flail" This is to certify that the thesis entitled Color—Form Salience and Stroop Interference in Mentally Retarded Clients: An Exploratory Study on the Relationships among Cognitive Styles, Multi— Dimensional Sorting Task Performance, and General Work Adjustment Characteristics presented by William Crimando has been accepted towards fulfillment of the requirements for Ph.D, degreein Counseling \ I 't i / V'v 1271 a; Major ' ofessor -/ 0-7639 A no.‘ C)Copyright by WILLIAM CRIMANDO 1980 COLOR-FORM SALIENCE AND STROOP INTERFERENCE IN MENTALLY RETARDED CLIENTS: AN EXPLORATORY STUDY ON THE RELATIONSHIPS AMONG COGNITIVE STYLES, MULTI- DIMENSIONAL SORTING TASK PERFORMANCE, AND GENERAL WORK ADJUSTMENT CHARACTERISTICS By William Crimando A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Counseling, Personnel Services and Educational Psychology 1980 ABSTRACT COLOR-FORM SALIENCE AND STROOP INTERFERENCE IN MENTALLY RETARDED CLIENTS: AN EXPLORATORY STUDY ON THE RELATIONSHIPS AMONG COGNITIVE STYLES, MULTI- DIMENSIONAL SORTING TASK PERFORMANCE, AND GENERAL WORK ADJUSTMENT CHARACTERISTICS By William Crimando The purpose of this study was to explore the inter- action of two aspects of cognitive style-—dimensiona1 salience and Stroop interference--as they relate to per— formance of a multidimensional sorting task among mentally retarded persons. It also sought to relate cognitive styles with overall work adjustment. It was thought that this would provide new information on the cognitive functioning of the retarded in habilitative facilities, thereby making it possible to devise more effective train- ing strategies for them. Seventy—three clients from four Michigan habilitative facilities were employed as subjects for this study. Thirty non-retarded subjects served as a comparison group for one of the hypotheses. The retarded subjects were assigned to four experimental groups based on their responses to three tests: the Color—Form Preference Test, the Stroop Color-Word Interference Test, and the Quick Test. Subjects in each group were trained and observed on a sorting task in which the color and shape of the objects were relevant to solution at alternate times. The dependent measures were the number of errors of each type: color or shape. The major hypotheses concerned differential per— formance of the sorting task across subjects in a 2x2x2 multivariate factorial design combining levels of each independent variable and a nuisance variable, order of presentation. First, it was thought that color-salient subjects would excel at sorting objects by color, while form—salient subjects would perform better when shape was relevant to task solution. It was also thought that sub- jects with low Stroop interference scores would perform better than those with relatively higher scores. Retarded subjects were also rated on the Counselor (Trainer) Questionnaire, regarding general work adjustment characteristics, and assigned an overall score. It was thought that degree of work adjustment would be associ- ated with different levels of the two cognitive styles. Finally, it was hypothesized that color-form salience would not be associated with the mental age of the subjects. In primary analysis of the data, only one of the hypotheses was supported, concerning differential performance of the task according to level of dimensional salience. Color—salient clients had fewer color than shape errors, while form—salient clients had fewer shape than color errors. While the hypothesis concerning the lack of association between dimensional salience and mental age was not supported entirely, it appeared that the relationship was minimal. Stroop interference did not appear to be associated with either task performance or overall work adjustment. Nor was there any relationship between salience and adjust- ment. Post—hoc analyses of the data revealed complex interactions between the two independent variables and sex and age, which may account for the lack of significant findings. It also appeared that certain items from the Counselor (Trainer) Questionnaire were correlated signi— ficantly with dimensional salience. Results were discussed in terms of their implications for work adjustment training and vocational evaluation. Finally, changes were suggested for future research based, in part, on some of the shortcomings of the present project, and on other questions related to learning style and performance. To Elizabeth Tomac My Best Friend ACKNOWLEDGMENTS I gratefully acknowledge the following people: Dr. James Engelkes and Dr. Marlene Pringle, both of whom chaired my committee during my tenure as a doctoral candidate, for their guidance and the opportunities they afforded me. Drs. Herbert Burks, Charles Coker, Carl Frost, and Bob Winborn, for their comments and guidance in the prepa- ration of this work, and for their willingness to serve on my committee. Dr. Charles Foster, of Michigan Association of Rehabilitation Facilities, for his egis and cooperation. Rita Charron and Daniel Devaney, of Ingham County Community Mental Health, and Dr. Bonnie Huntley and Ernestine Agnew of Kent County Community Mental Health, for their cooperation and assistance. The administrators of various facilities, Michael Mahoney, Anthony Smydra, Judith Touzeau, Roger Webb, and Terrance Weinberger, and their staffs, for making my stay in their centers so pleasant. All of the people who agreed to participate in this study. My research assistants, Karen Assenmacher, David Dubuc, Linda Rodgers, and Ruth Walkotten, who made my job immensely easier. Friends Ariel Anderson and Arthur Tabachneck, for the assistance and support that got me started and finished. Mrs. Jeanette Minkel, who spent long hours typing this volume, and Mr. Gary Nowak, who prepared the graphs used in Chapter IV. Finally, I wish to gratefully acknowledge the partial support for this research provided by the Research and Training Center of the University of Wisconsin — Stout, through Grant No. l6—P-56821/5, from the National Institute of Handicap Research of the Rehabilitation Services Administration. iv TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES LIST OF APPENDICES Chapter I. THE PROBLEM Introduction and Need Purpose . . Definition of Terms Overview II. REVIEW OF RELEVANT RESEARCH Dimensional Salience Stroop Interference Summary . . . III. METHODOLOGY Selection of the Sample . Characteristics of the Sample Phase One . . . . . . Phase Two Instrumentation . Color— Form Preference Test Construction Administration Scoring . Standardization Rationale for Construction Stroop Color- Word Interference Test Construction Scoring Administration Standardization Quick Test Administration Scoring Rationale for Inclusion Page vii KO \IGU‘IH l—' Chapter Counselor (Trainer) Questionnaire Scale Construction Scoring . Standardization . . Multidimensional Sorting Task . Administration and Scoring. Procedure . . . Design Hypotheses . Analysis of Data Summary . . IV. ANALYSIS OF RESULTS Demographic Data . Experimental Characteristics Tests of Hypotheses Hypothesis 1 Hypothesis 2 Hypothesis 3 Hypothesis 4 Hypothesis 5 Counselor (Trainer) Questionnaire Systematic Trends . . . . Sex . . . . . . Age . . . . . . Work Adjustment Summary of Results V. SUMMARY AND DISCUSSION Summary . . Problem . Design Results Discussion Implications for Future Research Conclusions . . . . . . . . BIBLIOGRAPHY APPENDICES vi Table LIST OF TABLES Results of Selection Procedures: Number of Clients Used at Each Phase and Attrition Test-Retest Reliability of the Color-Form Preference Test: Coefficients, Significance Levels, and Number of Clients . . . . . Demographic Characteristics of Retarded and Non- Retarded Subjects. Percentages and Raw Numbers . . . . . . . . . Experimental Characteristics of the Sample Summary of Multivariate Analysis of Variance for Test of Hypothesis 1 . Summary of Univariate Analyses of Variance for Hypothesis 1 . . Summary of Multivariate Analysis of Variance for Test of Hypothesis 2 . . . . . . . . Pearson's Product—Moment Correlations for Tests of Hypothesis 3 . Summary of the Correlations for Tests of Hypothesis 4 . . . . . Responses to Counselor (Trainer) Questionnaire Estimated Means for Color and Form Errors of Each Experimental Group Categorized by Sex Summary of Multivariate Analysis of Variance for Tests of Sex by Independent Variable Interaction . Estimated Means for Color and Form Errors of Each Experimental Group Categorized by Age. Page 31 37 61 64 68 68 71 72 73 75 78 81 82 Table Page 4.12 Summary of Multivariate Analysis of Variance for Test of Age by Independent Variable Interaction . . . . . . . . . 84 4.13 Sex and Age Differences in Overall Work Adjustment . . . . . . . . . . . . . . . . . 85 4.14 Correlations of Individual Items on Counselor (Trainer) Questionnaire with Dimensional Salience and Interference (CW-C) Scores . . . 86 viii LIST OF FIGURES Figure 3.1 Design of the Study 4.1 Estimated Cell Means for the Variables Mental Age and Interference Scores for the Four Experimental Groups 4.2 Estimated Cell Means for Pure Color and Pure Form Errors in the First Trial of the Sorting Task for Eight Experimental Groups 4.3 Graphs of Mean Color and Form Errors to Illustrate Sex by Independent Variable Interactions . . 4.4 Graphs of Mean Color and Form Errors to Illustrate Age by Independent Variable Interactions . ix Page 53 66 69 79 82 LIST OF APPENDICES Appendix A. Administrative Agreement B. Informed Consent C. Complete List of 15 Items of Color— Form Preference Test . . . D. Instructions to Client and Researcher for Color- Form Preference Test E. Color—Form Preference Test Scoresheet F. Facsimile of the Word, Color, and Color-Word Cards for Stroop Color—Word Preference Test . . . . . . . . . . . . . . G. Stroop Color-Word Interference Test Scoresheets H. Instructions to Client for Stroop Color-Word Interference Test . . . . . . . . . . I. Counselor (Trainer) Questionnaire J. Summary of Training Procedure Instructions to Research Assistants . . . . . . . . . K. Purpose of Study L. Combined Purpose and Informed Consent for Use at One Site Page 113 114 115 116 117 118 119 121 122 124 126 127 CHAPTER I THE PROBLEM Introduction and Need One of the primary tasks faced by contemporary voca- tional rehabilitation personnel is the habilitation and placement of persons who are severely disabled. Since the 1973 Rehabilitation Act (U.S. Congress, 93-112), which mandated a shift in target populations of service to severely handicapped, practitioners have been trying to devise new techniques for counseling, training, and other— wise serving these persons. These attempts have been accompanied by a wave of published material on experimentaL quasi—experimental, and correlational research describing and establishing the efficacy of those techniques. A cursory glance through a bibliography of research during the past five years would reveal a plethora of articles touting new methods for working with deaf persons, cancer patients, cardiac risk clients, and so on. One exception to this inundation of research has been in the area of mental retardation. In light of the physio- logical and behavioral problems often attending retardation, as well as societal attitudes toward them, retarded persons have certainly earned the dubious distinction of being among the most severely handicapped. Zigler (1978), however, described a "national crisis in mental retardation research 2 (p. 1),” pointing out the severe deficit in research directed toward the habilitation of these individuals. He highlighted several areas, among which are the areas of human learning and cognitive performance: More research is also needed on ways to enhance the behavioral functioning of retarded individuals. Many specific areas of inquiry that can lead to improvements have been listed by Haywood (1976). In the subject of human learning, for instance, we need to know what educational, social, and moti- vational factors influence learning and how these factors interact. We also need to determine what factors can enhance cognitive performance, and, thus, increase the number of retarded persons that may be educated and trained to do useful work. (p. 7) In light of this statement showing the extent to which more research is needed, and the general movement to devise new techniques or modify old ones to serve the severely handicapped, one would expect a representative amount of research on techniques for the retarded. This has not been the case, especially in rehabilitation and habili- tation services. The purpose of the project was to remedy this situation to some extent, by investigating two aspects of cognitive performance in retarded persons. Vocational rehabilitation is a learning experience; clients learn new modes of adapting to their handicaps, their environments, and the attitudes of those who have an impact on their lives. All persons bring a characteristic learning style to every situation. Learning, or cognitive, style consists of the tendencies a person has for attending to and categorizing one's environment, performing tasks in 3 a certain way, and interacting with others. Rusalem and Rusalem (in Rusalem & Malekin, 1976) reported that most rehabilitation failures occur because of the incompati- bility between instructional procedures used with a client and the learning style of the client. The same thing can be said of habilitation failures with the retarded. Con- versations with area habilitation personnel reveal that they concur with these findings. It is of utmost importance, they believe, to discover those specific condi- tions under which an individual client will learn, rather than presenting the same conditions for all and hoping that each client will respond equally to them. Rusalem and Rusalem are advocating, then, an individualization of the rehabilitation process. The move toward personalization is certainly not with— out precedent. In fact, individualization permeates the process: First, medical treatments and physical restora- tion are different for everyone, based on the particular needs of the client; second, since the 1973 Act, which required the development of Individualized Written Rehabili- tation Plans, policy—makers have been urging practitioners to assume nothing about clients, and to look upon each one as an individual (Wright, 1975); finally, job engineers look at the man-machine relationship of different jobs, and modify non-essential aspects to make them more compatible with the specific characteristics of any one client. 4 Bellamy, Horner, and Inman (1979) assume that "accommodation to individual differences is an integral aspect of service to severely retarded individuals (p. 6)," yet the inclination has been to treat all retarded persons as alike. The logical, though unfortunate, consequence of non-individualization is that millions of dollars are poured into treatment plans, training programs, and counseL- ing time with little or no assurance of success for any one client. Granted, in many cases the counselor or specialist notices the "symptoms” of an inadequate treatment plan in time to make the appropriate changes, symptoms such as extended periods of time in work activities programs when the potential for some form of regular employment clearly exists. But this trial-and-error approach costs time and money, and frequently results in public disillusionment wifli rehabilitation services. In this era of Proposition 13, sunset laws, and zero—based budgeting, habilitative facili— ties can ill afford an alienated public. A step toward personalization, by matching the habilitative process with the client's cognitive style, would do much to ensure con- tinued public support of services. There are many different constructs that have been called cognitive style; it is, indeed, a multifaceted characteristic. Some researchers (e.g., Rotter, 1966) have looked at broad aspects of cognitive set; a cognitive set is produced through specific learning and reinforcement experiences. A problem with this method is that it has been 5 difficult to devise a valid and reliable method of assess- ing broad constructs (Mischel, 1973). Other researchers, notably Piagetian developmentalists, have viewed the evolu- tion of cognitive styles in individuals as an ontogenetic process germane to the whole human species. Consequently, they have studied small, well-defined cognitive tasks, the successful completion of which marks an individual's posi- tion along a continuum of developmental stages. The research in this area, while highly successful, has by no means been conclusive. The significance of the distinction between cognitive sets and stages is important. "Set” implies that one could change a cognitive style by pro— viding a series of learning experiences. A developmental stage, however, is relatively more impervious to extra- ordinary attempts to accelerate it. Purpose The purpose of the present project was to explore the interaction of two aspects of cognitive style-—dimensional salience (Odom & Mumbauer, 1971) and Stroop interference (Dyer, 1973)-—and their effect on performance of a multi- dimensional sorting task among mentally retarded persons. Retarded clients in Michigan habilitative facilities were tested for color-form salience and high—low Stroop per- formance, and, on the basis of these, were divided into four groups. Each group was given a multi—dimensional sorting task. 6 The questions to be answered were: Is there differ— ential performance on the task across the groups? Does the evidence support the proponents of a cognitive set theory of cognitive style rather than those of a develop— mental stage theory? A final portion of the project dealt with the rela- tionship between cognitive style and habilitative performance. All clients tested were also rated on differ— ent aspects of work adjustment and an overall rating was assigned to each. These overall ratings were correlated with scores on both the dimensional salience and Stroop tasks. It was believed that a significant relationship would be found to exist between them. Definition of Terms Client: A retarded individual receiving vocational habilitative services in an habilitative or rehabilitative setting. Cognitive style: The tendencies one has for attend- ing to and conceptualizing the environment, performing tasks in a certain predictable manner, and interacting with others. Developmental stage: A universal, predetermined period in the cognitive growth of an individual, charac- terized by a specific mode of functioning. Each successive stage depends on the previous one, resulting in an invariant sequence of development. Dimensional salience: The tendency to attend selectively to one dimension (e.g., color, form, size, etc.) of a stimulus over others. Facility: An agency providing direct client services in the form of evaluation, treatment, and training. Mental retardation: ”Significantly subaverage general intellectual functioning existing concurrently with deficits in adaptive behavior, and manifested during the developmental period (Grossman, 1973, p. ll).'l Stroop phenomenon: The cognitive interference to naming colors resulting from their association with incon— gruent color names. Vocational habilitation: ”A process through which an individual is helped to develop work skills relevant to the job market and then to secure remunerative employment (Bellamy et a1., 1979, p. 16)." Overview The following chapters of this work will contain a description of the project as outlined in this chapter. In Chapter II, a review of literature relevant to dimensional salience and Stroop interference will be pre- sented. Methodology will be described in Chapter III; included will be explications of sample selection and instrumentation, presentation of the design, a statement of the hypotheses in testable form, and the statistical models used for data analysis. Chapter IV will consist of of the results, the 11minstatio of the I . ewzations for future projects on the same topic, and conclusions will be contained in Chapter V. CHAPTER II REVIEW OF RELEVANT LITERATURE A review of cognitive developmental psychology reveals a rather abundant body of literature on the two issues of main concern in the present study, dimensional salience and Stroop interference. Research applying either of these constructs to the mentally retarded, how— ever, has been noticeable by its absence. Neither has there been any attempt to study the interaction of the two constructs. The following review, therefore, will focus on the constructs separately, highlighting the literature which is more pertinent to the present project. Each expo- sition will begin with a brief description and history of the construct, and will end on relevant research. Dimensional Salience Investigators have long believed that when a learner is given a discrimination problem in which two or more dimensions (e.g., form, color, size, brightness, etc.) are presented, with only some of them relevant to solution of the problem, that performance will be better if the rele- vant dimension is salient for the learner. Dimensional salience indicates a tendency to attend selectively to one dimension over others. There is believed to be a develop- mental hierarchy for dimensional salience. Piaget (1952) suggested that the earliest mode of responding to external lO stimuli in normal children is random and purposeless, followed by attention to color at about age two, or the advent of the preconceptual stage. With further matura— tion, the salience of form as a basis for schemata (or categories) increases relative to color, until about the age of six, when it becomes dominant. There has been a host of persuasive cohort studies giving evidence for the existence of such a hierarchy, though the evidence is far from conclusive. As in all of the cognitive developmental tasks of Piaget, performance on a test for color-form salience would be indicative of the level of cognitive functioning, or intelligence, to which one has matured. In fact, some Piagetians have proposed that dimensional salience tasks are a more meaningful measure of intelligence for marginal populations (e.g., the mentally retarded), than more verbal tests which, they say, tend to overestimate developmental level by focusing only on superficial intellectual and social learning (Kay, 1977). The developmental hierarchy of Piaget has not gone unchallenged. Some early researchers (e.g., Brian & Goodenough, 1929) suggested that salience for form is dominant over color in infants. As the infant matures, color increases in salience; but by age six, color is replaced by form, again, as the dominant dimension. 