 u N H P 1 / 1 « H Nrt*M<»K)Nrtr'>nK>rt*lOK>IOMIONNnNrtNNK>K>*NNK)NI'INHH«*IOI'INN*0 WOOOHO^BWB^(Jl5l^NlflO'OOl|CNNtfHH\0’tl,)lAf'BH|OClNBBiOfflB^^IABC> NN*K)*HINrtMNNn**M*MMHrtl'lNI<lrtrtrtlOrtrtNnNNWHrt*rtNtONN*H *ONtOOO*0lOHNWi/l'O<»HO'Ol/) 01*<tC)OO*OCOrll/!eOOlOOinNrll/)VHOCOKlffii/!>fiN N im O tO IO N ^ M M K ltO ’ tfll'W tO IO H N ^ N K lN N rtrtlO N N tO N M tO N N ^ tfN K lN N ^ 'H M Q M /lt'^ O H fflM O llfltlH ^ ^ IO flO iO N N rt^ H V O N Sl^ B M fllO O O O B H r'O lN H H NWKllOIOHIO^NlONrt^^MKlMKlNH^N^NN^^lONMIOHNMION^tKlKlrtNlfllrtO 0'ON^f'l/llrtK)h'NOMl/ll/)OOtOi/)OOHOlOOO(ONOH/ 1(ONH^f'CO®COiOOllflOO®lflOOOOO K)N*KWION*K>MlO<»**K!«t**NH«NIOflH**tOtOK!IONNIONr,*t1Ne1Nn*H fl001tflOCO^O>06Q(0'OHO^N^OOO^^lfl'OONLn^tHOl/)K)f>NI/l^f'O O tO O O l/5fO O O fM (v IK I HNMflt'lNHNNNHtOlOtONKlW^NHtOHtOt'IQl'llONNNNNNNNHIOIONtONNlAH ONKlHOIWOWMOOlOlO^Mf'HHIOHr'^NlfllOMOlrtOOftlflMflNHOHH^OllOOlKlt' NKlKlNMNN^NNHtO^^lO^IOtONHKlNKlNNrttONNNKlNNKlKlNKlKlNNMNI/lO i0OMK)iCa0|iflNrtl> >BNBB^iOH^lfl'O^*^l‘WHlOK)B^I''H'O'OH^Bfi'OlOI> 'HH HtOMNNNOIONNNNrtNNrtNKlrlHNNNNNKlNHNHtONNNNrltOKlNNNN^N lflO N O \O B N N K im ^ M fll/I N iO 'tr tO N O 'fl# H fl^ B O lO lO H O 'O M M 'I I O iO N N ^ M O r H O HIONHHHHNrtHOHNNNNNlOHONrlNNHNHOHHKlHHHNNHNHHHHMO iOK|l/lNl/lHLni/5a4CTi'tOOOOtOt^UlNt/1inTtOiO>0(7i^,OlfflvOfflOOt^fOHU>HNKttvONrt\ON HNNHNflHflHHrlNIOMrtNNIOHONHNNONHHOHNHNNHHNHrtNNHtOH Klr l N l / l ( 0BU15lHJlHBl/l^OOHIv 'NBHBBl/l'fKllflf'’ tNBNBOt0NH'0HBt^N'0^ NNNHNNHMNHrlHMNMKlrtNHHlONNNHIONHrtNNHOrlNNNrtNHNOHH HBiAHN*Ht>.HBt"OOI'ONOt'>NClHBOMOMHHIO'0'OOlOHHO®OOMBOHI'lt') OHOOOHOHOOOOOOHHrlrlOOOHHHOrlOOOOHOHOrlOOOOOHONO , I I I I * T ) V * » ' H 6 0 W c • H 0 & t t t t t * I I * * * t t I ♦ I • I * • ' • * * ‘ HNl'Wl/l'ONBaOHNl'WlAOP'BSlOrlNt'Wl/IWNBfflOHNKWlA'Ot'BSlOHNt'W HHHHHHHHHHNNNNNNNNNNtOWfllOMKHOKIIOrt***** TABLE 8 EXTENDED MATRIX: FACTOR LOADINGS OF THE TRIALS ON THE ABILITY FACTORS [-rials CFC CSC CMU CMC CMS MSC MMC DFC DSU DSC DMC NFC NSC NMU NMC Sex h2 1 - IO - 1 5 IO 05 - 06 - O 1 1 1 - 06 11 08 -02 O 7 06 - 03 O 7 -05 IO 2 03 - 2 2 O 1 -07 -02 O 5 1 3 14 23 -Ol -04 -04 06 OO Ol - 11 1 6 3 O 2 - 06 1 2 -02 09 04 1 O 1 5 09 1 5 04 IO 11 - IO O 3 OO 1 3 4 IO -03 06 - 04 - O 1 IO O 8 09 - O 1 19 O 1 O 1 14 -06 O 1 -02 IO Symbo1i c 5 O 8 -02 O 5 - 07 O 7 1 O 14 O 3 1 1 20 O 1 O 2 24 04 1 1 -02 1 7 6 1 7 OO 1 2 OO 2 4 1 2 2 3 14 O 7 1 3 -02 O 1 1 3 O 1 O 3 -07 23 7 1 5 O 3 O 3 -03 1 2 IO 3 1 09 O 7 2 8 - 05 -04 19 O 4 -03 -02 28 Task 8 18 1 5 1 5 - 1 2 2 O O 8 20 O 8 06 19 OO - 1 5 11 O 4 1 5 1 5 31 9 13 O 3 06 08 2 4 IO IO O 3 - O 1 36 -06 O 5 1 7 O 7 O 8 O 2 28 io 1 2 20 O 5 06 2 3 IO 14 O 6 -05 26 - 04 OO 17 06 1 2 O 8 2 7 11 1 3 15 O 2 O 5 20 1 2 2 6 1 5 OO 24 -03 - Ol O 6 IO O 8 OO 2 7 12 -0 2 1 7 06 -0 3 1 7 05 2 5 1 3 19 14 06 - 04 1 5 06 06 14 2 5 1 IO - 09 - IO 09 - 20 06 O 2 1 2 O 2 17 1 5 IO 16 - 22 IO -03 23 2 -05 O 2 Ol O 7 - 14 IO OS 04 14 1 O 09 O 7 1 7 -04 O 1 O 2 11 3 14 OO -06 -02 - 06 13 O 2 O 7 -03 1 4 1 1 20 09 - O 1 06 - O 1 13 4 O 5 03 06 -08 O 3 1 7 20 -03 11 2 7 O 7 19 O 2 04 2 1 09 26 F i gura1 5 26 O 1 -O 1 O 8 -02 2 3 2 5 O 7 O 1 1 1 O 5 1 1 -02 1 4 09 -O 1 25 6 12 - O 1 O 7 O 7 O 3 34 03 O 7 O 3 16 03 O 6 20 16 14 O 8 26 7 24 1 O 04 O 8 -08 2 2 08 O 4 O 8 11 O 5 09 2 1 11 2 2 O 7 2 8 Task 8 13 2 3 O 3 IO 09 24 1 1 - Ol 04 12 2 4 2 2 O 8 - 06 17 1 3 34 9 14 1 5 09 1 1 1 1 2 2 23 12 O 1 1 1 06 12 09 OO 2 7 1 O 3 1 IO 1 2 13 IO 2 3 09 2 7 18 07 16 1 3 14 1 8 09 05 2 2 1 5 39 11 1 5 IO IO 19 O 7 20 1 8 O 1 20 IO O 1 O 3 1 4 O 3 15 2 3 30 1 2 2 1 IO 14 2 O O 8 O 7 19 08 13 19 O 2 2 1 16 O 3 03 O 8 30 1 Ol - 03 04 - 09 O 2 17 O 3 O 3 14 04 13 02 1 2 - 1 5 O 2 O 2 12 2 O 2 26 O 5 - 06 O 3 2 1 16 1 O 14 1 7 2 1 07 18 -05 O 2 -04 29 3 IO 06 O 3 1 2 O 2 1 8 2 4 12 09 16 O 1 05 19 -07 1 1 - 06 22 4 O 5 IO 11 - 04 04 2 1 17 1 5 14 1 2 1 2 - 09 18 06 -02 - 04 21 S emant i c 5 -03 1 4 1 3 16 O 5 1 1 20 06 04 18 O 7 IO 15 1 3 2 1 - 19 29 6 09 IO IO 04 03 1 8 30 1 7 1 4 19 O 3 OO 28 20 20 06 40 7 19 09 26 -02 09 31 2 5 1 5 1 6 O 7 1 2 - 13 1 3 - O 1 18 -08 42 Task 8 O 8 2 3 14 O 7 1 2 32 31 1 8 O 2 1 7 O 5 IO 12 14 2 7 - 04 48 9 06 26 2 1 IO 09 19 20 26 O 8 14 O 5 O 5 O 7 17 2 1 05 39 IO 05 2 5 19 1 5 - O 1 18 28 2 O -O 1 17 1 7 11 11 O 5 23 -04 4 1 11 12 20 1 7 1 3 O 1 23 31 1 2 -05 1 8 O 7 04 12 1 3 2 8 - 08 4 2 12 05 2 3 17 19 O 5 14 36 19 -07 1 3 IO 16 04 04 3 2 - 06 4 7 Note. Decimal points omitted. 53 .40 .35 .30 .25 .20 .15 .10 tn u 5 5 .05 .00 / V - .05 -.10 DSC DMC -.15 DFC -.20 _ DSU -.25 5 6 7 8 9 10 11 12 1 2 3 4 TRIALS Fig. 3.--Factor loadings of the trials for the symbolic-concept task on the divergent-production factors. LOADINGS 54 40 35 30 25 20 15 10 05 00 05 10 NFC NSC 15 NMU -.20 NMC -.25 1 2 6 7 3 5 8 9 10 11 TRIALS Fig. 4.--Factor loadings of the trials for the symbolic-concept task on the convergent-production factors. LOADINGS 55 .40 .35 30 25 20 15 .10 s .05 00 05 CMC 10 CSC CMU - .15 CFC 20 CMS 25 2 3 6 7 8 9 10 5 TRIALS Fig. 5.--Factor loadings of the trials for the symbolic-concept task on the cognition factors. LOADINGS 56 40 35 30 25 .20 15 10 .00 - .05 - .10 - .15 MMC - .20 _ MSC 1 2 5 6 7 8 9 10 11 12 3 4 TRIALS Fig. 6.--Factor loadings of the trials for the symbolic-concept task on the memory factors. LOADINGS 57 40 35 30 25 20 15 .10 05 00 05 DFC - .10 DSU 15 DSC DMC 20 -.25 9 10 11 12 6 7 8 3 4 5 TRIALS Fig. 7.--Factor loadings of the trials for the figural-concept task on the divergent-production factors. LOADINGS 58 .40 J .35 J .30 A 20 A 00 A NMU « NMC NFC NSC .25 - I 1 2 3 4 5 6 7 8 TRIALS Fig. 8.--Factor loadings of the trials for the figural-concept task on the convergent-production factors. LOADINGS 59 .40 25 20 - 15 - 10 - 05 - 00 _ 05 - CFC 10 - CSC CMU 15 - CMC 25 - CMS 25 - I 6 7 8 9 10 11 12 TRIALS Fig. 9.--Factor loadings of the trials for the figural-concept task on the cognition factors. LOADINGS 60 .40 .35 .30 .25 .20 .15 .10 .00 - .05 -.10 MMC - . 20 MSC -.25 5 6 7 8 9 10 11 12 1 2 3 4 TRIALS Fig. 10.--Factor loadings of the trials for the figural-concept task on the memory factors. LOADINGS 61 40 .35 - 30 - /V 00 - DFC DSU 15 - DSC 20 - DMC 25 1 2 3 4 5 6 7 8 9 TRIALS FiR* 11---Factor loadings of the trials for the mantic-concept task on the divergent-production factors. 62 .40 -| .35- .30 - .25 - .20 “ oo u .10 - ► — t ca < o . - a .00 - - .05 “ NFC NSC - .15 ~ NMU NMC - .25 2 3 4 5 6 7 1 8 9 10 11 12 TRIALS Fig. 12.--Factor loadings of the trials for the semantic-concept task on the convergent-production factors. LOADINGS 63 40 .35 30 25 20 .15 // 10 05 00 - .05 CFC CSC - .10 CMU - .15 CMC 20 CMS 25 5 6 7 8 9 1 2 3 4 TRIALS Fig. 13.--Factor loadings of the trials for the semantic-concept task on the cognition factors. LOADINGS 64 .40 .35 - 30 - MSC MMC 25 - I I 2 5 6 7 3 4 8 9 TRIALS Fig. 14.--Factor loadings of the trials for the semantic-concept task on the memory factors. 65 task. This seems especially true when the attributes are as unusual to as having the first three letters of a four- letter group in alphabetical order or beginning with a vowe1 . The relevance of MMC rather than MSC (see Fig. 6) is surprising. It was expected that the abilities with content similar to the task would be the important ones. Memory, however, was expected to play a major role in all of the tasks, and in the figural task this is also con firmed. Here, again (see Fig. 10), we see a crossing of contents as MSC rises to above 30 in the middle trials, and then gradually declines. The significance of memory is underscored in the semantic task (see Fig. 14). Both memory factors are in volved; MSC rises to a peak at trials 7 and 8 and then declines. MMC rises rapidly and maintains relatively high loadings throughout the task. NMC appears to play a significant role in the semantic task (see Fig. 12). It rises gradually over the trials, obtaining a loading of over .30 on the last trial. As pointed out earlier, the task is basically a convergent one, that is, forming a unique classification of the stimu li. As more information is gathered on each trial the task becomes more a problem of classification so this re sult was expected. However, the loadings on convergent production of the trials in the other tasks did not 66 conform to this expectation. The verbalization response for each task was scored by giving one point for each correctly labeled concept. Thus the maximum score for each task was four. The mean and standard deviation for each concept and each task are given in Table 9. The correlation coefficients of the verbalization scores with all the ability tasks were computed. This cor relation matrix is presented in Table 10. Vectors repres enting the verbalization responses for each task were then extended into the factor space of the ability measures. The loadings resulting from this procedure are presented in Table 11. A significant loading .32 of the verbalization responses of the symbolic task was obtained on DSC. A relatively high loading of .40 on CMC resulted for these responses in the figural task. The semantic task repre sented two factors. The verbalization responses for this task had a loading of .41 on CMC and .36 on MMC. These results fail to support the hypotheses that the verbalization responses for all three tasks would have significant loadings on NMU, the concept-naming factor. Both the verbalization responses of the figural and seman tic problems had high loadings on CMC. 67 TABLE 9 MEANS AND STANDARD DEVIATIONS OF THE CONCEPT VERBALIZATION SCORES Task Concept Mean Standard Deviation A. Double letter .78 .42 B. Letter S .21 .41 Symbolic C. Begins with a vowel .11 .31 Task D. Alphabetical order Total .60 1.70 .50 1.06 A. Intersecting lines .67 .47 B. Right angle lines .59 .49 Figural C. Dotted lines .92 .27 Task D. Parallel lines Total .45 2.63 .50 1.20 A. Leader .63 .48 B. Part .51 .50 Semantic C. Animal sounds . 