11 Finally, there are those who reject the notion of developmental stages, and tend to View dimensional salience or any other behavior, as more individually defined from specific learning experiences. Gaines (1970) averred that salience should more appropriately be considered a cogni- tive set, or a "predisposition to make predictable responses to given stimuli, but the source of this predisposition is appropriate training and experience (p. 980)." She pointed out that, by normal definition, a developmental stage can- not be changed by specific and appropriate training, and is, for all intents and purposes, irreversible. As evi- dence for her position, she noted several factors that affect dimensional salience. For one, culture plays a part; adult Africans consistently chose color (Suchman, 1966). Physiological status, too, affects responses; deaf children, 8—12 years old, were color dominant (Gaines, 1964). Finally, race and socio—economic status are likely to be correlated with color—form dominance (Marcus, West, & Gaines, 1968). In line with the developmental progression position, Melkman, Koriat, and Pardo (1976) hypothesized that children would show an early salience for form, switch to color, and then revert to form. They tested the color— form salience of 55 male and 54 female upper middle—class Israeli children, aged 2 to 4 years, 11 months. Their results substantiated their hypothesis, thus upholding early researchers (Brian & Goodenough, 1929). Cronbach's 12 alpha for six age cohorts ranged from .91 to .97, thus showing internal consistency. Test-retest reliability was more variable and, generally, not quite as favorable, ranging from .65 to .98. Fields (1973) studied color, size, and form prefer- ences in 128 infants, aged 4 months to 11 months, with a forced-choice,paired-comparison technique. In these types of designs, each dimension is tested separately by pre— senting the child with pairs of stimuli and noting to which he/she attends. On the basis of patterns of selective attention, it is decided whether the child is randomly attending to stimuli, perseverating on location, or actually attending to the relevant dimension. She found that, among the infants she tested, about one-third had form preferences, while less than one-third had color or size preferences, thus supporting, again, Brian and Goodenough (1929). Suchman and Trabasso (1966) studied color and form preference in 145 children, aged 3 to 6 years. They found that younger children aged 3 to 4 years, two months, had color preferences and older children aged 4 years, 3 months to 6 years, had form salience. Only 12 of the children showed mixed or no preference. Gaines (1964) tested the color-form dimensional sali— ence in 12 pairs of hearing and deaf children, aged 8 to 11 years. They were matched for age, sex, socio-economic status, and race. The hypothesis that deaf children, l3 regardless of age, would have color preferences, while hearing children would have form preferences, was supported. Color was preferred by 55% of the deaf children but only 23% of hearing children. In contrast to these studies, others have supported the Piagetian progression. Kay and Singh (1975) tested the feasibility of the use of color-form preferences for nosological and prognostic discrimination among individuals at three stages of cognitive dysfunction. They adminis— tered a color-form preference test to 33 schizophrenics and 26 non-psychotic patients, ranging from 18 to 47 years in age. Among the schizophrenic patients, 70% were identified as color salient or of mixed salience. The 30% identified as form salient were mostly acute-subacute schizophrenic and paranoid patients, while the color sali- ent were predominantly chronic schizophrenics. Finally, the nonpsychotic groups consisted largely of form salient persons. Kay and Singh believed this test to reveal a primitive and impoverished mode of thinking among the chronic schizophrenics, and supported a Piagetian interpre- tation of cognitive development. In another study, Kay (1977) tested 55 mentally retarded psychotics with a color—form preference test, and found a significant discrimination between color-form dominance over severity levels of retardation. In general, he found that the lower the functioning level of the sub- ject, the greater the likelihood of color dominance or 14 random responsiveness. In the same study, he found that color-form preference results identified individuals as lower on a cognitive functioning scale than did measured mental age (Quick Test; Ammons & Ammons, 1962). The evidence, then, seems to be divided among the three theories of color—form salience. The results of Fields (1973), Melkman et a1. (1976), and Suchman and Trabasso (1966) supported the Brian and Goodenough (1929) progression (i.e., form - color - form). Kay (1977) and Kay and Singh (1975) supported a Piagetian viewpoint (i.e., color-form), while Gaines (1964, 1970) affirmed the cogni- tive set position. What accounts for the disparity among results? How can three somewhat conflicting positions be reconciled into one practical theory of cognitive development? While a por- tion of the conflicting evidence may be attributed to methodological differences, it is believed that the answer lies elsewhere. Suchman and Trabasso (1966) provided a possible explanation. While their results supported the Brian and Goodenough (1929) position, they, themselves, favored the cognitive set theory, as revealed in other research (e.g., Gaines, 1964; 1970). They posited that evi- dence for cognitive stages can only be found in group averages, and that any one person's dimensional salience is more likely to be a matter of learning experiences. It is important, they believed, to discover those factors which would alter color and form preferences to a cognitive set 15 which is deviant from the prevailing style of one's develop— mental cohort. The results of Kay (1977) and Kay and Singh (1975) might easily be interpreted in that light, with psy- chosis being one of the contributors to this phenomenon. As stated earlier, it is believed that performance on color-form related problems would be better if salient dimensions were relevant to solution. Dimensional salience was shown to be a personological variable. Dimensional relevance, on the other hand, is a task variable implying that correct performance is tied to identifying and respond- ing to one dimension of a problem over others. The evidence here, too, is mixed. In two previously cited studies (Gaines, 1964; Melkman et al., 1976) performances on color and form discrimination, labeling, and identifica— tion tasks supported this hypothesis. Other researchers have tested this hypothesis with the performance of children on color—form tasks. Julius (1974) hypothesized that economically-disadvantaged children who had a strong preference for either form or color would find shifting to the non-preferred dimension more difficult than would economically disadvantaged children without strong preferences. Shift scores were used as a measure of the number of correct responses made when the shift from the preferred to the non—preferred dimension occurred. In all cases, children who showed extreme preferences for one or the other dimension had lower shift scores than those who showed moderate or no dimensional preference. Marantz 16 (1976), though, found no differences between performance in preferred and non-preferred dimensions in her study with 780 kindergarten children. Neither did Odom and Mumbauer (1971), who hypothesized that a significant cor- relation would be found between color—form salience and performance on concept identification tasks in which either color or form was relevant. The only correlation they, in fact, did find was between age and performance; young children consistently had more errors in the tasks than did older children. The discrepant evidence presented above echoes the admonition of Suchman and Trabasso (1966). It is point- less to merely label persons as deviant from their appropriate position along a continuum of developmental stages. What is important is to discover those factors which lead to that discrepancy. It was thought, in the present study, that the Stroop interference task would pro- vide one such factor. Stroop Interference The interest in relative speeds of color naming and color-word reading dates back to Cattell (1886). It was noted then, and has been observed time and again since, that color naming requires more time than colortword reading. In other words, it takes longer for a person to see a red color strip and vocalize "red" than it does for a person to see the word red” and say "red.” By far the 17 most interesting part of study of this phenomenon was introduced into American psychology by Stroop (1935), who noted differences in color naming. He found that the time that is required to name a color is substantially increased when that color is paired with an incongruent color-word. That is, a longer time is generally required to say ”red” if the stimulus is the color—word "green" printed in red ink, than if the stimulus is a red color strip. Jensen and Rohwer (1966) compiled a comprehensive review of the Stroop Color-Word Test. They offered three reasons for its extensive use in cognitive psychology: (a) it yields highly reliable and stable measures of individual differences on what seem to be three quite simple and basic aspects of human performance; (b) though there are reliable differences in each of the three time scores . . . (they) maintain the same rank order of magnitude for all subjects. (c) the test has been used in a large variety of studies and has shown significant correlations with a host of other, often more complex, psychological measurements. (p. 36) The usual form of the test is to give three cards to the testee, who is asked to read the color words or to name the colors printed on them. The first card, the word card (W), contains from three to five color names (e.g., blue, green, yellow, etc.) printed over and over. This card is read by the testee, who is timed and whose errors are noted. The second card is the color card (C); on this card the words are replaced by strips of the colors which 18 the words represent. Again the person is timed and errors are noted. The third card is the color—word (CW) or inter- ference card. On this card, the same words appear as on W, but they are printed in incongruent colors; that is, the word ”green," for example, would never be printed in green, but in every other color used. As on both other tasks, the testee is timed and errors are counted. These tasks yield three highly reliable time scores for C, W, and CW, and a number of derived scores based on a combination of the three. Examples of derived scores include: (1) CW—C (Callaway, 1959); (2)(CW—C)/W (Thurstone & Mellinger, 1953X and (3) W x(CW-C)/C (Thurstone, 1944). Though there have been few attempts to establish norms for the three basic scores, it usually takes one and one—half to two times as long to read the color card as the word card, and two and one-half to three times as long to read the color-word card as the word card. One of the derived scores is of primary importance to the present study: CW—C. This score has a heavy factor loading (.97) of what is believed to be the interference factor (Jensen & Rohwer, 1966). What this score is measur- ing is the ability to attend to the relevant dimension (color) in spite of the distraction caused by an interfer- ing dimension (incongruent word). Persons with low CW-C scores are better able to attend to the relevant dimension than people with high scores. 19 Like the color—form tests reviewed previously, the Stroop phenomenon has developmental and learning experience explanations. The reasons for which it takes more time to name colors than to read words have been attributed to differential practice in color naming and word reading. Jensen and Rohwer (1966) summarized nearly a half century of explanations, writing: Adults do not spontaneously react to every object or color they see by giving it a name, while the mere act of recognition of printed words implies a covert, if not overt, verbal response. Conse— quently, the habit strength for responding verbally to printed words is proved to be greater than the habit strength for verbally responding to objects and colors. (pp. 55—56) Stroop's own belief (1938) further refines the above explanation; he believed that only one dominant response habit is associated with each word, while colors are associated with a number of response tendencies which interfere with a quick response of the specific color—name requested. This explanation has been echoed by more recent researchers (e.g., Nealis, 1974; Schiller, 1966), believe that colors, geometric symbols, faces, etc., pro— vide more opportunity for attaching personal meaning and, consequently, allowing interference by competing responses. Words and letters, on the other hand, provide a more efficient coding mechanism, through cue redundancy. Each letter, part of a letter, or part of a word provides redundant information. For instance, if one were reading from the list of colors ”red, blue, orange . . . one 20 H II would need only see the o in orange to recognize the word. The rest of the word provides additional cues as to the specific response required. The color-word interference phenomenon has been explained in the same habit—strength terms. Stroop (1935) attributed CW interference to response competition between habits of unequal strength; word reading, a dominant habit, has to be inhibited in favor of the weaker habit, color naming. As proof of this he offered the evidence that reading the words on CW is not affected by the incongruent colors. Dyer and Severance (1972), Gumenik and Glass (1970), and Nealis (1974) found significant reverse inter— ference effects after intense color-naming training, there— by augmenting the applicability of the habit strength interpretations. Comalli, Wapner, and Werner (1962) believed that CW performance is a function of a person's ability to main— tain a course of action in the face of intrusion by other stimuli. As such, they concluded that it is an aspect of cognitive development, related to an increase in perceptual and cognitive differentiation and hierarchic integration with increasing maturity. They studied interference times and error scores of over 200 persons ranging from 7 to 80 years old. They found, indeed, that time and error scores were highest in younger testees, gradually declined between the ages of 13 and 44, and increased after that. Evidently the ability to overcome the interference effect is related 21 to the ability to differentiate more efficiently between the two semantically—related (though conflicting) proper— ties of the stimulus, that ability being a function of increasing cognitive development. Comalli et a1. believed this increasing efficiency in differentiation to be onto- genetically based; others (e.g., Gibson, 1969) attribute it to accumulated learning experience. After age 44, increased age brings, not an increase in cognitive maturity but a decrease in overall response efficiency due to the decline of personal tempo generally associated with age. These findings have been replicated in numerous studies using different portions of the age continuum (e.g., Schiller, 1964; Wise, Sutton, & Gibbons, 1975). Most Stroop research has been done with persons of normal intelligence, so that it is difficult to say whether intelligence confounds interference scores. Three studies are known in which persons with sub-normal intelli- gence were tested. Silverstein and Frankin (1965) were unable to find differences between mental age cohort groups because the regarded subjects kept losing their places on the three cards. Basset and Schellman (1976) overcame some of the weaknesses of the earlier study, and found that a significant interference effect did exist in retarded popu- lations. They did not, however, attempt to measure dif- ferences among varying intelligence levels. The one study in which normal and sub-normal intelligence levels were compared was completed by Uechi (1972). 22 Interference scores were found to be negatively correlated with intelligence in 38 normal and 48 mentally retarded persons. While the full report was unavailable for review, one can infer from the abstract that some attempt should be made to control for intelligence levels in research using the Stroop tasks. Jorgensen (1977) administered the Stroop interference task and the Speilberger State - Trait Anxiety Inventory (Speilberger & Gorsuch, 1966) to 50 female first—year psychology students. She found a significant curvilinear relationship between anxiety levels and Stroop performance. Women with medium anxiety levels had higher response latency (interference time) than did those with high or low anxiety levels. Since anxiety and performance, in general, have a curvilinear relationship which is the inverse of the one Jorgensen found (Anastasi, 1976; Fein, 1963) it would be advantageous to discover to what degree the Stroop and performance in habilitation centers are correlated, inas— much as it would provide a diagnostic device. The body of Stroop research has utilized the inter- ference task in three ways: (1) to describe a population (e.g., Comalli et al., 1962); (2) as an independent variable in correlational research (e.g., Jorgensen, 1977); and (3) as a dependent measure in experimental design (e.g., Coker, 1973). Few researchers have used the Stroop task as a concomitant variable, as it was employed in the present study. Two instances which bear special weight on 23 this project were presented by Broverman (1960a, 1960b). In the first, he divided 46 males into conceptually domi— nant and perceptual—motor dominant groups on the basis of a regression of their scores on C from their scores on W in the Stroop test. Conceptually dominant persons were those who were expected to excel at tasks which required coordination of previously learned rules with sensory stimuli, e.g., arithmetic problems. Perceptual-motor dominant males were expected to excel at tasks which required coordination of sensory stimuli with other sensory stimuli or motor performance, e.g., pressing a button when a light flashes. Each of these two groups was divided into strong automatizers and weak automatizers on the basis of regression scores of CW from C. The more CW was above that expected (i.e., the greater the interference score) the weaker the ability to automatize. Automatization refas to the degree to which one was expected to improve per- formance of a routine task after practice. Strong auto- matizers were expected to be able to routinize more dif— ficult tasks than weak automatizers. He then had each male perform tasks which would befit each of the four types (conceptually dominant, strong and weak automatizers; perceptual—motor dominant, strong and weak automatizers). As hypothesized, they were less distractable when perform- ing tasks which matched their cognitive styles. 24 In the second study (Broverman, 1960b), he divided 35 males as above, and again gave them tasks befitting their cognitive styles. He hypothesized that performance would be better if the task matched cognitive style. Results supported this hypothesis; conceptually dominant, strong automatizers excelled on the Verbal Meaning subtest of the Primary Mental Abilities (PMA) Test (Thurstone, 1944); perceptual-motor dominant, strong automatizers excelled on the PMA Space sub~test, etc. Since in the present project clients were divided on the basis of their interference scores, it is expected that a similar differ— ence in abilities will be noted on the sorting task. Summary In this chapter, the body of literature on cognitive developmental psychology was reviewed. Inasmuch as there has been no research on the interaction of the two main constructs of this study-—dimensional salience and Stroop interference--the two were reviewed separately. Research on dimensional salience, though abundant, leaves two rather basic questions unanswered. First, while most researchers would agree that there is a pro- gression of dimensional salience, there remains some question about the degree to which specific learning experiences are accountable and what part maturation plays. Some authors (e.g., Brian & Goodenough, 1929; Piaget, 1952) think that this progression is a universal, 25 ontogenetic process which will occur provided that there is no overall retardation of cognitive development. Others (e.g., Gaines, 1970) are of the opinion that dimensional salience is a cognitive set and as such is formed by, and can be altered by, specific learning experiences. Second, can differences in dimensional sali— ence lead to differences in performance on problems in which a preferred or non—preferred dimension is relevant to solution? The evidence here, too, has been conflicting. Some researchers (Gaines, 1964; Julius, 1974; Melkman et al., 1976) have found for the prevailing hypotheses-— that cognitive style is correlated with performance--while Marantz (1976) and Odom and Mumbauer (1971) found no such association. In the present study an attempt was made to shed new light to both of these questions, by studying the effects of cognitive interference on performance of such tasks. Stroop interference, too, has been a widely researched construct. Two explanations have been proposed for the effect. Some (e.g., Stroop, 1935) contend that the inter— ference is due to differential practice in word—reading and color-naming. If that were the case it would have little relevance for the present study. Others (e.g., Comalli et al., 1962) hold that CW performance is a func— tion of a person's ability to maintain a course of action in the face of intrusion by other stimuli. As such it 26 would be a more generalized variable and may be related to differential performance on other tasks. The intent of this study was to relate Stroop performance with color- form sorting performance and with general work adjustment characteristics, thus providing evidence to resolve the two viewpoints. Of greater importance than resolving viewpoints in cognitive-developmental psychology, is that this study involved an attempt to provide new insights for habilita- tive programming for the mentally retarded. It is the practice in facilities to provide the same types of train- ing, under the same conditions, for the same jobs to all clients, regardless of individual differences. This has often been due to budgetary problems; but, in most cases, it has been due to the lack of a compelling reason to change. There has been little evidence, thus far, that training to individual differences is any more effective than training to group similarities. The purpose of this study was to explore such reasons to change. If task training and work adjustment training are uniform for all clients, there are, theoretically, some clients who are performing poorly because their cognitive styles are not compatible with training strategies. While dimensional salience and Stroop interference relate, specifically, to task performance, failure on tasks could easily lead to client or trainer frustration, emotional .Igvftive styles affect performance may a evnse imp?» a and justification for true personalization of the habilitative process. . CHAPTER III METHODOLOGY This chapter contains the methodology employed in the present study. Included are selection procedures, charac— teristics of the sample, instrumentation, procedures, design, hypotheses studied, and the methods by which those hypotheses were tested. Rationales for selection pro- cedures, instrumentation, and statistical methods employed are presented in each of those sections. Selection of the Sample The project was divided into two phases, the pretest phase and the sorting task/posttest phase. Each phase had its own selection criteria. In the first phase, habilitation sites—-including adult activities centers, work activities centers, sheltered workshops, and comprehensive rehabilitation centers-—were contacted to secure the names of clients for possible inclusion in the project. Sites were selected from among those associated with Michigan Association of Rehabilitation Facilities, that provide clients with some form of on-site work activities. In order to minimize random effects due to site differences, clients were chosen from the fewest number of sites required to yield the desired number of clients. 28 29 On the basis of the above criteria, four sites were contacted and, after permission was secured from their respective Community Mental Health Boards, administrative agreements (Appendix A) were established with each. One board requested changes in both the agreement and the client's Informed Consent (Appendix B) before giving their permission. Following the procurement of the contracts, 110 clients who might meet the following criteria were sug— gested by site administrators. The population of interest included all clients who: 1. had been classified at borderline-retarded or below level (IQ=69 or below) could understand and speak standard English had normal color vision could read at the first-grade level or better LnJ-‘LJJN had adequate fine finger dexterity in at least one hand The first and second of these criteria were assessed by site staff. The last three were assessed by the administration of a short screening test, which required the clients to correctly identify by name, color strips in red, green, blue, orange, and yellow, read the above—listed words, and pick up a dime. Seventy—three clients passed the screening test and were pretested on the Color-Form Preference Test, Stroop Color—Word Interference Test, and the Quick Test. Of the 57 clients who were able to complete all three tests, 32 30 were selected for the second phase of the project. Inas- much as an established selection procedure for this type of project did not exist, the following procedure was set: 1. Clients were identified as being either color- or form—salient. There were 12 and 45 clients, respectively, in each group. 2. The color—salient clients were further classi— fied as being either low or high interference, dependent upon whether they scored in the lower or upper half of the range of interference scores. 3. The 20 form-salient clients who most nearly matched these groups on mental age (within one year) and interference scores (within 30 seconds) were selected and assigned to the high and low interference groups. When both conditions could not be satisfied, the former was given priority. The object of that process was to yield a 2x2 design (i.e., salience x interference) with proportional cell sizes of subjects who were matched on mental age. During the second phase of the project, however, three clients asked to be excused from further participation, as was their right. The results of the selection procedures are summarized in Table 3.1. 31 Table 3.l.--Results of Selection Procedures: Number of Clients Used at Each Phase and Attrition Number of Attrition Clients Attrition Rate Phase One Referred by sites 110 Passed screening test 73 37 .34 Completed 3 tests 57 16 .22 Phase Two Experimental groups Color—salient/low- interference 5 1 .17 Color—salient/high- interference 5 l .17 Form-salient/low— interference 10 O .00 Form-salient/high- interference 9 1 .10 32 As is evident from Table 3.1, the highest attrition rate occurred at the screening test. This was expected, and was the reason for which 110 names were secured initially. The second highest attrition occurred during the pretest. All of the 16 clients who failed to com- plete the three tests did so because of inability to do the Stroop. Two of these clients claimed not to have been able to read the Word card, although they had passed the screening test. There is no practical differential mor- tality among the experimental groups. The rates differ due to actual number of clients in each group. During the first phase, 2 high school graduates and 28 college students were asked to participate in the project. These were persons who could speak and under- stand standard English, had normal color vision, and could read at the first-grade level or better. These persons had the Color—Form Preference Test, Quick Test, and Stroop administered to them and served as a comparison group for Hypothesis 3. Due to the selection process employed in the two phases of the study, the final sample of subjects restricted direct generalization, and a cautious approach was taken in the interpretation of results. Characteristics of the Sample Phase One. The 73 clients who passed the three tests ranged in age from 20 to 69 years (R = 36 years). The 33 sample consisted of 28 males and 45 females. Their mental ages and IQ's (Quick Test) ranged from 3.5 to 18 years (X = 36 years) and from below 40 to 98 (K = 62), respectively. Their interference scores, explained else- where in this chapter, ranged from 5 to 432 seconds (X = 115 seconds); these scores were not available for 16 clients, due to inability to complete the Stroop. The 20 females and 10 males in the non-retarded group were 19 to 38 years old (Y = 28), with IQ‘s (Quick Test) from 75 to 135 (K = 108). Mental age is meaningless for them due to a ceiling of 19 years. The group was comprised of two high school graduates, four college undergraduates, and 24 master‘s and doctoral students. Their interference scores ranged from 1 to 107 seconds (X = 48 seconds). There were 23 form-salient and one color-salient persons in this group, while 6 showed no preference. Phase Two. The 29 clients who comprised the experi- mental groups consisted of 20 females and 9 males, 20 to 69 years old (X = 36). Mental ages ranged from 4.5 to 16 years (X = 8.4 years), and interference scores from 5 to 432 seconds (X = 105 seconds). Instrumentation Three instruments were used in the selection of sub- jects. One was a version of the color-form preference tests similar to those used in developmental research in 34 the past (e.g., Kay, 1977); Odom & Mumbauer, 1971). Another was the Stroop Color—Word Interference Test (Jensen & Rohwer, 1966). The third, the Quick Test (Ammons & Ammons, 1962) was used to block subjects on the basis of mental age. Additionally, the Counselor (Trainer) Questionnaire was developed and used to collect data con- cerning the relationship between cognitive styles and work adjustment. For the second phase of the project, a multi- dimensional sorting task was used to provide tests of the primary hypotheses. Color-Form Preference Test The Color—Form Preference Test essentially is a match— to-sample task in which there are no correct solutions. It was designed to test the relative color- or form-salience of the subjects. The test has been said to provide a measure of cognitive development, with color-salience characterizing a "primitive and impoverished mode of think- ing (Kay & Singh, 1977, p. 397).'' Construction. Fifteen unlined 3"x5" cards, on which had been affixed one of all possible combinations of five geometric symbols (square, circle, hexagon, rectangle, triangle) and three colors (red, blue, green),comprised the stimuli for this test. Each symbol was standardized in size (1%” from top to bottom) and centered on the cards; they were made of acetate (Chart-Pak CFOOS, CF055, CF103). 35 On each of 15 items, one of the cards was held by the subject and acted as the standard, while three others were placed in front of the client and acted as matching stimuli. One of the stimuli matched the standard on the basis of color and one on shape, while the third matched on neither color nor shape. The following rules guided presentation of the items: 1. Each of the 15 cards would be used as a standard once. 2. No pair of matching stimuli could ever appear twice; i.e., a red triangle and a blue square combination, for example, could only appear as matching stimuli in one item. 3. No color or symbol ever occupied a ''position" (client's left, right, or center) more than six times. 4. Color, form, and neither solutions occupied the three positions five times each. These rules were necessary to ensure that clients would not be mistakenly categorized as color- or form— salient on the basis of perseverant responses to the color, symbol, position, or patterns of stimuli. A complete list of the 15 items can be found in Appendix C. Administration. The standard was handed to the client and matching stimuli laid out in front to the left, right, and center. Testing was done in isolation. On each item, the testee was asked to look at the standard and to point to the one stimulus believed to match it. Two testees who objected that none of the cards matched were asked to pick 36 the card that matched the best. Complete verbal instruc- tions can be found in Appendix D. Scoring. An aggregate score of all color, form, and neither responses was computed. The testee was considered to be color— or form-salient if 10 or more consistent choices were made. A score sheet (Appendix E) was used to record responses. Standardization. Test-retest reliability (rxx) is of special importance in the study of traits. It has been reported as high as .98 for color-form tests (Melkman et al., 1976), Inasmuch as there is evidence that train— ing affects color-form salience (Suchman & Trabasso, 1966), 45 of the 57 clients were randomly assigned into one of three groups of 15 each. Ten of each group were from the experimental groups, and five from the remainder of the sample. One group (Group I) was retested before the sort- ing task, the second (Group 11) following it. In order to provide evidence that the results were not an artifact of the method of administration, the third group (Group III) was retested in a different way. Instead of being held by the client, the standard was placed in front of the client directly above the response opposite that which the client was categorized. That is, the standard was placed above the card which provided the color solution for form—salient clients, and vice-versa. In this way, an attempt was made to encourage the client to change cognitive style. Table 3.2 gives the results of retesting. 37 Table 3.2—~Test—Retest Reliability of the Color-Form Preference Test: Coefficients, Significance Levels, and Number of Clients rxx* Significance n Group I .9925 .001 13 Group II .5197 .028 14 Group III .8703 .001 15 7cretest interval of 2-6 weeks As can be seen in Table 3.2, rxx (.9925) was the highest for Group I clients who were retested prior to the sorting task. The substantial drop in rX for Group II to .5197 x was due, in part, to the responses of two persons who changed their responses on 10 of the 15 items. The effects of training may partially explain the drops in Group II and Group III. Any effect due to the different administra— tion to Group III clients is not reflected in the coefficient (rXX = .8703). Rationale for Construction. In surveying the litera- ture in color-form salience, it was discovered that there were no standardized tests of this type available. Most researchers have used specialized instruments, hand—made for their own projects. The only instrument that has ever approached being standardized on a similar sample was that of Kay (1975). His test, however, presented two (.AJ 57.) methodological discrepancies which precluded its use in the present project: 1. In the matching stimuli, the form and the neither choices were both the same color. This might encourage the client to just make the ”easier" choice, i.e., the color solution. 2. The scoring system was based on a Chi-square with 2 d.f. at an alpha level of .05. This would mean that a testee could make color responses to as many as seven items and still be categorized form. Because a test was needed that would discriminate well and would provide relatively pure color— and form- salient clients, it was decided that a specialized one would be constructed for the present project. Stroop Color-Word Interference Test The Stroop tasks provide three stable measures related to the cognitive functioning of the testee (Jensen & Rohwer, 1966). The most important of these is provided by the color-word score (CW); previously reported factor analyses using CW have yielded a heavy factor loading (.97) of what is believed to be the interference factor measur- ing the ability of the testee to attend to the relevant dimension in an identification task in the face of strong distraction. Construction. The Stroop tasks used in the present project were developed by Coker (1979). The test consisted of three 8%” x 11” vinyl—covered cards: the Word (W), Color (C) and Color—Word (CW) cards. Each card had a black background with 10 numbered rows of 10 stimulus strips 39 each, or 100 strips totally. Each row was underscored to aid the testee in keeping track of the row he/she was on. The stimulus strips on W consisted of the words "yellow," ”red," ”orange," "green," and "blue,” each printed in 18—point news gothic bold white print. The order of word presentation followed three rules: 1. No word followed itself. 2. Each word occurred twice in each row. 3. Each word was followed an equal number of times by the other four. The task on W was to read the words out loud. On C, words were replaced with groups of five x's in 18—point news gothic bold print in orange, red, blue, green, and yellow. The same rules as above were followed in the order of color-strip presentation. The task on C was to name out loud the colors of the strips of x's. On CW, stimulus strips were comprised of the words from W printed in the incongruent colors from C. For example, the word ”red” was printed in green. The task on CW was to name the color of the stimulus and not to read the word. For the above example, then, the correct respmnm would be ”green" and not "red.” The order of presentation on CW was critical to pre- vent unwanted influence of color or word patterns. To accomplish this, these five rules were used in construction: 40 1. No word was printed in a congruent color. 2. All words were printed in one of the four other colors an equal number of times. 3. Each word and color appeared twice in a row. 4. No word or color followed itself. 