73 .45 Task D. Food Total .77 2.65 .42 1.37 TABLE lO CORRELATION MATRIX OF CONCEPT-VERBALIZATION SCORES AND THE ABILITY TESTS Test Sy mb o1i c Task Figural Task Semantic Task 1 . 32 3 7 30 2 - 2 2 2 7 38 3 . 25 3 7 48 4 . 24 2 7 40 S . 38 54 55 6 . 37 46 47 7 . 34 44 40 8 . 39 51 51 9 - 34 4 2 43 lO . 24 32 43 11. 21 38 41 12 . 33 45 49 13. 47 44 52 14 . 37 41 48 15 . 37 3 7 40 16 . 39 45 45 17 . 25 28 36 18 . 34 37 43 19 . 23 19 15 20 . 1 2 19 23 21 . 40 49 50 22 . 3 7 36 36 23 . 34 2 7 30 24 . 27 29 32 25 . 19 1 7 20 26 . 39 50 49 2 7. 32 40 46 28 . 22 27 37 29 . 34 36 40 30 . 42 48 42 31 . 37 40 43 32 . 21 3 1 31 33 . 32 3 7 50 34 . 39 43 46 35 . 23 39 28 36 . 20 26 2 8 37 . 41 50 54 38 . 2 7 37 48 39 . 30 41 48 40 . 35 43 54 41 . 28 3 1 37 42 . 2 7 4 2 45 43 . 34 42 53 44 . - 04 1 2 - 09 Note. Decimal points omitted. TABLE 11 EXTENDED MATRIX: LOADINGS OF CONCEPT-VERBALIZATION RESPONSES ON THE ABILITY FACTORS Learning Task CFC CSC CMU CMC CMS MSC MMC DFC DSU DSC DMC NFC NSC NMU NMC Sex h2 Symbolic Task 18 23 08 18 13 12 10 12 04 32 05 08 16 10 12 08 35 Figural Task 20 18 05 40 20 15 09 15 13 18 13 15 14 10 19 18 52 Semantic Task 17 28 23 41 03 17 36 08 07 13 16 19 07 13 24 04 66 Note.--Decimal points omitted. O' to CHAPTER V DISCUSSION The Abilities The attempt to lend empirical support to the do main of the SI model that is concerned with classes was successful. Eleven of the 20 classes abilities depicted by the SI model were investigated and were identified in this study. Thirty-six tests were employed to identify these eleven factors. Thirty-four of these tests have loadings consistent with their hypothesized factor con tent; only two do not. The divergent-production-of-classes factors, DFC, DSC, and DMC, were defined entirely by the nine tests de signed to measure these three abilities. Seven of these tests were univocal, at least two for each of the three factors. DFC and DSC had previously been adequately mea sured by only two tests. The addition of Multiple Group ing of Figures for DFC and Multiple Grouping of Nonsense Words for DSC buttresses evidences for these two factors. Ten of the 11 tests designed to measure the con vergent production of classes have loadings on their hypothesized factors. Restricted Figural Classifications, 70 71 intended as a measure of NFC, did not load significantly on any factors in the analysis. NFC and NSC had not been identified previous to this study. NMC was previously identified (Merrifield, et al., 1960) but by only one good test. The addition of the new tests, Group Classification, Concept Grouping, and Largest Class, substantially strengthens support for this factor. Two of the NMC tests, two for NFC, and three for NSC were univocal in this analysis. MSC abd MMC are defined entirely by the tests hy pothesized to measure them. Each of these factors had been identified once, MMC by Brown, Guilford, § Hoepfner (1966) and MSC by Tenopyr, Guilford, § Hoepfner (1966). This analysis confirms their findings with the exception that one of Brown's MSC tests, Classified Information, had an additional significant loading on CMC. The three previously established cognition-of- classes factors, CFC, CSC, and CMC, were again identified. Nine of the ten tests hypothesized to measure the cognition-of-classes abilities perform as expected. Letter-Group Exclusion, hypothesized to measure CSC, loaded univocally on NSC. Two of the CFC tests, two of the CMC tests, and three of the CSC tests were univocal. CFC and CSC are given additional support in this study by the in clusion in the analysis of two newly developed tests. Figural Class Inclusion had a univocal loading on CFC; 72 Letter Classification loaded univocally on CSC. The four reference factors outside the classes do main, CMU, CMS, DSU, and NMU, were readily identified. Only one classes test has a significant loading on any of these factors. Concept Grouping has a substantial loading of .49 on CMU. The effort to keep the vocabulary employed in the classes tests at a level that would diminish mea surement of individual differences on CMU in the sample seems, for the most part, to have been successful. Appar ently this effort failed in the construction of Concept Grouping. None of the classes tests loaded significantly on the factor defined as Sex. Brown, et al. (1966) found a significant relationship between Sex and MMC. The results of their findings were not supported by this study. The separation of factors with respect to the type of test content defining them is a commonplace finding in the history of factor analysis of mental tests. However, in this study a stringent test of this result occurred. Several of the classes tests employed have identical prop erties except for the type of content. None of these tests have loadings on factors defined by tests with dif ferent contents. In fact, the significant loadings on each factor in this analysis are completely homogeneous with respect to content. Although the results of this study clearly favor 73 the SI hypotheses, two problems were encountered. First, the separation of the cognition factors from the produc tion factors was difficult. The factors were meaningfully separated, but there were a number of tests common to both types of factors. Some difficulties with the separation of cognition abilities from abilities in other operations have been encountered in other investigations of various domains of the SI model. Gershon, et al. (1963) found this to be true with respect to divergent production of figural products; Nihira, et al. (1964), with respect to the evalu ation of symbolic products. Although this could mean that this operation could be basic to the other operations in the model, its occurrence in tests designed to measure other operations could be simply due to faulty test con struction. In this study it is impossible to ascertain whether or not the tests that have significant loadings on the cognition factors were simply not adequately construc ted to control for variation due to cognition. Second, the tests that utilize an exclusion type of format do not seem to be the univocal measures of cog nition they were thought to be (Guilford § Merrifield, 1960; Guilford 6 Hoepfner, 1963). The figural and symbolic exclusion tests, Figure Exclusion and Letter-Group Exclu sion, loaded substantially on the convergent-production factors. Possibly this is due to the fact that a class has to be formed in order to exclude the member that does 74 not belong. The failure of. the semantic exclusion test, Word Classification, to load on NMC, however, does not support this conjecture. However, the classes in Word Classification are more easily recognized, while the at tributes utilized in Figure Exclusion and Letter-Group Exclusion are frequently not familiar to S. The Concept-Learning Tasks The relations of the abilities to the concept- learning tasks were reasonable, but certainly not dramatic. Aside from the memory abilities only two factors were of substantial importance, DSC in the symbolic task and NMC in the semantic task. Both of these results lacked sup port in the two parallel tasks with different contents in each case. Still, their significance should not be ignored. Both DSC and NMC have significant trends in tasks consistent with their content. In addition, their overall trends conform to the expectations from a content analysis of the tasks. As mentioned earlier, the tasks are basi cally convergent, that is, the formation of a unique classification into four mutually exclusive classes. As more information is gathered on each trial the task be comes more a problem of classification, and, consequently, convergent production was expected to have more importance as the task progressed. The loadings of the trials in the 75 semantic task on NMC conforms to this notion. However, since this trend is not evident in the figural and symbolic tasks, this result is at best open to some question. It was assumed that to arrive at a correct attri bute S would test hypotheses about what attributes were relevant to a certain letter (class label). These hy potheses would be formed in two ways. First, he would produce possible