5. Each word and color was followed by the others an equal number of times. By design, the word order was identical to that on W, while the color order matched that on C. The result of this was that the effects of differential speed of pro— nunciation and recognition among the three cards were minimized, and the scores more accurately reflected true interference. A facsimile of the three cards is included in Appendix F. Scoring. Testees were timed, and their errors counted, for the present project. The interference score used in the design was given by the difference between CW and C in seconds (CW — C). Additional restrictions were placed on the use of subjects' interference score. Anyone whose CW — C score was less than or equal to zero, or whose number of errors on any of the three cards was greater than 49, was excluded from consideration for further participa— tion; there were 16 excluded, in all. The purpose of these restrictions was to include only those persons for whom a time score would be meaningful. Stroop score sheets are included in Appendix G. Administration. Since the version of the Stroop obtained for this study was vinyl—covered to preserve the stimuli through repeated use, it was necessary to 41 administer it in a room with a strong, indirect light source. When that was not possible, the cards were propped to reduce glare, or a pivoting clamp light was used to augment room lighting, where appropriate. Beyond the problems that lighting might induce, there remained the effects of other distraction. It should be remembered that the Stroop is intended to measure ability to overcome distractions which might be brought about by cognitive interference. The effects of environmental distractions were not known during this project, but some of them were procedurally controlled for in any case. More than 50% of the retarded subjects were believed to be distractible by people's voices, noise other than voices, and/or movement within the field of vision. To reduce the effects due to voices and movement, clients were isolated during the administration. It was not always possible to reduce ambient noise due to the layout of some of the sites, although the isolation mitigated its effects. Instructions (Appendix H) were read to the client for each test. Before starting CW, the client was allowed to practice up to four times on the first line. Clients were. allowed to point to keep their place, though the decision to point or not to point was uniform on all three cards. Standardization. While no attempt was made to assess the validity of the Stroop, test-retest reliability was computed for the two post—task groups after an interval of two to six weeks. The resultant coefficient (rXX = .7970, 42 significance = .001, n = 25) may be low due to the length of the interval for some of the retesting. Quick Test Uechi (1972) found that Stroop performance was nega— tively correlated with intelligence; persons with sub- normal levels of intelligence had longer interference times than persons with normal intelligence. The coefficient between mental age and interference times in the present study reflects such a relationship (rxy = -.4l85, sig = .001, n = 87). Furthermore, it was expected that a similar relationship would be found to exist between intelligence and ability to perform multidimensional sort- ing. Therefore, the Quick Test (Ammons & Ammons, 1962) was used to create blocks of clients with equal intelli- gence levels. The Quick Test is a standardized, individual intelligence test, covering abilities from 2-year to superior adult levels. Administration. Three forms of the Quick Test were used. On each form, testees read, or had read to them, words of increasing difficulty from a list of 50. The testee was asked to indicate which of four drawings best exemplified each word. Testing with a given form con- tinued until six consecutive correct and six consecutive incorrect answers were given, or until the list was exhausted. 43 Scoring. Each correct answer was credited and counted. The sum of scores for the three forms were con- verted to mental age and tentative IQ's using charts provided in the manual. Rationale for Inclusion. The Quick Test has been shown to be a useful instrument for choosing subjects for assignment to experimental groups (Zimmerman, Schroll, Ackles, Barret, & Auster, 1978). Test—Retest reliability of .92 for male and female retarded children was reported by Rotatori (1978). Previous research (e.g., Ogilvie, 1965) reported the correlation between Quick Test IQ and WAIS Full Scale IQ (FSIQ) as high as .80. A similar coefficient was looked for in the present study when FSIQ was available. While the resultant coefficient (rXy = .4104, sig = .004, n = 42) was unfavorable, it was of little consequence since all of the subjects were assigned using the same criterion, and errors were assumed to be randomly distributed. Further- more, 18 of the WAIS scores were over three years old; scores that old generally would be considered invalid. Counselor (Trainer) Questionnaire A survey of practices in habilitation facilities revealed no instrument which could be used to collect data on the work adjustment of clients in the present study. Most facilities use non-standardized, in-house rating forum with unknown reliability. Standardized forms, such as the 44 McCarron—Dial Work Evaluation System (Dial & Swearingin, 1976) contain few items that would have been of value irrthe present study. A short questionnaire, the Counselor (Trainer) Questionnaire (Appendix I),was developed, there- fore, to investigate the possible relationship between cognitive styles and work adjustment. Scale Construction. The Counselor (Trainer) Question- naire, while similar to other rating systems used in facilities, was by no means exhaustive of the domain of behaviors that have been observed on such instruments. It covered basic issues of work adjustment, such as quantity and quality of work, relationships with supervisors, etc. It was also used to collect demographic data On clients as well as clients' sources of distraction for use in standard- ization of the other instruments. Items used on the questionnaire were of four types: 1. Items l~ll required the counselor to rate the client against a hypothetical, competitively employed worker according to the following criteria: 1 = better than would be expected from a competitively employed worker 2 = equal to that of a competitively employed worker 3 = slightly below that of a competitively employed person 4 = a severe deficiency in this characteristic exists in this client This rating format was chosen to lessen the possibility of the client being rated according to a non—comparable local workshop norm. One item each was chosen to cover basic issues of 45 work adjustment. Others which were thought to correlate with the Stroop and Color-Form tests were included, such as response to training, resistance to distraction while working, and frequency of emotional outbursts. 2. The next nine items pertained to the sources of distraction to which a client appeared to be vulnerable while working. Counselors were asked to check as many as applied. A few of these items were people's voices, noise other than voices, hunger, thirst, etc. 3. The next two items asked the counselor to assign an overall rating to the client. Item 21 required the counselor to choose one of the following recommendations for the client: a. He/she is ready for competitive employment. b. He/she will be ready for competitive employ- ment within a year. c. He/she is ready for sheltered employment. d. He/she will require work activities for an extended period of time. A fifth choice--I cannot make a decision at this time——was never used and subsequently was dropped from analysis. Item 22 asked the counselor to compare the overall competence of the client with other clients in the program, and assign him/her to one of four ranges: top %, top % but not top %, bottom % but not bottom %, bottom %. Scoring. Since all of the questions were rated 1-4 in terms of descending degrees of work adjustment, the scoring system used was the arithmetic sum of all of the items, with the exclusion of the items on sources of dis— tractions. The resultant score was considered to be ordinal data in all calculations using it. 46 Standardization. To assess content validity, a panel of four vocational trainers or counselors were asked, independently, to review the questionnaire. Three judges had master's degrees in Rehabilitation, and one in Educational Psychology. All had a minimum of one year of experience at their present jobs (K = 1.95 years). All of the items were judged by all reviewers as being related to the content area of work adjustment, with the exception of the items on sources of distraction. Given the nature of the questionnaire and facility practices, limited types of reliability were applicable. Overall work adjustment would be expected to change over time, so test-retest reliability was irrelevant. Because only one person, the immediate counselor or trainer, would be familiar enough with the client, inter-rater reliability was not useful. Coefficient alpha (Cronbach, 1951), however, provided a conservative estimate of the expected correlation between the questionnaire and a hypothetical equivalent form. Coefficient alpha for 13 items on the questionnaire (1—11, 21-22) was .8326 (n = 72). Multidimensional Sorting Task In order to provide data pertinent to tests of the main hypotheses, a task which featured form and color relevance at alternate times was devised. The task con- sisted of two identical boxes of small parts which were to 47 be sorted into two cases of six compartments. Each box, or set, contained 72 pieces which could vary along three dimensions: shape (form), color, and size. Standard hex nuts, square nuts, and washers were used to provide variation in shape. A set consisted of 12 of the above pieces in two sizes, large (3/4”) and small (9/16”). Of those 12, there were four each in red, blue, and green (Deshler K22—R9, K22-C7, and K22-G6, respectively). These colors were chosen to approximate those used in the Color-Form Preference Test. Plastic parts boxes with six compartments each were used to sort the pieces. Both boxes were identical. Affixed to each compartment was a picture of the type of piece which was to be sorted into it. These pictures were made of red, blue, and green acetate (Chart-Pak CF005, 055, and 103, respectively), cut to match the shapes and sizes of the small parts. The six pictures used were those of a large green square nut, a small blue washer, a large red hex nut, a large blue washer, a small red square nut, and a small green hex nut, in that order. The following rules guided presentation of the stimuli: 1. Each color was to be used twice. 2. Each form-size combination was to be used once. 3. No form or color was to be adjacent to itself; i.e., two red pictures could not be side—by-side, nor could two square nuts, etc. 48 The purpose of these rules was to facilitate discrimination among the pieces, and to mitigate the effects of carelessness. Administration and Scoring. The two boxes of parts were emptied onto a table in front of the subject, one each just right and left of center. The compartments were placed in back of the two piles. The task was to choose a piece from one pile and put it into the compartment with the matching picture, move to the other pile and do likewise, and to keep alternating until all 144 pieces were sorted. Subjects were instructed to sort the pieces on the left side according to their color and size and the color and size of the matching stimulus. The shape of the stimulus served as a distractor. The right pile was sorted on the basis of shape and size, with color serving as distractor. Before proceeding, subjects were given a brief demonstration of the task, showing where each type of piece went. They were then allowed to practice on a random assortment of 24 pieces, in order to become habituated to the alternating system and the method of reinforcement. Correct responses were reinforced with praise or a continua tion response. When the client made a mistake, he/she was told to try another box. If that choice was incorrect, a verbal prompt, such as ”Shape and size," was given, and the correct compartment was pointed out. 49 A trial consisted of a complete sorting of 144 pieces, using the above—described corrective procedures. Error tallies were kept for each trial. There were three types of errors, presented below; 1. Pure color errors could be made on the left side, and entailed putting a piece of the wrong color into a compartment. For example, a small, red washer should, appropriately, be matched with the small, red square. If it were matched with the small, blue washer, it would be scored as a color error. Its equivalent on the right side was the pure shape error. Size errors could be made on either side, and consisted of choosing the correct color (or shape) but the wrong size. In the above example, matching the small, red washer with the large, red hex nut would be an error of size. ”Neither“ errors could be made on either side and were usually errors of carelessness or inability to attend to the instructions. Match— ing a large, green, square nut with a small, blue washer would be scored as a "neither.” Two additional errors, whose types were ambiguous according to the above, were scored as follows: Color- size errors were categorized as color errors, while shape—size errors were shape errors. To illustrate, sort- ing a small, green washer on the right side with the large, green square stimulus would be scored as a form error, rather than a size error. Only the first error for any single piece was scored. To counterbalance the effects of stimulus presentation, half of the clients in each group started with a color piece, and half with a form piece. Trials continued in this fashion until a criterion of one trial with nine errors or less was reached. While the normal error rate 50 of sorting tasks is much less, this rate (6.25%) was established due to the complexity of the task, on the advice of a vocational trainer with three years of experience. Procedure Four graduate students in Counseling, Personnel Services and Educational Psychology, and one undergraduate student from Social Sciences were recruited to assist in data collection during Phase One. Each was individually trained in the proper administration of the three instru- ments, and was required to demonstrate competence by completing one administration under the supervision of the principal researcher. Training also included procedures for recruiting subjects and obtaining informed consent. Each site was prepared by a personal visit prior to the actual dates of data collection. Preparation included securing the names of possible subjects, the names and addresses of legal guardians when required, and arranging space. In addition, each client was scheduled for a 45— minute period, and schedules were arranged to minimize disruption of program activities. Sites were visited sequentially rather than simul- taneously. In other words, all clients from one site were met and tested before moving on to another site. This was done to facilitate scheduling and minimize transportation costs. 51 Research assistants were used at only one site due to lack of space at the other three. When assistants were used, clients were randomly assigned to them. Each client was met individually during the 45-minute period. Researchers carried out the following steps at Phase One: 1. Met contact person at site. Distributed Counselor (Trainer) Questionnaire to the appropriate site personnel, giving instructions in both its com- pletion and return. 2. Met with each client. A statement regarding the purpose of the study (Appendix K) was read to each client, and an informed consent agreement (Appendix B) was obtained from each. In the case of one site, the statement of purpose and informed consent were combined (Appendix L). 3. Administered short screening test. The remain— ing selection tests were administered to clients who passed the screening test, in this order: (a) Color—Form Preference Test, (b) Quick Test, (c) Stroop Color-Word Interference Test. 11 instructions were administered orally. 4. Met non—retarded subjects. Administered screen— ing tests in this order: (a) Color-Form Preference Test, (b) Quick Test, (c) Stroop Color-Word Interference Test. In the three-day-interim between Phases One and Two, subjects were assigned to experimental groups based on criteria detailed elsewhere in this chapter. Clients were given numbers during selection in order to carry out blind assignment. Phase Two activities included retesting of the Color- Form and Stroop tests, as well as observation on the sort— ing task. Again, to facilitate scheduling and transporta- tion, sites were completed sequentially. Prior to data 52 collection, sites were contacted by personal visit or telephone to schedule clients. Clients who participated in retesting alone were scheduled for one 15-30 minute period. Clients selected for the sorting task and retest- ing were scheduled for two one—hour periods over two days. Only the primary researcher was available for Phase Two. The researcher carried out the following steps: 1. Met individually with each client. a. Retest-only clients: were given the Color- Form and Stroop tests in that order. Only Groups II and III took the Stroop tasks again. b. Experimental groups: Group I clients took the Color-Form retest, followed by 1% hours of training and observation on the sorting task. Groups II and III were trained and observed for 1% hours, finishing with retests of the Color-Form and Stroop tests. Two hours was deemed as the maximum time that a client should be asked to participate, in order to prevent fatigue and disruption of programming. Some clients were able to reach criterion in that time. Most clients completed two trials, two clients completed three, and two others completed only one trial. 2. Collected completed Counselor (Trainer) Questionnaires. Design A 2x2x2 multivariate factorial design (Salience x Interference x Order) was used across clients in the sort- ing task, with order having been a counterbalanced nuisance variable. Within each salience x interference cell, clients were matched on mental age and randomly assigned with regard to order of presentation. The dependent variables were color and form errors on the first trial of the task. Due to a procedural error, one client who should have been trained and observed color-first was trained and observed form-first. This did not affect the independence of the design, however, since at no time was he treated as color-first; the only effect was to increase the size of one cell. This particular design can be illustrated as in Figure 3.1. Sl Color-salient SZ Form-salient 11 Low Stroop interference time 12 High Stroop interference time 0 Color-first sorting 02 Form-first sorting M1 Color-type errors M2 Form-type errors Figure 3.1-~Design of the Study 54 The nature of the study must be regarded as explora- tory due to the small sample size, and generalizability will be limited, accordingly. Hypotheses The hypotheses which were tested in this study are as follows: H 1: There will be a main effect for both levels of salience; that is, color-salient clients will have significantly fewer errors when color is relevant to the solution on the multidimensional sorting tasks than when form is relevant, while form salient clients will have significantly fewer errors when form is relevant than when color is relevant. There will be a main effect for Stroop inter— ference time. Significantly fewer errors on the sorting task will be made by clients who have low interference times than those who have high interference times. There will not be a significant correlation between dimensional salience, as measured by the Color—Form Preference Test, and mental age, as measured by the Quick Test. There will be a significant correlation between color—form salience and overall work adjustment, as measured by the Counselor (Trainer) Questionnaire. 55 H5: There will be a significant correlation between interference time and overall work adjustment. Analysis of Data Before formal hypothesis-testing, descriptive statistics were used to summarize the demographic charac- teristics and the sources of distraction of the clients used in both phases of this project, as well as normal subjects employed in Phase One. Analyses of variance were used to assess the effectiveness of the blocking pro- cedures. Multivariate analysis of covariance was used to test the hypothesis regarding differential performance on the sorting task due to dimensional salience. Actual inter- ference time was used as the covariate. Because learning the task and habituation would very likely mask initial differences due to cognitive style, an a priori decision was made to analyze results of the first trial alone. The dependent measures were the number of pure color and pure form errors. The second hypothesis, regarding differential performance on the sorting task due to Stroop interference, was analyzed using multivariate analysis of covariance. The results of all trials combined were analyzed. Dependent measures were the same as above. The degree of the relationship between color-form salience and mental age (H3) was obtained with Pearson's 56 Product—Moment Correlation, since the data were interval and a linear relationship was assumed, if one existed. The correlation was obtained on the entire sample of 103 persons, and the two subsamples were used as comparison groups and computed separately. Mental age was defined as mental age in years from the Quick Test, and color-form by the percentage of form responses given by each subject on the pretest. If higher mental age was positively related with form salience, as past research has indicated, per- cent of form responses should increase as mental age increases. A one—tailed test of significance was used to test the correlation. The fourth hypothesis, regarding the correlation between color—form salience and overall work adjustment, was obtained with the Goodman-Kruskal Gamma and Pearson's Product—Moment Correlation. Color-form salience was defined as percent of form responses on the pretest; overall work adjustment was defined as the score from the Counselor (Trainer) Questionnaire. There was no convenient method of obtaining a correlation between interval and ordinal data, so the two were used to provide conservative (Gamma) and liberal (Pearson's) estimates. There is no test of significance for Gamma, but the Pearson's was tested using a two-tailed test of significance. The hypothesis concerning the relationship between Stroop interference and overall work adjustment presented special problems. It was desired to include the 14 persons 57 who had not been able to complete CW due to excessive interference. There was, however, no convenient time score that could be assigned them while retaining the meaning of the rest of the scores. It was decided, there- fore, to divide interference scores into ranks of one standard deviation in width. The lowest rank was assigned to the lowest interference times. The highest rank was arbitrarily assigned to those 14 who could not complete CW. This was appropriate since ordinal data do not involve assumptions of equal differences. Overall work adjustment was defined as the score from the Counselor (Trainer) Questionnaire. Because both sets of data were in ordinal form, the correlation was obtained using Kendall's Tau corrected for ties (Marascuilo & McSweeney. 1977) and tested with a two—tailed test of significance. A family alpha level of .05 was set for each test. Family alpha is used in exploratory research, when greater power is needed to find significant differences where they actually exist; it obviates partitioning of alpha for each hypothesis. Summary Mentally retarded clients in four facilities associ— ated with Michigan Association of Rehabilitation Facilities were used as experimental subjects in a factorial design to explore the relationships between cognitive styles and sorting task and work adjustment performances. Normal subjects from an available sample were also used to test 58 the relationship between cognitive style and mental age. All subjects volunteered for the project. Clients were assigned to experimental groups based on their performances on three tests: a Color—Form Preference Test, the Stroop Color-Word Interference Test, and the Quick Test. Most clients were employed in retests of the first two measures. Clients were also rated for overall work adjustment on the Counselor (Trainer) Questionnaire. Demographic data and information to be used in the standardization of the first two tests were collected on this form. Experimental clients were trained and observed on a multi—dimensional sorting task on which the color and shape of the object were, alternately, relevant to solu- tion. Dependent measures consisted of the number of errors made during each experimental condition. Two primary hypotheses concerned differential performance in the sorting task due to two aspects of cognitive style, dimensional salience and Stroop inter- ference. It was thought that color-salient clients would perform better when color was relevant, while form-salient subjects would do better when form was relevant. Clients with lower interference times were hypothesized to do better overall than those with higher times. The remaining hypotheses concerned the relationships among cognitive style, mental age, and work adjustment. First, it was felt that dimensional salience would not be 59 correlated with mental age. It was also thought that relatively higher form salience would be associated with more successful work adjustment performance. Finally, it was felt that lower interference times would be found for clients with better work adjustment ratings, Statistical analyses were used to test these hypotheses. Descriptive statistics were used to summarize the other data collected in the study. CHAPTER IV ANALYSIS OF RESULTS Five areas were addressed in the analysis of results of the project, listed below: 1. The first area concerns a summary of the demographic data collected. Included in demographic data are sex, age, Quick Test IQ, and Quick Test Mental Age when applicable. 2. The next area consists of a description of the sample on experimental variables, dimensional salience, interference, and work adjustment. Included in this section is an analysis of the four main experimental groups to determine whether or not blocking was successful. 3. Formal testing of the hypotheses explored in this study comprises the third section. 4. In the fourth area a summary of the responses to the Counselor (Trainer) Questionnaire is presented. 5. The final area concerns exploratory analyses of some of the data to determine the presence of systematic trends. Demographic Data A total of 73 mentally retarded from four habilitative facilities in Michigan took part in Phase One of this study. Of these, 32 were selected to participate in the second phase, based on their pretest responses. 60 61 The primary reason for which a client was excluded from the selection process for Phase Two was inability to complete the Stroop tasks. Three clients who were selected for experimental groups asked to be excused from participa- tion, as was their right. In addition, 30 non-retarded subjects participated in testing at Phase One, to provide a comparison group for Hypothesis 3. These subjects came from an available sample of high school graduates and college students. Table 4.1 shows the demographic characteristics of the two comparison groups. Experimental Characteristics A summary of the experimental scores for the sample-- dimensional salience, interference, and work adjustment--is presented in Table 4.2. Of the 16 retarded persons for whom an interference score was incomputable, 2 were due to the client's unwilling- ness or inability to read the Word card (W), after having passed the screening test. The remaining 14 had too many errors on the Color—Word card (CW) to compute a meaningful interference score. One Counselor (Trainer) Questionnaire was missing from the 73 that were distributed. This was discovered after data collection, and after the client had left the facility. No effort was made to secure that information. 62 Aa v m.m mma ob oma AN v a.e mNH on ONH Aoav m.mm sea on one Aeav s.ee sea on OOH Am V s.e AN V m.m as ou om AHHV m.ma mm on em AH v m.m Aoav s.ea as on as AHHV m.wa me en es ANHV o.o~ mm on om fleas m.mN as on oe mm1mmwm1mmmmm A4 v e.m - am Am C e.e mm on me AHNV w.w~ we on mm ANHV o.oe Ammo o.eN mm on aw Away o.oe AeNV m.Nm mm on as mwm AoNv s.ee Amev e.ae masses AOHV m.mm Ammo m.mm mam: xmm sz s sz s maeeeee> muommnsm popumummucOz muomeSm popumumm muoflasz 3mm paw mmwqumoMmm ”muoomnsm popnmumMIGoz paw pmphmuwm mo moflumfimouompmflo camamuwoamonu.a.q wHLmH AmNV 5.5m - o.mH Au V m.m fie V e.m o.mH oh o.eH as V N.w m.ma on m.mH AmHV w.sH o.ma on o.HH AOHV s.ma m.oa on m.w AANV 0.5m o.m on o.e AmHV m.sa m.m on m.m Amumowv mw< Hmucmz ude on30 AzV x AzV N muommnsm pmphwummucoz muomnnsm popumumm maflmfluw> .emseaueoo--.a.e eaeee 64 AeHV m.HN maeeeeaeoeee ANV N.NN ANV o.oe AHHV H.mH - owe AeV e.ee AHV 0.0N Aw V o.HH e.meH on mNH AH V m.m AmV m.mm xNV o.oe AQHV s.ma e.eNH on OOH AN V o.o~ AaV o.o~ Am V m.NH m.mm on ma AeHV s.ee Am V o.om AHV o.o~ AOHV s.mH m.ea on om AmHV m.me Am V 0.0m AmV o.oe as V N.w m.me on mN Am V s.e AN V o.om Am V H.e a.eN on o Awpmoommv momwswmumumH Aw V 0.0N Homufloz AmNV s.es ARV o.ooa AoHV 0.00H AHeV e.mm shoe AH V m.m AmV 0.00H AmV c.ooH ANHV e.eH eoaoo toeeaaem HeeoamemEHQ AzV x AzV s AzV x AzV s AzV s AzV s eaeeeee> pmpnmumuucoz Q mmswuo Hmummefiummmm < popumumm HH< mHaEmm emu mo mofluwflnmuomumno Hmunmafluwaxm-n.m.q mHQMH 65 AC Hi: 3 V 1V4 wfimmflz Am V e.H - Hm AaV H.HH AHV o.oa AaV o.o~ AHV o.o~ Am V «.me em on He AeV s.ee AmV o.ow AeV o.ow AmV o.oe ANmV N.Ha oe on Hm “HV H.HH AaV o.oa AaV o.om AoHV s.ma om on o nemeehehee see: AzV s AzV a AzV s AzV s AzV s maeeeee> o o m a mmbono aficmfiuomxm popHmumu H2 Bingo: .N .q manna. 66 Analyses of variance were performed to determine whether or not there were any differences among the experi- mental groups on mental age and interference scores, and to determine the efficacy of blocking procedures. Figure 4.1 presents the estimated cell means for the two variables used in this analysis. There were no significant differences on mental age between the groups (F3,25 = .3786, significance = .7692), indicating that equating on mental age had been effective. There were significant differences on interference scores, as desired (F3,25 = 10.8, significance = .0001). Scheffe procedures were carried out to investigate where those differences lay, and it was discovered that the two color-salient groups were significantly different, but not the two form-salient groups. Color-Salient Form Salient KWA==9.2 years EMA==8.6 years low Interference XINT = 46.2 seconds XINT = 49.1 seconds r1= 5 n==10 EMA==7.23mmrs EMA==8.6)mmrs High Interference iINT = 215.8 seconds XINT = 139.5 seconds I1= 5 r1= 9 Figure 4.l.-—Estimated Cell Means for the variables Mental Age and Interference Scores for the Four Experinental Groups 67 Tests of Hypothesis This section consists of the results of statistical tests relevant to tests of each of the five hypotheses. Hypothesis 1. Null hypothesis: No differences will be found in color and form errors between color- and form— salient subjects, in a multi—dimensional sorting task. Alternate hypothesis: Color— and form-salient clients will show differential performance on the sorting task. Multivariate analysis of covariance was used to test this hypothesis. Since blocking for interference was ineffective, actual interference scores were used as the covariate. Table 4.3 contains the calculated F—ratios, significance of F, and the degrees of freedom used in the calculations. A Statistical Package for the Social Sciences (SPSS) multivariate technique was used because of correlated dependent variables and the unbalanced design. SPSS performs an exact least—squares test of the hypothesis adjusting for the unbalanced design. As presented in Table 4.3, the null hypothesis of no differences can be rejected. 68 Table 4.3.--Sunnary of Mhltivariate Analysis of Variance for Test of Hypothesis 1 Degrees of Source Freedom F—ratio Significance Salience 2,19 5.6576 .0118* Interference 2,19 .4280 .6580 Order 2,19 1.3601 .2805 Salience x Interference 2,19 1.5009 .2482 Salience x Order 2,19 3.2318 .0619 Interference x Order 2,19 .0665 .9359 Salience x Interference x Order 2,19 .3813 .6881 7CIndicates those F—ratios significant at the .05 alpha level or less Figure 4.2 provides the estimated cell means for color and form errors across the eight groups. To investigate where the differences were, the eight cells were collapsed into two, corresponding to the two levels of salience, and univariate analyses were performed on the two dependent measures. Table 4.4 gives the results of those analyses. Table 4.4.--Sunnary of Univariate Analyses of variance for Hypothesis 1 Degrees of Measure Freedom F—ratio Significance Color errors 1,20 5.6743 .027“ Form errors 1,20 5.4209 .0317' xIndicates those F-ratios significant at the .05 alpha level or less 69 Table 4.4 indicates that there were significant differences in both measures. Combining cell means showed that color-salient clients had fewer color than form errors, while form-salient subjects had fewer form than color errors. 31 32 01 O2 01 02 M1 14.0 4.0 6.2 8.5 11 M2 8.0 12.4 4 0 8 3 n = 1 n = 4 n = 6 n = 4 M1 7.0 3.3 9.0 10.8 I2 M2 2.8 6 5 1 4 5 0 n — 2 n = 3 n = 5 n = 4 Legend: 81: Color-Salient S2: Form—Salient 11: low Interference 12: High Interference 01: Color-first Sorting 02: Form—first Sorting 1%: Pure Color Errors M2: Pure Form Errors Figure 4.2 ——Estimated Cell MEans for Pure Color and Pure Form Errors in the First Trial of the Sorting Task for Eight Experimental Groups 70 Finally, eta (Marascuilo & McSweeney, 1977) was com- puted, giving a measure of the relationShip between the independent and dependent variables. The estimated degree of relationship between dimensional salience and color errors was .6597, while between salience and form errors, it was .7009. This meant that dimensional salience accounted for 43.5 and 49.1 percent of the variance of color and form errors, respectively, given by eta-squared. To summarize, the null hypothesis of no differences was rejected. Substantial differences according to level of color-form salience were found. Hypothesis 2. Null hypothesis: No differences in sorting will be found between low and high interference levels. Alternate hypothesis: There will be differential performance according to interference levels. Multivariate analysis of variance was used to test this hypothesis. This test was thought to provide the most powerful test for hypotheses on correlated means. Table 4.5 shows the results of this analysis, giving F—ratio, degrees of freedom, and significance. As indi— cated, there were no significant differences between the interference groups; therefore, the alternate hypothesis cannot be supported. There was, in fact, a significant main effect for salience which accounted for most of the differences. 71 Table 4.5.--Suunary of Mhltivariate Analysis of variance for Test of Hypothesis 2 Degrees of Source Freedom F-ratio Significance Interference 2,20 .8208 .4586 Salience 2,20 3.8641 0381* Order 2,20 1.4564 .2568 Interference x Salience 2,20 .7144 .5015 Interference x Order 2,20 .0133 .9862 Salience x order 2,20 2.1767 .1395 Interference x Salience x Order 2,20 .3024 .7024 4 Indicates F-ratio significant at the .05 alpha level or less Hypothesis 3. Null Hypothesis: There will be no correlation between mental age, as given by the Quick Test, and dimensional salience, as given by the Color-Form Preference Test. Alternate Hypothesis: There will be a positive rela— tionship between dimensional salience and mental age. Since it was, in fact, desired to reject the alternate hypothesis and to fail to reject the null, the most power— ful means of testing the hypothesis was used——a one—tailed test of Pearson's Product-Moment Correlation. Table 4.6 gives the results of the tests of the correlations taken on the entire sample, as well as the two comparison groups. 72 two comparison groups. Since mental age is meaningless for the non—retarded sample, the correlation between salience and Quick Test IQ is presented. As Table 4.6 indicates, there is no significant cor- relation between mental age, or IQ, and salience, when it is taken on the entire sample. The null hypothesis is rejected, however, for the correlation taken on the retarded comparison group (rMA,CFP = .2025, Significance = .043, n = 73). Table 4.6.—- Pearson's Product—Moment Correlations for Tests of Hypothesis 3. Group rMA,CFP rIQ,CFP Significance n Entire Sample .1422 .0763 103 .0932 .1910 90 Retarded .2025 .0430" 73 .1526 .1220 60 Non-Retarded --- --- --— .1009 .2980 30 7"Indicates correlations significant at the .05 alpha level or less Hypothesis 4. Null hypothesis: There will be no cor- relation between dimensional salience, as measured by the Color-Form Preference Test, and work adjustment, as measured by the Counselor (Trainer) Questionnaire. 73 Alternate hypothesis: There will be a significant correlation between dimensional salience and work adjust- ment. Goodman's and Kruskal's Gamma and Pearson's Product- Moment Correlation were used to provide a conservative and liberal estimate, respectively, of the relationship between the two variables. Gamma is used when both variables are ordered categories, while Pearson's is used when the variables are interval data and the relationship is expected to be linear. There is no test of significance for Gamma, but the Pearson's was tested using a two-tailed test of significance. The results are presented in Table 4.7. As indicated, the resultant coefficients are not large enough to reject the null hypothesis. Gamma (.0954) shows an inconsequential relationship between the two variables, which is echoed by the Pearson's. Table 4.7.-—Summary of the Correlations for Tests of Hypothesis 4. Statistic Coefficient n Significance Gamma. .0954 72 -—- Pearson's Product-Moment .0277 72 .817 74 Hypothesis 5. Null hypothesis: There will be no correlation between interference scores, as measured by the Stroop Color—Word Interference Test, and work adjust— ment, as measured by the Counselor (Trainer) Questionnaire. Alternate hypothesis: There will be a significant relationship between interference scores and work adjustment. Because both sets of data were ordinal, Kendall's Tau with correction for ties, tested with a two—tailed test of significance, was used to provide estimates of this relationship. The resultant coefficient (Tau = .0545, significance = .499, n = 72) shows a very weak relationship; therefore, the null hypothesis cannot be rejected. Counselor (Trainer) Questionnaire The responses to the Counselor (Trainer) Question- naire were collated and are summarized in Table 4.8. For ease in format, the items on sources of distraction were removed and the items were presented separately; all items are abbreviated. There were 72 returned questionnaires out of 73 distributed. The remaining one was for a client who left programming just following data collection, before the discovery was made. 75 .wqflmmde 4 new mouwmmCONummDv teaspoon Nu no pommmx Am4Vw.s4 AemVo.oe 4e4Vo.eN 4o VN.w metronome mo soeoeooee .44 444Vm.mm A4eVN.om 4N4Ve.o4 AN V4.4 eo4oomeom4e on ooeeoo4oor .04 As Vo.m Aw V0.44 ANmVN.4s Aw V0.44 ooemeeooo< .m 4m Vm.o AmmVN.me A4me.Ne Am V4.4 4044eenoese .w Am V4.4 44eVN.om AwNVe.wm we4e4eeo on ooeoeoom .4 444Vm.m4 AmeVo.4o Am4Vw.s4 memes o>4n4ooeoh-eoz .o AN V4.4 AemVo.oe ANme.me As Vm.m oxoeu osao4ooeem .m Am Vm.N4 Aome.ee ANNVo.sm whoruoeroo e043 oeo4oo4om .e 4m Vm.o AoeVm.em AoNVe.mm x4 Ve.4 eom4>eoesm e043 oeo4oe4om .m Am V4.4 AseVe.eo 4NNV4.om rues 4o 4044650 .N 4w4Vs.eN 44me.eo Am V4.e r463 we so4eeeoo .4 sz a N sz m N AZV N N AZV N msoom Deoumou wmfiumm wHHmCGOHumoDO AuochMHV hoawmcsoo on momcoamomuu.w.e mHDmH run 76 memmHE H can moHHMCCONuman pmcnnumn NN mo powmmk AemV o.oe ArmV 4.4m eoeoo .om AwNV e.wm AeeV m.oe opener ooeeeeeoooeeH .m4 Aw V 0.44 AeoV 4.4w ome4ee .w4 4m V m.N4 AmoV m.ow homage .s4 4w4V s.eN AemV o.es oesoeeoeeos .64 404V s.m4 446V o.ew we4oem44 .m4 448V w.sm AomV 4.44 eo4os> e4eo43 oeoem>oz .e4 x4eV N.om A4mV m.Ne moo4o> teen porno oo4oz .m4 4mmV o.ms 4o4V o.eN ooo4o> m.o4eooe .44 AZV II N AZV III N mCOHuomHumHo oz mos omfioamom oeso4oeoo--.w.e 64469 u.» 77 .ooHBu xsmao amoH mQB Eoufl wash 0.» u.» wdflwwfifi 4 paw moHHMGEOHumoSU pocusuou Nu mo pommm v... Ae Vm.m 404Ve.44 444V4.me 4o4Vo.44 emcee .44 44 V4.4 Aome.wo 4e4V4.44 Ao V4.4 treo4ooeeoesooom .44 42V 4 42V 4 42V N 42V 4 osooe octocoo e m 4 4 wcHumm eoee4oeoo--.w.e 64484 w.» 78 Systematic Trends A post-hoc analysis of the data was performed to deter- mine whether differences among the four experimental groups could mask a main effect for interference, or confound main effects for dimensional salience. Sgp. While a chi-square with 3 degrees of freedom showed that sex was not nested (chi-square = 3.23, significance = .358), it was possible that sex differences among the groups could account for some of the differences in performance of the sorting task. To investigate this, color and form errors were categorized by sex within the four groups. Means of the sexes were calculated and are presented in Table 4.9. Table 4.9.