classes, in which the stimulus presented to him could belong. Second, S would remember stimuli associated with a particular letter and would produce at tributes in common with those stimuli and the stimulus he was confronted with at the time. Both of these processes can be subsumed under the concept of divergent production of classes. The fact that the divergent production of classes was found to be relevant only with the symbolic task can be partially explained. The correct attributes in the semantic task, leader, animal sounds, food, and part, are commonly encountered. Possibly the obviousness of these attributes reduced the significance of individual differ ences with regard to producing semantic attributes or classes. This could also be true with the figural task, in which the correct attributes were right angle, inter secting lines, dotted line, and parallel lines. This ex planation seems more plausible when we recall the attri butes of the symbolic task. Words with the letter S, 76 words with two letters the same, words beginning with a vowel, and words with the first three letters in alpha betic order are attributes which are probably infrequently encountered or utilized by S. The importance of memory in all the tasks is cer tainly underscored. One of the memory abilities involved in this study is relevant in both the figural and symbolic tasks, and both memory abilities show a relationship to scores from the semantic task. There is, however, a curious crossing of content. MMC is relevant to perfor mance on the symbolic task and MSC seems to play a role in connection with the figural task. If Ss verbalized the attributes as they went from trial to trial, the signifi cance of MMC in the symbolic task would make sense. How ever, the presence of MSC in connection with the figural task does not seem interpretable, although it should be pointed out that a memory-for-figural-classes ability was not represented in this analysis. Substantial loadings on MSC also occurred in the semantic task where the relevance of MMC was also clearly demonstrable. The importance of memory is certainly no surprise. The significant consideration to the present author is that if we can measure memory or various kinds of memory as those depicted by the SI model, and if we can demon strate that there is a relationship between memory and concept learning, as in this study, then it should be 77 possible to incorporate the individual-differences measure of memory into our theories and investigations of concept learning. The verbalization responses obtained at the end of each of the three learning tasks did not load on the conceptual-naming factor, NMU, as expected. Two of the loadings of the verbalization tasks were on factors which were relevant to the trial scores for those same tasks. DSC is represented by substantial loadings of both the ex tended trial scores and the verbalization scores for the symbolic task. The same principle is also true of the semantic task with respect to MMC. These results, if con sidered alone, would imply that the abilities involved in the concept-learning process are the same, whether learn ing is measured by the number of correct classifications of the elements or by verbalization of the concepts at the end of the task. However, the high loadings of the fig ural and semantic verbalization scores on CMC seem to imply that verbalizing the concept demands an additional ability. That is, the cognition of a meaningful attribute among the stimuli is necessary to verbalize the concept. The failure of the verbalization scores for the symbolic task to load on CMC weakens this interpretation. Relation to Other Studies Due to the differences in type of concept-learning 78 tasks employed, it is difficult to compare the results of this study with those from many other investigations. Two previous investigations (Allison, 1960; Manley, 1965) are relevant by intent. Utilizing an interbattery method of factor analysis, Allison found a factor which he called conceptual learning. This factor was primarily defined by parameters taken from two tasks, one semantic, the other figural, similar to those used in this study. This factor was related to most of the ability reference factors in cluded in the study. He concluded that the unifying characteristic of the factors correlating with the con- ceptual-learning factor was that the process of concep tualization or thinking was required. In this study no general overall relation was found between the mental abil ities and the concept-learning tasks. Furthermore, the interbattery factor which correlated the highest with his conceptual-learning factor was a factor called intellec tual ability. This factor was defined primarily by verbal tests such as a vocabulary test, a sentence-completion test, and a test consisting of both verbal analogies and sentence completion. The factor most similar to intellec tual ability, in this study, was CMU. CMU was not found to be related to any of the tasks utilized in this study. Although the conceptual-learning factor was led by parameters from concept-learning tasks, the factor had several significant loadings from a variety of other 79 learning tasks. The complexity of this factor could ac count for the discrepancies between the results of this study and Allison's. As mentioned earlier, Duncanson (1964) favored a similar interpretation when accounting for his lack of establishing any significant relationships between intellectual abilities and the concept-learning task he employed. The learning parameter indicating whether a sub ject was learning faster during the first or second half of the trials for the figural-concept task led a small factor with a loading of .36. Allison did not interpret this factor, but it is interesting to note that its high est correlation with any of the ability factors was .30 (with rote memory). In this study MSC was the dominant ability in the figural task. Although a memory ability, it is not to be identified with Allison's "rote memory." In a recent investigation, Manley (1965) used the verbal-concept task developed by Allison. Unfortunately he treated the error scores for the four concepts sepa rately as independent measures in the factor analysis. The scores were necessarily interdependent. If one con cept is learned, the probability of making an error on the other three is reduced, and, in fact, when three are learned the probability of making an error on the fourth concept should be zero. These four concepts, of course, intercorrelated very highly and defined a very strong 80 factor, which was unrelated to the ability measures. How ever, examination of the correlation matrix shows the con cepts to correlate substantially with a number of reason ing tests. A reasoning factor, CMS, was defined in the present study, but did not show any significant relation ship with scores from the concept-learning tasks. CHAPTER VI SUMMARY AND CONCLUSIONS Current theories of concept learning have focussed on the processes intervening between the stimulus and the response. Although intellectual abilities have been mea sured for decades, their relevance to these processes have been largely ignored. The significance of intellectual abilities has recently been emphasized by a theory of in telligence (Guilford, 1967), in which an entire domain of abilities concerning concepts is postulated. The present study attempted to provide an empirical foundation for these abilities and to investigate their relationship to performance on concept-learning tasks. Forty-three mental aptitude tests were employed to measure fifteen abilities postulated by the structure-of- intellect theory. A factor analysis of these tests iden tified all of the hypothesized factors. Four of these were reference factors and eleven were factors measuring abilities involving classes. The factors separated with regard to the type of content: symbolic, figural, and semantic; and with respect to the operation: cognition, convergent production, divergent production, and memory. 