--Estimated Means for Color and Form Errors of Each Experimental Group Categorized by Sex MEan Color Errors Mean Form Errors Group Male Female Male Female A. Color—salient/ low-Interference 4.0 7.3 11.5 11.3 B. Color-salient/ High-Interference 0.0 6.0 4.0 7.3 C. Formrsalient/ low-Interference 13.0 5.6 10.5 2.8 D. Form—salient/ High-Interference 8.8 11.0 2.0 4.3 79 As can be seen in Table 4.9, there were sex differ- ences within the cells. To investigate whether there might be a sex-by—independent variable interaction, the groups were collapsed and their means graphed. Figure 4.3 contains the graphs for color and form errors, first by combining Groups A with C and B with D to illustrate sex by interference, then by combining A with B and C with D to illustrate sex-by-salience. Figure 4.3 shows the possibility of sex by independent variable interactions for both variables. High-inter— ference males performed better than low-interference males (a3 and a4), while females performed about equally in both groups (a1 and a2). Color-salient males did better than form-salient males (b3), while females stayed about the same, with regard to color errors only (bl). The reverse is seen with regard to form errors; color- salient females had more errors than did form—salient females (b2), while males performed about equally (b4). To investigate further the possibility of interaction, a multivariate analysis of variance was performed, substi— tuting sex for order of presentation as the nuisance variable. Table 4.10 contains the resultant F—ratios, significance levels, and degrees of freedom. 80 a1) Female color fl . 0 -.__~ a2) Female form 11 n: 5 a3) Male color \ a4) Male form A,C B,D Groups a) Sex4by-Interference b3) Male color bl) Female color a . 3 b4) Male form e1 1 \ e; \ b2) Female form A,B C,D Groups b) Sex—by-Salience Figure 4 3.——Graphs of Mean Color and Form Errors to Illustrate Sexeby—Independent Variable Interactions 81 Table 4.10.-— Summary of Multivariate Analysis of Variance for Tests of Sex-by-Independent Variable Interaction Degrees of Source Freedom F—ratio Significance Salience 2,20 3.1220 .0627 Interference 2,20 .6887 .5138 Sex 2,20 .1881 .8300 Salience x Interference 2,20 1.0021 .3848 Interference x Sex 2,20 2.9556 .0751 Salience x Sex 2,20 2.6096 .0984 Salience x Interference x Sex 2,20 .7203 .4988 Table 4.10 shows the possibility of sex-by—independent variable interactions with the four groups intact. Both interference and salience appear to be vulnerable to sex interactions. Agg. There was also a possibility of the presence of age—by-independent variable interactions. To investigate this, subjects in each cell were classified into one of two age levels, according to whether they were below or above the median age for the four groups (33.5 years). A chi— square with 3 degrees of freedom was calculated, and showed that the groups did not differ on age (chi-square = 2.96, significance = .396). The means of the age groups were calculated and are given in Table 4.11. 82 Table 4.11.--Esthmated Means for Color and Form Errors of Each Experimental Group Categorized by Age Mean Color Errors Mean Form Errors Group 20-33 34-69 20-33 34—69 A. Color-Salient/ Irmhlntenfinence 7.0 4.5 5.0 21.0 B. Color-Salient/ Highéhnrrference 4.5 5.0 5.4 1.7 C. Form-Salient/ Iow-Luxnference 8.7 3.3 3.0 9.0 D. FormrSalient/ Highéhmenfinence 8.3 10.5 .7 4.2 As can be seen in 4.11, there were age differences within the cells. To investigate the possible inter- actions, the groups were collapsed, as before, and means of the age groups graphed. Figure 4.4 includes the graphs for color and form errors, first by combining A with C and B with D to illustrate age by interference, then A with B and C with D to illustrate age-by-salience interactions. Figure 4.4 represents a possible age by interference interaction with regard to color errors; low—interference, younger subjects performed poorer than did high—inter- ference, younger clients, while the reverse is seen for the older clients (bl and b3). The interaction with regard to form errors is due, primarily, to the steep gradient for older clients. Low-interference clients performed much 83 \ a1) 34-69 color a3) 20-33 color Mean Errors I . \ TTTTT lca4) 20—33 form -- a2) 34—69 form -L. A.C B,D Groups a) Age-by-Interference \. bl) 34—69 color b3) 20-33 color b2) 34-69 form / lban Errors ‘~ b4) 20—33 form i % A,B C,D Groups b) Age-by-Salience Figure 4.4.--Graphs of Mean Color and Form Errors to Illustrate Age—by—Independent Variable Interactions 84 poorer than did high interference clients (b2 and b4). There is no age by salience interaction illustrated with regard to either variable. To investigate these interactions, age replaced order of presentation in a 2x2x2 multivariate analysis of variance. Table 4.12 shows the F-ratios, degrees of free— dom, and significance levels for that test. Table 4.12.--Sunnary of Multivariate Analysis of Variance for Test of Age—by—Independent Variable Interactions Degrees of Source Freedom F—ratio Significance Salience 2,20 5.2977 0143"" Interference 2,20 1.2630 .3439 Age 2,20 2.6062 .0987 Salience x Age 2,20 3.5550 0478* Interference x Age 2,20 1.1941 .3236 Salience x Interference 2,20 1.9108 .1740 Salience x Interference x Age 2,20 2.8983 .0785 7': Indicates F-ratios significant at the .05 level or less Table 4.12 shows the possibility of a salience—by—age interaction, with interference-by-age disappearing when dm groups are kept intact. Work Adjustment. Sex and age differences were investi- gated with regard to overall work adjustment to determine whether they could account for the failure to find a cor- relation. A chi-square was performed for each test. 85 dividing age at the median (34.7 years). The results of each analysis are given in Table 4.13. Table 4.13.—-Sex and Age Differences in Overall Work Adjustment Variable Chi-square Degrees of Freedom Significance Sex 17.41 20 .607 Age 21.79 20 .352 As can be seen, there is virtually no relationship between either age or sex and overall work adjustment which could account for the failure to reject the null for either H4 or H5. Individual items on the Counselor (Trainer) Question- naire were analyzed to determine if any of them correlated with dimensional salience or interference. Kendall's Tau with correction for ties was computed for each item. Table 4.14 presents the results of that analysis. As noted, no items correlated with interference. Only one item, quantity of work, even approached signifi- cance (Tau = .1087, significance = .089). There is a strong likelihood that this is due to chance. Three items correlated with dimensional salience, performance of repetitive tasks (Tau = .1593, significance = .024), recommendation (Tau = .1619, significance = .023), and range (Tau = .1576, 86 0064 Ho Ho>m4 usaam mo. um unmoflwflmw4w mdbflumHoHHoo moumonGHx 004. 4400.- 44 0040. 0404.- 60660 .44 000. 0000. 04 6040. 0404.- 66466066666660 .44 400. 0440.- 44 040. 0040. 060466666 46 466660666 .44 004. 4000. 44 000. 0440.- 66466666040 66 6666604060 .04 440. 0040.- 44 440. 0044. 6666066660 .0 004. 0000. 44 000. 0400.- 46446666660 .0 044. 0040. 44 400. 0040.- 06464646 06 60666060 .4 404. 0000. 44 040. 0000.- 06406.6 64446466664602 .0 004. 0000. 44 .4040. 0004. 064066 6444646644644 .0 444. 4040. 44 444. 0000.- 0464663-66 6643 066466460 .0 004. 4000. 44 000. 0000.- 4604446660 6643 066466460 .0 404. 4040. 44 000. 0444. 4463 46 4644660 .4 000. 4004. 44 404. 0040.- 4463 06 46466660 .4 666664446040 664 2 666664446040 664 6664 mug-me-kuC-H $04-84.“me mmuoom Ao-BDV oocmuowumucH new monoHHmm HmmO4mcm54o £403 oufimmc04umono Aeomwmuhv moHommDoo co mEmuH Hmnp4>HpGH mo msowuwamuuoo-u.04.0 oHDmH 87 significance = .026). Two other items, quality of work (Tau = .1118, significance = .083) and attendance (Tau = .1186, significance = .071), approach the signifi— cant level. Better performance of repetitive tasks, higher quality of work, and better attendance appear to be associated with color-salient clients, while form-salient subjects were given higher recommendations and rated better against group norms. Summary of Results 1. There were no significant differences among the experimental groups on mental age. 2. There were significant differences between color- salient/high—and-low-interference groups, but not between form—salient groups, with regard to interference scores. 3. There was a main effect for dimensional salience on the multidimensional sorting task. Color-salient sub- jects performed better when color was relevant to solution; form—salient clients did better when shape was relevant. 4. There was no difference between high and low interference levels on performance of the sorting task. 5. There was no correlation between mental age and color-form salience for the entire sample, but there was a small, but significant positive relationship for the retarded comparison group. 88 6. There was no correlation between work adjustment, as measured by the Counselor (Trainer) Questionnaire, and dimensional salience. 7. There was no correlation between work adjustment as measured by the Counselor (Trainer) Questionnaire, and Stroop interference. 8. A variety of analyses were performed to explore possible significant trends in the data: (a) There is a possibility of sex-by—salience and sex—by—interference interactions on the sorting task; (b) there is a possi- bility of age-by-salience interactions on the sorting task; (c) there are no sex or age differences in overall work adjustment; and (d) some items on the Counselor (Trainer) Questionnaire correlate with salience better than others, while interference does not relate to any item. CHAPTER V Summary and Discussion Summary Problem. The purpose of this study was to explore the interaction of two aspects of cognitive style-- dimensional salience (Odom & Mumbauer, 1971) and Stroop interference (Dyer, l973)——and their effects on performance of a multidimensional sorting task among mentally retarded persons. Another purpose was to relate the two aspects of learning style with overall work adjustment. The questions to be answered were: Is there differential performance on the task across the groups? Can these differences be attributed to effects of cognitive style? Does the evi- dence support the proponents of a cognitive set theory of cognitive style rather than those of a developmental stage theory? Can differences in work adjustment performance be attributed to learning style? Most researchers would agree that the evolution of a cognitive style is a developmental process. Research in both of the aspects studied in the present project, how- ever, has left some rather basic questions either unanswered or with conflicting answers. First, there is a question about the degree to which specific learning experience and maturation account for the progression of 89 90 dimensional salience. Some researchers (e.g., Piaget, 1952) think that the progression is a universal, ontogene— tic process, while others (e.g., Gaines, 1970) feel that salience is a cognitive set which is formed by, and can be altered by, specific learning experiences. There are also questions about whether differences in both dimensional salience and Stroop interference can lead to differences in other tasks. The research here, too, is conflicting. Some researchers (e.g., Gaines, 1964) have found that dimen- sional salience is related to performance of tasks which feature preferred and non-preferred dimensions; others (e.g., Marantz, 1976) have found no relationship. The same controversy exists for the Stroop tasks. The evidence is equivocal about whether Stroop interference is associated with differential performance on other tasks. In the present study the aim was to combine the two aspects of learning style and look at their effects on task performance and overall work adjustment in a mentally retarded population. This would provide valuable informa— tion for the controversies on cognitive style, but, pri— marily, it would be beneficial to know for habilitative programming for the retarded. By showing that cognitive styles affect work adjustment performance, the results would promote the individualization of habilitation plans, a process desired by many, but practiced by few. 91 Design. Mentally retarded clients from Michigan habilitative facilities were used as subjects in a factorial design to explore the relationship between two aspects of cognitive styles, and sorting task and overall work adjust- ment performance. Non-retarded subjects from an available sample were also used to test the relationship between cognitive style and mental age. All subjects were volunteers. Clients were assigned to experimental groups based on their performances on three tests: a Color—Form Preference Test, the Stroop Color-Word Interference Test, and the Quick Test. They were also observed by their counselors using the Counselor (Trainer) Questionnaire, designed to collect information about their overall work adjustment. Experimental clients were trained and observed on a multidimensional sorting task on which the color and shape of the object were, alternatively, relevant to solution. Dependent measures consisted of the number of errors made during each experimental condition. Two hypotheses related to the effects of the two aspects of cognitive style on sorting performance. It was thought that color-salient clients would perform better when color was relevant to task solution, if resistance to cognitive interference was controlled for, and that form- salient clients would perform better when form was relevann under the same condition. It was also hypothesized that 92 persons with low Stroop scores, scores which measured resistance to cognitive interference, would have fewer errors than those with high scores. The remaining hypotheses concerned the relationship of cognitive style to mental age and work adjustment. It was hypothesized that dimensional salience would not correlate significantly with mental age. Finally, it was felt that different levels of dimensional salience and Stroop interference would be associated with varying degrees of work adjustment success. Results. The hypotheses regarding differential performance of the sorting task according to dimensional salience was supported. Clients who preferred color in the pretest had fewer errors when color was relevant to task solution than when form was relevant. Form-salient clients showed similar tendencies on the task; they had fewer shape errors than color errors. Practical signifi- cance was relatively high; dimensional salience accounted for about half of the variance in performing errors. The hypothesis concerning similar differential per: formance on the task according to interference level was not supported. There were no significant differences between the subjects in low and high interference groups. Post—hoe analyses of the data to determine signifi- cant trends revealed what may be complex interactions of both independent and dependent variables with sex and age. 93 Interference level appeared to affect males more than females, in the opposite way, however, than what was originally thought. High-interference males performed better than low-interference males. Sex-by—salience interactions were different with each dependent measure. Color-salient males had fewer color errors than form- salient males, but the two groups were equal with regard to form errors. Form-salient females had fewer shape errors than did color—salient females, but the two groups were equal with regard to color errors. Similar complex inter- actions may exist with regard to age. Color—form salience was not related to mental age when the correlation was taken on the entire sample. However, there was a small, but significant positive rela— tionship for the retarded sub-sample. Neither dimensional salience nor interference was significantly related to overall work adjustment, nor were there any sex or age differences on that factor. Some individual items on the Counselor (Trainer) Questionnaire were related to color-form salience. Color-salient clients were rated better on performance of repetitive tasks, quality of work, and attendance, while form—salient clients were given better recommendations and higher ratings against local norms. 94 Discussion Before a formal interpretation and discussion of the results is undertaken, a brief overview of the general utility of the study is in order. Exploratory research is used when little is known about treatment effects, and when the population from which the sample is drawn is limited. Its goal is not so much to generalize to a larger population, but to exercise experimental control over the variance, thereby making complete descrip- tions of the outcomes possible. By doing this, the problems associated with a certain line of research may be discovered, so that future researchers can try to overcome them. A cautious approach, therefore, will be taken in the discussion of the findings and implications of the present study. However, discretionary use of the Bridge Argument of Cornfield and Tukey (1956), which allows generalization of results back to an hypothesized population with similar characteristics, is not inappropriate. Should future researchers wish to use the Bridge, it is hoped that the description of the sample provided in the present study will suffice. That color— and form—salience is associated with differential performance of the sorting task has been seen. Color—salient clients clearly had fewer color than shape errors, and form-salient clients did better when shape was 95 relevant to task solution. This echoes Gaines (1964), Melkman et al. (1976), and others whose research produced similar findings. But what of the relevance of this issue to current habilitative practice? Even narrowly defined, color- form salience has its implications. First, sorting tasks, such as the one used on this project, are typical of the jobs that retarded persons perform in facilities. Currently, everyone is trained to do a job in the same way, that way being to sort according to gradations in some dimension, such as size, color, or shape. What the present study has given evidence to is that the inability of some clients to perform well may not be due to their lack of dis— crimination, but to their tendencies to attend to some dimension other than the one that is specifically relevant to task solution. By knowing the color-form salience of non-performers, trainers might be able to increase discrhn— ination of one dimension by adding, and later fading, a redundant preferred dimension. Or if a job allows dis— crimination on several dimensions, a trainer would need, only to shift instruction to a preferred dimension for a particular client. Second, there are commercialized vocational evaluation systems which use similar color-form sorting tasks to hypothesize about the overall job readi- ness and vocational potential of clients. This practice 96 may become suspect in light of these findings: that people attend differently to those dimensions, and that this has little to do with the overall cognitive development of the client. Furthermore, these tests have been normed on groups whose color-form salience was unknown. If the group had predominate preference for one dimension, e.g., form, then color preferrers are being compared against an unrepresentative norm group. It would seem that developers of evaluation systems need to consider the learning styles of their target and norm groups in construction. One way would be to include people of different levels of salience in norm groups; another would be to provide different methods of administration for clients with different learns ing styles. The foregoing argument must be used cautiously, how- ever, in view of the small sample size of the current study, and, especially, in light of the complex inter- actions associated with sex and age. Interactions such as these have plagued researchers since cognitive styles were first studied, and the present project lends no exception. That sex and