81 82 Three concept-learning tasks, utilizing the same three types of content as the tests, were administered. Vectors representing the number of correct responses for each trial and the number of correctly verbalized concepts for each task were extended into the factor matrix. Although the relationship of the abilities to the concept-learning tasks was short of expectations, the re sults clearly indicate that certain abilities are relevant to certain tasks. DSC and MMC were found to have signifi cant relationships at various stages in the symbolic- concept-learning task; MSC in the figural task; and MMC, MSC, and NMC in the semantic task. An additional ability, CMC, was found to have a strong relationship with the scores of the verbalizations of the concepts for the seman tic and figural tasks. The results of this study are different from those of the two previous studies (Allison, 1960; Manley, 1965) that used tasks similar to those employed in this investi gation. However, those studies utilized somewhat different procedures of analysis and different ability tests; con sequently, it is difficult to make direct comparisons. It is felt that the tests and procedures used in the present study are the more appropriate for this type of investiga tion. The major tenet of this investigation was that an individual performs intellectual processes to learn a 83 concept. Thus a comprehensive theory of concept learning would require that these processes be identified and mea sured. The relationships between intellectual abilities and concept learning established in this study offer the possibility that the processes involved in a concept- learning task can be subsumed under intellectual abilities measured by means of individual differences. This study is at best only a beginning. But, by utilizing the procedures of this study, further investiga tions could ascertain a more precise description of the intellectual abilities involved in concept learning. Then by examination of the functional relationships between the measures of the abilities and the variables now utilized in theories and investigations of concept learning, it seems evident that we could only improve our conceptions and theories concerning this elusive phenomenon. REFERENCES 84 85 REFERENCES Allison, R. B. Learning parameters and human abilities. ONR Technical Report. Princeton, N. J.: Educa tional Testing Service, 1960. Bourne, L. E., Jr., 8 Restle, F. Mathematical theory of concept identification. Psychological Review, 1959, 66, 278-296. Brown, S. W., Guilford, J. P., 8 Hoepfner, R. A factor analysis of semantic-memory abilities. Reports from the Psychological Laboratory, No. 37. Los Angeles: University of Southern California, 1966. Bruner, J. S., Goodnow, J. J., 6 Austin, G. A. A study of thinking. New York: Wiley, 1956. Bunderson, V. C. Transfer of mental abilities at differ ent stages of practice in the solution of concept problems. Unpublished doctoral dissertation, Princeton University, Princeton, N. J., 1965. Cliff, N. Orthogonal rotation to congruence. Psychomet- rika, 1966, 31, 33-42. Dixon, W. J. Biomedical Computer Programs. Los Angeles: Health Science Computing Facility, University of California, Los Angeles, 1965. Duncanson, J. P. Intelligence and the ability to learn. 86 Research Bulletin (RB-64-29). Princeton, N. J.: Educational Testing Service, 1964. Dwyer, P. S. The determination of the factor loadings of a given test from the known factor loadings of other tests. Psychometrika, 1937, 2^ 173-178. Ferguson, G. A. On learning and human ability. Canadian Journal of Psychology, 1954, l[, 95-114. Ferguson, G. A. On transfer and the abilities of man. Canadian Journal of Psychology, 1956, 10^, 121-131. Fleishman, E. A. Abilities at different stages of prac tice in rotary pursuit performance. Journal of Experimental Psychology, 1960, 60, 162-171. Fleishman, E. A., 6 Hempel, W. E., Jr. Changes in factor structure of a complex psychomotor test as a function of practice. Psychometrika, 1954, 19, 239-252. Fleishman, E. A., 6 Hempel, W. E., Jr. The relation be tween abilities and improvement with practice in a visual discrimination reaction task. Journal of Experimental Psychology, 1955, 4£, 301-312. Games, P. A. A factorial analysis of verbal learning tasks. Journal of Experimental Psychology, 1962, 63, 1-11. Gershon, A., Guilford, J. P., 6 Merrifield, P. R. Figural and symbolic divergent-production abilities in adolescent and adult populations. Reports from 87 the Psychological Laboratory, No. 29. Los Angeles: University of Southern California, 1963. Guilford, J. P. The structure of intellect. Psychological Bulletin, 19S6, 53, 267-293. Guilford, J. P. Fundamental statistics in psychology and education, 4th ed. New York: McGraw-Hill, 1965. Guilford, J. P. A revised structure of intellect. Re ports from the Psychological Laboratory, No. 19. Los Angeles: University of Southern California, 1957. Guilford, J. P. Three faces of intellect. American Psy chologist , 1959, ,14, 469-479. Guilford, J. P. An informational view of mind. Journal of Psychological Research, 1962, 6^, 1-10. Guilford, J. P. The nature of human intelligence. New York: McGraw-Hill, 1967. (In press) Guilford, J. P., Berger, R. M., § Christensen, P. R. A factor-analytic study of planning. II. Admini stration of tests and analysis of results. Re ports from the Psychological Laboratory, No. 12. Los Angeles: University of Southern California, 1955. Guilford, J. P., § Christensen, P. R. A factor-analytic study of verbal fluency. Reports from the Psycho logical Laboratory, No. 17. Los Angeles: Univer sity of Southern California, 1956. 88 Guilford, J. P., Frick, J. W., Christensen, P. R., § Merrifield, P. R. A factor-analytic study of flexibility in thinking. Reports from the Psycho logical Laboratory, No. 18. Los Angeles: Univer sity of Southern California, 1957. Guilford, J. P., Green, R. F., Christensen, P. R., Hertzka, A. F., § Kettner, N. W. A factor-analytic study of Navy reasoning tests with the Air Force Aircrew Classification Battery. Reports from the Psychological Laboratory, No. 6. Los Angeles: University of Southern California, 1952. Guilford, J. P., Hertzka, A. F., § Christensen, P. R. A factor-analytic study of evaluative abilities. II. Administration of tests and analysis of re sults. Reports from the Psychological Laboratory, No. 9. Los Angeles: University of Southern Cali fornia, 1953. Guilford, J. P., § Hoepfner, R. Current summary of structure-of-intellect factors and suggested tests. Reports from the Psychological Laboratory, No. 30. Los Angeles: University of Southern California, 1963. Guilford, J. P., 8 Hoepfner, R. Structure-of-intellect tests and factors. Reports from the Psychological Laboratory, No. 36. Los Angeles: University of Southern California, 1966. 89 Guilford, J. P., Kettner, N. W., § Christensen, P. R. A factor-analytic investigation of the factor called general reasoning. Reports from the Psychological Laboratory, No. 14. Los Angeles: University of Southern California, 1955. Guilford, J. P., Kettner, N. W., 8 Christensen, P. R. A factor-analytic study across the domains of rea soning, creativity, and evaluation. II. Admini stration of tests and analysis of results. Re ports from the Psychological Laboratory, No. 16. Los Angeles: University of Southern California, 1956. Guilford, J. P., § Merrifield, P. R. The structure of in tellect model: its uses and implications. Re ports from the Psychological Laboratory, No. 24. Los Angeles: University of Southern California, 1960. Guilford, J. P., Merrifield, P. R., Christensen, P. R., § Frick, J. W. An investigation of symbolic factors of cognition and convergent production. Reports from the Psychological Laboratory, No. 23. Los Angeles: University of Southern California, 1960. Guilford, J. P., Merrifield, P. R., § Cox, Anna B. Crea tive thinking in children at the junior high school levels. Reports from the Psychological Laboratory, No. 26. Los Angeles: University of 90 Southern California, 1961. Guilford, J. P., Wilson, R. C., § Christensen, P. R. A factor-analytic study of creative thinking. II. Administration of tests and analysis of results. Reports from the Psychological Laboratory, No. 8. Los Angeles: University of Southern California, 1952 . Hoepfner, R., 8 Guilford, J. P. Figural, symbolic, and semantic factors of creative potential in ninth- grade students. Reports from the Psychological Laboratory, No. 35. Los Angeles: University of Southern California, 1965. Hoepfner, R., Guilford, J. P., § Merrifield, P. R. A fac tor analysis of the symbolic-evaluation abilities. Reports from the Psychological Laboratory, No. 33. Los Angeles: University of Southern California, 1964. Hunt, E. B. Concept learning. New York: Wiley, 1962. Kendler, T. S. Concept formation. Annual Review of Psy chology , 1961, 12^, 447-472. Kendler, H. H., § D'Amato, M. F. A comparison of reversal shifts and non-reversal shifts in human concept formation behavior. Journal of Experimental Psy chology, 1955, 49, 165-174. Manley, M. B. A factor analytic study of three types of concept attainment tasks. Research Bulletin 91 (RB-65-31), Princeton, N. J.: Educational Test ing Service, 1965. Marks, A., Guilford, J. P., § Merrifield, P. R, A study of military leadership in relation to selected in tellectual factors. Reports from the Psychologi cal Laboratory, No. 21. Los Angeles: University of Southern California, 1959. Merrifield, P. R., Guilford, J. P., Christensen, P. R., § Frick, J. W. A factor-analytic study of problem solving abilities. Reports from the Psychological Laboratory, No. 22. Los Angeles: University of Southern California, 1960. Mosier, C. I. A note on Dwyer: the determination of the factor loadings of a given test. Psychometrika, 1938, 3, 297-299. Nihira, K., Guilford, J. P., Hoepfner, R., $ Merrifield, P. R. A factor analysis of the semantic- evaluation abilities. Reports from the Psycholo gical Laboratory, No. 32. Los Angeles: Univer sity of Southern California, 1964. O'Sullivan, Maureen, Guilford, J. P., § deMille, R. The measurement of social intelligence. Reports from the Psychological Laboratory, No. 34. Los Angeles: University of Southern California, 1965. Petersen, H., Guilford, J. P., Hoepfner, R., § Merrifield, P. R. Determination of "structure-of-intellect" Stake, Tenopyr Tilton, Tucker, Tucker, 92 abilities involved in ninth-grade algebra and general mathematics. Reports from the Psychologi cal Laboratory, No. 31. Los Angeles: University of Southern California, 1963. .. E. Learning parameters, aptitudes, and achieve ments. Psychometric Monograph, 1961, No. 9. Mary, Guilford, J. P., § Hoepfner, R. A factor analysis of symbolic memory abilities. Reports from the Psychological Laboratory, No. 38. Los Angeles: University of Southern California, 1966. J. W. Intelligence test scores as indicative of ability to learn. Educational and Psychological Measurement, 1949, 9, 291-296. L. R. Determination of parameters of functional relation by factor analysis. Psychometrika, 1958, 23, 19-23. L. R. Determination of generalized learning curves by factor analysis. ONR Technical Report. Princeton, N. J.: Educational Testing Service, 1960 . APPENDIX A DESCRIPTION OF TESTS 93 APPENDIX DESCRIPTION O F TESTS 1. Alternate Letter Groups - DFC03B. 1 Find letters of the alphabet that belong to a class because of a communality of shape or form, Possible groups: Sample: Given AH VTC Score: Number of acceptable responses. Parts; 2; items per part: 2; working time: 6 minutes. (a) A H V T (all letters made of straight lines) <b) A H T (all letters have horizontal lines ) 2. Alternate Uses - DlVlC 0 3C ( SSC ). List a s many as six uses for an object, othe r than thee o m m o n use which is stated. Sample: Given: A N E W S P AP ER , (used for reading). Uses: Score: Number of possible, different uses listed. Parts: 2; items per part: 3; working Time: 8 minutes, 3. Classified Information - M M.C 01A. Recognize classes of words s imi Lar to tho given on a previously studiedpage. Sample study item: SILK Sample test items. W O O L C O T T O N ICE N Y L O N F E L T SLE ET Answers: Yes, No. R A Y O N C OT TON FE LT S N O W ICE SLEET Score: Number of correct responses minus the number of wrong responses. Parts: 2; items per study page: 15; items per test page: 30; working time : 6 minutes. attribute of each class so formed is also an attribute of the target word. Sample: 1. tar 4. log 7. gasoline C Q> 2. silver 5. ink 8. gold 3. raven 6. copper 9. kerosene Score: Number of correct classes. Parts: 2; items per part; 3; working time: 6 minutes. 5. Figural Class Inclusion - C FC 04 A. Give n twofigures thathave a common figural pr ope rty, select fr o m five alter - natives the one figure that contains the same property. the W'ords into cla s ses so that the C la s s I: / ©T Class II: 4- r 9 Class III: 6 S GIVEN FIGURES A L T E R N A T I V E S 1 . Sample : 4 < Vi CA D X f — \ A C 2 ) B - ©[ > 0 D E V Answers: 1. D , 2. E. Score: Number oj correct responses* minus n ne - l ■ . >u r t h The number ->? wr.-n^ re spor.se - . . V ■ it *r a i H i * • r a r t . a , < , r » . s -i r N K : A } : 1 It . I i r v * • s i . Parts: 2; items per part: 12; working time: 8 minutes. m ha«ed juon c o m m o n propert i*• » i„___ > t; 4 . ____ _ • — - . i . . * • * - 5 1 \r - APPENDIX DESCRIPTION O F TESTS 1. Alternate Letter Groups - DFC03B. Find letters of the alphabet that belong to a class because of a communa lit y of shape or form. Sample: Given AH VTC Score: Number of acceptable responses. Parts: 2; items per part: 2; working time: 6 minutes. asibie groups: (a ) A H V T (all lettersmadeof straight line s ) {b) A H T (all letters have horizontal lines ) 2. Alternate Uses - D M C 0 3 C (SSC ). Li s t a s many as six uses for an obje ct, othe r thanthe c o m m o n use which is stated. Sample: Given: A N E W S P A P E R . (used for reading). Uses: b ^ _____________________________ Score: Number of possible, different uses listed. Parts: 2; items per part: 3; working time: 8 minutes. -3. Classified Information - M M C 01 A. Recognize classes of words similar to tho^e given on a previou sly studiedpage. Sample test items. Sample study item: SILK W O O L NY LON R A Y O N C O T T O N F E L T SNO W ICE S LE ET Answers : Yes, No. Score: Number of correct responses minus the number of wrong responses. Parts: 2; items per study page: 15; items per test page: 30; working time: 6 minutes. 4. Concept Grouping - NMC02A. Given a target word and a list of words, classify the words into classes so that the attribute of each class so formed is also an attribute of the target word. Sample: 1. tar 4. log 7. gasoline 2. silver 5. ink 8. gold 3. raven 6. copper 9. kerosene Score: Number of correct classes. Parts: 2; items per part: 3; working time: 6 minutes. COAL Class I: Class II: Class III: / . . A r ■ 9 A 6 S 5. Figural Class Inclusion - CFC04A. Given twofigures that have a common figural property, select from five alter - natives the one figure that contains the same property. GIVEN FIGURES A L T E R N A TIVES Sample : 2 . > ® ixi AA V ' i D E h~\ Ae 6 / o \ E V Answers: 1, D; 2, E. Score: Number of correct responses minus one-fourth the number of wrong responses. Parts: 2; items per part: 12; working time: 8 minutes. 6. Figural Hierarchical Grouping - NFC02A. Placefigures into a hierarchical system based uponcommon properties. Sample: 1 A 5 □ A V (D-KD Score: Number of correctly classified figures. Parts: 2; items per part: 2; working time: 10 minutes. The code imme diate ly following eachtest name indicates the hypothesized factor content of the test at the stage of teste on st ruction. Additional codes areas fo llow s : SSC - copyright by She r idan Supply Co. , Beverly Hills , California, adapted wit h pe rmis si on; L L T - adapte d wit h pe r mis s ion from a te st by L. L. Thur stone ; U N C - adapte d with permis si on from a test developed at the University of North Carolina. 1 7. Figure Classification - CFC01A. Recognize classes of figure S-a m pie i t e m s : then assign given figures to the classes. □ A L T E R N A T I V E S B. C. E. Answers: 1, C; Z, A. Sc or e : Numbe r of correct responses minus one-fourth the number of in correct re spons e s. Pa rt s : 2 ; ite ms pe r part: 10; working time: 8 minutes. 8. FiRure-Concept Grouping - NFC03A. Given a target figure and a group of figure s , classify the figures intoclasses so that the attribute of each class formed is also an attribute of the target figure. mple 3 1 4 6 8 T A R G E T Class I: C la s s II: Class III: 9 s.,6 ,T 3,4- 8 Score: Number of correct c las ses. Parts: 2; items per part: 3; working time: 8 minutes . 9. Figure Exclusion which is different. CFC03A. Given five figures. Sample : show that you see what four have in common by excluding the one An s we r : D. Score: Number of correct re sponse s minus one-fourth the number of wrong responses. Parts: 2; items per part: 14; working time: 9 minutes. 10. Figure Grouping - NFC01A. Given a set of twelve figures, group the figures into four distinct classes using each figure once. Samp l c » © ‘ O 10 l i Z P ' e y 12 Class I: 9 Class II: 7 , * 2 - C las s III: 3, $ . / o Class IV: 4- 6 / / Score: Number of figures correctly classified. Parts: 2; items per part: 3; working time: 8 minutes. Sa mp 1 e : T A R G E T A bargain store T A R G E T B Class B / 3 In T rad io s extant oc e an fear 5 sugar s c r atch like c at t g r oup s of words. s se s. 2 c loth consent tear ice scale plastic cheap re cord hate tool bumpe r c lock fruler^ bite dog thimb le cook sew gauge button knife captain agree cut 8 savings can approve bank Score: Number of correctly classified word groups. Parts: 2; items per part: 2; working time: 10 minutes 1 2. Largest Class - N M C 04A. Given a list of words, formthe largest possible class of words with the remaining words also making a class. Sample : 1. button 4. zippe r 7. mai Ibox Sc ore: Numbe r of correct classes. 2. staple 5. filing cabinet 8. scotch tape Fa r t s : 2; items per part: 4; working time : 8 3. pur se 6. paperc lip 9. pocket minutes. 1 3. Lette r Classification - CSC06A. Recognize classes of nonsense words, then assign given nonsense words to the c la s se s. Sample : Items A It e r nat i ve s 1. A L F OSTE IB M R A. LSUC 2. C F C O A Q O Q H C H Y B. C. W A W O D X T E An s we r s : 1. D; 2, B. D. O E M A Score: Number of correct responses minus one-fourth the number of wrong responses. E. zsu Parts: 2; items per part: 10; working time: 8 minutes. 1 4. Le tie r - G one ept Grouping - NSC 0 2A, Given a li st of n on s en se words and a target nonsense word, classify the words into classes so that the attribute of each class so formed is also an attribute of the target word. Sample: 1. A M K T B L E T 2. SBN 3. T F T 4. QIF C la s s I: / ^ 5. BYS Class II: % Class III: 2 . ^ 7 6. G H H 7. R D B 8. LLS Score: Number of correct classes. 9. C V O Parts: 2; items per part: 3; working time: 8 minutes . 