salience interacted differently with each dependent measure, that three—way interactions were noted between age and the two independent variables, and that only a slight association was found between salience and more conventional measures of cognitive development, all indicate that the progression of dimensional salience, and 97 its effects on performance, are neither just an invariant and universal ontogenetic process, nor just the result of past training and experiences. This supports previous researchers (e.g., Gaines, 1970) who noted cultural, physiological, and socio-economic differences in saliences. This would also explain discrepancies in the findings of past research on the effects of cognitive style. Few researchers have controlled sex or age, though Odom and Mumbauer (1971) did note age differences in performance. In a broader sense, color—form salience may have relevance for work adjustment itself. While no associa— tion was found between overall work adjustment and dimensional salience, a serendipitous finding was that certain aspects of work adjustment may be associated with it. Perhaps the failure to find a relationship was due to viewing work adjustment as a single entity, just as it would be inappropriate to call any one phenomenon cogni- tive style. Examination of those items which show promise may be enlightening. The three items which were most closely associated with color—salience--performance of repetitive tasks, attendance, and quality of work——have a singular theme, in that persons rated highly on them might also be described as having a rigid sense of order, or as paying close attention to detail. Perhaps Kay and Singh's (1975) finding of a ”primitive and impoverished mode of thinking" 98 (p. 397) among chronic schizophrenics can be translated into inflexibility or compulsiveness among persons without pathological processes. Indicative of the possibility that color-salient clients are not performing as well in programs as do the form—salient, is that they received lower ratings on nearly every other item, specifically on the recommendation and rating against local norms. Color—salience represented a variation from the norm in this study; only 12 of 73 clients and 1 out of 30 non-retarded subjects had color preferences. Since training strategies are developed for the average client, the possibility exists that the strate- gies were developed for form-salient clients by form- salient trainers. It is not unlikely that clients who were not producing as much as others, or who had poorer response to training, would be down-graded in other areas, as poor performance created frustration in both client and trainer, poor relations between these clients and others, and so on. This is speculative, of course, in view of the small relationships between salience and even the most signifi- cant of the items. In no way should one conclude that color-salience causes poor work adjustment, or that it is the only thing associated with adjustment. The findings do show, however, that salience could be one of the con— tributors to work adjustment success, either directly, or in interactions with factors beyond the scope of the present study. 99 The failure to support hypotheses concerning Stroop interference could be due to any of four reasons: (1) that the cognitive interference involved is only task- related and has nothing to do with performance on other tasks or work adjustment; (2) that the score used for categorizing clients as low or high interference was inappropriate; (3) that, among form-salient clients in the sorting task, there were no clearly defined high—and-low- interference groups; and (4) that the effects of inter- ference are seen only in interaction with other factors. Each reason warrants further comment. It is possible that Stroop interference has very limited effects which are task—specific and which do not generalize to other types of cognitive functioning. The body of literature reviewed for this project offers no direct statement on this possibility. Most Stroop researchers have studied the phenomenon in and for itself; relatively few have looked at broader implications of cog- nitive interference. Yet, the examples presented (Broverman, 1960a, 1960b; Jorgensen, 1977), while not allowing complete rejection of this possibility, do per- mit its relegation to least likely status. The scoring system chosen for the present project—— the difference between CW and C in seconds, and the exclusion of clients with over 49 incorrect on any card-- was chosen because of its factor loading of what has been 100 interpreted as the interference factor, and its conceptu— ally logical computation. Many others have been established (c f., Jensen & Rohwer, 1966) and could have been used, but were rejected in favor of CW-C. This could have been an inappropriate choice, in that it does not make full use of all of the scores on the test. Perhaps a scoring system that used all three scores, and the error scores, would have provided a more meaningful and appropriate way of classifying clients. The failure to find an interference main effect on the sorting task could be due to the lack of clear dif- ferentiation between high- and low—interference groups among form-salient clients, which would have had the effect of creating three non-comparable groups, i.e., color— salient/high- and low—interference, and form—salient/nedium- interference, thereby making significant findings impos- sible and uninterpretable, had they existed. The explanation which seems the most reasonable, given the present findings, is that the effects of inter- ference are seen only in interactions with other factors. Post-hoe findings included sex-by-interference level interactions; the association between sorting task per- formance and interference level was strongest in males. There were also possible interactions with age. Jorgensen (1977) got around this by choosing from a narrow age band of college undergraduate females, while Broverman (1960a; 101 1960b) used only males. The expense of such restrictions in sample selection for this study, in terms of testing time, expansion of geographical scope, and transportation costs, would have been prohibitive. It seems, though. that the effects of interference are only seen in the presence of other conditions, such as sex, age, and other factors not investigated in this study, including task complexity and cultural background. While the exact nature of the effects of interference remain unknown, one can speculate about them. The best estimate of them is provided in the post—hoe analyses per- formed in this study. High-interference males performed the sorting task better than did low-interference males, and interference had little association with performance in females. It is possible that high-interference males found the task challenging and, therefore, more engaging, while low-interference males found it too easy and lost interest in it, with a corresponding drop in performance. This would be supported by one of the Broverman studies cited earlier (19603). Broverman found that males were less distractible in tasks that matched their inter- ference cognitive style. Extrapolating to work adjustment, one could conclude that low—interference clients require a certain level of ambiguity. This would lead them to lose interest in any part of work adjustment which was 102 highly structured and offered little opportunity for flexibility or personal interpretation. A second possibility is that there is a complex curvilinear relationship between interference and per- formance, rather than the linear one assumed in this study. Successful performance of the types of tasks displayed here might be associated with a moderate level of interference, with deviations either way decreasing performance. This may relate to the curvilinear rela- tionship found between interference and anxiety (Jorgensen, 1977), and the similar relationship between anxiety and performance (Fein, 1963). Either of these explanations is plausible in the absence of any clear-cut answer in the present findings. It remains for future researchers to decide which of them, if any, provides the best model for the effects of cognitive interference on work adjustment and task performance. Implications for Future Research Those who wish to continue research in the topics of this study may find that certain methodological changes would increase the likelihood of finding relationships between these aspects of cognitive style, and work adjust- ment and task performance. Some of these changes have been implied in the preceding discussion. 103 Certain instrumentation changes could be made with little difficulty. One would be to use a different derived interference score with which to select the sample. There are some derived scores which make more use of all of the individual scores, including the score on W and the error scores. An example would be the one proposed by Thurstone and Mellinger (1953), given by (CW—C)/W; furthermore, the sample could be limited to clients who had even fewer than the 49 errors per card as was the cut- off in this study. Another instrumentation change would be to use a different measure of work adjustment. The most appro— priate measures would break down work adjustment into its parts (e.g., quality and quantity of work, tardiness), and would have criterion—related validity by using actual performance records rather than rating systems. This, of course, assumes that reliable records exist for each aspect of work adjustment one wished to study. A final change would entail finding a different task with which to measure the effects of cognitive style on performance. The sorting task developed for this project may not have been complex enough to "activate” individual differences. It is reasonable to assume that a larger sample size would increase the power to find significant relationships providing the total variance did not change. This could be 104 done by choosing more of the same type of client used in this project. Post—hoc analyses showed that there were differences in performance associated with interference; they just were not large enough. Concurrent with increas- ing the sample size would be to control for possible sex and age differences, by using a single sex within a narrow age band for all groups. Additionally, significant dif- ferences between the groups on interference scores should be sought, thus strengthening one weakness in the present project. This could be accomplished by selecting from extremes, such as the bottom and top 27% of interference scores; though this would, again, assume a linear rela- tionship between interference and performance. The expense involved in imposing the above conditions on future research would be prohibitive for all but the most liberal budgets, and require an almost unlimited population from which to draw. One way to circumvent this would be to break down this study into parts and study them separately. One might choose, for example, only form—salient clients and study interference on them, or study interference independent of salience. The loss in scope incurred in doing this would be offset by the gain in control over undesirable influences. A more attractive, and highly feasible, change would be provided by performing a series of intensive design studies (Anton, 1978) with single, well-chosen subjects. 105 One could, then, provide needed controls over unwanted variance (the variance of one person is zero), while retaining the scope of the present study. Different com- binations of salience, interference levels, task complexity, sex, and age could be studied until a pattern began to emerge. At that time, when the effects of cog— nitive style were more clear—cut, formal hypothesis- testing with comparative group designs could be performed. It appears that cognitive style, especially dimensional salience, is related to task performance, and possibly to work adjustment. It seems appropriate, then, to look at the relationships of other aspects of learning style to that performance. What, for instance, are the effects of mode of instruction—~e.g., group vs. individual instruction, written vs. ora1—-on performance? Is discovery learning possible, or does the client always have to be taught by someone else? What type of clients will benefit from role-play, modeling, chaining, and so on, and for which clients will these methods be inappro- priate? Any number of combinations of independent variables could be studied, given the availability of instrumentation. It may also be fruitful to attempt to generalize cogni— tive interference, which the Stroop is believed to measure, to other sets of stimuli. Separate instruments could be devised to measure shape—word and size-word interference, for example. A stimulus on a shape-word test, for instance, 106 H would be a triangle with the word square” printed inside of it. By testing other types of interference, one could ascertain whether this cognitive interference was task- specific or, in the traditional sense of the word, a trait. That is, if there was no interference on the other tests, one might feel justified in believing that the interference was task—specific and not an indication of one's overall ability to overcome cognitive distraction. Such tests may offer additional diagnostic data, provided that a signifi- cant correlation were found between performance on them and general work adjustment characteristics. No matter which aspect of cognitive style is chosen, however, the complex interactions that can be found with sex, age, or other factors should be kept in mind, and measures to control them, such as the ones mentioned above, should be taken. Conclusions The hypotheses studied in this project concerned the effects of two aspects of cognitive style-—dimensional salience and Stroop interference——on task performance and overall work adjustment. It was thought that differential performance would be associated with the various combina- tions of the two independent variables. With one exception these hypotheses were not supported; that exception con- cerned dimensional salience and multidimensional sorting task performance. It was found that, consistent with past 107 research, color-salient clients had fewer errors when color was solution-relevant than when form was, and that form-salient subjects had fewer errors when shape was relevant. Cautious interpretation of both those results, and the lack of significant findings for the other hypothe— sis, was required given the exploratory nature of the study, and the complex interactions found between the variables and sex and age. Changes in instrumentation and methodology that may provide more clear-cut conclusions were suggested, among which were: (1) a change in the interference score used for selection; (2) different measures of work adjustment; (3) larger samples studied with further restriction in sample selection; (4) comparative group designs with more limited scope; and (5) a series of single-subject intensive designs which would retain the scope desired of the present project, while increasing the control and generalizability of such research. Because it seemed that cognitive style was associated with task performance, other lines of research were sug— gested. These included studying other aspects of cognitive style, such as mode of instruction, and their effects on performance. It was offered that, in all such studies, the possible interactions that cognitive styles might have with sex, age, task complexity, and other factors be kept in mind and controlled for, where possible. Y H m G O I L B I B BIBLIOGRAPHY Ammons, R4 & Ammons, C. The Quick Test: Provisional Manual, Psychological Reports: Monograph Supplement, 1962, I—VII, 111-161. Anastasi, A. Psychological Testing. New York: Macmillan, 1976 Anton, J. Intensive experimental design: A model for the counselor/researcher. Personnel and Guidance Journal, 1978, 5Q, 273-278. Bassett, J. & Schellman, G. Performance of retardates on the Stroop Color-Word Test. Perceptual and Motor Skills, 1976, 43, 1259-1262. Bellamy, G. T., Horner, R. H., & Inman, D. P. Vocational Habilitation of Severely Retarded Adults. Baltimore: University Park Press, 1979. Brian, C4 & Goodenough, F. The relative potency of color and form at various ages. Journal of Experimental Psychology, 1929, 12, l97—213 Broverman, D. Cognitive style and intra-individual variation in abilities. Journal of Personality, 1960, 28, 240—256 (b). Broverman, D. Dimensions of cognitive styles. Journal of Personality, 1960, 28, 167-185 (a). Callaway, E. The influence of amobarbital (amylobarbitone) and methamphetamine on the focus of attention. Journal of Mental Science, 1959, 105, 382- 392. Carlisle, H. L. Quick Test performance by institutional retardates. Psychological Reports, 1965, 11, 489-490. Cattell, J. M. The time it takes to see and name objects. Mind, 1886, 11, 63—65. Coker, C. C. Preparation and safety in relation to the behavioral aftereffects of predictable and unpre- dictable aversive situations. Unpublished doctoral dissertation, Bowling Green State University, 1973. 108 109 Coker, C. C. Stroop Color Word Test. Construction Report, Menomonie, University of Wisconsin—Stout, 1979. Comalli, D., Wapner, S., & Werner, M. Interference effects of Stroop color and word test in children, adulthood, and aging. Journal of Genetic Psychology, 1962, 100, 47—53. Cornfield, J., & Tukey, J. Average values of mean squares in factorials. Annals of Mathematical Statistics, 1956, 22, 907—949. Cronbach, L. J. Coefficient alpha and the internal structure of tests. Psychometrika, 1951, 28, 297—334. Dial, J., & Swearingen, S. The prediction of sheltered workshop performance: special application of the McCarron-Dial-Work Evaluation System. Vocational Evaluation and Work Adjustment Bulletin, 24-33. Dyer, F. The Stroop phenomenon and its use in the study of perceptual, cognitive, and response processes. Memory and Cognition, 1973, 2, 106-120. Dyer, F., & Severance, L. Effects of irrelevant colors in reading of color names: A controlled replication of the ”reversed Stroop” effect. Psychonomic Science, 1972, 28, 336-338. Fein, L. Evidence of curvilinear relationship between IPAT anxiety and achievement at nursing school. Journal of Clinical Psychology, 1963, 28, 374-376. Fields, M. Color, size, and form preferences in infants. Unpublished doctoral dissertation, University of Arkansas, 1973. Gaines, R. Color-form preference and color-form discrimina— tion ability of deaf and hearing children. Perceptual and Motor Skills, 1964, 28, 70. Gaines, R. Children's selective attention to stimuli: Stage or set? Child Development, 1970, 32, 979-991. Gibson, E. Principles of perceptual learning and develop- ment. New York: Appleton-Century—Crofts, 1969. Grossman, H. (Ed.) Manual on terminology and classifica— tion in mental retardation (1973 revision). Washington, D.C.: American Association of Mental Deficiency, 1973. 110 Gumenik, W., & Glass, N. Effects of reducing the read— ability of the words in the Stroop color-word test. Psychonomic Science, 1970, 28, 247-248. Haywood, H. The ethics of doing research . . . and not doing it. American Journal of Mental Deficiency, 1976, 82, 311—317. Jensen, A. R., & Rohwer, W. The Stroop Color—Word Test: A review. Acta Psychologica, 1966, 28, 36-93. Jorgensen, C. Visual set and anxiety in the Stroop phenomenon. Perceptual and Motor Skills, 1977, 33, 659—667. Julius, A. A study of the relationship between color, form, and function preference and sorting flexibility of preschool children. Unpublished doctoral disserta- tion, Columbia University, 1974. Kay, S. Developmental assessment of cognitive style in mentally retarded psychotics. Journal of Clinical Psychology, 1977, 88, 953-958. Kay, 8., & Singh, M. A developmental approach to delineate components of cognitive dysfunction in schizophrenia. British Journal of Social and Clinical Psychology, 1975, 28, 387-399. Marantz, B. Effect of cue salience, cue dimension, and dimensional preference in class inclusion ability in young children. Unpublished doctoral dissertation, New York University, 1976. Marascuilo, L., & McSweeney, M. Nonparametric and distribution-free methods for the social sciences. Monterey: Brooks/Cole, 1977. Marcus, M., West, K., & Gaines, R. Color—form preference in young children of lower socioeconomic status and its relation to cognitive style. Paper presented at Brain Behavior Research Center seminar, Sonoma State Hospital, Eldridge, California, 1968. Melkman, R., Koriat, H., & Pardo, K. Preference for color and form in preschoolers as related to color and form differentiation. Child Development, 1976, 32, 1045—1050. Mischel, W. Toward a cognitive social learning reconcep- tualization of personality. Psychological Review, 1973, 88, 252-283. 111 Nealis, P. Reversal of the Stroop test: Interference in word reading. Perceptual and Motor Skills, 1974, 88, 379-382. Odom, R., & Mumbauer, C. Dimensional salience and identification of the relevant dimension in problem solving: A developmental study. Developmental Psychology, 1971, 8, 135—140. Ogilvie, R. D. Correlations between the Quick Test (QT) and the Wechsler Adult Intelligence Scale (WAIS) as used in a clinical setting. Psychological Reports, 1965, 28, 497-498. Piaget, J. The origins of intelligence in children. New York: International University Press, 1952. Rotatori, A. F. Test-retest reliability of the Quick Test for mentally retarded children. Perceptual and Motor Skills, 1978, 88, 162. Rotter, J. Generalized expectancies for internal versus external control of reinforcement. Ps cholo ical Monographs: General & Applied, 1966, 88, I-EB. Rusalem, H., & Rusalem, H. Learning Capacities Project; In Malekin, D.,and Rusalem, H. (Eds.) Contemporar vocational rehabilitation. New York: New York University Press, 1976. Schiller, P. Developmental study of color-word interference. Journal of Experimental Psychology, 1966, 22, 105—108. Silverstein, A., & Frankin, R. Performance of the mentally retarded on the Stroop Color-Word Test. Perceptual and Motor Skills, 1965, 22, 618. Speilberger, C., & Gorsuch, R. Mediating Processes in Verbal Conditioning. Unpublished report to the National Institutes of Mental Health. Vanderbilt University, 1966. Stroop, J. Factors affecting speed in serial verbal reactions. Psychological Monographs, 1938, 88, 38-48. Stroop, J. Studies of interference in serial verbal reactions. Journal of Experimental Psychology, 1935, 28, 643—662. Suchman, R. G. Cultural differences in children's color and form preferences. Journal of Social Psychology, 1966, 28, 3-10. 112 Suchman, R. G., & Trabasso, T. Color and forms preference in young children. Journal of Experimental Child Psychology, 1966, 8, 177-187. Thurstone, L. L. A factorial study of perception. Chicago: University of Chicago Press, 1944. Thurstone, L. L., & Mellinger, J. J. The Stroo Test. Chapel Hill: University of North Carolina, 1953. Uechi, Y. Cognitive interference and intelligence: Reexamination of the measures of the Stroop Color— Word Test. Japanese Journal of Educational Psychology, 1972, 28, 92-100. U.S. Congress: Ninety-third session. Vocational Rehabilitation Act of 1973; 93-112. Wise, L., Sutton, J., & Gibbons, P. Decrement in Stroop interference time with age. Perceptual and Motor Skills, 1975, 32, 149-150. Wright, B. Social-psychological leads to enhance rehabili— tation effectiveness. Rehabilitation Counseling Bulletin, 1975, 8, 214—223. Zigler, E. National crisis in mental retardation research. American Journal of Mental Deficiency, 1978, 88, 1—8. Zimmerman, R. B., Schroll, E. P., Ackles, P., Barrett, R., & Auster, M. Performance of institutional retardates in standard and new focus of the Quick Test. Perceptual and Motor Skills, 1978, 88, 263-266. APPENDICES APPENDIX A. ADMINISTRATIVE AGREEMENT 113 ADMINISTRATIVE AGREEMENT This document indicates the agreement of the Director of: Name of facility Address of facility to allow research (to be supervised by William Crimando, Researcher, Michigan State University) to be conducted in the aforementioned facility. The following list of responsibilities constitutes an agreement between the above facility and the researcher from Michigan State University to insure the con- tinuity of this study and the protection of the confidentiality of subjects of the study on client cognitive styles. The Director of agrees to: 1. Allow a researcher to meet with selected clients attending the facility during the time period July 15, 1979 through January 15, 1980. 2. Provide access to information contained in the files of said clients. 3. Provide access to information to be supplied by the counselor and/or trainer of said clients. 4. Allow the use of said information for the use of performing quantitative analyses for any and all subsequent research reports. 5. Provide other additional information pertinent to the study as requested. The researcher from Michigan State University agrees to: 1. Take full responsibility for protecting the confidentiality of the indi- viduals who are used in any of the selection, observation, and subsequent data analyses. 2. Secure from each client a consent form for participation in the study. 3. Take full responsibility for protecting the confidentiality of test and survey results obtained on each individual participating in the study. 4. Inform the facility of the availability of the results of this study and provide any necessary assistance in interpretation of the study's findings. These agreements shall be in force until the final report is written. Signed: Date (Director) Date (Researcher, Michigan State University) APPENDIX B. INFORMED CONSENT 114 INFORMED CONSENT AGREEMENT STUDY ON CLIENT COGNITIVE STYLES I, , have had the purposes of this project explained to me. I understand that the general purpose of the procedures to be used in this project is the better understanding of how people learn. I understand that the personal information to be collected during the course of this project is essential to the project and this information is confidential and will not be released to anyone without my express written permission. I give this vocational habilitation facility, , permission to obtain any necessary information from my file and records. In any research report prepared subsequent to this project, I will not be identified by name, and other identifying information will be changed so as to protect my identity. I understand that I can stop participating in the study at any time during the study. This consent agreement terminates March 1, 1980, but the terms of confidentiality are extended indefinitely. Signed Date Witness I certify that I have read this document, or had it read to me, prior to my signing it. Signed APPENDIX C. COMPLETE LIST OF 15 ITEMS OF THE COLOR-FORM PREFERENCE TEST Complete List 115 of Items on Color—Form Preference Test Standard 1. G. Triangle 2. R. Hexagon 3 R. Triangle 4 G. Hexagon 5. B. Triangle 6 R. Rectangle 7 B. Hexagon 8 B. Square 9 G. Circle 10. B. Rectangle 11. G. Rectangle 12. R. Square 13. R. Circle 14. B. Circle 15. G. Square Legend: G Green R Red B Blue Rectangle Rectangle Square . Hexagon . Rectangle . Hexagon . Circle Square Hexagon Triangle Triangle Square Circle . Circle . Triangle Matching Stimuli R. Triangle Hexagon Triangle Square Hexagon Rectangle Square Rectangle . Circle Circle . Circle Hexagon . Rectangle . Triangle Circle Circle Square Hexagon Rectangle Triangle Square Hexagon Circle Triangle . Rectangle . Rectangle Triangle Square . Hexagon Square APPENDIX D. INSTRUCTIONS TO CLIENT AND RESEARCHER FOR COLOR-FORM PREFERENCE TEST 116 CFP INSTRUCTIONS I am going to show you some cards. Each time I do, I would like you to point to the card on the table that looks like the card in your hand. There will be no cards on the table that look exactly like the one in your hand. Just point to the card that you think looks the most like the one in your hand. Are you ready? Let's start with these! Place the three matching cards on the table, side—by—side, directly in front of the client, and then give the client the standard. If you are sitting directly across from the client, your right will be LEFT on the chart below. Circle each response on the CFP Scoresheet. If the client balks at the instructions, try to determine if he/she understood them. You may have to reassure the client that it doesn't matter that no card looks exactly like the standard. Standard EEEI QEEEEE EEEEE 1 ll 3 14 2 6 7 15 3 15 5 2 4 2 8 11 5 10 2 1 6 7 11 12 7 14 12 4 8 15 10 13 9 4 13 5 10 1 l4 6 ll 3 9 10 12 8 4 3 13 9 6 8 14 13 l 7 15 5 9 12 APPENDIX E. COLOR-FORM PREFERENCE TEST SCORESHEET 117 CFP SCORESHEET NAME DATE OF BIRTH DATE For each item, circle the letter in the column(left, center, or right) to which the client points. w Standard EEEE CENTER EEEEI 1 C F N 2 C F N 3 N F C 4 F N C 5 C N F 6 N F C 7 C N F 8 F C N 9 C F N 10 N C F 11 N C F 12 F N C 13 F C N 14 F N C 15 N C F COLUMN TOTAL F total C total N total APPENDIX F. FACSIMILE OF THE WORD, COLOR, AND COLOR-WORD CARDS FOR THE STROOP COLOR—WORD PREFERENCE TEST 00 l l .Hasaflum map mo mHoHoo Hmsuom man unwmmummu mmwmsuGona CH muouuma oLH Boaaow smouw wSHm mwcmuo no com ”m mung; ”3 “pomwmq ume woflmummumucH pHanHoaoo moouum ago so mpumo puozuwoaoo can .HoHoo .pH03 osu mo moHHEHmomm on Amv Ame Adv va Ame on va on Amv pow mmfimuo BoHHo% mafia Cwouw Boaam% pow owcmuo ammuw oDHn .N .Amv Ame va on on va Amv on on Amv wmcwno Boaamk Comuw oSHn non Comuw mwcmuo msafl wow Boaawx .H mm coopw pop mafia BOHHO% woes pm“ Bofifioz mwcmuo BoHHw% :wwuw ownmpo 3oHHw% Bo-w> wmcmuo cowwm masque w5~n comhm mafia mafia sofifimz mafia :wwhm wsfin omcmuo :woum wmcmuo wow compm Lupwfi wo wumm mmcwpo somum owcwho Sofiawz mafia BoHHmz compm omcmpo wDHn 3o-w% moan pop osap coopw cmwpm oSHL Bofiawz no» coopw mmcmpo mzo Hmmmmmmoom mOOMHm pop cmmum mmcwwo mafia cmmhm omauo Bo-oz wow mmcwpo mass BoHHmz new cowpm wmcwpo Bofiawz cmmpw no» mEmc m.ucwflfio «Ens Boaflom moan comma pop mmcmuo 3o-m> pop wmcwpo mafia Bo-m> .cn wean no o:~£ Bo-m> wow nomad omcmpo pop mafia BoHwa mwcmpo compw .cfi M“ may :wwuw mafia wmcmpo Bo-o> was coopm mag: BoHwa wwcmuo.¢ coohw mafia Bofifiox mwcapo twp omcmpo .cowum new Bofifimz moan .w :oouw wmcwuo osfln Bo-o> vow mDHQ new mmcwuo meum BoHHoz .n omcwpo omen 30-m> vow cowuw no» mafia mmcwuo cmouw BoHwa .0 won nowhw meapo moan Sofiaww no» smwum Bofifioz mmcmpo mafia .m wmcquo Bofipwz pop cmmuw mafia wow omcmuo mafia Bo-mx comuw .q omcwuo sofifiwz cowum no» oDHQ owcwpo mop compw mafia Bofifiwx .m owCQHo mafia no» :mopm onHo> mafia :ownw Bo-o> wmcmpo non .N mafia wok 3oflfiwx :wouw wmcmuo Bo-o> mafia cmwuw omcwuo top .H pump :uuwn we mama mews m.u:oHHo mmm:H\oze Ewarmmmoom mOOMHm APPENDIX H. INSTRUCTIONS TO CLIENT FOR STROOP COLOR-WORD INTERFERENCE TEST 121 STROOP INSTRUCTIONS Card One Here is a list of the color words blue green, red, yellow, and orange. When I tell you to begin, I would like you to start here(point), read each word in the row out loud, and then go on to the next row. Keep reading until you get to the end of the list(point). Read as quickly as you can without getting mixed up. Point to the words to keep your place. Ready? Begin! Card Two Here is a card with the colors blue, red, green, yellow, and orange. When I tell you to begin, I would like you to start here(point), name each color in the row, and then go on just like you did on the last card. Keep naming the colors until you get to the end of the list. Name the colors as fast as you can without getting mixed up. Point to them to keep your place. Ready? Begin! Card Three Here is a card with some more of the same colors. When I tell you to begin, I would like you to look at each word and tell me the color that is there. Don't tell me the word, just name the color. Start here and go on until you're finished(point). Name the colors as fast as you can without getting mixed up. (At this point, have client try out the first line to see if the directions were understood. Repeat first line until you are sure, or until the client has attempted the line four times. If after the client has tried four timess/he is still getting mixed up, make a note of it and go on as if the client has understood) Ready? Begin! APPENDIX I. COUNSELOR (TRAINER) QUESTIONNAIRE 122 COUNSELOR(TRAINER) QUESTIONNAIRE Please complete the following form for each client chosen for the study. You may look in the records of the client should that be necessary to answer the questions. Answer each question to the best of your ability. Counselor's name Facility Client's name Sex Date of birth A. Please rate the client on each of the below-listed points using the following rating system: l...better than would be expected from a competitively employed worker 2...equal to that of a competitively employed worker 3...s1ight1y below that of a competitively employed person 4...a severe deficiency in this characteristic exists in the client CIRCLE ONE 1. quantity of work...............................................1 2 3 4 2. quality of work......................................... ....... l 2 3 4 3. relations or cooperation with supervisors......................l 2 3 4 4. relations or cooperation with co-workers.......................1 2 3 4 5. performance of repetitive tasks............ ........ ...... ..... .1 2 3 4 6. performance of tasks requiring variety and change in methods...l 2 3 4 7. response to training...................... ...... ... ..... .......1 2 3 4 8. punctuality for all programming................................l 2 3 4 9. daily attendance...............................................1 2 3 4 10. resistance to distraction while w0rking........................1 2 3 4 ll. frequency of emotional outbursts...............................1 2 3 4 B. To which of the below-listed distractions is the client vulnerable while working? CHECK AS MANY AS APPLY people's voices noise other than voices movement or activity within field of vision change in lighting change in temperature hunger thirst distractions of an indeterminate nature other sourcesCPLEASE LIST) C. What recommendation would you make about this client? (CHECK ONE) 1. He/she is ready for competitive employment........... ..... . ................ 2. He/she will be ready for competitive employment within a year............. 3. He/she is ready for sheltered employment. . ................ . .......... ... 4. He/she will require work activities for an extended period of time. ........ 5. I cannot make a decision at this time. ..................... ...... ....... D. [‘11 123 Counselor(trainer) questionnaire p. 2 Consider this client's overall competence, either in reference to the above factors, or to factors not mentioned(e.g., responsibility, initiative, etc.). How well does the client compare to other clients in the program? (CHECK ONE) In the top %............................ In the top % but not in the top %....... In the bottom % but not in the bottom % In the bottom %......................... If this information is available in the client's files, please complete the following: Full Scale IQ or Mental Age yrs. mos. Date tested Name of test Social Age yrs. * mos. Date tested Name of test APPENDIX J. SUMMARY OF TRAINING PROCEDURE: INSTRUCTIONS TO RESEARCH ASSISTANTS 124 Summary of Training Procedures Instructions to Research Assistants When you arrive at the testing area, do the following: a. Clear a surface for testing and accessible seating arrangements for yourself and the client. You should be able to sit across from the client. b. Check for all materials, including: Pretest: 5 color strips (red, blue, green, yellow, orange) 5 word strips (red, blue, green, yellow, orange) 1 dime Color-Form Test: 15 cards instruction sheet Quick Test: 3 forms instruction card Stroop Test: book with three cards instruction card List of names of clients you are to see One scoresheet for each test and one Informed Consent per client One stopwatch c. Check the Stroop cards for glare. If there is a glare, prop them up until it disappears. d. Call your first client. Introduce yourself to client as follows: Hi, (name)! I'm (name) from Michigan State University, and I'd like to see you for a few minutes to talk about a project I'm work- ing on. Read the statement regarding the purpose of the project, to be found at the beginning of the Informed Consent. When you have finished, elicit the client's questions and participa— tion. If participation is desired, elicit the client's full name and birthdate, writing those in the appropriate spaces. Read the section on Informed Consent (with the exception of the witness qualifier) and both of you sign the form. Administer the short screening test. 125 Administer the remaining tests to the clients who pass the screening test in the following order: Color-Form Preference Test Quick Test Stroop Color-Word Interference Test Record all responses on the score sheets. Any clients wishing not to participate, or failing to com— plete any test, should be thanked and excused. Those who are able to complete the test should be thanked and told that someone will contact them about participation in the next part of the project. Return all materials at the end of the day. Extra scoresheets are available if you need them. APPENDIX K. PURPOSE OF STUDY 126 STATEMENT OF PURPOSE OF STUDY ON CLIENT COGNITIVE STYLES We would like you to take part in a study being run by researchers at Michigan State University. The study will help us better understand how people like your- self learn. If you decide to take part in the study, you will be given a number of tasks today and again in a couple of weeks. Your performance on these tasks will be kept private and won't be shared with anyone unless you ask us to give them to someone else. We hope that this study will help us know more about how people learn. We believe that if we know how people learn, we can come up with ways to help them to learn easier. Your participation in this study is important because of some of your experiences here at . Other people here and at workshops (name of facility) all over Michigan are taking part in this study. We hope that by combining all the information from everybody, including you, we will be able to make learning hov to do a job easier for people in the future. APPENDIX L. COMBINED PURPOSE AND INFORMED CONSENT FOR USE AT ONE SITE STATEMENT OF PURPOSE AND INFORMED CONSENT STUDY ON CLIENT BEAENING STTIES we are looking at how people like you learn Jobs. All people want to learn how to do a job. It's harder for some people. But if we can change our ways of teaching jobs, people might learn faster. We are going to work with people here at . If we find out how they learn, we might be able to come up with ways to help them. we would like you to help us. First, we need you to sign this form. I will tell you what will go on if you work with us. We have some tasks that show how people learn. You will do some of them. It will take from one to three hours to do them. You will not have to work for more than one hour a day. Someone will be with you and will write down how you do. This will be a secret. No one will know how you did unless you want us to tell them. we will also ask your supervisor here to fill out a form. It will tell us what kind of a worker you are, and some things from your file here. We need this so we can find out how people who learn in different ways do at different times. This will also be a secret, but if you want to see it you can. All you have to do is ask. We will keep everything locked up so that no one else can see it. You can stop helping us anytime you want to. If you want to, just tell me. What- ever you decide will be okay. It will not change what goes on here, but you might lose some work while you are helping us. I, , want to help out on this project: ‘birthdate 1. I know I can find out more about this project anytime I want to. 2. I can stop anytime. All I have to do is ask. 3. The things that are learned about me will be a secret. If they are used in a report after the project, no one will know it was me. 12‘ . I know I might lose some work while I help out, but only for a few hours. 5. I want to volunteer, so that I, or people like me, might be helped to learn better in the future. Client's signature Date _____ Witness Legal guardian Date Witness is responsible to assure that if client signs, he was competent to give informed consent (3330.7003) (R330.6013(S) (a) — (c) Michigan Department of Mental Health Emergen- cy Rules or if guardian signed, documentation is on file indicating that the court has empowered the guardian with the authority. If the witness does not feel the client is competent, refer to R330.6011(3) — (h). '_- This consent agreement terminates march 1, 1980, but the terms of confidentiality are extended indefinitely. 127