1 5. Letter - Group Ex c lusion - C SC 01B ( L L T ). Choose the group of letters that is different from t he other three g roups . Sample: (1) (2) Answer: 3. A A B C A C A D , . . Sc or e : Numbe r of correct responses minus one-third the (3) <4) ^ number of incorrect responses. A C S H A A C G Parts: 2; items per part: 20; working time: 8 minutes 16. Letter Grouping - NSC01A. Given a listof nonsensev/ords, groupthem into four classes u s ing each word only once . Samp le: 1. L X D 7. G O G 8. L Z Q 9. O F Q E E B RIR Class I: / T / o 3. Class II: <? 4. B C D lO. M A A 1 1. SUS 12. L W F K U RST Class III: 7 6. Class IV: -'T V / / Sc ore : Number of correctly classified nonsense words. Fa rt s : 2 ; items per part: 3; wo rking time: 9minutes. 1 7. N-le m o r y for Nonsense Word Classes - MSC02B. Indicate whichof four nonsense words given in each i t e m on a test page represents a class given on a previous study pace. Sample Test Items: Sample Study Items: N E C G U Z 1. 1) GIS N E P G A Z 2) G O Z N E F G Y Z 3) LO Z 4) M O Z Answers: 1, 2; 2. 3. 2. 1) N O P 2) N A R Score: Number of correct responses minus one-third the number of wrong responses. 3) N E R Parts: i; items per study page: 10; items per test page: 10; working time: 4 1 /2 minutes. 4) N U P 18. Memo ry for Word Classes - MSC04A. Indicate whether or not each of a number of words presented on a test page represents a class given on a previous study page. Sample Study Items: pan test Sample Test Items: 1. west 2. boat 3. can ran pest fan lest Answers: 1, Yes; 2, No; 3, Yes. Score: Number of correct responses minus number of wrong responses. Parts: 2; items per study page: 10; items per test page: 20; working time: 9 minutes. to C r » 19. Multiple- Figural Similantieti - DFC07A. Given a set of three figural objects that car* be conceived as representing different classes, select single figures that can be classified with the set. each for a different reason. Sa m p 31 GIVEN CLASS This class A LT E R N A TIVES 2 Score: Number of correctly chosen alternative s. Parts: 2; items per part: 3 ; •working time: 8 minutes. Multiple Grouping - DMCOZC. Arrange given words into several different meaningful groups. Sample : 1. a r r ow Class A : / 2 S 7 2. 3. be e c rocodi 1 e Class B : 3 ( o 4. 5. fish kite Class C : 2 3 4- 7 Score: Number of acceptable classes . 6. 7. sailboat spa r r ow C la s s D. 3 4-3 7 Parts; 2; items per part: 1; time: 4 minutes. working 2 1 . Multiple Grouping of Figures - - DFC08A. Given a number of figures. group and reg roup the m into as many diffe rent classes as possible. Sample : 1 2 © J r ) 4 s < c ‘ Q © C LASSES /, 3 , (n_ _ A 3 © Score: Number of acceptable different classes. Parts: 2; items per part: 1; working time: 8 minutes. 22. Multiple Grouping of Nonsense Words - DSC05A. Given a list of nonsense words, form as many different classes a s po ssible Sample : Nonsense Word List C la s s e s 1. R U A T W S 2. FJOSUX 3. E J L O R U 4. A A K N P B 5. BOOQIC 6. HIOST V / S Score: Number of correct classes. Parts: 2; items per part: 1; working time: 8 minutes. 23. Multiple Letter Similarities - DSC04A. Given a set of three groups of letters that can be conceived as representing different classes, specify alternative groups of letters that can be classified with the set for different reasons. This class is like Alternatives: <5~ -h ( n __________________ Sa mple : Given Class Alternatives IPEC 1 . FOQI 2. Z H E M 3. LAO 4. MKI CA 5. E I M C K 6. IJUME 7. N W R O 8. GOINU Score: Number of correctly chosen alternatives. Part s : 2; items per part: working time: 8 minutes 24. Name Grouping - DSC 023. Classify a group of c o m m o n na me s into several g roups ba s e d upon the diff e r ent alphabet ic properties they have in common. Sample: Classes 1. G E R T R U D E 2. BILL 3. A L E X 4. CARRIE 5. B E L L E 6. D O N / 3 - 4 - 2 4-S- Score: Number of acceptable classes , Parts: 2; items per part: 1; working time: 6 minutes. 25. Naming Meaningful Trends - N M U 0 4 A (UNC). Recognize and express a trend in a group of words. Sample: mouse rat lion pig c u m horse elephant f - ~ . - Score: Number of correctly specified trends . f i„ £ A_____ ?_____________ Parts: 1; items: 10; working time: 3 minutes . 26. Number Classification - CSC03C. Recognize classes of three numbers, then a s sign given number s to the classes . Sample: 1. 44 55 33 Alternatives 2. 10 45 15 A. 42 1 B. 5 3 C. 219 Score: Number of items right minus one-fourth of the P . 2 2 number wrong. E. 25 Parts: 2; items per part: 10; working time: 6 minutes . Answers: 1, D ; 2, E. 27. Number-Group Naming - CSC05B. State what it is that three given numbers have in common. Sample s : 35 6 76 1 10 75 6 5 16 1 Score: Number of correctly named groups. Parts: 1; items: 12; working time: 3 minutes 28. Picture Class Memory - M1V1C03B. Indicate whether or not a giver two-element class represents the same concept as one given on a previously studied page. Sample Study Item: Sample Test Items ' / • f \ \ Ss/ c— ) Answers: Yes, No Score: Number of correct responses minus the number of wrong responses. Parts: 1; items per study page: 11; items per test page: 22; working time: 3 minutes. Pi c tu r e - Gr oup Naming - N M U 0 3 A (UNC). Pr Sample : Score: Numbe r of correct names . Parts: 1; items: 9; working time : 2 m i n ut e s . 30. Problem Solving - CMS05A, Solve verbally stated arithmetic problems where the numerical calculations are minimi zed. Simple: A ship can cruise front L to S before its fuel supply is exhausted. To what A. Between N and O point could it cruise and return with B. To exactly O the same amount of fuel? C. Between O and P P. To exactly P Answer: C. E. Between P and Q Score: Number of correct responses minus one-fourth the number ofwrong re sponse s . Parts: 1; items: 10; working time : 10 minutes. 31. Restricted Figural Classifications - NFC04A, Classify a given set of figures so that each figure is a m e m b e r of two classes. , T / /r~ Class I: / . CJ ^ O _______________ Sample : 1 z--- 7 V --N_ 4 5 / \ 6 > @ /°\ c Class II: Cj Class III: ^ Class IV: / Z . S Score: Number of figures correctly classified twice. Sample : of two classes. 1. 2. A V F B SCP Z Class I: 3. M W D N Class II: 4. 5. P O Y T G X K H C 1 3. S s III 6. W PI R Class IV Parts: 2; items per part: 4; working time: 8 minutes . ify a given list of nonsense words so that each word is a /. C? / <5 y > Z J - S ' Z 4- L > Score: Number of nonsense words correctly classified twice. Pa rts: 2; items per par t: 3; working time: 8 minutes. to ©v 33. Sentence Classification - CMC03A. Designate sentences of two short paragraphs as conveying either (A) fact. (B) possibility, or (C ) name. Sample: 1. The natives of New Zealand have wooden houses Answers: 1, A; 2, C; 3, B. which meet the requirements for cool climate, _ _ _ Sc ore: Number of correct responses minu s one - half 2. ihe Rarotongan word van means ''mud. " ---- the number of wrong responses. 3. The gods informed the people of Tahiti of the Parts: 2; items per part: 15; working time : 8 disaster. minutes. 34. Ship Destination Test - CMSOZD (SSC). Kind the distance from a ship to a port, taking into account the influence of an increasing number of variables. Score: Number of correct responses minus one-fourth the number of wrong responses. Parts: 1; items: 24; working time: 8 minutes. 35. Suff ijte s - DSUOIA (LLT). Write words ending with a specified suffix, such as able. Score: Number of correctly listed words. Parts: 1; items: 1; working time: 3 minutes. 36. Utility Test - D M C 0 1A. lust as many uses as possible for a common object. Score: Number of shifts in category in a series of acceptable responses. Parts: 2; items per part: 1; working time: 10 minutes. 37. Ve rbal C las sif icat ion - C M C 0 2 B (LLT). Assign words to one of two classes, or to neither, each class being represented by a set of four words. Sample: C O W _______ desk ty*- T A B L E H O R S E v X sheep_______CHAIR G O A T _______ rocker ^ B O O K C A S E D O G ________ tree_______ L A M P \/ cat________ nose dre s se r y' Score: Number of correct responses. donkey ______ Parts: 2; items per part: 5; working time: 8 minutes . 38. Verbal Comprehension - C M U 0 2C (SSC ). Select from a group, a word that means about the same a s a given word . Sample: E A R T H A. sugar B. farm C. sun D. soil E. horse Answer: D . Score: Number of correct responses minus one-fourth of the number wrong. j Parts: 1; items: 24; working time: 4 minutes. 39. Word Classification - CMC01B. Select the one word in a set of four that does not belong to the class on the basis of meaning. Sample: A. horse B. cow C. man D. flower Answer: D . Score: Number of items right minus one-third of the number wrong. Parts: 1; items: 20; working time: 5 minutes. 40. Word Completion - CMU01B. Write acceptable meanings for given words. Sample: y / ] Score: Number of acceptable definitions written. C O U R A G E O U S . n Score: N u m b e r oi acceptaDie aeiimtions written. Parts: 1; items: 20; working time: 5 l/Z minutes . 41. Word Fluency - DSU02A (SSC). Write words containing one specified letter, such as O. Score: Number of different words written containing the specified letter. Parts: 2; items per part: 1; working time: 4 minutes. 42. Word-Group Naming - NMU02A. Give a class name to a group of five words. Sample: knife pan bowl rolling pin strainer , - . . _ ^ K e r Score: Number of correct names. 7le. Parts: 1; items: 16; working time: 6 minutes . 43. Word Grouping . 1 ICO 1 B . Given twelve common words, put them into four, and only four, classes, leaving no extra words. Sample: 1. blue 5. larger 9 . ope ne r C las s I: /. /<-> Score: Number of words correctly 2. curte r 6. tight 10. orange Class II: O r / £. classified. 3. driver 7. litt le 11. re dde r C las s III: (n Parts: 2, item s pe r part: 2; working 4. heavy 8. long 1 2. short Class IV: g ■5 G time: 6 minutes. APPENDIX B INSTRUCTIONS AND SAMPLE STIMULI FOR CONCEPT-LEARNING TASKS 97 98 Instructions for Semantic Problem In this problem you will be presented with 96 dif ferent nonsense words. These nonsense words foTm four distinct classes represented by the capital letters A, B, C, and D. All of the nonsense words associated with a par ticular letter have something in common. There are 24 words associated with the capital letter A, 24 with B, 24 wijth C, and 24 with D. Your task is to learn to assign the correct letter to the nonsense words, by figuring out what common property is associated with each letter. For example, any nonsense word with two vowels might always be associated with A, and any nonsense word with the first and last letters the same might always be associated with B. The classes are distinct; that is, no word will be associated with more than one letter. The classes you are to identify in the problem are different from those just mentioned. On the first page of the problem booklet you will be presented with a nonsense word followed by the letters A, B, C, and D. Choose or guess the letter you think is 99 associated with this particular nonsense word. After you have circled your choice, turn the page and the nonsense word will be presented again, but followed this time by the correct letter. On the same page a new nonsense word will be presented, and again you are to choose A, B, C, or D. The next page will have the letter correctly associated with this word. This process will be repeated until all 96 non sense words have been presented. During the first few trials you probably will have to rely on guessing. How ever, on later trials you should be able to make accurate predictions since all of the nonsense words associated with a particular letter have something in common. You will be instructed when to turn each page, when to examine the answer, and when to look at the new nonsense word. It is very important that you follow these instructions. Do not leave any page blank. Always circle one of the four letters. If you do not know the correct answer make your best guess. STOP HERE. WAIT FOR FURTHER INSTRUCTIONS. Instructions for Figural Problem In this problem you will be presented with 96 dif ferent figures. These figures form four distinct classes represented by the capital letters A, B, C, and D. All of 100 the figures associated with a particular letter have some thing in common. There are 24 figures associated with the capital letter A, 24 with B, 24 with C, and 24 with D. Your task is to learn to assign the correct letter to the figures, by figuring out what common property is associated with each letter. For example, figures with curved lines might be associated with A, and figures with three parts might be associated with B. The classes are distinct; that is, no figure will be associated with more than one letter. The classes you are to identify in the problem are different from those just mentioned. On the first page of the problem booklet you will be presented with a figure followed by the letters A, B, C, and D. Choose or guess the letter you think is asso ciated with this particular figure. After you have circled your choice, turn the page and the figure will be presented again, but followed this time by the correct letter. On the same page a new figure will be presented, and again you are to choose A, B, C, or D. The next page will have the letter correctly associated with this figure. This process will be repeated until all 96 figures have been presented. During the first few trials you probably will have to rely on guessing. However, on later trials you should be able to make accurate predictions since all of the figures associated with a particular letter have something in common. 101 You will be instructed when to turn each page, when to examine the answer, and when to look at the new figure. It is very important that you follow these instructions. Do not leave any page blank. Always circle one of the four letters. If you do not know the correct answer make your best guess. STOP HERE. WAIT FOR FURTHER INSTRUCTIONS. Instructions for Semantic Problem In this problem you will be presented with 96 groups of four words each. These word groups form four separate and distinct classes. The four classes are labeled A, B, C, and D. All of the word groups associated with a particular letter have something in common. You are to learn which group belongs with each letter. For example, look at the following word groups: flower joy think funny ship bright red horse book steel milk hate A B yellow window *• throw B camp anger tunnel ®• smoke A B sheet blue 4• glue B stop cough Groups 1 and 2 are followed by A. This means that groups 1 and 2 have something in common. Groups 3 and 4 have something different in common, and are followed by B. Now look at groups 5 and 6. Does group 5 belong 102 with the groups represented by A or with those represented by B? For group 5, B has been circled because group 5 contains the name of a color, red, as do the other groups represented by B (yellow in group 3 and blue in group 4). A is circled for group 6. The group of words represented by A contain a word which indicates an emotion (joy in group 1, hate in group 2, and anger in group 6). In the problem, as in the above example, each word group is associated with one letter on the basis of the similarity of one of the words in the group. There will be 96 different word groups--24 go with the letter A, 24 with B, 24 with C, and 24 with D. Your task is to learn to assign the correct letter to each word group. On the first page of the problem booklet you will be presented with a group of words followed by A, B, C, and D. Choose or guess the letter you think goes with this group, and circle it. When instructed, turn the page. The same word group will be presented again, followed by the letter with which it is correctly associated. On the same page a new group of four words will be presented, and again you are to choose A, B, C, or D. The next page will have the correct letter associated with this group. This process will be repeated for all 96 word groups. Although on the first few pages you will have to rely on guessing, on the later pages you should have enough information to choose the correct letter. 103 You will be instructed when to turn each page, when to examine the answer, and when to look at the next word group. It is very important that you follow these instructions. STOP HERE. WAIT FOR FURTHER INSTRUCTIONS. Sample Pages from the Problem Booklets PROBLEM 100 101 LLAM A B C D PROBLEM ZO O 201 r t v f — V'. A B C D PROBLEM 300 301 fence quiet sour director A B C D Answer 101 L L A M A 102 201 x Z V s \ Answer fence Answer 301 quiet A sour director 202 302 SOZF A B C D A B C D part extreme A „ _ „ A B C D repeat garden 105 APPENDIX C ORDER OF ADMINISTRATION OF TESTS AND CONCEPT TASKS 106 107 Session Session one Letter Group Exclusion Figure Classification Alternate Uses Letter-Concept Grouping Verbal Classification Classified Information Memory of Nonsense Word Classes Symbolic-Concept Task Memory of Word Classes* Picture Class Memory* two Problem Solving Sentence Classification Word Fluency Restricted Figural Classifications Concept Grouping Number-Group Naming Figural-Concept Task Utility Test Figural Hierarchical Grouping Letter-Grouping Group Classification Word Classification Number Classification Multiple Figural Similarities Suffixes Figure-Concept Grouping Multiple Grouping of Nonsense Words Because of time scheduling, Memory of Word Classes and Picture Class Memory were administered at the beginning of session two for the juniors. 108 Session three Verbal Comprehension Figure Exclusion Multiple Grouping Multiple Letter Similarities Largest Class Word-Group Naming Semantic-Concept Task Figural Class Inclusion Letter Classification Ship Destination Test Name Grouping Naming Meaningful Trends Alternate Letter Groups Restricted Symbolic Classifications Multiple Grouping of Figures Word Grouping Word Completion Figure Grouping Picture-Group Nam'Mig

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

A Multidimensional Scaling Of Mood Expressions

PDF

The Effect Of Subject Sophistication On Ratio And Discrimination Scales

PDF

A Factor Analysis Of The Symbolic-Evaluation Abilities

PDF

A Factor Analysis Of The Semantic-Evaluation Abilities

PDF

Personality Variables And Intellectual Abilities As Determinants Of Concept Learning

PDF

Figural And Symbolic Divergent-Production Abilities In Adults And Adolescents

PDF

A Factor-Analytic Study Of Problem-Solving Abilities

PDF

Factorial Stability As A Function Of Analytical Rotational Method, Type Of Simple Structure, And Size Of Sample

PDF

Effects Of Different Treatment Procedures On Reading Ability And Anxiety Level In Children With Learning Difficulties

PDF

A Rotational Approach To Psychological Invariance

PDF

A Preliminary Study Of The Dimensions Of Future Time Perspective

PDF

Intellect After Lobotomy In Schizophrenia: A Factor-Analytic Study

PDF

The Influence Of Order Information And Stimulus Display Time On Short-Term Retention

PDF

The Influence Of Communality And N On The Sampling Distributions Of Factor Loadings

PDF

Reward Expectancy Strength As Related To The Magnitude Of Frustration In Children

PDF

Role Playing Ability And Social Adjustment In Children

PDF

Some Aspects Of The Dimensionality Of School Adjustment Of Fifth Grade Boys

PDF

A Factor Analysis Of The Figural-Evaluation Abilities

PDF

A Monte Carlo Evaluation Of Interactive Multidimensional Scaling

PDF

On The Relationship Between Anxiety And Aggression In Nine-Year-Old Boys

Asset Metadata

Creator Dunham, Jack Lewis (author)

Core Title The Role Of Intellectual Abilities In Concept Learning

Degree Doctor of Philosophy

Degree Program Psychology

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag OAI-PMH Harvest,psychology, general

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Guilford, Joy P. (committee chair), Cliff, Norman (committee member), Wolf, Richard M. (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c18-107479

Unique identifier UC11360795

Identifier 6702106.pdf (filename),usctheses-c18-107479 (legacy record id)

Legacy Identifier 6702106.pdf

Dmrecord 107479

Document Type Dissertation

Rights Dunham, Jack Lewis

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA